**Meet the editor**

Dr Dong-Hyun Kim is an associate professor of the Department of Electrical and Electronic Engineering at Yonsei University located in Seoul, Korea. He received his Bachelor's degree from Yonsei University. He obtained his M.S. and Ph.D degree from Stanford University in the Department of Electrical Engineering specializing in 'Technical developments for 1H in-vivo Magnetic Resonance Spectroscopic Imaging". After completion of his training, he joined the faculty of Radiology at the University of California, San Francisco where he continued his research on Magnetic Resonance Spectroscopy (MRS) and Magnetic Resonance Spectroscopic Imaging (MRSI). Since 2006, he has been teaching and conducting research at Yonsei University. He has worked in the field of Magnetic Resonance Imaging (MRI) and MRS/MRSI for over ten years and has published numerous peer reviewed articles on the topic.

Contents

**Preface IX** 

**Part 1 MRS Inside the Clinic 1** 

Bjørnar Hassel

Fahmy Aboul-Enein

Beata Tarnacka

Chapter 4 **Wilson's Disease in Brain Magnetic Resonance Spectroscopy 73** 

> **Disease Progression 93**  Münire Kılınç Toprak, Banu Çakir, E.Meltem Kayahan Ulu and Zübeyde Arat

Chapter 1 **Quantification Improvements of 1**

Chapter 2 **13C Magnetic Resonance Spectroscopy in** 

**Neurobiology - Its Use in Monitoring Brain Energy Metabolism and in Identifying Novel Metabolic Substrates and Metabolic Pathways 29** 

Chapter 3 **MR Spectroscopy in Multiple Sclerosis - A New Piece of the Puzzle or Just a New Puzzle 47** 

Chapter 5 **MRS in MS, With and Without Interferon Beta 1a** 

Chapter 6 **Magnetic Resonance Spectroscopy (MRS) in Kidney** 

**Obstructive Pulmonary Disease 123** 

**Treatment, to Define the Dynamic Changes of Metabolites in the Brain, and to Monitor** 

**Transplantation: Interest and Perspectives 105**  Bon Delphine, Seguin François and Hauet Thierry

Chapter 7 **Acute Effects of Branched-Chain Amino Acid Ingestion on Muscle pH during Exercise in Patients with Chronic** 

Tomoko Kutsuzawa, Daisaku Kurita and Munetaka Haida

**H MRS Signals 3** 

Maria I. Osorio-Garcia, Anca R. Croitor Sava, Diana M. Sima, Flemming U. Nielsen, Uwe Himmelreich and Sabine Van Huffel

### Contents

#### **Preface XI**

#### **Part 1 MRS Inside the Clinic 1**


#### X Contents

#### **Part 2 MRS Beyond the Clinic 141**

Chapter 8 **NMR Spectroscopy as a Tool to Provide Mechanistic Clues About Protein Function and Disease Pathogenesis 143**  Benjamin Bourgeois, Howard J. Worman and Sophie Zinn-Justin

#### Chapter 9 **Structural and Vibrational Properties and NMR Characterization of (2'-furyl)-Imidazole Compounds 167**  Ana E. Ledesma, Juan Zinczuk, Juan J. López González and Silvia A. Brandán


### Preface

Magnetic Resonance Spectroscopy (MRS) is a unique tool to probe the biochemistry in vivo providing metabolic information non-invasively. Applications using MRS has been found over a broad spectrum in investigating the underlying structures of compounds as well as in determining disease states. In this book, topics of MRS both relevant to the clinic and also those that are beyond the clinical arena are covered. The book consists of two sections. The first section is entitled 'MRS inside the clinic' and is focused on clinical applications of MRS. For clinical routine usage of MRS, accurate quantification methods are necessary. The first chapter starts with the various quantification methods used in 1H MRS. This is followed by an in depth coverage of MRS used in different clinical settings. Diseases that are covered include MRS for Multiple Sclerosis, Wilson's disease, kidney transplantation, chronic obstructive pulmonary disease. In addition, 13C MRS is also an important application in studying the energy metabolism in neurobiology. This is also covered in the book. In the second section of the book entitled 'MRS beyond the clinic', topics related either directly or indirectly to the clinic settings are discussed. These include a variety of NMR applications to probe chemical structures further using MRS. Our hope is that through this book, readers can understand the broad applications that NMR and MRS can offer and also that there are enough references to guide the readers for further study in this important topic.

> **Donghyun Kim**  School of Electrical and Electronic Engineering, Yonsei University, Korea

**Part 1** 

**MRS Inside the Clinic** 

**Part 1** 

**MRS Inside the Clinic** 

**1. Introduction**

metabolite concentrations.

the relevant information for diagnosis purposes.

**Quantification Improvements of <sup>1</sup>**

*In vivo* 1H Magnetic Resonance Spectroscopy (MRS) and Magnetic Resonance Spectroscopic Imaging (MRSI) signals contain relevant metabolic information commonly used as biomarkers of diseases that can provide complementary information for the diagnosis. These signals are measured in the time domain, however, their representation in the frequency domain provides a better visualization of the metabolites in the form of resonances (peaks). In 1H MRS(I), not only metabolites but also water and macromolecule/lipid resonances can be observed. In practice, the concentrations of metabolites and the presence of macromolecule/lipids contain

Maria I. Osorio-Garcia1, Anca R. Croitor Sava1, Diana M. Sima1, Flemming

U. Nielsen2, Uwe Himmelreich2 and Sabine Van Huffel1

*IBBT - K.U. Leuven Future Health Department* 

*1Dept. Electrical Engineering, ESAT-SCD, Katholieke Universiteit Leuven* 

*2Biomedical Nuclear Magnetic Resonance Unit, Katholieke Universiteit Leuven* 

**1**

*Belgium* 

**H MRS Signals** 

The technique used to determine metabolic contributions is called quantification and different methods have been proposed in the time or the frequency domain. In these methods, the metabolite concentrations are estimated, *e.g.*, using a linear combination of individual peaks or a basis set of known metabolite profiles to fit the MR signal (Poullet et al., 2007; Provencher, 1993; 2001; Ratiney et al., 2004; Slotboom et al., 1998). No matter which quantification method is used, the quality of MRS signals is important for obtaining accurate estimates of the

Although *in vivo* MRS(I) acquisitions provide non-invasively metabolic information, different limitations related to the specifications of the acquisition protocol, the localization of the voxels of interest and the homogeneity of the magnetic field in the selected region make the accurate quantification of these signals still a challenge. Therefore, advanced and efficient measurements have been developed to decrease the scanning time and to increase the spectral resolution. Nevertheless, it is essential to perform a series of preprocessing steps to improve spectral quality of MRS signals before performing quantification. Successful combinations between advanced acquisition techniques and appropriate signal processing increase the potential of integrating MR techniques in the clinical diagnosis routines (Chu et al., 2003;

Devos et al., 2004; Huang et al., 2001; Ruiz-Pena et al., 2004; Vermathen et al., 2003).

In the next sections, we introduce the MRS(I) signals together with some of the most common preprocessing and quantification methods. In particular, we focus on recent methods and approaches to improve the quantification of *in vivo* MRS(I) signals, performed with

#### **Quantification Improvements of <sup>1</sup> H MRS Signals**

Maria I. Osorio-Garcia1, Anca R. Croitor Sava1, Diana M. Sima1, Flemming U. Nielsen2, Uwe Himmelreich2 and Sabine Van Huffel1 *1Dept. Electrical Engineering, ESAT-SCD, Katholieke Universiteit Leuven IBBT - K.U. Leuven Future Health Department 2Biomedical Nuclear Magnetic Resonance Unit, Katholieke Universiteit Leuven Belgium* 

#### **1. Introduction**

*In vivo* 1H Magnetic Resonance Spectroscopy (MRS) and Magnetic Resonance Spectroscopic Imaging (MRSI) signals contain relevant metabolic information commonly used as biomarkers of diseases that can provide complementary information for the diagnosis. These signals are measured in the time domain, however, their representation in the frequency domain provides a better visualization of the metabolites in the form of resonances (peaks). In 1H MRS(I), not only metabolites but also water and macromolecule/lipid resonances can be observed. In practice, the concentrations of metabolites and the presence of macromolecule/lipids contain the relevant information for diagnosis purposes.

The technique used to determine metabolic contributions is called quantification and different methods have been proposed in the time or the frequency domain. In these methods, the metabolite concentrations are estimated, *e.g.*, using a linear combination of individual peaks or a basis set of known metabolite profiles to fit the MR signal (Poullet et al., 2007; Provencher, 1993; 2001; Ratiney et al., 2004; Slotboom et al., 1998). No matter which quantification method is used, the quality of MRS signals is important for obtaining accurate estimates of the metabolite concentrations.

Although *in vivo* MRS(I) acquisitions provide non-invasively metabolic information, different limitations related to the specifications of the acquisition protocol, the localization of the voxels of interest and the homogeneity of the magnetic field in the selected region make the accurate quantification of these signals still a challenge. Therefore, advanced and efficient measurements have been developed to decrease the scanning time and to increase the spectral resolution. Nevertheless, it is essential to perform a series of preprocessing steps to improve spectral quality of MRS signals before performing quantification. Successful combinations between advanced acquisition techniques and appropriate signal processing increase the potential of integrating MR techniques in the clinical diagnosis routines (Chu et al., 2003; Devos et al., 2004; Huang et al., 2001; Ruiz-Pena et al., 2004; Vermathen et al., 2003).

In the next sections, we introduce the MRS(I) signals together with some of the most common preprocessing and quantification methods. In particular, we focus on recent methods and approaches to improve the quantification of *in vivo* MRS(I) signals, performed with

(*a*)

Cr

m−Ins

Cho

H MRS Signals 5

Cr

1.5 T

Lip

Fig. 1. (*a*) T2-weighted MRI of a brain tumor patient overlaid with a grid of voxels corresponding to an MRSI measurement. (*b*) Real part of the spectrum of a single voxel *in vivo* MRS signal acquired with a 1.5 T Philips NT Gyroscan (Philips Medical Systems, Best, The Netherlands) with acquisition parameters: PRESS sequence, repetition time (TR)=6 s, echo time (TE)=23 ms, bandwidth (SW)=1 KHz, number of points (NDP)=512 points and 64 averages. (*c*) Real part of the spectrum of a signal from a selected multi-voxel (MRS(I)) measurement acquired with a 3.0 T Philips Achieva (Philips Medical Systems, Best, The Netherlands) with acquisition parameters: PRESS sequence, TR=2 s, TE=35 ms, SW=2 KHz and NDP=2048 points and 1 average. These measurements were done at the University

(*b*) (*c*)

phase, the time delay between the excitation and detection, the flip-angle variation across the spectrum, and the phase shifts from the filter employed to reduce noise outside the spectral bandwidth, produce different phase distortions (Chen et al., 2002). As a solution, phase correction approaches have been successfully used to improve the visualization of MR spectra. In practice, the phase correction consists of two components, one frequency-dependent (first order phase) and one frequency-independent (zero order phase). In particular, the zero-order phase correction consists of the multiplication of the complex spectrum by a complex phase factor equal to the initial phase of the FID (Jiru, 2008). On the other hand, adjusting the first order phase corresponds to the modification of the begin time of the MR signal (Chen et al., 2002). Although some quantification

4.5 4 3.5 3 2.5 2 1.5 1 0.5 ppm

Gln,Glu

NAA

3.0 T

Lip

Lac,Ala,Lip

Lac,Ala,Lip

4.5 4 3.5 3 2.5 2 1.5 1 0.5 ppm

Hospital of the Katholieke Universiteit Leuven, Belgium.

Gln,Glu

NAA

Cr

m−InsCho

Quantification Improvements of 1

Cr

the quantification method AQSES (Poullet et al., 2007): *a*) filtering of residual water, *b*) lineshape estimation, *c*) baseline estimation and *d*) inclusion of spatial constraints in MRSI quantification.

#### **2. Description of 1H MRS(I) signals and preprocessing methods**

#### **2.1 1H MRS(I) signals**

These signals are measured in the time domain and have ideally a decaying shape, also called FID, which can be represented as a sum of complex-damped exponentials. They can be measured from a specific anatomical region (single voxel MRS), or from an entire organ overlaid by a grid of multiple voxels (MRSI). In order to observe the contribution of individual metabolites, MRS(I) signals are transformed to the frequency domain using the Fast Fourier Transform (FFT), producing a spectrum where the metabolite resonances can be visualized. Typically, a water suppression technique is applied as part of the acquisition protocol in order to enhance the visualization of the metabolites of interest. Additionally, an unsuppressed water signal is always measured, which is commonly used as a reference for phase, eddy currents and lineshape corrections. Depending on the acquisition protocol, single voxel signals have commonly a better signal-to-noise (SNR) ratio and thinner linewidths than multi-voxel signals, however, the advantage of multi-voxel signals is the possibility to study complete anatomical organs using one acquisition. See Fig. 1 for an example of a single and multi-voxel acquisition.

#### **2.2 Preprocessing 1H MRS(I) signals**

Preprocessing of MRS signals aims at the improvement of signal quality in order to accurately extract relevant information about the metabolites via the technique called quantification. Some essential preprocessing procedures in 1H MRS(I) are:


the quantification method AQSES (Poullet et al., 2007): *a*) filtering of residual water, *b*) lineshape estimation, *c*) baseline estimation and *d*) inclusion of spatial constraints in MRSI

These signals are measured in the time domain and have ideally a decaying shape, also called FID, which can be represented as a sum of complex-damped exponentials. They can be measured from a specific anatomical region (single voxel MRS), or from an entire organ overlaid by a grid of multiple voxels (MRSI). In order to observe the contribution of individual metabolites, MRS(I) signals are transformed to the frequency domain using the Fast Fourier Transform (FFT), producing a spectrum where the metabolite resonances can be visualized. Typically, a water suppression technique is applied as part of the acquisition protocol in order to enhance the visualization of the metabolites of interest. Additionally, an unsuppressed water signal is always measured, which is commonly used as a reference for phase, eddy currents and lineshape corrections. Depending on the acquisition protocol, single voxel signals have commonly a better signal-to-noise (SNR) ratio and thinner linewidths than multi-voxel signals, however, the advantage of multi-voxel signals is the possibility to study complete anatomical organs using one acquisition. See Fig. 1 for an example of a single and

Preprocessing of MRS signals aims at the improvement of signal quality in order to accurately extract relevant information about the metabolites via the technique called quantification.

• **Time circular shift**. The digital filter and decimation in the Bruker scanners distort the beginning of the FIDs, as a consequence, the MR spectrum is filled with wiggles and the metabolite resonances are hard to identify. These first data points are initially zero and then increase, until the actual FID starts (Cobas & Sardina, 2003). A solution to this distortion is a time circular shift. This step is performed in the time domain and consists of removing

• **Frequency alignment**. Due to variations in the physiological and experimental conditions (*e.g.*, temperature and pH), MR resonances experiment a frequency shift. Specially in multivariate analysis, this peak misalignment needs to be corrected. Thus, the FIDs are first transformed to the frequency domain and the spectra are shifted such that some recognizable peaks reach the desired frequency locations (Veselkov et al., 2009). In 1H MRS, the resonance frequency of known metabolites can be used for shifting the full spectra. For instance, the peak of N-Acetyl-Aspartate (NAA) is known to be located at 2.01 ppm. Moreover, the displacement of individual spectral peaks from one metabolite can be corrected using advanced frequency alignment techniques, such as, quantum mechanics approaches and advanced warping algorithms (Giskeødegård et al.,

• **Phase correction**. Ideally, MRS spectra should have zero-phase (*i.e.*, all peaks are pointing upwards), however, differences between the reference phase and the receiver detector

data points from the beginning and adding them to the end of the signal.

**2. Description of 1H MRS(I) signals and preprocessing methods**

quantification.

**2.1 1H MRS(I) signals**

multi-voxel acquisition.

**2.2 Preprocessing 1H MRS(I) signals**

2010; Lazariev et al., 2011).

Some essential preprocessing procedures in 1H MRS(I) are:

Fig. 1. (*a*) T2-weighted MRI of a brain tumor patient overlaid with a grid of voxels corresponding to an MRSI measurement. (*b*) Real part of the spectrum of a single voxel *in vivo* MRS signal acquired with a 1.5 T Philips NT Gyroscan (Philips Medical Systems, Best, The Netherlands) with acquisition parameters: PRESS sequence, repetition time (TR)=6 s, echo time (TE)=23 ms, bandwidth (SW)=1 KHz, number of points (NDP)=512 points and 64 averages. (*c*) Real part of the spectrum of a signal from a selected multi-voxel (MRS(I)) measurement acquired with a 3.0 T Philips Achieva (Philips Medical Systems, Best, The Netherlands) with acquisition parameters: PRESS sequence, TR=2 s, TE=35 ms, SW=2 KHz and NDP=2048 points and 1 average. These measurements were done at the University Hospital of the Katholieke Universiteit Leuven, Belgium.

phase, the time delay between the excitation and detection, the flip-angle variation across the spectrum, and the phase shifts from the filter employed to reduce noise outside the spectral bandwidth, produce different phase distortions (Chen et al., 2002). As a solution, phase correction approaches have been successfully used to improve the visualization of MR spectra. In practice, the phase correction consists of two components, one frequency-dependent (first order phase) and one frequency-independent (zero order phase). In particular, the zero-order phase correction consists of the multiplication of the complex spectrum by a complex phase factor equal to the initial phase of the FID (Jiru, 2008). On the other hand, adjusting the first order phase corresponds to the modification of the begin time of the MR signal (Chen et al., 2002). Although some quantification

the MR signals as a sum of complex-damped exponentials. Extensions of this method for quantification and peak suppression have been successfully applied to MR signals (Chen et al., 2004; Laudadio et al., 2004). Alternative to HLSVD, other filtering approaches have also been used for this purpose (Antoine et al., 2000; Sundin et al., 1999). Additional to the residual water, other unwanted resonances (*e.g.*, reference peaks from *in vitro* phantom solutions) may also need to be suppressed from the spectrum (Cabanes et al.,

H MRS Signals 7

Several quantification methods have been developed in the time and the frequency domain to determine metabolite concentrations. Thus, depending on the nuclei, the complexity of the signal and the *prior* information available, MR signals can be quantified with the methods listed below. An extended review of time- and frequency domain methods has been given in (Mierisová & Ala-Korpela, 2001; Vanhamme et al., 2001) and more recently in (Poullet et al.,

• Time- and frequency domain fitting using a linear combination of individual peaks/profiles to fit the spectra QUEST (Ratiney et al., 2004), AQSES (Poullet et al., 2007)

• Time-domain estimation of parameters using *prior knowledge* (Soher et al., 1998;

• Time-domain non-iterative fitting methods such as HLSVD (Barkhuijsen et al., 1987; Chen et al., 1996; Dologlou et al., 1998; Laudadio et al., 2002; Pijnappel et al., 1992;

• Time-domain variable projection (VARPRO) (Cavassila et al., 1999; van der Veen et al.,

• Time domain fitting of one peak at a time and wavelet modeling for the baseline

• Principal Component Analysis (PCA), Independent Component Analysis (ICA)

• Constrained least squares (TARQUIN) (Reynolds et al., 2006; Wilson et al., 2011)

2001).

2008).

1988)

**3. Quantification**

Quantification Improvements of 1

LCModel (Provencher, 1993; 2001)

• Semi-parametric fitting (Elster et al., 2005)

(Dong et al., 2006; Romano et al., 2002)

• Genetic algorithms (Metzger et al., 1996) • Fast Padé Transform (Belki´c & Belki´c, 2006)

• Sparse representation (Guo et al., 2010)

(Hao et al., 2009; Stoyanova & Brown, 2001)

• Circular fitting (Gabr et al., 2006)

van den Boogaart, 1997)

Young et al., 1998), AMARES (Vanhamme et al., 1997)

• Iterative time- and frequency domain fitting (Slotboom et al., 1998)

• Artificial Neural Networks (Bhat et al., 2006; Hiltunen et al., 2002)

methods (Poullet et al., 2007; Provencher, 2001; Ratiney et al., 2005) are able to take into consideration phase distortions and provide accurate metabolite estimates even if the spectra are not zero-phased, other quantification methods, such as peak integration, require zero-phased spectra in order to obtain reliable metabolite estimates.


the MR signals as a sum of complex-damped exponentials. Extensions of this method for quantification and peak suppression have been successfully applied to MR signals (Chen et al., 2004; Laudadio et al., 2004). Alternative to HLSVD, other filtering approaches have also been used for this purpose (Antoine et al., 2000; Sundin et al., 1999). Additional to the residual water, other unwanted resonances (*e.g.*, reference peaks from *in vitro* phantom solutions) may also need to be suppressed from the spectrum (Cabanes et al., 2001).

#### **3. Quantification**

4 Magnetic Resonance Spectroscopy

• **Eddy current corrections**. These currents are induced by the rapid switching of the magnetic field gradient in the magnet coils and surrounding metal structures (de Graaf, 1998). Because these currents are caused by the scanner and can not always be avoided, a spectral correction is necessary. Klose (Klose, 1990) has developed a method to correct eddy currents, called ECC, by point-wise dividing the water suppressed signal by the

• **Lineshape correction/estimation**. The lineshape of MR signals is determined by the decay function (damping) of the time domain signal and refers to the shape of the peaks in the frequency domain. Ideally, it is represented by a Lorentzian, Gaussian or Voigt function, corresponding in the time domain to an exponentially decaying sinusoid. However, when magnetic field perturbations and field inhomogeneities are present, these lineshapes are disturbed (in symmetry and linewidth) and as a consequence, the ideal model used in the quantification method is unable to yield accurate metabolite estimates. Therefore, different hardware and mathematical approaches have been developed to cope with this problem. For instance, shimming techniques (*i.e.*, coils adjustment) are applied during the MR acquisition in order to correct field inhomogeneities and thus improve spectral quality (Blamire et al., 1996; Gruetter, 1993; Juchem et al., 2010). Moreover, local magnetic field susceptibility problems caused by measurements near an air cavity such as sinuses, may not always be corrected with shimming techniques making lineshape distortions unavoidable. Notwithstanding, some preprocessing methods have been successfully used based on the deconvolution of the spectra using either the unsuppressed water signal or a reference peak from the experimental signal (Bartha et al., 2000; de Graaf et al., 1990; Metz et al., 2000). Alternatively, other lineshape estimation methods based on self-deconvolution have been proposed in cases when a reference signal is not available

• **Baseline correction/estimation**. 1H MRS signals measured at short TE contain not only information from metabolites but also from lipids and macromolecules which may affect the baseline of the spectra. This baseline influences the quantification causing the mis-fitting of individual peaks or metabolite profiles. Therefore, different correction algorithms have been proposed to estimate and remove the baseline contribution from the spectrum (Chang et al., 2007; Cobas et al., 2006; Xi & Rocke, 2008). Alternatively, other methods have been proposed to estimate the baseline using special acquisition protocols or characterizing each macromolecule/lipid component to finally consider them in the quantification procedure (Behar & Ogino, 1993; Hofmann et al., 2001; Knight-Scott, 1999). • **Filtering of residual water**. 1H MRS signals contain a water peak thousands of units larger than those of metabolites. Therefore, in order to visualize metabolites, this water peak needs to be suppressed. Although the water resonance is partially suppressed during acquisition, the residual water requires a reliable filter for further elimination without affecting the metabolite resonances. Moreover, especially at short TE, the tail of this residual water may overlap with the baseline of macromolecule/lipids and thus, pose problems for accurate quantification. As a solution, the Hankel Lanczos Singular Value Decomposition (HLSVD) method proposed by (Pijnappel et al., 1992) parametrizes

require zero-phased spectra in order to obtain reliable metabolite estimates.

phase term of the water unsuppressed signal measured at the same location.

(Maudsley, 1995; Popa et al., 2011; Sima et al., 2009).

methods (Poullet et al., 2007; Provencher, 2001; Ratiney et al., 2005) are able to take into consideration phase distortions and provide accurate metabolite estimates even if the spectra are not zero-phased, other quantification methods, such as peak integration,

> Several quantification methods have been developed in the time and the frequency domain to determine metabolite concentrations. Thus, depending on the nuclei, the complexity of the signal and the *prior* information available, MR signals can be quantified with the methods listed below. An extended review of time- and frequency domain methods has been given in (Mierisová & Ala-Korpela, 2001; Vanhamme et al., 2001) and more recently in (Poullet et al., 2008).


5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0

individual corrected components

5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 ppm

filtered original signal (blue) − filtered estimated signal

<sup>0</sup> <sup>1</sup> <sup>2</sup> <sup>3</sup> <sup>4</sup> <sup>5</sup>

Fig. 2. Quantification results window for AQSES in SPID (left side) of an *in vivo* 1H MRS signal from mouse brain acquired with a 9.4 T Bruker Biospec small animal MR scanner (Bruker BioSpin MRI, Ettlingen, Germany) with acquisition parameters: PRESS sequence, TR=4 s, TE=12 ms, SW=4 KHz, NDP=2048 points and 256 averages, volume of interest (VOI)= 3 <sup>×</sup> 1.75 <sup>×</sup> 1.75 *mm*3. The basis set of 16 metabolites used for quantification was measured *in vitro*. Right: plots of residue, splines baseline, estimated metabolites and

As described in section 2.2, the water component (*w*(*t*) from Eq.(1)) can be filtered using HLSVD, however, appropriate filtering of this resonance in the case of distorted signals is often inaccurate. When HLSVD with one component (*i.e.*, a model order *K* = 1) is used to fit one non-Lorentzian peak, the result obtained by subtracting that modeled peak from the original signal produces a non-flat residual. In an ideal case, only white noise should be left

The state-of-the-art methods proposed for determination of an optimal model order *K* in order to provide a reliable water peak removal have been described and studied on simulated, noiseless or non-overlapping peaks (Cabanes et al., 2001; Hu et al., 2010; Lin et al., 1997; Papy et al., 2007; Van Huffel & Vandewalle, 1991). For instance, (Cabanes et al., 2001) presented a method for optimal residual water removal of *in vivo* 1H MRS signals of

original with estimated signal.

Quantification Improvements of 1

**3.2.1 Filtering of residual water**

in the residual.

baseline

5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0

residuals (filtered original signal − filtered estimated signal − baseline)

H MRS Signals 9

#### **3.1 Automated Quantification of Short echo time MRS signals (AQSES)**

We present in more details the method AQSES, which will be used in the next sections to illustrate several recent improvements in the field of MRS(I) signal processing. This time-domain quantification method has been especially developed for short TE MRS signals, where a mathematical model fits to a basis set of predefined metabolite profiles. From this fit, metabolite amplitudes are obtained, which represent the weighting coefficients of a linear combination of corrected metabolite profiles with Lorentzian lineshapes used for quantification. Finally, these values are proportional to the concentrations of all estimated metabolites. AQSES is available in the Java open source software AQSES GUI (De Neuter et al., 2007; Poullet et al., 2007), as a quantification method inside the Matlab® graphical user interface SPID (Poullet, 2008) and as a plug-in in the jMRUI software package (version 4.1) (Stefan et al., 2009). Fig. 2 shows the results window for a quantification made in SPID.

The model describing a short TE *in vivo* time-domain MRS signal *y*(*t*) as a combination of several metabolite profiles is:

$$y(t) = \sum\_{k=1}^{K} a\_k e^{(j\phi\_k)} e^{(-d\_k t + 2\pi j f\_l t)} v\_k(t) + B(t) + w(t) + \varepsilon(t) \tag{1}$$

where *K* is the number of metabolites (*k* = 1, . . . , *K*), *vk*(*t*) the metabolite profile in the basis set measured as individual phantoms or simulated using quantum mechanics, *ak* the amplitudes, *φ<sup>k</sup>* the phase shifts, *dk* the damping corrections, *fk* the frequency shifts due to field inhomogeneity, *<sup>j</sup>* <sup>=</sup> √−1, *<sup>B</sup>*(*t*) is the baseline due to macromolecule/lipid contamination (in AQSES it is fitted via nonparametric modeling using a basis of splines (Eilers & Marx, 1996)), *w*(*t*) the water resonance (filtered either before quantification as explained in section 3.2.1, or during quantification as in (Sundin et al., 1999)) and *�*(*t*) denotes white noise with standard deviation *σ*.

In order to assess quantification results, the Cramér-Rao lower bounds (CRLB) (Cavassila et al., 2001) are computed to provide an indication about the uncertainty and reliability of the estimated amplitudes (concentrations). Small CRLB values may (but not necessarily) indicate good parameter estimates and are proportional to the variance of the residue obtained from subtracting the fitted signal and the baseline from the original signal. Thus, acceptable CRLB should normally be below 40%.

#### **3.2 Methods for quantification improvement**

Although the quantification of *in vivo* MRS signals can often be reliably done using model (1) in AQSES, several issues appear when high field, short TE signals or multi-voxel MRSI data are being quantified. In this section, we present several quantification improvement methods implemented in AQSES (Croitor Sava, 2011; Osorio-Garcia, 2011). Whereas the filter of residual water is performed as a preprocessing step, the lineshape and baseline estimation methods are included inside the quantification method. Finally, a modified version of AQSES for MRSI data, which includes spatial knowledge is presented.

We present in more details the method AQSES, which will be used in the next sections to illustrate several recent improvements in the field of MRS(I) signal processing. This time-domain quantification method has been especially developed for short TE MRS signals, where a mathematical model fits to a basis set of predefined metabolite profiles. From this fit, metabolite amplitudes are obtained, which represent the weighting coefficients of a linear combination of corrected metabolite profiles with Lorentzian lineshapes used for quantification. Finally, these values are proportional to the concentrations of all estimated metabolites. AQSES is available in the Java open source software AQSES GUI (De Neuter et al., 2007; Poullet et al., 2007), as a quantification method inside the Matlab® graphical user interface SPID (Poullet, 2008) and as a plug-in in the jMRUI software package (version 4.1) (Stefan et al., 2009). Fig. 2 shows the results window for a quantification made in

The model describing a short TE *in vivo* time-domain MRS signal *y*(*t*) as a combination of

where *K* is the number of metabolites (*k* = 1, . . . , *K*), *vk*(*t*) the metabolite profile in the basis set measured as individual phantoms or simulated using quantum mechanics, *ak* the amplitudes, *φ<sup>k</sup>* the phase shifts, *dk* the damping corrections, *fk* the frequency shifts due to field inhomogeneity, *<sup>j</sup>* <sup>=</sup> √−1, *<sup>B</sup>*(*t*) is the baseline due to macromolecule/lipid contamination (in AQSES it is fitted via nonparametric modeling using a basis of splines (Eilers & Marx, 1996)), *w*(*t*) the water resonance (filtered either before quantification as explained in section 3.2.1, or during quantification as in (Sundin et al., 1999)) and *�*(*t*) denotes white noise with standard

In order to assess quantification results, the Cramér-Rao lower bounds (CRLB) (Cavassila et al., 2001) are computed to provide an indication about the uncertainty and reliability of the estimated amplitudes (concentrations). Small CRLB values may (but not necessarily) indicate good parameter estimates and are proportional to the variance of the residue obtained from subtracting the fitted signal and the baseline from the original signal.

Although the quantification of *in vivo* MRS signals can often be reliably done using model (1) in AQSES, several issues appear when high field, short TE signals or multi-voxel MRSI data are being quantified. In this section, we present several quantification improvement methods implemented in AQSES (Croitor Sava, 2011; Osorio-Garcia, 2011). Whereas the filter of residual water is performed as a preprocessing step, the lineshape and baseline estimation methods are included inside the quantification method. Finally, a modified version of AQSES

*vk*(*t*) + *B*(*t*) + *w*(*t*) + *�*(*t*) (1)

*e*(−*dkt*+2*πj fkt*)

**3.1 Automated Quantification of Short echo time MRS signals (AQSES)**

SPID.

deviation *σ*.

several metabolite profiles is:

*y*(*t*) =

*K* ∑ *k*=1

Thus, acceptable CRLB should normally be below 40%.

for MRSI data, which includes spatial knowledge is presented.

**3.2 Methods for quantification improvement**

*ake*(*jφk*)

Fig. 2. Quantification results window for AQSES in SPID (left side) of an *in vivo* 1H MRS signal from mouse brain acquired with a 9.4 T Bruker Biospec small animal MR scanner (Bruker BioSpin MRI, Ettlingen, Germany) with acquisition parameters: PRESS sequence, TR=4 s, TE=12 ms, SW=4 KHz, NDP=2048 points and 256 averages, volume of interest (VOI)= 3 <sup>×</sup> 1.75 <sup>×</sup> 1.75 *mm*3. The basis set of 16 metabolites used for quantification was measured *in vitro*. Right: plots of residue, splines baseline, estimated metabolites and original with estimated signal.

#### **3.2.1 Filtering of residual water**

As described in section 2.2, the water component (*w*(*t*) from Eq.(1)) can be filtered using HLSVD, however, appropriate filtering of this resonance in the case of distorted signals is often inaccurate. When HLSVD with one component (*i.e.*, a model order *K* = 1) is used to fit one non-Lorentzian peak, the result obtained by subtracting that modeled peak from the original signal produces a non-flat residual. In an ideal case, only white noise should be left in the residual.

The state-of-the-art methods proposed for determination of an optimal model order *K* in order to provide a reliable water peak removal have been described and studied on simulated, noiseless or non-overlapping peaks (Cabanes et al., 2001; Hu et al., 2010; Lin et al., 1997; Papy et al., 2007; Van Huffel & Vandewalle, 1991). For instance, (Cabanes et al., 2001) presented a method for optimal residual water removal of *in vivo* 1H MRS signals of

9 8 7 6 5 4 3 2 1 0 −1 ppm

The lineshape of a peak in the frequency domain is determined by the decay function (damping) of the time domain signal. Self-deconvolution methods make use of the fact that within a measurement the most important factor that determines the decay rate is the local field heterogeneity, thus, all metabolites are distorted in the same way and therefore, a common lineshape can be estimated. Nevertheless, the computation of this lineshape produces a noisy function that needs to be converted into a smooth function. Methods in the literature differ in the approaches used for smoothing this noisy function. Here, we present a lineshape estimation method for correcting lineshape distortions during quantification with AQSES, where the exponential dampings *e*(−*dkt*) in Eq.(1) are replaced by the common factor

Fig. 3. Top: Real part of the spectrum of an *in vivo* MRS signal measured at 1.5 T with acquisition parameters: Philips scanner, PRESS sequence, TR=6 s, TE=23 ms, SW=1 KHz, NDP=512 points and 64 averages. Bottom: filtered signals using HLSVD-PRO with the model orders of *K* = 25 and overestimated as *K* = 44. All modeled peaks in the frequency

region outside the interval [0 ppm, 4.3 ppm] were suppressed.

*y*(*t*) = *g*(*t*)

considering the damping estimates.

*K* ∑ *k*=1

*ake*(*jφk*)

The AQSES Lineshape algorithm for lineshape estimation is described below.

*e*(2*πj fkt*)

**Step 1. Initial fitting.** Quantification of the signal assuming a Lorentzian lineshape to extract the spectral parameters: amplitudes, frequencies, phases and dampings using the model in Eq. (1). Then, the signal is reconstructed from the estimated spectral parameters without

*g*(*t*) of arbitrary shape:

Quantification Improvements of 1

9 8 7 6 5 4 3 2 1 0 −1

**In vivo 1.5 T**

H MRS Signals 11

Fixed 25

Overestimation

*vk*(*t*) + *B*(*t*) + *�*(*t*) (2)

human brain using HLSVD with a model order of 25 (*K* = 25). Nevertheless, suppression of unwanted resonances of *in vivo* MRS signal using HLSVD may require more than 25 components. In fact, it is certain that a higher model order may better fit one non-Lorentzian peak. In such cases, the model order in HLSVD can be overestimated to adequately fit the experimental signal with distorted shape. Finally, all components encountered in the selected filtering region are suppressed without affecting the other resonances.

Therefore, we present a heuristic approach to estimate the model order by overestimating the number of components in HLSVD (Osorio-Garcia, 2011). In this chapter, we use the Lanczos algorithm with partial reorthogonalization, HLSVD-PRO (Laudadio et al., 2004). First, the tail of the time domain signal is truncated in order to obtain by Fourier transformation a spectrum with lower, but still adequate, spectral resolution; this truncation starts at the point at which the FID completely decays into the noise. Then, the signal is transformed to the frequency domain and an estimated number for the model order is obtained by counting the number of spectral points larger in absolute value than a noise-related threshold. We illustrate in Fig. 3 to 6 good peak suppression results on short TE *in vivo* and *in vitro* 1H MRS signals, as follows:


#### **3.2.2 Lineshape estimation**

The metabolite profiles used for quantification have ideally a Lorentzian shape, however, the lineshape of the metabolite resonances in an *in vivo* signal might be distorted. We aim to estimate the lineshape distortion from the *in vivo* signal and impose the same distortion to each individual metabolite to ensure the same lineshape in both the *in vivo* and the model spectra. Here, we present a method (AQSES Lineshape) to estimate a common lineshape which has been successfully used in simulated, *in vitro* and *in vivo* signals (Osorio-Garcia et al., 2011; Sima et al., 2009).

In the presence of magnetic field inhomogeneities due to field perturbations or tissue heterogeneities, the symmetry and linewidth of the lineshapes are disturbed. Therefore, the fitting of the MR signal using an ideal lineshape (*e.g.*, Lorentzian, Gaussian or Voigt) becomes unreliable. Although some preprocessing methods can correct lineshape distortions, the use of a separate reference acquisition signal or a well-separated reference peak may limit their applicability.

human brain using HLSVD with a model order of 25 (*K* = 25). Nevertheless, suppression of unwanted resonances of *in vivo* MRS signal using HLSVD may require more than 25 components. In fact, it is certain that a higher model order may better fit one non-Lorentzian peak. In such cases, the model order in HLSVD can be overestimated to adequately fit the experimental signal with distorted shape. Finally, all components encountered in the selected

Therefore, we present a heuristic approach to estimate the model order by overestimating the number of components in HLSVD (Osorio-Garcia, 2011). In this chapter, we use the Lanczos algorithm with partial reorthogonalization, HLSVD-PRO (Laudadio et al., 2004). First, the tail of the time domain signal is truncated in order to obtain by Fourier transformation a spectrum with lower, but still adequate, spectral resolution; this truncation starts at the point at which the FID completely decays into the noise. Then, the signal is transformed to the frequency domain and an estimated number for the model order is obtained by counting the number of spectral points larger in absolute value than a noise-related threshold. We illustrate in Fig. 3 to 6 good peak suppression results on short TE *in vivo* and *in vitro* 1H MRS signals, as follows:

• **1.5 T signals**. Filtering of residual water and reference peaks for an *in vivo* (normal human brain) and an *in vitro* (NAA) MRS signal obtained at 1.5 T was done by suppressing all peaks in the frequency region outside the intervals [0 ppm, 4.3 ppm] and [1 ppm, 4.44 ppm], respectively. For the *in vivo* signal in Fig. 3, results for *K* = 25 and *K* overestimated (*K* = 44) show a good water suppression, but the water region is flatter when *K* = 44. Fig. 4 illustrates the incomplete filtering of the reference peaks at 0 ppm and 8.4 ppm corresponding to solvents added to the NAA phantom solution when *K* = 25, but

• **9.4 T signals**. Filtering of residual water and reference peaks for an *in vivo* (mouse brain) and an *in vitro* (Glucose (Glc)) MRS signal obtained at 9.4 T was done by suppressing all peaks in the frequency region outside the intervals [0 ppm, 4.3 ppm] and [2.95 ppm, 4.44 ppm], respectively. A residual reference peak located at 2.8 ppm contained in the *in vitro* signal could only be suppressed when *K* was overestimated. Figures 5 and 6

The metabolite profiles used for quantification have ideally a Lorentzian shape, however, the lineshape of the metabolite resonances in an *in vivo* signal might be distorted. We aim to estimate the lineshape distortion from the *in vivo* signal and impose the same distortion to each individual metabolite to ensure the same lineshape in both the *in vivo* and the model spectra. Here, we present a method (AQSES Lineshape) to estimate a common lineshape which has been successfully used in simulated, *in vitro* and *in vivo* signals (Osorio-Garcia et al., 2011;

In the presence of magnetic field inhomogeneities due to field perturbations or tissue heterogeneities, the symmetry and linewidth of the lineshapes are disturbed. Therefore, the fitting of the MR signal using an ideal lineshape (*e.g.*, Lorentzian, Gaussian or Voigt) becomes unreliable. Although some preprocessing methods can correct lineshape distortions, the use of a separate reference acquisition signal or a well-separated reference peak may limit their

filtering region are suppressed without affecting the other resonances.

this is not the case when *K* is overestimated, thus, *K* = 309.

show the results for both signals.

**3.2.2 Lineshape estimation**

Sima et al., 2009).

applicability.

Fig. 3. Top: Real part of the spectrum of an *in vivo* MRS signal measured at 1.5 T with acquisition parameters: Philips scanner, PRESS sequence, TR=6 s, TE=23 ms, SW=1 KHz, NDP=512 points and 64 averages. Bottom: filtered signals using HLSVD-PRO with the model orders of *K* = 25 and overestimated as *K* = 44. All modeled peaks in the frequency region outside the interval [0 ppm, 4.3 ppm] were suppressed.

The lineshape of a peak in the frequency domain is determined by the decay function (damping) of the time domain signal. Self-deconvolution methods make use of the fact that within a measurement the most important factor that determines the decay rate is the local field heterogeneity, thus, all metabolites are distorted in the same way and therefore, a common lineshape can be estimated. Nevertheless, the computation of this lineshape produces a noisy function that needs to be converted into a smooth function. Methods in the literature differ in the approaches used for smoothing this noisy function. Here, we present a lineshape estimation method for correcting lineshape distortions during quantification with AQSES, where the exponential dampings *e*(−*dkt*) in Eq.(1) are replaced by the common factor *g*(*t*) of arbitrary shape:

$$y(t) = g(t) \sum\_{k=1}^{K} a\_k e^{(j\phi\_k)} e^{(2\pi j f\_k t)} v\_k(t) + \mathcal{B}(t) + \varepsilon(t) \tag{2}$$

The AQSES Lineshape algorithm for lineshape estimation is described below.

**Step 1. Initial fitting.** Quantification of the signal assuming a Lorentzian lineshape to extract the spectral parameters: amplitudes, frequencies, phases and dampings using the model in Eq. (1). Then, the signal is reconstructed from the estimated spectral parameters without considering the damping estimates.

9 8 7 6 5 4 3 2 1 0 −1 ppm

was included to simulate a damping different from the ideal Lorentzian. Fig. 8 shows the quantification results obtained for a set of simulated signals with triangular and eddy current distortions having small and large dampings. The top row shows the fits with AQSES and AQSES Lineshape with the corresponding residuals when simulating the signals with a small (left) and a large (right) damping. The residual corresponding to AQSES contains some patterns corresponding to metabolite contributions that were not correctly quantified due to the Lorentzian lineshape model, whereas the residuals corresponding to AQSES Lineshape show a nearly flat line containing white noise that may be attributed to the estimated lineshape model. At the bottom of each plot, the amplitude

• *In vitro* **signals.** An *in vitro* signal containing Ala, Cr, Gln, Glu, Lac, NAA and Tau was acquired using the default shimming technique with linewidth=1.36 Hz and SNR=22. The magnetic field was afterwards intentionally distorted by mis-setting the shim current of the X coil in order to simulate lineshape distortions caused by incorrect shimming; two distorted signals were then acquired. Results of quantification of *in vitro* signals are shown in Fig. 9. The undistorted *in vitro* signal is fitted identically by AQSES and AQSES Lineshape, *i.e.*, AQSES Lineshape reports convergence after the first iteration (results not shown). For the distorted signals, the resonances of Cr at 3 ppm and 3.9 ppm and the one from NAA at 2 ppm are not very well fitted with AQSES, while AQSES Lineshape is able to fit these peaks. This is due to the fact that the lineshape distortions have a shape different

Fig. 5. Top: Real part of the spectrum of an *in vivo* MRS signal measured at 9.4 T with acquisition parameters: PRESS sequence with implemented pre-delay OVS as well as the water suppression method VAPOR, FASTMAP shimming correction, TR=4 s, TE=12 ms, SW=4 KHz, 2048 points and 256 averages. Bottom: Filtered signals with the model orders of *K* = 25 and overestimated as *K* = 134. All modeled peaks in the frequency region outside the

interval [0 ppm,4.3 ppm] were suppressed.

Quantification Improvements of 1

estimates using AQSES and AQSES Lineshape are shown.

from the typical Lorentzian type considered by AQSES.

9 8 7 6 5 4 3 2 1 0 −1

**In vivo 9.4 T**

H MRS Signals 13

Fixed 25

Overestimation

Fig. 4. Top: Real part of the spectrum of *in vitro* NAA measured at 1.5 T with acquisition parameters: Philips scanner, PRESS sequence, TR=6 s, TE=23 ms, SW=1 KHz, NDP=2048 points and 128 averages. Bottom: filtered signals using HLSVD-PRO with the model orders of *K* = 25 and overestimated as *K* = 309. All modeled peaks in the frequency region outside the interval [1 ppm, 4.44 ppm] were suppressed. Residual resonances are visible for *K* = 25 in the water and reference peak regions, *i.e.*, 4.7 ppm, 0 ppm and 8.44 ppm.

**Step 2. Damping estimation.** Computation of the damping function *g*(*t*) as:

$$\log(t) = \frac{y(t) - B(t)}{\sum\_{k=1}^{K} a\_k e^{(j\phi\_k)} e^{(2\pi j f\_k t)} v\_k(t)}\tag{3}$$

where *y*(*t*) is the experimental signal, *B*(*t*) is the current baseline estimate, *K* is the number of metabolites, *vk*(*t*) the metabolite signal *k* in the basis set, and the amplitudes *ak*, frequency shifts *fk* and phase shift *φ<sup>k</sup>* are estimated from a previous AQSES iteration from **Step 1** or **Step 4**.


Results of the lineshape estimation are shown below.

• **Simulated signals.** The simulated MRS signals were generated as a linear combination of 7 *in vitro* measured metabolites: Alanine (Ala), Creatine (Cr), Glutamine (Gln), Glutamate (Glu), Lactate (Lac), N-Acetyl-Aspartate (NAA), and Taurine (Tau). Then a distortion

**NAA**

Fixed 25

Overestimation

*<sup>k</sup>*=<sup>1</sup> *ake*(*jφk*)*e*(2*πj fkt*)*vk*(*t*) (3)

9 8 7 6 5 4 3 2 1 0 −1 ppm

Fig. 4. Top: Real part of the spectrum of *in vitro* NAA measured at 1.5 T with acquisition

NDP=2048 points and 128 averages. Bottom: filtered signals using HLSVD-PRO with the model orders of *K* = 25 and overestimated as *K* = 309. All modeled peaks in the frequency region outside the interval [1 ppm, 4.44 ppm] were suppressed. Residual resonances are visible for *K* = 25 in the water and reference peak regions, *i.e.*, 4.7 ppm, 0 ppm and 8.44 ppm.

> *<sup>g</sup>*(*t*) = *<sup>y</sup>*(*t*) <sup>−</sup> *<sup>B</sup>*(*t*) ∑*<sup>K</sup>*

where *y*(*t*) is the experimental signal, *B*(*t*) is the current baseline estimate, *K* is the number of metabolites, *vk*(*t*) the metabolite signal *k* in the basis set, and the amplitudes *ak*, frequency shifts *fk* and phase shift *φ<sup>k</sup>* are estimated from a previous AQSES iteration

**Step 3. Smoothing.** Outliers caused by numerical instability and division by small numbers are reduced using local regression (LOESS). This method assigns lower weight to outliers

**Step 4. Estimate.** Spectral analysis is carried out again after point-wise multiplying the

• **Simulated signals.** The simulated MRS signals were generated as a linear combination of 7 *in vitro* measured metabolites: Alanine (Ala), Creatine (Cr), Glutamine (Gln), Glutamate (Glu), Lactate (Lac), N-Acetyl-Aspartate (NAA), and Taurine (Tau). Then a distortion

**Steps 2-4** are repeated until a residual smaller than a chosen threshold is obtained or a

in the regression (Cleveland, 1979) and allows a robust smoothing. (See Fig. 7)

original metabolite basis set with the new smoothed function *g*(*t*) from **Step 3**.

convergence of amplitude estimates is reached.

Results of the lineshape estimation are shown below.

parameters: Philips scanner, PRESS sequence, TR=6 s, TE=23 ms, SW=1 KHz,

**Step 2. Damping estimation.** Computation of the damping function *g*(*t*) as:

from **Step 1** or **Step 4**.

9 8 7 6 5 4 3 2 1 0 −1

Fig. 5. Top: Real part of the spectrum of an *in vivo* MRS signal measured at 9.4 T with acquisition parameters: PRESS sequence with implemented pre-delay OVS as well as the water suppression method VAPOR, FASTMAP shimming correction, TR=4 s, TE=12 ms, SW=4 KHz, 2048 points and 256 averages. Bottom: Filtered signals with the model orders of *K* = 25 and overestimated as *K* = 134. All modeled peaks in the frequency region outside the interval [0 ppm,4.3 ppm] were suppressed.

was included to simulate a damping different from the ideal Lorentzian. Fig. 8 shows the quantification results obtained for a set of simulated signals with triangular and eddy current distortions having small and large dampings. The top row shows the fits with AQSES and AQSES Lineshape with the corresponding residuals when simulating the signals with a small (left) and a large (right) damping. The residual corresponding to AQSES contains some patterns corresponding to metabolite contributions that were not correctly quantified due to the Lorentzian lineshape model, whereas the residuals corresponding to AQSES Lineshape show a nearly flat line containing white noise that may be attributed to the estimated lineshape model. At the bottom of each plot, the amplitude estimates using AQSES and AQSES Lineshape are shown.

• *In vitro* **signals.** An *in vitro* signal containing Ala, Cr, Gln, Glu, Lac, NAA and Tau was acquired using the default shimming technique with linewidth=1.36 Hz and SNR=22. The magnetic field was afterwards intentionally distorted by mis-setting the shim current of the X coil in order to simulate lineshape distortions caused by incorrect shimming; two distorted signals were then acquired. Results of quantification of *in vitro* signals are shown in Fig. 9. The undistorted *in vitro* signal is fitted identically by AQSES and AQSES Lineshape, *i.e.*, AQSES Lineshape reports convergence after the first iteration (results not shown). For the distorted signals, the resonances of Cr at 3 ppm and 3.9 ppm and the one from NAA at 2 ppm are not very well fitted with AQSES, while AQSES Lineshape is able to fit these peaks. This is due to the fact that the lineshape distortions have a shape different from the typical Lorentzian type considered by AQSES.

4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 ppm

Residual AQSES Lineshape

Quantification Improvements of 1

Residual AQSES

0

Ala

in Fig. 10.

**3.2.3 Baseline estimation**

Cre

Gln

Glu

Lac

simulated signals using AQSES and AQSES Lineshape.

NAA

Simulated AQSES AQSES Lineshape

Tau

quantification results. These signals were obtained as the linear combination of 7 metabolites (Ala, Cr, Gln, Glu, Lac, NAA, and Tau) measured *in vitro* at 9.4 T with acquisition parameters: PRESS sequence, TR=8 s, TE=20 ms, SW=4 KHz and NDP=2048 points and 64 averages. Top left: small damping (to simulate *in vitro* signals). Top right: large damping (to simulate *in vivo* signals). The bottom plots represent the amplitude estimates for the corresponding

• *In vivo* **signals.** An *in vivo* signal was acquired using the default shimming technique. Then, after modifying the first and second order shim coils *X* and *Z*2, two mis-shimmed signals were also acquired. The undistorted *in vivo* signal was fitted similarly by AQSES and AQSES Lineshape, *i.e.*, AQSES Lineshape reports convergence after the first iteration (results not shown). Results of quantification for the two distorted *in vivo* signals are shown

The contribution from macromolecules and lipids may vary depending on the anatomical region or due to disease (tumor or metabolic disease), providing potentially useful diagnostic information, but also creating a sort of baseline in the spectra. As a consequence, some of these macromolecule/lipid resonances also overlap with metabolite peaks and it is necessary to account for these contributions during the quantification. Therefore, several baseline methods

Fig. 8. Simulated MRS signals with lineshape distortions and the corresponding

0

Ala

Cre

Gln

Glu

Lac

NAA

Simulated AQSES Lineshape AQSES

Tau

0.5

1

Amplitude [a.u.]

1.5

0.5

Amplitude [a.u.]

1

1.5

Simulated AQSES AQSES Lineshape

> 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 ppm

Residual AQSES Lineshape

H MRS Signals 15

Residual AQSES

Simulated AQSES AQSES Lineshape

Fig. 6. Top: Real part of the spectrum of *in vitro* Glc measured at 9.4 T with acquisition parameters: PRESS sequence with implemented pre-delay OVS as well as the water suppression method VAPOR, FASTMAP shimming correction, TR=4 s, TE=12 ms, SW=4 KHz, 6144 points and 64 averages. Bottom: filtered signal with the model orders of *K* = 25 and overestimated as *K* = 126. All modeled peaks in the frequency region outside the interval [2.95 ppm,4.44 ppm] were suppressed. Residual resonances are visible at the region around 2.8 ppm (between the two vertical dashed lines).

Fig. 7. Time domain signal of the resulting lineshape for simulated (left), *in vitro* (middle) and *in vivo* (right) signals. The signal labeled as 'Original' corresponds to the *g*(*t*) function calculated with the ratio formula in Eq.(3) and the smoothed signal is its final denoised version after convergence.

**Glc**

Fixed 25

Overestimation

200 400 600 800 Time [sec]

In vivo

Original Smoothed

9 8 7 6 5 4 3 2 1 0

1000 2000 3000 4000 5000 Time [sec]

Fig. 7. Time domain signal of the resulting lineshape for simulated (left), *in vitro* (middle) and *in vivo* (right) signals. The signal labeled as 'Original' corresponds to the *g*(*t*) function calculated with the ratio formula in Eq.(3) and the smoothed signal is its final denoised

Fig. 6. Top: Real part of the spectrum of *in vitro* Glc measured at 9.4 T with acquisition parameters: PRESS sequence with implemented pre-delay OVS as well as the water suppression method VAPOR, FASTMAP shimming correction, TR=4 s, TE=12 ms, SW=4 KHz, 6144 points and 64 averages. Bottom: filtered signal with the model orders of *K* = 25 and overestimated as *K* = 126. All modeled peaks in the frequency region outside the interval [2.95 ppm,4.44 ppm] were suppressed. Residual resonances are visible at the region

In vitro

Original Smoothed

around 2.8 ppm (between the two vertical dashed lines).

Original Smoothed

500 1000 1500 Time [sec]

version after convergence.

Simulated

9 8 7 6 5 4 3 2 1 0

ppm

Fig. 8. Simulated MRS signals with lineshape distortions and the corresponding quantification results. These signals were obtained as the linear combination of 7 metabolites (Ala, Cr, Gln, Glu, Lac, NAA, and Tau) measured *in vitro* at 9.4 T with acquisition parameters: PRESS sequence, TR=8 s, TE=20 ms, SW=4 KHz and NDP=2048 points and 64 averages. Top left: small damping (to simulate *in vitro* signals). Top right: large damping (to simulate *in vivo* signals). The bottom plots represent the amplitude estimates for the corresponding simulated signals using AQSES and AQSES Lineshape.

• *In vivo* **signals.** An *in vivo* signal was acquired using the default shimming technique. Then, after modifying the first and second order shim coils *X* and *Z*2, two mis-shimmed signals were also acquired. The undistorted *in vivo* signal was fitted similarly by AQSES and AQSES Lineshape, *i.e.*, AQSES Lineshape reports convergence after the first iteration (results not shown). Results of quantification for the two distorted *in vivo* signals are shown in Fig. 10.

#### **3.2.3 Baseline estimation**

The contribution from macromolecules and lipids may vary depending on the anatomical region or due to disease (tumor or metabolic disease), providing potentially useful diagnostic information, but also creating a sort of baseline in the spectra. As a consequence, some of these macromolecule/lipid resonances also overlap with metabolite peaks and it is necessary to account for these contributions during the quantification. Therefore, several baseline methods

have been developed based on advanced acquisition techniques using inversion recovery (*i.e.*, metabolite-nulled spectrum) (Cudalbu et al., 2009; Kunz et al., 2010; Mlynárik et al., 2008; Pfeuffer et al., 1999), parametric (Bartha et al., 1999; Seeger et al., 2003) and non-parametric estimation methods (Poullet et al., 2007; Provencher, 2001; Ratiney et al., 2004). Fig. 11 shows an *in vivo* MRS signal from a mouse brain obtained at 9.4 T and the macromolecule/lipid

Here, we present a method that extracts characteristic information from a set of inversion recovery MM signals. Thus, individual macromolecule/lipid components are computed and

Original

H MRS Signals 17

MM

ppm

4.5 4 3.5 3 2.5 2 1.5 1 0.5 ppm

Fig. 11. Real part of the spectrum of an *in vivo* MRS signal and an MM signal obtained from a mouse brain. 'Original' acquisition parameters: Bruker 9.4 T scanner, PRESS sequence, TR=4 s, TE=12 ms, SW=4 KHz and 256 averages. 'MM' acquisition parameters: Bruker 9.4 T

Characterization of the macromolecular baseline was addressed by modeling *m* previously identified profiles shown in Fig. 12 left. Thus, a set of 10 inversion recovery signals was first measured and *m* individual lipid and macromolecular peaks were identified as *prior knowledge*. These resonances were observed at frequency locations around: 0.89 (MM1), 1.20 (MM2), 1.36 (MM3), 1.63 (MM4), 2.02 (MM5), 2.29 (MM6), 2.65 (MM7), 3.03 (MM8), 3.21 (MM9), 3.75 (MM10), and 4.31 ppm (MM11). The spectral fitting method AMARES (Vanhamme et al., 1997) in jMRUI (Stefan et al., 2009) was used to quantify the set of inversion recovery signals as a sum of 11 damped sinusoids and the mean of amplitudes, frequency locations and linewidths were used to create the individual MM resonances shown in Fig. 12 right. The final MM components are included in the basis set of AQSES subject to small

When lineshape and baseline estimations are considered in the quantification, the model in

where *y*(*t*) is the experimental signal, *B*(*t*) is the non-parametric (spline) baseline from the previous iteration, *MMi*(*t*) is the set of modeled profiles, *K* is the number of metabolites, *vk*(*t*)

*<sup>i</sup>*=<sup>1</sup> *<sup>a</sup>*˜*ie*(*jφ*˜*i*)*e*(2*π<sup>j</sup>* ˜

*fit*)*MMi*(*t*) (4)

scanner, TI=800 ms, TR=3 s, NDP=2048 points and 1024 averages.

*<sup>g</sup>*(*t*) = *<sup>y</sup>*(*t*) <sup>−</sup> *<sup>B</sup>*(*t*)

*<sup>k</sup>*=<sup>1</sup> *ake*(*jφk*)*e*(2*πj fkt*)*vk*(*t*) + <sup>∑</sup>*<sup>m</sup>*

included in the metabolite basis set used in the AQSES quantification method.

baseline obtained via inversion recovery.

Quantification Improvements of 1

parameter variations.

Eq. (3) changes and *g*(*t*) is estimated as:

∑*K*

Fig. 9. Spectra of *in vitro* MRS signals measured with a Bruker scanner at 9.4 T containing 7 metabolites: Ala, Cr, Gln, Glu, Lac, NAA, and Tau. Acquisition parameters: PRESS sequence, TR=8 s, TE=20 ms, SW=4 KHz and NDP=2048 points and 128 averages. Two distorted signals were acquired by mis-setting the shim current of the X coil, Left: 'in vitro 2' with linewidth=3.92 Hz and SNR=20. Right: 'in vitro 3' with linewidth=6.52 Hz and SNR=18. Both were quantified by AQSES and AQSES Lineshape.

Fig. 10. Rat brain *in vivo* spectra from the right hemisphere of the thalamus measured with a Bruker scanner at 9.4 T with acquisition parameters: PRESS sequence, TR=8 s, TE=20 ms, SW=4 KHz, NDP=2048 points and 128 averages. 'In vivo 2' and 'in vivo 3' were acquired by mis-setting the shimming parameters of the first and second order shim coils *X* and *Z*2. The linewidth for these signals was 27.8 Hz and 39.69 Hz, respectively with SNR=20. Both were quantified by AQSES and AQSES Lineshape.

Fig. 9. Spectra of *in vitro* MRS signals measured with a Bruker scanner at 9.4 T containing 7 metabolites: Ala, Cr, Gln, Glu, Lac, NAA, and Tau. Acquisition parameters: PRESS sequence, TR=8 s, TE=20 ms, SW=4 KHz and NDP=2048 points and 128 averages. Two distorted signals were acquired by mis-setting the shim current of the X coil, Left: 'in vitro 2' with linewidth=3.92 Hz and SNR=20. Right: 'in vitro 3' with linewidth=6.52 Hz and SNR=18.

Residual AQSES

Residual AQSES Lineshape

Residual AQSES

Fig. 10. Rat brain *in vivo* spectra from the right hemisphere of the thalamus measured with a Bruker scanner at 9.4 T with acquisition parameters: PRESS sequence, TR=8 s, TE=20 ms, SW=4 KHz, NDP=2048 points and 128 averages. 'In vivo 2' and 'in vivo 3' were acquired by mis-setting the shimming parameters of the first and second order shim coils *X* and *Z*2. The linewidth for these signals was 27.8 Hz and 39.69 Hz, respectively with SNR=20. Both were

Residual AQSES Lineshape

4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 ppm

4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 ppm

In vitro 3 AQSES AQSES Lineshape

In vivo 3 AQSES AQSES Lineshape

In vitro 2 AQSES AQSES Lineshape

4.5 4 3.5 3 2.5 2 1.5 1 ppm

4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 ppm

quantified by AQSES and AQSES Lineshape.

Both were quantified by AQSES and AQSES Lineshape.

In vivo 2 AQSES AQSES Lineshape

Residual AQSES Residual AQSES Lineshape

Residual AQSES

Residual AQSES Lineshape

have been developed based on advanced acquisition techniques using inversion recovery (*i.e.*, metabolite-nulled spectrum) (Cudalbu et al., 2009; Kunz et al., 2010; Mlynárik et al., 2008; Pfeuffer et al., 1999), parametric (Bartha et al., 1999; Seeger et al., 2003) and non-parametric estimation methods (Poullet et al., 2007; Provencher, 2001; Ratiney et al., 2004). Fig. 11 shows an *in vivo* MRS signal from a mouse brain obtained at 9.4 T and the macromolecule/lipid baseline obtained via inversion recovery.

Here, we present a method that extracts characteristic information from a set of inversion recovery MM signals. Thus, individual macromolecule/lipid components are computed and included in the metabolite basis set used in the AQSES quantification method.

Fig. 11. Real part of the spectrum of an *in vivo* MRS signal and an MM signal obtained from a mouse brain. 'Original' acquisition parameters: Bruker 9.4 T scanner, PRESS sequence, TR=4 s, TE=12 ms, SW=4 KHz and 256 averages. 'MM' acquisition parameters: Bruker 9.4 T scanner, TI=800 ms, TR=3 s, NDP=2048 points and 1024 averages.

Characterization of the macromolecular baseline was addressed by modeling *m* previously identified profiles shown in Fig. 12 left. Thus, a set of 10 inversion recovery signals was first measured and *m* individual lipid and macromolecular peaks were identified as *prior knowledge*. These resonances were observed at frequency locations around: 0.89 (MM1), 1.20 (MM2), 1.36 (MM3), 1.63 (MM4), 2.02 (MM5), 2.29 (MM6), 2.65 (MM7), 3.03 (MM8), 3.21 (MM9), 3.75 (MM10), and 4.31 ppm (MM11). The spectral fitting method AMARES (Vanhamme et al., 1997) in jMRUI (Stefan et al., 2009) was used to quantify the set of inversion recovery signals as a sum of 11 damped sinusoids and the mean of amplitudes, frequency locations and linewidths were used to create the individual MM resonances shown in Fig. 12 right. The final MM components are included in the basis set of AQSES subject to small parameter variations.

When lineshape and baseline estimations are considered in the quantification, the model in Eq. (3) changes and *g*(*t*) is estimated as:

$$g(t) = \frac{y(t) - B(t)}{\sum\_{k=1}^{K} a\_k e^{(j\phi\_k)} e^{(2\pi jf\_k t)} v\_k(t) + \sum\_{i=1}^{m} \mathfrak{a}\_i e^{(j\tilde{\phi}\_i)} e^{(2\pi i \tilde{j} \tilde{f}\_i t)} M M\_i(t)} \tag{4}$$

where *y*(*t*) is the experimental signal, *B*(*t*) is the non-parametric (spline) baseline from the previous iteration, *MMi*(*t*) is the set of modeled profiles, *K* is the number of metabolites, *vk*(*t*)

4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 ppm

Quantification Improvements of 1

the smoothness of individual parameters is adjustable:

baseline.

Fig. 13. AQSES fit for the mean spectra of short TE 1H MRS *in vivo* from mouse brains. Real part of the *in vivo* spectrum (noisy signal) together with the fit using AQSES (bold line on top of the signal), MM (dash-dotted curve below the signal), spline baseline (smooth curve below the MM signal) and the residual (noisy line beneath). Quantification was made accounting for the baseline in two ways: (*i*) considering a metabolite-nulled signal computed as the mean of all measured signals (left) and (*ii*) considering 5 MM profiles obtained via AMARES (right). Additionally, a very smooth splines function was used to fit any broadening of the

data on a voxel-by-voxel basis can suffer from accuracy limitations that are inherent to maximum likelihood estimation, and, moreover, it can be prone to faulty convergence to local minima. Since MRSI provides spatial knowledge, in order to improve quantification results, and thus, metabolite estimates in MRSI data, a new metabolite quantification method, in which the available spatial information is exploited, has been proposed. The method, called AQSES-MRSI, is a modified version of AQSES for MRSI data (Croitor Sava et al., 2011).

AQSES-MRSI starts by individually fitting each signal in the grid using nonlinear least squares (Poullet et al., 2007; Sima & Van Huffel, 2007) to extract the spectral parameters, frequencies and dampings, from each voxel which will be further used as *prior knowledge*. Then, for the quantification of each voxel *c* within the MRSI grid, spatial information is taken into account in more steps. Moreover, several sweeps are performed through the grid and at each run some hyper-parameters may be tuned. Firstly, the starting values for the nonlinear parameters, *θc* (vector containing damping corrections and frequency shifts for voxel *c*) are optimized by setting them to the median of the parameter values from the considered neighbors *θs*(*s* = 1, . . . , *S*, where *S* is the total number of voxels in the considered neighborhood). Secondly, optimized bounds on the parameters' variability are computed so that the parameters of the neighboring voxels do not present a high variability. Thirdly, a penalty term that promotes a spatially smooth spectral parameter map for the frequency shifts and damping corrections are imposed, while allowing complete freedom to the metabolite amplitudes. The weight on

4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 ppm

H MRS Signals 19

Fig. 12. **Left:** Real part of an *in vivo* metabolite-nulled MM spectrum acquired with a Bruker scanner at 9.4 T with acquisition parameters: TI=800 ms, TR=3 s, NDP=2048 points and 1024 averages. Spectra were post-processed with a 15 Hz Lorentzian line broadening. Due to a shorter *T*1, the marked (Cr + PCr) resonance at 3.9 ppm was not completely minimized, therefore this peak was filtered out using HLSVD-PRO (top curve). Macromolecular resonances are labeled as MM from 1 until 11 at the following central frequencies: 0.89, 1.20, 1.36, 1.63, 2.02, 2.29, 2.65, 3.03, 3.21, 3.75, 4.31 ppm; the mean of 10 filtered MM signals is shown in the bottom curve. **Right:** Real part of the MM spectrum obtained with AMARES using *prior knowledge* from the measured inversion recovery signals. This plot shows the individual MMs computed based on the mean of the AMARES estimates of all measured MMs. Small variations are expected for different signals which are then corrected by AQSES.

the metabolite signal *<sup>k</sup>* in the basis set, and the amplitudes *ak*, *<sup>a</sup>*˜*i*, frequency shifts *fk*, ˜ *fi* and phase shift *φk*, *φ*˜*<sup>i</sup>* are estimated from the previous iteration of the AQSES Lineshape algorithm.

To avoid mis-quantification due to overlapping of MM resonances with metabolites, we combined MM3 with MM5-MM11 to obtain a single profile of seven resonances located in the region between 2 ppm - 4.1 ppm where all important metabolites are resonating. Fig. 13 shows the results of fitting using the mean of inversion recovery measured baselines and the mean of individual MMs obtained with AMARES (*i.e.*, five MM profiles).

#### **3.2.4 Inclusion of spatial constraints in MRSI quantification**

Magnetic Resonance Spectroscopic Imaging (MRSI) provides MR spectra from multiple adjacent voxels within a body volume represented as a 2 or 3 dimensional matrix, allowing measurement of the distribution of metabolites over this volume. Commonly, for estimating the metabolite concentrations, the signals within an MRSI grid are analyzed on a single voxel basis by quantifying each signal individually. To this aim, methods such as QUEST (Ratiney et al., 2005), AQSES (Poullet et al., 2007) or AMARES (Vanhamme et al., 1999) may be considered (the list can be extended, see the methods listed at the beginning of section 3). Compared with *in vitro*, *ex vivo* and single voxel *in vivo* MRS signals, *in vivo* MRSI data have a lower quality due to the spatial/spectral trade-off. Moreover, the magnetic field inhomogeneities, relatively low signal-to-noise ratio (SNR) and physiological motion that might appear during an MRSI acquisition compromise the spectral resolution and lead to strongly overlapping metabolite peaks. Therefore, quantifying metabolites within MRSI

Fig. 12. **Left:** Real part of an *in vivo* metabolite-nulled MM spectrum acquired with a Bruker scanner at 9.4 T with acquisition parameters: TI=800 ms, TR=3 s, NDP=2048 points and 1024 averages. Spectra were post-processed with a 15 Hz Lorentzian line broadening. Due to a shorter *T*1, the marked (Cr + PCr) resonance at 3.9 ppm was not completely minimized, therefore this peak was filtered out using HLSVD-PRO (top curve). Macromolecular resonances are labeled as MM from 1 until 11 at the following central frequencies: 0.89, 1.20, 1.36, 1.63, 2.02, 2.29, 2.65, 3.03, 3.21, 3.75, 4.31 ppm; the mean of 10 filtered MM signals is shown in the bottom curve. **Right:** Real part of the MM spectrum obtained with AMARES using *prior knowledge* from the measured inversion recovery signals. This plot shows the individual MMs computed based on the mean of the AMARES estimates of all measured MMs. Small variations are expected for different signals which are then corrected by AQSES.

the metabolite signal *<sup>k</sup>* in the basis set, and the amplitudes *ak*, *<sup>a</sup>*˜*i*, frequency shifts *fk*, ˜

mean of individual MMs obtained with AMARES (*i.e.*, five MM profiles).

**3.2.4 Inclusion of spatial constraints in MRSI quantification**

phase shift *φk*, *φ*˜*<sup>i</sup>* are estimated from the previous iteration of the AQSES Lineshape algorithm. To avoid mis-quantification due to overlapping of MM resonances with metabolites, we combined MM3 with MM5-MM11 to obtain a single profile of seven resonances located in the region between 2 ppm - 4.1 ppm where all important metabolites are resonating. Fig. 13 shows the results of fitting using the mean of inversion recovery measured baselines and the

Magnetic Resonance Spectroscopic Imaging (MRSI) provides MR spectra from multiple adjacent voxels within a body volume represented as a 2 or 3 dimensional matrix, allowing measurement of the distribution of metabolites over this volume. Commonly, for estimating the metabolite concentrations, the signals within an MRSI grid are analyzed on a single voxel basis by quantifying each signal individually. To this aim, methods such as QUEST (Ratiney et al., 2005), AQSES (Poullet et al., 2007) or AMARES (Vanhamme et al., 1999) may be considered (the list can be extended, see the methods listed at the beginning of section 3). Compared with *in vitro*, *ex vivo* and single voxel *in vivo* MRS signals, *in vivo* MRSI data have a lower quality due to the spatial/spectral trade-off. Moreover, the magnetic field inhomogeneities, relatively low signal-to-noise ratio (SNR) and physiological motion that might appear during an MRSI acquisition compromise the spectral resolution and lead to strongly overlapping metabolite peaks. Therefore, quantifying metabolites within MRSI

5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 ppm

Individual MMs

MM1 MM2 MM4 MM11 MM3, MM5−10

*fi* and

4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 ppm

MM6

MM5

MM4 MM3 MM2 MM1

MM11 MM10 MM9

MM8 MM7

**Cr: 3.9 ppm**

Fig. 13. AQSES fit for the mean spectra of short TE 1H MRS *in vivo* from mouse brains. Real part of the *in vivo* spectrum (noisy signal) together with the fit using AQSES (bold line on top of the signal), MM (dash-dotted curve below the signal), spline baseline (smooth curve below the MM signal) and the residual (noisy line beneath). Quantification was made accounting for the baseline in two ways: (*i*) considering a metabolite-nulled signal computed as the mean of all measured signals (left) and (*ii*) considering 5 MM profiles obtained via AMARES (right). Additionally, a very smooth splines function was used to fit any broadening of the baseline.

data on a voxel-by-voxel basis can suffer from accuracy limitations that are inherent to maximum likelihood estimation, and, moreover, it can be prone to faulty convergence to local minima. Since MRSI provides spatial knowledge, in order to improve quantification results, and thus, metabolite estimates in MRSI data, a new metabolite quantification method, in which the available spatial information is exploited, has been proposed. The method, called AQSES-MRSI, is a modified version of AQSES for MRSI data (Croitor Sava et al., 2011).

AQSES-MRSI starts by individually fitting each signal in the grid using nonlinear least squares (Poullet et al., 2007; Sima & Van Huffel, 2007) to extract the spectral parameters, frequencies and dampings, from each voxel which will be further used as *prior knowledge*. Then, for the quantification of each voxel *c* within the MRSI grid, spatial information is taken into account in more steps. Moreover, several sweeps are performed through the grid and at each run some hyper-parameters may be tuned. Firstly, the starting values for the nonlinear parameters, *θc* (vector containing damping corrections and frequency shifts for voxel *c*) are optimized by setting them to the median of the parameter values from the considered neighbors *θs*(*s* = 1, . . . , *S*, where *S* is the total number of voxels in the considered neighborhood). Secondly, optimized bounds on the parameters' variability are computed so that the parameters of the neighboring voxels do not present a high variability. Thirdly, a penalty term that promotes a spatially smooth spectral parameter map for the frequency shifts and damping corrections are imposed, while allowing complete freedom to the metabolite amplitudes. The weight on the smoothness of individual parameters is adjustable:

Fig. 15. Metabolic maps obtained after applying AQSES-MRSI. The color scheme is relative to each metabolite. The patient is diagnosed to have glioblastoma tumor (lower right corner

H MRS Signals 21

AQSES-MRSI, overlapping peaks or peaks of compounds present at low concentration can

Fig. 14 presents four simulated MRSI data sets of 10×10 voxels generated with different levels of inhomogeneities within the parameter maps. 75% of each simulated MRSI grid contains a region with normal-tissue-like spectra and 25% of the grid presents tumor-tissue-like spectra (damping, frequency and amplitude values were set to values corresponding to the signals measured in normal brain and tumor region, respectively). The simulated MRSI signals were obtained as a linear combination of 11 metabolite profiles: NAA, myo-inositol (Myo), Cr, Phosphocholine (PCh), Glu, Lac, Ala, Glc, Tau plus two simulated lipids profiles located at 1.3 ppm (Lip1) and 0.9 ppm (Lip2). These profiles were selected from a measured database acquired with a 1.5 T Philips NT Gyroscan using a PRESS sequence with TE=23 ms, and a PRESS box of 2 <sup>×</sup> <sup>2</sup> <sup>×</sup> <sup>2</sup> *cm*3. The results illustrated in Fig. 14 show that regardless of the degree of inhomogeneity, AQSES-MRSI outperforms the single voxel approach, AQSES.

Results on an *in vivo* MRSI case (see Fig. 15) show that AQSES-MRSI method provides a much less noisy spatial metabolite distribution compared to single voxel approaches (Croitor Sava et al., 2011). Well-contoured metabolic maps are obtained for all illustrated metabolites. Even metabolites that are difficult to quantify with conventional approaches (see

When using HLSVD for filtering the water and unwanted resonances in *in vitro* and *in vivo* MRS(I) signals, the model order plays an important role. Compared to other methods, the overestimation approach is able to suppress distorted lineshape peaks in the presence of low or high SNR. Although we focused on 1H MRS signals, potential implementation with other

of the MRI image).

Quantification Improvements of 1

be better resolved than in single-voxel approaches.

Glu) are well estimated with AQSES-MRSI.

**4. Discussion and conclusions**

$$\min\_{\theta\_c} \frac{1}{N} \sum\_{t=t\_0}^{t\_{N-1}} \left| y\_c(t) - \hat{y}\_c(t, \theta\_c) \right|^2 + \sigma^2 \sum\_{\theta\_c \neq \theta\_s} \varepsilon\_s \theta\_{cs} \left\| W(\theta\_c - \theta\_s) \right\|\_2^2 \tag{5}$$

where the signal *yc* (*t*) corresponds to voxel *c* in the grid and the model *y*ˆ*<sup>c</sup>* (*t*, *θ<sup>c</sup>* ) is considered as a weighted sum of metabolite signals with nonlinear corrections *θc*, similarly to the model in Eq. (1). The second term, called penalty term, encourages a smooth solution for the problem. *εs* accounts for the trade-off between an optimal fitting of the current signal and the penalty, *σ*<sup>2</sup> is an estimate of the noise variance computed from the tail of the signal in time domain in voxel *c*, *βcs* is a weighting scalar which gives the influence of the parameters *θs* on the parameters *<sup>θ</sup><sup>c</sup>* (as described below), *<sup>W</sup>* is a diagonal weighting matrix, with *<sup>W</sup>* <sup>∈</sup> *<sup>R</sup>Km*×*Km*, which accounts for the scale differences between parameters, where *K* is the number of metabolite profiles and *m* is the number of parameters per metabolite.

AQSES-MRSI's performance on simulated MRSI data with several types of disturbances and on short echo time *in vivo* proton MRSI data showed improved metabolites estimates compared to quantification of each voxel signal individually, see Fig. 14 and 15. With

Fig. 14. First row: simulated MRSI grids with different parameter maps: A-D. 2nd and 3rd row illustrate color maps with the values of the error in estimating the metabolite concentrations, computed as a root mean square error (RMSE computed as in (Croitor Sava et al., 2011)), for each voxel for AQSES and AQSES-MRSI, respectively.

 2

where the signal *yc* (*t*) corresponds to voxel *c* in the grid and the model *y*ˆ*<sup>c</sup>* (*t*, *θ<sup>c</sup>* ) is considered as a weighted sum of metabolite signals with nonlinear corrections *θc*, similarly to the model in Eq. (1). The second term, called penalty term, encourages a smooth solution for the problem. *εs* accounts for the trade-off between an optimal fitting of the current signal and the penalty, *σ*<sup>2</sup> is an estimate of the noise variance computed from the tail of the signal in time domain in voxel *c*, *βcs* is a weighting scalar which gives the influence of the parameters *θs* on the parameters *<sup>θ</sup><sup>c</sup>* (as described below), *<sup>W</sup>* is a diagonal weighting matrix, with *<sup>W</sup>* <sup>∈</sup> *<sup>R</sup>Km*×*Km*, which accounts for the scale differences between parameters, where *K* is the number of

AQSES-MRSI's performance on simulated MRSI data with several types of disturbances and on short echo time *in vivo* proton MRSI data showed improved metabolites estimates compared to quantification of each voxel signal individually, see Fig. 14 and 15. With

Fig. 14. First row: simulated MRSI grids with different parameter maps: A-D. 2nd and 3rd

row illustrate color maps with the values of the error in estimating the metabolite concentrations, computed as a root mean square error (RMSE computed as in (Croitor Sava et al., 2011)), for each voxel for AQSES and AQSES-MRSI, respectively.

<sup>+</sup> *<sup>σ</sup>*<sup>2</sup> ∑ *θc*�=*θ<sup>s</sup>* *εsβcs* 

*W*(*θ<sup>c</sup>* − *θs*)

 2

<sup>2</sup> (5)

min *θc*

1 *N* *tN*−<sup>1</sup> ∑ *t*=*t*<sup>0</sup>

 

*yc*(*t*) <sup>−</sup> *<sup>y</sup>*ˆ*c*(*t*, *<sup>θ</sup>c*)

metabolite profiles and *m* is the number of parameters per metabolite.

Fig. 15. Metabolic maps obtained after applying AQSES-MRSI. The color scheme is relative to each metabolite. The patient is diagnosed to have glioblastoma tumor (lower right corner of the MRI image).

AQSES-MRSI, overlapping peaks or peaks of compounds present at low concentration can be better resolved than in single-voxel approaches.

Fig. 14 presents four simulated MRSI data sets of 10×10 voxels generated with different levels of inhomogeneities within the parameter maps. 75% of each simulated MRSI grid contains a region with normal-tissue-like spectra and 25% of the grid presents tumor-tissue-like spectra (damping, frequency and amplitude values were set to values corresponding to the signals measured in normal brain and tumor region, respectively). The simulated MRSI signals were obtained as a linear combination of 11 metabolite profiles: NAA, myo-inositol (Myo), Cr, Phosphocholine (PCh), Glu, Lac, Ala, Glc, Tau plus two simulated lipids profiles located at 1.3 ppm (Lip1) and 0.9 ppm (Lip2). These profiles were selected from a measured database acquired with a 1.5 T Philips NT Gyroscan using a PRESS sequence with TE=23 ms, and a PRESS box of 2 <sup>×</sup> <sup>2</sup> <sup>×</sup> <sup>2</sup> *cm*3. The results illustrated in Fig. 14 show that regardless of the degree of inhomogeneity, AQSES-MRSI outperforms the single voxel approach, AQSES.

Results on an *in vivo* MRSI case (see Fig. 15) show that AQSES-MRSI method provides a much less noisy spatial metabolite distribution compared to single voxel approaches (Croitor Sava et al., 2011). Well-contoured metabolic maps are obtained for all illustrated metabolites. Even metabolites that are difficult to quantify with conventional approaches (see Glu) are well estimated with AQSES-MRSI.

#### **4. Discussion and conclusions**

When using HLSVD for filtering the water and unwanted resonances in *in vitro* and *in vivo* MRS(I) signals, the model order plays an important role. Compared to other methods, the overestimation approach is able to suppress distorted lineshape peaks in the presence of low or high SNR. Although we focused on 1H MRS signals, potential implementation with other

Bartha, R., Drost, D. J. & Williamson, P. C. (1999). Factors affecting the quantification of short

Bartha, R., Drost, D., Menon, R. & Williamson, P. (2000). Spectroscopic lineshape correction by

Behar, K. L. & Ogino, T. (1993). Characterization of macromolecule resonances in the 1H NMR

Belki´c, D. & Belki´c, K. (2006). *In vivo* Magnetic Resonance Spectroscopy by the fast Padé

Bhat, H., Sajja, B. R. & Narayana, P. A. (2006). Fast quantification of proton Magnetic

Blamire, A. M., Rothman, D. L. & Nixon, T. (1996). Dynamic shim updating: A new

Cabanes, E., Confort-Gouny, S., Fur, Y. L., Simond, G. & Cozzone, P. J. (2001). Optimization of

spectra of the human brain, *Journal of Magnetic Resonance* 150(2): 116 – 125. Cavassila, S., Deval, S., Huegen, C., van Ormondt, D. & Graveron-Demilly, D. (1999). The

Chang, D., Banack, C. D. & Shah, S. L. (2007). Robust baseline correction algorithm for signal

Chen, H., Van Huffel, S., van Ormondt, D. & de Beer, R. (1996). Parameter estimation with

Chen, L., Weng, Z., Goh, L. & Garland, M. (2002). An efficient algorithm for automatic

Chu, W. C. W., Chik, K., Chan, Y., Yeung, D. K. W., Roebuck, D. J., Howard, R. G., Li, C.

Cleveland, W. S. (1979). Robust locally weighted regression and smoothing scatterplots,

Cobas, J. C., Bernstein, M. A., Martín-Pastor, M. & Tahoces, P. G. (2006). A new

MR imaging and 1H MR spectroscopy, *Radiology* 229(3): 659–669.

*Journal of the American Statistical Association* 74(368): pp. 829–836.

data, *Journal of Magnetic Resonance* 183(1): 145 – 151.

Resonance Spectroscopy signals, *Investigative Radiology* 34: 242–246. Cavassila, S., Deval, S., Huegen, C., van Ormondt, D. & Graveron-Demilly, D. (2001).

dense NMR spectra, *Journal of Magnetic Resonance* 187(2): 288 – 292.

and tissues, *Journal of Magnetic Resonance* 171(1): 143 – 150.

*of Magnetic Resonance* 158(1-2): 164 – 168.

spectrum of rat brain, *Magnetic Resonance in Medicine* 30(1): 38–44.

transform, *Physics in Medicine and Biology* 51(5): 1049.

*in Biomedicine* 12(4): 205–216.

Quantification Improvements of 1

*Resonance* 183(1): 110 – 122.

36(1): 159–165.

14(4): 278–283.

*Resonance in Medicine* 44(4): 641–645.

echo *in vivo* 1H MR spectra: prior knowledge, peak elimination, and filtering, *NMR*

H MRS Signals 23

QUECC: Combined QUALITY deconvolution and eddy current correction, *Magnetic*

Resonance Spectroscopic Imaging with artificial neural networks, *Journal of Magnetic*

approach towards optimized whole brain shimming, *Magnetic Resonance in Medicine*

residual water signal removal by HLSVD on simulated short echo time proton MR

beneficial influence of prior knowledge on the quantitation of *In vivo* Magnetic

Cramér-Rao bounds: an evaluation tool for quantitation, *NMR in Biomedicine*

*Prior Knowledge* of known signal poles for the quantification of NMR spectroscopy data in the time domain, *Journal of Magnetic Resonance, Series A* 119(2): 225 – 234. Chen, J.-H., Sambol, E. B., Kennealey, P. T., O'Connor, R. B., DeCarolis, P. L., Cory, D. G. &

Singer, S. (2004). Water suppression without signal loss in HR-MAS 1H NMR of cells

phase correction of NMR spectra based on entropy minimization -ACME, *Journal*

& Metreweli, C. (2003). White matter and cerebral metabolite changes in children undergoing treatment for acute lymphoblastic leukemia: Longitudinal study with

general-purpose fully automatic baseline-correction procedure for 1D and 2D NMR

nuclei is also feasible. Thus, in MRS signals a good peak suppression is essential to achieve accurate quantification results and is therefore important for a correct usage of the signals.

Even though increasing importance has been given to the lineshape of MR signals, many *in vivo* studies seem to still ignore this problem. For not heavily distorted MRS signals, the Lorentzian lineshape is a good approximation to the shape of the experimental data. However, it is advantageous to rely on a method that takes any shape into account. Therefore, AQSES Lineshape works iteratively and evaluates the performance of tuning a more flexible lineshape starting with the Lorentzian model. As a result, quantification of distorted signals with lineshape estimation showed good quantification results for simulated, *in vitro* and *in vivo* distorted signals.

Although the MM signal acquired by inversion recovery is known to provide a good approximation of the macromolecular contamination, it is also true that it requires a long acquisition time and is not reproducible when the conditions of the region of interest are affected by acquisition problems and various diseases. Moreover, it also contains some unsuppressed metabolites. Nevertheless, the baseline information acquired with this method is a good approximation to the expected MM contributions. Therefore, the presented approach includes *prior knowledge* about the frequency locations, amplitudes and linewidths of individual MM profiles, which are beneficial for the quantification method.

Exploiting spatial *prior knowledge* when analyzing MRSI data has been tackled previously (Bao & Maudsley, 2007; Kelm, 2007; Pels, 2005). In all these approaches a common quantification solution is formulated for the whole MRSI grid. Still, due to the heterogeneity of the tissue that characterizes brain tumors, and to the variations induced by magnetic field inhomogeneities, a common optimization over the whole MRSI array may not always provide satisfactory results. With AQSES-MRSI the bounds on the relevant values of the parameters are iteratively adapted, and/or the parameters of the model function take at each iteration new starting values for each voxel. Such a dynamic approach shows improved quantification results and demonstrates that considering spatial information can improve the estimation of metabolite levels.

#### **5. Acknowledgments**

Research supported by: GOA MaNet, CoE EF/05/006 Optimization in Engineering (OPTEC), PFV/10/002 (OPTEC), FWO postdoc grants, IBBT, IUAP P6/04 (DYSCO, 'Dynamical systems, control and optimization', 2007-2011), FAST (FP6-MC-RTN-035801), K.U. Leuven Center of Excellence 'MoSAIC'.

#### **6. References**


nuclei is also feasible. Thus, in MRS signals a good peak suppression is essential to achieve accurate quantification results and is therefore important for a correct usage of the signals. Even though increasing importance has been given to the lineshape of MR signals, many *in vivo* studies seem to still ignore this problem. For not heavily distorted MRS signals, the Lorentzian lineshape is a good approximation to the shape of the experimental data. However, it is advantageous to rely on a method that takes any shape into account. Therefore, AQSES Lineshape works iteratively and evaluates the performance of tuning a more flexible lineshape starting with the Lorentzian model. As a result, quantification of distorted signals with lineshape estimation showed good quantification results for simulated, *in vitro* and *in*

Although the MM signal acquired by inversion recovery is known to provide a good approximation of the macromolecular contamination, it is also true that it requires a long acquisition time and is not reproducible when the conditions of the region of interest are affected by acquisition problems and various diseases. Moreover, it also contains some unsuppressed metabolites. Nevertheless, the baseline information acquired with this method is a good approximation to the expected MM contributions. Therefore, the presented approach includes *prior knowledge* about the frequency locations, amplitudes and linewidths

Exploiting spatial *prior knowledge* when analyzing MRSI data has been tackled previously (Bao & Maudsley, 2007; Kelm, 2007; Pels, 2005). In all these approaches a common quantification solution is formulated for the whole MRSI grid. Still, due to the heterogeneity of the tissue that characterizes brain tumors, and to the variations induced by magnetic field inhomogeneities, a common optimization over the whole MRSI array may not always provide satisfactory results. With AQSES-MRSI the bounds on the relevant values of the parameters are iteratively adapted, and/or the parameters of the model function take at each iteration new starting values for each voxel. Such a dynamic approach shows improved quantification results and demonstrates that considering spatial information can improve the estimation of

Research supported by: GOA MaNet, CoE EF/05/006 Optimization in Engineering (OPTEC), PFV/10/002 (OPTEC), FWO postdoc grants, IBBT, IUAP P6/04 (DYSCO, 'Dynamical systems, control and optimization', 2007-2011), FAST (FP6-MC-RTN-035801), K.U. Leuven Center of

Antoine, J.-P., Coron, A. & Dereppe, J.-M. (2000). Water peak suppression: Time-frequency vs

Bao, Y. & Maudsley, A. (2007). Improved reconstruction for MR Spectroscopic Imaging, *IEEE*

Barkhuijsen, H., de Beer, R. & van Ormondt, D. (1987). Improved algorithm for noniterative

time-domain model fitting to exponentially damped magnetic resonance signals,

time-scale approach, *Journal of Magnetic Resonance* 144(2): 189 – 194.

*Transactions on Medical Imaging* 26(5): 686 –695.

*Journal of Magnetic Resonance* 73(3): 553 – 557.

of individual MM profiles, which are beneficial for the quantification method.

*vivo* distorted signals.

metabolite levels.

**5. Acknowledgments**

Excellence 'MoSAIC'.

**6. References**


independent component analysis (ICA) on childhood brain tumour 1H magnetic

H MRS Signals 25

R. A. (2002). Quantification of human brain metabolites from *in vivo* 1H NMR magnitude spectra using automated artificial neural network analysis, *Journal of*

macromolecule baseline in localized 1H-MR spectra of human brain, *Magnetic*

M. B. (2001). Brain metabolite concentration and dementia severity in Alzheimer's

homogenization of the human prefrontal cortex with a set of localized electrical coils,

macromolecule resonances in short echo time 1H NMR spectroscopy of human brain,

R. (2010). Diffusion-weighted spectroscopy: A novel approach to determine macromolecule resonances in short-echo time 1H-MRS, *Magnetic Resonance in*

Lanczos algorithms for blackbox MRS data quantitation, *Journal of Magnetic Resonance*

Subspace-based MRS data quantitation of multiplets using *prior knowledge*, *Journal*

I. J., van Ormondt, D. & Graveron-Demilly, D. (2011). Optimization of metabolite basis-sets prior to quantitation: a quantum mechanics approach, *Proc. of the 19th*

for noisy NMR signals: Applications to delayed acquisition data, *Journal of Magnetic*

Hiltunen, Y., Kaartinen, J., Pulkkinen, J., Häkkinen, A.-M., Lundbom, N. & Kauppinen,

Hofmann, L., Slotboom, J., Boesch, C. & Kreis, R. (2001). Characterization of the

Hu, S.-L. J., Bao, X. & Li, H. (2010). Model order determination and noise removal for modal parameter estimation, *Mechanical Systems and Signal Processing* 24(6): 1605 – 1620. Huang, W., Alexander, G. E., Chang, L., Shetty, H. U., Krasuski, J. S., Rapoport, S. I. & Schapiro,

Jiru, F. (2008). Introduction to post-processing techniques, *European journal of radiology*

Juchem, C., Nixon, T. W., McIntyre, S., Rothman, D. L. & de Graaf, R. A. (2010). Magnetic field

Kelm, B. (2007). *Evaluation of Vector-Valued Clinical Image Data Using Probabilistic Graphical*

Klose, U. (1990). *In vivo* proton spectroscopy in presence of eddy currents, *Magnetic Resonance*

Knight-Scott, J. (1999). Application of multiple inversion recovery for suppression of

Kunz, N., Cudalbu, C., Mlynárik, V., Hüppi, P. S., Sizonenko, S. V. & Gruetter,

Laudadio, T., Mastronardi, N., Vanhamme, L., Hecke, P. V. & Huffel, S. V. (2002). Improved

Laudadio, T., Selén, Y., Vanhamme, L., Stoica, P., Hecke, P. V. & Huffel, S. V. (2004).

Lazariev, A., Allouche, A.-R., Aubert-Frécon, M., Fauvelle, F., Elbayed, K., Piotto, M., Namer,

Maudsley, A. A. (1995). Spectral lineshape determination by self-deconvolution, *Journal of*

*International Society of Magnetic Resonance in Medicine - ISMRM 2011*. Lin, Y.-Y., Hodgkinson, P., Ernst, M. & Pines, A. (1997). A novel detection-estimation scheme

resonance spectra, *NMR in Biomedicine* 22(8): 809–818.

*Magnetic Resonance* 154(1): 1 – 5.

*Resonance in Medicine* 46(5): 855–863.

disease, *Neurology* 57(4): 626–632.

*in Medicine* 14(1): 26 – 30.

*Medicine* 64(4): 939–946.

*Resonance* 128(1): 30 – 41.

*of Magnetic Resonance* 168(1): 53 – 65.

*Magnetic Resonance, series B* 106(1): 47 – 57.

157(2): 292 – 297.

*Magnetic Resonance in Medicine* 63(1): 171–180.

*Journal of Magnetic Resonance* 140(1): 228 – 234.

*Models: Quantification and Pattern Recognition*, PhD thesis.

67(2): 202–217.

Quantification Improvements of 1


Cobas, J. C. & Sardina, F. J. (2003). Nuclear Magnetic Resonance data processing. MestRe-C:

Croitor Sava, A. R., Sima, D. M., Poullet, J.-B., Wright, A. J., Heerschap, A. &

Croitor Sava, A. R. (2011). *Signal processing and classification for Magnetic Resonance Spectroscopy*

Cudalbu, C., Mlynárik, V., Xin, L. & Gruetter, R. (2009). Quantification of *in vivo*

de Graaf, A. A., van Dijk, J. E. & BoéE, W. M. M. J. (1990). Quality: quantification improvement

de Graaf, R. (1998). *In vivo NMR Spectroscopy – Principles and Techniques*, Wiley: Chichester. De Neuter, B., Luts, J., L., V., Lemmerling, P. & Van Huffel, S. (2007). Java-based framework

Devos, A., Lukas, L., Suykens, J., Vanhamme, L., Tate, A., Howe, F., Majós, C., Moreno-Torres,

Dologlou, I., Van Huffel, S. & Van Ormondt, D. (1998). Frequency-selective MRS

Dong, Z., Dreher, W. & Leibfritz, D. (2006). Toward quantitative short-echo-time *in vivo*

Eilers, P. H. C. & Marx, B. D. (1996). Flexible smoothing with B-splines and penalties, *Statistical*

Elster, C., Schubert, F., Link, A., Walzel, M., Seifert, F. & Rinneberg, H. (2005). Quantitative

Gabr, R. E., Ouwerkerk, R. & Bottomley, P. A. (2006). Quantifying *in vivo* MR spectra with

Giskeødegård, G. F., Bloemberg, T. G., Postma, G., Sitter, B., Tessem, M.-B., Gribbestad, I. S.,

Gruetter, R. (1993). Automatic localized *in vivo* adjustment of all first- and second- order shim

Guo, Y., Ruan, S., Landré, J. & Constans, J. M. (2010). A sparse representation method

Hao, J., Zou, X., Wilson, M. P., Davies, N. P., Sun, Y., Peet, A. C. & Arvanitis, T. N.

uncertainties, *Magnetic Resonance in Medicine* 53(6): 1288–1296.

circles, *Journal of Magnetic Resonance* 179(1): 152 – 163.

coils, *Magnetic Resonance in Medicine* 29(6): 804–811.

signals, *Computer Methods and Programs in Biomedicine* 85: 129–137.

19A(2): 80–96.

24(7): 824–835.

13(3): 343–357.

130(2): 238 – 243.

55(6): 1441–1446.

*Science* 11: 89–121.

683(1): 1 – 11.

*Engineering* 57(7): 1620 –1627.

*with clinical applications*, PhD thesis.

*Technology* 20(10): 104034 (7pp).

A software package for desktop computers, *Concepts in Magnetic Resonance Part A*

Van Huffel, S. (2011). Exploiting spatial information to estimate metabolite levels in two-dimensional MRSI of heterogeneous brain lesions, *NMR in Biomedicine*

short echo-time proton magnetic resonance spectra at 14.1 T using two different approaches of modelling the macromolecule spectrum, *Measurement Science and*

by converting lineshapes to the Lorentzian type, *Magnetic Resonance in Medicine*

for processing and displaying short-echo-time Magnetic Resonance Spectroscopy

A., van der Graaf, M., Arús, C. & Huffel, S. V. (2004). Classification of brain tumours using short echo time 1H MR spectra, *Journal of Magnetic Resonance* 170(1): 164 – 175.

data quantification with frequency *Prior Knowledge*, *Journal of Magnetic Resonance*

proton MR spectroscopy without water suppression, *Magnetic Resonance in Medicine*

Magnetic Resonance Spectroscopy: Semi-parametric modeling and determination of

Bathen, T. F. & Buydens, L. M. (2010). Alignment of high resolution magic angle spinning magnetic resonance spectra using warping methods, *Analytica Chimica Acta*

for Magnetic Resonance Spectroscopy quantification, *IEEE Transactions on Biomedical*

(2009). A comparative study of feature extraction and blind source separation of

independent component analysis (ICA) on childhood brain tumour 1H magnetic resonance spectra, *NMR in Biomedicine* 22(8): 809–818.


Reynolds, G., Wilson, M., Peet, A. & Arvanitis, T. N. (2006). An algorithm for the automated

Romano, R., Motta, A., Camassa, S., Pagano, C., Santini, M. T. & Indovina, P. L. (2002). A

Ruiz-Pena, J. L., Pinero, P., Sellers, G., Argente, J., Casado, A., Foronda, J., Ucles, A. &

Seeger, U., Klose, U., Mader, I., Grodd, W. & Nägele, T. (2003). Parameterized evaluation of

Sima, D. M., Osorio-Garcia, M. I., Poullet, J.-B., Suvichakorn, A., Antoine, J.-P., Huffel,

Sima, D. M. & Van Huffel, S. (2007). Separable nonlinear least squares fitting with linear

Stefan, D., Di Cesare, F., Andrasescu, A., Popa, E., Lazariev, A., Vescovo, E., Strbak, O.,

Sundin, T., Vanhamme, L., Van Hecke, P., Dologlou, I. & Van Huffel, S. (1999). Accurate

van den Boogaart, A. (1997). Quantitative data analysis of *in vivo* MRS data sets, *Magnetic*

van der Veen, J. W., de Beer, R., Luyten, P. R. & van Ormondt, D. (1988). Accurate

Van Huffel, S. & Vandewalle, J. (1991). *The Total Least Squares Problem: Computational Aspects*

Vanhamme, L., Huffel, S. V., Hecke, P. V. & van Ormondt, D. (1999). Time-domain

Vanhamme, L., Sundin, T., Van Hecke, P. & Van Huffel, S. (2001). MR spectroscopy

*prior knowledge*, *Magnetic Resonance in Medicine* 6(1): 92–98.

software package, *Measurement Science and Technology* 20(10): 104035. Stoyanova, R. & Brown, T. R. (2001). NMR spectral quantitation by Principal Component

56(6): 1211–1219.

Quantification Improvements of 1

*Resonance* 155(2): 226 – 235.

and spectroscopy, *BMC Neurology* 4(1): 8.

*Magnetic Resonance in Medicine* 40(6): 822–831.

Analysis, *NMR in Biomedicine* 14(4): 271–277.

*Journal of Magnetic Resonance* 140(1): 120 – 130.

*Resonance in Chemistry* 35(13): 146–152.

*and Analysis*, SIAM.

*Resonance in Medicine* 49(1): 19–28.

*Technology* 20(10): 104031.

quantitation of metabolites in *in vitro* NMR signals, *Magnetic Resonance in Medicine*

H MRS Signals 27

new time-domain frequency-selective quantification algorithm, *Journal of Magnetic*

Izquierdo, G. (2004). Magnetic resonance spectroscopy of normal appearing white matter in early relapsing-remitting multiple sclerosis: correlations between disability

macromolecules and lipids in proton MR Spectroscopy of brain diseases, *Magnetic*

S. V. & van Ormondt, D. (2009). Lineshape estimation for Magnetic Resonance Spectroscopy (MRS) signals: self-deconvolution revisited, *Measurement Science and*

bound constraints and its application in magnetic resonance spectroscopy data quantification, *Journal of Computational and Applied Mathematics* 203: 264–278. Slotboom, J., Boesch, C. & Kreis, R. (1998). Versatile frequency domain fitting using time domain models and *prior knowledge*, *Magnetic Resonance in Medicine* 39(6): 899–911. Soher, B. J., Young, K., Govindaraju, V. & Maudsley, A. A. (1998). Automated spectral analysis

III: Application to *in vivo* proton MR spectroscopy and spectroscopic imaging,

Williams, S., Starˇcuk, Z., Cabanas, M., van Ormondt, D. & Graveron-Demilly, D. (2009). Quantitation of Magnetic Resonance Spectroscopy signals: the jMRUI

quantification of 1H spectra: From finite impulse response filter design for solvent suppression to parameter estimation, *Journal of Magnetic Resonance* 139(2): 189 – 204.

quantification of *in vivo* 31P NMR signals using the variable projection method and

quantification of series of biomedical magnetic resonance spectroscopy signals,

quantitation: a review of time-domain methods, *NMR in Biomedicine* 14(4): 233–246.


Metz, K. R., Lam, M. M. & Webb, A. G. (2000). Reference deconvolution: A simple

Metzger, G. J., Patel, M. & Hu, X. (1996). Application of genetic algorithms to spectral quantification, *Journal of Magnetic Resonance, Series B* 110(3): 316 – 320. Mierisová, S. & Ala-Korpela, M. (2001). MR Spectroscopy quantitation: a review of frequency

Mlynárik, V., Cudalbu, C., Xin, L. & Gruetter, R. (2008). 1H NMR Spectroscopy of rat brain

Osorio-Garcia, M. (2011). *Advanced signal processing for Magnetic Resonance Spectroscopy*, PhD

Osorio-Garcia, M. I., Sima, D. M., Nielsen, F. U., Himmelreich, U. & Van Huffel, S.

Papy, J.-M., De Lathauwer, L. & Van Huffel, S. (2007). A shift invariance-based order-selection technique for exponential data modelling, *IEEE Signal Processing Letters* 14(7). Pels, P. (2005). *Analysis and Improvement of Quantification Algorithms for Magnetic Resonance*

Pfeuffer, J., Tkáˇc, I., Provencher, S. & Gruetter, R. (1999). Toward an *in vivo* neurochemical

Pijnappel, W. W. F., van den Boogaart, A., de Beer, R. & van Ormondt, D. (1992). SVD-based

Popa, E., Karras, D., Mertzios, B., Sima, D., de Beer, R., van Ormondt, D. & Graveron-Demilly,

Poullet, J.-B., Sima, D. M., Simonetti, A. W., De Neuter, B., Vanhamme, L., Lemmerling, P. &

Provencher, S. W. (2001). Automatic quantitation of localized *in vivo* 1H spectra with LCModel,

Ratiney, H., Coenradie, Y., Cavassila, S., van Ormondt, D. & Graveron-Demilly, D.

Ratiney, H., Sdika, M., Coenradie, Y., Cavassila, S., Ormondt, D. v. & Graveron-Demilly, D.

(2004). Time-domain quantitation of 1H short echo-time signals: background accommodation, *Magnetic Resonance Materials in Physics, Biology and Medicine*

(2005). Time-domain semi-parametric estimation based on a metabolite basis set,

Spectroscopy, *Concepts in Magnetic Resonance* 12(1): 21–42.

domain methods, *NMR in Biomedicine* 14(4): 247–259.

*Journal of Magnetic Resonance* 194(2): 163 – 168.

estimation, *Journal of Chemometrics* 25(4): 183–192.

rat brain, *Journal of Magnetic Resonance* 141(1): 104 – 120.

Poullet, J.-B. (2008). Spid: Simulation Package based on *in vitro* databases. URL: *http://homes.esat.kuleuven.be/ biomed/software.php*

NMR spectra, *Magnetic Resonance in Medicine* 30(6): 672–679.

thesis.

*Spectroscopy*, PhD thesis.

(Epub) ahead of print, 2011.

*NMR Biomed.* 14(4): 260–264.

*NMR in Biomedicine* 18(1): 1–13.

16: 284–296.

97(1): 122 – 134.

and effective method for resolution enhancement in Nuclear Magnetic Resonance

*in vivo* at 14.1 Tesla: Improvements in quantification of the neurochemical profile,

(2011). Quantification of Magnetic Resonance Spectroscopy signals with lineshape

profile: Quantification of 18 metabolites in short-echo-time 1H NMR spectra of the

quantification of Magnetic Resonance signals, *Journal of Magnetic Resonance (1969)*

D. (2011). Semi-parametric estimation without searching in function space. Application to *in vivo* metabolite quantitation, *Measurement Science and Technology*

Van Huffel, S. (2007). An automated quantitation of short echo time MRS spectra in an open source software environment: AQSES, *NMR in Biomedicine* 20(5): 493–504. Poullet, J.-B., Sima, D. M. & Van Huffel, S. (2008). MRS signal quantitation: A review of timeand frequency-domain methods, *Journal of Magnetic Resonance* 195(2): 134 – 144. Provencher, S. W. (1993). Estimation of metabolite concentrations from localized *in vivo* proton


**2** 

Bjørnar Hassel

*Norway* 

**13C Magnetic Resonance Spectroscopy in** 

**Neurobiology - Its Use in Monitoring Brain** 

*Department of Neurohabilitation, Oslo University Hospital-Ullevål, Oslo* 

*Norwegian Defense Research Establishment, Kjeller* 

**Energy Metabolism and in Identifying Novel** 

**Metabolic Substrates and Metabolic Pathways** 

A wealth of information on brain metabolism has been gathered from studies in which 13Clabeled metabolic substrates (often glucose) have been administered to human subjects or experimental animals. 13C Magnetic resonance spectroscopy (MRS1) of the brain (or extracts of brain) has shown to what degree the 13C-labeled compound has been metabolized, and, to some extent, along which metabolic pathways. The latter interpretation derives from the fact that 13C MRS shows the 13C labeling not of individual compounds, e.g. amino acids, but of individual carbon positions in those different compounds (Fig. 1). Because some enzymes of amino acid metabolism have a cell-specific expression, it is possible to study the metabolic activity of individual cell types or the transfer of amino acids between them. Finally, 13C MRS allows studies of a host of substrates to examine their potential as metabolic substrates for the brain. An extensive review, which included an overview of technical aspects of 13C MRS in studies of brain and cultured brain cells, appeared recently (Rodrigues et al., 2009).

Glucose is considered the physiologic energy substrate of the brain and is used by all brain cells, neurons and glial cells alike. Glucose consumption is tightly coupled to the need for energy (ATP), which in the brain largely reflects neuronal activity, or more specifically, depolarization of neuronal cell membranes. The membrane potential derives from the ionic gradients created by ion pumps that transport sodium and potassium against their concentration gradients across cell membranes, an activity that is fuelled by ATP (Attwell and Laughlin, 2001). Therefore, alterations in neuronal activity will be reflected in changes in glucose consumption, and, conversely, constraints in the availability of glucose and oxygen will limit neuronal activity. 13C MRS after administration of 13C-labeled glucose to

1 Abbreviations:ATP: adenosine triphosphate, CoA: coenzyme A, GABA: -aminobutyrate, i.v.

intravenously, MRS: magnetic resonance spectroscopy, PET; positron emission tomography, TCA cycle:

**2. 13C MRS for the visualization of brain energy metabolism** 

**1. Introduction** 

tricarboxylic acid cycle.


## **13C Magnetic Resonance Spectroscopy in Neurobiology - Its Use in Monitoring Brain Energy Metabolism and in Identifying Novel Metabolic Substrates and Metabolic Pathways**

Bjørnar Hassel

*Department of Neurohabilitation, Oslo University Hospital-Ullevål, Oslo Norwegian Defense Research Establishment, Kjeller Norway* 

#### **1. Introduction**

26 Magnetic Resonance Spectroscopy

28 Magnetic Resonance Spectroscopy

Vanhamme, L., van den Boogaart, A. & Van Huffel, S. (1997). Improved method for accurate

Vermathen, P., Laxer, K. D., Schuff, N., Matson, G. B. & Weiner, M. W. (2003). Evidence

Wilson, M., Reynolds, G., Kauppinen, R. A., Arvanitis, T. N. & Peet, A. C. (2011). A

Xi, Y. & Rocke, D. (2008). Baseline correction for NMR spectroscopic metabolomics data

Young, K., Govindaraju, V., Soher, B. J. & Maudsley, A. A. (1998). Automated spectral analysis

Spectroscopic Imaging - Initial experience, *Radiology* 226(1): 195–202. Veselkov, K. A., Lindon, J. C., Ebbels, T. M. D., Crockford, D., Volynkin, V. V., Holmes, E.,

*Magnetic Resonance* 129(1): 35 – 43.

analysis, *BMC Bioinformatics* 9(1): 324.

*Chemistry* 81(1): 56–66.

*Medicine* 40(6): 812–815.

and efficient quantification of MRS data with use of *Prior Knowledge*, *Journal of*

of neuronal injury outside the medial temporal lobe in temporal lobe epilepsy: N-Acetylaspartate concentration reductions detected with multisection Proton MR

Davies, D. B. & Nicholson, J. K. (2009). Recursive segment-wise peak alignment of biological 1H NMR spectra for improved metabolic biomarker recovery, *Analytical*

constrained least-squares approach to the automated quantitation of *in vivo* 1H Magnetic Resonance Spectroscopy data, *Magnetic Resonance in Medicine* 65(1): 1–12.

I: Formation of *a priori* information by spectral simulation, *Magnetic Resonance in*

A wealth of information on brain metabolism has been gathered from studies in which 13Clabeled metabolic substrates (often glucose) have been administered to human subjects or experimental animals. 13C Magnetic resonance spectroscopy (MRS1) of the brain (or extracts of brain) has shown to what degree the 13C-labeled compound has been metabolized, and, to some extent, along which metabolic pathways. The latter interpretation derives from the fact that 13C MRS shows the 13C labeling not of individual compounds, e.g. amino acids, but of individual carbon positions in those different compounds (Fig. 1). Because some enzymes of amino acid metabolism have a cell-specific expression, it is possible to study the metabolic activity of individual cell types or the transfer of amino acids between them. Finally, 13C MRS allows studies of a host of substrates to examine their potential as metabolic substrates for the brain. An extensive review, which included an overview of technical aspects of 13C MRS in studies of brain and cultured brain cells, appeared recently (Rodrigues et al., 2009).

#### **2. 13C MRS for the visualization of brain energy metabolism**

Glucose is considered the physiologic energy substrate of the brain and is used by all brain cells, neurons and glial cells alike. Glucose consumption is tightly coupled to the need for energy (ATP), which in the brain largely reflects neuronal activity, or more specifically, depolarization of neuronal cell membranes. The membrane potential derives from the ionic gradients created by ion pumps that transport sodium and potassium against their concentration gradients across cell membranes, an activity that is fuelled by ATP (Attwell and Laughlin, 2001). Therefore, alterations in neuronal activity will be reflected in changes in glucose consumption, and, conversely, constraints in the availability of glucose and oxygen will limit neuronal activity. 13C MRS after administration of 13C-labeled glucose to

<sup>1</sup> Abbreviations:ATP: adenosine triphosphate, CoA: coenzyme A, GABA: -aminobutyrate, i.v.

intravenously, MRS: magnetic resonance spectroscopy, PET; positron emission tomography, TCA cycle: tricarboxylic acid cycle.

13C Magnetic Resonance Spectroscopy in Neurobiology - Its Use in Monitoring

Brain Energy Metabolism and in Identifying Novel Metabolic Substrates and Metabolic Pathways 31

Fig. 2. Simplified scheme of glycolysis and the TCA cycle. Glucose is metabolized glycolytically to pyruvate, which may undergo decarboxylation in the pyruvate dehydrogenase reaction to become acetyl-CoA. Acetyl-CoA enters the TCA cycle to condense with oxaloacetate (OAA) and form citrate. The two carbons from acetyl-CoA that contribute to citrate (and hence α-ketoglutarate, glutamate, glutamine and GABA) are shown in italics. Due to the symmetry of succinate these two carbons are distributed evenly between the C1+C2 positions and the C3+C4 positions in succinate and in fumarate, malate and oxaloacetate, which are formed downstream of succinate. Pyruvate may also become

Oxaloacetate equilibrates with aspartate, and whereas oxaloacetate at a tissue concentration of ~10 nmol/g brain tissue (Siesjö, 1978) defies detection by 13C MRS, aspartate is readily detected, especially its 2nd and 3rd carbon positions, the labeling of which reflects the passage of 13C through the TCA cycle (Figs. 1 and 2). It has been shown with various techniques that aspartate is concentrated in GABAergic neurons (Ottersen and Storm-Mathisen, 1985; Hassel et al., 1992; 1995b). With 13C MRS this has been verified by blocking the TCA cycle of GABAergic neurons specifically; this leads to a marked reduction in the 13C labeling of aspartate from [1-13C]glucose (Hassel and Sonnewald, 1995a; Johannessen et al., 2001).

Inhibition of TCA cycle activity by drugs or toxins may reduce the 13C labeling of glutamate, GABA, or glutamine from [13C]glucose (e.g. Hassel and Sonnewald, 1995a). But because TCA cycle activity is closely coupled to neuronal activity, any influence that reduces brain activity (e.g. anesthetics, antiepileptic drugs) will reduce glucose metabolism in the brain

aminated to alanine or reduced to lactate.

human volunteers or to experimental animals is well suited to investigate the relationship between neuronal activity, energy requirements, and glucose metabolism, because several of the downstream metabolites of glucose (lactate, alanine, glutamate, glutamine GABA, aspartate – see Fig. 1) can be detected by 13C MRS.

Fig. 1. 13C MR spectrum of a brain extract from a mouse that received 150 µmol [U- 13C] glucose intravenously in the awake state and was sacrificed after 5 minutes. Peak numbers: 1, lactate C3; 2, glutamate C4; 3, GABA C2; 4, glutamine C4, 5, aspartate C3, 6: lactate C2; 7, alanine C3, 8: glutamate C3, which is labeled after a full turn of the TCA cycle. \*Internal standard (dioxane). Note the strong 13C labeling of glutamate C4 and lactate C3 from [U-13C]glucose. Reprinted from Nguyen et al., 2003, with permission from the publisher.

Administration of 13C-labeled glucose intravenously (i.v.) or intraperitoneally leads to predominant labeling of glutamate in the brain (Figs. 1 and 2). This is so, because glutamate is present in concentrations that allow detection by 13C MRS, and because virtually all glucose that enters the brain is metabolized through glutamate: glutamate equilibrates with α-ketoglutarate of the tricarboxylic acid (TCA) cycle, through which glucose is metabolized oxidatively. Briefly, 13C-glucose is metabolized to 13C-acetyl-CoA, which reacts with oxaloacetate to form citrate. Acetyl-CoA gives rise to the 1st and 2nd carbons of citrate, which correspond to the 5th and 4th carbon positions in α-ketoglutarate and glutamate (Fig. 2). Logically, these two positions are the first to be labeled from glucose, provided that uniformly labeled glucose, [U-13C]glucose is being used. If [1-13C]glucose is used, which is often the case, the 2nd position of acetyl-CoA becomes labeled, and hence the 4th position of glutamate (e.g. Fitzpatrick et al., 1990; Shank et al., 1993; Mason et al., 1999; see Figs. 1 and 2).

However, when α-ketoglutarate is metabolized further through the TCA cycle, succinate is formed. Succinate has four carbon atoms and is symmetrical, so the label is now scrambled between the 1st + 2nd positions and the 3rd + 4th positions (or the 2nd and 3rd positions if [1- 13C]glucose is being used). From succinate fumarate, malate, and oxaloacetate are formed before another molecule of acetyl-CoA enters the TCA cycle to form citrate. The labeling of the 2nd and 3rd positions in succinate, fumarate, malate, and oxaloacetate correspond to the 2nd and 3rd positions in glutamate. Labeling of these positions in glutamate reflects passage of 13C through the TCA cycle (Fig. 1 and 2). From the labeling of the 2nd and 3rd vs. the 4th positions in glutamate from [1-13C]glucose an impression of the TCA cycle rate may be gained, or the rate may even be calculated (Hassel et al., 1995a; 1997; Mason et al., 1999; Hyder et al., 2003).

human volunteers or to experimental animals is well suited to investigate the relationship between neuronal activity, energy requirements, and glucose metabolism, because several of the downstream metabolites of glucose (lactate, alanine, glutamate, glutamine GABA,

Fig. 1. 13C MR spectrum of a brain extract from a mouse that received 150 µmol [U- 13C] glucose intravenously in the awake state and was sacrificed after 5 minutes. Peak numbers: 1, lactate C3; 2, glutamate C4; 3, GABA C2; 4, glutamine C4, 5, aspartate C3, 6: lactate C2; 7, alanine C3, 8: glutamate C3, which is labeled after a full turn of the TCA cycle. \*Internal standard (dioxane). Note the strong 13C labeling of glutamate C4 and lactate C3 from [U-13C]glucose. Reprinted from Nguyen et al., 2003, with permission from the publisher.

Administration of 13C-labeled glucose intravenously (i.v.) or intraperitoneally leads to predominant labeling of glutamate in the brain (Figs. 1 and 2). This is so, because glutamate is present in concentrations that allow detection by 13C MRS, and because virtually all glucose that enters the brain is metabolized through glutamate: glutamate equilibrates with α-ketoglutarate of the tricarboxylic acid (TCA) cycle, through which glucose is metabolized oxidatively. Briefly, 13C-glucose is metabolized to 13C-acetyl-CoA, which reacts with oxaloacetate to form citrate. Acetyl-CoA gives rise to the 1st and 2nd carbons of citrate, which correspond to the 5th and 4th carbon positions in α-ketoglutarate and glutamate (Fig. 2). Logically, these two positions are the first to be labeled from glucose, provided that uniformly labeled glucose, [U-13C]glucose is being used. If [1-13C]glucose is used, which is often the case, the 2nd position of acetyl-CoA becomes labeled, and hence the 4th position of glutamate (e.g.

However, when α-ketoglutarate is metabolized further through the TCA cycle, succinate is formed. Succinate has four carbon atoms and is symmetrical, so the label is now scrambled between the 1st + 2nd positions and the 3rd + 4th positions (or the 2nd and 3rd positions if [1- 13C]glucose is being used). From succinate fumarate, malate, and oxaloacetate are formed before another molecule of acetyl-CoA enters the TCA cycle to form citrate. The labeling of the 2nd and 3rd positions in succinate, fumarate, malate, and oxaloacetate correspond to the 2nd and 3rd positions in glutamate. Labeling of these positions in glutamate reflects passage of 13C through the TCA cycle (Fig. 1 and 2). From the labeling of the 2nd and 3rd vs. the 4th positions in glutamate from [1-13C]glucose an impression of the TCA cycle rate may be gained, or the rate may even be calculated (Hassel et al., 1995a; 1997; Mason et al., 1999;

Fitzpatrick et al., 1990; Shank et al., 1993; Mason et al., 1999; see Figs. 1 and 2).

Hyder et al., 2003).

aspartate – see Fig. 1) can be detected by 13C MRS.

Fig. 2. Simplified scheme of glycolysis and the TCA cycle. Glucose is metabolized glycolytically to pyruvate, which may undergo decarboxylation in the pyruvate dehydrogenase reaction to become acetyl-CoA. Acetyl-CoA enters the TCA cycle to condense with oxaloacetate (OAA) and form citrate. The two carbons from acetyl-CoA that contribute to citrate (and hence α-ketoglutarate, glutamate, glutamine and GABA) are shown in italics. Due to the symmetry of succinate these two carbons are distributed evenly between the C1+C2 positions and the C3+C4 positions in succinate and in fumarate, malate and oxaloacetate, which are formed downstream of succinate. Pyruvate may also become aminated to alanine or reduced to lactate.

Oxaloacetate equilibrates with aspartate, and whereas oxaloacetate at a tissue concentration of ~10 nmol/g brain tissue (Siesjö, 1978) defies detection by 13C MRS, aspartate is readily detected, especially its 2nd and 3rd carbon positions, the labeling of which reflects the passage of 13C through the TCA cycle (Figs. 1 and 2). It has been shown with various techniques that aspartate is concentrated in GABAergic neurons (Ottersen and Storm-Mathisen, 1985; Hassel et al., 1992; 1995b). With 13C MRS this has been verified by blocking the TCA cycle of GABAergic neurons specifically; this leads to a marked reduction in the 13C labeling of aspartate from [1-13C]glucose (Hassel and Sonnewald, 1995a; Johannessen et al., 2001).

Inhibition of TCA cycle activity by drugs or toxins may reduce the 13C labeling of glutamate, GABA, or glutamine from [13C]glucose (e.g. Hassel and Sonnewald, 1995a). But because TCA cycle activity is closely coupled to neuronal activity, any influence that reduces brain activity (e.g. anesthetics, antiepileptic drugs) will reduce glucose metabolism in the brain

13C Magnetic Resonance Spectroscopy in Neurobiology - Its Use in Monitoring

Brain Energy Metabolism and in Identifying Novel Metabolic Substrates and Metabolic Pathways 33

the glutamine-derived glutamate becomes decarboxylated to GABA by glutamic acid decarboxylase (GAD), an enzyme that in the brain is expressed in inhibitory GABAergic neurons (Fonnum et al., 1970). Therefore, when administration of 13C-labeled acetate to experimental animals leads to 13C labeling of GABA, this labeling illustrates a series of metabolic and transport-related events (Fig. 3) that highlights the role of (astrocytic)

glutamine as a precursor of (neuronal) glutamate and GABA (Hassel et al., 1997).

Fig. 3. Metabolism of [13C]acetate in astrocytes with formation of [13C]glutamine, which is transferred to neurons, where it is deamidated to [13C]glutamate. In GABAergic neurons [13C]glutamate is decarboxylated to [13C]GABA. The spectrum is from a brain extract from a mouse that received [1,2-13C]acetate, 150 mol i.v. and was sacrificed after 15 minutes. Peaks

are: 1, alanine C-3; 2, lactate C-3; 3, GABA C-3; 5, glutamine C-3; 6, glutamate C-3; 7, glutamine C-4; 8, glutamate C-4; 10, GABA C-2; 11, aspartate C-3; 12, GABA C-4. Note the strong labeling of glutamine compared to results with [13C]glucose (Fig. 1) (From Hassel et

al., 1997, with permission from the publisher). Thus, formation of [13C]GABA from

membranes, neuronal glutaminase activity, and glutamate decarboxylase activity.

[13C]acetate illustrates astrocytic uptake and oxidative metabolism of [13C]acetate, glutamine synthetase activity, the activity of glutamine transporters at the astrocytic and neuronal cell

and hence 13C labeling of glucose metabolites. Therefore, a reduction in the labeling of cerebral metabolites from [13C]glucose must be interpreted with some caution. Such a reduction does not necessarily imply a direct effect, e.g. of a drug, on TCA cycle activity, for instance through enzyme inhibition.

Conversely, any influence that leads to increased activation of the brain may increase glucose metabolism and increase 13C labeling from [13C]glucose. Such effects have been seen in human subjects and in rats (Hyder et al., 1996; Patel et al., 2004). However, it should be noted that although previous studies on a variety of experimentally induced conditions, including epileptic seizures and stroke, have shown increased uptake of the glucose analogue deoxyglucose by positron emission tomography (PET) or autoradiographic methods, data obtained with 13C MRS seem to indicate that the oxidative metabolism of glucose may be only little affected or even reduced (Petroff et al., 2002; Eloqayli et al., 2004; Nguyen et al., 2007a; Pan et al., 2008; Håberg et al., 2009). The same is true for other conditions that would be expected to entail an increased energy demand, e.g. traumatic brain injury (Bartnik et al., 2007; Scafidi et al., 2009). Deoxyglucose is a substrate for the glucose transporter in the blood-brain barrier and in neurons, as well as for the first enzyme of glucose metabolism, hexokinase, which phopsphorylates glucose and deoxyglucose to their 6-phospho- derivatives. But once phophorylated deoxyglucose is not metabolized further and accumulates intracellularly; this allows its detection by PET or autoradiography. It follows that the deoxyglucose-based methods strictly detect the transport and initial phosphorylation of deoxyglucose, whereas 13C MRS detection of glutamate labeling from [13C]glucose reflects both glycolytic and oxidative metabolic activity in the brain.

The possibility of visualizing brain energy metabolism appeared to pave the way for studies in which brain activity, including sensory activation and thought processes, could be monitored by 13C MRS. It soon became clear, however, that the increase in glucose metabolism caused by various forms of physiological activation, was quite small, and not always readily detectable on the background of a high basal metabolic rate in the unanesthetized brain (Shulman et al., 2004).

#### **3. Visualization of the metabolic interplay between astrocytes and neurons**

The cellular complexity of the brain makes it difficult to assess the metabolic activity of individual cell types separately. However, because some metabolic pathways are unique to certain cells, and because some metabolic substrates are metabolized selectively by certain cell types, some degree of metabolic dissection of the brain is made possible. First we consider the transfer of glutamine from astrocytes to neurons as a precursor for glutamate and GABA in neurons. Thereafter we discuss the transfer of glutamate from neurons to astrocytes during glutamatergic neurotransmission.

In the brain, acetate is metabolized oxidatively by astrocytes and not by neurons. Because astrocytes express glutamine synthetase (Martinez-Hernandez et al., 1977), isotopically labeled acetate (given i.v. or intracerebrally) leads to preferential labeling of glutamine (van den Berg et al., 1969; Hassel et al., 1997; Lebon et al., 2002; Deelchand et al., 2009). Glutamine is exported from astrocytes to the extracellular fluid, from where it is taken up by neurons, both glutamatergic and GABAergic. Inside neurons glutamine becomes deamidated to glutamate by phosphate-activated glutaminase (Kvamme et al., 2001), a mitochondrial enzyme, which in the brain appears to be expressed only by neurons. In GABAergic neurons

and hence 13C labeling of glucose metabolites. Therefore, a reduction in the labeling of cerebral metabolites from [13C]glucose must be interpreted with some caution. Such a reduction does not necessarily imply a direct effect, e.g. of a drug, on TCA cycle activity, for

Conversely, any influence that leads to increased activation of the brain may increase glucose metabolism and increase 13C labeling from [13C]glucose. Such effects have been seen in human subjects and in rats (Hyder et al., 1996; Patel et al., 2004). However, it should be noted that although previous studies on a variety of experimentally induced conditions, including epileptic seizures and stroke, have shown increased uptake of the glucose analogue deoxyglucose by positron emission tomography (PET) or autoradiographic methods, data obtained with 13C MRS seem to indicate that the oxidative metabolism of glucose may be only little affected or even reduced (Petroff et al., 2002; Eloqayli et al., 2004; Nguyen et al., 2007a; Pan et al., 2008; Håberg et al., 2009). The same is true for other conditions that would be expected to entail an increased energy demand, e.g. traumatic brain injury (Bartnik et al., 2007; Scafidi et al., 2009). Deoxyglucose is a substrate for the glucose transporter in the blood-brain barrier and in neurons, as well as for the first enzyme of glucose metabolism, hexokinase, which phopsphorylates glucose and deoxyglucose to their 6-phospho- derivatives. But once phophorylated deoxyglucose is not metabolized further and accumulates intracellularly; this allows its detection by PET or autoradiography. It follows that the deoxyglucose-based methods strictly detect the transport and initial phosphorylation of deoxyglucose, whereas 13C MRS detection of glutamate labeling from

[13C]glucose reflects both glycolytic and oxidative metabolic activity in the brain.

The possibility of visualizing brain energy metabolism appeared to pave the way for studies in which brain activity, including sensory activation and thought processes, could be monitored by 13C MRS. It soon became clear, however, that the increase in glucose metabolism caused by various forms of physiological activation, was quite small, and not always readily detectable on the background of a high basal metabolic rate in the un-

**3. Visualization of the metabolic interplay between astrocytes and neurons**  The cellular complexity of the brain makes it difficult to assess the metabolic activity of individual cell types separately. However, because some metabolic pathways are unique to certain cells, and because some metabolic substrates are metabolized selectively by certain cell types, some degree of metabolic dissection of the brain is made possible. First we consider the transfer of glutamine from astrocytes to neurons as a precursor for glutamate and GABA in neurons. Thereafter we discuss the transfer of glutamate from neurons to

In the brain, acetate is metabolized oxidatively by astrocytes and not by neurons. Because astrocytes express glutamine synthetase (Martinez-Hernandez et al., 1977), isotopically labeled acetate (given i.v. or intracerebrally) leads to preferential labeling of glutamine (van den Berg et al., 1969; Hassel et al., 1997; Lebon et al., 2002; Deelchand et al., 2009). Glutamine is exported from astrocytes to the extracellular fluid, from where it is taken up by neurons, both glutamatergic and GABAergic. Inside neurons glutamine becomes deamidated to glutamate by phosphate-activated glutaminase (Kvamme et al., 2001), a mitochondrial enzyme, which in the brain appears to be expressed only by neurons. In GABAergic neurons

instance through enzyme inhibition.

anesthetized brain (Shulman et al., 2004).

astrocytes during glutamatergic neurotransmission.

the glutamine-derived glutamate becomes decarboxylated to GABA by glutamic acid decarboxylase (GAD), an enzyme that in the brain is expressed in inhibitory GABAergic neurons (Fonnum et al., 1970). Therefore, when administration of 13C-labeled acetate to experimental animals leads to 13C labeling of GABA, this labeling illustrates a series of metabolic and transport-related events (Fig. 3) that highlights the role of (astrocytic) glutamine as a precursor of (neuronal) glutamate and GABA (Hassel et al., 1997).

Fig. 3. Metabolism of [13C]acetate in astrocytes with formation of [13C]glutamine, which is transferred to neurons, where it is deamidated to [13C]glutamate. In GABAergic neurons [13C]glutamate is decarboxylated to [13C]GABA. The spectrum is from a brain extract from a mouse that received [1,2-13C]acetate, 150 mol i.v. and was sacrificed after 15 minutes. Peaks are: 1, alanine C-3; 2, lactate C-3; 3, GABA C-3; 5, glutamine C-3; 6, glutamate C-3; 7, glutamine C-4; 8, glutamate C-4; 10, GABA C-2; 11, aspartate C-3; 12, GABA C-4. Note the strong labeling of glutamine compared to results with [13C]glucose (Fig. 1) (From Hassel et al., 1997, with permission from the publisher). Thus, formation of [13C]GABA from [13C]acetate illustrates astrocytic uptake and oxidative metabolism of [13C]acetate, glutamine synthetase activity, the activity of glutamine transporters at the astrocytic and neuronal cell membranes, neuronal glutaminase activity, and glutamate decarboxylase activity.

13C Magnetic Resonance Spectroscopy in Neurobiology - Its Use in Monitoring

Brain Energy Metabolism and in Identifying Novel Metabolic Substrates and Metabolic Pathways 35

Fig. 4. Uptake of transmitter glutamate by astrocytes. In this experiment mice received the glia-specific metabolic inhibitor fluoroacetate, which blocks the astrocytic TCA cycle at the level of aconitase. Thus, the ability of astrocytes to form glutamate and hence glutamine from α-ketoglutarate from their own TCA cycle was blocked. Even so, glutamine was

An enigma of brain metabolism is why acetate is oxidized only by astrocytes and not by neurons. Waniewski and Martin (1998) provided evidence that acetate is taken up across the cell membrane only in astrocytes. This finding was unexpected, since acetate is a small monocarboxylate, for which neurons are richly equipped with transporters (Bergersen, 2007). Nguyen et al (2007b) later found that neurons take up propionate (which is closely related to acetate), which suggested that even acetate is taken up (but not metabolized) by neurons. Later it was found that administration of propionate to mice leads to an increase in brain GABA levels, suggesting an inhibitory action of propionate on the GABA-degrading enzyme GABA transaminase (Hassel et al., unpublished). GABA transaminase is an intramitochondrial enzyme in GABAergic neurons (Schousboe et al., 1974), and its inhibition by a small monocarboxylate such as propionate suggests that propionate (and acetate) is taken up by neurons and neuronal mitochondria, but that the latter are deficient in the enzyme that converts acetate and propionate into their mitochondrial CoA derivatives.

Because glutamate is readily labeled from 13C-labeled precursors, there has been considerable interest in identifying transmitter glutamate with 13C MRS. Glutamate is the main excitatory neurotransmitter in the central nervous system. While glutamate serves many functions in the brain, it has been estimated that 2-20% of the total amount of brain glutamate serves a neurotransmitter function (Hassel and Dingledine, 2006). After release from synaptic nerve endings glutamate is largely taken up by astrocytes that surround the synaptic cleft, and converted into glutamine (Danbolt, 2001). This process may be illustrated with 13C MRS. Administration of 13C-labeled glucose to experimental animals leads to rapid 13C labeling of glutamate and, after a delay, of glutamine. This labeling of glutamine to a large extent reflects uptake of (neuronal) transmitter glutamate into astrocytes. This was shown in an experiment in which mice had their astrocytic TCA cycle inhibited by the gliaspecific metabolic inhibitor fluoroacetate (Hassel et al., 1997). The mice could not form glutamate or glutamine through their TCA cycle activity, but they still produced 13Cglutamine from 13C-glucose. Formation of glutamine from 13C-glucose was interpreted to reflect formation of 13C-glutamate from 13C-glucose in neurons, release of 13C-labeled transmitter glutamate from neurons, uptake into astrocytes of the released 13C-labeled transmitter glutamate and subsequent amidation by glutamine synthetase (Fig. 4).

The exchange between neurons and astrocytes of glutamate and glutamine has been the subject of many studies, in both human subjects and experimental animals (Shen et al., 1999; see Rothman et al., 1999; 2003, for review). Interpretation of 13C MRS after infusion of 13Clabeled glucose into awake humans indicated that the flux of glutamate from neurons to astrocytes was close to 50% of the cerebral metabolism of glucose. This would suggest a massive transfer of glutamate from neurons to astrocytes. However, if this is true, it is still not known whether all this glutamate is transmitter glutamate in the sense that it originates from presynaptic vesicles.

In 13C MRS studies of the brain, acetate is used as a glial substrate, and glucose is used as an energy substrate for all brain cells, but with a dominant contribution from (glutamatergic) neurons. As will become evident below ('Identification of alternative substrates for brain energy metabolism'), lactate and pyruvate have become regarded as more purely neuronal substrates that are used by glutamatergic and GABAergic neurons alike, whereas glycerol appears to be a substrate specifically used by GABAergic neurons (Nguyen et al., 2003). Thus, some degree of metabolic dissection of the brain is made possible by combining the administration of an appropriate 13C-labeled energy substrate with 13C MRS.

An enigma of brain metabolism is why acetate is oxidized only by astrocytes and not by neurons. Waniewski and Martin (1998) provided evidence that acetate is taken up across the cell membrane only in astrocytes. This finding was unexpected, since acetate is a small monocarboxylate, for which neurons are richly equipped with transporters (Bergersen, 2007). Nguyen et al (2007b) later found that neurons take up propionate (which is closely related to acetate), which suggested that even acetate is taken up (but not metabolized) by neurons. Later it was found that administration of propionate to mice leads to an increase in brain GABA levels, suggesting an inhibitory action of propionate on the GABA-degrading enzyme GABA transaminase (Hassel et al., unpublished). GABA transaminase is an intramitochondrial enzyme in GABAergic neurons (Schousboe et al., 1974), and its inhibition by a small monocarboxylate such as propionate suggests that propionate (and acetate) is taken up by neurons and neuronal mitochondria, but that the latter are deficient in the enzyme that

Because glutamate is readily labeled from 13C-labeled precursors, there has been considerable interest in identifying transmitter glutamate with 13C MRS. Glutamate is the main excitatory neurotransmitter in the central nervous system. While glutamate serves many functions in the brain, it has been estimated that 2-20% of the total amount of brain glutamate serves a neurotransmitter function (Hassel and Dingledine, 2006). After release from synaptic nerve endings glutamate is largely taken up by astrocytes that surround the synaptic cleft, and converted into glutamine (Danbolt, 2001). This process may be illustrated with 13C MRS. Administration of 13C-labeled glucose to experimental animals leads to rapid 13C labeling of glutamate and, after a delay, of glutamine. This labeling of glutamine to a large extent reflects uptake of (neuronal) transmitter glutamate into astrocytes. This was shown in an experiment in which mice had their astrocytic TCA cycle inhibited by the gliaspecific metabolic inhibitor fluoroacetate (Hassel et al., 1997). The mice could not form glutamate or glutamine through their TCA cycle activity, but they still produced 13Cglutamine from 13C-glucose. Formation of glutamine from 13C-glucose was interpreted to reflect formation of 13C-glutamate from 13C-glucose in neurons, release of 13C-labeled transmitter glutamate from neurons, uptake into astrocytes of the released 13C-labeled

transmitter glutamate and subsequent amidation by glutamine synthetase (Fig. 4).

from presynaptic vesicles.

The exchange between neurons and astrocytes of glutamate and glutamine has been the subject of many studies, in both human subjects and experimental animals (Shen et al., 1999; see Rothman et al., 1999; 2003, for review). Interpretation of 13C MRS after infusion of 13Clabeled glucose into awake humans indicated that the flux of glutamate from neurons to astrocytes was close to 50% of the cerebral metabolism of glucose. This would suggest a massive transfer of glutamate from neurons to astrocytes. However, if this is true, it is still not known whether all this glutamate is transmitter glutamate in the sense that it originates

In 13C MRS studies of the brain, acetate is used as a glial substrate, and glucose is used as an energy substrate for all brain cells, but with a dominant contribution from (glutamatergic) neurons. As will become evident below ('Identification of alternative substrates for brain energy metabolism'), lactate and pyruvate have become regarded as more purely neuronal substrates that are used by glutamatergic and GABAergic neurons alike, whereas glycerol appears to be a substrate specifically used by GABAergic neurons (Nguyen et al., 2003). Thus, some degree of metabolic dissection of the brain is made possible by combining the

administration of an appropriate 13C-labeled energy substrate with 13C MRS.

converts acetate and propionate into their mitochondrial CoA derivatives.

Fig. 4. Uptake of transmitter glutamate by astrocytes. In this experiment mice received the glia-specific metabolic inhibitor fluoroacetate, which blocks the astrocytic TCA cycle at the level of aconitase. Thus, the ability of astrocytes to form glutamate and hence glutamine from α-ketoglutarate from their own TCA cycle was blocked. Even so, glutamine was

13C Magnetic Resonance Spectroscopy in Neurobiology - Its Use in Monitoring

astrocytes as the main metabolic compartment.

from the publisher.

Brain Energy Metabolism and in Identifying Novel Metabolic Substrates and Metabolic Pathways 37

a predominantly neuronal metabolism of these subtrates (Fig. 5). Even octanoate has been shown to be metabolized by the brain (Ebert et al., 2003), suggesting a certain capacity for fatty acid metabolism in the brain; the preferential labeling of glutamine pointed to

Fig. 5. 13C MR spectra of brain extracts of mice that received [3-13C]pyruvate or [2- 13C]pyruvate i.v. The (labeling of) individual carbon positions in certain metabolites is indicated. Wake mice received 9 mmoles/kg [3-13C]pyruvate (upper panel) or [2-

13C]pyruvate (lower panel) i.v. Survival time was 5 min. [3-13C]Pyruvate labels the C3 of lactate (Lac) and alanine (Ala), the C4 of glutamate (Glu) and glutamine (Gln), and GABA C2 (but not glutamate C5 or the C1 of GABA or lactate; insert). [2-13C]Pyruvate labels lactate and alanine C2, and glutamate C5, GABA C1 and lactate C1 (insert). D is dioxane (internal standard). The peak corresponding to Glu C4 in lower spectrum represents naturally abundant 13C-glutamate. Reprinted from Gonzalez et al., 2005, with permission

13C-Labeled acetate also yields good labeling of cerebral amino acids, predominantly glutamine, after i.v. injection, a finding that points to their metabolism in astrocytes (see above, 3. Visualization of the metabolic interplay between astrocytes and neurons). The labeling of amino acids in the brain from [13C]acetate is on a level similar to that achieved with [13C]glucose (Hassel et al., 1995a;1997), which suggests that acetate could support brain

energy metabolism; this has been shown in hypoglycemic mice (Urion et al., 1979).

formed from [13C]glucose, presumably because 13C-labeled transmitter glutamate was still being released from neurons and taken up by astrocytes with subsequent formation of [13C]glutamine. The upper spectrum is from the brain of a mouse that received [1,2- 13C]acetate and [1-13C]glucose. The doubly labeled acetate gives rise to the double peaks in glutamine C4 (peak 7). The lower spectrum is from a mouse, whose astrocytic TCA ycle was blocked by fluoroacetate; here only glutamine labeled from [1-13C]glucose (single peak 7) can be seen (From Hassel et al., 1997, with permission from the publisher).

#### **4. Identification of alternative substrates for brain energy metabolism**

Serum glucose is considered the physiological energy substrate of the brain. Even so, a number of substrates may feed into the glycolytic pathway or the TCA cycle of brain cells. For example, lactate, which was shown by 13C MRS to be a substrate for brain metabolism in mice (Hassel and Bråthe, 2000), was recently shown by 13C MRS to contribute importantly to energy metabolism in the human brain under physiological conditions (Gallagher et al., 2009; Boumezbeur et al., 2010). The ability of a substrate to function as a metabolic substrate for brain cells depends on its ability to cross the blood-brain barrier, i.e. on the presence of specific transporter proteins in the barrier. Further, brain cells must have the necessary transporters to take the substrates up from the extracellular fluid as well as the enzymes necessary for their metabolism. A substrate that is not metabolized in the brain after i.v. injection may prove to be metabolized after intracerebral injection, a finding that points to transport limitations at the blood-brain barrier; this holds true for dicarboxylates like fumarate and malate (Hassel et al., 2002); if injected directly into the brain parenchyma, these substrates are taken up into astrocytes and rapidly metabolized through the TCA cycle to glutamine, indicating the presence of the plasma membrane dicarboxylate transporter predominantly in astrocytes. A similar situation exists for the monosaccharide fructose (Hassel et al., submitted), which hardly crosses the blood-brain barrier at all. This can be shown by giving mice [13C]fructose i.v.; in contrast to liver, the brain does not accumulate the injected [13C]fructose, as can be shown by 13C MRS of brain and liver extracts. However, if fructose is injected into the brain parenchyma, it becomes metabolized. The oxidation of fructose by brain tissue would explain the ability of fructose to support energy-requiring processes, such a axonal activity, in vitro (Meakin et al., 2007).

Even though a substrate crosses the blood-brain barrier, the enzymatic machinery required for its metabolism may not be present in the brain. An example is propionate, which readily crosses the blood-brain barrier, but hardly becomes metabolized in the brain (Nguyen et al., 2007b; Morland et al., in prep). The enzymes required for the initial metabolism of propionate, propionyl-CoA synthetase is expressed at low levels, making it difficult to detect propionate metabolites by 13C MRS after i.v. administration of [13C]propionate to experimental animals. Cerebral metabolism of propionate may be detected with radiolabeled propionate, however, owing to the higher sensitivity of radiodetective methods (scintillation counting) than of 13C MRS (Nguyen et al., 2007b).

With the use of 13C MRS -hydroxybutyrate (Künnecke et al., 1993; Pan et al., 2002; Andrews et al., 2009), lactate (Hassel and Bråthe, 2000; Tyson et al., 2003; Boumezbeur et al., 2010), and pyruvate (Gonzalez et al., 2005) have been shown to be energy substrates for the brain after i.v. injection, pointing to the existence of the appropriate monocarboxylate transporters at the blood-brain barrier. The predominant labeling of glutamate (over glutamine) points to a predominantly neuronal metabolism of these subtrates (Fig. 5). Even octanoate has been shown to be metabolized by the brain (Ebert et al., 2003), suggesting a certain capacity for fatty acid metabolism in the brain; the preferential labeling of glutamine pointed to astrocytes as the main metabolic compartment.

36 Magnetic Resonance Spectroscopy

formed from [13C]glucose, presumably because 13C-labeled transmitter glutamate was still being released from neurons and taken up by astrocytes with subsequent formation of [13C]glutamine. The upper spectrum is from the brain of a mouse that received [1,2-

13C]acetate and [1-13C]glucose. The doubly labeled acetate gives rise to the double peaks in glutamine C4 (peak 7). The lower spectrum is from a mouse, whose astrocytic TCA ycle was blocked by fluoroacetate; here only glutamine labeled from [1-13C]glucose (single peak 7)

Serum glucose is considered the physiological energy substrate of the brain. Even so, a number of substrates may feed into the glycolytic pathway or the TCA cycle of brain cells. For example, lactate, which was shown by 13C MRS to be a substrate for brain metabolism in mice (Hassel and Bråthe, 2000), was recently shown by 13C MRS to contribute importantly to energy metabolism in the human brain under physiological conditions (Gallagher et al., 2009; Boumezbeur et al., 2010). The ability of a substrate to function as a metabolic substrate for brain cells depends on its ability to cross the blood-brain barrier, i.e. on the presence of specific transporter proteins in the barrier. Further, brain cells must have the necessary transporters to take the substrates up from the extracellular fluid as well as the enzymes necessary for their metabolism. A substrate that is not metabolized in the brain after i.v. injection may prove to be metabolized after intracerebral injection, a finding that points to transport limitations at the blood-brain barrier; this holds true for dicarboxylates like fumarate and malate (Hassel et al., 2002); if injected directly into the brain parenchyma, these substrates are taken up into astrocytes and rapidly metabolized through the TCA cycle to glutamine, indicating the presence of the plasma membrane dicarboxylate transporter predominantly in astrocytes. A similar situation exists for the monosaccharide fructose (Hassel et al., submitted), which hardly crosses the blood-brain barrier at all. This can be shown by giving mice [13C]fructose i.v.; in contrast to liver, the brain does not accumulate the injected [13C]fructose, as can be shown by 13C MRS of brain and liver extracts. However, if fructose is injected into the brain parenchyma, it becomes metabolized. The oxidation of fructose by brain tissue would explain the ability of fructose to support energy-requiring

Even though a substrate crosses the blood-brain barrier, the enzymatic machinery required for its metabolism may not be present in the brain. An example is propionate, which readily crosses the blood-brain barrier, but hardly becomes metabolized in the brain (Nguyen et al., 2007b; Morland et al., in prep). The enzymes required for the initial metabolism of propionate, propionyl-CoA synthetase is expressed at low levels, making it difficult to detect propionate metabolites by 13C MRS after i.v. administration of [13C]propionate to experimental animals. Cerebral metabolism of propionate may be detected with radiolabeled propionate, however, owing to the higher sensitivity of radiodetective methods

With the use of 13C MRS -hydroxybutyrate (Künnecke et al., 1993; Pan et al., 2002; Andrews et al., 2009), lactate (Hassel and Bråthe, 2000; Tyson et al., 2003; Boumezbeur et al., 2010), and pyruvate (Gonzalez et al., 2005) have been shown to be energy substrates for the brain after i.v. injection, pointing to the existence of the appropriate monocarboxylate transporters at the blood-brain barrier. The predominant labeling of glutamate (over glutamine) points to

can be seen (From Hassel et al., 1997, with permission from the publisher).

processes, such a axonal activity, in vitro (Meakin et al., 2007).

(scintillation counting) than of 13C MRS (Nguyen et al., 2007b).

**4. Identification of alternative substrates for brain energy metabolism** 

Fig. 5. 13C MR spectra of brain extracts of mice that received [3-13C]pyruvate or [2- 13C]pyruvate i.v. The (labeling of) individual carbon positions in certain metabolites is indicated. Wake mice received 9 mmoles/kg [3-13C]pyruvate (upper panel) or [2- 13C]pyruvate (lower panel) i.v. Survival time was 5 min. [3-13C]Pyruvate labels the C3 of lactate (Lac) and alanine (Ala), the C4 of glutamate (Glu) and glutamine (Gln), and GABA C2 (but not glutamate C5 or the C1 of GABA or lactate; insert). [2-13C]Pyruvate labels lactate and alanine C2, and glutamate C5, GABA C1 and lactate C1 (insert). D is dioxane (internal standard). The peak corresponding to Glu C4 in lower spectrum represents naturally abundant 13C-glutamate. Reprinted from Gonzalez et al., 2005, with permission from the publisher.

13C-Labeled acetate also yields good labeling of cerebral amino acids, predominantly glutamine, after i.v. injection, a finding that points to their metabolism in astrocytes (see above, 3. Visualization of the metabolic interplay between astrocytes and neurons). The labeling of amino acids in the brain from [13C]acetate is on a level similar to that achieved with [13C]glucose (Hassel et al., 1995a;1997), which suggests that acetate could support brain energy metabolism; this has been shown in hypoglycemic mice (Urion et al., 1979).

13C Magnetic Resonance Spectroscopy in Neurobiology - Its Use in Monitoring

take up glutamate or GABA from the synaptic cleft.

or HCO3-

represented in italics.

Brain Energy Metabolism and in Identifying Novel Metabolic Substrates and Metabolic Pathways 39

rise to 13C-labeled pyruvate in the brain (glucose, pyruvate, lactate, alanine, glycerol) may enter the TCA cycle through two different enzymatic routes. The first, which gives rise to acetyl-CoA, is pyruvate dehydrogenase, which removes a carboxylic group from pyruvate in the form of CO2. The second route involves the addition of a carboxylic group (from CO2

part of Fig. 7); these reactions are known as pyruvate carboxylation. In astrocytes, the reaction is catalyzed by pyruvate carboxylase, in neurons, it may be catalyzed by malic enzyme. These reactions are 'anaplerotic', meaning that they function to 'fill up' the TCA cycle with intermediates. They are thought to compensate for the loss of α-ketoglutarate inherent in export of glutamine from astrocytes and for the similar loss of α-ketoglutarate inherent in glutamatergic and GABAergic neurotransmission when (See below) astrocytes

Fig. 7. Reversible pyruvate carboxylation and equilibration of malate with fumarate, which

carboxylation to [1-13C]malate, presumably by malic enzyme. [1-13C]Malate equilibrates with

explains rapid formation of 13CO2 from [1-13C]pyruvate. [1-13C]Pyruvate undergoes

the symmetrical fumarate, so that both [1-13C]malate and [4-13C]malate are formed. Decarboxylation of these malates produces 13CO2 as well as unlabeled CO2. The 13C is

) to pyruvate to produce oxaloacetate or its immediate precursor malate (See upper

An interesting example of the power of 13C MRS in elucidating cell-specific metabolism in the CNS is the avid labeling of GABA from [13C]glycerol (Nguyen et al., 2003); this labeling greatly exceeded that of glutamate and glutamine (Fig. 6), pointing to GABAergic neurons as the predominant cell compartment for glycerol metabolism. A later 13C MRS study showed that glycerol is being produced from glucose by both neurons and astrocytes and that this production increases during hypoxia or epileptic seizures (Nguyen et al., 2007c). The two findings put together (formation of glycerol in both neurons and astrocytes and its metabolism by GABAergic neurons) could point to a 'glycerol cycle' in the brain.

Fig. 6. 13C MR spectra of brain extracts from mice that received [U- 13C]glycerol i.v. Wake mice received 150 µmol [U-13C]glycerol and were killed at 5 or 15 min. Peak numbers: 1, lactate C3; 2, glutamate C4; 3, GABA C2; 4, glycerol C1 + C3. \*Internal standard (dioxane). Note the strong labeling of GABA C2 relative to glutamate C4 at 15 min after injection of [U-13C]glycerol in comparison with the much higher labeling of glutamate from 13C-labeled glucose or pyruvate (Figs. 1 and 5). Reprinted from Nguyen et al., 2003, with permission from the publisher.

#### **5. Identification of metabolic pathways from the 13C labeling of specific carbon positions in glutamate and glutamine**

In the second paragraph of this paper ('13C MRS for the visualization of brain energy metabolism') the 13C labeling of glutamate from [13C]glucose through TCA cycle activity is described. In fact, any 13C-labeled substrate that gives rise to 13C-labeled acetyl-CoA will label glutamate and glutamine in positions that reflect the entry of label into the TCA cycle and its subsequent turning in the cycle. This applies to 13C-labeled lactate, pyruvate, alanine, beta-hydroxybutyrate, acetate, butyrate, glycerol, and others. However, substrates that give

An interesting example of the power of 13C MRS in elucidating cell-specific metabolism in the CNS is the avid labeling of GABA from [13C]glycerol (Nguyen et al., 2003); this labeling greatly exceeded that of glutamate and glutamine (Fig. 6), pointing to GABAergic neurons as the predominant cell compartment for glycerol metabolism. A later 13C MRS study showed that glycerol is being produced from glucose by both neurons and astrocytes and that this production increases during hypoxia or epileptic seizures (Nguyen et al., 2007c). The two findings put together (formation of glycerol in both neurons and astrocytes and its

metabolism by GABAergic neurons) could point to a 'glycerol cycle' in the brain.

Fig. 6. 13C MR spectra of brain extracts from mice that received [U- 13C]glycerol i.v. Wake mice received 150 µmol [U-13C]glycerol and were killed at 5 or 15 min. Peak numbers: 1, lactate C3; 2, glutamate C4; 3, GABA C2; 4, glycerol C1 + C3. \*Internal standard (dioxane). Note the strong labeling of GABA C2 relative to glutamate C4 at 15 min after injection of [U-13C]glycerol in comparison with the much higher labeling of glutamate from 13C-labeled glucose or pyruvate (Figs. 1 and 5). Reprinted from Nguyen et al., 2003, with permission

**5. Identification of metabolic pathways from the 13C labeling of specific** 

In the second paragraph of this paper ('13C MRS for the visualization of brain energy metabolism') the 13C labeling of glutamate from [13C]glucose through TCA cycle activity is described. In fact, any 13C-labeled substrate that gives rise to 13C-labeled acetyl-CoA will label glutamate and glutamine in positions that reflect the entry of label into the TCA cycle and its subsequent turning in the cycle. This applies to 13C-labeled lactate, pyruvate, alanine, beta-hydroxybutyrate, acetate, butyrate, glycerol, and others. However, substrates that give

**carbon positions in glutamate and glutamine** 

from the publisher.

rise to 13C-labeled pyruvate in the brain (glucose, pyruvate, lactate, alanine, glycerol) may enter the TCA cycle through two different enzymatic routes. The first, which gives rise to acetyl-CoA, is pyruvate dehydrogenase, which removes a carboxylic group from pyruvate in the form of CO2. The second route involves the addition of a carboxylic group (from CO2 or HCO3- ) to pyruvate to produce oxaloacetate or its immediate precursor malate (See upper part of Fig. 7); these reactions are known as pyruvate carboxylation. In astrocytes, the reaction is catalyzed by pyruvate carboxylase, in neurons, it may be catalyzed by malic enzyme. These reactions are 'anaplerotic', meaning that they function to 'fill up' the TCA cycle with intermediates. They are thought to compensate for the loss of α-ketoglutarate inherent in export of glutamine from astrocytes and for the similar loss of α-ketoglutarate inherent in glutamatergic and GABAergic neurotransmission when (See below) astrocytes take up glutamate or GABA from the synaptic cleft.

Fig. 7. Reversible pyruvate carboxylation and equilibration of malate with fumarate, which explains rapid formation of 13CO2 from [1-13C]pyruvate. [1-13C]Pyruvate undergoes carboxylation to [1-13C]malate, presumably by malic enzyme. [1-13C]Malate equilibrates with the symmetrical fumarate, so that both [1-13C]malate and [4-13C]malate are formed. Decarboxylation of these malates produces 13CO2 as well as unlabeled CO2. The 13C is represented in italics.

13C Magnetic Resonance Spectroscopy in Neurobiology - Its Use in Monitoring

markedly (Shank et al., 1993).

2011; Park et al., 2011).

393.

**7. References** 

Brain Energy Metabolism and in Identifying Novel Metabolic Substrates and Metabolic Pathways 41

substrates has been done prior to i.v. injection. The MRS signal may thus be enhanced more than 10,000 times (Ardenkjaer-Larsen et al., 2003). In some studies pyruvate, which is labeled in the carboxylic position, [1-13C]pyruvate, has been used (Hurd et al., 2010; Marjańska et al., 2010; Mayer et al., 2011). The [13C]pyruvate is hyperpolarized immediately before i.v. injection. The hyperpolarization is lost within some seconds, however, and the detection of metabolism of pyruvate is restricted to 13C-labeled lactate, alanine, and CO2 (see Fig. 2). Formation of 13CO2 has been interpreted to reflect pyruvate dehydrogenase activity, i.e. the formation of acetyl-CoA, which may enter the TCA cycle to form citrate. But in the brain the activity of pyruvate dehydrogenase is quite low (Morland et al., 2007), and formation of TCA cycle intermediates and related amino acids takes several minutes (Gonzalez et al., 2005), so a more likely explanation for the formation of 13CO2 within a few seconds is reversible pyruvate carboxylation (Fig. 3): [13C]pyruvate is carboxylated to [13C]malate, which equilibrates with (the symmetrical) fumarate, so that 13C is distributed evenly between the two carboxylic groups of fumarate. [13C]Fumarate equilibrates with malate, which may become decarboxylated to pyruvate and CO2. Some of this CO2 will then be 13C-labeled. This series of reactions has been described in the brain with the use of [2- 13C]pyruvate and [3-13C]pyruvate (Gonzalez et al., 2005), and they occur much more rapidly than the pyruvate dehydrogenase reaction. Another issue in studies that use hyperpolarized substrates is the use of anesthesia, which reduces brain metabolism of 13C-labeled substrates

Hyperpolarized [13C]pyruvate may prove useful in the diagnostic workup of malignant tumors, including brain tumors, which show a greater tendency to convert [13C]pyruvate into [13C]lactate than the normal surrounding tissue. Some recent studies even suggest the possibility of monitoring tumor response to therapy from the reduced conversion of pyruvate into lactate within a day's time after irradiation or chemotherapy (Day et al., 2007;

Andrews MT, Russeth KP, Drewes LR, Henry PG. (2009) Adaptive mechanisms regulate

Ardenkjaer-Larsen JH, Fridlund B, Gram A, Hansson G, Hansson L, Lerche MH, Servin R,

Attwell D, Laughlin SB. (2001) An energy budget for signaling in the grey matter of the

Bartnik BL, Lee SM, Hovda DA, Sutton RL. (2007) The fate of glucose during the period of

Bergersen LH. (2007) Is lactate food for neurons? Comparison of monocarboxylate

Boumezbeur F, Petersen KF, Cline GW, Mason GF, Behar KL, Shulman GI, Rothman DL.

liquid-state NMR. Proc. Natl. Acad. Sci. U. S. A. 100:10158-10163.

transporter subtypes in brain and muscle. Neuroscience 145:11-19.

brain. J. Cereb. Blood Flow Metab. 21:1133-1145.

Neurotrauma. 24:1079-1092.

preferred utilization of ketones in the heart and brain of a hibernating mammal during arousal from torpor. Am. J. Physiol. Regul. Integr. Comp. Physiol. 296:R383-

Thaning M, Golman K. (2003) Increase in signal-to-noise ratio of > 10,000 times in

decreased metabolism after fluid percussion injury: a 13C NMR study. J.

(2010) The contribution of blood lactate to brain energy metabolism in humans

When pyruvate is 13C-labeled these reactions lead to the formation of 13C-labeled malate and oxaloacetate. Pyruvate labeled in the 3rd carbon (which originates from glucose labeled in the 1st position, [1-13C]glucose) leads to formation of malate or oxaloacetate labeled in the 3rd carbon. This carbon position corresponds to the 2nd carbon in glutamate and glutamine. In several studies the greater labeling of the 2nd than the 3rd carbon in glutamine has been taken as evidence of pyruvate carboxylase activity in astrocytes (Shank et al., 1993; Hassel et al., 1995a; Serres et al., 2008). A similar preferential labeling of the 2nd over the 3rd carbon in glutamate has not been consistently found, apparently pointing to the absence of pyruvate carboxylation in neurons. However, oxaloacetate and malate tend to equilibrate with the symmetrical fumarate, leading to scrambling of label between the two carbon position. 13C MRS evidence of such scrambling in neurons has been reported in both cultured neurons (Merle et al., 1996) and in vivo (Hassel et al., 2000; Gonzalez et al., 2005). Another way to study pyruvate carboxylation is to use glucose labeled in the 2nd position. Glucose labeled in this position ([2-13C]glucose) does not label acetyl-CoA, but it may label the 2nd and 3rd positions in glutamate and glutamine through pyruvate carboxylation. Mason et al (2007) showed in awake human subjects the flux through pyruvate carboxylation to be approximately 6% of that through pyruvate dehydrogenase (the activity of which was studied with [1-13C]glucose). This value was in good agreement with previous 13C MRS studies in rats (Shank et al., 1993) and with previous estimates based on radiolabeling studies (Van den Berg, 1973).

Formation of pyruvate from TCA cycle intermediates malate or oxaloacetate (through the decarboxylating activities of malic enzyme and phosphoenolpyruvate carboxykinase, respectively) is termed pyruvate recycling. This process takes place in the brain (Cerdan et al., 1990; Hassel and Sonnewald, 1995b; Cruz et al., 1998) as it does in liver. In one study this process was identified in brain by 13C MRS of brain extracts from rats that had received [13C]acetate, which labeled lactate (Hassel and Sonnewald, 1995b). Such labeling can only occur through formation of pyruvate from TCA cycle intermediates. When pyruvate thus formed (and labeled) enters the TCA cycle through pyruvate dehydrogenase a distinct labeling pattern may be seen in glutamate (Cerdan et al., 1990; Håberg et al., 1998). Valid criticism of the interpretation of previous studies has come from Serres et al. (2008) who pointed out that hepatic gluconeogenesis with 13C labeling of serum glucose from [13C]acetate may influence findings in the brain. However, pyruvate recycling also occurs in cultured brain cells (astrocytes), in which hepatic gluconeogenesis is not an issue (Sonnewald et al., 1996).

With the use of 13C MRS Gruetter and colleagues were able to determine the concentration and turnover of glycogen in the intact brain (van Heeswijk et al., 2010), and they also described the formation of N-acetyl-aspartate, which is present in the brain at high concentrations, from [13C]glucose, and they determined the turnover rate of this presumed 'inert' compound (Choi et al., 2004).

#### **6. The use of hyperpolarization of 13C to increase MRS sensitivity**

A problem with 13C MRS is the low sensitivity of the method. This weakness implies that 13C-labeled metabolic substrates have to be given in large amounts to achieve detection of their metabolites. In contrast, radiolabeling studies are done with minute quantities of substrate. In an attempt to improve 13C MRS sensitivity, hyperpolarization of 13C-labeled substrates has been done prior to i.v. injection. The MRS signal may thus be enhanced more than 10,000 times (Ardenkjaer-Larsen et al., 2003). In some studies pyruvate, which is labeled in the carboxylic position, [1-13C]pyruvate, has been used (Hurd et al., 2010; Marjańska et al., 2010; Mayer et al., 2011). The [13C]pyruvate is hyperpolarized immediately before i.v. injection. The hyperpolarization is lost within some seconds, however, and the detection of metabolism of pyruvate is restricted to 13C-labeled lactate, alanine, and CO2 (see Fig. 2). Formation of 13CO2 has been interpreted to reflect pyruvate dehydrogenase activity, i.e. the formation of acetyl-CoA, which may enter the TCA cycle to form citrate. But in the brain the activity of pyruvate dehydrogenase is quite low (Morland et al., 2007), and formation of TCA cycle intermediates and related amino acids takes several minutes (Gonzalez et al., 2005), so a more likely explanation for the formation of 13CO2 within a few seconds is reversible pyruvate carboxylation (Fig. 3): [13C]pyruvate is carboxylated to [13C]malate, which equilibrates with (the symmetrical) fumarate, so that 13C is distributed evenly between the two carboxylic groups of fumarate. [13C]Fumarate equilibrates with malate, which may become decarboxylated to pyruvate and CO2. Some of this CO2 will then be 13C-labeled. This series of reactions has been described in the brain with the use of [2- 13C]pyruvate and [3-13C]pyruvate (Gonzalez et al., 2005), and they occur much more rapidly than the pyruvate dehydrogenase reaction. Another issue in studies that use hyperpolarized substrates is the use of anesthesia, which reduces brain metabolism of 13C-labeled substrates markedly (Shank et al., 1993).

Hyperpolarized [13C]pyruvate may prove useful in the diagnostic workup of malignant tumors, including brain tumors, which show a greater tendency to convert [13C]pyruvate into [13C]lactate than the normal surrounding tissue. Some recent studies even suggest the possibility of monitoring tumor response to therapy from the reduced conversion of pyruvate into lactate within a day's time after irradiation or chemotherapy (Day et al., 2007; 2011; Park et al., 2011).

#### **7. References**

40 Magnetic Resonance Spectroscopy

When pyruvate is 13C-labeled these reactions lead to the formation of 13C-labeled malate and oxaloacetate. Pyruvate labeled in the 3rd carbon (which originates from glucose labeled in the 1st position, [1-13C]glucose) leads to formation of malate or oxaloacetate labeled in the 3rd carbon. This carbon position corresponds to the 2nd carbon in glutamate and glutamine. In several studies the greater labeling of the 2nd than the 3rd carbon in glutamine has been taken as evidence of pyruvate carboxylase activity in astrocytes (Shank et al., 1993; Hassel et al., 1995a; Serres et al., 2008). A similar preferential labeling of the 2nd over the 3rd carbon in glutamate has not been consistently found, apparently pointing to the absence of pyruvate carboxylation in neurons. However, oxaloacetate and malate tend to equilibrate with the symmetrical fumarate, leading to scrambling of label between the two carbon position. 13C MRS evidence of such scrambling in neurons has been reported in both cultured neurons (Merle et al., 1996) and in vivo (Hassel et al., 2000; Gonzalez et al., 2005). Another way to study pyruvate carboxylation is to use glucose labeled in the 2nd position. Glucose labeled in this position ([2-13C]glucose) does not label acetyl-CoA, but it may label the 2nd and 3rd positions in glutamate and glutamine through pyruvate carboxylation. Mason et al (2007) showed in awake human subjects the flux through pyruvate carboxylation to be approximately 6% of that through pyruvate dehydrogenase (the activity of which was studied with [1-13C]glucose). This value was in good agreement with previous 13C MRS studies in rats (Shank et al., 1993) and with

previous estimates based on radiolabeling studies (Van den Berg, 1973).

(Sonnewald et al., 1996).

'inert' compound (Choi et al., 2004).

Formation of pyruvate from TCA cycle intermediates malate or oxaloacetate (through the decarboxylating activities of malic enzyme and phosphoenolpyruvate carboxykinase, respectively) is termed pyruvate recycling. This process takes place in the brain (Cerdan et al., 1990; Hassel and Sonnewald, 1995b; Cruz et al., 1998) as it does in liver. In one study this process was identified in brain by 13C MRS of brain extracts from rats that had received [13C]acetate, which labeled lactate (Hassel and Sonnewald, 1995b). Such labeling can only occur through formation of pyruvate from TCA cycle intermediates. When pyruvate thus formed (and labeled) enters the TCA cycle through pyruvate dehydrogenase a distinct labeling pattern may be seen in glutamate (Cerdan et al., 1990; Håberg et al., 1998). Valid criticism of the interpretation of previous studies has come from Serres et al. (2008) who pointed out that hepatic gluconeogenesis with 13C labeling of serum glucose from [13C]acetate may influence findings in the brain. However, pyruvate recycling also occurs in cultured brain cells (astrocytes), in which hepatic gluconeogenesis is not an issue

With the use of 13C MRS Gruetter and colleagues were able to determine the concentration and turnover of glycogen in the intact brain (van Heeswijk et al., 2010), and they also described the formation of N-acetyl-aspartate, which is present in the brain at high concentrations, from [13C]glucose, and they determined the turnover rate of this presumed

A problem with 13C MRS is the low sensitivity of the method. This weakness implies that 13C-labeled metabolic substrates have to be given in large amounts to achieve detection of their metabolites. In contrast, radiolabeling studies are done with minute quantities of substrate. In an attempt to improve 13C MRS sensitivity, hyperpolarization of 13C-labeled

**6. The use of hyperpolarization of 13C to increase MRS sensitivity** 


13C Magnetic Resonance Spectroscopy in Neurobiology - Its Use in Monitoring

Effects of fluorocitrate. Neurochem. Res. 20:413-420.

fluoroacetate. J. Cereb. Blood Flow Metab. 17:1230-1238.

pyruvate recycling in brain. Dev. Neurosci. 20:389-398.

occlusion in rat. J. Neurochem. 109 Suppl 1:174-181.

activated glutaminase. J. Neurosci. Res. 66:951-958.

astrocytic oxidative metabolism. J. Neurosci. 22:1523-1531.

pyruvate. Magn. Reson. Med. 63:1137-1143.

pyruvate carboxylation. J. Cereb. Blood Flow Metab. 20:327-336.

1191.

pp. 267-290.

12:3729-3732.

Biomed. 6:264-277.

Neurochem. 65:2227-2234.

Brain Energy Metabolism and in Identifying Novel Metabolic Substrates and Metabolic Pathways 43

Hassel B, Westergaard N, Schousboe A, Fonnum F. (1995b) Metabolic differences between

Hassel B, Sonnewald U. (1995a) Selective inhibition of the tricarboxylic acid cycle of

Hassel B, Sonnewald U. (1995b) Glial formation of pyruvate and lactate from TCA cycle

Hassel B, Bachelard H, Jones P, Fonnum F, Sonnewald U. (1997) Trafficking of amino acids

Hassel B, Bråthe A. (2000) Cerebral metabolism of lactate in vivo: evidence for neuronal

Hassel B, Dingledine R. (2006) Glutamate. In GJ Siegel (ed) *Basic Neurochemistry*. Elsevier,

Hurd RE, Yen YF, Mayer D, Chen A, Wilson D, Kohler S, Bok R, Vigneron D, Kurhanewicz

Hyder F, Chase JR, Behar KL, Mason GF, Siddeek M, Rothman DL, Shulman RG. (1996)

Håberg AK, Qu H, Sonnewald U. (2009) Acute changes in intermediary metabolism in

Johannessen CU, Qu H, Sonnewald U, Hassel B, Fonnum F. (2001) Estimation of aspartate

Künnecke B, Cerdan S, Seelig J. (1993) Cerebral metabolism of [1,2-13C2]glucose and [U-

Kvamme E, Torgner IA, Roberg B. (2001) Kinetics and localization of brain phosphate

Lebon V, Petersen KF, Cline GW, Shen J, Mason GF, Dufour S, Behar KL, Shulman GI,

detected with 1H[13C]NMR. Proc. Natl. Acad. Sci. U. S. A. 93:7612-7617. Hyder F, Brown P, Nixon TW, Behar KL. (2003) Mapping cerebral glutamate 13C turnover and oxygen consumption by in vivo NMR. Adv. Exp. Med. Biol. 530:29-39. Håberg A, Qu H, Bakken IJ, Sande LM, White LR, Haraldseth O, Unsgård G, Aasly J,

primary cultures of astrocytes and neurons from cerebellum and cerebral cortex.

GABAergic neurons with 3-nitropropionic acid in vivo. J. Neurochem. 65:1184-

intermediates: implications for the inactivation of transmitter amino acids? J.

between neurons and glia in vivo. Effects of inhibition of glial metabolism by

J, Tropp J, Spielman D, Pfefferbaum A. (2010) Metabolic imaging in the anesthetized rat brain using hyperpolarized [1-13C] pyruvate and [1-13C] ethyl

Increased tricarboxylic acid cycle flux in rat brain during forepaw stimulation

Sonnewald U. (1998) In vitro and ex vivo 13C-NMR spectroscopy studies of

cerebellum and contralateral hemisphere following middle cerebral artery

synthesis in GABAergic neurons in mice by 13 C NMR spectroscopy. Neuroreport

13C4]3-hydroxybutyrate in rat brain as detected by 13C NMR spectroscopy. NMR

Rothman DL. (2002) Astroglial contribution to brain energy metabolism in humans revealed by 13C nuclear magnetic resonance spectroscopy: elucidation of the dominant pathway for neurotransmitter glutamate repletion and measurement of

measured by dynamic 13C nuclear magnetic resonance spectroscopy. J. Neurosci. 30:13983-13991.


Cerdan S, Künnecke B, Seelig J. (1990) Cerebral metabolism of [1,2-13C2]acetate as detected

Choi IY, Gruetter R. (2004) Dynamic or inert metabolism? Turnover of N-acetyl aspartate

Cruz F, Scott SR, Barroso I, Santisteban P, Cerdán S. (1998) Ontogeny and cellular

Day SE, Kettunen MI, Gallagher FA, Hu DE, Lerche M, Wolber J, Golman K, Ardenkjaer-

Day SE, Kettunen MI, Cherukuri MK, Mitchell JB, Lizak MJ, Morris HD, Matsumoto S,

Deelchand DK, Shestov AA, Koski DM, Uğurbil K, Henry PG. (2009) Acetate transport and

Ebert D, Haller RG, Walton ME. (2003) Energy contribution of octanoate to intact rat brain

Eloqayli H, Dahl CB, Götestam KG, Unsgård G, Sonnewald U. (2004) Changes of glial-

Fitzpatrick SM, Hetherington HP, Behar KL, Shulman RG. (1990) The flux from glucose to

Fonnum F, Storm-Mathisen J, Walberg F. (1970) Glutamate decarboxylase in inhibitory

Gallagher CN, Carpenter KL, Grice P, Howe DJ, Mason A, Timofeev I, Menon DK,

high-resolution nuclear magnetic resonance study. Brain 132:2839-2849. Gonzalez SV, Nguyen NH, Rise F, Hassel B. (2005) Brain metabolism of exogenous

Hassel B, Paulsen RE, Johnsen A, Fonnum F. (1992) Selective inhibition of glial cell

Hassel B, Sonnewald U, Fonnum F. (1995a) Glial-neuronal interactions as studied by

metabolism in vivo by fluorocitrate. Brain Res. 576:120-124.

spectroscopic study. J. Neurochem. 64:2773-2782.

by in vivo and in vitro 13C NMR. J. Biol. Chem. 265:12916-12926.

Danbolt NC. (2001) Glutamate uptake. Prog. Neurobiol. 65:1-105.

spectroscopic imaging. Magn. Reson. Med. 65:557-563.

pentylenetetrazole. Neurochem. Int. 45:739-745.

pyruvate. J. Neurochem. 95:284-293.

spectroscopy. J. Cereb. Blood Flow Metab. 10:170-179.

utilization in the rat brain. J. Neurochem. 109 Suppl 1:46-54.

30:13983-13991.

91:778-787.

13:1382-1387.

23:5928-5935.

Res. 20:259-275.

2619.

measured by dynamic 13C nuclear magnetic resonance spectroscopy. J. Neurosci.

and glutathione from D-[1-13C]glucose in the rat brain in vivo. J. Neurochem.

localization of the pyruvate recycling system in rat brain. J. Neurochem. 70:2613-

Larsen JH, Brindle KM. (2007) Detecting tumor response to treatment using hyperpolarized 13C magnetic resonance imaging and spectroscopy. Nat. Med.

Koretsky AP, Brindle KM. (2011) Detecting response of rat C6 glioma tumors to radiotherapy using hyperpolarized [1- 13C]pyruvate and 13C magnetic resonance

metabolism measured by 13C nuclear magnetic resonance spectroscopy. J. Neurosci.

neuronal interaction and metabolism after a subconvulsive dose of

glutamate in the rat brain in vivo as determined by 1H-observed, 13C-edited NMR

neurons. A study of the enzyme in Purkinje cell axons and boutons in the cat. Brain

Kirkpatrick PJ, Pickard JD, Sutherland GR, Hutchinson PJ. (2009) The human brain utilizes lactate via the tricarboxylic acid cycle: a 13C-labelled microdialysis and

cerebral metabolism of [2-13C]acetate and [1-13C]glucose: an ex vivo 13C NMR


13C Magnetic Resonance Spectroscopy in Neurobiology - Its Use in Monitoring

Metab. 22:890-898.

985.

53:631-645.

Rev. Physiol. 65:401-27.

Neurochem. 109 Suppl 1:189-197.

Cereb. Blood Flow Metab. 28:712-724.

J. Neurochem. 101:1428-1240.

Neurochem. 61:315-323.

67:2566-2572.

Acad. Sci. U. S. A. 96:8235-8240.

Siesjö BK. (1978) *Brain Energy Metabolism*, New York.

Brain Energy Metabolism and in Identifying Novel Metabolic Substrates and Metabolic Pathways 45

Pan JW, de Graaf RA, Petersen KF, Shulman GI, Hetherington HP, Rothman DL. (2002) [2,4-

Park I, Bok R, Ozawa T, Phillips JJ, James CD, Vigneron DB, Ronen SM, Nelson SJ. (2011)

Petroff OA, Errante LD, Rothman DL, Kim JH, Spencer DD. (2002) Glutamate-glutamine

Rodrigues TB, Fonseca CP, Castro MM, Cerdán S, Geraldes CF. (2009) 13C NMR tracers in

Rothman DL, Sibson NR, Hyder F, Shen J, Behar KL, Shulman RG. (1999) In vivo nuclear

Rothman DL, Behar KL, Hyder F, Shulman RG. (2003) In vivo NMR studies of the glutamate

Scafidi S, O'Brien J, Hopkins I, Robertson C, Fiskum G, McKenna M. (2009) Delayed cerebral

Serres S, Bezancon E, Franconi JM, Merle M. (2007) Brain pyruvate recycling and peripheral

Serres S, Raffard G, Franconi JM, Merle M. (2008) Close coupling between astrocytic and

Shank RP, Leo GC, Zielke HR. (1993) Cerebral metabolic compartmentation as revealed by

Shen J, Petersen KF, Behar KL, Brown P, Nixon TW, Mason GF, Petroff OA, Shulman GI,

Shulman RG, Rothman DL, Behar KL, Hyder F. (2004) Energetic basis of brain activity:

Sonnewald U, Westergaard N, Jones P, Taylor A, Bachelard HS, Schousboe A. (1996)

implications for neuroimaging. Trends Neurosci. 27:489-495.

cycling in the epileptic human hippocampus. Epilepsia 43:703-710.

Philos. Trans. R. Soc. Lond. B Biol. Sci. 354:1165-1177.

13C2 ]-beta-Hydroxybutyrate metabolism in human brain. J. Cereb. Blood Flow

Detection of early response to temozolomide treatment in brain tumors using hyperpolarized 13C MR metabolic imaging. J. Magn. Reson. Imaging 33:1284-1290. Patel AB, de Graaf RA, Mason GF, Kanamatsu T, Rothman DL, Shulman RG, Behar KL.

(2004) Glutamatergic neurotransmission and neuronal glucose oxidation are coupled during intense neuronal activation. J. Cereb. Blood Flow Metab. 24:972-

neurochemistry: implications for molecular imaging. Q. J. Nucl. Med. Mol. Imaging

magnetic resonance spectroscopy studies of the relationship between the glutamate-glutamine neurotransmitter cycle and functional neuroenergetics.

neurotransmitter flux and neuroenergetics: implications for brain function. Annu.

oxidative glucose metabolism after traumatic brain injury in young rats. J.

metabolism: an NMR analysis ex vivo of acetate and glucose metabolism in the rat.

neuronal metabolisms to fulfill anaplerotic and energy needs in the rat brain. J.

nuclear magnetic resonance analysis of D-[1-13C]glucose metabolism. J.

Shulman RG, Rothman DL. (1999) Determination of the rate of the glutamate/glutamine cycle in the human brain by in vivo 13C NMR. Proc. Natl.

Metabolism of [U-13C5] glutamine in cultured astrocytes studied by NMR spectroscopy: first evidence of astrocytic pyruvate recycling. J. Neurochem.


Marjańska M, Iltis I, Shestov AA, Deelchand DK, Nelson C, Uğurbil K, Henry PG. (2010) In

Martinez-Hernandez A, Bell KP, Norenberg MD. (1977) Glutamine synthetase: glial

Mason GF, Pan JW, Chu WJ, Newcomer BR, Zhang Y, Orr R, Hetherington HP. (1999)

Mason GF, Petersen KF, de Graaf RA, Shulman GI, Rothman DL. (2007) Measurements of

Mayer D, Yen YF, Takahashi A, Josan S, Tropp J, Rutt BK, Hurd RE, Spielman DM,

Merle M, Martin M, Villégier A, Canioni P. (1996) [1-13C]glucose metabolism in brain cells:

Morland C, Henjum S, Iversen EG, Skrede KK, Hassel B. (2007) Evidence for a higher

Nguyen NH, Bråthe A, Hassel B. (2003) Neuronal uptake and metabolism of glycerol and

Nguyen N, Gonzalez SV, Rise F, Hassel B. (2007a) Cerebral metabolism of glucose and

Nguyen NH, Morland C, Gonzalez SV, Rise F, Storm-Mathisen J, Gundersen V, Hassel B.

Ottersen OP, Storm-Mathisen J. (1985) Different neuronal localization of aspartate-like and

Pan JW, Williamson A, Cavus I, Hetherington HP, Zaveri H, Petroff OA, Spencer DD. (2008)

Neurometabolism in human epilepsy. Epilepsia. 49 Suppl 3:31-41.

spectroscopy and [1-13C] and [2-13C] glucose. J. Neurochem. 100:73-86. Mason GF, Petersen KF, Lebon V, Rothman DL, Shulman GI. (2006) Increased brain

13C]pyruvate. J. Magn. Reson. 206:210-218.

localization in brain. Science 195:1356-1358.

insert. Magn. Reson. Med. 65:1228-1233.

granule cells. Dev. Neurosci. 18:460-468.

Neurochem. Int. 50:703-709.

study. Neurotoxicology 28:13-18.

aminobutyrate. Neuroscience 16:589-606.

Neurochem. 85:831-842.

101:1694-1700.

Metab. 19:1179-1188.

934.

vivo 13C spectroscopy in the rat brain using hyperpolarized [1-13C]pyruvate and [2-

Measurement of the tricarboxylic acid cycle rate in human grey and white matter in vivo by 1H-[13C] magnetic resonance spectroscopy at 4.1T. J. Cereb. Blood Flow

the anaplerotic rate in the human cerebral cortex using 13C magnetic resonance

monocarboxylic acid transport and utilization in type 1 diabetes. Diabetes 55:929-

Pfefferbaum A. (2011) Dynamic and high-resolution metabolic imaging of hyperpolarized [1-13C]-pyruvate in the rat brain using a high-performance gradient

isotopomer analysis of glutamine from cerebellar astrocytes and glutamate from

glycolytic than oxidative metabolic activity in white matter of rat brain.

the neuronal expression of mitochondrial glycerol-3-phosphate dehydrogenase. J.

pyruvate in soman poisoning. A 13C nuclear magnetic resonance spectroscopic

(2007b) Propionate increases neuronal histone acetylation, but is metabolized oxidatively by glia. Relevance for propionic acidemia. J. Neurochem. 101:806-814. Nguyen NH, Gonzalez SV, Hassel B. (2007c) Formation of glycerol from glucose in rat brain

and cultured brain cells. Augmentation with kainate or ischemia. J. Neurochem.

glutamate-like immunoreactivities in the hippocampus of rat, guinea-pig and Senegalese baboon (Papio papio), with a note on the distribution of gamma-


**3** 

*Austria* 

Fahmy Aboul-Enein

*SMZ-Ost Donauspital, Department of Neurology* 

**MR Spectroscopy in Multiple Sclerosis - A New** 

*Systems do not exist in Nature but only in man's minds.1*

In the late '80s magnetic resonance imaging (MRI) and magnetic resonance spectroscopy (MRS) revolutionized the care and monitoring of patients with multiple sclerosis (MS). For the first time detailed, high resolution images of the brain, spinal cord and lesions could be made by MRI *in vivo*, and followed over time. However, the changes detectable on T1- or T2-weighted magnetic resonance images are non-specific for the underlying pathology, especially in MS patients, and slow down the enthusiasm. They may reflect edema, demyelination, axonal loss, inflammation, gliosis, remyelination, or Wallerian

Similar it is with MRS. Early on it was suggested that metabolites within the tissue may reflect certain cellular compartments and tissue conditions in health and disease. First reports about *in vivo* detection of brain metabolites in humans were published in 1989 (Bruhn et al., 1989; Frahm et al., 1989a; 1989b). Combined with conventional MRI technique spatial anatomical and chemical information about the analysed tissue may be achieved. MRS was approved by the FDA in 1996. However, only few metabolites are detectable (table 1), limiting the main applications of MRS yet to (1) brain tumors (low grade vs. high grade gliomas, metastasis, lymphoma, meningeoma), (2) follow-up of tumors (under chemotherapy and radiotherapy), (2) infectious, mostly focal CNS processes such as pyogenic abscesses, toxoplasmosis, tuberculosis, cryptococcosis, and (3) hepatic encephalopathy, other metabolic disorders, inborn errors of metabolism or hypoxic encephalopathy. In some selected MS cases MRS might be useful to differentiate tumor-like MS brain lesions or Balo-like MS lesions from tumor lesions (table 1) or confluent MS lesions

Well characterized peaks obtained at long echo times (more than TE 135ms, highlighted in grey) are: (1) at 2.0 ppm ('NAA-peak'; N-acetyl-aspartate (NAA)), (2) at 3.0 ppm ('Cr-peak'; creatine (Cr)), (3) at 3.2 ppm ('Cho-peak'; choline (cho)), (4) at 1.3 ppm ('Lac-peak'; lactate

**1. Introduction** 

degeneration in MS.

from leucodystrophies or leucoencephalopathies.

(Lac)), (5) at 1.48ppm ('Alanine Peak'; alanine (Ala)),

<sup>1</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

**Piece of the Puzzle or Just a New Puzzle** 


## **MR Spectroscopy in Multiple Sclerosis - A New Piece of the Puzzle or Just a New Puzzle**

Fahmy Aboul-Enein

*SMZ-Ost Donauspital, Department of Neurology Austria* 

*Systems do not exist in Nature but only in man's minds.1*

#### **1. Introduction**

46 Magnetic Resonance Spectroscopy

Tyson RL, Gallagher C, Sutherland GR. (2003) 13C-Labeled substrates and the cerebral metabolic compartmentalization of acetate and lactate. Brain Res. 992:43-52. Urion D, Vreman HJ, Weiner MW. (1979) Effect of acetate on hypoglycemic seizures in mice.

Van den Berg, C. J. (1973) *Metabolic compartmentation in the brain* (Balázs R and Cremer JE,

van Heeswijk RB, Morgenthaler FD, Xin L, Gruetter R. (2010) Quantification of brain

Waniewski RA, Martin DL. (1998) Preferential utilization of acetate by astrocytes is

glycogen concentration and turnover through localized 13C NMR of both the C1

Diabetes 28:1022-1026.

eds), Macmillan, London, pp. 137–166.

and C6 resonances. NMR Biomed. 23:270-276.

attributable to transport. J. Neurosci. 18:5225-5233.

In the late '80s magnetic resonance imaging (MRI) and magnetic resonance spectroscopy (MRS) revolutionized the care and monitoring of patients with multiple sclerosis (MS). For the first time detailed, high resolution images of the brain, spinal cord and lesions could be made by MRI *in vivo*, and followed over time. However, the changes detectable on T1- or T2-weighted magnetic resonance images are non-specific for the underlying pathology, especially in MS patients, and slow down the enthusiasm. They may reflect edema, demyelination, axonal loss, inflammation, gliosis, remyelination, or Wallerian degeneration in MS.

Similar it is with MRS. Early on it was suggested that metabolites within the tissue may reflect certain cellular compartments and tissue conditions in health and disease. First reports about *in vivo* detection of brain metabolites in humans were published in 1989 (Bruhn et al., 1989; Frahm et al., 1989a; 1989b). Combined with conventional MRI technique spatial anatomical and chemical information about the analysed tissue may be achieved. MRS was approved by the FDA in 1996. However, only few metabolites are detectable (table 1), limiting the main applications of MRS yet to (1) brain tumors (low grade vs. high grade gliomas, metastasis, lymphoma, meningeoma), (2) follow-up of tumors (under chemotherapy and radiotherapy), (2) infectious, mostly focal CNS processes such as pyogenic abscesses, toxoplasmosis, tuberculosis, cryptococcosis, and (3) hepatic encephalopathy, other metabolic disorders, inborn errors of metabolism or hypoxic encephalopathy. In some selected MS cases MRS might be useful to differentiate tumor-like MS brain lesions or Balo-like MS lesions from tumor lesions (table 1) or confluent MS lesions from leucodystrophies or leucoencephalopathies.

Well characterized peaks obtained at long echo times (more than TE 135ms, highlighted in grey) are: (1) at 2.0 ppm ('NAA-peak'; N-acetyl-aspartate (NAA)), (2) at 3.0 ppm ('Cr-peak'; creatine (Cr)), (3) at 3.2 ppm ('Cho-peak'; choline (cho)), (4) at 1.3 ppm ('Lac-peak'; lactate (Lac)), (5) at 1.48ppm ('Alanine Peak'; alanine (Ala)),

<sup>1</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

MR Spectroscopy in Multiple Sclerosis - A New Piece of the Puzzle or Just a New Puzzle 49

Brain metabolites were suggested as markers for (1) neurons (NAA), (2) energy metabolism (Cr), (3) cell membranes or cell membrane turnover (Cho), (4) anaerobic glykolysis, i.e. tissue necrosis (Lac, lipids, cytosolic amino acids, Ala, acetate, succinate), (5) astrogliosis (mIns), and (6) myelin break down (lipids, cytosolic amino acids). Abnormally strong or weak peaks or certain patterns of several peaks may be indicative for various pathological

For instance, lac is normally only detectable if the lac concentration in the CNS parenchyma has reached at least the lac concentration in the cerebrospinal fluid (CSF). Lac is the end product of anaerobic glycolysis and may occur in severe tissue necrosis or abscesses (table 1). The most discussed MRS peaks (or brain metabolites) yet are, NAA, Cho, Crea and mIns.

In normal CNS tissue the highest peak in the proton spectrum is found at 2.0 ppm, resembling the sharp resonance signal from NAA, largely composed of NAA itself, and to a lesser proportion out of N-acetyl-aspartate-glutamate. NAA is an amino-acid derivative synthesized from L-aspartate and acetyl-CoA. NAA was found to be mainly expressed in the mitochondria of neurons, their cell bodies, dendrites and axons, and thus was suggested to indicate either the structural or functional integrity of neurons or axons. However, its function remains unknown. In principal three scenarios, where a reduction of the NAA peak might occur, are discussed: (1) if neurons or axons are damaged irreversible, and degenerate, (2) if their density is reduced relatively in the tissue due to oedema, or (3) if

Cho and Cho-containing phospholipids hardly resonate under normal conditions because they are mainly insoluble and immobile as they constitute all cell membranes and myelin. In tumors, inflammation, infarcts, leucoencephalopathy, leucodystrophies or in certain MS lesions, high cell membrane turnover due to cellular mitosis and cell death, or myelin break down may lead to higher concentration of soluble, freely mobile Cho, which may be detected at 3.2 ppm on the frequency scale. If freely mobile Cho molecules are abundant and

Note that severe tissue destruction, i.e. necrosis, where cell membranes are heavily damaged, torn asunder and single cell membrane compounds released into the necrosis,

The 'Cr-peak' is composed of Cr and phosphor-creatine, and lies at 3.0 ppm on the frequency scale, representing a marker for cellular energy metabolism. As the Cr concentration was found relatively stable throughout the CNS, **and** found relatively resistant to change, Cr is often used as internal standard, and thus the signals strength of the other metabolites are expressed as ratio to Cr. However, under certain conditions such as higher age, tumors, infarcts or trauma Cr is found decreased. (Sometimes Cr was also found

may produce other resonances or peaks (such as free lipids, lac, acetate, succinate).

processes, and may help to interpret changes found by conventional MRI.

**1.1.1 N-acetyl-aspartate (NAA), at 2.0 ppm** 

their function is impaired, only.

**1.1.2 Choline (Cho), at 3.2 ppm** 

**1.1.3 Creatine (Cr), at 3.0 ppm** 

resonate, an increased 'Cho-peak' may be found.

increased in the NAWM of MS patients, see below).


Table 1. Brain metabolites

And at short echo times (less than TE 30ms) are : (6) at 3.5 ppm ('mIns-Peak'; myo-inositol, (mIns)), (7) at 0.9-1.2 ppm ('free lipids peak').

Normal peak levels are given in [mM].

Brain metabolites were suggested as markers for (1) neurons (NAA), (2) energy metabolism (Cr), (3) cell membranes or cell membrane turnover (Cho), (4) anaerobic glykolysis, i.e. tissue necrosis (Lac, lipids, cytosolic amino acids, Ala, acetate, succinate), (5) astrogliosis (mIns), and (6) myelin break down (lipids, cytosolic amino acids). Abnormally strong or weak peaks or certain patterns of several peaks may be indicative for various pathological processes, and may help to interpret changes found by conventional MRI.

For instance, lac is normally only detectable if the lac concentration in the CNS parenchyma has reached at least the lac concentration in the cerebrospinal fluid (CSF). Lac is the end product of anaerobic glycolysis and may occur in severe tissue necrosis or abscesses (table 1).

The most discussed MRS peaks (or brain metabolites) yet are, NAA, Cho, Crea and mIns.

#### **1.1.1 N-acetyl-aspartate (NAA), at 2.0 ppm**

48 Magnetic Resonance Spectroscopy

**(metabolite) ppm [mM] indicates possible to find in…** 

0.9 - products of proteolysis

(highly specific)

(neutrophil cells)

anaerobic glycolysis; inverted double peak (TE136ms)

(TE 136ms)

product of propionic acid fermentation and mixed acid fermentation (anaerobic bacteria)

excitatory neurotransmitter

product of propionic acid fermentation and mixed acid fermentation

metabolism

cell membranes (cell turnover, cell destruction)

mainly in astrocytes

And at short echo times (less than TE 30ms) are : (6) at 3.5 ppm ('mIns-Peak'; myo-inositol,

brain tumors, abscesses, tissue necrosis

abscesses, neurocysticerkosis (but not in neoplasms!)

Balo like MS lesions, malignant tumors, infarcts, abscesses, mitochondrial disorders

> meningeomas brain abscesses

abscesses, neurocysticercosis

decreased in tumors, and any process with tissue destruction

increased in stroke, lymphoma, hepatic encephalopathy, metabolic disorders

> abscesses, neurocysticercosis

mostly stable, used as reference peak

tumors, lymphomas, stroke, infectious processes, MS

gliosis, hepatic encephalopathy, pontine myelinolysis, MS

lipids 0.9-1.2 - tissue necrosis

(Ala) 1.48 - inverted double peak

(NAA) 2.0 7-17 neurons, axons

6-12 and 3-12

(Cr) 3.0 4.5-10.5 cell energy/

(mIns) 3.5 4.0-9.0 glucose metabolism

**marker** 

**cytosolic amino acids**  (valine, leucine, and iso-leucine)

**lactate** 

**alanine** 

**N-acetylaspartate**

**glutamate and glutamine**  (glx)

**Creatine** 

**Choline** 

**myo-Inositol** 

Table 1. Brain metabolites

(Lac) 1.2 -

**acetate** 1.5 -

2.2-2.7

**succinate** 2.4 -

(Cho) 3.2 0.5-3.0

(mIns)), (7) at 0.9-1.2 ppm ('free lipids peak').

Normal peak levels are given in [mM].

In normal CNS tissue the highest peak in the proton spectrum is found at 2.0 ppm, resembling the sharp resonance signal from NAA, largely composed of NAA itself, and to a lesser proportion out of N-acetyl-aspartate-glutamate. NAA is an amino-acid derivative synthesized from L-aspartate and acetyl-CoA. NAA was found to be mainly expressed in the mitochondria of neurons, their cell bodies, dendrites and axons, and thus was suggested to indicate either the structural or functional integrity of neurons or axons. However, its function remains unknown. In principal three scenarios, where a reduction of the NAA peak might occur, are discussed: (1) if neurons or axons are damaged irreversible, and degenerate, (2) if their density is reduced relatively in the tissue due to oedema, or (3) if their function is impaired, only.

#### **1.1.2 Choline (Cho), at 3.2 ppm**

Cho and Cho-containing phospholipids hardly resonate under normal conditions because they are mainly insoluble and immobile as they constitute all cell membranes and myelin. In tumors, inflammation, infarcts, leucoencephalopathy, leucodystrophies or in certain MS lesions, high cell membrane turnover due to cellular mitosis and cell death, or myelin break down may lead to higher concentration of soluble, freely mobile Cho, which may be detected at 3.2 ppm on the frequency scale. If freely mobile Cho molecules are abundant and resonate, an increased 'Cho-peak' may be found.

Note that severe tissue destruction, i.e. necrosis, where cell membranes are heavily damaged, torn asunder and single cell membrane compounds released into the necrosis, may produce other resonances or peaks (such as free lipids, lac, acetate, succinate).

#### **1.1.3 Creatine (Cr), at 3.0 ppm**

The 'Cr-peak' is composed of Cr and phosphor-creatine, and lies at 3.0 ppm on the frequency scale, representing a marker for cellular energy metabolism. As the Cr concentration was found relatively stable throughout the CNS, **and** found relatively resistant to change, Cr is often used as internal standard, and thus the signals strength of the other metabolites are expressed as ratio to Cr. However, under certain conditions such as higher age, tumors, infarcts or trauma Cr is found decreased. (Sometimes Cr was also found increased in the NAWM of MS patients, see below).

MR Spectroscopy in Multiple Sclerosis - A New Piece of the Puzzle or Just a New Puzzle 51

*In biological sciences, the role of a method is even more important than in other sciences, because of the immense complexity of the phenomena and the countless sources of error, which complexity brings into experimentation.2*

**1.2.1 'Differential diagnosis of suspected multiple sclerosis: A consensus approach'**  Miller DH, Weinshenker BG, Filippi M, Banwell BL, Cohen JA, Freedman MS, Galetta SL, Hutchinson M, Johnson RT, Kappos L, Kira J, Lublin FD, McFarland HF, Montalban X,

P 1164, […] *Although symptoms and signs of a monophasic illness have been an essential prerequisite for diagnosis of CIS, there is an exceptional scenario that the Panel feels warrants inclusion as CIS Type 5 (Table 2): patients who have no symptoms or only non-specific symptoms (e.g., headache, dizziness), but have MRI evidence for multifocal abnormalities typical for demyelination. Such patients are increasingly identified using MRI for incidental indications (e.g., headache) especially with high-field strength magnets with greater sensitivity for such lesions [17]. Current criteria preclude a diagnosis of MS without objective clinical evidence for CNS abnormality and the ability to establish a confident diagnosis of MS in such individuals and their natural history* 

P 1172, […] *Disease biomarkers will aid enormously in differential diagnosis. Accurate and sensitive disease markers – imaging or laboratory based – may provide non-invasive aids to differential diagnosis. The discovery of NMO-IgG that helps to distinguish NMO from MS is a good example. Relevant imaging advances may include non-conventional MR techniques to quantitate change in normal appearing white matter that may be relatively specific for MS, and high-field MRI to better* 

Interestingly, the reference No. 17 (title 'Cerebral Cortical Lesions in Multiple Sclerosis Detected by MR Imaging at 8 Tesla') is a case report where the *post mortem/ex vivo* findings of a single patient, a male aged 42 years, were described (Kangarlu et al., 2007). Coronal slices of the patient's formalin fixed brain were examined using MRI at 8T. Details about the

**'Cerebral cortical lesions in multiple sclerosis detected by MR imaging at 8 Tesla.'**

*A 42-year-old man with severe disabilities secondary to MS died of aspiration pneumonia. In the year before his death, he had been seen at our medical center on 2 occasions, mostly for palliative care. He was bed-bound and very dependent for all activities of daily living. He was paraplegic with additional severe weakness of the upper extremities, which were also severely ataxic. The ataxia also affected his trunk and prevented him from sitting unassisted in a standard wheelchair. He had complete bilateral internuclear ophthalmoplegia, vertical nystagmus, and severe* 

*visualize Dawson's fingers [17] or cortical lesions that may be specific to MS.* […]

Kangarlu A, Bourekas EC, Ray-Chaudhury A, Rammohan KW.

<sup>2</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

AJNR American Journal of Neuroradiology, Vol. 28, No. 2, pp. 262-266.

disease course, or how the diagnosis 'MS' were made was not specified, p262, […]

**1.2 Recently the pivotal role of advanced MR technique was emphasized** 

Panitch H, Richert JR, Reingold SC, Polman CH. Multiple Sclerosis, Vol. 14, No. 9, pp. 1157-1174.

*should be addressed through prospective studies* [18]. […]

#### **1.1.4 Myo-Inositol (mINS), at 3.5 ppm**

The 'mIns-peak' is believed to resemble an 'activated' state of higher cell metabolism or proliferation of astrocytes. mIns is a sugar-like molecule which may be crucial for the osmotic regulation in the CNS parenchyma. An elevated 'mIns-peak' may be interpreted as higher astroglial activity and proliferation or gliosis.

But MRS allows a limited view only. From a scientific point of view, it is hardly conceivable that a few metabolites really allow the complete representation of complex biological processes of individual cells within the spectroscopied CNS tissue. Note that the function of the metabolites and that their (resonance) characteristics under certain conditions are still largely unknown.

One of the most important and controversially discussed hypotheses today is that 'early axonal damage already exists in very early stages of (all) MS patients'. This view may open the way for 'an (very) early treatment' of MS patients, also if clinical symptoms are minor or even absent (Miller et al., 2008; Gilmore et al., 2010). This hypothesis comes mainly from MRI and MRS findings reporting reduced NAA levels in the normal appearing white matter (NAWM). NAA levels are suggested as marker for neuronal integrity and function and should reflect the burden of disease in MS patients (figure 1).

Fig. 1. A hypothesis, nothing else. **a,** It remains unclear, simply hypothetically, whether progressive, i.e. steadily ongoing 'silent', axonal damage exists in all MS patients, or not and if so, whether it can be reliable detected and monitored over time. It was suggested that NAA levels in the NAWM decline in some MS patients continuously. This NAA decline would be expected more than naturally occurring declines by age. **b,** Objective and neutral interpretation of the data allows also a second scenario, where brain volume (and NAA concentration?) only slightly decreases with age, or secondary progression. And during relapses temporarily a decrease of NAA can be found, that may recover fully.

However, due to the most often relapsing-remitting course of MS patients, with remission over years and sometimes over decades, early treatment per se must be carefully considered and risks and benefits carefully weighed.

*In biological sciences, the role of a method is even more important than in other sciences, because of the immense complexity of the phenomena and the countless sources of error, which complexity brings into experimentation.2*

#### **1.2 Recently the pivotal role of advanced MR technique was emphasized**

#### **1.2.1 'Differential diagnosis of suspected multiple sclerosis: A consensus approach'**

Miller DH, Weinshenker BG, Filippi M, Banwell BL, Cohen JA, Freedman MS, Galetta SL, Hutchinson M, Johnson RT, Kappos L, Kira J, Lublin FD, McFarland HF, Montalban X, Panitch H, Richert JR, Reingold SC, Polman CH.

Multiple Sclerosis, Vol. 14, No. 9, pp. 1157-1174.

50 Magnetic Resonance Spectroscopy

The 'mIns-peak' is believed to resemble an 'activated' state of higher cell metabolism or proliferation of astrocytes. mIns is a sugar-like molecule which may be crucial for the osmotic regulation in the CNS parenchyma. An elevated 'mIns-peak' may be interpreted as

But MRS allows a limited view only. From a scientific point of view, it is hardly conceivable that a few metabolites really allow the complete representation of complex biological processes of individual cells within the spectroscopied CNS tissue. Note that the function of the metabolites and that their (resonance) characteristics under certain conditions are still

One of the most important and controversially discussed hypotheses today is that 'early axonal damage already exists in very early stages of (all) MS patients'. This view may open the way for 'an (very) early treatment' of MS patients, also if clinical symptoms are minor or even absent (Miller et al., 2008; Gilmore et al., 2010). This hypothesis comes mainly from MRI and MRS findings reporting reduced NAA levels in the normal appearing white matter (NAWM). NAA levels are suggested as marker for neuronal integrity and function and

Fig. 1. A hypothesis, nothing else. **a,** It remains unclear, simply hypothetically, whether progressive, i.e. steadily ongoing 'silent', axonal damage exists in all MS patients, or not and if so, whether it can be reliable detected and monitored over time. It was suggested that NAA levels in the NAWM decline in some MS patients continuously. This NAA decline would be expected more than naturally occurring declines by age. **b,** Objective and neutral interpretation of the data allows also a second scenario, where brain volume (and NAA concentration?) only slightly decreases with age, or secondary progression. And during

However, due to the most often relapsing-remitting course of MS patients, with remission over years and sometimes over decades, early treatment per se must be carefully considered

relapses temporarily a decrease of NAA can be found, that may recover fully.

and risks and benefits carefully weighed.

**1.1.4 Myo-Inositol (mINS), at 3.5 ppm** 

largely unknown.

higher astroglial activity and proliferation or gliosis.

should reflect the burden of disease in MS patients (figure 1).

P 1164, […] *Although symptoms and signs of a monophasic illness have been an essential prerequisite for diagnosis of CIS, there is an exceptional scenario that the Panel feels warrants inclusion as CIS Type 5 (Table 2): patients who have no symptoms or only non-specific symptoms (e.g., headache, dizziness), but have MRI evidence for multifocal abnormalities typical for demyelination. Such patients are increasingly identified using MRI for incidental indications (e.g., headache) especially with high-field strength magnets with greater sensitivity for such lesions [17]. Current criteria preclude a diagnosis of MS without objective clinical evidence for CNS abnormality and the ability to establish a confident diagnosis of MS in such individuals and their natural history should be addressed through prospective studies* [18]. […]

P 1172, […] *Disease biomarkers will aid enormously in differential diagnosis. Accurate and sensitive disease markers – imaging or laboratory based – may provide non-invasive aids to differential diagnosis. The discovery of NMO-IgG that helps to distinguish NMO from MS is a good example. Relevant imaging advances may include non-conventional MR techniques to quantitate change in normal appearing white matter that may be relatively specific for MS, and high-field MRI to better visualize Dawson's fingers [17] or cortical lesions that may be specific to MS.* […]

Interestingly, the reference No. 17 (title 'Cerebral Cortical Lesions in Multiple Sclerosis Detected by MR Imaging at 8 Tesla') is a case report where the *post mortem/ex vivo* findings of a single patient, a male aged 42 years, were described (Kangarlu et al., 2007). Coronal slices of the patient's formalin fixed brain were examined using MRI at 8T. Details about the disease course, or how the diagnosis 'MS' were made was not specified, p262, […]

**'Cerebral cortical lesions in multiple sclerosis detected by MR imaging at 8 Tesla.'** Kangarlu A, Bourekas EC, Ray-Chaudhury A, Rammohan KW.

AJNR American Journal of Neuroradiology, Vol. 28, No. 2, pp. 262-266.

*A 42-year-old man with severe disabilities secondary to MS died of aspiration pneumonia. In the year before his death, he had been seen at our medical center on 2 occasions, mostly for palliative care. He was bed-bound and very dependent for all activities of daily living. He was paraplegic with additional severe weakness of the upper extremities, which were also severely ataxic. The ataxia also affected his trunk and prevented him from sitting unassisted in a standard wheelchair. He had complete bilateral internuclear ophthalmoplegia, vertical nystagmus, and severe* 

 <sup>2</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

MR Spectroscopy in Multiple Sclerosis - A New Piece of the Puzzle or Just a New Puzzle 53

Fig. 2. 'The chicken or the egg causality dilemma' in MS lesion evolution.

around focal MS lesions; **lime,** hypothesized invisible, yet not determined,

may cause substantial bystander tissue damage, or both.

**Red,** enhancing (active) MR lesions; **blue,** old(er), at least not enhancing lesions**; green,** in MR invisible inflammation or an underlying, yet not determined, pathogenetic mechanism

compartmentalized focal or diffuse inflammation or yet not determined pathogenetic process**. a-e, the disease might be triggered from outside the CNS. a,** normal brain without any lesion or invisible disease activity; **b,** the naïve brain with two enhancing lesions (red) only; **c,** brain with typical periventricular and subcortical old (not enhancing) white matter (WM) lesions only; **d**, same scenario as in c, but with new enhancing lesions (red). **e**, older non enhancing WM lesions (blue) of which one enhances in its right anterior margin (red margin) with ongoing (invisible) activity (green) beyond the in MR visible lesion border and two new enhancing (red) with invisible activity in the left hemisphere. **f-o, the disease might be already compartmentalized within the CNS from the very beginning, but invisible with MRI (lime): f-j,** focally, but widely dispersed; **k-o,** diffusely within the CNS. **f-o,** show the same enhancing (red) and old/er (blue) in MRI detectable lesions as presented in *a-d*. In both scenarios, the in MRI detectable lesions may resemble only the 'tip of the iceberg', leading to over time and space disseminated focal lesions with BBB alterations. In other words, the hypothesized underlying, invisible disease activity may lead from inside out to focal 'eruptions' with BBB alterations and influx of inflammation, or might attract inflammatory cells which lead to the BBB breakdown. The infiltrating inflammatory cells might be primarily harmful to the CNS, or otherwise, they might be primarily beneficial but

*titubation of his head. His ability to communicate was intact, though his speech was barely intelligible. His mentation seemed intact; he remained completely oriented and participated in all decisions of his care, including his desire to have his brain evaluated for scientific endeavors after death. He needed to be fed through a percutaneous gastrostomy tube. Cognitive status could not be formally evaluated because of the severity of his disabilities. The duration of his disease was estimated to be approximately 12 years. At the time of his demise, he was considered to be in the advanced stages of the secondary-progressive form of MS. His general health was otherwise excellent, and he had no known cardiovascular, cerebrovascular, hypertensive, or diabetes-related disabilities.* […].

[…]

For reference No. 18 see, Lebrun at al., 2008 and the comment by Chattaway, 2008.

[…]

As critical readers we are aware that we must read more than the headlines and abstracts of the papers and that the hypothesis that 'early axonal damage already exists in very early stages of (all) MS patients' has yet to be proven (Figure 1), and that the very controversial discussion is still ongoing (e.g. Chattaway, 2008 and 2010; Gilmore et al., 2010; Lebrun et al., 2008; Miller et al., 2008). In any case, we must keep in mind that direct scientific evidence is still lacking.

> *I consider the hospital as the antechamber of medicine. It is the first place where the physician makes his observations. But the laboratory is the temple of the science of medicine.3*

#### **2. Multiple sclerosis**

Multiple sclerosis (MS) is a chronic idiopathic disease of the central nervous system (CNS). Inflammation, demyelination and axonal injury are most typical pathological features, but their underlying pathogenetic mechanisms are still unclear (Barnett et al., 2009; Hibberd, 1994; Lassmann et al., 2007). They are probably complex and heterogeneous, and may be triggered outside the CNS or conversely, may be triggered within the CNS. And they may initiate focally and further disperse, or they may affect the whole CNS diffusely at once. In either case, blood brain barrier (BBB) alterations are conceivable, focally or diffusely, visible or invisible (Figure 2). But at least the primary cause of MS, and whether BBB alterations are the initial detectable pathological event in the evolution of the MS lesions, or why and whatever, the MS lesions are caused by, remains unknown (Aboul-Enein & Lassmann, 2006; Allen & McKeown, 1979; Allen et al., 1989; Barnes et al., 1991; Barnett et al., 2004; Barnett & Prineas, 2009; Henderson et al., 2009; Höftberger et al., 2004; Lassmann et al., 2007).

> *If an idea presents itself to us, we must not reject it simply, because it does not agree with the logical deductions of a reigning theory.4 It is what we know already that often prevents us from learning.5*

<sup>3</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

<sup>4</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

<sup>5</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

*titubation of his head. His ability to communicate was intact, though his speech was barely intelligible. His mentation seemed intact; he remained completely oriented and participated in all decisions of his care, including his desire to have his brain evaluated for scientific endeavors after death. He needed to be fed through a percutaneous gastrostomy tube. Cognitive status could not be formally evaluated because of the severity of his disabilities. The duration of his disease was estimated to be approximately 12 years. At the time of his demise, he was considered to be in the advanced stages of the secondary-progressive form of MS. His general health was otherwise excellent, and he had no known cardiovascular, cerebrovascular, hypertensive, or diabetes-related* 

For reference No. 18 see, Lebrun at al., 2008 and the comment by Chattaway, 2008.

As critical readers we are aware that we must read more than the headlines and abstracts of the papers and that the hypothesis that 'early axonal damage already exists in very early stages of (all) MS patients' has yet to be proven (Figure 1), and that the very controversial discussion is still ongoing (e.g. Chattaway, 2008 and 2010; Gilmore et al., 2010; Lebrun et al., 2008; Miller et al., 2008). In any case, we must keep in mind that direct

> *I consider the hospital as the antechamber of medicine. It is the first place where the physician makes his observations. But the laboratory is the temple of the science of medicine.3*

Multiple sclerosis (MS) is a chronic idiopathic disease of the central nervous system (CNS). Inflammation, demyelination and axonal injury are most typical pathological features, but their underlying pathogenetic mechanisms are still unclear (Barnett et al., 2009; Hibberd, 1994; Lassmann et al., 2007). They are probably complex and heterogeneous, and may be triggered outside the CNS or conversely, may be triggered within the CNS. And they may initiate focally and further disperse, or they may affect the whole CNS diffusely at once. In either case, blood brain barrier (BBB) alterations are conceivable, focally or diffusely, visible or invisible (Figure 2). But at least the primary cause of MS, and whether BBB alterations are the initial detectable pathological event in the evolution of the MS lesions, or why and whatever, the MS lesions are caused by, remains unknown (Aboul-Enein & Lassmann, 2006; Allen & McKeown, 1979; Allen et al., 1989; Barnes et al., 1991; Barnett et al., 2004; Barnett &

Prineas, 2009; Henderson et al., 2009; Höftberger et al., 2004; Lassmann et al., 2007).

*If an idea presents itself to us, we must not reject it simply, because it does not agree with the logical deductions of a reigning theory.4 It is what we know already that often prevents us from learning.5*

*disabilities.* […].

scientific evidence is still lacking.

<sup>3</sup> **Claude Bernard** (French physiologist, 1813 – 1878) <sup>4</sup> **Claude Bernard** (French physiologist, 1813 – 1878) <sup>5</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

**2. Multiple sclerosis** 

[…]

[…]

Fig. 2. 'The chicken or the egg causality dilemma' in MS lesion evolution. **Red,** enhancing (active) MR lesions; **blue,** old(er), at least not enhancing lesions**; green,** in MR invisible inflammation or an underlying, yet not determined, pathogenetic mechanism around focal MS lesions; **lime,** hypothesized invisible, yet not determined, compartmentalized focal or diffuse inflammation or yet not determined pathogenetic process**. a-e, the disease might be triggered from outside the CNS. a,** normal brain without any lesion or invisible disease activity; **b,** the naïve brain with two enhancing lesions (red) only; **c,** brain with typical periventricular and subcortical old (not enhancing) white matter (WM) lesions only; **d**, same scenario as in c, but with new enhancing lesions (red). **e**, older non enhancing WM lesions (blue) of which one enhances in its right anterior margin (red margin) with ongoing (invisible) activity (green) beyond the in MR visible lesion border and two new enhancing (red) with invisible activity in the left hemisphere. **f-o, the disease might be already compartmentalized within the CNS from the very beginning, but invisible with MRI (lime): f-j,** focally, but widely dispersed; **k-o,** diffusely within the CNS. **f-o,** show the same enhancing (red) and old/er (blue) in MRI detectable lesions as presented in *a-d*. In both scenarios, the in MRI detectable lesions may resemble only the 'tip of the iceberg', leading to over time and space disseminated focal lesions with BBB alterations. In other words, the hypothesized underlying, invisible disease activity may lead from inside out to focal 'eruptions' with BBB alterations and influx of inflammation, or might attract inflammatory cells which lead to the BBB breakdown. The infiltrating inflammatory cells might be primarily harmful to the CNS, or otherwise, they might be primarily beneficial but may cause substantial bystander tissue damage, or both.

MR Spectroscopy in Multiple Sclerosis - A New Piece of the Puzzle or Just a New Puzzle 55

sodium channels may be restored, or the conduction block causing mediators may be removed (Smith, 2007). Of course, other mediators derived from any other tissue or incorporated from outside, and being not associated with the body's immune system might be worth considering. But neither remyelination nor gliosis, both can be extensive in autopsies, correlate to permanent functional loss, if at all, only the irreversible axonal loss was found to correlate. And pathologists have only a very limited point of view (as neurologists and radiologists, too!). They can merely try to reconstruct the *in vivo* situation in most of the cases after long disease courses with very specific markers which however, allow mostly no further statement

On a more philosophical - someone might argue - on a very 'realistic' level, we must remember that we merely know nothing about the disease's origin or the disease's specific processes. Direct scientific evidence is lacking for almost every little so far identified experimental, pathological, clinical, radiological or epidemiological fragment. We even do not know whether MS is one disease or a whole spectrum of disorders, or vice versa, whether the same pathogenetic mechanisms may have different effects in different individuals, or at different times in one individual MS patient. Otherwise not a specific primary cause itself, but the lack of specific resistant factor(s) or repair mechanisms might be responsible that some individuals become ill (and diagnosed), whereas all the others remain healthy. In other words, how many healthy people (Also you, dear readers!) could have a considerable lesion load or brain atrophy in MRI suggestive for MS, but without any further consequence as they feel healthy (and never see a doctor). Patients diagnosed with MS may be that ones with lesions mainly in eloquent areas. Thus they may resemble only a very small portion, the so called 'tip of the iceberg', but of course, they shape our

However, by occasion some patients, who receive a MRI of their brain (because of a wide range of 'unspecific' symptoms, and because MRI is now widely available), show asymptomatic lesions suggestive for MS (Chataway, 2010). Recently, these 'patients' were classified as radiological isolated syndrome (RIS) or clinical isolated syndrome (CIS) type 5 (Miller et al., 2008). It is controversially discussed, whether these 'patients' should be followed closely, or even treated (Chataway, 2008, 2010; Gilmore et al., 2010; Lebrun et al.,

The link between disease onset and possible irreversible end-stage parameters such as glial scar tissue formation and brain and spinal cord atrophy is still missing. MRI and magnetic resonance spectroscopy (MRS) allows to visualize changes during individual disease courses in vivo, but they provide at least information from magnetic fields of certain

> *A fact in itself is nothing. It is valuable only for the idea attached to it, or for the proof which it furnishes.7*

molecules, mostly hydrogen nuclei (protons, or 1H) of water molecules, nothing else.

about the in vivo integrity and function over the patients' lifetime.

2008; Miller et al., 2008; Sellner et al., 2010; Siva et al., 2009).

<sup>7</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

**2.3 The clinical-radiological paradox of MS** 

view about MS.

#### **2.1 Diagnostic criteria and the disease course of MS**

The diagnosis 'MS' is a clinical one, if stringent diagnostic criteria are fulfilled and if there is no better explanation for the clinical presentation (Mc Donald et al., 2001; Polman et al. 2005, 2010; Poser et al., 1983; Schuhmacher et al., 1965). The diagnosis 'MS' is based on the clinical evidence for typical disseminated CNS lesions, and may be supported by cerebrospinal fluid (CSF) analysis, magnetic resonance imaging (MRI), and evoked potentials. However, all findings are non-specific by themselves, but overall, they might be typical for the diagnosis 'MS'. Up till now, a reliable specific para-clinical marker neither for the diagnosis nor for monitoring disease activity of MS could yet be established (e.g. Aboul-Enein et al., 2010; Gilmore et al., 2010; Kuhle et al., 2007; Serbecic et al., 2010, 2011).

In most of the cases MS starts between the age of 20 to 40 years and follows a relapsingremitting course with clear defined relapses with no apparent clinical deterioration between the relapses. Each MS patient follows his/her individual disease course (Gilmore et al., 2010). It is unclear why MS sometimes follows a benign disease course with clinical remission over decades, or changes after an uncertain period of time into secondary progression with or without superimposed relapses (SPMS, secondary progressive MS), or follows rarely a progressive disease course right from the disease onset (PPMS, primary progressive MS) (Lublin & Reingold, 1996).

To sum up, no clinical or paraclinical parameter could yet be found, which allows a reliable prognosis of the 'relapse-free intervals', the relapses, the sequelae and the possible 'conversion into secondary progression'. In particular, neither the frequency nor severity of relapses nor disability in the first years after onset nor the lesion load in magnetic resonance images (MRI) correlates strictly with the disease activity or its impact on the individual clinical disease course (e.g. Brex et al., 2002; Fisniku et al., 2008; Gilmore et al., 2010; Weinshenker et al., 1989a, 1989b). And axonal alterations or axonal loss, which is yet suggested as main pathological substrate, unfortunately, provides only some explanation.

*In teaching man, experimental science results in lessening his pride more and more by proving to him every day that primary causes, like the objective reality of things, will be hidden from him forever and that he can only know relations.6*

#### **2.2 The clinical-pathological paradox of MS**

Pathologists and neurologists have known for a long time that a considerable amount of CNS lesions with demyelination, axonal changes, inflammation and gliosis might be found, although clinical signs and symptoms were absent or should have never occurred (Lumdsen, 1970). A plausible explanation could be that in such cases the lesions were small, or that they occurred in non-eloquent CNS areas, but in any way, it means that recovery of function is possible. Axonal loss of function may be caused either by structural damage such as demyelination and (secondary) axonal injury (McDonald and Sears, 1970) or may be caused without any detectable structural damage by certain immune mediators only, which might interfere with axonal conductance (Aboul-Enein et al., 2006; Smith & Lassmann, 2002). Block of conduction is in principal reversible as axons may be remyelinated and the axonal

<sup>6</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

sodium channels may be restored, or the conduction block causing mediators may be removed (Smith, 2007). Of course, other mediators derived from any other tissue or incorporated from outside, and being not associated with the body's immune system might be worth considering. But neither remyelination nor gliosis, both can be extensive in autopsies, correlate to permanent functional loss, if at all, only the irreversible axonal loss was found to correlate. And pathologists have only a very limited point of view (as neurologists and radiologists, too!). They can merely try to reconstruct the *in vivo* situation in most of the cases after long disease courses with very specific markers which however, allow mostly no further statement about the in vivo integrity and function over the patients' lifetime.

#### **2.3 The clinical-radiological paradox of MS**

54 Magnetic Resonance Spectroscopy

The diagnosis 'MS' is a clinical one, if stringent diagnostic criteria are fulfilled and if there is no better explanation for the clinical presentation (Mc Donald et al., 2001; Polman et al. 2005, 2010; Poser et al., 1983; Schuhmacher et al., 1965). The diagnosis 'MS' is based on the clinical evidence for typical disseminated CNS lesions, and may be supported by cerebrospinal fluid (CSF) analysis, magnetic resonance imaging (MRI), and evoked potentials. However, all findings are non-specific by themselves, but overall, they might be typical for the diagnosis 'MS'. Up till now, a reliable specific para-clinical marker neither for the diagnosis nor for monitoring disease activity of MS could yet be established (e.g. Aboul-Enein et al., 2010;

In most of the cases MS starts between the age of 20 to 40 years and follows a relapsingremitting course with clear defined relapses with no apparent clinical deterioration between the relapses. Each MS patient follows his/her individual disease course (Gilmore et al., 2010). It is unclear why MS sometimes follows a benign disease course with clinical remission over decades, or changes after an uncertain period of time into secondary progression with or without superimposed relapses (SPMS, secondary progressive MS), or follows rarely a progressive disease course right from the disease onset (PPMS, primary

To sum up, no clinical or paraclinical parameter could yet be found, which allows a reliable prognosis of the 'relapse-free intervals', the relapses, the sequelae and the possible 'conversion into secondary progression'. In particular, neither the frequency nor severity of relapses nor disability in the first years after onset nor the lesion load in magnetic resonance images (MRI) correlates strictly with the disease activity or its impact on the individual clinical disease course (e.g. Brex et al., 2002; Fisniku et al., 2008; Gilmore et al., 2010; Weinshenker et al., 1989a, 1989b). And axonal alterations or axonal loss, which is yet suggested as main pathological substrate, unfortunately, provides only some explanation. *In teaching man, experimental science results in lessening his pride more and more by proving to him every day that primary causes, like the objective reality of things, will be hidden from him forever and that he can only know relations.6*

Pathologists and neurologists have known for a long time that a considerable amount of CNS lesions with demyelination, axonal changes, inflammation and gliosis might be found, although clinical signs and symptoms were absent or should have never occurred (Lumdsen, 1970). A plausible explanation could be that in such cases the lesions were small, or that they occurred in non-eloquent CNS areas, but in any way, it means that recovery of function is possible. Axonal loss of function may be caused either by structural damage such as demyelination and (secondary) axonal injury (McDonald and Sears, 1970) or may be caused without any detectable structural damage by certain immune mediators only, which might interfere with axonal conductance (Aboul-Enein et al., 2006; Smith & Lassmann, 2002). Block of conduction is in principal reversible as axons may be remyelinated and the axonal

**2.1 Diagnostic criteria and the disease course of MS** 

Gilmore et al., 2010; Kuhle et al., 2007; Serbecic et al., 2010, 2011).

progressive MS) (Lublin & Reingold, 1996).

**2.2 The clinical-pathological paradox of MS** 

<sup>6</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

On a more philosophical - someone might argue - on a very 'realistic' level, we must remember that we merely know nothing about the disease's origin or the disease's specific processes. Direct scientific evidence is lacking for almost every little so far identified experimental, pathological, clinical, radiological or epidemiological fragment. We even do not know whether MS is one disease or a whole spectrum of disorders, or vice versa, whether the same pathogenetic mechanisms may have different effects in different individuals, or at different times in one individual MS patient. Otherwise not a specific primary cause itself, but the lack of specific resistant factor(s) or repair mechanisms might be responsible that some individuals become ill (and diagnosed), whereas all the others remain healthy. In other words, how many healthy people (Also you, dear readers!) could have a considerable lesion load or brain atrophy in MRI suggestive for MS, but without any further consequence as they feel healthy (and never see a doctor). Patients diagnosed with MS may be that ones with lesions mainly in eloquent areas. Thus they may resemble only a very small portion, the so called 'tip of the iceberg', but of course, they shape our view about MS.

However, by occasion some patients, who receive a MRI of their brain (because of a wide range of 'unspecific' symptoms, and because MRI is now widely available), show asymptomatic lesions suggestive for MS (Chataway, 2010). Recently, these 'patients' were classified as radiological isolated syndrome (RIS) or clinical isolated syndrome (CIS) type 5 (Miller et al., 2008). It is controversially discussed, whether these 'patients' should be followed closely, or even treated (Chataway, 2008, 2010; Gilmore et al., 2010; Lebrun et al., 2008; Miller et al., 2008; Sellner et al., 2010; Siva et al., 2009).

The link between disease onset and possible irreversible end-stage parameters such as glial scar tissue formation and brain and spinal cord atrophy is still missing. MRI and magnetic resonance spectroscopy (MRS) allows to visualize changes during individual disease courses in vivo, but they provide at least information from magnetic fields of certain molecules, mostly hydrogen nuclei (protons, or 1H) of water molecules, nothing else.

> *A fact in itself is nothing. It is valuable only for the idea attached to it, or for the proof which it furnishes.7*

<sup>7</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

MR Spectroscopy in Multiple Sclerosis - A New Piece of the Puzzle or Just a New Puzzle 57

immune cells, down-regulation of adhesion molecules on immune cells or endothelial cells, or both, or another yet not determined mechanism, that seals the BBB (Engelhardt, 2006;

Most of the acute gadolinium enhancing MS lesions change their MR characteristics in further follow up. Some convert to mildly or severly hpointense T1 lesions which might reflect severe underlying tissue damage with marked axonal loss and correlating reduced NAA levels (e.g. van Walderveen et al., 1998), others convert to persistent T2 lesions after several months. The latter may appear hyperintense in T1 weighted images or become undectable in T1 weighted images. In persistent T2 lesions, no matter what behavior they had in T1, nearly all thinkable metabolite constellations were described (increase, decrease, no changes of NAA, Cr, Cho, mIns etc., e.g. Davie et al., 1997; van Walderveen et al., 1997.). Reduced NAA levels might reflect axonal injury, dysfunction or loss. They might be temporarily or permanently, within lesions or remote from lesions due to secondary axonal degeneration or Wallerian degeneration. (Increased NAA levels have yet been described in Canavan's disease only). Elevated Cr, mIns levels were suggested to reflect 'higher metabolism' in astrocytes, and increased Cho was suggested to reflect higher cell membrane turnover or proliferation of astrocytes within MS lesions or NAWM. Decreased levels might be due to decreased cell numbers or lower 'cellular energy state'. These divergent results may be explained by the very small and heterogeneous patient cohorts (e.g. 2 RRMS and 12 SPMS aged from 28 to 61 years, and 4 controls aged 33 to 48 years in van Walderveen et al., 1999), by the use of a 1.5 Tesla 1H-MRS and by the measurement of metabolite ratios instead

Only single reports of histopathological correlations to *in vivo* MR data have been published (Bitsch et al., 1997) that are of most importance for our understanding, but must be interpreted very carefully. Biopsies are only very seldom performed in MS patients, mostly if tumor lesions must be ruled out or verified. Whether demyelinating tumor-like MS lesions have the same characteristics and behavior like common unspecific smaller MS lesions, is unclear. Moreover, it is still unknown, if MS is one disease or a syndrome of many disorders, or if the same trigger causes the same pathological changes in different MS patients or if the same trigger at different time points leads to different types of lesions in one individual MS patient. Therefore, any general interpretation must be done with caution. *Men who have excessive faith in their theories or ideas are not only ill prepared for making discoveries. They also make very poor observations.9*

MRI allows the detailed visualization of the anatomical and pathological structures based on different signals of freely mobile hydrogen protons between different tissues or anatomical

Gelati et al., 2002; Leussink et al., 2001; Prat et al., 2002).

**3.2 Others than acute enhancing lesions** 

of absolute metabolite concentrations.

compartments.

**4. Of what we believe to see with MRT and MRS** 

<sup>9</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

#### **3. Of what we believe about MRI lesion evolution**

#### **3.1 Acute enhancing lesions**

It is believed, that the alteration of the BBB may be the earliest event in formation of a focal MS lesion which may be visualized by gadolinium enhancement in T1-weighted MRI images with high sensitivity. Prerequisite for this is that gadolinium must be able to reach the specific region and that the BBB alterations are rather gross ('BBB leakage', in a strict sense). Gadolinium enhancement was suggested to be more sensitive in detecting disease activity than either clinical examination or T2 weighted MRI (Miller et al., 1993). Theoretically, all relapses should be accompanied by new, active enhancing MRI lesions (Figure 1b, d, e g, i, j, l, n o). But, enhancement is dependent on the dose and the technique used, and dependent on the lesion size and location. This must be borne in mind when different studies are compared.

Higher doses of gadolinium (triple dose, 0.3mmol/kg body weight), longer delay between injection and the MRI scan (30-60 minutes), and the combination with a magnetization transfer pulse which suppresses the normal brain tissue, but spares areas with enhancement, can increase the sensitivity of MRI, i.e. make an higher number of enhancing MS lesions visualized (Filippi et al., 1996; Miller et al., 1998). Caveat is the increase of false positive results. However it remains unclear, if all lesions, in particular, those with only subtle BBB alterations can be detected hereby (Barnes et al., 1991; Kwon & Prineas et al., 1994, Waubant, 2006). Very often neurological symptoms may occur without any (correlating) acute enhancing MRI lesions, especially in the spinal cord. And most confusing is that acute enhancing MRI lesions that are clinically silent, may be found between relapses, even though they are located in eloquent areas (Davie et al., 1994; Schubert et al., 2002).

> *Put off your imagination as you take off your overcoat, when you enter the laboratory.8*

In contrast to usual CNS immune surveillance with regular, always occurring transmigration of only few immune cells (Aboul-Enein et al., 2004), it is supposed that the lesion evolution starts with large numbers of immune cells that cross the BBB into the CNS parenchyma, and thus may lead to BBB leakage (Barnes et al., 1991; Hawkins et al., 1991). This is mainly supported by animal models of experimental autoimmune encephalomyelitis (EAE) that however, must not be uncritically translated to humans. In EAE huge masses of CNS antigen specific immune cells or antibodies, or both, are injected, and initiate an acute, but mostly monophasic inflammatory disease in the CNS (Flugel et al., 2001, 2007; Floris et al., 2004; Hawkins et al., 1991; Kawakami et al., 2005). The situation in humans with MS remains unclear. It is even unknown whether the inflammation is primary, whether the infiltrating immune cells themselves cause the tissue damage, or whether some other primary event within the CNS leads to the actual lesion formation and attracts the inflammatory cells into the CNS. The inflammation may be beneficial or conversely, may cause some bystander damage (Barnett & Prineas, 2004; Henderson et al., 2009; Hohlfeld et al., 2006). Anyway, corticosteroids may decrease or may completely suppress the enhancement of MS lesions for several weeks (Barkhof et al., 1991, 1994; Burnham et al., 1991). Conceivable are various mechanisms of action such as apoptosis of circulating

<sup>8</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

immune cells, down-regulation of adhesion molecules on immune cells or endothelial cells, or both, or another yet not determined mechanism, that seals the BBB (Engelhardt, 2006; Gelati et al., 2002; Leussink et al., 2001; Prat et al., 2002).

#### **3.2 Others than acute enhancing lesions**

56 Magnetic Resonance Spectroscopy

It is believed, that the alteration of the BBB may be the earliest event in formation of a focal MS lesion which may be visualized by gadolinium enhancement in T1-weighted MRI images with high sensitivity. Prerequisite for this is that gadolinium must be able to reach the specific region and that the BBB alterations are rather gross ('BBB leakage', in a strict sense). Gadolinium enhancement was suggested to be more sensitive in detecting disease activity than either clinical examination or T2 weighted MRI (Miller et al., 1993). Theoretically, all relapses should be accompanied by new, active enhancing MRI lesions (Figure 1b, d, e g, i, j, l, n o). But, enhancement is dependent on the dose and the technique used, and dependent on the lesion size and location. This must be borne in mind when

Higher doses of gadolinium (triple dose, 0.3mmol/kg body weight), longer delay between injection and the MRI scan (30-60 minutes), and the combination with a magnetization transfer pulse which suppresses the normal brain tissue, but spares areas with enhancement, can increase the sensitivity of MRI, i.e. make an higher number of enhancing MS lesions visualized (Filippi et al., 1996; Miller et al., 1998). Caveat is the increase of false positive results. However it remains unclear, if all lesions, in particular, those with only subtle BBB alterations can be detected hereby (Barnes et al., 1991; Kwon & Prineas et al., 1994, Waubant, 2006). Very often neurological symptoms may occur without any (correlating) acute enhancing MRI lesions, especially in the spinal cord. And most confusing is that acute enhancing MRI lesions that are clinically silent, may be found between relapses, even

though they are located in eloquent areas (Davie et al., 1994; Schubert et al., 2002).

*Put off your imagination as you take off your overcoat, when you enter the laboratory.8* In contrast to usual CNS immune surveillance with regular, always occurring transmigration of only few immune cells (Aboul-Enein et al., 2004), it is supposed that the lesion evolution starts with large numbers of immune cells that cross the BBB into the CNS parenchyma, and thus may lead to BBB leakage (Barnes et al., 1991; Hawkins et al., 1991). This is mainly supported by animal models of experimental autoimmune encephalomyelitis (EAE) that however, must not be uncritically translated to humans. In EAE huge masses of CNS antigen specific immune cells or antibodies, or both, are injected, and initiate an acute, but mostly monophasic inflammatory disease in the CNS (Flugel et al., 2001, 2007; Floris et al., 2004; Hawkins et al., 1991; Kawakami et al., 2005). The situation in humans with MS remains unclear. It is even unknown whether the inflammation is primary, whether the infiltrating immune cells themselves cause the tissue damage, or whether some other primary event within the CNS leads to the actual lesion formation and attracts the inflammatory cells into the CNS. The inflammation may be beneficial or conversely, may cause some bystander damage (Barnett & Prineas, 2004; Henderson et al., 2009; Hohlfeld et al., 2006). Anyway, corticosteroids may decrease or may completely suppress the enhancement of MS lesions for several weeks (Barkhof et al., 1991, 1994; Burnham et al., 1991). Conceivable are various mechanisms of action such as apoptosis of circulating

**3. Of what we believe about MRI lesion evolution** 

**3.1 Acute enhancing lesions** 

different studies are compared.

<sup>8</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

Most of the acute gadolinium enhancing MS lesions change their MR characteristics in further follow up. Some convert to mildly or severly hpointense T1 lesions which might reflect severe underlying tissue damage with marked axonal loss and correlating reduced NAA levels (e.g. van Walderveen et al., 1998), others convert to persistent T2 lesions after several months. The latter may appear hyperintense in T1 weighted images or become undectable in T1 weighted images. In persistent T2 lesions, no matter what behavior they had in T1, nearly all thinkable metabolite constellations were described (increase, decrease, no changes of NAA, Cr, Cho, mIns etc., e.g. Davie et al., 1997; van Walderveen et al., 1997.). Reduced NAA levels might reflect axonal injury, dysfunction or loss. They might be temporarily or permanently, within lesions or remote from lesions due to secondary axonal degeneration or Wallerian degeneration. (Increased NAA levels have yet been described in Canavan's disease only). Elevated Cr, mIns levels were suggested to reflect 'higher metabolism' in astrocytes, and increased Cho was suggested to reflect higher cell membrane turnover or proliferation of astrocytes within MS lesions or NAWM. Decreased levels might be due to decreased cell numbers or lower 'cellular energy state'. These divergent results may be explained by the very small and heterogeneous patient cohorts (e.g. 2 RRMS and 12 SPMS aged from 28 to 61 years, and 4 controls aged 33 to 48 years in van Walderveen et al., 1999), by the use of a 1.5 Tesla 1H-MRS and by the measurement of metabolite ratios instead of absolute metabolite concentrations.

Only single reports of histopathological correlations to *in vivo* MR data have been published (Bitsch et al., 1997) that are of most importance for our understanding, but must be interpreted very carefully. Biopsies are only very seldom performed in MS patients, mostly if tumor lesions must be ruled out or verified. Whether demyelinating tumor-like MS lesions have the same characteristics and behavior like common unspecific smaller MS lesions, is unclear. Moreover, it is still unknown, if MS is one disease or a syndrome of many disorders, or if the same trigger causes the same pathological changes in different MS patients or if the same trigger at different time points leads to different types of lesions in one individual MS patient. Therefore, any general interpretation must be done with caution.

> *Men who have excessive faith in their theories or ideas are not only ill prepared for making discoveries. They also make very poor observations.9*

#### **4. Of what we believe to see with MRT and MRS**

MRI allows the detailed visualization of the anatomical and pathological structures based on different signals of freely mobile hydrogen protons between different tissues or anatomical compartments.

<sup>9</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

MR Spectroscopy in Multiple Sclerosis - A New Piece of the Puzzle or Just a New Puzzle 59

Only **38** original articles providing MRS data (NAA/Cr ratios and/or NAA absolute concentrations) of RRMS or CIS patients could be further analysed. Other search results had to be excluded because they included MRS studies on SPMS or PPMS patients only (Leary et al., 1999; Cucurella et al., 2000; Sastre-Garriga et al., 2005), or did not compare the NAWM of matched controls (Hiehle et al., 1994), or were reviews, post mortem studies, MRS studies on patients with neuromyelitis optica, systemic lupus erythematodes or relatives of MS patients

A meta-analysis of the remaining 38 identified MRS studies could not be performed as they differ from study design (inclusion criteria and technical parameters), and do not provide all necessary data such as detailed demographic data and individual values for metabolite ratios and concentrations. However, following characteristics of the NAWM of CIS and RRMS patients could be identified (table 2, references are listed chronologically by

who were not afflicted with MS.

Table 2. Synopsis of literature research.

publication date).

The signals from freely mobile hydrogen protons are up to 1000 fold more than the resonances from other brain metabolites. The selective suppression of the signal from hydrogen protons 'unmasks' the spectra of various specific metabolites within the magnetic resonance frequency spectrum, and thus allows to detect and separate them, as each brain metabolite resonates at a characteristic position along the frequency scale expressed in parts per million (ppm). The sharpness and the amplitude of a specific peak along the frequency scale is determined by the specific metabolite itself, its concentration and most importantly, by its mobility. The brain metabolites must be freely mobile to produce well detectable peaks. In general, the more freely mobile specific metabolites in a specific volume of interest (VOI) resonate the higher their resulting peak is. If the specific peak is low, or even absent, this could mean that the protons in the examined VOI are either not occurring or not freely mobile or might require specific MR technique/parameters to be detected.

For instance, lipids are abundant within the brain, but under normal conditions immobile, as they constitute mainly highly compacted myelin. If myelin breaks down, they become mobile and may produce a 'free lipid peak' in the frequency spectrum at 0.9-1.2 ppm.

'NAA-peak' reduction or reduction of NAA absolute concentration in the spectroscopied tissue may occur due to axonal injury, damage or even loss, and might be irreversible or reversible. Infarcts, tumors, traumatic injury may lead to severe tissue destruction, heavy axonal damage and loss of axons, and thus may lead to irreversible NAA reduction. In the case of temporarily NAA reduction, various scenarios are conceivable: firstly, the NAA reduction correlates with structural axonal damage, i.e. axonal loss that in principle may be restored by remyelination and regain more or less the same NAA concentrations as before. Secondly, NAA reduction might reflect mitochondrial dysfunction in neurons or axons, whatever the underlying cause might be (e.g. inflammatory mediators, or any other molecule which might interfere with mitochondrial function), and might be restored when the interfering stimulus or agent is removed. The numbers of axons can be unchanged, or can be markedly reduced but have the same NAA levels, i.e. more 'active mitochondria' per axon or neuron. In other words, the loss of axons might be compensated with increase of numbers of mitochondria or increase of the NAA content in axonal mitochondria. (It was suggested that NAA might play a crucial role in the mitochondrial function).

A tumor or tumor-like brain lesion may either show profound reduction of NAA and relative high Cr concentrations in a first baseline scan. Several weeks later the NAA levels might be found still reduced or restored. In the first case, a tumor must be considered, whereas in the latter case a demyelinating tumor-like lesion with subsequent remyelination seems conceivable. (Tumors do not recover spontaneously).

> *A discovery is generally an unforeseen relation not included in theory, for otherwise it would be foreseen.10*

#### **5. Of what could be found with MRS in early stages of RRMS or even in CIS**

Medline Database Research using following search term/key words {"multiple sclerosis" AND (NAWM OR "normal appearing white matter") AND ("N-acetyl-aspartate" OR NAA) and ("spectroscopy") and "absolute concentration"} yielded 67 results (until March 2011).

<sup>10</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

The signals from freely mobile hydrogen protons are up to 1000 fold more than the resonances from other brain metabolites. The selective suppression of the signal from hydrogen protons 'unmasks' the spectra of various specific metabolites within the magnetic resonance frequency spectrum, and thus allows to detect and separate them, as each brain metabolite resonates at a characteristic position along the frequency scale expressed in parts per million (ppm). The sharpness and the amplitude of a specific peak along the frequency scale is determined by the specific metabolite itself, its concentration and most importantly, by its mobility. The brain metabolites must be freely mobile to produce well detectable peaks. In general, the more freely mobile specific metabolites in a specific volume of interest (VOI) resonate the higher their resulting peak is. If the specific peak is low, or even absent, this could mean that the protons in the examined VOI are either not occurring or not freely

For instance, lipids are abundant within the brain, but under normal conditions immobile, as they constitute mainly highly compacted myelin. If myelin breaks down, they become mobile and may produce a 'free lipid peak' in the frequency spectrum at 0.9-1.2 ppm.

'NAA-peak' reduction or reduction of NAA absolute concentration in the spectroscopied tissue may occur due to axonal injury, damage or even loss, and might be irreversible or reversible. Infarcts, tumors, traumatic injury may lead to severe tissue destruction, heavy axonal damage and loss of axons, and thus may lead to irreversible NAA reduction. In the case of temporarily NAA reduction, various scenarios are conceivable: firstly, the NAA reduction correlates with structural axonal damage, i.e. axonal loss that in principle may be restored by remyelination and regain more or less the same NAA concentrations as before. Secondly, NAA reduction might reflect mitochondrial dysfunction in neurons or axons, whatever the underlying cause might be (e.g. inflammatory mediators, or any other molecule which might interfere with mitochondrial function), and might be restored when the interfering stimulus or agent is removed. The numbers of axons can be unchanged, or can be markedly reduced but have the same NAA levels, i.e. more 'active mitochondria' per axon or neuron. In other words, the loss of axons might be compensated with increase of numbers of mitochondria or increase of the NAA content in axonal mitochondria. (It was

A tumor or tumor-like brain lesion may either show profound reduction of NAA and relative high Cr concentrations in a first baseline scan. Several weeks later the NAA levels might be found still reduced or restored. In the first case, a tumor must be considered, whereas in the latter case a demyelinating tumor-like lesion with subsequent remyelination

> *A discovery is generally an unforeseen relation not included in theory, for otherwise it would be foreseen.10*

**5. Of what could be found with MRS in early stages of RRMS or even in CIS**  Medline Database Research using following search term/key words {"multiple sclerosis" AND (NAWM OR "normal appearing white matter") AND ("N-acetyl-aspartate" OR NAA) and ("spectroscopy") and "absolute concentration"} yielded 67 results (until March 2011).

mobile or might require specific MR technique/parameters to be detected.

suggested that NAA might play a crucial role in the mitochondrial function).

seems conceivable. (Tumors do not recover spontaneously).

<sup>10</sup> **Claude Bernard** (French physiologist, 1813 – 1878)

Only **38** original articles providing MRS data (NAA/Cr ratios and/or NAA absolute concentrations) of RRMS or CIS patients could be further analysed. Other search results had to be excluded because they included MRS studies on SPMS or PPMS patients only (Leary et al., 1999; Cucurella et al., 2000; Sastre-Garriga et al., 2005), or did not compare the NAWM of matched controls (Hiehle et al., 1994), or were reviews, post mortem studies, MRS studies on patients with neuromyelitis optica, systemic lupus erythematodes or relatives of MS patients who were not afflicted with MS.

A meta-analysis of the remaining 38 identified MRS studies could not be performed as they differ from study design (inclusion criteria and technical parameters), and do not provide all necessary data such as detailed demographic data and individual values for metabolite ratios and concentrations. However, following characteristics of the NAWM of CIS and RRMS patients could be identified (table 2, references are listed chronologically by publication date).


Table 2. Synopsis of literature research.

MR Spectroscopy in Multiple Sclerosis - A New Piece of the Puzzle or Just a New Puzzle 61

**Seven** out of 12 MRS studies reported **increased** levels of mIns in the NAWM of in a total of **96 CIS** or **116 RRMS** patients (Kapeller et al., 2001; Chard et al., 2002; Fernando et al., 2004;

Keep in mind, that in most of the cases statistical characteristics of groups of MS patients were reported. Only very rarely the presentation of the data allow the readers their own interpretation because data are summarized in tables (with means, standard errors or standard deviations, range) and bar charts, instead of tables and scatter plots presenting all data for each individual patients. That groups of patients differ significantly, does not necessarily mean that all patients differ significantly from age and sex matched normal controls. They even may lie within normal range. For instance, the NAA levels of RRMS patients and even a proportion of SPMS patients were found within the range of NAA absolute concentrations that were found in healthy controls (Aboul-Enein et al., 2010; fig.3 squares, RRMS patients (n=27); circles, SPMS patients (n=10); triangles, controls (n=8); bars, means). However the mean NAA levels of SPMS patients were found significantly

Fig. 3. MRS absolute concentrations of brain metabolites in the NAWM of MS patients and controls (from Aboul-Enein F et al., 2010). White, NAA [mM]; grey, Cr [mM]; black, Cho [mM].

Currently the potpourri of data allows no reliable general information or even evidence about metabolite changes in the NAWM that were hypothesized to occur in all CIS or RRMS patients. Sophisticated meta-analysis is yet not possible as different MR methods were used, and individual data such as age, sex, disease duration, disease activity, lesion load, therapy, and metabolites concentrations cannot be extracted. Moreover, it remains unclear, what the specific metabolites reflect in the tissue of individual MS patients. In any case, we must be

**5.3.2 Significant changes of mIns ratios or mIns concentrations** 

decreased compared to controls and RRMS patients.

0,0

**6. Conclusion** 

2,0

4,0

6,0

8,0

Concentrations [mM]

10,0

12,0

14,0

16,0

He et al., 2005; Vrenken et al., 2005; Srinivasan et al., 2005; Kirov et al., 2008).

#### **5.1 Reported NAA changes**

#### **5.1.1 Decreased NAA ratios or NAA concentrations**

**Twenty-four MRS studies** were published with decreased ratios of NAA to Cr or Cho, or decreased NAA absolute concentrations in the NAWM in a total of **56 CIS** or **418 RRMS** patients (Davie et al., 1994; Husted et al., 1994; Peters et al., 1995; Fu et al., 1998; Schiepers et al., 1997; Foong J et al., 1999; Sarchielli et al., 1999; van Walderveen et al., 1999; Suhy et al., 2000; de Stefano et al., 2001; Tedeschi et al., 2001; Kapeller et al., 2001; Chard et al., 2002; de Stefano et al., 2002; Casanova et al. 2003; Inglese et al., 2004; Adalsteinsson et al., 2004; He et al., 2005; Staffen et al., 2005; Tiberio et al., 2006; Sijens et al., 2006; Pascual et al., 2007; Wattjes et al., 2008; Wattjes et al., 2008).

#### **5.1.2 No significant changes of NAA ratios or NAA concentrations**

**Fourteen MRS studies** were published with no significant NAA changes in the NAWM in a total of **184 CIS** or **263 RRMS** patients (Davie et al., 1997; Sarchielli et al., 1998; Brex et al., 1999; Kapeller et al., 2002; Fernando et al., 2004; Ruiz-Pena et al., 2004; Vrenken et al., 2005; Mathiesen et al., 2005; Srinivasan et al., 2005; Khan et al., 2005; Khan et al., 2008; Kirov et al., 2008; Bellmann-Strobl et al., 2009; Aboul-Enein et al., 2010).

#### **5.2 Reported Cho changes**

#### **5.2.1 No significant changes of Cho ratios or Cho concentrations**

**Twenty-nine MRS studies** were published with no significant Cho changes in the NAWM in a total of **190 CIS** or **513 RRMS** patients (Husted et al., 1994; Peters et al., 1995; Schiepers et al., 1997;Davie et al., 1997; Sarchielli et al., 1999; Foong J et al., 1999; Sarchielli et al., 1999; Brex et al., 1999; van Walderveen et al., 1999; Suhy et al., 2000; de Stefano et al., 2001; Kapeller et al., 2001; Chard et al., 2002; de Stefano et al., 2002; Kapeller et al., 2002; Casanova et al. 2003; Inglese et al., 2004; Adalsteinsson et al., 2004; Fernando et al., 2004; Ruiz-Pena et al., 2004; Vrenken et al., 2005; Mathiesen et al., 2005; Khan et al., 2005; Sijens et al., 2006; Wattjes et al., 2008; Wattjes et al., 2008; Khan et al., 2008; Bellmann-Strobl et al., 2009; Aboul-Enein et al., 2010).

#### **5.2.2 Significant changes of Cho ratios or Cho concentrations**

In **two MRS studies** (**8** and **19 RRMS** patients) Cho was found decreased (Davie et al., 1994; Tedeschi et al., 2001) and in **4 MRS studies** (**9, 16**, **43** and **21 RRMS** patients) Cho levels were found increased (He et al., 2005; Srinivasan et al., 2005; Pascual et al., 2007; Kirov et al., 2008).

#### **5.3 Reported mIns changes**

#### **5.3.1 No significant changes of mIns ratios or mIns concentrations**

**Five** out of 12 MRS studies reported no significant mINS changes in the NAWM of in a total of **76 CIS** or **17 RRMS** patients (Brex et al., 1999; Sarchielli et al., 1998; Sijens et al., 2006; Wattjes et al., 2008a; Wattjes et al., 2008b).

#### **5.3.2 Significant changes of mIns ratios or mIns concentrations**

**Seven** out of 12 MRS studies reported **increased** levels of mIns in the NAWM of in a total of **96 CIS** or **116 RRMS** patients (Kapeller et al., 2001; Chard et al., 2002; Fernando et al., 2004; He et al., 2005; Vrenken et al., 2005; Srinivasan et al., 2005; Kirov et al., 2008).

Keep in mind, that in most of the cases statistical characteristics of groups of MS patients were reported. Only very rarely the presentation of the data allow the readers their own interpretation because data are summarized in tables (with means, standard errors or standard deviations, range) and bar charts, instead of tables and scatter plots presenting all data for each individual patients. That groups of patients differ significantly, does not necessarily mean that all patients differ significantly from age and sex matched normal controls. They even may lie within normal range. For instance, the NAA levels of RRMS patients and even a proportion of SPMS patients were found within the range of NAA absolute concentrations that were found in healthy controls (Aboul-Enein et al., 2010; fig.3 squares, RRMS patients (n=27); circles, SPMS patients (n=10); triangles, controls (n=8); bars, means). However the mean NAA levels of SPMS patients were found significantly decreased compared to controls and RRMS patients.

Fig. 3. MRS absolute concentrations of brain metabolites in the NAWM of MS patients and controls (from Aboul-Enein F et al., 2010). White, NAA [mM]; grey, Cr [mM]; black, Cho [mM].

#### **6. Conclusion**

60 Magnetic Resonance Spectroscopy

**Twenty-four MRS studies** were published with decreased ratios of NAA to Cr or Cho, or decreased NAA absolute concentrations in the NAWM in a total of **56 CIS** or **418 RRMS** patients (Davie et al., 1994; Husted et al., 1994; Peters et al., 1995; Fu et al., 1998; Schiepers et al., 1997; Foong J et al., 1999; Sarchielli et al., 1999; van Walderveen et al., 1999; Suhy et al., 2000; de Stefano et al., 2001; Tedeschi et al., 2001; Kapeller et al., 2001; Chard et al., 2002; de Stefano et al., 2002; Casanova et al. 2003; Inglese et al., 2004; Adalsteinsson et al., 2004; He et al., 2005; Staffen et al., 2005; Tiberio et al., 2006; Sijens et al., 2006; Pascual et al., 2007; Wattjes

**Fourteen MRS studies** were published with no significant NAA changes in the NAWM in a total of **184 CIS** or **263 RRMS** patients (Davie et al., 1997; Sarchielli et al., 1998; Brex et al., 1999; Kapeller et al., 2002; Fernando et al., 2004; Ruiz-Pena et al., 2004; Vrenken et al., 2005; Mathiesen et al., 2005; Srinivasan et al., 2005; Khan et al., 2005; Khan et al., 2008; Kirov et al.,

**Twenty-nine MRS studies** were published with no significant Cho changes in the NAWM in a total of **190 CIS** or **513 RRMS** patients (Husted et al., 1994; Peters et al., 1995; Schiepers et al., 1997;Davie et al., 1997; Sarchielli et al., 1999; Foong J et al., 1999; Sarchielli et al., 1999; Brex et al., 1999; van Walderveen et al., 1999; Suhy et al., 2000; de Stefano et al., 2001; Kapeller et al., 2001; Chard et al., 2002; de Stefano et al., 2002; Kapeller et al., 2002; Casanova et al. 2003; Inglese et al., 2004; Adalsteinsson et al., 2004; Fernando et al., 2004; Ruiz-Pena et al., 2004; Vrenken et al., 2005; Mathiesen et al., 2005; Khan et al., 2005; Sijens et al., 2006; Wattjes et al., 2008; Wattjes et al., 2008; Khan et al., 2008; Bellmann-Strobl et al., 2009; Aboul-

In **two MRS studies** (**8** and **19 RRMS** patients) Cho was found decreased (Davie et al., 1994; Tedeschi et al., 2001) and in **4 MRS studies** (**9, 16**, **43** and **21 RRMS** patients) Cho levels were found increased (He et al., 2005; Srinivasan et al., 2005; Pascual et al., 2007;

**Five** out of 12 MRS studies reported no significant mINS changes in the NAWM of in a total of **76 CIS** or **17 RRMS** patients (Brex et al., 1999; Sarchielli et al., 1998; Sijens et al., 2006;

**5.1 Reported NAA changes** 

et al., 2008; Wattjes et al., 2008).

**5.2 Reported Cho changes** 

Enein et al., 2010).

Kirov et al., 2008).

**5.3 Reported mIns changes** 

Wattjes et al., 2008a; Wattjes et al., 2008b).

**5.1.1 Decreased NAA ratios or NAA concentrations** 

**5.1.2 No significant changes of NAA ratios or NAA concentrations** 

**5.2.1 No significant changes of Cho ratios or Cho concentrations** 

**5.2.2 Significant changes of Cho ratios or Cho concentrations** 

**5.3.1 No significant changes of mIns ratios or mIns concentrations** 

2008; Bellmann-Strobl et al., 2009; Aboul-Enein et al., 2010).

Currently the potpourri of data allows no reliable general information or even evidence about metabolite changes in the NAWM that were hypothesized to occur in all CIS or RRMS patients. Sophisticated meta-analysis is yet not possible as different MR methods were used, and individual data such as age, sex, disease duration, disease activity, lesion load, therapy, and metabolites concentrations cannot be extracted. Moreover, it remains unclear, what the specific metabolites reflect in the tissue of individual MS patients. In any case, we must be

MR Spectroscopy in Multiple Sclerosis - A New Piece of the Puzzle or Just a New Puzzle 63

Barkhof F, Hommes OR, Scheltens P, Valk J. (1991). Quantitative MRI changes in

Barkhof F, Tas MW, Frequin ST, Scheltens P, Hommes OR, Nauta JJ, Valk J. (1994). Limited

Barnett MH, Prineas JW. (2004). Relapsing and remitting multiple sclerosis: pathology of the

Barnett MH, Parratt JD, Pollard JD, Prineas JW. (2009). MS: is it one disease? *Int MS J,* Vol.

Benedetti B, Rovaris M, Rocca MA, Caputo D, Zaffaroni M, Capra R, Bertolotto A, Martinelli

Bitsch A, Bruhn H, Vougioukas V, Stringaris A, Lassmann H, Frahm J, Brück W. (1999).

Brex PA, Gomez-Anson B, Parker GJ, Molyneux PD, Miszkiel KA, Barker GJ, MacManus

Brex PA, Ciccarelli O, O'Riordan JI, Sailer M, Thompson AJ, Miller DH. (2002). A

*N Engl J Med,* Vol. 346, No. 3, (January 2002), pp. 158-164, ISSN 0028-4793. Bruhn H, Frahm J, Gyngell ML, Merboldt KD, Hänicke W, Sauter R. (1989). Cerebral

Burnham JA, Wright RR, Dreisbach J, Murray RS. (1991). The effect of high-dose steroids on

Casanova B, Martínez-Bisbal MC, Valero C, Celda B, Martí-Bonmatí L, Pascual A, Landente

Chard DT, Griffin CM, McLean MA, Kapeller P, Kapoor R, Thompson AJ, Miller DH. (2002).

*Neurol,* Vol. 250, No. 1, (January 2003), pp. 22-28, ISSN 0340-5354.

No. 9, (September 1991), pp. 1349-1354. ISSN 0028-3878.

(June 1991), pp. 1271-1280; ISSN 0006-8950.

16, No. 2, (June 2009), pp. 57-65, ISSN 1352-8963.

(October 1999), pp. 1619-1627, ISSN 0195-6108.

(June 1999), pp. 16-22, ISSN 1300-1817.

2009), pp. 789-794, ISSN 1352-4585.

5134.

0028-3878.

0364-5134.

ISSN 1522-2594.

in multiple sclerosis. *Ann Neurol,* Vol. 36, No. 1, (July 1994), pp. 76-82, ISSN 0364-

gadolinium-DTPA enhancement after high-dose intravenous methylprednisolone in multiple sclerosis. *Neurology,* Vol. 41, No. 8, (August 1991), pp. 1219-1222, ISSN

duration of the effect of methylprednisolone on changes on MRI in multiple sclerosis. *Neuroradiology,* Vol. 36, No. 5, (July 1994), pp. 382-387, ISSN 0028-3940. Barnes D, Munro PM, Youl BD, Prineas JW, McDonald WI. (1991) The longstanding MS

lesion. A quantitative MRI and electron microscopic study. *Brain,* Vol. 114, No. 3,

newly forming lesion. *Ann Neurol,* Vol. 55, No. 4, (April 2004), pp. 458-468, ISSN

V, Comi G, Filippi M. (2009). In-vivo evidence for stable neuroaxonal damage in the brain of patients with benign multiple sclerosis. *Mult Scler,* Vol. 15, No. 7, (July

Inflammatory CNS demyelination: histopathologic correlation with in vivo quantitative proton MR spectroscopy. *AJNR Am J Neuroradiol*, Vol. 20, No. 9,

DG, Davie CA, Plant GT, Miller DH. (1999). Proton MR spectroscopy in clinically isolated syndromes suggestive of multiple sclerosis. *J Neurol Sci,* Vol. 166, No. 1,

longitudinal study of abnormalities on MRI and disability from multiple sclerosis.

metabolism in man after acute stroke: new observations using localized proton NMR spectroscopy. *Magn Reson Med*, Vol. 9, No. 1, (January 1989), pp. 126-131,

MRI gadolinium enhancement in acute demyelinating lesions. *Neurology,* Vol. 41,

L, Coret F. (2003). Evidence of Wallerian degeneration in normal appearing white matter in the early stages of relapsing-remitting multiple sclerosis: a HMRS study. *J* 

Brain metabolite changes in cortical grey and normal-appearing white matter in

aware that the magnetic resonances of freely mobile hydrogen protons or certain other freely mobile metabolites are converted to images, and thus 'are made visible for the human's eye'.

#### **7. Acknowledgment**

The author is thankful to Atbin for substantial help. In love, dedicated to my little daughter and my wife.

#### **8. References**


aware that the magnetic resonances of freely mobile hydrogen protons or certain other freely mobile metabolites are converted to images, and thus 'are made visible for the

The author is thankful to Atbin for substantial help. In love, dedicated to my little daughter

Aboul-Enein F, Krssák M, Höftberger R, Prayer D, Kristoferitsch W. (2010). Reduced NAA-

Aboul-Enein F, Bauer J, Klein M, Schubart A, Flügel A, Ritter T, Kawakami N, Siedler F,

Aboul-Enein F, Lassmann H. (2005). Mitochondrial damage and histotoxic hypoxia: a

Aboul-Enein F, Weiser P, Höftberger R, Lassmann H, Bradl M. (2006). Transient axonal

Aboul-Enein F, Krssák M, Höftberger R, Prayer D, Kristoferitsch W. (2010). Diffuse white

Adalsteinsson E, Langer-Gould A, Homer RJ, Rao A, Sullivan EV, Lima CA, Pfefferbaum A,

Allen IV, McKeown SR. (1979). A histological, histochemical and biochemical study of the

Allen IV, Glover G, Anderson R. (1981). Abnormalities in the macroscopically normal white

Arnold DL, Matthews PM, Francis G, Antel J. (1990). Proton magnetic resonance

Arnold DL, Riess GT, Matthews PM, Francis GS, Collins DL, Wolfson C, Antel JP. (1994).

levels in the NAWM of patients with MS is a feature of progression. A study with quantitative magnetic resonance spectroscopy at 3 Tesla. *PLoS One,* Vol. 5, No. 7,

Linington C, Wekerle H, Lassmann H, Bradl M. (2004). Selective and antigendependent effects of myelin degeneration on central nervous system inflammation. *J Neuropathol Exp Neurol,* Vol. 63, No. 12, (December 2004), pp. 1284-1296, ISSN

pathway of tissue injury in inflammatory brain disease? *Acta Neuropathol,* Vol. 109,

injury in the absence of demyelination: a correlate of clinical disease in acute experimental autoimmune encephalomyelitis. *Acta Neuropathol,* Vol. 111, No. 6,

matter damage is absent in neuromyelitis optica. *AJNR Am J Neuroradiol,* Vol. 31,

Atlas SW. (2003). Gray matter N-acetyl aspartate deficits in secondary progressive but not relapsing-remitting multiple sclerosis. *AJNR Am J Neuroradiol*, Vol. 24, No.

macroscopically normal white matter in multiple sclerosis. *J Neurol Sci,* Vol. 41, No.

matter in cases of mild or spinal multiple sclerosis (MS). *Acta Neuropathol,* Vol. 7,

spectroscopy of human brain in vivo in the evaluation of multiple sclerosis: assessment of the load of disease. *Magn Reson Med,* Vol. 14 , No 1., (April 1990), pp.

Use of proton magnetic resonance spectroscopy for monitoring disease progression

human's eye'.

and my wife.

**8. References** 

**7. Acknowledgment** 

0022-3069.

(July 2010), e11625, ISSN 1932-6203.

No. 1, (January 2005), pp. 49-55, ISSN 0001-6322.

No. 1, (January 2010), pp. 76-79, ISSN 0195-6108.

10, (November 2003), pp. 1941-1945, ISSN 0195-6108.

No. Supplement, (1981), pp. 176-178, ISSN 0001-6322.

1. (March 1979), pp. 81-91, ISSN 1300-1817.

154-159, ISSN 1522-2594.

(June 2006), pp. 539-547, ISSN 0001-6322.

in multiple sclerosis. *Ann Neurol,* Vol. 36, No. 1, (July 1994), pp. 76-82, ISSN 0364- 5134.


MR Spectroscopy in Multiple Sclerosis - A New Piece of the Puzzle or Just a New Puzzle 65

Flügel A, Berkowicz T, Ritter T, Labeur M, Jenne DE, Li Z, Ellwart JW, Willem M, Lassmann

Flügel A, Odoardi F, Nosov M, Kawakami N. (2007). Autoaggressive effector T cells in the

Foong J, Rozewicz L, Davie CA, Thompson AJ, Miller DH, Ron MA. (1999). Correlates of

Frahm J, Bruhn H, Gyngell ML, Merboldt KD, Hänicke W, Sauter R. (1989a). Localized high-

Frahm J, Bruhn H, Gyngell ML, Merboldt KD, Hänicke W, Sauter R. (1989b). Localized

Fu L, Matthews PM, De Stefano N, Worsley KJ, Narayanan S, Francis GS, Antel JP, Wolfson

Gelati M, Corsini E, De Rossi M, Masini L, Bernardi G, Massa G, Boiardi A, Salmaggi A.

Grossman RI, Lenkinski RE, Ramer KN, Gonzalez-Scarano F, Cohen JA. (1992). MR proton

Hawkins CP, Mackenzie F, Tofts P, du Boulay EP, McDonald WI. (1991). Patterns of blood-

He J, Inglese M, Li BS, Babb JS, Grossman RI, Gonen O. (2005). Relapsing-remitting multiple

Henderson AP, Barnett MH, Parratt JD, Prineas JW. (2009). Multiple sclerosis: distribution of

*Arch Neurol,* Vol. 59, No. 5, (May 2002), pp. 774-780, ISSN 0003-9942. Gilmore CP, Cottrell DA, Scolding NJ, Wingerchuk DM, Weinshenker BG, Boggild M.

*Scler,* Vol. 16, No. 6, (June 2010), pp. 756-759, ISSN 1352-4585.

(November-December 1992), pp. 1535-1343, ISSN 0195-6108.

(April 1991), pp. 801-810. ISSN 0006-8950.

2005), pp. 211-217, ISSN 0033-8419.

2009); pp. 739-753, ISSN 0364-5134.

7613.

86-97, ISSN 0165-5728.

ISSN 1522-2594.

8950.

No. 1, (Winter 1999), pp. 45-50, ISN 0895-0172.

11, No. 1, (July 1989), pp. 47-63, ISSN 1522-2594.

H, Wekerle H. (2001). Migratory activity and functional changes of green fluorescent effector cells before and during experimental autoimmune encephalomyelitis. *Immunity,* Vol. 14, No. 5, (May 2001), pp. 547-560, ISSN 1074-

course of experimental autoimmune encephalomyelitis visualized in the light of two-photon microscopy. *J Neuroimmunol,* Vol. 191, No. 1-2, (November 2007), pp.

executive function in multiple sclerosis: the use of magnetic resonance spectroscopy as an index of focal pathology. *J Neuropsychiatry Clin Neurosci*, Vol. 11,

resolution proton NMR spectroscopy using stimulated echoes: initial applications to human brain in vivo. *Magn Reson Med,* Vol. 9, No. 1, (January 1989), pp. 79-93,

proton NMR spectroscopy in different regions of the human brain in vivo. Relaxation times and concentrations of cerebral metabolites. *Magn Reson Med,* Vol.

C, Arnold DL. (1998). Imaging axonal damage of normal-appearing white matter in multiple sclerosis. *Brain,* Vol. 121, No. 1, (January 1998), pp. 103-113, ISSN 0006-

(2002). Methylprednisolone acts on peripheral blood mononuclear cells and endothelium in inhibiting migration phenomena in patients with multiple sclerosis.

(2010). A window of opportunity for no treatment in early multiple sclerosis? *Mult* 

spectroscopy in multiple sclerosis. *AJNR Am J Neuroradiol,* Vol. 13, No. 6,

brain barrier breakdown in inflammatory demyelination. *Brain,* Vol. 114, No. 2,

sclerosis: metabolic abnormality in nonenhancing lesions and normal-appearing white matter at MR imaging: initial experience. *Radiology,* Vol. 234, No. 1, (January

inflammatory cells in newly forming lesions. *Ann Neurol,* Vol. 66, No. 6, (December

clinically early relapsing-remitting multiple sclerosis. *Brain,* Vol. 125, No. 10, (October 2002), pp. 2342-2352, ISSN 0006-8950.


Chataway J. (2008). When the MRI scan suggests multiple sclerosis but the symptoms do

Chataway J. (2010). When confronted by a patient with the radiologically isolated syndrome. *Pract Neurol,* Vol. 10, No. 5, (October 2010), pp. 271-277, ISSN 1474-7758. Davie CA, Hawkins CP, Barker GJ, Brennan A, Tofts PS, Miller DH, McDonald WI. (1994).

Davie CA, Barker GJ, Thompson AJ, Tofts PS, McDonald WI, Miller DH. (1997). 1H

De Stefano N, Matthews PM, Fu L, Narayanan S, Stanley J, Francis GS, Antel JP, Arnold DL.

De Stefano N, Narayanan S, Francis GS, Arnaoutelis R, Tartaglia MC, Antel JP, Matthews

De Stefano N, Narayanan S, Francis SJ, Smith S, Mortilla M, Tartaglia MC, Bartolozzi ML,

Engelhardt B. (2006). Molecular mechanisms involved in T cell migration across the blood-

Falini A, Calabrese G, Filippi M, Origgi D, Lipari S, Colombo B, Comi G, Scotti G. (1998).

Filippi M, Yousry T, Campi A, Kandziora C, Colombo B, Voltz R, Martinelli V, Spuler S,

Floris S, Blezer EL, Schreibelt G, Döpp E, van der Pol SM, Schadee-Eestermans IL, Nicolay

*Brain,* Vol. 127, No. 3, (March 2004), pp. 616-627, ISSN 0006-8950.

*Neurology*, Vol. 46, No. 2, (February 1996), pp. 379-384, ISSN 0028-3878. Fisniku LK, Brex PA, Altmann DR, Miszkiel KA, Benton CE, Lanyon R, Thompson AJ,

Vol. 59, No. 10, (October 2002), pp. 1565-1571, ISSN 0003-9942.

19, No. 2, (February 1998), pp. 223-229, ISSN 0195-6108.

*Brain,* Vol. 117, (February 1994), pp. 49-58, ISSN 0006-8950.

Vol. 63, No. 6, (December 1997), pp. 736-742, ISSN 0022-3050.

(October 2002), pp. 2342-2352, ISSN 0006-8950.

0022-3050.

8950.

9564.

pp. 65-70, ISSN 0003-9942.

808-817, ISSN 0006-8950.

clinically early relapsing-remitting multiple sclerosis. *Brain,* Vol. 125, No. 10,

not. *J Neurol Neurosurg Psychiatry*, Vol. 79, No. 2, (February 2008), pp. 112-113, ISSN

Serial proton magnetic resonance spectroscopy in acute multiple sclerosis lesions.

magnetic resonance spectroscopy of chronic cerebral white matter lesions and normal appearing white matter in multiple sclerosis. *J Neurol Neurosurg Psychiatry,* 

(1998). Axonal damage correlates with disability in patients with relapsingremitting multiple sclerosis. Results of a longitudinal magnetic resonance spectroscopy study. *Brain,* Vol. 121, No. 8, (August 1998), pp. 1469-1477, ISSN 0006-

PM, Arnold DL. (2001). Evidence of axonal damage in the early stages of multiple sclerosis and its relevance to disability. *Arch Neurol,* Vol. 58, No. 1, (January 2001),

Guidi L, Federico A, Arnold DL. (2002). Diffuse axonal and tissue injury in patients with multiple sclerosis with low cerebral lesion load and no disability. *Arch Neurol,*

brain barrier. *J Neural Transm,* Vol. 113, No. 4, (April 2006), pp. 477-485, ISSN 0300-

Benign versus secondary-progressive multiple sclerosis: the potential role of proton MR spectroscopy in defining the nature of disability. *AJNR Am J Neuroradiol*, Vol.

Bressi S, Scotti G, Comi G. (1996). Comparison of triple dose versus standard dose gadolinium-DTPA for detection of MRI enhancing lesions in patients with MS.

Miller DH. (2008). Disability and T2 MRI lesions: a 20-year follow-up of patients with relapse onset of multiple sclerosis. *Brain,* Vol. 131, No. 3, (March 2008), pp.

K, Dijkstra CD, de Vries HE. (2004). Blood-brain barrier permeability and monocyte infiltration in experimental allergic encephalomyelitis: a quantitative MRI study.


MR Spectroscopy in Multiple Sclerosis - A New Piece of the Puzzle or Just a New Puzzle 67

Kirov II, Patil V, Babb JS, Rusinek H, Herbert J, Gonen O. (2009). MR spectroscopy indicates

Khan O, Shen Y, Caon C, Bao F, Ching W, Reznar M, Buccheister A, Hu J, Latif Z, Tselis A,

Khan O, Shen Y, Bao F, Caon C, Tselis A, Latif Z, Zak I. (2008). Long-term study of brain 1H-

Kwon EE, Prineas JW. (1994). Blood-brain barrier abnormalities in longstanding multiple

Lassmann H, Brück W, Lucchinetti CF. (2007). The immunopathology of multiple sclerosis:

Lebrun C, Bensa C, Debouverie M, De Seze J, Wiertlievski S, Brochet B, Clavelou P, Brassat

Leussink VI, Jung S, Merschdorf U, Toyka KV, Gold R. (2001). High-dose

Lublin FD, Reingold SC. (1996).Defining the clinical course of multiple sclerosis: results of

Lumsden CE. (1970). The neuropathology of multiple sclerosis, In: *Handbook of Clinical* 

Mathiesen HK, Tscherning T, Sorensen PS, Larsson HB, Rostrup E, Paulson OB, Hanson LG.

Vol. 80, No. 12, (December 2009), pp. 1330-1336, ISSN 0022-3050.

6, (December 2005), pp. 646-651, ISSN 1352-4585.

No. 6, (November 1994), pp. 625-636, ISSN 0022-3069.

46, No. 4, (April 1996), pp. 907-911, ISSN 0028-3878.

0072-9752, Amsterdam, North-Holland.

(April 2005), pp. 750-759, ISSN 1522-2594.

1999), pp. 1023-1026, ISSN 0340-5354.

0028-4793.

3050.

9942.

diffuse multiple sclerosis activity during remission. J Neurol Neurosurg Psychiatry,

Lisak R. (2005). Axonal metabolic recovery and potential neuroprotective effect of glatiramer acetate in relapsing-remitting multiple sclerosis. *Mult Scler,* Vol. 11, No.

MRS study in multiple sclerosis: effect of glatiramer acetate therapy on axonal metabolic function and feasibility of long-Term H-MRS monitoring in multiple sclerosis. *J Neuroimaging,* Vol. 18, No. 3, (July 2008), pp. 314-319, ISSN 1051-2284. Kuhle J, Pohl C, Mehling M, Edan G, Freedman MS, Hartung HP, Polman CH, Miller DH,

Montalban X, Barkhof F, Bauer L, Dahms S, Lindberg R, Kappos L, Sandbrink R. (2007). Lack of association between antimyelin antibodies and progression to multiple sclerosis. *N Engl J Med,* Vol. 356, No. 4, (January 2007), pp. 371-378, ISSN

sclerosis lesions. An immunohistochemical study. *J Neuropathol Exp Neurol,* Vol. 53,

an overview. *Brain Pathol,* Vol. 17, No. 2, (April 2007), pp. 210-218, ISSN 1015-6305. Leary SM, Davie CA, Parker GJ, Stevenson VL, Wang L, Barker GJ, Miller DH, Thompson

AJ. (1999). 1H magnetic resonance spectroscopy of normal appearing white matter in primary progressive multiple sclerosis. *J Neurol*, Vol. 246, No. 11, (November

D, Labauge P, Roullet E; CFSEP. (2008). Unexpected multiple sclerosis: follow-up of 30 patients with magnetic resonance imaging and clinical conversion profile. *J Neurol Neurosurg Psychiatry,* Vol. 79, No. 2, (February 2008), pp. 195-198, ISSN 0022-

methylprednisolone therapy in multiple sclerosis induces apoptosis in peripheral blood leukocytes. *Arch Neurol,* Vol. 58, No. 1, (January 2001), pp. 91-97, ISSN 0003-

an international survey. National Multiple Sclerosis Society (USA) Advisory Committee on Clinical Trials of New Agents in Multiple Sclerosis. *Neurology,* Vol.

*Neurology,* Vinken PJ, and Bruyn GW (Eds.), Vol. 9, pp. 217-309, Elsevier, ISSN

(2005). Multi-slice echo-planar spectroscopic MR imaging provides both global and local metabolite measures in multiple sclerosis. *Magn Reson Med,* Vol. 53, No. 4,


Hibberd PL. (1994). Use and misuse of statistics for epidemiological studies of multiple

Höftberger R, Aboul-Enein F, Brueck W, Lucchinetti C, Rodriguez M, Schmidbauer M,

Hohlfeld R, Kerschensteiner M, Stadelmann C, Lassmann H, Wekerle H. (2006). The

*Pathol,* Vol. 14, No. 1, (January 2004), pp. 43-50, ISSN 1015-6305.

*Neurology,* Vol. 43, No. 4, (April 1993), pp. 655-661, ISSN 0028-3878. Jacobs L, Kinkel PR, Kinkel WR. (1986). Silent brain lesions in patients with isolated

*Arch Neurol,* Vol. 43, No. 5, (May 1986), pp. 452-455, ISSN 0003-9942. Jackson JA, Leake DR, Schneiders NJ, Rolak LA, Kelley GR, Ford JJ, Appel SH, Bryan RN.

*Neuroradiol*, Vol. 28, No. 2, (February 2007), pp. 262-266, ISSN 0195-6108. Kapeller P, McLean MA, Griffin CM, Chard D, Parker GJ, Barker GJ, Thompson AJ, Miller

Kapeller P, Brex PA, Chard D, Dalton C, Griffin CM, McLean MA, Parker GJ, Thompson AJ,

Kawakami N, Nägerl UV, Odoardi F, Bonhoeffer T, Wekerle H, Flügel A. (2005). Live

Kimura H, Grossman RI, Lenkinski RE, Gonzalez-Scarano F. (1996). Proton MR

17, No. 8, (September 1996), pp. 1539-1547, ISSN 0195-6108.

(February 2001), pp. 131-138, ISSN 0340-5354.

2002), pp. 207-210, ISSN 0340-5354.

2005), pp. 1805-1814, ISSN: 0022-1007.

0364-5134.

sclerosis. *Ann Neurol,* Vol. 36, No. Suppl. 2, (December 1994), pp. S218-230, ISSN

Jellinger K, Lassmann H. (2004). Expression of major histocompatibility complex class I molecules on the different cell types in multiple sclerosis lesions. *Brain* 

neuroprotective effect of inflammation: implications for the therapy of multiple sclerosis. *Neurol Sci,* Vol. 27, No. S1, (March 2006), pp. S1-S7, ISSN 1590-1874. Husted CA, Goodin DS, Hugg JW, Maudsley AA, Tsuruda JS, de Bie SH, Fein G, Matson

GB, Weiner MW. (1994). Biochemical alterations in multiple sclerosis lesions and normal-appearing white matter detected by in vivo 31P and 1H spectroscopic imaging. *Ann Neurol.* Vol. 36, No. 2, (August 1994), pp. 157-165, ISSN 0364-5134. IFNB Multiple Sclerosis Study Group. (11993). Interferon beta-1b is effective in relapsing-

remitting multiple sclerosis. I. Clinical results of a multicenter, randomized, double-blind, placebo-controlled trial. The IFNB Multiple Sclerosis Study Group.

idiopathic optic neuritis. A clinical and nuclear magnetic resonance imaging study.

(1985). Magnetic resonance imaging in multiple sclerosis: results in 32 cases. *AJNR Am J Neuroradiol,* Vol. 6, No. 2, (March-April 1985), pp. 171-176, ISSN 0195-6108. Kangarlu A, Bourekas EC, Ray-Chaudhury A, Rammohan KW. (2007). Cerebral cortical

lesions in multiple sclerosis detected by MR imaging at 8 Tesla. *AJNR Am J* 

DH. (2001). Preliminary evidence for neuronal damage in cortical grey matter and normal appearing white matter in short duration relapsing-remitting multiple sclerosis: a quantitative MR spectroscopic imaging study. *J Neurol,* Vol. 248, No. 2,

Miller DH. Quantitative 1H MRS imaging 14 years after presenting with a clinically isolated syndrome suggestive of multiple sclerosis. *Mult Scler,* Vol. 8, No. 3, (May

imaging of effector cell trafficking and autoantigen recognition within the unfolding autoimmune encephalomyelitis lesion. *J Exp Med,* Vol. 201, No. 11, (June

spectroscopy and magnetization transfer ratio in multiple sclerosis: correlative findings of active versus irreversible plaque disease. *AJNR Am J Neuroradiol,* Vol.


MR Spectroscopy in Multiple Sclerosis - A New Piece of the Puzzle or Just a New Puzzle 69

*Ann Neurol*, Vol. 69, No. 2, (February 2011), pp. 292-302, ISSN 0364-5134. Poser CM, Paty DW, Scheinberg L, McDonald WI, Davis FA, Ebers GC, Johnson KP, Sibley

Prat A, Biernacki K, Lavoie JF, Poirier J, Duquette P, Antel JP. (2002). Migration of multiple

Richards TL. Proton MR spectroscopy in multiple sclerosis: value in establishing diagnosis,

Rooney WD, Goodkin DE, Schuff N, Meyerhoff DJ, Norman D, Weiner MW. (1997). 1H

Roser W, Hagberg G, Mader I, Brunnschweiler H, Radue EW, Seelig J, Kappos L. (1995).

Ruiz-Peña JL, Piñero P, Sellers G, Argente J, Casado A, Foronda J, Uclés A, Izquierdo G.

Sarchielli P, Presciutti O, Pelliccioli GP, Tarducci R, Gobbi G, Chiarini P, Alberti A,

Sastre-Garriga J, Ingle GT, Chard DT, Ramió-Torrentà L, McLean MA, Miller DH,

Schubert F, Seifert F, Elster C, Link A, Walzel M, Mientus S, Haas J, Rinneberg H. (2002).

Schuhmacher GA, Beebe G, Kibler RF, Kurland LT, Kurtzke JF, McDowell F, Nagler B,

Arch Neurol, Vol. 62, No. 4, (April 2005), pp. 569-573, ISSN 0003-9942. Schiepers C, Van Hecke P, Vandenberghe R, Van Oostende S, Dupont P, Demaerel P,

*Scler,* Vol. 3, No. 1, (February 1997), pp. 8-17, ISSN 1352-4585.

157, No. 5, (November 1991), pp. 1073-1078, ISSN 1546-3141.

64, No. 2, (February 1998), pp. 204-212, ISSN 0022-3050.

1983), pp. 227-231, ISSN 0364-5134.

4, (August 1997), pp. 231-237, ISSN

811-817, ISSN 1522-2594.

0006-8950.

2002), pp. 213-222,

(March 2002), pp. 391-397, ISSN 0003-9942.

.Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria.

WA, Silberberg DH, Tourtellotte WW. (1983). New diagnostic criteria for multiple sclerosis: guidelines for research protocols. *Ann Neurol,* Vol. 13, No. 3, (March

sclerosis lymphocytes through brain endothelium. *Arch Neurol,* Vol. 59, No. 3,

monitoring progression, and evaluating therapy. (1991). *AJR Am J Roentgenol,* Vol.

MRSI of normal appearing white matter in multiple sclerosis. *Mult Scler,* Vol. 3, No.

Proton MRS of gadolinium-enhancing MS plaques and metabolic changes in normal-appearing white matter. *Magn Reson Med,* Vol. 33, No. 6, (June 1995), pp.

(2004). Magnetic resonance spectroscopy of normal appearing white matter in early relapsing-remitting multiple sclerosis: correlations between disability and spectroscopy. *BMC Neurol*, Vol. 10, No. 6, (June 2004), pp. 4-8, ISSN 1471-2377. Sarchielli P, Presciutti O, Tarducci R, Gobbi G, Alberti A, Pelliccioli GP, Orlacchio A, Gallai

V. (1998). 1H-MRS in patients with multiple sclerosis undergoing treatment with interferon beta-1a: results of a preliminary study. *J Neurol Neurosurg Psychiatry,* Vol.

Vicinanza F, Gallai V. (1999). Absolute quantification of brain metabolites by proton magnetic resonance spectroscopy in normal-appearing white matter of multiple sclerosis patients. *Brain,* Vol 122, No. 3, (March 1999), pp. 513-521, ISSN

Thompson AJ. (2005). Metabolite changes in normal-appearing gray and white matter are linked with disability in early primary progressive multiple sclerosis.

Bormans G, Carton H. (1997). Positron emission tomography, magnetic resonance imaging and proton NMR spectroscopy of white matter in multiple sclerosis. *Mult* 

Serial 1H-MRS in relapsing-remitting multiple sclerosis: effects of interferon-beta therapy on absolute metabolite concentrations. MAGMA, Vol. 14, No. 3, (June

Sibley WA, Tourtellotte WW, Willmon TL. (1965). Problems of experimental trials


Matthews PM, Pioro E, Narayanan S, De Stefano N, Fu L, Francis G, Antel J, Wolfson C,

McDonald WI, Sears TA. (1970). The effects of experimental demyelination on conduction in

McDonald WI, Compston A, Edan G, Goodkin D, Hartung HP, Lublin FD, McFarland HF,

Miller DH, Grossman RI, Reingold SC, McFarland HF. (1998). The role of magnetic

Miller DH, Weinshenker BG, Filippi M, Banwell BL, Cohen JA, Freedman MS, Galetta SL,

Narayanan S, Fu L, Pioro E, De Stefano N, Collins DL, Francis GS, Antel JP, Matthews PM,

Pascual AM, Martínez-Bisbal MC, Boscá I, Valero C, Coret F, Martínez-Granados B, Marti-

Paty DW, Li DK. (1993). Interferon beta-1b is effective in relapsing-remitting multiple

Peters AR, Geelen JA, den Boer JA, Prevo RL, Minderhoud JM, 's Gravenmade EJ. (1995). A

imaging. *Mult Scler,* Vol. 1, No. 1, (April 1995), pp. 25-31, ISSN 1352-4585. Polman CH, Reingold SC, Edan G, Filippi M, Hartung HP, Kappos L, Lublin FD, Metz LM,

Polman CH, Reingold SC, Banwell B, Clanet M, Cohen JA, Filippi M, Fujihara K, Havrdova

119, No. 3, (June 1999), pp. 715-722, ISSN 0006-8950.

Vol. 121, No. 1, (January 1998), pp. 3-24, ISSN 0006-8950.

No. 9, (November 2008), pp. 1157-1174, ISSN 1352-4585.

(March 1997), pp. 385-391, ISSN 0364-5134.

No. 1, (July 2007), pp. 63-67, ISSN 0028-3878.

0006-8950.

0028-3878.

ISSN 0364-5134.

127, ISSN 0364-5134.

Arnold DL. (1996). Assessment of lesion pathology in multiple sclerosis using quantitative MRI morphometry and magnetic resonance spectroscopy. *Brain,* Vol.

the central nervous system. *Brain,* Vol. 93, No. 3, (March 1970), pp. 583-598. ISSN

Paty DW, Polman CH, Reingold SC, Sandberg-Wollheim M, Sibley W, Thompson A, van den Noort S, Weinshenker BY, Wolinsky JS. (2001). Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. *Ann Neurol,* Vol. 50, No. 1, (July 2001), pp121-

resonance techniques in understanding and managing multiple sclerosis. *Brain,*

Hutchinson M, Johnson RT, Kappos L, Kira J, Lublin FD, McFarland HF, Montalban X, Panitch H, Richert JR, Reingold SC, Polman CH. (2008). Differential diagnosis of suspected multiple sclerosis: a consensus approach. *Mult Scler,* Vol. 14,

Arnold DL. (1997). Imaging of axonal damage in multiple sclerosis: spatial distribution of magnetic resonance imaging lesions. *Ann Neurol,* Vol. 41, No. 3,

Bonmati L, Mir A, Celda B, Casanova B. (2007). Axonal loss is progressive and partly dissociated from lesion load in early multiple sclerosis. *Neurology,* Vol. 69,

sclerosis. II. MRI analysis results of a multicenter, randomized, double-blind, placebo-controlled trial. UBC MS/MRI Study Group and the IFNB Multiple Sclerosis Study Group. *Neurology,* Vol. 43, No. 4, (April 1993), pp. 662-667, ISSN

study of multiple sclerosis patients with magnetic resonance spectroscopic

McFarland HF, O'Connor PW, Sandberg-Wollheim M, Thompson AJ, Weinshenker BG, Wolinsky JS. (2005). Diagnostic criteria for multiple sclerosis: 2005 revisions to the "McDonald Criteria". *Ann Neurol,* Vol. 58, No. 6, (December 2005), pp. 840-846,

E, Hutchinson M, Kappos L, Lublin FD, Montalban X, O'Connor P, Sandberg-Wollheim M, Thompson AJ, Waubant E, Weinshenker B, Wolinsky JS. (2011) .Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria. *Ann Neurol*, Vol. 69, No. 2, (February 2011), pp. 292-302, ISSN 0364-5134.


MR Spectroscopy in Multiple Sclerosis - A New Piece of the Puzzle or Just a New Puzzle 71

Tedeschi G, Bonavita S, McFarland HF, Richert N, Duyn JH, Frank JA. (2001). Proton MR

Tiberio M, Chard DT, Altmann DR, Davies G, Griffin CM, McLean MA, Rashid W, Sastre-

Tourbah A, Stievenart JL, Iba-Zizen MT, Zannoli G, Lyon-Caen O, Cabanis EA. (1996). In

Tourbah A, Stievenart JL, Gout O, Fontaine B, Liblau R, Lubetzki C, Cabanis EA, Lyon-Caen

van Walderveen MA, Barkhof F, Pouwels PJ, van Schijndel RA, Polman CH, Castelijns JA.

Vrenken H, Barkhof F, Uitdehaag BM, Castelijns JA, Polman CH, Pouwels PJ. (2005). MR

Wattjes MP, Harzheim M, Lutterbey GG, Bogdanow M, Schmidt S, Schild HH, Träber F.

Wattjes MP, Harzheim M, Lutterbey GG, Bogdanow M, Schild HH, Träber F. (2008). High

Waubant E. (2006). Biomarkers indicative of blood-brain barrier disruption in multiple

Weinshenker BG, Bass B, Rice GP, Noseworthy J, Carriere W, Baskerville J, Ebers GC.

Weinshenker BG, Bass B, Rice GP, Noseworthy J, Carriere W, Baskerville J, Ebers GC.

(January 2000), pp. 148-155, ISSN 1352-4585.

2002), pp. 37-42, Jan;44(1):37-42, ISSN 0028-3940.

(February 2006), pp. 224-230, ISSN 1351-5101.

(September 1999), pp. 1091-1097, ISSN 0028-3878.

Vol. 46, No. 1, (July 1999), pp. 79-87, ISSN 0364-5134.

49-55, ISSN 0150-9861.

pp. 256-266, ISSN 1522-2594.

3940.

ISSN 0340-5354.

ISSN 0006-8950.

pp. 1419-1428, ISSN 0006-8950.

lesions in primary progressive and relapsing-remitting MS. *Mult Scler,* Vol. 6, No. 3,

spectroscopic imaging in multiple sclerosis. *Neuroradiology,* Vol. 44, No. 1, (January

Garriga J, Thompson AJ, Miller DH. (2006). Metabolite changes in early relapsingremitting multiple sclerosis. A two year follow-up study. *J Neurol,* Vol. 253, No. 2,

vivo localized NMR proton spectroscopy of normal appearing white matter in patients with multiple sclerosis. *J Neuroradiol,* Vol. 23, No. 2, (September 1996), pp.

O. (1999). Localized proton magnetic resonance spectroscopy in relapsing remitting versus secondary progressive multiple sclerosis. *Neurology,* Vol. 53, No. 5,

(1999). Neuronal damage in T1-hypointense multiple sclerosis lesions demonstrated in vivo using proton magnetic resonance spectroscopy. *Ann Neurol,*

spectroscopic evidence for glial increase but not for neuro-axonal damage in MS normal-appearing white matter. *Magn Reson Med,* Vol. 53, No. 2, (February 2005),

(2008). Prognostic value of high-field proton magnetic resonance spectroscopy in patients presenting with clinically isolated syndromes suggestive of multiple sclerosis. *Neuroradiology,* Vol. 50, No. 2, (February 2008), pp. 123-129, ISSN 0028-

field MR imaging and 1H-MR spectroscopy in clinically isolated syndromes suggestive of multiple sclerosis: correlation between metabolic alterations and diagnostic MR imaging criteria. *J Neurol,* Vol. 255, No. 1, (January 2008), pp. 56-63,

sclerosis. *Dis Markers,* Vol. 22, No. 4, (November 2006), pp. 235-244, ISSN 0278-0240.

(1989a). The natural history of multiple sclerosis: a geographically based study. I. Clinical course and disability. *Brain,* Vol. 112, No. 1, (February 1989), pp. 133-146,

(1989b). The natural history of multiple sclerosis: a geographically based study. 2. Predictive value of the early clinical course. *Brain*, Vol. 112, No. 6, (December 1989),

of therapy in multiple sclerosis: report by the panel on the evaluation of experimental trials of therapy in multiple sclerosis. *Ann N Y Acad Sci,* Vol. 31., no. 122, (March 1965), pp. 552-556, ISSN 0077-8923.


Sellner J, Schirmer L, Hemmer B, Mühlau M. (2010). The radiologically isolated syndrome:

Serbecic N, Aboul-Enein F, Beutelspacher SC, Graf M, Kircher K, Geitzenauer W, Brannath

Serbecic N, Aboul-Enein F, Beutelspacher SC, Vass C, Kristoferitsch W, Lassmann H, Reitner

Siger-Zajdel M, Selmaj K. (2006). Proton magnetic resonance spectroscopy of normal

*Neurol Sci,* Vol. 144, No. 1-2, (December 1996), pp. 182-190, ISSN 1300-1817. Siva A, Saip S, Altintas A, Jacob A, Keegan BM, Kantarci OH. (2009) Multiple sclerosis risk

Smith KJ, Lassmann H. (2002). The role of nitric oxide in multiple sclerosis. *Lancet Neurol,* 

Smith KJ. (2007). Sodium channels and multiple sclerosis: roles in symptom production,

Sriram S, Steiner I. (2005). Experimental allergic encephalomyelitis: A misleading model of

Srinivasan R, Sailasuta N, Hurd R, Nelson S, Pelletier D. (2005). Evidence of elevated

Staffen W, Zauner H, Mair A, Kutzelnigg A, Kapeller P, Stangl H, Raffer E, Niederhofer H,

Suhy J, Rooney WD, Goodkin DE, Capizzano AA, Soher BJ, Maudsley EE, Waubant E,

*Brain,* Vol. 128, No. 5, (May 2005), pp. 1016-1025, ISSN 0006-8950.

(Summer 2005), pp. 357-363, ISN 0895-0172.

Vol. 1, No. 4, (August 2002), pp. 232-241, ISSN 1474-4422.

*Eur J Neurol,* Vol. 13, No. 3, (March 2006), pp. 296-298, ISSN 1351-5101. Simone IL, Federico F, Trojano M, Tortorella C, Liguori M, Giannini P, Picciola E, Natile G,

years. *PLoS One*. Vol. 6, No. 5, (May 2011), e19843, ISSN 1932-6203.

122, (March 1965), pp. 552-556, ISSN 0077-8923.

Vol.5, No. 11, (November 2010), e19843, ISSN 1932-6203.

2010), pp. 1602-1611, ISSN 0340-5354.

489-495, ISSN 0938-7994.

1015-6305.

0364-5134.

of therapy in multiple sclerosis: report by the panel on the evaluation of experimental trials of therapy in multiple sclerosis. *Ann N Y Acad Sci,* Vol. 31., no.

take action when the unexpected is uncovered? *J Neurol,* Vol. 257, No. 10, (October

W, Lang P, Kristoferitsch W, Lassmann H, Reitner A, Schmidt-Erfurth U. (2010). Heterogeneous pattern of retinal nerve fiber layer in multiple sclerosis. High resolution optical coherence tomography: potential and limitations. *PLoS One*.

A, Schmidt-Erfurth U. (2011). High resolution spectral domain optical coherence tomography (SD-OCT) in multiple sclerosis: the first follow up study over two

appearing white matter in asymptomatic relatives of multiple sclerosis patients.

Livrea P. (1996). High resolution proton MR spectroscopy of cerebrospinal fluid in MS patients. Comparison with biochemical changes in demyelinating plaques. *J* 

in radiologically uncovered asymptomatic possible inflammatory-demyelinating disease. *Mult Scler,* Vol. 15, No. 8, (August 2009), pp. 918-927, ISSN 1352-4585. Sijens PE, Mostert JP, Oudkerk M, De Keyser J. (2006). (1)H MR spectroscopy of the brain in

multiple sclerosis subtypes with analysis of the metabolite concentrations in gray and white matter: initial findings. *Eur Radiol,* Vol. 16, No. 2, (February 2006), pp.

damage and therapy. *Brain Pathol,* Vol. 17, No. 2, (April 2007), pp. 230-242, ISSN

multiple sclerosis. *Ann Neurol,* Vol. 58, No. 6, (December 2005), pp. 939–945, ISSN

glutamate in multiple sclerosis using magnetic resonance spectroscopy at 3 T.

Ladurner G. (2005). Magnetic resonance spectroscopy of memory and frontal brain region in early multiple sclerosis. *J Neuropsychiatry Clin Neurosci,* Vol. 17, No. 3,

Andersson PB, Weiner MW. (2000). 1H MRSI comparison of white matter and

lesions in primary progressive and relapsing-remitting MS. *Mult Scler,* Vol. 6, No. 3, (January 2000), pp. 148-155, ISSN 1352-4585.


**4** 

Beata Tarnacka

*Poland* 

**Wilson's Disease in Brain Magnetic** 

Wilson's disease (WD) is an inherited disorder of copper metabolism first defined by Dr Samuel Alexander Kinnier Wilson in 1912 (Wilson, 1912). WD is caused by mutations to the gene coding for ATPase copper transporting beta polypeptide (ATP7B), which is located on 13 chromosome 13 and expressed in the liver (Bull et al.,1993; Loudianos et al.,1999; Yamaguchi et al.,1993; Frydman et al., 1985). The disease is inherited in an autosomal

WD is present worldwide in most populations, and particularly in those in which consanguineous marriage is common. The disease frequency is estimated to be between 1 in 5,000 and 1 in 30,000, and the heterozygous carriers (Hzc) frequency is approximately 1 in 90

In the ATP7B gene of WD a variety of mutations were defected. These defects include insertion, deletion, splice site and point mutations. In most ethnic groups, one or a small number of these ATP7B gene mutations are predominant, in addition to many other rare mutations. In Europeans and North Americans, two ATP7B point mutations: His1069Gln and Gly1267Arg account for 38% of mutations described in WD (Thomas et al., 1995). There is still no favorable correlation between phenotype and genotype, but frame shift deletions and nonsense mutations that cause a truncation of the translated protein product usually result in a severe form of the disease because of loss of the functional protein. The knowledge of the prevalence of mutations is helpful in achieving rapid mutational screening. Mutations in ATP7B cause a reduction in the conversion of apoceruloplasmin into

A failure to excrete copper into the biliary canaliculi leads to its toxic accumulation in the hepatocytes (Schilsky&Tavill, 2003). The copper excess may damage hepatic mitochondria and oxidative damage to hepatic cells and cause the spillage of copper into the blood, thereby overloading other organs such as the brain, kidney and red blood cells, initiating toxic damage (Schilsky&Tavill, 2003). In the early stages of WD diffuse cytoplasmic copper accumulation can be seen only by special immunohistochemical stains for copper detecting. The early accumulation of copper is associated with hepatic steatosis. The ultrastructural abnormalities range from enlargement and separation of the mitochondrial inner and outer membranes with widening of the intercristal spaces, to increases in the density and

ceruloplasmin, which, is therefore present at low levels in WD patients.

**1. Introduction**

recessive manner.

(Figus et al., 1995).

 **Resonance Spectroscopy** 

*Mazovian Center of Rehabilitation, Konstancin Jeziorna* 

Young IR, Hall AS, Pallis CA, Legg NJ, Bydder GM, Steiner RE. (1981). Nuclear magnetic resonance imaging of the brain in multiple sclerosis. *Lancet.* Vol. 318, No. 8255 (November 1981), pp. 1063-1066, ISSN 0140-6736.

## **Wilson's Disease in Brain Magnetic Resonance Spectroscopy**

Beata Tarnacka

*Mazovian Center of Rehabilitation, Konstancin Jeziorna Poland* 

#### **1. Introduction**

72 Magnetic Resonance Spectroscopy

Young IR, Hall AS, Pallis CA, Legg NJ, Bydder GM, Steiner RE. (1981). Nuclear magnetic

(November 1981), pp. 1063-1066, ISSN 0140-6736.

resonance imaging of the brain in multiple sclerosis. *Lancet.* Vol. 318, No. 8255

Wilson's disease (WD) is an inherited disorder of copper metabolism first defined by Dr Samuel Alexander Kinnier Wilson in 1912 (Wilson, 1912). WD is caused by mutations to the gene coding for ATPase copper transporting beta polypeptide (ATP7B), which is located on 13 chromosome 13 and expressed in the liver (Bull et al.,1993; Loudianos et al.,1999; Yamaguchi et al.,1993; Frydman et al., 1985). The disease is inherited in an autosomal recessive manner.

WD is present worldwide in most populations, and particularly in those in which consanguineous marriage is common. The disease frequency is estimated to be between 1 in 5,000 and 1 in 30,000, and the heterozygous carriers (Hzc) frequency is approximately 1 in 90 (Figus et al., 1995).

In the ATP7B gene of WD a variety of mutations were defected. These defects include insertion, deletion, splice site and point mutations. In most ethnic groups, one or a small number of these ATP7B gene mutations are predominant, in addition to many other rare mutations. In Europeans and North Americans, two ATP7B point mutations: His1069Gln and Gly1267Arg account for 38% of mutations described in WD (Thomas et al., 1995). There is still no favorable correlation between phenotype and genotype, but frame shift deletions and nonsense mutations that cause a truncation of the translated protein product usually result in a severe form of the disease because of loss of the functional protein. The knowledge of the prevalence of mutations is helpful in achieving rapid mutational screening. Mutations in ATP7B cause a reduction in the conversion of apoceruloplasmin into ceruloplasmin, which, is therefore present at low levels in WD patients.

A failure to excrete copper into the biliary canaliculi leads to its toxic accumulation in the hepatocytes (Schilsky&Tavill, 2003). The copper excess may damage hepatic mitochondria and oxidative damage to hepatic cells and cause the spillage of copper into the blood, thereby overloading other organs such as the brain, kidney and red blood cells, initiating toxic damage (Schilsky&Tavill, 2003). In the early stages of WD diffuse cytoplasmic copper accumulation can be seen only by special immunohistochemical stains for copper detecting. The early accumulation of copper is associated with hepatic steatosis. The ultrastructural abnormalities range from enlargement and separation of the mitochondrial inner and outer membranes with widening of the intercristal spaces, to increases in the density and

Wilson's Disease in Brain Magnetic Resonance Spectroscopy 75

cerebellum and cerebral cortex can be seen (Hoogenraad, 1997). MRI is much more sensitive method for revealing abnormalities in WD (Hoogenraad, 1997;. Sinha et al., 2006). On T1 weighted images, generalized brain atrophy is seen in about 75% of cases, and hypointensities in the basal ganglia. On T2-weighted images, hyperintensity in the basal ganglia, white matter, thalamus or brainstem can be seen (Figure 1). These abnormalities are caused by neuronal loss, gliosis, degeneration of fibers, and vacuolization associated with increased water content in the brain. Signal abnormalities vary according to the stage of the disease, and can be reversible with therapy in the early stages. In some cases, T2-weighted images show hypointensity in the basal ganglia (globus pallidus) region may be as a result

In some WD patients with WD MRI shows a typical pallidal hyperintensity on T1-weighted images (Cordoba et al., 2003) (Figure 3). This abnormality appears to be secondary to the accumulation of manganese in basal ganglia because of portal-systemic shunting (Cordoba

Fig. 1. Axial T2 - weighed MRI image of a 31- years – old men with WD, MRI was done before treatment. The picture shows bilateral hyperintensities in putamen, caudatum and

thalami and degree of diffuse brain atrophy.

of deposition of iron in exchange for copper after chelation (Figure 2).

et al., 2003).

granularity of the matrix. During the disease progression periportal inflammation mononuclear cellular infiltration lobular necrosis and bridging fibrosis occur. The areas of the brain main affected in WD are the basal ganglia (lenticular nuclei), which macroscopically appear brown in color because of copper deposition (Scheinberg& Sternlieb, 1984). In the early stages of the WD, proliferation of large protoplasmic astrocytes such as Opalski cells and Alzheimer cells occurs. During the disease progression the degeneration occurs leading to necrosis, gliosis and cystic changes. The degeneration can be seen in the brainstem, thalamus, cerebellum and cerebral cortex. During the WD progression, copper deposits can lead to vacuolar degeneration in proximal renal tubular cells, appearance of the 'Kayser–Fleischer' (KF) ring in Descemet's membrane. Copper from the hepatic cells degradation can be released into the circulation and can damage red blood cells, thereby inducing hemolysis (Schilsky&Tavill, 2003).

The majority of patients with WD present or with hepatic or neuropsychiatric symptoms, and with either clinically asymptomatic or symptomatic liver involvement. The remaining patients, may present with symptoms attributable to the involvement of other organs, as acute non-immunological hemolytic anemia, osteoarthritis, arrhythmias, rheumatic-feverlike manifestation, renal function abnormalities, primary or secondary amenorrhea, repeated and unexplained spontaneous abortions. Patients with hepatic WD (hWD) usually present in late childhood or adolescence, and exhibit features of acute hepatitis, fulminant hepatic failure, or progressive chronic liver disease in the form of either chronic active hepatitis or cirrhosis of the macronodular type (Schilsky&Tavill, 2003;. Hoogenraad, 1997). The mean age of onset of neurological WD (nWD) is the second to third decade (Hoogenraad, 1997). Most of patients present with extrapyramidal, cerebellar and cerebralrelated symptoms. Tremor, gait and speech disturbances are the most common initial presentation symptoms. Patients may present also present with dystonia. About 30% of patients experience psychiatric disturbances (Hoogenraad, 1997). These disturbances can manifest as changes in school-related or work-related performance, attention deficit hyperactivity disorder, impulsivity, paranoid psychosis, obsessive behavior, depression, suicidal tendencies or bizarre behavior, and can occur early or late in the disease course.

The WD diagnosis and monitoring is based on history, physical, biochemical, liver biopsy and genetic mutations and imaging assessment. The KF ring is an important marker in nWD examination. The WD is biochemically characterized by low ceruloplasmin and total serum copper levels, increased 24-hour urinary copper excretion, and abnormally high hepatic copper content in the liver biopsy particularly in children performed. Another noninvasive method for assessing copper metabolism is incorporation of radioactive copper into the hepatocytes used with favorable results in Poland specially in presymptomatic gene mutation carriers, siblings of WD patients diagnosis (Członkowska et al. , 1973). The mutation screening to identify defects in the ATP7B gene can confirmation of WD diagnosis, but will not necessarily detect all disease producing mutations.

#### **2. Neuroimaging**

#### **2.1 Magnetic resonance imaging**

Neuroimaging plays an important role in the diagnosis and monitoring of WD patients. In brain CT scans hypodensities and atrophy of the bilateral basal ganglia, brainstem,

granularity of the matrix. During the disease progression periportal inflammation mononuclear cellular infiltration lobular necrosis and bridging fibrosis occur. The areas of the brain main affected in WD are the basal ganglia (lenticular nuclei), which macroscopically appear brown in color because of copper deposition (Scheinberg& Sternlieb, 1984). In the early stages of the WD, proliferation of large protoplasmic astrocytes such as Opalski cells and Alzheimer cells occurs. During the disease progression the degeneration occurs leading to necrosis, gliosis and cystic changes. The degeneration can be seen in the brainstem, thalamus, cerebellum and cerebral cortex. During the WD progression, copper deposits can lead to vacuolar degeneration in proximal renal tubular cells, appearance of the 'Kayser–Fleischer' (KF) ring in Descemet's membrane. Copper from the hepatic cells degradation can be released into the circulation and can damage red blood

The majority of patients with WD present or with hepatic or neuropsychiatric symptoms, and with either clinically asymptomatic or symptomatic liver involvement. The remaining patients, may present with symptoms attributable to the involvement of other organs, as acute non-immunological hemolytic anemia, osteoarthritis, arrhythmias, rheumatic-feverlike manifestation, renal function abnormalities, primary or secondary amenorrhea, repeated and unexplained spontaneous abortions. Patients with hepatic WD (hWD) usually present in late childhood or adolescence, and exhibit features of acute hepatitis, fulminant hepatic failure, or progressive chronic liver disease in the form of either chronic active hepatitis or cirrhosis of the macronodular type (Schilsky&Tavill, 2003;. Hoogenraad, 1997). The mean age of onset of neurological WD (nWD) is the second to third decade (Hoogenraad, 1997). Most of patients present with extrapyramidal, cerebellar and cerebralrelated symptoms. Tremor, gait and speech disturbances are the most common initial presentation symptoms. Patients may present also present with dystonia. About 30% of patients experience psychiatric disturbances (Hoogenraad, 1997). These disturbances can manifest as changes in school-related or work-related performance, attention deficit hyperactivity disorder, impulsivity, paranoid psychosis, obsessive behavior, depression, suicidal tendencies or bizarre behavior, and can occur early or late in the disease course.

The WD diagnosis and monitoring is based on history, physical, biochemical, liver biopsy and genetic mutations and imaging assessment. The KF ring is an important marker in nWD examination. The WD is biochemically characterized by low ceruloplasmin and total serum copper levels, increased 24-hour urinary copper excretion, and abnormally high hepatic copper content in the liver biopsy particularly in children performed. Another noninvasive method for assessing copper metabolism is incorporation of radioactive copper into the hepatocytes used with favorable results in Poland specially in presymptomatic gene mutation carriers, siblings of WD patients diagnosis (Członkowska et al. , 1973). The mutation screening to identify defects in the ATP7B gene can confirmation of WD diagnosis,

Neuroimaging plays an important role in the diagnosis and monitoring of WD patients. In brain CT scans hypodensities and atrophy of the bilateral basal ganglia, brainstem,

cells, thereby inducing hemolysis (Schilsky&Tavill, 2003).

but will not necessarily detect all disease producing mutations.

**2. Neuroimaging** 

**2.1 Magnetic resonance imaging** 

cerebellum and cerebral cortex can be seen (Hoogenraad, 1997). MRI is much more sensitive method for revealing abnormalities in WD (Hoogenraad, 1997;. Sinha et al., 2006). On T1 weighted images, generalized brain atrophy is seen in about 75% of cases, and hypointensities in the basal ganglia. On T2-weighted images, hyperintensity in the basal ganglia, white matter, thalamus or brainstem can be seen (Figure 1). These abnormalities are caused by neuronal loss, gliosis, degeneration of fibers, and vacuolization associated with increased water content in the brain. Signal abnormalities vary according to the stage of the disease, and can be reversible with therapy in the early stages. In some cases, T2-weighted images show hypointensity in the basal ganglia (globus pallidus) region may be as a result of deposition of iron in exchange for copper after chelation (Figure 2).

In some WD patients with WD MRI shows a typical pallidal hyperintensity on T1-weighted images (Cordoba et al., 2003) (Figure 3). This abnormality appears to be secondary to the accumulation of manganese in basal ganglia because of portal-systemic shunting (Cordoba et al., 2003).

Fig. 1. Axial T2 - weighed MRI image of a 31- years – old men with WD, MRI was done before treatment. The picture shows bilateral hyperintensities in putamen, caudatum and thalami and degree of diffuse brain atrophy.

Wilson's Disease in Brain Magnetic Resonance Spectroscopy 77

mediated neuronal dipeptide N-acetylaspartylglutamate biosynthesis; NAA provides a source of acetate for myelin lipid synthesis and is involved in energy metabolism in neuronal mitochondria-reflect improvement of neuronal energetic (Moffett et al.,2007).

Fig. 3. Axial T1 - weighed MRI image of a 45- years – old WD patient with hepatic cirrhosis

1H-MRS is a technique that could help distinguish between brain changes caused by HE and those related to copper toxicosis in WD. To elucidate the pathomechanism of the cerebral pathology of WD a study using 1H-MRS in 37 newly diagnosed WD patients in the globus pallidus and thalamus examined bilaterally was performed (Tarnacka et al, 2009a). The calculations were performed for: mI, Cho, creatine (Cr), NAA, lipid (Lip), Glx. In all WD patients a significantly decreased mI/Cr and NAA/Cr ratio level and an increased Lip/Cr ratio in the pallidum was observed. Analysis revealed a significantly increased Glx/Cr and Lip/Cr ratio in the thalamus. In the pallidum of nWD patients, Cho/Cr and Glx/Cr and Lip/Cr ratios were higher than in control subjects, and the NAA/Cr was significantly lower (Table 1). In hWD patients, the mI/Cr and Cho/Cr and NAA/Cr ratio levels were lower

with subclinical hepatic encephalopathy.

Fig. 2. Axial T2 - weighed MRI image of a 47- years – old women with WD treated for more than 10 years. The picture shows persistent hyperintensities in putamen, caudatum globus pallidus and degree of diffuse brain atrophy and hypointensity in the basal ganglia.

Proton Magnetic Resonance spectroscopy (1H-MRS) is a practical research tool for elucidating the pathophysiology underlying certain diseases (Rudkin & Arnold, 1999). In patients with hepatic encephalopathy (HE), 1H-MRS has been used to detect metabolic abnormalities in the brain with very high sensitivity (Ross et al., 1999). The biochemical alterations that were detected in HE included an increase in cerebral glutamine compounds (Glx) and a decrease in myoinositol (mI) and choline (Cho) metabolites (Cordoba et al., 2002). Other metabolites detectable in vivo by 1H-MRS are N-acetylaspartate (NAA), considered to be a marker of neuronal health, and creatine (Cr), often used as an internal standard against which the resonance intensities of other metabolites are normalized. The role of NAA in nervous system can be related to: action as an organic osmolyte that counters the "anion deficit" in neurons, or a cotransport substrate for a proposed "molecular water pump" that removes metabolic water from neurons; NAA is a precursor for the enzyme-

Fig. 2. Axial T2 - weighed MRI image of a 47- years – old women with WD treated for more than 10 years. The picture shows persistent hyperintensities in putamen, caudatum globus pallidus and degree of diffuse brain atrophy and hypointensity in the basal ganglia.

Proton Magnetic Resonance spectroscopy (1H-MRS) is a practical research tool for elucidating the pathophysiology underlying certain diseases (Rudkin & Arnold, 1999). In patients with hepatic encephalopathy (HE), 1H-MRS has been used to detect metabolic abnormalities in the brain with very high sensitivity (Ross et al., 1999). The biochemical alterations that were detected in HE included an increase in cerebral glutamine compounds (Glx) and a decrease in myoinositol (mI) and choline (Cho) metabolites (Cordoba et al., 2002). Other metabolites detectable in vivo by 1H-MRS are N-acetylaspartate (NAA), considered to be a marker of neuronal health, and creatine (Cr), often used as an internal standard against which the resonance intensities of other metabolites are normalized. The role of NAA in nervous system can be related to: action as an organic osmolyte that counters the "anion deficit" in neurons, or a cotransport substrate for a proposed "molecular water pump" that removes metabolic water from neurons; NAA is a precursor for the enzymemediated neuronal dipeptide N-acetylaspartylglutamate biosynthesis; NAA provides a source of acetate for myelin lipid synthesis and is involved in energy metabolism in neuronal mitochondria-reflect improvement of neuronal energetic (Moffett et al.,2007).

Fig. 3. Axial T1 - weighed MRI image of a 45- years – old WD patient with hepatic cirrhosis with subclinical hepatic encephalopathy.

1H-MRS is a technique that could help distinguish between brain changes caused by HE and those related to copper toxicosis in WD. To elucidate the pathomechanism of the cerebral pathology of WD a study using 1H-MRS in 37 newly diagnosed WD patients in the globus pallidus and thalamus examined bilaterally was performed (Tarnacka et al, 2009a). The calculations were performed for: mI, Cho, creatine (Cr), NAA, lipid (Lip), Glx. In all WD patients a significantly decreased mI/Cr and NAA/Cr ratio level and an increased Lip/Cr ratio in the pallidum was observed. Analysis revealed a significantly increased Glx/Cr and Lip/Cr ratio in the thalamus. In the pallidum of nWD patients, Cho/Cr and Glx/Cr and Lip/Cr ratios were higher than in control subjects, and the NAA/Cr was significantly lower (Table 1). In hWD patients, the mI/Cr and Cho/Cr and NAA/Cr ratio levels were lower

Wilson's Disease in Brain Magnetic Resonance Spectroscopy 79

Fig. 4. b. 1H-MRS images of patient with neurological WD presentation compared with control subject. Proton spectrum shows expression of Glx, Lip; NAA was significantly reduced. The axial image shows the inset from where the spectrum was recorded.

Fig. 4. c. 1H-MRS images of presymptomatic WD patients compared with control subject. Proton spectrum shows expression of Lip; Cr was significantly reduced. The axial image

shows the inset from where the spectrum was recorded.

than in controls (Table 1). The Cho/Cr and Lip/Cr ratios were higher in the thalami of nWD patients, and Lip/Cr ratios were higher than controls' in patients only with hWD. On the Figure 4 the examples of 1H-MRS spectra in nWD and hWD and presymptomatic patients compared with control subjects are provided.


1 H-MRS study pallidum


1 H-MRS.thalamus

\*statistically higher than in control group

¶ statistically lower than in control group

¥ statistically lower compared with thalami

Table 1. The mean relative metabolic ratios in (left and right) pallidum and (right and left) thalamus in neurological (nWD) and hepatic (hWD) and presymptomatic subgroups.

Fig. 4. a. 1H-MRS images of patient with hepatic WD presentation compared with control subject. Proton spectrum shows expression of Glx; Cho and NAA were significantly reduced. The axial image shows the inset from where the spectrum was recorded.

than in controls (Table 1). The Cho/Cr and Lip/Cr ratios were higher in the thalami of nWD patients, and Lip/Cr ratios were higher than controls' in patients only with hWD. On the Figure 4 the examples of 1H-MRS spectra in nWD and hWD and presymptomatic patients

Patients: mI/Cr Cho/Cr Glx/Cr NAA/Cr Lip/Cr **nWD** 0,66 0,31 0,89 0,30\* 1,26 0,51\* 1,43 0,42 ¥¶ 0,94 0,46\* **hWD** 0,50 0,27¶¥ 0,68 0,24¶ 1,38 0,39\* 1,36 0,39¶¥ 0,81 0,31\* presymptomatic 0,35 0,10 0,66 0,09 1,37 0,24 1,59 0,14 0,65 0,08 Control group 0,72 0,15 0,78 0,15 1,11 0,25 1,61 0,26 0,47 0,20

Patients: mI/Cr Cho/Cr Glx/Cr NAA/Cr Lip/Cr **nWD** 0,74 0,30 1.01 0,27\* 1,19 0,32 1,91 0,56 0,74 0,32\* **hWD** 0,85 0,27 1.07 0,24 1.21 0,39 1,73 0,39 0.78 0,31\* presymptomatic 0,760,21 0,910,23 1,130,20 1,690,32 0,500,22 Control group 0,82 0.20 0,83 0,14 1,11 0,21 1,68 0,24 0,50 0,20

Table 1. The mean relative metabolic ratios in (left and right) pallidum and (right and left) thalamus in neurological (nWD) and hepatic (hWD) and presymptomatic subgroups.

Fig. 4. a. 1H-MRS images of patient with hepatic WD presentation compared with control subject. Proton spectrum shows expression of Glx; Cho and NAA were significantly reduced. The axial image shows the inset from where the spectrum was recorded.

compared with control subjects are provided.

1 H-MRS study pallidum

1 H-MRS.thalamus

\*statistically higher than in control group ¶ statistically lower than in control group ¥ statistically lower compared with thalami

Fig. 4. b. 1H-MRS images of patient with neurological WD presentation compared with control subject. Proton spectrum shows expression of Glx, Lip; NAA was significantly reduced. The axial image shows the inset from where the spectrum was recorded.

Fig. 4. c. 1H-MRS images of presymptomatic WD patients compared with control subject. Proton spectrum shows expression of Lip; Cr was significantly reduced. The axial image shows the inset from where the spectrum was recorded.

Wilson's Disease in Brain Magnetic Resonance Spectroscopy 81

bradykinesia, dysarthria, hypomimia, or rigidity. Some authors have speculated that the neurological symptoms of WD can be caused by concomitant liver disease (Victor, 1999).

These MRS findings can prove that in newly diagnosed nWD patients in the brain a portosystemic shunting changes, but also neurodegenerative pattern associated with Cho, Lip and NAA/Cr depletion nWD can coexist. Verma compared brain metabolite alternations in patients with acute liver failure, acute-on-chronic liver failure and chronic liver disease (Verma et al, 2008). He found that NAA/Cr ratio was significantly decreased in acute-onchronic liver failure and chronic liver disease. This fact can prove that in chronic liver disease an mitochondrial dysfunction can be detected due to neurotoxic effect. It is also worthy to mentioned that in patients with hepatic and neurological impairment in WD a significant negative correlation in the pallidum between the clinical status and the NAA/Cr

From the 37 patients newly diagnosed, we followed 17WD cases for more than one year period (Tarnacka et al, 2008). In the 1H-MRS done during the follow-up of all WD patients, significantly lower levels of mI/Cr and higher levels of Lip/Cr compared to the control group were persistently noted (Figure 5a). In patients with hepatic signs with improvement, a statistically significant increase of mI/Cr and Glx/Cr in follow- up 1H-MRS was observed (about one year post-treatment, Figure 5b). In patients with neurological improvement after treatment in the follow-up 1H-MRS, a statistically significant increase of NAA/Cr was noted (Table 5c). During neurological deterioration in one case, a decrease of Glx/Cr and NAA/Cr was seen, in contrast to another neurologically impaired patient with liver failure exacerbation, where a decrease of

**WD First Study WD Second Study Control Group**

a

c c

b

mI/Cr and increase of Glx/Cr was observed (Tarnacka et al, 2009).

b. significantly higher than in controls (p<0.0001) and compared with first study p<0.005,

(a)

**mI/Cr Cho/Cr Glx/Cr NAA/Cr Lip/Cr**

was detected (Tarnacka et al, 2009).

a. significantly lower than in controls (p<0.005),

0,00 0,20 0,40 0,60 0,80 1,00 1,20 1,40 1,60 1,80 2,00

a a

c. significantly higher than in controls (p<0.005),

In those study the thalamic changes compared with the basal ganglia were more sensitive to ongoing degenerative changes and portal-systemic encephalopathy.

Spectroscopic changes in presymtomatic patients were performed only on 4 cases, because of it was very difficult to find patients, who did not have any clinical and biochemical (liver failure) abnormalities. In those patients significantly lower levels of mI/Cr and Cho/Cr and increase of Glx/Cr ratios in basal ganglia. In thalami no changes were seen. Those findings can suggest that in presymptomatic patients in MRS early encephalopatic changes with decrease of mI can be noticed.

The decreased Cho/Cr and mI/Cr in the pallidum of hWD patients could be indicative of minimal HE. Mioinositol is one of the chief compatible organic osmolytes responsible for equilibrating an increased intracellular tonicity (Cho, is believed to be also an osmolyte) (Danielsen & Ross, 1999). The mI/Cr and Cho/Cr reduction was also reported by Kraft in one *de novo* WD patient with hepatic disease (Kraft E et al.,1999). In newly diagnosed WD patients a reduction of NAA/Cr in the pallidum of hWD patients was noted (Tarnacka et al, 2009). This could indicate that in these patients, 1H-MRS detects a combination of early encephalopathic and neurodegenerative changes, if NAA can be considered a neuronal marker.

Newly diagnosed WD patients with neurological impairment showed an increased level of Cho/Cr and Lip/ Cr in the pallidum and thalamus; furthermore, Glx/Cr was increased and NAA/Cr was decreased in the pallidum. The Cho peak is considered a potential biomarker for the status of membrane phospholipid metabolism (Danielsen & Ross, 1999), so that an elevated Cho signal most likely reflects an increase in membrane turnover. In pathologies characterised by membrane breakdown, such as neurodegeneration, bound Cho moieties may be liberated into the free Cho pool (Danielsen & Ross, 1999). It is possible that the increase of Cho/Cr and Lip/Cr ratios in the pallidum and the thalami can reflect an increase in membrane turnover caused by free copper accumulation, or can be associated with gliosis because Cho is present in high concentrations in oligodendrocytes (Van der Hart et al., 2000). In nWD patients, the increased level of Glx/Cr in the pallidum, can be related to the shunting of ammonia to the brain and glutamine accumulation. The elevation of Glx, can be also related to neuronal energy impairment. Removal of glucose greatly accelerates glutamate transamination to aspartate in brain synaptosomes, suggesting that glutamate could be an energy source in brain under some circumstances (Erecinska et al, 1988). Aspartate transaminase is the enzyme that accomplishes the task of glutamate transamination. It consumes oxaloacetate and glutamate to produce aspartate and alphaketoglutarate, which can directly enter to the TCA cycle (Moffett et al, 2007). Most studies concerning the relationship between neurospectroscopic abnormalities and the neurological HE manifestation found an association between the presence of HE and Glx/Cr level (Kreis et al., 1992;. Cordoba et al., 2001). In nWD patients no other metabolite changes specific for HE such as mI/Cr and Cho/Cr reduction was detected, so it is conceivable that because of chronic liver failure, in nWD patients the compensatory mechanism to counteract intracellular hypertonicity can take place, with no mI depletion, but with an increase of Glx. This finding, could indicate, that in newly diagnosed nWD patients an HE could exist. Patients with liver cirrhosis do present with parkinsonian signs including: tremor,

In those study the thalamic changes compared with the basal ganglia were more sensitive to

Spectroscopic changes in presymtomatic patients were performed only on 4 cases, because of it was very difficult to find patients, who did not have any clinical and biochemical (liver failure) abnormalities. In those patients significantly lower levels of mI/Cr and Cho/Cr and increase of Glx/Cr ratios in basal ganglia. In thalami no changes were seen. Those findings can suggest that in presymptomatic patients in MRS early encephalopatic changes with

The decreased Cho/Cr and mI/Cr in the pallidum of hWD patients could be indicative of minimal HE. Mioinositol is one of the chief compatible organic osmolytes responsible for equilibrating an increased intracellular tonicity (Cho, is believed to be also an osmolyte) (Danielsen & Ross, 1999). The mI/Cr and Cho/Cr reduction was also reported by Kraft in one *de novo* WD patient with hepatic disease (Kraft E et al.,1999). In newly diagnosed WD patients a reduction of NAA/Cr in the pallidum of hWD patients was noted (Tarnacka et al, 2009). This could indicate that in these patients, 1H-MRS detects a combination of early encephalopathic and neurodegenerative changes, if NAA can be considered a neuronal

Newly diagnosed WD patients with neurological impairment showed an increased level of Cho/Cr and Lip/ Cr in the pallidum and thalamus; furthermore, Glx/Cr was increased and NAA/Cr was decreased in the pallidum. The Cho peak is considered a potential biomarker for the status of membrane phospholipid metabolism (Danielsen & Ross, 1999), so that an elevated Cho signal most likely reflects an increase in membrane turnover. In pathologies characterised by membrane breakdown, such as neurodegeneration, bound Cho moieties may be liberated into the free Cho pool (Danielsen & Ross, 1999). It is possible that the increase of Cho/Cr and Lip/Cr ratios in the pallidum and the thalami can reflect an increase in membrane turnover caused by free copper accumulation, or can be associated with gliosis because Cho is present in high concentrations in oligodendrocytes (Van der Hart et al., 2000). In nWD patients, the increased level of Glx/Cr in the pallidum, can be related to the shunting of ammonia to the brain and glutamine accumulation. The elevation of Glx, can be also related to neuronal energy impairment. Removal of glucose greatly accelerates glutamate transamination to aspartate in brain synaptosomes, suggesting that glutamate could be an energy source in brain under some circumstances (Erecinska et al, 1988). Aspartate transaminase is the enzyme that accomplishes the task of glutamate transamination. It consumes oxaloacetate and glutamate to produce aspartate and alphaketoglutarate, which can directly enter to the TCA cycle (Moffett et al, 2007). Most studies concerning the relationship between neurospectroscopic abnormalities and the neurological HE manifestation found an association between the presence of HE and Glx/Cr level (Kreis et al., 1992;. Cordoba et al., 2001). In nWD patients no other metabolite changes specific for HE such as mI/Cr and Cho/Cr reduction was detected, so it is conceivable that because of chronic liver failure, in nWD patients the compensatory mechanism to counteract intracellular hypertonicity can take place, with no mI depletion, but with an increase of Glx. This finding, could indicate, that in newly diagnosed nWD patients an HE could exist. Patients with liver cirrhosis do present with parkinsonian signs including: tremor,

ongoing degenerative changes and portal-systemic encephalopathy.

decrease of mI can be noticed.

marker.

bradykinesia, dysarthria, hypomimia, or rigidity. Some authors have speculated that the neurological symptoms of WD can be caused by concomitant liver disease (Victor, 1999).

These MRS findings can prove that in newly diagnosed nWD patients in the brain a portosystemic shunting changes, but also neurodegenerative pattern associated with Cho, Lip and NAA/Cr depletion nWD can coexist. Verma compared brain metabolite alternations in patients with acute liver failure, acute-on-chronic liver failure and chronic liver disease (Verma et al, 2008). He found that NAA/Cr ratio was significantly decreased in acute-onchronic liver failure and chronic liver disease. This fact can prove that in chronic liver disease an mitochondrial dysfunction can be detected due to neurotoxic effect. It is also worthy to mentioned that in patients with hepatic and neurological impairment in WD a significant negative correlation in the pallidum between the clinical status and the NAA/Cr was detected (Tarnacka et al, 2009).

From the 37 patients newly diagnosed, we followed 17WD cases for more than one year period (Tarnacka et al, 2008). In the 1H-MRS done during the follow-up of all WD patients, significantly lower levels of mI/Cr and higher levels of Lip/Cr compared to the control group were persistently noted (Figure 5a). In patients with hepatic signs with improvement, a statistically significant increase of mI/Cr and Glx/Cr in follow- up 1H-MRS was observed (about one year post-treatment, Figure 5b). In patients with neurological improvement after treatment in the follow-up 1H-MRS, a statistically significant increase of NAA/Cr was noted (Table 5c). During neurological deterioration in one case, a decrease of Glx/Cr and NAA/Cr was seen, in contrast to another neurologically impaired patient with liver failure exacerbation, where a decrease of mI/Cr and increase of Glx/Cr was observed (Tarnacka et al, 2009).

a. significantly lower than in controls (p<0.005),

b. significantly higher than in controls (p<0.0001) and compared with first study p<0.005,

c. significantly higher than in controls (p<0.005),

Wilson's Disease in Brain Magnetic Resonance Spectroscopy 83

The alternations of NAA/Cr ratio in neurologically impaired patients and mI/Cr and Glx/Cr in patients with liver failure could be a sensitive marker of the clinical recovery and

A decrease in relative NAA concentrations has been observed in pathological processes known to involve neuronal loss (Rudkin & Arnold, 1999). NAA is synthesized by neuronal mitochondria, which are sensitive to injury (Bates et al., 1996). A reversible decrease of NAA could well be caused by mitochondrial toxin 3-nitropropionate (Dauntry et al., 2000). Mitochondrial proteins seem to be the target of copper toxicity, because a decrease in the

mI/Cr Cho/Cr Glx/Cr NAA/Cr Lip/Cr

All 0.660.30 0.890.21a 1.350.20b 1.440.35 0.930.48c CS 0.830.14 0.750.16 1.090.31 1.610.22 0.450.18 hWD 0.630.33 0.790.38 1.330.47 1.470.38 0.990.57d CS 0.660.14 0.760.16 1.090.31 1.580.22 0.430.18 nWD 0.720.21 0.960.36e 1.360.44f 1.410.42 0.890.48g CS 0.950.41 0.750.12 1.090.23 1.630.28 0.460.22

(a)

mI/Cr Cho/Cr Glx/Cr NAA/Cr Lip/Cr

All 0.680.30 0.790.36 1.100.37 1.360.26a 0.940.31b CS 0.710.18 0.760.16 1.130.21 1.650.21 0.470.16 hWD 0.640.22 0.580.21c 1.310.18 1.450.50 0.870.50 d CS 0.780.28 0.930.24 1.200.27 1.560.31 0.500.22 nWD 0.750.34 1.030.30e 1.180.32 1.300.25f 0.860.29g CS 0.710.17 0.650.16 1.030.22 1.600.29 0.470.20

(b) Table 2. The mean metabolites ratios compared with control subjects (CS) in all WD patients, in hWD and nWD subgroups in: a) patients with improvement of clinical status, b) patients

deterioration in those WD patients.

a. significantly higher than in controls (p<0.0002) b. significantly higher than in controls (p<0.0005) c. significantly higher than in controls (p<0.00002) d. significantly higher than in controls (p<0.000000) e. significantly higher than in controls (p<0.00002) f. significantly higher than in controls (p<0.004) g. significantly higher than in controls (p<0.0003)

a. significantly lower than in controls (p<0.01), b. significantly higher than in controls (p<0.003) c. significantly lower than in controls (p<0.001) d. significantly higher than in controls (p<0.0003) e. significantly higher than in controls (p<0.0003) f. significantly lower than in controls (p<0.002) g. significantly higher than in controls (p<0.003)

with no improvement of clinical status.

a. significantly lower than in controls, p<0.005,

b. significantly higher compared with first study p<0.04,

c. significantly higher than in controls, p<0.00006 and compared with first study p<0.002,

d. significantly higher than in controls, p<0.005

b. significantly higher compare with first study, p<0.01 c. significantly higher than in controls, p<0.02

(c)

Fig. 5. The mean metabolic ratios in patients with improvement: a) all 15 WD patients b) hepatic subgroup (n=9), c) neurological subgroup (n=6), in the first and second 1H-MRS test compared to the age-matched control subjects. Statistical analysis in text.

a. significantly lower than in controls, p<0.005,

d. significantly higher than in controls, p<0.005

a. significantly lower than in controls, p<0.05,

c. significantly higher than in controls, p<0.02

b. significantly higher compare with first study, p<0.01

b. significantly higher compared with first study p<0.04,

c. significantly higher than in controls, p<0.00006 and compared with first study p<0.002,

(b)

(c) Fig. 5. The mean metabolic ratios in patients with improvement: a) all 15 WD patients b) hepatic subgroup (n=9), c) neurological subgroup (n=6), in the first and second 1H-MRS test

compared to the age-matched control subjects. Statistical analysis in text.

The alternations of NAA/Cr ratio in neurologically impaired patients and mI/Cr and Glx/Cr in patients with liver failure could be a sensitive marker of the clinical recovery and deterioration in those WD patients.

A decrease in relative NAA concentrations has been observed in pathological processes known to involve neuronal loss (Rudkin & Arnold, 1999). NAA is synthesized by neuronal mitochondria, which are sensitive to injury (Bates et al., 1996). A reversible decrease of NAA could well be caused by mitochondrial toxin 3-nitropropionate (Dauntry et al., 2000). Mitochondrial proteins seem to be the target of copper toxicity, because a decrease in the


a. significantly higher than in controls (p<0.0002) b. significantly higher than in controls (p<0.0005) c. significantly higher than in controls (p<0.00002) d. significantly higher than in controls (p<0.000000) e. significantly higher than in controls (p<0.00002) f. significantly higher than in controls (p<0.004)

g. significantly higher than in controls (p<0.0003)

(a)


a. significantly lower than in controls (p<0.01),

b. significantly higher than in controls (p<0.003)

c. significantly lower than in controls (p<0.001)

d. significantly higher than in controls (p<0.0003) e. significantly higher than in controls (p<0.0003)

f. significantly lower than in controls (p<0.002)

g. significantly higher than in controls (p<0.003)

(b)

Table 2. The mean metabolites ratios compared with control subjects (CS) in all WD patients, in hWD and nWD subgroups in: a) patients with improvement of clinical status, b) patients with no improvement of clinical status.

Wilson's Disease in Brain Magnetic Resonance Spectroscopy 85

changes in Hzc was performed (Tarnacka et.al, 2009b). We observed statistically significant higher ratios of Glx/Cr and Lip/Cr in 1 H-MRS in Hzc in the pallidum (Table 3). Glx is a glial marker; the increase of Glx/Cr could correlate with the glial proliferation as a result of the protective function of the astrocytes to copper and iron accumulation. The significantly higher level of Lip/Cr may reflect the liberation of lipids from membranes during their breakdown as a reaction to copper and iron overload. Our results may instead suggest that WD Hz carriers accumulate free copper or iron or both in basal ganglia and thalami. Hzc in our study were much older than WD cases are and in an animal model of WD copper and iron levels were found to be augmented during aging in striatum and substantia nigra (Kim et al., 2001). It is possible that during aging in Hz there is an increase of copper and iron accumulation because of ATP7b and ceruolplasmin impairment. There is accumulating evidence that ceruloplasmin, a copper protein with ferroxidase activity, plays an important role in iron metabolism and is also the principal plasma copper-binding protein (Harris et al., 1998). In our Hz the ceruloplasmin level was lower in only four cases, but ceruloplasmin levels in the blood are influenced by acute-phase reactions. The copper accumulation in the brain can also be caused by disruption in its excretion. In animal models, heterozygous mice have a reduced ability to excrete copper, which may indicate that half of the normal liver ATP7b copper transporter activity is insufficient to deal with a large amount of copper intake (Cheach et al., 2007). Like iron, copper is a heavy metal that can induce the Fenton reaction via production of free radicals and copper and iron accumulation increased with aging might

mI/Cr Cho/Cr Glx/Cr NAA/Cr Lip/Cr

Hz: 0.730,19 0.810,13 1.380,30\* 1.50,28 0.960,40\*\*

**subjects** 0.82 0,82 0.78 0,16 1.11 0,26 1.62 0,27 0.48 0,20

Hz 0.90 0,27 0. 95 0,23 1.44 0,40¥ 1.93 0,35 1.08 0,40¶

Control subjects 0.82 0,21 0.84 0,15 1.12 0,21 1.69 0,25 0.51 0,20

Table 3. The mean metabolites ratios in heterozygotes (Hz) and control subjects from left and right globus pallidus (Gp) and right and left thalamus (Th) with standard deviations.

The treatment of WD patients may take many forms, from drug therapy (chelating agents as d-penicillamine (d-p) or trientine, zinc therapy, tetrathiomolibdate) to liver transplantation

have deleterious of liver and brain via oxidative stress.

\*significantly higher than in controls, p<0.00006 \*\* significantly higher than in controls, p<0.00002 ¥ significantly higher than in controls, p<0.0006 ¶ significantly higher than in controls, p<0.000001

**2.1.2 MRS versus Wilson's disease treatment** 

1 H-MRS study

1 H-MRS study

Gp

Th

**Control** 

levels of the subunits of some of the complexes of the respiratory chain occurs (complex I being the most affected) (Aricello et al., 2005). It is possible that in WD, functional changes in neurons affecting oxidative metabolism may result in the reversible changes in relative concentration of NAA are reported in those study here. These findings suggest, that the early diagnosis and treatment of WD patients can reduce neurons metabolic disturbances and avoid their degeneration process. It is a very useful information which can prove that in nWD patients with improvement of clinical signs in the brain reversible functional changes of neurons could be detected.

In WD patients treated for long time in nWD and hWD subgroups with improvement or no improvement of clinical status the MRS findings in Table 2 are provided (Tarnacka et al., 2010). In those study we investigated 4 hWD patients, with no improvement and 8 with marked improvement; and 8 nWD patients with marked improvement and 7 with no improvement of clinical status. In hWD patients with improvement the MRS did not show any important pathological changes and in the nWD with improvement significantly higher Cho/Cr, Glx/Cr and Lip/Cr ratios levels compared with controls were noted (Table 2a). In hWD patients with no improvement the lower Cho/Cr and in nWD significantly lower NAA/Cr and higher Cho/Cr and Lip/Cr ratios were detected (Table 2b).

In all WD patients who showed improvement after treatment, MRS revealed a higher level of Cho/Cr, Glx/Cr and Lip/Cr ratios. In our previous study, we demonstrated a decrease in NAA/Cr ratio in all symptomatic patients, which increased after one year of treatment (Tarnacka et al, 2008). In patients treated for a longer time with improvement of clinical status no NAA/Cr decrease was seen; this could suggest that in WD patients the NAA/Cr depletion is reversible during the first few years of treatment and that treatment with z-s or d-p can protect them from further WD progression. The Cho/Cr and Glx/Cr elevation may mirror the glial proliferation that can be detected in patients even after long-term treatment (Horoupian et al., 1988). In this study, the persistence of NAA/Cr was related to a lack of improvement in neurological status (Tarnacka et al., 2010), what is in concordance with earlier reports in the literature showing a lower level of NAA in the striatum in neurologically impaired patients who received treatment (Page et al,.2004; Lucato et al., 2005). The lower level of NAA/Cr in the pallidum in these patients may suggest that, in patients with neurological involvement, persistent neuronal dysfunction may occur as a result of copper and/or iron deposition. NWD patients with no improvement initially had lower ceruloplasmin levels compared with those in whom neurological recovery was noted. It is possible that in these patients the iron metabolism could also be disturbed because of ceruloplasmin deficiency, which can be enhanced during chelation treatment (Medici et al.,2007).

#### **2.1.1 MRS in heterozygous WD gene carriers**

Wilson's disease a genetic disease due to mutation of both alleles of "Wilson's disease" gene ATP7b. The number of heterozygotes – having only one faulty copy of the gene – is high, around 1-2% of the human population (Johnson, 2001). Heterozygote carriers should not have symptoms of hepatic, neurological dysfunctions, but in other recessive disorders such as phenyloketonuria, some deviations in brain function in the literature were reported (Vogel, 1984). Because 1 H-MRS is able to detect brain abnormalities that are invisible in clinical assessment and MRI a study using this method to investigate the probability of brain

levels of the subunits of some of the complexes of the respiratory chain occurs (complex I being the most affected) (Aricello et al., 2005). It is possible that in WD, functional changes in neurons affecting oxidative metabolism may result in the reversible changes in relative concentration of NAA are reported in those study here. These findings suggest, that the early diagnosis and treatment of WD patients can reduce neurons metabolic disturbances and avoid their degeneration process. It is a very useful information which can prove that in nWD patients with improvement of clinical signs in the brain reversible functional changes

In WD patients treated for long time in nWD and hWD subgroups with improvement or no improvement of clinical status the MRS findings in Table 2 are provided (Tarnacka et al., 2010). In those study we investigated 4 hWD patients, with no improvement and 8 with marked improvement; and 8 nWD patients with marked improvement and 7 with no improvement of clinical status. In hWD patients with improvement the MRS did not show any important pathological changes and in the nWD with improvement significantly higher Cho/Cr, Glx/Cr and Lip/Cr ratios levels compared with controls were noted (Table 2a). In hWD patients with no improvement the lower Cho/Cr and in nWD significantly lower

In all WD patients who showed improvement after treatment, MRS revealed a higher level of Cho/Cr, Glx/Cr and Lip/Cr ratios. In our previous study, we demonstrated a decrease in NAA/Cr ratio in all symptomatic patients, which increased after one year of treatment (Tarnacka et al, 2008). In patients treated for a longer time with improvement of clinical status no NAA/Cr decrease was seen; this could suggest that in WD patients the NAA/Cr depletion is reversible during the first few years of treatment and that treatment with z-s or d-p can protect them from further WD progression. The Cho/Cr and Glx/Cr elevation may mirror the glial proliferation that can be detected in patients even after long-term treatment (Horoupian et al., 1988). In this study, the persistence of NAA/Cr was related to a lack of improvement in neurological status (Tarnacka et al., 2010), what is in concordance with earlier reports in the literature showing a lower level of NAA in the striatum in neurologically impaired patients who received treatment (Page et al,.2004; Lucato et al., 2005). The lower level of NAA/Cr in the pallidum in these patients may suggest that, in patients with neurological involvement, persistent neuronal dysfunction may occur as a result of copper and/or iron deposition. NWD patients with no improvement initially had lower ceruloplasmin levels compared with those in whom neurological recovery was noted. It is possible that in these patients the iron metabolism could also be disturbed because of ceruloplasmin deficiency, which can be enhanced

Wilson's disease a genetic disease due to mutation of both alleles of "Wilson's disease" gene ATP7b. The number of heterozygotes – having only one faulty copy of the gene – is high, around 1-2% of the human population (Johnson, 2001). Heterozygote carriers should not have symptoms of hepatic, neurological dysfunctions, but in other recessive disorders such as phenyloketonuria, some deviations in brain function in the literature were reported (Vogel, 1984). Because 1 H-MRS is able to detect brain abnormalities that are invisible in clinical assessment and MRI a study using this method to investigate the probability of brain

NAA/Cr and higher Cho/Cr and Lip/Cr ratios were detected (Table 2b).

during chelation treatment (Medici et al.,2007).

**2.1.1 MRS in heterozygous WD gene carriers** 

of neurons could be detected.

changes in Hzc was performed (Tarnacka et.al, 2009b). We observed statistically significant higher ratios of Glx/Cr and Lip/Cr in 1 H-MRS in Hzc in the pallidum (Table 3). Glx is a glial marker; the increase of Glx/Cr could correlate with the glial proliferation as a result of the protective function of the astrocytes to copper and iron accumulation. The significantly higher level of Lip/Cr may reflect the liberation of lipids from membranes during their breakdown as a reaction to copper and iron overload. Our results may instead suggest that WD Hz carriers accumulate free copper or iron or both in basal ganglia and thalami. Hzc in our study were much older than WD cases are and in an animal model of WD copper and iron levels were found to be augmented during aging in striatum and substantia nigra (Kim et al., 2001). It is possible that during aging in Hz there is an increase of copper and iron accumulation because of ATP7b and ceruolplasmin impairment. There is accumulating evidence that ceruloplasmin, a copper protein with ferroxidase activity, plays an important role in iron metabolism and is also the principal plasma copper-binding protein (Harris et al., 1998). In our Hz the ceruloplasmin level was lower in only four cases, but ceruloplasmin levels in the blood are influenced by acute-phase reactions. The copper accumulation in the brain can also be caused by disruption in its excretion. In animal models, heterozygous mice have a reduced ability to excrete copper, which may indicate that half of the normal liver ATP7b copper transporter activity is insufficient to deal with a large amount of copper intake (Cheach et al., 2007). Like iron, copper is a heavy metal that can induce the Fenton reaction via production of free radicals and copper and iron accumulation increased with aging might have deleterious of liver and brain via oxidative stress.


\*significantly higher than in controls, p<0.00006

\*\* significantly higher than in controls, p<0.00002

¥ significantly higher than in controls, p<0.0006

¶ significantly higher than in controls, p<0.000001

Table 3. The mean metabolites ratios in heterozygotes (Hz) and control subjects from left and right globus pallidus (Gp) and right and left thalamus (Th) with standard deviations.

#### **2.1.2 MRS versus Wilson's disease treatment**

The treatment of WD patients may take many forms, from drug therapy (chelating agents as d-penicillamine (d-p) or trientine, zinc therapy, tetrathiomolibdate) to liver transplantation

Wilson's Disease in Brain Magnetic Resonance Spectroscopy 87

a)

b) Fig. 6. The mean metabolites ratios before treatment with d-p (a) and z-s (b), after minimum

a. statistically higher than in controls ( p<0,0003), b. statistically higher than before treatment (p<0,01), c. statistically higher than in controls (p<0,00004),

a. significantly higher than in controls (p<0,003), b. significantly higher than in controls (p<0,0002), c. significantly higher than in controls (p<0,001).

one year of treatment and in control subjects.

(LT). The aim of medical treatment for WD is to remove the toxic deposit of copper from the body to produce a negative copper balance, and to prevent its reaccumulation. Successful therapy is measured in terms of a restoration of normal levels of free serum copper and its excretion in the urine. In 9 patients treated for more than one year with d-p in author's preliminary study, statistically significant increase of Glx/Cr and NAA/Cr ratios was noticed in contrary to 8 patients treated with zinc sulphate (z-s) were decrease of ml/Cr and NAA/Cr was detected (Figure 6a,b). D-penicillamine seams to be more effective (increase of NAA/Cr) during one year treatment, compared with z-s, but those study was performed on a quite small sample of patients. Those findings can also confirm Brewers observations, that zinc compounds are slow acting drugs and the decrease of mI/Cr can reflect the metabolic deterioration during zinc therapy (Brewer, 2005).

Current therapeutic strategies are often effective in the treatment of WD and provide a negative copper balance, however the course of WD in some treated patients may be unpredictable (Członkowska et al., 2005). The effectiveness of chelation or zinc therapy should therefore be monitored. So far, the only useful tests of monitoring the treatment efficiency are laboratory and imaging methods. Biochemical tests used in monitoring WD treatment include plasma liver tests, caeruloplasmin, copper and zinc in serum, and urinary copper and zinc excretion tests. These methods may sometimes be not sufficiently in particular at the beginning of the treatment when most of treatment side-effects occur as iatrogenic worsening during d-p therapy. In this stage wee postulated that 1H-MRS can be used in early stages of treatment monitoring when the alternations of NAA, mI and Glx can mirror therapeutic response or deterioration (Tarnacka et al, 2008). We think that 1H-MRS can also be helpful in differential diagnosis of clinical deterioration in patients with WD.

In patients with progressive liver failure or acute liver failure from fulminant hepatitis with or without intravenous hemolysis, orthotropic hepatic transplantation is an efficient treatment (Hoogenraad, 1997). Hepatic transplantation is also indicated in the absence of liver failure in patients with neurological WD in whom chelation therapy has proved ineffective, and significant improvements in neurological features have been reported (Polson et al., 1987). We presented a study of spectroscopic changes in globus pallidus in 3 patients with WD undergoing LT. The first case was a patient with neurological and hepatic impairment who was undergoing LT because of liver cirrhosis exacerbation. The second and third patients were subjects only with hepatic signs and the LT was performed because of liver failure. In the patient with neurological signs the MRS was performed 4 months after LT because there was no neurological improvement, and again one year after that. In the first MRS in those patient there was an increase in Glx/Cr, NAA/Cr and Lip/Cr ratios levels compared with the controls (Figure 7a). After one year's observation time an improvement of clinical status was observed and no changes of metabolites ratios were seen(Figure 7a). Before LT in the patient with liver cirrhosis mI/Cr ratio level was very low, in both patients with liver failure a decreased level of Glx/Cr, NAA/Cr and increased Lip/Cr ratios levels compared with controls were found (Figure 7b,c). After LT in those patients an increase of mI, Glx/Cr and NAA/Cr was seen (Figure 7.bc). Those data provide that after LT in those WD patients a renormalization of brain metabolites changes detected in 1 H-MRS could be seen.

(LT). The aim of medical treatment for WD is to remove the toxic deposit of copper from the body to produce a negative copper balance, and to prevent its reaccumulation. Successful therapy is measured in terms of a restoration of normal levels of free serum copper and its excretion in the urine. In 9 patients treated for more than one year with d-p in author's preliminary study, statistically significant increase of Glx/Cr and NAA/Cr ratios was noticed in contrary to 8 patients treated with zinc sulphate (z-s) were decrease of ml/Cr and NAA/Cr was detected (Figure 6a,b). D-penicillamine seams to be more effective (increase of NAA/Cr) during one year treatment, compared with z-s, but those study was performed on a quite small sample of patients. Those findings can also confirm Brewers observations, that zinc compounds are slow acting drugs and the decrease of mI/Cr can reflect the metabolic

Current therapeutic strategies are often effective in the treatment of WD and provide a negative copper balance, however the course of WD in some treated patients may be unpredictable (Członkowska et al., 2005). The effectiveness of chelation or zinc therapy should therefore be monitored. So far, the only useful tests of monitoring the treatment efficiency are laboratory and imaging methods. Biochemical tests used in monitoring WD treatment include plasma liver tests, caeruloplasmin, copper and zinc in serum, and urinary copper and zinc excretion tests. These methods may sometimes be not sufficiently in particular at the beginning of the treatment when most of treatment side-effects occur as iatrogenic worsening during d-p therapy. In this stage wee postulated that 1H-MRS can be used in early stages of treatment monitoring when the alternations of NAA, mI and Glx can mirror therapeutic response or deterioration (Tarnacka et al, 2008). We think that 1H-MRS can also be helpful in differential diagnosis of clinical deterioration in patients

In patients with progressive liver failure or acute liver failure from fulminant hepatitis with or without intravenous hemolysis, orthotropic hepatic transplantation is an efficient treatment (Hoogenraad, 1997). Hepatic transplantation is also indicated in the absence of liver failure in patients with neurological WD in whom chelation therapy has proved ineffective, and significant improvements in neurological features have been reported (Polson et al., 1987). We presented a study of spectroscopic changes in globus pallidus in 3 patients with WD undergoing LT. The first case was a patient with neurological and hepatic impairment who was undergoing LT because of liver cirrhosis exacerbation. The second and third patients were subjects only with hepatic signs and the LT was performed because of liver failure. In the patient with neurological signs the MRS was performed 4 months after LT because there was no neurological improvement, and again one year after that. In the first MRS in those patient there was an increase in Glx/Cr, NAA/Cr and Lip/Cr ratios levels compared with the controls (Figure 7a). After one year's observation time an improvement of clinical status was observed and no changes of metabolites ratios were seen(Figure 7a). Before LT in the patient with liver cirrhosis mI/Cr ratio level was very low, in both patients with liver failure a decreased level of Glx/Cr, NAA/Cr and increased Lip/Cr ratios levels compared with controls were found (Figure 7b,c). After LT in those patients an increase of mI, Glx/Cr and NAA/Cr was seen (Figure 7.bc). Those data provide that after LT in those WD patients a renormalization of brain metabolites changes detected

deterioration during zinc therapy (Brewer, 2005).

with WD.

in 1 H-MRS could be seen.

a)

b)

Fig. 6. The mean metabolites ratios before treatment with d-p (a) and z-s (b), after minimum one year of treatment and in control subjects.

Wilson's Disease in Brain Magnetic Resonance Spectroscopy 89

**I study**

**II study**

**contr.**

**mI Glx NAA Lip**

Fig. 7. The mean metabolites ratios in globus pallidus in 3 patients undergoing liver transplantation, in first patient a) second- b) third- c), explanation in the text.

c)

**I**n conclusion, the results of the present studies appear to prove that in WD the astrocytic abnormalities occur primarily with dysfunction of the astrocytic-neuronal interactions. This process leads to metabolic derangement in neurons. During the disease progression and increasing copper deposition in the brain the compensation reactions as metallothionein synthesis become inefficient and the degenerative changes of astrocytes prevail over their detoxicative possibilities. The copper liberates from damaged astrocytes, and can also damage the neurons directly. Free copper can induce the neuronal mitochondria dysfunction causing the impaired NAA synthesis, but on this stage the destructive process can be reversible, because the increase of NAA/Cr can be observed in most treated patients. This is a very important conclusion, because it shows that the first year of treatment are pivotal in the therapeutic process. In majority of patients in this stage the central nervous system damage can be reversible. These findings emphasize the need for early diagnosis and treatment of WD. Effective antioxidant therapy introduced in early stages of WD might have modifying effect of metabolic disturbances in neurons. The neuronal dysfunction lasting longer can cause non reversible degenerative process like in patients treated for long time with no neurological improvement. Irreversible astrocytic dysfunction caused by copper can be related with disturbed ceruloplasmin synthesis causing iron overload and lead to lack of neurological improvement in some WD patients. The 1H-MRS technique can be useful in treatment monitoring in patients with WD in early stages (until one year). MRS seems to be more effective method in comparison with MRI

**0**

**3. Conclusion** 

**0,2**

**0,4**

**0,6**

**0,8**

**1**

**1,2**

**1,4**

**1,6**

a)

b)

Fig. 7. The mean metabolites ratios in globus pallidus in 3 patients undergoing liver transplantation, in first patient a) second- b) third- c), explanation in the text.

#### **3. Conclusion**

88 Magnetic Resonance Spectroscopy

**I study**

**II study**

**contr.**

**I study**

**II study**

**contr.**

**mI Glx NAA Lip**

**mI Glx NAA Lip**

b)

a)

**0**

**0**

**0,2**

**0,4**

**0,6**

**0,8**

**1**

**1,2**

**1,4**

**1,6**

**1,8**

**0,2**

**0,4**

**0,6**

**0,8**

**1**

**1,2**

**1,4**

**1,6**

**1,8**

**2**

**I**n conclusion, the results of the present studies appear to prove that in WD the astrocytic abnormalities occur primarily with dysfunction of the astrocytic-neuronal interactions. This process leads to metabolic derangement in neurons. During the disease progression and increasing copper deposition in the brain the compensation reactions as metallothionein synthesis become inefficient and the degenerative changes of astrocytes prevail over their detoxicative possibilities. The copper liberates from damaged astrocytes, and can also damage the neurons directly. Free copper can induce the neuronal mitochondria dysfunction causing the impaired NAA synthesis, but on this stage the destructive process can be reversible, because the increase of NAA/Cr can be observed in most treated patients. This is a very important conclusion, because it shows that the first year of treatment are pivotal in the therapeutic process. In majority of patients in this stage the central nervous system damage can be reversible. These findings emphasize the need for early diagnosis and treatment of WD. Effective antioxidant therapy introduced in early stages of WD might have modifying effect of metabolic disturbances in neurons. The neuronal dysfunction lasting longer can cause non reversible degenerative process like in patients treated for long time with no neurological improvement. Irreversible astrocytic dysfunction caused by copper can be related with disturbed ceruloplasmin synthesis causing iron overload and lead to lack of neurological improvement in some WD patients. The 1H-MRS technique can be useful in treatment monitoring in patients with WD in early stages (until one year). MRS seems to be more effective method in comparison with MRI

Wilson's Disease in Brain Magnetic Resonance Spectroscopy 91

Frydman, M.; Bonne-Tamir, B. & Farrer, L. et al. (1985) Assignment of the gene for Wilson

Harris, L.; Klomb, L. & Giltin J. (1998). Aceruloplasminemia: an inherited neurodegenerative disease with impairment of iron homeostasis. *Am J Clin Nutr* 67 (suppl):972S-977S. Hoogenraad, T. (1997). Wilson's Disease. London Philadelphia Toronto Sydney Tokyo: *WB.* 

Horoupian, D.; Sternlieb, I. & Scheinberg, I. (1988). Neuropathological findings in

Johnson, S. (2001). Is Parkinson's disease the heterozygote form of Wilson's disease: PD=1/2

Kim, H.; Kim, D. & Lee, B. et al. (2001). Diffusion-weighted image and MR spectroscopic analysis of a case of MELAS with repeated attacks. *Yonsei Med J* 42:128-33. Kraft, E. Trenkwalder, C. & Then Berg, F. et al. (1999). Magnetic resonance proton

Kreis, R.; Ross. B. &, Farrow, N. et al. (1992). Metabolic disorders of the brain in chronic hepatic encephalopathy detected with H-1 MR spectroscopy. *Radiology* 182:19-27. Loudianos, G.; Dessi, V. & Lovicu, M. et al. (1999) Molecular characterization of Wilson

Lucato, L.; Barbosa, M. & Machado, A. et al. (2005). Proton MR spectroscopy in Wilson

Medici, V.; Di Leo, V. & Lamboglia, F. et al. (2007). Effect of penicillamine and zinc on iron

Moffett, J.; Ross, B. & Arun, P. et al. (2007). N -acetylaspartate in CNS: from neurodiagnostic

Page, R.; Davie, F. & MacManus, D. et al. (2004). Clinical correlation of brain MRI and MRS abnormalities in patients with Wilson's disease. *Neurology* 63:638-643. Polson, R.; Rolles, K. & Calne, R. et al. (1994). Reversal of severe neurological manifestations of Wilson's disease following orthoptic liver transplantation. *Q J Med* 244:685-691. Ross, B.; Jacobson, S. , & Villami, F. et al. (1994). Subclinical hepatic encephalopathy: proton

Rudkin, T. & Arnold, D. (1999). Proton magnetic resonance spectroscopy for the diagnosis

Schilsky, M. & Tavill, A. (2003) Wilson disease. In Disease of the Liver, edn 9, 1169–1186

Sinha, S.; Taly, A. & Ravishankar, S. et al. (2006). Wilson's disease: cranial MRI observations

Tarnacka, B.; Szeszkowski, W. & Gołębiowski, M. et al. (2008). MR spectroscopy in monitoring the treatment of Wilson's disease patients. *Mov Disord* 23:1560-6. Tarnacka, B.; Szeszkowski, W. & Gromadzka, G. et al. (2010). Brain Proton Magnetic

Spectroscopy in long-term treatment of Wilson's disease patients. *Metab Brain Dis*

metabolizm in Wilson's disease. *Scand J Gastroenterol* 42:1495-1500.

disease in the Sardinian population—evidence of a founder effect. *Hum Mutat* 14:

spectroscopy of brain in Wilson's disease. *J Neurol* 246:693-9.

disease: analysis of 36 cases. *Am J Neuroradiol* 26:1066-1071.

to neurobiology. *Progress in Neurobiol* 81:89-131.

MR spectroscopic abnormalities. *Radiology* 93:457-63.

and clinical correlation. *Neuroradiology* 48:613-21.

and management of cerebral disorders. *Arch Neurol* 56:919-26. Scheinberg, I. & Sternlieb, I. (1984) Wilson disease. Philadelphia: *WB Saunders*

(Eds Schiff ER et al.) Philadelphia: *Lippincott Williams & Wilkins*

82: 1891–1821

67.

294–303.

25:325-9.

*Saunders Company* Ttd.

WD? *Medical Hypotheses*;56:171-173.

disease to chromosome 13: linkage to the esterase D locus. *Proc Natl Acad Sc USA*

penicillamine – treated patients with Wilson's disease. Clinical *Neuropathology* 7:62-

in monitoring effectiveness of treatment during first year. MRS could be helpful in differential diagnosis of the clinical deterioration in treated WD patients. The alternations of NAA/Cr ratio in neurologically impaired patients and mI/Cr and Glx/Cr in patients with liver failure could be a sensitive markers of the clinical recovery and deterioration in those WD patients. MR spectroscopy could be a useful complementary tool in other monitoring treatment methods in WD patients.

#### **4. Acknowledgment**

This work was supported by grant from the Polish Ministry of Education and Science (3P05B/119/23).

#### **5. References**


in monitoring effectiveness of treatment during first year. MRS could be helpful in differential diagnosis of the clinical deterioration in treated WD patients. The alternations of NAA/Cr ratio in neurologically impaired patients and mI/Cr and Glx/Cr in patients with liver failure could be a sensitive markers of the clinical recovery and deterioration in those WD patients. MR spectroscopy could be a useful complementary tool in other

This work was supported by grant from the Polish Ministry of Education and Science

Aricello, M.; Rotillo, G. & Rossi, L. (2005). Copper-dependent toxicity in SH-SY5Y

Bates, T.; Strangward M. & Keelan J. (1996). Inhibition of N-acetylaspartate production:

Brewer, G. (2005). Neurologically presenting Wilson's disease: epidemiology,

Bull, P.; Thomas, G. & Rommens, J. et al. (1993) The Wilson disease gene is a putative copper transporting P-type ATPase similar to the Menkes gene. *Nat Genet* 5: 327–337 Cheach, D.; Deal, Y. & WRIGHT, P. et al. Heterozygous tx mice have an increased sensitivity

Cordoba, J.; Alonso, J. & Rovira, A. et al. (2001). The development of low grade cerebral

Cordoba, J.; Sanpedro, F. & Alonso, J. et al. (2002). 1 H magnetic resonance in the study of hepatic encephalopthy in humans. *Metabolic Brain Disease* 17:415-29. Członkowska, A.; Galewicz, A. & Rodo M et al. (1973). Observation on copper metabolism

Członkowska, A.; Tarnacka, B. & Litwin, T. et al (2005). Wilson's disease - cause of mortality in 164 patients during 1992-2003 observation period. *J Neurol* 252:698-703. Danielsen, E. & Ross B. (1999). Magnetic resonance spectroscopy diagnosis of neurological

Dauntry, C.; Vaufrey, F. & Brouillet, E. et al. (2000). Early N-acetylaspartate depletion is a

Figus, A.; Angius, A. & Loudianos, G. et al. (1995) Molecular pathology and haplotype

metabolizm: studies with [15N] glutamate*. J Neurochem* 51:892-902.

marker of neuronal dysfunction in rats and primates chronically treated with the mitochondrial toxin 3-nitropropionic acid. *J Cereb Blood Flow Metab* 20:789-799. Erecinska, M.; Zaleska, M. & Nissim, I. et al. (1988). Glucose and synaptosomal glutamate

analysis of Wilson disease in Mediterranean populations. *Am J Hum Genet* 57: 1318–

implications for 1 H-MRS studies in vivo. *Neuroreport* 7:1397-1400.

pathophysiology and treatment. *CNS Drugs*;19:185-92.

abnormalities after liver transplantation. *J Hepatol* 3:596-604.

in Wilson's disease. Acta Univ Caro Med 56:175-178.

diseases. New York: *Marcel Dekker,*

neuroblastoma cells involves mitochondrial damage. *Biochem Biophys Res Commun* 

to copper loading: Implications for Wilson's disease carriers. *Biometals* 2007;20:751-

edema in cirrhosis is supported by the evolution of 1H magnetic resonance

monitoring treatment methods in WD patients.

**4. Acknowledgment** 

327:454-459.

760 .

1324.

(3P05B/119/23).

**5. References** 


**5** 

*Turkey* 

**MRS in MS, With and Without Interferon** 

**Changes of Metabolites in the Brain, and to Monitor Disease Progression** 

*1Baskent University, Faculty of Medicine, Department of Neurology,* 

*2Baskent University, Faculty of Medicine, Department of Radiology, Ankara* 

*3Baskent University, Faculty of Medicine, Department of Internal Medicine, Ankara* 

Multiple sclerosis (MS) is a common cause of chronic neurological disability in young adults. Magnetic Resonance Imaging (MRI) readily identifies multifocal white matter lesions(WML) that represent areas of demyelination and is thus useful in supporting the diagnosis of MS, even after a single clinical episode(McDonald et al., 2001). MRI is also widely used to monitor disease progression in natural history and in therapeutic trial studies. However, conventional MRI does not detect the subtle histopathological changes that are described in the normal-appearing white matter (NAWM) in patients with MS(Allen & McKeown, 1979; Bitsch et al., 1999). MRS has been used in the evaluation of MS to help further define the nature of the lesions revealed by T2-weighted MRI(Arnold, Matthews, Francis, O'Connor, & Antel, 1992; Arnold et al., 1994; Bitsch et al., 1999; Davie et al., 1997; De Stefano et al., 1995; Fu et al., 1998; Gonen et al., 2000; Matthews et al., 1996; Narayana, Doyle, Lai, & Wolinsky, 1998; Pan et al., 1996; Rooney et al., 1997; Sarchielli et al., 1999; Tourbah et al., 1999; van Walderveen, Barkhof et al., 1999). The measurement of N-acetly aspartic acid (NAA) is thought to be a marker of axonal loss, damage or dysfunction. Many studies (Arnold et al., 1992; Arnold et al., 1994; Davie et al., 1997; De Stefano et al., 1995; Husted et al., 1994; Narayana et al., 1998,( Allen & McKeown, 1979; De Stefano et al., 1995; Fu et al., 1998; Husted et al., 1994; Narayana et al., 1998) have shown a reduction in the absolute concentration of NAA or the NAA/Cr ratio in MS lesions. However, pathological(Allen & McKeown, 1979) and quantitative MRSI studies(Armspach, Gounot, Rumbach, & Chambron, 1991; Filippi et al., 1995; Gasperini et al., 1996; Loevner et al., 1995) have shown that in patients with clinically definite MS, abnormalities also occur in the NAWM(Arnold et al., 1992; Davie et al., 1997; Davie et al.,

**1. Introduction** 

 \*

Corresponding Author

Münire Klnç Toprak1,\*, Banu Çakir2,

*Fevzi Cakmak Caddesi, Bahcelievler, Ankara* 

E.Meltem Kayahan Ulu2 and Zübeyde Arat3

**Beta 1a Treatment, to Define the Dynamic** 


## **MRS in MS, With and Without Interferon Beta 1a Treatment, to Define the Dynamic Changes of Metabolites in the Brain, and to Monitor Disease Progression**

Münire Klnç Toprak1,\*, Banu Çakir2,

E.Meltem Kayahan Ulu2 and Zübeyde Arat3 *1Baskent University, Faculty of Medicine, Department of Neurology, Fevzi Cakmak Caddesi, Bahcelievler, Ankara 2Baskent University, Faculty of Medicine, Department of Radiology, Ankara 3Baskent University, Faculty of Medicine, Department of Internal Medicine, Ankara Turkey* 

#### **1. Introduction**

92 Magnetic Resonance Spectroscopy

Tarnacka, B.; Szeszkowski, W. & Gołębiowski, M. et al. (2009). Metabolic changes in 37

Tarnacka, B.; Szeszkowski, W. & Gromadzka, G. et al. (2009). Heterozygous carriers for

Thomas, G.; Forbes, J. & Roberts E. et al. (1995) The Wilson disease gene: spectrum of

Van der Hart, M.; Czeh, B. & Biurrun, G. et al. (2000). P receptor antagonist an

Victor, M. (1999). Neurologic manifestation of gastrointestinal diseases. In Asbury AK.

Vogel, F. (1984). Clinical consequences of heterozygosity for autosomal-recessive diseases.

Wilson, S. (1912) Progressive lenticular degeneration: a familial nervous disease associated

Yamaguchi, Y.; Heiny, M. & Gitlin, J. et al. (1993) Isolation and characterization of a human

mutations and their consequences. *Nat Genet* 9: 210–217.

Spectroscopy. *Park Rel Dis* 15:582-6.

*Saunders, Philadelphia*, pp 1470-82.

with cirrhosis of the liver. *Brain* 34: 295–509

*Clinical Genetics* 25:381-415.

24:463-8.

7:933-41.

197: 271–277

newly diagnosed Wilson's disease patients assessed by Magnetic Resonance

Wilson's disease - magnetic spectroscopy changes in the brain. *Metab Brain Dis*

cloimipramine prevent stress-induced alternations in cerebral metabolites, cytogenesis in the dentate gyrus and hippocampal volume. *Molecular Psychiatry*

McKann CM, McDonald W. Diseases of the nervous system. Clinical Neurobiology,

liver cDNA as a candidate gene for Wilson disease. *Biochem Biophys Res Common*

Multiple sclerosis (MS) is a common cause of chronic neurological disability in young adults. Magnetic Resonance Imaging (MRI) readily identifies multifocal white matter lesions(WML) that represent areas of demyelination and is thus useful in supporting the diagnosis of MS, even after a single clinical episode(McDonald et al., 2001). MRI is also widely used to monitor disease progression in natural history and in therapeutic trial studies. However, conventional MRI does not detect the subtle histopathological changes that are described in the normal-appearing white matter (NAWM) in patients with MS(Allen & McKeown, 1979; Bitsch et al., 1999). MRS has been used in the evaluation of MS to help further define the nature of the lesions revealed by T2-weighted MRI(Arnold, Matthews, Francis, O'Connor, & Antel, 1992; Arnold et al., 1994; Bitsch et al., 1999; Davie et al., 1997; De Stefano et al., 1995; Fu et al., 1998; Gonen et al., 2000; Matthews et al., 1996; Narayana, Doyle, Lai, & Wolinsky, 1998; Pan et al., 1996; Rooney et al., 1997; Sarchielli et al., 1999; Tourbah et al., 1999; van Walderveen, Barkhof et al., 1999). The measurement of N-acetly aspartic acid (NAA) is thought to be a marker of axonal loss, damage or dysfunction. Many studies (Arnold et al., 1992; Arnold et al., 1994; Davie et al., 1997; De Stefano et al., 1995; Husted et al., 1994; Narayana et al., 1998,( Allen & McKeown, 1979; De Stefano et al., 1995; Fu et al., 1998; Husted et al., 1994; Narayana et al., 1998) have shown a reduction in the absolute concentration of NAA or the NAA/Cr ratio in MS lesions. However, pathological(Allen & McKeown, 1979) and quantitative MRSI studies(Armspach, Gounot, Rumbach, & Chambron, 1991; Filippi et al., 1995; Gasperini et al., 1996; Loevner et al., 1995) have shown that in patients with clinically definite MS, abnormalities also occur in the NAWM(Arnold et al., 1992; Davie et al., 1997; Davie et al.,

<sup>\*</sup> Corresponding Author

MRS in MS, With and Without Interferon Beta 1a Treatment,

to Define the Dynamic Changes of Metabolites in the Brain, and to Monitor Disease Progression 95

Fig. 1. a-f: Axial fluid-attenuated, inversion recovery (FLAIR) MR image of a 50-year-old woman receiving INF-1a treatment shows a voxel placed over a WM lesion in the right centrum semiovale (1a). The multivoxel CSI MRI spectrum of this lesion is obtained during study months zero, 1, 3, 6, and 12 (b-f). The results of proton MRSI show that there is a

statistically significant increment in absolute NAA and Cho values.

1994; Fu et al., 1998; Husted et al., 1994; Narayana et al., 1998). Although the principal finding of MRS studies in patients with MS was a decrease in the NAA peak area, opposite results were obtained for other metabolites(Davie et al., 1994; Davies, Newcombe, Williams, McDonald, & Clark, 1995; De Stefano et al., 1995; Husted et al., 1994; Larsson et al., 1991). Despite the difficulties in interpreting data obtained in vivo, MRS provides direct information about metabolic variations and the damage to or integrity of myelin and axons, which are not revealed by traditional MRI(Arnold et al., 1994; Ferguson, Matyszak, Esiri, & Perry, 1997; Fernando et al., 2004).

In patients with relapsing remitting multiple sclerosis (RRMS), interferons reduce the attack frequency of the disease, the mean attack duration, the annual lesion burden, and the frequency of gadolinium (Gd) enhancement on MRI. However the mechanism of action of interferons in patients with MS and the effect of interferons on axonal injury remain undefined. We questioned whether treatment with interferon beta-1a (INF-1a) would enable the recovery of injured axons. To determine this, we used MRSI to examine changes in the metabolic peaks of neuronal markers NAA, Cho(Choline), and Cr(Creatine) after treatment with INF-1a(Rebif 44 g) in a small group of patients with RRMS. An untreated group was studied for comparison, and healthy volunteers were used as controls.

#### **2. Materials and methods**

#### **2.1 Patients and methods**

We studied 10 patients (7 women 3 men) with clinically definite RRMS (median age, 39 years; age range 22-50 years). The median expanded status scale score (EDSS) was 2.1 (range 1.0-3.0), and the median duration of disease was 4.7 years (range, 1-20 years). In the 2 years before their participation in the study, all patients had had at least one relapse. None of the patients had undergone immunosuppressant therapy ever before, or corticosteroid treatment within 2 months before the initiation of the study. Five patients elected not to be

1994; Fu et al., 1998; Husted et al., 1994; Narayana et al., 1998). Although the principal finding of MRS studies in patients with MS was a decrease in the NAA peak area, opposite results were obtained for other metabolites(Davie et al., 1994; Davies, Newcombe, Williams, McDonald, & Clark, 1995; De Stefano et al., 1995; Husted et al., 1994; Larsson et al., 1991). Despite the difficulties in interpreting data obtained in vivo, MRS provides direct information about metabolic variations and the damage to or integrity of myelin and axons, which are not revealed by traditional MRI(Arnold et al.,

In patients with relapsing remitting multiple sclerosis (RRMS), interferons reduce the attack frequency of the disease, the mean attack duration, the annual lesion burden, and the frequency of gadolinium (Gd) enhancement on MRI. However the mechanism of action of interferons in patients with MS and the effect of interferons on axonal injury remain undefined. We questioned whether treatment with interferon beta-1a (INF-1a) would enable the recovery of injured axons. To determine this, we used MRSI to examine changes in the metabolic peaks of neuronal markers NAA, Cho(Choline), and Cr(Creatine) after treatment with INF-1a(Rebif 44 g) in a small group of patients with RRMS. An untreated

We studied 10 patients (7 women 3 men) with clinically definite RRMS (median age, 39 years; age range 22-50 years). The median expanded status scale score (EDSS) was 2.1 (range 1.0-3.0), and the median duration of disease was 4.7 years (range, 1-20 years). In the 2 years before their participation in the study, all patients had had at least one relapse. None of the patients had undergone immunosuppressant therapy ever before, or corticosteroid treatment within 2 months before the initiation of the study. Five patients elected not to be

group was studied for comparison, and healthy volunteers were used as controls.

**2. Materials and methods 2.1 Patients and methods** 

1994; Ferguson, Matyszak, Esiri, & Perry, 1997; Fernando et al., 2004).

Fig. 1. a-f: Axial fluid-attenuated, inversion recovery (FLAIR) MR image of a 50-year-old woman receiving INF-1a treatment shows a voxel placed over a WM lesion in the right centrum semiovale (1a). The multivoxel CSI MRI spectrum of this lesion is obtained during study months zero, 1, 3, 6, and 12 (b-f). The results of proton MRSI show that there is a statistically significant increment in absolute NAA and Cho values.

MRS in MS, With and Without Interferon Beta 1a Treatment,

between the untreated and treated MS patients (Table 2).

NAA: N-acetyl aspartate, Cho: Choline, Cr: Creatine

and untreated patients and controls

therapy once daily for 5 days.

throughout the study.

**3.2 MRSI results** 

to Define the Dynamic Changes of Metabolites in the Brain, and to Monitor Disease Progression 97

clinical MS attack at the initiation of the study in 3 patients during the sixth month of the study. The results of serial MR examinations of another patient revealed new T2 lesions without Gd-enhancement that were present without evidence of a clinical MS attack from the initiation of the study until the twelfth study month. However, during the twelfth month, that patient experienced a clinical MS attack (moderate paresis [4/5] of the left leg), and a new Gd-enhancing lesion was identified. And only that patient received pulse steroid

Five other patients received INF-1a treatment (subcutaneous INF-1a 44 µg –Rebif-3 times per week). Among the subjects who received INF, two patients had Gd-enhancing new T2 lesions on almost every MRI examination, another patient had a Gd-enhancing lesion only on the last MRI examination of the study, and the remaning patients exhibited no new lesions. None of those patients experienced a clinical MS attack accompanying the MRI activities. There was no significant change in the EDSS scores of the patients in either group

There was no statistically significant difference between the basal metabolite levels in the NAWM of patients with MS and the healthy control subjects (*P*> .005) (Table 1). There was also no statistically significant difference between the basal metabolite peaks in the WML

**NAWM metabolic peaks NAA Basal CHO Basal CR Basal**  Untreated Mean 6,5±4,6 4,3±1,9 3,9±1,8 Median 4,8950 3,4650 3,1650 Treated Mean 7,1±2,3 4,5±1,4 4,3±1,3 Median 7,8100 4,7550 4,5200 Control Mean 6,8±0,4 4,1±0,7 4,0±0,2 Median 6,9250 4,2900 3,9800 **WML metabolic peaks NAA Basal CHO Basal CR Basal**  Untreated Mean 6,03,7 4,41,9 3,81,8 Median 4,8900 4,6700 3,1900 Treated Mean 6,82,5 4,81,6 3,91,2 Median 5,4800 5,0950 3,7900

Table 1. Basal normal-appearing white matter and white matter metabolite levels in treated

We analyzed whether any difference emerged in the NAWM metabolite peaks between the treated and untreated patients throughout the study period and found that a significant difference existed in NAWM NAA peaks (P = .003); Cho peaks (P = .001); and Cr peaks (P = .006) beginning from the first month of the study. When we evaluated the groups separately to assess when that difference emerged between the basal and the following months' NAA, Cho and Cr peaks we noted that in the treatment group, the NAA peaks began to increase during the sixth month (P = .000); Cho peaks increased during the sixth and twelfth months (P = .000); and an increment in the Cr peak was found in the third (P< .04), sixth (P = .001), and twelfth

treated. Five other patients received INF-1a treatment. All patients were evaluated at the beginning of the study and during the first, third, sixth, and twelfth month. During their follow-up evaluations, neurological examinations that included EDSS assessment were performed in addition to MR studies with gadopentetate dimegluminelinium (Gd) 0,1 mmol/kg intravenously, and MRSI (Figure 1a-f). A relapse was defined as the appearance of new neurological symptoms or worsening of prior neurological symptoms of 48 hours' duration in addition to objective evidence of a change in neurological status of a patient who has been clinically stable for the 4 prior weeks. All patients suspected of having an attack were examined within 7 days of the onset of symptoms and were treated with intravenous methyprednisolone (IVMP) at a dose of 1 g/day for 5 days. At the beginning of the study, MRI and MRSI studies were also performed on 6 healthy, age-matched and sex-matched control subjects (4 women, 2 men; median age, 33.6 years, age range, 27-48 years) who had neither systemic nor neurologic disease. The local ethics committee approved the study.

#### **2.1.1 MRI studies**

All MRI and MRSI studies were performed with a 1.5 Tesla MR unit (Siemens Magnetom Symphony, Erlangen, Germany) with a standard quadrature head coil. To detect focal WM lesions, we used conventional MRI, including axial T1 weighted (TR 586 ms, TE 15 ms, 5 mm slice thickness), fluid-attenuated, inversion recovery (FLAIR) (TR 9000 ms, TE 98 ms, 5 mm slice thickness), axial and sagittal proton density (TR 2200 ms, TE 30 ms, 5 mm slice thickness), T2-weighted spin echo sequences (TR 2200 ms, TE 80 ms, 5 mm slice thickness) and post-gadolinium axial and sagittal T1 weighted (TR 779 ms, TE 15 ms, 5 mm slice thickness) spin echo sequences. MRI images were acquired with a matrix size of 145x256 mm and FOVof 230x230 mm. During the same session, water-suppressed proton MRSI was performed by means of multivoxel chemical shift imaging (TR/TE: 1500/135, voxel volume 1 cm3). Metabolite peaks of NAA, Cho, and Cr were assessed in both WML -without Gd enhancement- and in NAWM.

#### **2.1.2 Statistics**

SPSS 11 for Windows was used for statistical analysis. The Kruskal-Wallis Test was used to compare the basal NAWM metabolite peaks among treated, untreated and healthy control subjects. The Mann-Whitney U test was used to compare the basal WML metabolite peaks in treated and untreated patients. Then, multivariate analysis of variance (MANOVA) was used to compare the NAWM and WML metabolite peaks in treated and untreated patients throughout the study period. The Wilcoxon signed rank test was used to compare changes in the NAWM and WML metabolite peaks during study months zero, 1, 3, 6, and 12. Results with p< .05 were considered as significant.

#### **3. Results**

#### **3.1 Treatment effect as assessed by relapse rate, and new T2 lesions with and without GD enhancement**

Five patients (3 women and 2 men) elected not to be treated. Four of those 5 patients experienced no attacks during the 12 months of the study. However, the results of MRI examination revealed new Gd-enhancing lesions without a trace of an accompanying clinical MS attack at the initiation of the study in 3 patients during the sixth month of the study. The results of serial MR examinations of another patient revealed new T2 lesions without Gd-enhancement that were present without evidence of a clinical MS attack from the initiation of the study until the twelfth study month. However, during the twelfth month, that patient experienced a clinical MS attack (moderate paresis [4/5] of the left leg), and a new Gd-enhancing lesion was identified. And only that patient received pulse steroid therapy once daily for 5 days.

Five other patients received INF-1a treatment (subcutaneous INF-1a 44 µg –Rebif-3 times per week). Among the subjects who received INF, two patients had Gd-enhancing new T2 lesions on almost every MRI examination, another patient had a Gd-enhancing lesion only on the last MRI examination of the study, and the remaning patients exhibited no new lesions. None of those patients experienced a clinical MS attack accompanying the MRI activities. There was no significant change in the EDSS scores of the patients in either group throughout the study.

#### **3.2 MRSI results**

96 Magnetic Resonance Spectroscopy

treated. Five other patients received INF-1a treatment. All patients were evaluated at the beginning of the study and during the first, third, sixth, and twelfth month. During their follow-up evaluations, neurological examinations that included EDSS assessment were performed in addition to MR studies with gadopentetate dimegluminelinium (Gd) 0,1 mmol/kg intravenously, and MRSI (Figure 1a-f). A relapse was defined as the appearance of new neurological symptoms or worsening of prior neurological symptoms of 48 hours' duration in addition to objective evidence of a change in neurological status of a patient who has been clinically stable for the 4 prior weeks. All patients suspected of having an attack were examined within 7 days of the onset of symptoms and were treated with intravenous methyprednisolone (IVMP) at a dose of 1 g/day for 5 days. At the beginning of the study, MRI and MRSI studies were also performed on 6 healthy, age-matched and sex-matched control subjects (4 women, 2 men; median age, 33.6 years, age range, 27-48 years) who had neither systemic nor neurologic disease. The local ethics committee approved the study.

All MRI and MRSI studies were performed with a 1.5 Tesla MR unit (Siemens Magnetom Symphony, Erlangen, Germany) with a standard quadrature head coil. To detect focal WM lesions, we used conventional MRI, including axial T1 weighted (TR 586 ms, TE 15 ms, 5 mm slice thickness), fluid-attenuated, inversion recovery (FLAIR) (TR 9000 ms, TE 98 ms, 5 mm slice thickness), axial and sagittal proton density (TR 2200 ms, TE 30 ms, 5 mm slice thickness), T2-weighted spin echo sequences (TR 2200 ms, TE 80 ms, 5 mm slice thickness) and post-gadolinium axial and sagittal T1 weighted (TR 779 ms, TE 15 ms, 5 mm slice thickness) spin echo sequences. MRI images were acquired with a matrix size of 145x256 mm and FOVof 230x230 mm. During the same session, water-suppressed proton MRSI was performed by means of multivoxel chemical shift imaging (TR/TE: 1500/135, voxel volume 1 cm3). Metabolite peaks of NAA, Cho, and Cr were assessed in both WML -without Gd

SPSS 11 for Windows was used for statistical analysis. The Kruskal-Wallis Test was used to compare the basal NAWM metabolite peaks among treated, untreated and healthy control subjects. The Mann-Whitney U test was used to compare the basal WML metabolite peaks in treated and untreated patients. Then, multivariate analysis of variance (MANOVA) was used to compare the NAWM and WML metabolite peaks in treated and untreated patients throughout the study period. The Wilcoxon signed rank test was used to compare changes in the NAWM and WML metabolite peaks during study months zero, 1, 3, 6, and 12. Results

**3.1 Treatment effect as assessed by relapse rate, and new T2 lesions with and without** 

Five patients (3 women and 2 men) elected not to be treated. Four of those 5 patients experienced no attacks during the 12 months of the study. However, the results of MRI examination revealed new Gd-enhancing lesions without a trace of an accompanying

**2.1.1 MRI studies** 

enhancement- and in NAWM.

with p< .05 were considered as significant.

**2.1.2 Statistics** 

**3. Results** 

**GD enhancement** 

There was no statistically significant difference between the basal metabolite levels in the NAWM of patients with MS and the healthy control subjects (*P*> .005) (Table 1). There was also no statistically significant difference between the basal metabolite peaks in the WML between the untreated and treated MS patients (Table 2).


NAA: N-acetyl aspartate, Cho: Choline, Cr: Creatine

Table 1. Basal normal-appearing white matter and white matter metabolite levels in treated and untreated patients and controls

We analyzed whether any difference emerged in the NAWM metabolite peaks between the treated and untreated patients throughout the study period and found that a significant difference existed in NAWM NAA peaks (P = .003); Cho peaks (P = .001); and Cr peaks (P = .006) beginning from the first month of the study. When we evaluated the groups separately to assess when that difference emerged between the basal and the following months' NAA, Cho and Cr peaks we noted that in the treatment group, the NAA peaks began to increase during the sixth month (P = .000); Cho peaks increased during the sixth and twelfth months (P = .000); and an increment in the Cr peak was found in the third (P< .04), sixth (P = .001), and twelfth

MRS in MS, With and Without Interferon Beta 1a Treatment,

untreated group (*P* > .05) (Graphic 2).

**4. Discussion** 

concentrations.

to Define the Dynamic Changes of Metabolites in the Brain, and to Monitor Disease Progression 99

We found a significant difference between the changes in WML NAA (*P*= .001), Cho (*P* = .025); and Cr (*P* = .031) peaks of untreated and treated patients, beginning from the first month. When we evaluated the groups separately to determine when the difference emerged between the basal WML NAA peaks and the following months' NAA peaks, we noted that the NAA peaks started to increase during the sixth month in the treatment group (*P* = .001); the Cho peak started to increase during the first month (*P* = .03); and the Cr peak began to increase during the third month in the treatment group (*P* = .04). However, NAA, Cho or Cr peaks did not vary significantly in WML over the entire period of study in the

In this study, variations in cerebral metabolites were assessed during treatment with INF- 1a in a small group of patients with RRMS, and the results were compared with those in untreated patients matched for age and disability. We used the absolute quantification of the metabolites. The use of ratios is potentially less sensitive to the effects of disease in MS than is the use of absolute values; measurement error may be increased when two measures rather than one are required(Fernando et al., 2004). The use of the ratio to Cr also requires the assumption that Cr is normal in the patients' NAWM. However, there have been reports of abnormalities of Cr levels in demyelinating lesions and in the NAWM in patients with MS(Brex et al., 1999; Pan et al., 1996; Rooney et al., 1997; Suhy et al., 2000; van Walderveen, Truyen et al., 1999). Pan and colleagues found an isolated increase in Cr without a change in the level of Cho or NAA in the NAWM(Pan et al., 1996)of patients with MS. An increase in Cr was also found in a clinically isolated syndrome (CIS) group(Fernando et al., 2004). In that study, an increase in Cr was found in a subgroup of subjects with abnormal T2 weighted MRI results and in those with MS defined according to the McDonald criteria(Fernando et al., 2004). Those studies suggest that Cr may not be a reliable internal standard for use in the investigation of the effects of demylinating disease upon metabolite

We have found no significant difference between the basal metabolite peak areas of NAA, Cho and Cr in the NAWM of patients with MS and that of healthy controls (*P* >.005). This was attributed to the patients' having been on relatively early stages of the disease (primarily 2-5 years after diagnosis) and their low EDSS levels. Those factors indicate that widespread axonal damage either had not occurred or was not detectactable at this stage and that the reduced NAA found in established MS is acquired later in the disease course. In a study, when 15 clinically isolated syndrome patients with T2 abnormalities were considered separately, their NAWM NAA concentrations were not found to be significantly different from those in controls(Brex et al., 1999). In our study, there was no significant difference between NAWM metabolite peaks of the treated and untreated groups in the results of the initial MRSI (*P*> .005) showing that the 2 patient groups were homogenous. We found that in the treatment group, NAWM NAA and Cho peaks began to increase during the sixth month, and that Cr peaks began to increase during the third month. No change was observed in the untreated group. When we evaluated WML metabolite changes, we found that in the treatment group, NAA and Cr peaks began to increase during the third month and Cho peaks began to increase during the first month. No changes in the

metabolite peak areas of WML were observed in the untreated group.

study months (P = .00). However, in the untreated group, none of the NAA, Cho or Cr peaks (in NAWM) varied significantly during the entire study period (P> .05) (Graphic 1).

Graphic 1.

Graphic 2.

We found a significant difference between the changes in WML NAA (*P*= .001), Cho (*P* = .025); and Cr (*P* = .031) peaks of untreated and treated patients, beginning from the first month. When we evaluated the groups separately to determine when the difference emerged between the basal WML NAA peaks and the following months' NAA peaks, we noted that the NAA peaks started to increase during the sixth month in the treatment group (*P* = .001); the Cho peak started to increase during the first month (*P* = .03); and the Cr peak began to increase during the third month in the treatment group (*P* = .04). However, NAA, Cho or Cr peaks did not vary significantly in WML over the entire period of study in the untreated group (*P* > .05) (Graphic 2).

#### **4. Discussion**

98 Magnetic Resonance Spectroscopy

study months (P = .00). However, in the untreated group, none of the NAA, Cho or Cr peaks

(in NAWM) varied significantly during the entire study period (P> .05) (Graphic 1).

Graphic 1.

Graphic 2.

In this study, variations in cerebral metabolites were assessed during treatment with INF- 1a in a small group of patients with RRMS, and the results were compared with those in untreated patients matched for age and disability. We used the absolute quantification of the metabolites. The use of ratios is potentially less sensitive to the effects of disease in MS than is the use of absolute values; measurement error may be increased when two measures rather than one are required(Fernando et al., 2004). The use of the ratio to Cr also requires the assumption that Cr is normal in the patients' NAWM. However, there have been reports of abnormalities of Cr levels in demyelinating lesions and in the NAWM in patients with MS(Brex et al., 1999; Pan et al., 1996; Rooney et al., 1997; Suhy et al., 2000; van Walderveen, Truyen et al., 1999). Pan and colleagues found an isolated increase in Cr without a change in the level of Cho or NAA in the NAWM(Pan et al., 1996)of patients with MS. An increase in Cr was also found in a clinically isolated syndrome (CIS) group(Fernando et al., 2004). In that study, an increase in Cr was found in a subgroup of subjects with abnormal T2 weighted MRI results and in those with MS defined according to the McDonald criteria(Fernando et al., 2004). Those studies suggest that Cr may not be a reliable internal standard for use in the investigation of the effects of demylinating disease upon metabolite concentrations.

We have found no significant difference between the basal metabolite peak areas of NAA, Cho and Cr in the NAWM of patients with MS and that of healthy controls (*P* >.005). This was attributed to the patients' having been on relatively early stages of the disease (primarily 2-5 years after diagnosis) and their low EDSS levels. Those factors indicate that widespread axonal damage either had not occurred or was not detectactable at this stage and that the reduced NAA found in established MS is acquired later in the disease course. In a study, when 15 clinically isolated syndrome patients with T2 abnormalities were considered separately, their NAWM NAA concentrations were not found to be significantly different from those in controls(Brex et al., 1999). In our study, there was no significant difference between NAWM metabolite peaks of the treated and untreated groups in the results of the initial MRSI (*P*> .005) showing that the 2 patient groups were homogenous. We found that in the treatment group, NAWM NAA and Cho peaks began to increase during the sixth month, and that Cr peaks began to increase during the third month. No change was observed in the untreated group. When we evaluated WML metabolite changes, we found that in the treatment group, NAA and Cr peaks began to increase during the third month and Cho peaks began to increase during the first month. No changes in the metabolite peak areas of WML were observed in the untreated group.

MRS in MS, With and Without Interferon Beta 1a Treatment,

reactive features like those in the NAWM of patients with MS.

design because of ethical problems.

monitor the disease progression in MS.

**5. Conclusions** 

**6. References** 

81-91.

to Define the Dynamic Changes of Metabolites in the Brain, and to Monitor Disease Progression 101

Although an increase in the Cr level has been found in the NAWM of patients with MS(Brex et al., 1999; Rooney et al., 1997; Suhy et al., 2000; van Walderveen, Barkhof et al., 1999), some experts do not agree on the fluctuations of that metabolite in MS lesions. However, some authors have reported a decrease in total Cr(van Walderveen, Barkhof et al., 1999); others have reported an increase(Kapeller et al., 2001; Suhy et al., 2000). It has been suggested that apart from the axonal damage occurring in the NAWM of patients with MS, there is a proliferation of glial cells that produces an increase in the Cr level(Pan et al., 1996; van Walderveen, Barkhof et al., 1999). When working at 4.1tesla with 0.5 voxels, Pan and colleagues found an isolated increase in the Cr level without concomitant changes in Cho or NAA levels in the WM far from the WM lesions and suggested that in NAWM, there may have been reactive astrocytosis without inflammatory changes. It has been postulated that an increase in the Cr level may reflect an astrocytic (gliotic) or oligodendrocytic remyelinating process. Histological evidence confirms that gliosis is the most common abnormality in macroscopically normal WM of patients with MS. Thus the MRSI abnormalities observed in such patients may be due to the effect of microscopic lesions below the resolution of the images(Rooney et al., 1997; Suhy et al., 2000). However, the increase in Cr along with Cho levels in NAWM and WM may also be accounted for an increase in metabolism, which might be expected to occur with inflammatory or glial

There are some limitations of our study: The small number of patients, the short duration of study, an open label design, and the patients' being either in relatively early stages of disease or experiencing a relatively benign course of MS. However the numbers of patients in relatively early stages of the disease are more or less similar between the groups, and again the ones with a relatively benign form are again similarly distributed between the groups. Therefore we do not consider that this would have caused a bias in our results. It would have been better if the study were performed with a greater number of patients along with a longer study period. However we could not perform such a more desirable study

Our results showed that NAA, Cho and Cr in a VOI that predominantly contains NAWM increased in patients with MS after 1 year of treatment with INF-1a. The same effect was also observed in WML after treatment. This evidence supports the theory that although axonal pathology begins early in the course of MS, INF-1a therapy may reverse part of that early axonal and metabolic dysfunction, and thus inhibit permanent axonal loss. Follow-up studies could help to define the dynamic changes of metabolites in MS plaques and to

Allen, I. V., & McKeown, S. R. (1979). A histological, histochemical and biochemical study of

the macroscopically normal white matter in multiple sclerosis. *J Neurol Sci, 41*(1),

There are a number of possible explanations for the increase in NAA that occurred between the sixth and twelfth months of treatment. Chronic low-grade encephalitis is probably present in patients with MS. There is pathological evidence of an increase in inflammatory cells in the NAWM of such patients(Allen & McKeown, 1979). The results of studies using Gd-enhanced MRI suggest that the blood-brain barrier is diffusely impaired in patients with MS. This chronic, low-level inflammation could result in widespread, partially reversible axonal damage and loss, and INF could induce its reversal.

A number of tissue injury mediating factors have been implicated in the production of the symptoms of MS(Moreau et al., 1996). INF directly or indirectly inhibits both the production of a number of those associated inflammatory factors (including interferon gamma) and the ability of those factors to stimulate nitric oxide release from astrocytes(Stewart et al., 1997). Elevated nitric oxide production and the consequent damage to mitochondria in surrounding neurons may be mitigated by INF and this could be a mechanism of action of INF in patients with MS(Stewart, Land, Clark, & Heales, 1998). NAA is produced by neuronal mitochondria(Patel & Clark, 1979), this could explain the recovery of NAA in patients with MS who are treated with INF. The increase in NAA could also be associated with the reversal of axonal metabolic dysfunction that is associated with in apparent myelin pathology in the NAWM.

In this study, one of the principal effects, which seems to be due to INF-1a, was an increase in the total Cho peak, which in the NAWM became evident during the sixth month of treatment and persisted through the twelfth month, and in the WML became evident during the first month of treatment then persisted throughout the study period. Choline-containing compounds are considered to be turnover products of cell membranes, and an increase in the level of Cho seems to be related primarily to inflammatory and glial cell proliferation rather than to myelin destruction. This observation was confirmed in a histopathological and spectroscopic correlative study in EAE, the results of which showed that a high Cho was associated with inflammation and not with demyelination. Furthermore, glial proliferation in bioptic samples from patients with MS was consistent with an in vivo increase in Cho(Brenner et al., 1993). An increased level of Cho has been described in MS lesions in a variety of studies(Bitsch et al., 1999; Davie et al., 1994; Kapeller et al., 2001; Narayana et al., 1998). Myelin and membrane constituents make a major contribution to the constituent of Cho; therefore, increased lesion cellularity and turn over may account for the observed increase in the level of Cho. However, an increase in Cho has been interpreted as an index of active or recent demyelination by some authors(Confort-Gouny et al., 1993; Davie et al., 1994; Kapeller et al., 2001; Larsson et al., 1991), an association suggested by the finding of an abundance of Cho-containing compounds in myelin and in all cell membranes, including those of inflammatory cells, in patients with MS(Brenner et al., 1993). This increase in the level of Cho may occur without significant clinical deterioration or disease activity. In a recent study, an increase in metabolite concentrations of Cho and Cr was found in CIS NAWM(Fernando et al., 2004). H-MRS-visible Cho containing compounds are derived primarily from cell membranes and may be elevated when there is increased cell turnover. Because the relation between the variations in Cho levels and the pathological changes in underlying lesions is not very well known at the moment, it cannot be excluded that the rise in Cho peak could express an increased turnover of myelin in the plaques and NAWM examined, a turnover that may not be the result of new demyelination, but the consequence of the remodeling of plaques or the effects of a remyelination process.

Although an increase in the Cr level has been found in the NAWM of patients with MS(Brex et al., 1999; Rooney et al., 1997; Suhy et al., 2000; van Walderveen, Barkhof et al., 1999), some experts do not agree on the fluctuations of that metabolite in MS lesions. However, some authors have reported a decrease in total Cr(van Walderveen, Barkhof et al., 1999); others have reported an increase(Kapeller et al., 2001; Suhy et al., 2000). It has been suggested that apart from the axonal damage occurring in the NAWM of patients with MS, there is a proliferation of glial cells that produces an increase in the Cr level(Pan et al., 1996; van Walderveen, Barkhof et al., 1999). When working at 4.1tesla with 0.5 voxels, Pan and colleagues found an isolated increase in the Cr level without concomitant changes in Cho or NAA levels in the WM far from the WM lesions and suggested that in NAWM, there may have been reactive astrocytosis without inflammatory changes. It has been postulated that an increase in the Cr level may reflect an astrocytic (gliotic) or oligodendrocytic remyelinating process. Histological evidence confirms that gliosis is the most common abnormality in macroscopically normal WM of patients with MS. Thus the MRSI abnormalities observed in such patients may be due to the effect of microscopic lesions below the resolution of the images(Rooney et al., 1997; Suhy et al., 2000). However, the increase in Cr along with Cho levels in NAWM and WM may also be accounted for an increase in metabolism, which might be expected to occur with inflammatory or glial reactive features like those in the NAWM of patients with MS.

There are some limitations of our study: The small number of patients, the short duration of study, an open label design, and the patients' being either in relatively early stages of disease or experiencing a relatively benign course of MS. However the numbers of patients in relatively early stages of the disease are more or less similar between the groups, and again the ones with a relatively benign form are again similarly distributed between the groups. Therefore we do not consider that this would have caused a bias in our results. It would have been better if the study were performed with a greater number of patients along with a longer study period. However we could not perform such a more desirable study design because of ethical problems.

#### **5. Conclusions**

100 Magnetic Resonance Spectroscopy

There are a number of possible explanations for the increase in NAA that occurred between the sixth and twelfth months of treatment. Chronic low-grade encephalitis is probably present in patients with MS. There is pathological evidence of an increase in inflammatory cells in the NAWM of such patients(Allen & McKeown, 1979). The results of studies using Gd-enhanced MRI suggest that the blood-brain barrier is diffusely impaired in patients with MS. This chronic, low-level inflammation could result in widespread, partially reversible

A number of tissue injury mediating factors have been implicated in the production of the symptoms of MS(Moreau et al., 1996). INF directly or indirectly inhibits both the production of a number of those associated inflammatory factors (including interferon gamma) and the ability of those factors to stimulate nitric oxide release from astrocytes(Stewart et al., 1997). Elevated nitric oxide production and the consequent damage to mitochondria in surrounding neurons may be mitigated by INF and this could be a mechanism of action of INF in patients with MS(Stewart, Land, Clark, & Heales, 1998). NAA is produced by neuronal mitochondria(Patel & Clark, 1979), this could explain the recovery of NAA in patients with MS who are treated with INF. The increase in NAA could also be associated with the reversal of axonal metabolic dysfunction that is associated

In this study, one of the principal effects, which seems to be due to INF-1a, was an increase in the total Cho peak, which in the NAWM became evident during the sixth month of treatment and persisted through the twelfth month, and in the WML became evident during the first month of treatment then persisted throughout the study period. Choline-containing compounds are considered to be turnover products of cell membranes, and an increase in the level of Cho seems to be related primarily to inflammatory and glial cell proliferation rather than to myelin destruction. This observation was confirmed in a histopathological and spectroscopic correlative study in EAE, the results of which showed that a high Cho was associated with inflammation and not with demyelination. Furthermore, glial proliferation in bioptic samples from patients with MS was consistent with an in vivo increase in Cho(Brenner et al., 1993). An increased level of Cho has been described in MS lesions in a variety of studies(Bitsch et al., 1999; Davie et al., 1994; Kapeller et al., 2001; Narayana et al., 1998). Myelin and membrane constituents make a major contribution to the constituent of Cho; therefore, increased lesion cellularity and turn over may account for the observed increase in the level of Cho. However, an increase in Cho has been interpreted as an index of active or recent demyelination by some authors(Confort-Gouny et al., 1993; Davie et al., 1994; Kapeller et al., 2001; Larsson et al., 1991), an association suggested by the finding of an abundance of Cho-containing compounds in myelin and in all cell membranes, including those of inflammatory cells, in patients with MS(Brenner et al., 1993). This increase in the level of Cho may occur without significant clinical deterioration or disease activity. In a recent study, an increase in metabolite concentrations of Cho and Cr was found in CIS NAWM(Fernando et al., 2004). H-MRS-visible Cho containing compounds are derived primarily from cell membranes and may be elevated when there is increased cell turnover. Because the relation between the variations in Cho levels and the pathological changes in underlying lesions is not very well known at the moment, it cannot be excluded that the rise in Cho peak could express an increased turnover of myelin in the plaques and NAWM examined, a turnover that may not be the result of new demyelination, but the consequence

axonal damage and loss, and INF could induce its reversal.

with in apparent myelin pathology in the NAWM.

of the remodeling of plaques or the effects of a remyelination process.

Our results showed that NAA, Cho and Cr in a VOI that predominantly contains NAWM increased in patients with MS after 1 year of treatment with INF-1a. The same effect was also observed in WML after treatment. This evidence supports the theory that although axonal pathology begins early in the course of MS, INF-1a therapy may reverse part of that early axonal and metabolic dysfunction, and thus inhibit permanent axonal loss. Follow-up studies could help to define the dynamic changes of metabolites in MS plaques and to monitor the disease progression in MS.

#### **6. References**

Allen, I. V., & McKeown, S. R. (1979). A histological, histochemical and biochemical study of the macroscopically normal white matter in multiple sclerosis. *J Neurol Sci, 41*(1), 81-91.

MRS in MS, With and Without Interferon Beta 1a Treatment,

sclerosis. *Magn Reson Med, 22*(1), 23-31.

measurements. *Radiology, 196*(2), 511-515.

sclerosis. *Brain, 119 ( Pt 1)*, 225-237.

*Reson Med, 36*(1), 72-77.

*Brain, 122 ( Pt 3)*, 513-521.

*36*(2), 157-165.

127.

539-546.

*3*(4), 231-237.

to Define the Dynamic Changes of Metabolites in the Brain, and to Monitor Disease Progression 103

Gasperini, C., Horsfield, M. A., Thorpe, J. W., Kidd, D., Barker, G. J., Tofts, P. S., et al. (1996).

Kapeller, P., McLean, M. A., Griffin, C. M., Chard, D., Parker, G. J., Barker, G. J., et al. (2001).

Loevner, L. A., Grossman, R. I., Cohen, J. A., Lexa, F. J., Kessler, D., & Kolson, D. L. (1995).

Matthews, P. M., Pioro, E., Narayanan, S., De Stefano, N., Fu, L., Francis, G., et al. (1996).

Moreau, T., Coles, A., Wing, M., Isaacs, J., Hale, G., Waldmann, H., et al. (1996). Transient

Narayana, P. A., Doyle, T. J., Lai, D., & Wolinsky, J. S. (1998). Serial proton magnetic

Patel, T. B., & Clark, J. B. (1979). Synthesis of N-acetyl-L-aspartate by rat brain mitochondria

Rooney, W. D., Goodkin, D. E., Schuff, N., Meyerhoff, D. J., Norman, D., & Weiner, M. W.

Sarchielli, P., Presciutti, O., Pelliccioli, G. P., Tarducci, R., Gobbi, G., Chiarini, P., et al. (1999).

quantitative MR spectroscopic imaging study. *J Neurol, 248*(2), 131-138. Larsson, H. B., Christiansen, P., Jensen, M., Frederiksen, J., Heltberg, A., Olesen, J., et al.

Macroscopic and microscopic assessments of disease burden by MRI in multiple sclerosis: relationship to clinical parameters. *J Magn Reson Imaging, 6*(4), 580-584. Gonen, O., Catalaa, I., Babb, J. S., Ge, Y., Mannon, L. J., Kolson, D. L., et al. (2000). Total brain N-acetylaspartate: a new measure of disease load in MS. *Neurology, 54*(1), 15-19. Husted, C. A., Goodin, D. S., Hugg, J. W., Maudsley, A. A., Tsuruda, J. S., de Bie, S. H., et al.

(1994). Biochemical alterations in multiple sclerosis lesions and normal-appearing white matter detected by in vivo 31P and 1H spectroscopic imaging. *Ann Neurol,* 

Preliminary evidence for neuronal damage in cortical grey matter and normal appearing white matter in short duration relapsing-remitting multiple sclerosis: a

(1991). Localized in vivo proton spectroscopy in the brain of patients with multiple

Microscopic disease in normal-appearing white matter on conventional MR images in patients with multiple sclerosis: assessment with magnetization-transfer

Assessment of lesion pathology in multiple sclerosis using quantitative MRI morphometry and magnetic resonance spectroscopy. *Brain, 119 ( Pt 3)*, 715-722. McDonald, W. I., Compston, A., Edan, G., Goodkin, D., Hartung, H. P., Lublin, F. D., et al.

(2001). Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. *Ann Neurol, 50*(1), 121-

increase in symptoms associated with cytokine release in patients with multiple

resonance spectroscopic imaging, contrast-enhanced magnetic resonance imaging, and quantitative lesion volumetry in multiple sclerosis. *Ann Neurol, 43*(1), 56-71. Pan, J. W., Hetherington, H. P., Vaughan, J. T., Mitchell, G., Pohost, G. M., & Whitaker, J. N.

(1996). Evaluation of multiple sclerosis by 1H spectroscopic imaging at 4.1 T. *Magn* 

and its involvement in mitochondrial/cytosolic carbon transport. *Biochem J, 184*(3),

(1997). 1H MRSI of normal appearing white matter in multiple sclerosis. *Mult Scler,* 

Absolute quantification of brain metabolites by proton magnetic resonance spectroscopy in normal-appearing white matter of multiple sclerosis patients.


Armspach, J. P., Gounot, D., Rumbach, L., & Chambron, J. (1991). In vivo determination of

Arnold, D. L., Matthews, P. M., Francis, G. S., O'Connor, J., & Antel, J. P. (1992). Proton

Arnold, D. L., Riess, G. T., Matthews, P. M., Francis, G. S., Collins, D. L., Wolfson, C., et al.

Bitsch, A., Bruhn, H., Vougioukas, V., Stringaris, A., Lassmann, H., Frahm, J., et al. (1999).

Brex, P. A., Gomez-Anson, B., Parker, G. J., Molyneux, P. D., Miszkiel, K. A., Barker, G. J., et

Confort-Gouny, S., Vion-Dury, J., Nicoli, F., Dano, P., Donnet, A., Grazziani, N., et al. (1993).

Davie, C. A., Barker, G. J., Thompson, A. J., Tofts, P. S., McDonald, W. I., & Miller, D. H.

Davie, C. A., Hawkins, C. P., Barker, G. J., Brennan, A., Tofts, P. S., Miller, D. H., et al. (1994).

Davies, S. E., Newcombe, J., Williams, S. R., McDonald, W. I., & Clark, J. B. (1995). High

De Stefano, N., Matthews, P. M., Antel, J. P., Preul, M., Francis, G., & Arnold, D. L. (1995).

Ferguson, B., Matyszak, M. K., Esiri, M. M., & Perry, V. H. (1997). Axonal damage in acute

Fernando, K. T., McLean, M. A., Chard, D. T., MacManus, D. G., Dalton, C. M., Miszkiel, K.

Filippi, M., Campi, A., Dousset, V., Baratti, C., Martinelli, V., Canal, N., et al. (1995). A

Fu, L., Matthews, P. M., De Stefano, N., Worsley, K. J., Narayanan, S., Francis, G. S., et al.

*Magn Reson Imaging, 9*(1), 107-113.

demyelinating plaques. *Ann Neurol, 31*(3), 235-241.

the Cho:Cr ratio. *Magn Reson Med, 29*(6), 737-745.

multiple sclerosis. *J Neurol Sci, 166*(1), 16-22.

*J Neurol Sci, 118*(2), 123-133.

*Psychiatry, 63*(6), 736-742.

*Brain, 117 ( Pt 1)*, 49-58.

disability. *Ann Neurol, 38*(6), 901-909.

sclerosis. *Brain, 121 ( Pt 1)*, 103-113.

multiple sclerosis lesions. *Brain, 120 ( Pt 3)*, 393-399.

multiple sclerosis. *Neurology, 45*(3 Pt 1), 478-482.

suggestive of multiple sclerosis. *Brain, 127*(Pt 6), 1361-1369.

*64*(2), 742-748.

progression in multiple sclerosis. *Ann Neurol, 36*(1), 76-82.

multiexponential T2 relaxation in the brain of patients with multiple sclerosis.

magnetic resonance spectroscopic imaging for metabolic characterization of

(1994). Use of proton magnetic resonance spectroscopy for monitoring disease

Inflammatory CNS demyelination: histopathologic correlation with in vivo quantitative proton MR spectroscopy. *AJNR Am J Neuroradiol, 20*(9), 1619-1627. Brenner, R. E., Munro, P. M., Williams, S. C., Bell, J. D., Barker, G. J., Hawkins, C. P., et al.

(1993). The proton NMR spectrum in acute EAE: the significance of the change in

al. (1999). Proton MR spectroscopy in clinically isolated syndromes suggestive of

A multiparametric data analysis showing the potential of localized proton MR spectroscopy of the brain in the metabolic characterization of neurological diseases.

(1997). 1H magnetic resonance spectroscopy of chronic cerebral white matter lesions and normal appearing white matter in multiple sclerosis. *J Neurol Neurosurg* 

Serial proton magnetic resonance spectroscopy in acute multiple sclerosis lesions.

resolution proton NMR spectroscopy of multiple sclerosis lesions. *J Neurochem,* 

Chemical pathology of acute demyelinating lesions and its correlation with

A., et al. (2004). Elevated white matter myo-inositol in clinically isolated syndromes

magnetization transfer imaging study of normal-appearing white matter in

(1998). Imaging axonal damage of normal-appearing white matter in multiple


**6** 

**Magnetic Resonance Spectroscopy** 

**(MRS) in Kidney Transplantation:** 

Bon Delphine, Seguin François and Hauet Thierry

*CHU Poitiers, Pole UBM, Service de Biochimie, Poitiers IBISA, Domaine Expérimental du Magneraud, Surgères* 

*Université de Poitiers, Faculté de Médecine et de Pharmacie, Poitiers* 

Currently, in biology, Magnetic Resonance Spectroscopy (MRS) is widely used in metabonomics for diagnosis and prognostication in a wide range of studies from brain (Blasco et al., 2010) to leg (Borel et al., 2009) pathologies. Between both, the topic of interest in this section: the kidney. We will especially focus on kidney transplantation. Indeed, in France as well as in Europe, the number of patients awaiting transplant is still rising while the number of transplantations performed remains stable and even decreases (Figure 1.).

Patient awaiting transplant and transplants performed between 2003 and 2010 in France Awaiting transplant Transplant performed

2003 2004 2005 2006 2007 2008 2009 2010

Fig. 1. Evolution of the number of patients on the waiting list for transplant and the number

of transplantations performed between 2003 and 2010 in France (from Agence de la

**1. Introduction** 

0

Biomédecine Rapport annuel 2011).

**Interest and Perspectives** 

*Inserm U927, Poitiers* 

*France* 


### **Magnetic Resonance Spectroscopy (MRS) in Kidney Transplantation: Interest and Perspectives**

Bon Delphine, Seguin François and Hauet Thierry *Inserm U927, Poitiers Université de Poitiers, Faculté de Médecine et de Pharmacie, Poitiers CHU Poitiers, Pole UBM, Service de Biochimie, Poitiers IBISA, Domaine Expérimental du Magneraud, Surgères France* 

#### **1. Introduction**

104 Magnetic Resonance Spectroscopy

Stewart, V. C., Giovannoni, G., Land, J. M., McDonald, W. I., Clark, J. B., & Heales, S. J.

Suhy, J., Rooney, W. D., Goodkin, D. E., Capizzano, A. A., Soher, B. J., Maudsley, A. A., et al.

Tourbah, A., Stievenart, J. L., Gout, O., Fontaine, B., Liblau, R., Lubetzki, C., et al. (1999).

van Walderveen, M. A., Truyen, L., van Oosten, B. W., Castelijns, J. A., Lycklama a Nijeholt,

secondary progressive multiple sclerosis. *Neurology, 53*(5), 1091-1097. van Walderveen, M. A., Barkhof, F., Pouwels, P. J., van Schijndel, R. A., Polman, C. H., &

and relapsing-remitting MS. *Mult Scler, 6*(3), 148-155.

inflammatory activity. *Arch Neurol, 56*(3), 345-351.

gamma induction of nitric oxide synthase. *J Neurochem, 68*(6), 2547-2551. Stewart, V. C., Land, J. M., Clark, J. B., & Heales, S. J. (1998). Pretreatment of astrocytes with

*Neurochem, 70*(1), 432-434.

*Neurol, 46*(1), 79-87.

(1997). Pretreatment of astrocytes with interferon-alpha/beta impairs interferon-

interferon-alpha/beta prevents neuronal mitochondrial respiratory chain damage. *J* 

(2000). 1H MRSI comparison of white matter and lesions in primary progressive

Localized proton magnetic resonance spectroscopy in relapsing remitting versus

Castelijns, J. A. (1999). Neuronal damage in T1-hypointense multiple sclerosis lesions demonstrated in vivo using proton magnetic resonance spectroscopy. *Ann* 

G. J., van Waesberghe, J. H., et al. (1999). Development of hypointense lesions on T1-weighted spin-echo magnetic resonance images in multiple sclerosis: relation to

> Currently, in biology, Magnetic Resonance Spectroscopy (MRS) is widely used in metabonomics for diagnosis and prognostication in a wide range of studies from brain (Blasco et al., 2010) to leg (Borel et al., 2009) pathologies. Between both, the topic of interest in this section: the kidney. We will especially focus on kidney transplantation. Indeed, in France as well as in Europe, the number of patients awaiting transplant is still rising while the number of transplantations performed remains stable and even decreases (Figure 1.).

Patient awaiting transplant and transplants performed between 2003 and 2010 in France

Fig. 1. Evolution of the number of patients on the waiting list for transplant and the number of transplantations performed between 2003 and 2010 in France (from Agence de la Biomédecine Rapport annuel 2011).

Magnetic Resonance Spectroscopy (MRS) in Kidney Transplantation: Interest and Perspectives 107

diabetic nephropathy induced by streptozotocin, the authors performed a holistic metabolic analysis in order to elucidate the mechanism of this disease. For that purpose, they realised 1H liquid NMR spectra of urine on one hand and kidney extracts on the other. Elevated level of glucose in diabetic rats affected the ketone pathway, fatty acid oxidation, tricarboxylic acid cycle and glycolosis. Those metabolic changes led to decrease of energy production, hence aggravating kidney damage. Moreover, in this study the authors assessed osmolyte

Another cause of kidney failure is autosomal dominant polycystic kidney diseases (ADPKD). Some candidate urinary protein biomarkers such as KIM1 or NGAL have already been evaluated for this nephropathology. However, to improve the quality of diagnosis for ADPKD, complete 1H and 13C NMR fingerprint of urine were put in place (Gronwald et al., 2011). Elevated levels of proteins and methanol were found in patients with ADPKD

NMR is well suited for the study of neprotoxicity. In 1998, the team of Nicholson reported urine metabolic profiles in 15 groups of animal treated with nephrotoxic molecules and compared spectra with control. They thus highlighted endogenous biomarkers of nephrotoxic insult (Holmes et al., 1998). Citrate, 2-oxoglutarate and hippurate were the most discriminating metabolites between treated group and control group whereas formate and dimethylamine showed almost not change. Other studies on administration of nephrotoxic compounds realised with urine and plasma of Sprague-Dawley rats focused on lanthanum, a rare earth compound used as fertilizer with a higher accumulation rate and a lower metabolic rate (Feng et al., 2002) or thioacetamide (Waters et al., 2005). In the first study, the authors showed that a 6-months long ingestion of La3+-induced nephro- and hepato-toxicity which could be highlighting by using NMR. From a dose of 10 mg/kg of La(NO3)3+, all the evolving quantified metabolites involved Krebs cycle intermediates or amino acids that are potential NMR markers for La3+-induced proximal tubular lesions. Moreover, decrease of creatinine in urine highlighted a low glomerular filtration. A second study focused on energy intermediairy (citrate, lactate, succinate) and lipid metabolisms. The work also realised on intact kidney tissue using hrmas NMR spectroscopy (Garrod et al., 2001; Wang et al., 2006) established changes in the spectral profile of renal papilla involved the marked depletion of several renal

Finally, the effects of cyclosporine A, the basis for most immunosuppressive protocols and used for allograft recipients on kidneys were evaluated (Lenz et al., 2004; Serkova et al., 2003). This treatment is well known for its nephrotoxicity. Cyclosporine A induced a decrease of poly unsaturated fatty acid and an increase of lipid peroxidation which suggests the use of an alternative pathway for energy production when the oxidative mitochondrial pathway is inhibited. This lipid peroxidation is responsive for kidney damage especially in the cortex and medulla. Cyclosporine A also leads to osmolyte regulation (taurine, betaine

The isolated perfused pig kidney (IPK) was used to mimic the ischemia reperfusion episode occurring during transplantation. This model was destined to assess initial renal function

osmolytes such as glycerophosphocholine, betaine, and myo-inositol.

metabolism.

receiving medication against hypertension.

and trimethylamine-N-oxide (TMAO)).

**2.2 Kidney transplantation analysis** 

**2.2.1 Isolated Perfused Kidney (IPK) of pig** 

Although campaigns for organ donation increase the number of donors, donated organs are most often in marginal conditions because the donors often present co-morbidity factors (diabetes, obesity, cardiovascular disease, hypertension, poor renal function). In the case of the kidney, the percentage of successful grafts tends to decrease. These factors weaken the organ and render it more susceptible to develop lesions following ischemia reperfusion, the unavoidable syndrome encountered during organ preservation. As the extend of these lesions is correlated to graft outcome, it is necessary to improve the entire sequence of ischemia-reperfusion and in particular to find ways to improve storage conditions and decrease the rate of delayed graft function after kidney transplantation.

In kidney transplantation, delayed graft function is mainly due to the intensity of ischemia reperfusion lesions associated with the hypothermic and hypoxic conditions during preservation. These lesions are exacerbated at the moment of transplantation with reoxygenation that occurs during the reperfusion. In this light, tissue damages identified by the detection of endogenous metabolic changes in plasma and urine would permit a better estimation of graft quality and improved management of the graft, to the direct benefit of the patient. High Resolution Magnetic Resonance Spectroscopy (HR MRS) in these fluids could permit the identification of biomarkers indicative of specific kidney lesions occurring during an ischemia-reperfusion sequence.

The purpose of the chapter is to make a mini-review about thematics involving kidney transplantation from animals, mainly large white pig, to humans in order to highlight the potential of MRS, high resolution or imaging, and what perspectives could be envisioned to improve comprehension of kidney transplantation.

#### **2. High resolution NMR**

#### **2.1 Kidney analysis**

Before considering transplantation, it is important to review nephrologic situations and nephropathologies that have been widely analyzed by MRS to find biomarkers of pathologies associated to kidney.

In 1999, Garrod and al. established the biochemical composition of normal inner and outer renal cortex and renal papilla from rats (Garrod et al., 1999). This work realised with 1H liquid NMR spectroscopy on tissue extracts and with 1H hrmas (high resolution magic angle spining) NMR spectroscopy on crude tissue allowed the direct observation of metabolites into kidney and showed the metabolic composition of the 3 main areas of the kidney. They established the 1H (hydrogen-1) and 13C (carbon-13) chemical shift assignments of metabolites found in cortex and/or papilla. Still with 1H hrmas NMR spectroscopy, a team realised the metabolic profiling of normal and hypertensive rat kidney (Huhn et al., 2004). They showed that better differentiation between both groups occurred in the kidney cortex. It has also been shown that NMR can be used to compare intact kidney, blood plasma and urine of 4 species of rodent. They thus showed that metabolic data acquired on laboratory animals can only be extended to wild species with the use of a lot of precaution (Griffin et al., 2000).

On those bases, and among pathologies affecting the kidney, diabetic nephropathy is one of the lethal manifestations of diabetic systemic disease (Zhao et al., 2011). Using a rat model of diabetic nephropathy induced by streptozotocin, the authors performed a holistic metabolic analysis in order to elucidate the mechanism of this disease. For that purpose, they realised 1H liquid NMR spectra of urine on one hand and kidney extracts on the other. Elevated level of glucose in diabetic rats affected the ketone pathway, fatty acid oxidation, tricarboxylic acid cycle and glycolosis. Those metabolic changes led to decrease of energy production, hence aggravating kidney damage. Moreover, in this study the authors assessed osmolyte metabolism.

Another cause of kidney failure is autosomal dominant polycystic kidney diseases (ADPKD). Some candidate urinary protein biomarkers such as KIM1 or NGAL have already been evaluated for this nephropathology. However, to improve the quality of diagnosis for ADPKD, complete 1H and 13C NMR fingerprint of urine were put in place (Gronwald et al., 2011). Elevated levels of proteins and methanol were found in patients with ADPKD receiving medication against hypertension.

NMR is well suited for the study of neprotoxicity. In 1998, the team of Nicholson reported urine metabolic profiles in 15 groups of animal treated with nephrotoxic molecules and compared spectra with control. They thus highlighted endogenous biomarkers of nephrotoxic insult (Holmes et al., 1998). Citrate, 2-oxoglutarate and hippurate were the most discriminating metabolites between treated group and control group whereas formate and dimethylamine showed almost not change. Other studies on administration of nephrotoxic compounds realised with urine and plasma of Sprague-Dawley rats focused on lanthanum, a rare earth compound used as fertilizer with a higher accumulation rate and a lower metabolic rate (Feng et al., 2002) or thioacetamide (Waters et al., 2005). In the first study, the authors showed that a 6-months long ingestion of La3+-induced nephro- and hepato-toxicity which could be highlighting by using NMR. From a dose of 10 mg/kg of La(NO3)3+, all the evolving quantified metabolites involved Krebs cycle intermediates or amino acids that are potential NMR markers for La3+-induced proximal tubular lesions. Moreover, decrease of creatinine in urine highlighted a low glomerular filtration. A second study focused on energy intermediairy (citrate, lactate, succinate) and lipid metabolisms. The work also realised on intact kidney tissue using hrmas NMR spectroscopy (Garrod et al., 2001; Wang et al., 2006) established changes in the spectral profile of renal papilla involved the marked depletion of several renal osmolytes such as glycerophosphocholine, betaine, and myo-inositol.

Finally, the effects of cyclosporine A, the basis for most immunosuppressive protocols and used for allograft recipients on kidneys were evaluated (Lenz et al., 2004; Serkova et al., 2003). This treatment is well known for its nephrotoxicity. Cyclosporine A induced a decrease of poly unsaturated fatty acid and an increase of lipid peroxidation which suggests the use of an alternative pathway for energy production when the oxidative mitochondrial pathway is inhibited. This lipid peroxidation is responsive for kidney damage especially in the cortex and medulla. Cyclosporine A also leads to osmolyte regulation (taurine, betaine and trimethylamine-N-oxide (TMAO)).

#### **2.2 Kidney transplantation analysis**

106 Magnetic Resonance Spectroscopy

Although campaigns for organ donation increase the number of donors, donated organs are most often in marginal conditions because the donors often present co-morbidity factors (diabetes, obesity, cardiovascular disease, hypertension, poor renal function). In the case of the kidney, the percentage of successful grafts tends to decrease. These factors weaken the organ and render it more susceptible to develop lesions following ischemia reperfusion, the unavoidable syndrome encountered during organ preservation. As the extend of these lesions is correlated to graft outcome, it is necessary to improve the entire sequence of ischemia-reperfusion and in particular to find ways to improve storage conditions and

In kidney transplantation, delayed graft function is mainly due to the intensity of ischemia reperfusion lesions associated with the hypothermic and hypoxic conditions during preservation. These lesions are exacerbated at the moment of transplantation with reoxygenation that occurs during the reperfusion. In this light, tissue damages identified by the detection of endogenous metabolic changes in plasma and urine would permit a better estimation of graft quality and improved management of the graft, to the direct benefit of the patient. High Resolution Magnetic Resonance Spectroscopy (HR MRS) in these fluids could permit the identification of biomarkers indicative of specific kidney lesions occurring

The purpose of the chapter is to make a mini-review about thematics involving kidney transplantation from animals, mainly large white pig, to humans in order to highlight the potential of MRS, high resolution or imaging, and what perspectives could be envisioned to

Before considering transplantation, it is important to review nephrologic situations and nephropathologies that have been widely analyzed by MRS to find biomarkers of

In 1999, Garrod and al. established the biochemical composition of normal inner and outer renal cortex and renal papilla from rats (Garrod et al., 1999). This work realised with 1H liquid NMR spectroscopy on tissue extracts and with 1H hrmas (high resolution magic angle spining) NMR spectroscopy on crude tissue allowed the direct observation of metabolites into kidney and showed the metabolic composition of the 3 main areas of the kidney. They established the 1H (hydrogen-1) and 13C (carbon-13) chemical shift assignments of metabolites found in cortex and/or papilla. Still with 1H hrmas NMR spectroscopy, a team realised the metabolic profiling of normal and hypertensive rat kidney (Huhn et al., 2004). They showed that better differentiation between both groups occurred in the kidney cortex. It has also been shown that NMR can be used to compare intact kidney, blood plasma and urine of 4 species of rodent. They thus showed that metabolic data acquired on laboratory animals can only be extended to wild species with the use of a lot

On those bases, and among pathologies affecting the kidney, diabetic nephropathy is one of the lethal manifestations of diabetic systemic disease (Zhao et al., 2011). Using a rat model of

decrease the rate of delayed graft function after kidney transplantation.

during an ischemia-reperfusion sequence.

**2. High resolution NMR** 

pathologies associated to kidney.

of precaution (Griffin et al., 2000).

**2.1 Kidney analysis** 

improve comprehension of kidney transplantation.

#### **2.2.1 Isolated Perfused Kidney (IPK) of pig**

The isolated perfused pig kidney (IPK) was used to mimic the ischemia reperfusion episode occurring during transplantation. This model was destined to assess initial renal function

Magnetic Resonance Spectroscopy (MRS) in Kidney Transplantation: Interest and Perspectives 109

glomerular filtration rate (GFR), tubular reabsortion of sodium and lactate dehydrogenase

The IPK model, which provided ex vivo information, was converted to a preclinical model with 3-months old large white pig. Briefly, the basic model was to remove the kidney, preserve it 24 hours and then transplant it in the same animal with a contralateral nephrectomy. We followed the animal for 3 months after transplantation and we studied

In this model, damage biomarkers appearing after an ischemia reperfusion episode were analyzed in order to confirm their pertinence with the follow up of the recovery of function. To summarize, metabolites quantified with liquid NMR were TMAO, correlated to medullar lesions, acetate correlated to cortical lesions and lactate linked to global ischemia. Amino acids such as alanine or valine could be associated with proximal tubules dysfunction. Creatinine measured in both plasma and urine allowed the determination of GFR. All these

Fig. 3. Levels of trimethylamine-N-oxide (A) and dimethylamine (B) in urine and changes of trimethylamine-N-oxide in plasma (C). Levels of citrate (D), acetate (E), and lactate (F) excretion in urine after autotransplantation. °P , 0.05, °° P , 0.01 (Hauet et al., 2000b).

levels (Hauet et al., 2000d).

**2.2.2 In vivo autotransplantation pig model** 

biochemical parameter in order to correlate them with MRS data.

metabolites were highlighted in different studies summarized below.

after different preservation condition. The schematic diagram of the isolated perfusion method is illustrated in the Figure 2. Based on an isolated perfused kidney, HR MRS injury biomarkers of an ischemia-reperfusion sequence were identified in the urine. Spectra modifications compared to histological or biochemical analysis showed changes of some metabolites excretion.

Fig. 2. Schematic diagram illustrating the isolated perfusion method in which A is the peristaltic pump, B, the oxygenator, C, the filter, D, the pressure transducer, E, the flow meter and electromagnetic flow probe (Hauet et al., 2000d).

One of the first studies using HR MRS on this model was performed to mimic the non-heartbeating donor situation. The kidney underwent a warm ischemic period before the hypothermic period of conservation and reperfusion (Hauet et al., 1997b). In this study, 3 groups of pigs were used. In the control group, kidneys were flushed with cold heparinized saline and immediately perfused. Kidney of a second group underwent a cold flush and 24 hours of cold-storage preservation (CSP) and finally reperfusion. The last group was composed of kidneys with 30 minutes of warm ischemia, 24 hours of CSP and reperfusion. The reperfusion occurred at 37°C and was performed with Kreb's solution supplemented with 22 amino acids. Among the interesting metabolites, TMAO and lactate were higher in the group with warm ischemia and related to damage of medullar cells and to tubular dysfunction respectively. Another study performed with the IPK was performed to assess the effect of cold storage conservation time with EuroCollins solution (EC) (Hauet et al., 1997a). Perfusion time was pushed to 72 hours and the degree of proximal tubule cell damage was increased with prolonged cold ischemia (Goujon et al., 1999). Urine content after a delay of 24 hours of perfusion was not significantly different from urine content at the beginning of the perfusion. After 48 hours, urine profile was significantly different. Rises of TMAO/creatinine and lactate/creatinine ratio measured with NMR were in accordance with elevated levels of beta-N-acetylglucosaminidase and lactate dehydrogenase. The same protocol was used to compare EC to University of Wisconsin solution (UW) (Hauet et al., 2000c; Hauet et al., 1999). Metabolites released into the conservation solution analyzed with NMR also confirmed classical biochemical parameters such as perfusion flow rate, glomerular filtration rate (GFR), tubular reabsortion of sodium and lactate dehydrogenase levels (Hauet et al., 2000d).

#### **2.2.2 In vivo autotransplantation pig model**

108 Magnetic Resonance Spectroscopy

after different preservation condition. The schematic diagram of the isolated perfusion method is illustrated in the Figure 2. Based on an isolated perfused kidney, HR MRS injury biomarkers of an ischemia-reperfusion sequence were identified in the urine. Spectra modifications compared to histological or biochemical analysis showed changes of some

Fig. 2. Schematic diagram illustrating the isolated perfusion method in which A is the peristaltic pump, B, the oxygenator, C, the filter, D, the pressure transducer, E, the flow

One of the first studies using HR MRS on this model was performed to mimic the non-heartbeating donor situation. The kidney underwent a warm ischemic period before the hypothermic period of conservation and reperfusion (Hauet et al., 1997b). In this study, 3 groups of pigs were used. In the control group, kidneys were flushed with cold heparinized saline and immediately perfused. Kidney of a second group underwent a cold flush and 24 hours of cold-storage preservation (CSP) and finally reperfusion. The last group was composed of kidneys with 30 minutes of warm ischemia, 24 hours of CSP and reperfusion. The reperfusion occurred at 37°C and was performed with Kreb's solution supplemented with 22 amino acids. Among the interesting metabolites, TMAO and lactate were higher in the group with warm ischemia and related to damage of medullar cells and to tubular dysfunction respectively. Another study performed with the IPK was performed to assess the effect of cold storage conservation time with EuroCollins solution (EC) (Hauet et al., 1997a). Perfusion time was pushed to 72 hours and the degree of proximal tubule cell damage was increased with prolonged cold ischemia (Goujon et al., 1999). Urine content after a delay of 24 hours of perfusion was not significantly different from urine content at the beginning of the perfusion. After 48 hours, urine profile was significantly different. Rises of TMAO/creatinine and lactate/creatinine ratio measured with NMR were in accordance with elevated levels of beta-N-acetylglucosaminidase and lactate dehydrogenase. The same protocol was used to compare EC to University of Wisconsin solution (UW) (Hauet et al., 2000c; Hauet et al., 1999). Metabolites released into the conservation solution analyzed with NMR also confirmed classical biochemical parameters such as perfusion flow rate,

meter and electromagnetic flow probe (Hauet et al., 2000d).

metabolites excretion.

The IPK model, which provided ex vivo information, was converted to a preclinical model with 3-months old large white pig. Briefly, the basic model was to remove the kidney, preserve it 24 hours and then transplant it in the same animal with a contralateral nephrectomy. We followed the animal for 3 months after transplantation and we studied biochemical parameter in order to correlate them with MRS data.

In this model, damage biomarkers appearing after an ischemia reperfusion episode were analyzed in order to confirm their pertinence with the follow up of the recovery of function. To summarize, metabolites quantified with liquid NMR were TMAO, correlated to medullar lesions, acetate correlated to cortical lesions and lactate linked to global ischemia. Amino acids such as alanine or valine could be associated with proximal tubules dysfunction. Creatinine measured in both plasma and urine allowed the determination of GFR. All these metabolites were highlighted in different studies summarized below.

Fig. 3. Levels of trimethylamine-N-oxide (A) and dimethylamine (B) in urine and changes of trimethylamine-N-oxide in plasma (C). Levels of citrate (D), acetate (E), and lactate (F) excretion in urine after autotransplantation. °P , 0.05, °° P , 0.01 (Hauet et al., 2000b).

Magnetic Resonance Spectroscopy (MRS) in Kidney Transplantation: Interest and Perspectives 111

with perchloric acid of kidney tissue samples 24 hours after transplantation were performed and hydrosoluble metabolites were analyzed with 1H liquid NMR spectroscopy. In parallel, blood samples were analyzed. TMAO again gave the best results. Its decreased levels in the blood of NAC group compared to control group were associated with a decrease of allantoin in the blood and kidney tissue. This latter metabolite results from the free radical action of urate and its presence indicates free radical activities. Pretreatment with NAC reduced radical excretion, hence reduced free radical activity and thus tissue damages (Fuller et al., 2004).

Use of HR MRS in Human to follow transplanted kidney recovery of function highlighted several issues. The main problem came from the drug treatment received by patients after transplantation. We observed overlapping between the signals of the metabolites of interest and of the drugs. In these conditions it is often difficult to extract information about the kidney function. Moreover a high variability, in addition of the individual variability, is often observed. This variability, mainly due to diet**,** may cause urine profile modifications which can hide variations of the metabolite of interest in relation with the kidney injuries. One solution could be to introduce a 2-days controlled diet before each sampling to minimise those environmental variation. However, an older study performed on 33 patients who underwent primary renal allograft transplantation, using the same immunosuppressive regiment and of which 57.6% achieve immediate graft function, showed that patients with good graft function and normal patients had urinary TMAO/creatinine level inferior to 200 µmol/mM while patients with graft dysfunction had urinary TMAO/creatinine level of 410

Hypothermic machine perfusion (MP) is increasingly being preferred to CSP. Indeed, MP has been showed to offer protection to the organ since the 80's (Alijani et al., 1985; Kwiatkowski et al., 2007; Wight et al., 2003) with benefits such as reduction of delayed graft function and increased survival. However, the mechanisms involved in these improvements are still unclear and need to be elucidated. Biomarkers such as lactate dehydrogenase, total glutathione-S-transferase or N-acetyl--D-glucosaminidase measured at the end of the perfusion in the machine perfusates are used by clinicians. Even if those markers are independently efficient to predict graft outcome, they are not efficient enough to be taken into account in the decision to transplant or discard a donation after brain death or controlled donation after cardiac death (Moers et al., 2010; Siew et al., 2011). Again the potential of multivariate analysis offered by NMR could be used to predict graft outcome. In a recent study in a pig model of autotransplantation of livers, the effect of warm ischemia on

Currently, our team is attempting to answer the question: can 1H HR NMR metabonomic analysis during machine perfusion be used to predict graft outcome? We thus work on perfusates from different protocols of kidney transplantation using MP. We evaluate the effect of warm ischemia, duration of ischemia, types of solution and temperatures of preservation and finally effect of adjuvant in the perfusion solution. We showed, in a small

metabolites contents into perfusats was measured (Liu et al., 2009).

**2.2.3 Application to Human** 

± 102 µmol/mM (Foxall et al., 1993).

**2.3 Perspectives** 

**2.3.1 Perfusate analysis** 

In 2000, comparison of EC and UW solution with CSP, already performed on IPK of pig, was performed on this preclinical model of kidney autotransplantation (Hauet et al., 2000b). The control group, undergoing uninephrectomy, presented 100% survival at day 14 whereas only 75% and 66 % of UW and EC pigs survived, respectively. Glutathione-S-transferase and creatinine clearance were significantly different between EC and UW groups with a light advantages for UW compared to control. Results obtained with MRS on urine and plasma are presented in figure 3. Delayed graft function was associated with the release of osmolytes such as TMAO in urine and in plasma. Moreover, dimethylamine, acetate, citrate and lactate level in the UW group were always closer to control group levels compared to levels in the EC group. These data outline the effect of UW preservation against renal medullar injury and impairment of oxidative metabolism. Here, strength of MRS was highlighted: in one experiment, with a 200 to 500 µL volume of sample, detection and quantitation of large number of metabolites can be made. This allows for a multivariable analysis which is an important factor for diagnostic and prognostic purpose.

Using the same experimental design, the impact of the preservation solution was demonstrated and the addition of colloids (polyethylene glycol (PEG)) into an extracellular solution showed better results (Faure et al., 2002). Indeed, in the urine spectra, TMAO/creatine ratio were significantly higher in kidney preserved with EC, UW or a solution containing PEG-50 than in kidney preserved with a solution containing PEG-30. This demonstrated that PEG reduced damages to the renal medulla during conservation, in a dose-dependent manner. Citrate was not detected during the first postoperative week in urine from kidneys preserved in EC. Its excretion was detected significantly earlier in urine from the ICPEG30 group than from the UW and ICPEG50 groups. This suggested that PEG associated to an intracellular solution improved oxidative metabolism and reduced renal medulla injury, also in a concentration-dependent manner. Furthermore, the addition of antioxidative (Baumert et al., 1999) or anticoagulant (Favreau et al., 2010) molecules, trimetazidine and melagatran respectively, into the preservation solution increased the performance of the hypothermic conservation. The effect of trimetazidine during conservation with EC or UW showed interesting results. The early excretion of acetate in the trimetazidine supplemented group demonstrated an efficient function recovery of the citric acid cycle compared with standard solutions (Hauet et al., 2000a). Combination of PEG and trimetazidine in conservation solution was the next step of long term time study which included 18 groups also exploring cold ischemia time (Doucet et al., 2004). Biological parameters indicated that reduced delayed graft was found when combining PEG and trimetazidine during CSP. Histological staining for CD4+ positive cells indicated that PEG reduced CD4+ positive cells infiltration and trimetazidine was efficient in reducing the inflammatory reaction. The NMR analysis was again in accordance with these results. Acetate excretion between day 1 and day 14 was improved in all experimental group preserved with trimetazidine while urinary TMAO was reduced in all trimetazidine group and we noticed presence of TMAO in plasma in all group except from control. Combination of trimetazidine with PEG seemed to be the best option for kidney conservation and MRS brought supplementary information to classical biochemical parameters such as GFR or sodium excretion. In addition, this study underlined that MRS permits the performance of a multivariate analysis in a single experiment.

Use of conditioning drugs before conservation was also investigated. Donor pretreatment with N-acetylcysteine (NAC) has been shown to ameliorate acute renal failure. Extractions with perchloric acid of kidney tissue samples 24 hours after transplantation were performed and hydrosoluble metabolites were analyzed with 1H liquid NMR spectroscopy. In parallel, blood samples were analyzed. TMAO again gave the best results. Its decreased levels in the blood of NAC group compared to control group were associated with a decrease of allantoin in the blood and kidney tissue. This latter metabolite results from the free radical action of urate and its presence indicates free radical activities. Pretreatment with NAC reduced radical excretion, hence reduced free radical activity and thus tissue damages (Fuller et al., 2004).

#### **2.2.3 Application to Human**

110 Magnetic Resonance Spectroscopy

In 2000, comparison of EC and UW solution with CSP, already performed on IPK of pig, was performed on this preclinical model of kidney autotransplantation (Hauet et al., 2000b). The control group, undergoing uninephrectomy, presented 100% survival at day 14 whereas only 75% and 66 % of UW and EC pigs survived, respectively. Glutathione-S-transferase and creatinine clearance were significantly different between EC and UW groups with a light advantages for UW compared to control. Results obtained with MRS on urine and plasma are presented in figure 3. Delayed graft function was associated with the release of osmolytes such as TMAO in urine and in plasma. Moreover, dimethylamine, acetate, citrate and lactate level in the UW group were always closer to control group levels compared to levels in the EC group. These data outline the effect of UW preservation against renal medullar injury and impairment of oxidative metabolism. Here, strength of MRS was highlighted: in one experiment, with a 200 to 500 µL volume of sample, detection and quantitation of large number of metabolites can be made. This allows for a multivariable

Using the same experimental design, the impact of the preservation solution was demonstrated and the addition of colloids (polyethylene glycol (PEG)) into an extracellular solution showed better results (Faure et al., 2002). Indeed, in the urine spectra, TMAO/creatine ratio were significantly higher in kidney preserved with EC, UW or a solution containing PEG-50 than in kidney preserved with a solution containing PEG-30. This demonstrated that PEG reduced damages to the renal medulla during conservation, in a dose-dependent manner. Citrate was not detected during the first postoperative week in urine from kidneys preserved in EC. Its excretion was detected significantly earlier in urine from the ICPEG30 group than from the UW and ICPEG50 groups. This suggested that PEG associated to an intracellular solution improved oxidative metabolism and reduced renal medulla injury, also in a concentration-dependent manner. Furthermore, the addition of antioxidative (Baumert et al., 1999) or anticoagulant (Favreau et al., 2010) molecules, trimetazidine and melagatran respectively, into the preservation solution increased the performance of the hypothermic conservation. The effect of trimetazidine during conservation with EC or UW showed interesting results. The early excretion of acetate in the trimetazidine supplemented group demonstrated an efficient function recovery of the citric acid cycle compared with standard solutions (Hauet et al., 2000a). Combination of PEG and trimetazidine in conservation solution was the next step of long term time study which included 18 groups also exploring cold ischemia time (Doucet et al., 2004). Biological parameters indicated that reduced delayed graft was found when combining PEG and trimetazidine during CSP. Histological staining for CD4+ positive cells indicated that PEG reduced CD4+ positive cells infiltration and trimetazidine was efficient in reducing the inflammatory reaction. The NMR analysis was again in accordance with these results. Acetate excretion between day 1 and day 14 was improved in all experimental group preserved with trimetazidine while urinary TMAO was reduced in all trimetazidine group and we noticed presence of TMAO in plasma in all group except from control. Combination of trimetazidine with PEG seemed to be the best option for kidney conservation and MRS brought supplementary information to classical biochemical parameters such as GFR or sodium excretion. In addition, this study underlined that MRS permits the performance of a

Use of conditioning drugs before conservation was also investigated. Donor pretreatment with N-acetylcysteine (NAC) has been shown to ameliorate acute renal failure. Extractions

analysis which is an important factor for diagnostic and prognostic purpose.

multivariate analysis in a single experiment.

Use of HR MRS in Human to follow transplanted kidney recovery of function highlighted several issues. The main problem came from the drug treatment received by patients after transplantation. We observed overlapping between the signals of the metabolites of interest and of the drugs. In these conditions it is often difficult to extract information about the kidney function. Moreover a high variability, in addition of the individual variability, is often observed. This variability, mainly due to diet**,** may cause urine profile modifications which can hide variations of the metabolite of interest in relation with the kidney injuries. One solution could be to introduce a 2-days controlled diet before each sampling to minimise those environmental variation. However, an older study performed on 33 patients who underwent primary renal allograft transplantation, using the same immunosuppressive regiment and of which 57.6% achieve immediate graft function, showed that patients with good graft function and normal patients had urinary TMAO/creatinine level inferior to 200 µmol/mM while patients with graft dysfunction had urinary TMAO/creatinine level of 410 ± 102 µmol/mM (Foxall et al., 1993).

#### **2.3 Perspectives**

#### **2.3.1 Perfusate analysis**

Hypothermic machine perfusion (MP) is increasingly being preferred to CSP. Indeed, MP has been showed to offer protection to the organ since the 80's (Alijani et al., 1985; Kwiatkowski et al., 2007; Wight et al., 2003) with benefits such as reduction of delayed graft function and increased survival. However, the mechanisms involved in these improvements are still unclear and need to be elucidated. Biomarkers such as lactate dehydrogenase, total glutathione-S-transferase or N-acetyl--D-glucosaminidase measured at the end of the perfusion in the machine perfusates are used by clinicians. Even if those markers are independently efficient to predict graft outcome, they are not efficient enough to be taken into account in the decision to transplant or discard a donation after brain death or controlled donation after cardiac death (Moers et al., 2010; Siew et al., 2011). Again the potential of multivariate analysis offered by NMR could be used to predict graft outcome. In a recent study in a pig model of autotransplantation of livers, the effect of warm ischemia on metabolites contents into perfusats was measured (Liu et al., 2009).

Currently, our team is attempting to answer the question: can 1H HR NMR metabonomic analysis during machine perfusion be used to predict graft outcome? We thus work on perfusates from different protocols of kidney transplantation using MP. We evaluate the effect of warm ischemia, duration of ischemia, types of solution and temperatures of preservation and finally effect of adjuvant in the perfusion solution. We showed, in a small

Magnetic Resonance Spectroscopy (MRS) in Kidney Transplantation: Interest and Perspectives 113

Magnetic resonance imaging (MRI) is a non invasive technique bringing essential information in the diagnostic of kidney transplant rejection since it can provide data on the anatomic and functional status of the transplanted kidney (Beckmann et al., 2000). Briefly, anatomic MRI is applied to evaluate the transplanted organ by assessing structural and anatomical changes or by measuring the graft volume. The most popular imaging sequences are the T1- and T2-weighted spin echo sequences and the T1-weighted gradient echo sequence. However, it would really be informative to add other acquisition sequences to obtain functional information about the transplanted kidney, if the magnet system is available during enough time to perform these exams. Indeed, further information can be

In addition of the magnetic resonance angiography to image blood vessels, quantification of the renal blood flow can be performed with dynamic MRI (Montet et al., 2003). Using a contrast agent, an association of a lower MRI perfusion with chronic allograft nephropathy severity was found (Pereira et al., 2011). Perfusion was determined in chosen area of the kidney, by measuring enhanced signal in the region of interest during the injection of the contrast agent. These results led to an homogeneous estimation of the perfusion, and the ability to evaluate the entire kidney with perfusion MRI may have benefits in establishing which areas of the transplant have a reduction in perfusion. The limitation in obtaining an entire kidney image is the necessity to use a cardiac gating to avoid pulsatility in the arterial curve. Another problem is the potential nephrotoxicity of the gadolinium which is generally used as contrast agent in a chelated form. Using low dose three dimensional magnetic resonance renography, the mean transit time of the tracer for the different compartment of the kidney (vascular, tubular and collecting system) was measured different in normal kidneys, transplanted kidneys with acute tubular necrosis and transplanted kidney with acute rejection (Yamamoto et al., 2011) indicating that a multicompartimental tracer kinetic renal model may

obtained on perfusion and functionality of the graft and the integrity of the tissue.

help to differentiate acute rejection from acute tubular necrosis in transplanted kidney.

Ultra-small superparamagnetic particles of iron oxides (USPIO) could be used as contrast agent in clinical medicine for in vivo MRI due to their properties to locally perturb magnetic field and novel particles are in development (Mills et al., 2011). Their toxicity is not well known and liquiq 1H NMR analysis of plasma and urine demonstrated metabolic changes (Feng et al., 2010). Energenetic and lipidic pathways such as glucose and amino acid metabolisms were affected by USPIO administration. Whereas no change has been found in histopathological analysis of tissue between two types of USPIO administration, coated with dextran or uncoated, coated-USPIO administration led to acetate, unsaturated fatty acids and some amino acids elevation in the 1H hrmas kidney profile. It also led to decrease of triacylglycerol, myo-niositol or taurine, for instance (Feng et al., 2011). Hence, this sort of

A noninvasive magnetic resonance arterial spin labeling (ASL) technique has been performed to evaluate kidney perfusion without contrast agents (Artz et al., 2011). ASL uses the blood as an endogenous contrast agent allowing perfusion measurements without gadolinium injection. Using a flow-sensitive alternating inversion recovery (FAIR) sequence, inflowing blood is selectively labeled to have an opposite magnetization compared to the

**3. Magnetic resonance in vivo** 

tracer has to be used with precaution.

**3.1 MRI** 

cohort of pig with two distinct graft outcomes, that a correlation between creatininemia at day 7 after transplantation and choline concentration after 24 hours of perfusion existed. Other metabolites like TMAO, lactate or glutathione showed the same correlation (Figure 4.). Metabonomic NMR analysis of machine perfusates could thus be use to predict graft outcome. A larger study is required to validate this finding, especially since this technique is easily transposable to the clinic.

Fig. 4. 1H HR NMR spectra of kidney perfusion solution (a) at the beginning of the perfusion and (b) after 24 hours of perfusion. (Ace: acetate, Ala: alanine, Cho: choline, Cr: creatine, Crine: creatinine, Glu: glutamate, GSH: reduced glutathione, GSSG: oxidized glutathione, TMAO: trimethylamine-N-oxide, Val: valine).

#### **2.3.2 <sup>1</sup> H hrmas NMR on kidney tissue or cells**

In an hrmas probe, adapted for classical spectrometer, samples are spun with high speed (2000 – 5000 Hz) at a 54.7° angle. This rotation leads to decreased heterogeneity of samples and allows direct sample analysis. 1H hrmas NMR is thus particularly adapted for tissue or cells without sample pre-treatment (Desmoulin et al., 2008). Addition of deuterated water into phosphate buffer solution is the only required modification of the samples.

In regards to the in vitro analysis of ischemia reperfusion injuries, our team developed an endothelial cell model mimicking transplantation through a hypoxia-reoxygenation episode in which conditions of hypoxia and/or reoxygenation can be modulated by changing temperature of hypoxia, atmosphere or conservation solutions. 1H hrmas NMR is a precious tool for this application but needs improvement for cell analysis. Indeed, the main problem resides on cell quantity, especially after hypoxia when cell number is inferior to 1 million as it is the case in our experiments. Use of the rotor allows for limitation of volume inside the tube and is one of the solutions we are examining. On the other hand, after 30 minutes spinning at 2500 Hz at 4°C, when cells were replaced into culture medium, 40 to 60 % of cells were still alive and restarted growing normally after 24 hours. Thus, we could imagine the use of 1H hrmas NMR to study all hypoxia-reoxygenation sequence on the same cells. In parallel, supernatant can also been analyzed in order to follow the cell metabolism.

Concerning tissue analysis, care has to be taken during tissue collection. Indeed, as already mentioned, because of its histological heterogeneity the kidney has different metabolites composition depending on the area (Garrod et al., 1999). For autotransplantation models as in clinical studies, a defined protocol for biopsy must be respected by all involved parties.

#### **3. Magnetic resonance in vivo**

#### **3.1 MRI**

112 Magnetic Resonance Spectroscopy

cohort of pig with two distinct graft outcomes, that a correlation between creatininemia at day 7 after transplantation and choline concentration after 24 hours of perfusion existed. Other metabolites like TMAO, lactate or glutathione showed the same correlation (Figure 4.). Metabonomic NMR analysis of machine perfusates could thus be use to predict graft outcome. A larger study is required to validate this finding, especially since this technique is

Fig. 4. 1H HR NMR spectra of kidney perfusion solution (a) at the beginning of the perfusion and (b) after 24 hours of perfusion. (Ace: acetate, Ala: alanine, Cho: choline, Cr: creatine, Crine: creatinine, Glu: glutamate, GSH: reduced glutathione, GSSG: oxidized glutathione,

In an hrmas probe, adapted for classical spectrometer, samples are spun with high speed (2000 – 5000 Hz) at a 54.7° angle. This rotation leads to decreased heterogeneity of samples and allows direct sample analysis. 1H hrmas NMR is thus particularly adapted for tissue or cells without sample pre-treatment (Desmoulin et al., 2008). Addition of deuterated water

In regards to the in vitro analysis of ischemia reperfusion injuries, our team developed an endothelial cell model mimicking transplantation through a hypoxia-reoxygenation episode in which conditions of hypoxia and/or reoxygenation can be modulated by changing temperature of hypoxia, atmosphere or conservation solutions. 1H hrmas NMR is a precious tool for this application but needs improvement for cell analysis. Indeed, the main problem resides on cell quantity, especially after hypoxia when cell number is inferior to 1 million as it is the case in our experiments. Use of the rotor allows for limitation of volume inside the tube and is one of the solutions we are examining. On the other hand, after 30 minutes spinning at 2500 Hz at 4°C, when cells were replaced into culture medium, 40 to 60 % of cells were still alive and restarted growing normally after 24 hours. Thus, we could imagine the use of 1H hrmas NMR to study all hypoxia-reoxygenation sequence on the same cells. In

into phosphate buffer solution is the only required modification of the samples.

parallel, supernatant can also been analyzed in order to follow the cell metabolism.

Concerning tissue analysis, care has to be taken during tissue collection. Indeed, as already mentioned, because of its histological heterogeneity the kidney has different metabolites composition depending on the area (Garrod et al., 1999). For autotransplantation models as in clinical studies, a defined protocol for biopsy must be respected by all involved parties.

easily transposable to the clinic.

TMAO: trimethylamine-N-oxide, Val: valine).

**H hrmas NMR on kidney tissue or cells** 

**2.3.2 <sup>1</sup>**

b

a

Magnetic resonance imaging (MRI) is a non invasive technique bringing essential information in the diagnostic of kidney transplant rejection since it can provide data on the anatomic and functional status of the transplanted kidney (Beckmann et al., 2000). Briefly, anatomic MRI is applied to evaluate the transplanted organ by assessing structural and anatomical changes or by measuring the graft volume. The most popular imaging sequences are the T1- and T2-weighted spin echo sequences and the T1-weighted gradient echo sequence. However, it would really be informative to add other acquisition sequences to obtain functional information about the transplanted kidney, if the magnet system is available during enough time to perform these exams. Indeed, further information can be obtained on perfusion and functionality of the graft and the integrity of the tissue.

In addition of the magnetic resonance angiography to image blood vessels, quantification of the renal blood flow can be performed with dynamic MRI (Montet et al., 2003). Using a contrast agent, an association of a lower MRI perfusion with chronic allograft nephropathy severity was found (Pereira et al., 2011). Perfusion was determined in chosen area of the kidney, by measuring enhanced signal in the region of interest during the injection of the contrast agent. These results led to an homogeneous estimation of the perfusion, and the ability to evaluate the entire kidney with perfusion MRI may have benefits in establishing which areas of the transplant have a reduction in perfusion. The limitation in obtaining an entire kidney image is the necessity to use a cardiac gating to avoid pulsatility in the arterial curve. Another problem is the potential nephrotoxicity of the gadolinium which is generally used as contrast agent in a chelated form. Using low dose three dimensional magnetic resonance renography, the mean transit time of the tracer for the different compartment of the kidney (vascular, tubular and collecting system) was measured different in normal kidneys, transplanted kidneys with acute tubular necrosis and transplanted kidney with acute rejection (Yamamoto et al., 2011) indicating that a multicompartimental tracer kinetic renal model may help to differentiate acute rejection from acute tubular necrosis in transplanted kidney.

Ultra-small superparamagnetic particles of iron oxides (USPIO) could be used as contrast agent in clinical medicine for in vivo MRI due to their properties to locally perturb magnetic field and novel particles are in development (Mills et al., 2011). Their toxicity is not well known and liquiq 1H NMR analysis of plasma and urine demonstrated metabolic changes (Feng et al., 2010). Energenetic and lipidic pathways such as glucose and amino acid metabolisms were affected by USPIO administration. Whereas no change has been found in histopathological analysis of tissue between two types of USPIO administration, coated with dextran or uncoated, coated-USPIO administration led to acetate, unsaturated fatty acids and some amino acids elevation in the 1H hrmas kidney profile. It also led to decrease of triacylglycerol, myo-niositol or taurine, for instance (Feng et al., 2011). Hence, this sort of tracer has to be used with precaution.

A noninvasive magnetic resonance arterial spin labeling (ASL) technique has been performed to evaluate kidney perfusion without contrast agents (Artz et al., 2011). ASL uses the blood as an endogenous contrast agent allowing perfusion measurements without gadolinium injection. Using a flow-sensitive alternating inversion recovery (FAIR) sequence, inflowing blood is selectively labeled to have an opposite magnetization compared to the

Magnetic Resonance Spectroscopy (MRS) in Kidney Transplantation: Interest and Perspectives 115

relation with the oxygen consumption in the tissue. In a murine model, BOLD MRI of a clamped kidney showed the decreased oxygenation of all regions. 24 hours after reperfusion, a lower reoxygenation of the outer medulla than control was found, and a higher reoxygenation of the cortex and an identical reoxygenation of the inner medulla were determined (Oostendorp et al., 2011). These results highlighted the potential of BOLD MRI for the detection of change in kidney tissue oxygenation. Non invasive exploration of human kidneys showed that oxygenation was lower in medulla than in the cortex, confirming a physiological medullar hypoxia (Malvezzi et al., 2009; Thoeny et al., 2006). A few days post transplantation, this technique showed an increase of T2 relaxation times in the cortex and medulla of the transplanted kidney, compared to the donated kidney of living donor before collection, demonstrating increased oxygenation (Malvezzi et al., 2009). A few months after transplantation, an increased oxygen content in the medulla was observed despite an unchanged perfusion fraction (Thoeny et al., 2006). Comparing allografts with normal function or acute tubular necrosis, a decreased blood flow and an increased oxygen bioavailability in the medulla of acutely rejecting kidneys were observed, suggesting a greater decrease in oxygen use (Sadowski et al., 2010). No significant modification of BOLD and DW MRI were observed in patients with acute tubular necrosis compared to normal functioning allografts. These results indicate the potential of these MRI techniques to longitudinally follow transplanted kidneys and

As an anecdote, MRI was tested for the counting the kidney glomerules number and size distribution in normal rat kidneys using ferritin as labeling and compared to standard stereological evaluation (Heilmann et al., 2011). This study showed that this ex vivo analysis

Magnetic resonance spectroscopy of Phosphorus-31 (31P) can be performed in kidney during the preservation period and in vivo after transplantation. In these two cases, 31P MRS brings information about the energetic metabolism. This non invasive methods allows the observation of a limited number of metabolites which are mainly: adenosine triphospate (ATP) which gives three signals corresponding to the three phosphorus nuclei in the ATP molecule ( and phosphorus corresponding to the position of the phosphorus nucleus in the molecule); inorganic phosphate (Pi) which has a chemical shift varying with the intracellular pH; phosphomonoesters (PME), mainly including phosphocholine and phosphoethanolamine that is highly concentrated in the renal cortex; phosphodiesters (PDE), corresponding to the resonances of glycerophosphocholine (GPC) and glycerolphosphoethanolamine (GPE) which can be used as an indicator of the physiological integrity of the organ (Wolff &Balaban, 1988). Phosphocreatine is not seen in NMR spectrum

Ex vivo or in vivo, 31P MRS does not allow an absolute quantification thus only signal ratios can be used. These ratios were studied in the preservation period and after the transplantation of the kidney, and compared with the kidney graft function. During the preservation period, 31P MRS was performed ex vivo with the kidney in the preservation solution and in the apparatus of conservation. The ATP disappeared quickly after the collection and perfusion with a cold preservation solution, in spite of the fact that the kidney was immediately cooled to 4°C and maintained to this temperature. The most significant

of entire kidney was less time consuming than stereological method.

of kidney, and when it appears in vivo, it comes from the surrounding tissues.

obtain information on function.

**3.2 31P MRS** 

destination tissue. The difference between a labeled image and a non labeled image can be used to calculate tissue perfusion. The first results obtained indicated that medullar perfusion was systematically lower in transplanted versus native kidneys irrespectively of estimated glomerular filtration rate (eGFR). Cortical perfusion was lower for transplanted kidneys when eGFR was > 60 ml/min per 1.73 m² and correlated with eGFR in both native and transplanted kidneys. These results were obtained with a one model compartment requiring rapid water exchange assumption between the intravascular and extravascular space. A two compartments model would allow more accurate perfusion quantification, but would require measurements which are not possible in a clinically feasible scan time.

Diffusion weighted (DW) MRI, which was established for the tissue characterization and lesion detection (Thoeny &De Keyzer, 2011), is particularly interesting for kidney function exploration because of its high blood flow and water transport functions. Simple acquisition is performed with a bipolar gradient (diffusion gradient) dedicated to the assessment of fluid movements. When molecules of water moves, the absence of rephasement of the nuclear spin appears as a loss of signal intensity on the image. If diffusion is restricted, molecules of water reduce their movement and may be refocused by the second gradient impulse (Palmucci et al., 2011). Different intensities of diffusion gradient can be used to determine the Apparent Diffusion Coefficient (ADC). The DW MRI techniques yields a total ADC in each voxel explored that provides information on diffusion properties of water, including contribution from micro-circulation (Le Bihan et al., 1988). Thus image processing allows to separate the micro-circulation information, quantified with the perfusion fraction (FP) which reflects micro-circulation of blood and movement of fluids in predefined structure such as tubules and glomeruli, and the primarily pure diffusion with the perfusion-free diffusion (ADCD) (Eisenberger et al., 2010). The potential of ADC has been studied in kidney transplantation to determined modification in case of graft rejection or acute tubular necrosis. In patients with stable allograft function (posttransplant time between 8 and 10 months), total ADC and ADCD were found identical in the cortex and medulla of the transplanted kidney while values were higher in cortex than in medulla of healthy volunteers. Cortical total ADC and ADCD were higher in native kidneys than in transplanted kidneys (Thoeny et al., 2006), and a difference in ADCD was observed between patients with normal clearance (>60 mL/min) and low clearance (>30 mL/min) where ADCD were the lowest (Palmucci et al., 2011). In allografts with stable function early after transplantation (posttransplantation time of 10 days) ADCD and FP were also found identical in cortex and medulla, but the FP was strongly reduced in the cortex and medulla of renal transplants with allograft rejection and acute tubular necrosis. FP correlated with eGFR, while no correlation were found between eGFR and ADCD or total ADC (Eisenberger et al., 2010). These results show that DW MRI is a promising non-invasive method for detection or monitoring of functional derangements early after kidney transplantation.

The blood oxygen level dependent (BOLD) MRI is a functional MRI technique allowing assessment of the tissue oxygenation. When the kidney is the explored organ, it allows the differentiation of specific anatomical regions (I.E cortex, inner and outer medulla). BOLD MRI is based on susceptibility differences between oxyhemoglobin and deoxyhemoglobin which induce image contrast differences, generally obtained with a gradient echoes sequence. Oxyhemoglogin is diamagnetic, and has no effect on the NMR signal of the surrounding water, by contrast deoxyhemoglobin is paramagnetic, increasing its magnetic properties and modifying the signal relaxation of the surrounding water molecules. Higher concentrations of deoxyhemoglobin induces shorter apparent T2 relaxation time, which is in relation with the oxygen consumption in the tissue. In a murine model, BOLD MRI of a clamped kidney showed the decreased oxygenation of all regions. 24 hours after reperfusion, a lower reoxygenation of the outer medulla than control was found, and a higher reoxygenation of the cortex and an identical reoxygenation of the inner medulla were determined (Oostendorp et al., 2011). These results highlighted the potential of BOLD MRI for the detection of change in kidney tissue oxygenation. Non invasive exploration of human kidneys showed that oxygenation was lower in medulla than in the cortex, confirming a physiological medullar hypoxia (Malvezzi et al., 2009; Thoeny et al., 2006). A few days post transplantation, this technique showed an increase of T2 relaxation times in the cortex and medulla of the transplanted kidney, compared to the donated kidney of living donor before collection, demonstrating increased oxygenation (Malvezzi et al., 2009). A few months after transplantation, an increased oxygen content in the medulla was observed despite an unchanged perfusion fraction (Thoeny et al., 2006). Comparing allografts with normal function or acute tubular necrosis, a decreased blood flow and an increased oxygen bioavailability in the medulla of acutely rejecting kidneys were observed, suggesting a greater decrease in oxygen use (Sadowski et al., 2010). No significant modification of BOLD and DW MRI were observed in patients with acute tubular necrosis compared to normal functioning allografts. These results indicate the potential of these MRI techniques to longitudinally follow transplanted kidneys and obtain information on function.

As an anecdote, MRI was tested for the counting the kidney glomerules number and size distribution in normal rat kidneys using ferritin as labeling and compared to standard stereological evaluation (Heilmann et al., 2011). This study showed that this ex vivo analysis of entire kidney was less time consuming than stereological method.

### **3.2 31P MRS**

114 Magnetic Resonance Spectroscopy

destination tissue. The difference between a labeled image and a non labeled image can be used to calculate tissue perfusion. The first results obtained indicated that medullar perfusion was systematically lower in transplanted versus native kidneys irrespectively of estimated glomerular filtration rate (eGFR). Cortical perfusion was lower for transplanted kidneys when eGFR was > 60 ml/min per 1.73 m² and correlated with eGFR in both native and transplanted kidneys. These results were obtained with a one model compartment requiring rapid water exchange assumption between the intravascular and extravascular space. A two compartments model would allow more accurate perfusion quantification, but would require measurements which are not possible in a clinically feasible scan time.

Diffusion weighted (DW) MRI, which was established for the tissue characterization and lesion detection (Thoeny &De Keyzer, 2011), is particularly interesting for kidney function exploration because of its high blood flow and water transport functions. Simple acquisition is performed with a bipolar gradient (diffusion gradient) dedicated to the assessment of fluid movements. When molecules of water moves, the absence of rephasement of the nuclear spin appears as a loss of signal intensity on the image. If diffusion is restricted, molecules of water reduce their movement and may be refocused by the second gradient impulse (Palmucci et al., 2011). Different intensities of diffusion gradient can be used to determine the Apparent Diffusion Coefficient (ADC). The DW MRI techniques yields a total ADC in each voxel explored that provides information on diffusion properties of water, including contribution from micro-circulation (Le Bihan et al., 1988). Thus image processing allows to separate the micro-circulation information, quantified with the perfusion fraction (FP) which reflects micro-circulation of blood and movement of fluids in predefined structure such as tubules and glomeruli, and the primarily pure diffusion with the perfusion-free diffusion (ADCD) (Eisenberger et al., 2010). The potential of ADC has been studied in kidney transplantation to determined modification in case of graft rejection or acute tubular necrosis. In patients with stable allograft function (posttransplant time between 8 and 10 months), total ADC and ADCD were found identical in the cortex and medulla of the transplanted kidney while values were higher in cortex than in medulla of healthy volunteers. Cortical total ADC and ADCD were higher in native kidneys than in transplanted kidneys (Thoeny et al., 2006), and a difference in ADCD was observed between patients with normal clearance (>60 mL/min) and low clearance (>30 mL/min) where ADCD were the lowest (Palmucci et al., 2011). In allografts with stable function early after transplantation (posttransplantation time of 10 days) ADCD and FP were also found identical in cortex and medulla, but the FP was strongly reduced in the cortex and medulla of renal transplants with allograft rejection and acute tubular necrosis. FP correlated with eGFR, while no correlation were found between eGFR and ADCD or total ADC (Eisenberger et al., 2010). These results show that DW MRI is a promising non-invasive method for detection or

monitoring of functional derangements early after kidney transplantation.

The blood oxygen level dependent (BOLD) MRI is a functional MRI technique allowing assessment of the tissue oxygenation. When the kidney is the explored organ, it allows the differentiation of specific anatomical regions (I.E cortex, inner and outer medulla). BOLD MRI is based on susceptibility differences between oxyhemoglobin and deoxyhemoglobin which induce image contrast differences, generally obtained with a gradient echoes sequence. Oxyhemoglogin is diamagnetic, and has no effect on the NMR signal of the surrounding water, by contrast deoxyhemoglobin is paramagnetic, increasing its magnetic properties and modifying the signal relaxation of the surrounding water molecules. Higher concentrations of deoxyhemoglobin induces shorter apparent T2 relaxation time, which is in Magnetic resonance spectroscopy of Phosphorus-31 (31P) can be performed in kidney during the preservation period and in vivo after transplantation. In these two cases, 31P MRS brings information about the energetic metabolism. This non invasive methods allows the observation of a limited number of metabolites which are mainly: adenosine triphospate (ATP) which gives three signals corresponding to the three phosphorus nuclei in the ATP molecule ( and phosphorus corresponding to the position of the phosphorus nucleus in the molecule); inorganic phosphate (Pi) which has a chemical shift varying with the intracellular pH; phosphomonoesters (PME), mainly including phosphocholine and phosphoethanolamine that is highly concentrated in the renal cortex; phosphodiesters (PDE), corresponding to the resonances of glycerophosphocholine (GPC) and glycerolphosphoethanolamine (GPE) which can be used as an indicator of the physiological integrity of the organ (Wolff &Balaban, 1988). Phosphocreatine is not seen in NMR spectrum of kidney, and when it appears in vivo, it comes from the surrounding tissues.

Ex vivo or in vivo, 31P MRS does not allow an absolute quantification thus only signal ratios can be used. These ratios were studied in the preservation period and after the transplantation of the kidney, and compared with the kidney graft function. During the preservation period, 31P MRS was performed ex vivo with the kidney in the preservation solution and in the apparatus of conservation. The ATP disappeared quickly after the collection and perfusion with a cold preservation solution, in spite of the fact that the kidney was immediately cooled to 4°C and maintained to this temperature. The most significant

Magnetic Resonance Spectroscopy (MRS) in Kidney Transplantation: Interest and Perspectives 117

*in vivo* **et** *in vitro models.*

**Metabolomic analysis Quantitation**

Fig. 5. Summary of samples for liquid or hrmas MRS analysis to study ischemia-reperfusion

From liquid NMR analysis of urine in IPK or perfusate in an autotransplantation model of large white pig, to hrmas analysis of cell culture or biopsy from Human (abstracted in the figure 5.) through imaging, MRS, combined with metabolomic analysis is a powerful tool for the study of transplantation. Despite all the difficulties in regards to kidney heterogeneity or low detection limit for cell analysis, MRS is still an interesting implement deserving of amelioration. Advances concerning imaging and contrast agents as well as transposition of hrmas to imaging are also the key for better diagnosis, prognosis and transplantation

We thank all people working in the INSERM U927 unit and allow us to access in vitro or in vivo samples: William Hebrard, Sébastien Giraud, Frédéric Favreau, Nicolas Chatauret and Raphael Thuillier, for his grammatical help especially. Thank to Nadège Boildieu, technician

 **Hrmas NMR** ✔ Kidney tissue ✔Endothelial cells

**Liquid NMR :** ✔ Plasma ✔ Urine ✔ Preservation solution ✔ Tissue extract ✔ Cellular extract ✔ Culture supernatant

phenomena occurring during kidney transplantation.

performed almost all the NMR acquisition in the team.

ADC Apparent Diffusion Coefficient ADCD Perfusion-free diffusion

ADPKD Autosomal dominant polycystic kidney diseases

**4. Conclusion** 

comprehension.

**5. Acknowledgment** 

**6. Abbreviations** 

1H Hydrogen-1 13C Carbon-13 31P Phosphorus-31

indicator of graft quality was the ratio of PME to Pi (PME/Pi), determined in human cadaveric kidneys prior to planned transplantation, which has been correlated with tubular necrosis and delayed graft function (Bretan et al., 1989; Hene et al., 1994). A high PME/Pi ratio was associated with the best renal function after transplantation; a decrease of this ratio (decrease of PME and/or increase of Pi) correlated with prolonged acute tubular necrosis. In regards to the determination of this ratio in the preserved kidney, the difficulty was the presence of a phosphate buffer in the preservation solution, this Pi signal overlapping the intracellular inorganic phosphate. Determination of this ratio can be obtained using Chemical Shift Imaging (CSI). This technique uses spatial resolution to obtain a 2D-CSI spectra, where each spectrum comes from a voxel. The interest is to minimize the Pi signal of the preservation solution in the voxel's content. In these conditions, a better correlation was obtained between PME/Pi ratio and serum creatinine 14 days after transplantation (Niekisch et al., 2004). A decrease of the PME/Pi ratio was shown as a monoexponential time dependent function in pig kidneys (von Elverfeldt et al., 2007). This decrease shows the influence of preservation time on transplantation outcome, and the authors determined a different decay constant with the preservation solution (comparison between two solutions) suggesting a predictive indicator of the preservation quality with time. This PME/Pi ratio indicator can be obtained with the kidney staying in its preservation container, at preservation temperature. No injection or biopsy is necessary, allowing easy integration of this exam in the period between collection and transplantation; the only condition being a non-magnetic container to be introduced in the magnet of a magnetic resonance apparatus (whole body scanner MR for this exam).

After transplantation it is possible to perform an in vivo 31P MRS of the kidney, which remains fully non invasive as it does not require injection nor anesthesia. More information can be obtained from the 31P spectrum of kidney since in addition to the PME and PDE signals, ATP signals can be observed and intracellular pH can be determined with the chemical shift of the Pi. When spectrum is obtained from the whole kidney, -ATP to Pi ratio measured 2 or 3 years after transplantation has been shown to be an excellent indicator of long term survival of the transplanted kidney (3 years when ratio > 1.2 and 5 years when ration > 2.5) (Seto et al., 2001). When MRS was performed in the 4 weeks after transplantation, the PME to PDE ratio seemed to be an indicator of allograft function (Klemm et al., 1998), showing a good correlation with serum creatinine levels. High PME/PDE ratio reflects the cell membrane regeneration process occurring during the first weeks after ischemia; PME level is considered as a reflection of the membrane phospholipid metabolism with the degree of cell growth and regeneration, while in contrast PDE level correspond to intermediate metabolites formed during cell membrane degradation. A decreased PME/PDE ratio was found in patients with delayed graft function due to acute tubular necrosis, but was not observed in patient with allograft rejection. A better localization of the voxel for 31P MRS could be obtained with CSI technique (Vyhnanovska et al., 2011), allowing the discard of surrounding tissue from the kidney signal in the spectra and the reduction in result dispersion, even if resolution did not permit to distinguish cortex from medulla in the spectra. In these conditions it is possible to distinguish different types of kidney failure after transplantation: (i) acute tubular necrosis, characterized by a decrease of PDE/-ATP, PDE/Pi and PME/Pi compared to controls; (ii) acute rejection episodes (ARE), characterized by an increase of PME/-ATP and PME/Pi ratio compare to controls. This method permitted the demonstration that late kidney graft dysfunction patients exhibited a higher Pi/-ATP and lower PDE/Pi and PME/Pi ratio compared with ARE patients.

Fig. 5. Summary of samples for liquid or hrmas MRS analysis to study ischemia-reperfusion phenomena occurring during kidney transplantation.

#### **4. Conclusion**

116 Magnetic Resonance Spectroscopy

indicator of graft quality was the ratio of PME to Pi (PME/Pi), determined in human cadaveric kidneys prior to planned transplantation, which has been correlated with tubular necrosis and delayed graft function (Bretan et al., 1989; Hene et al., 1994). A high PME/Pi ratio was associated with the best renal function after transplantation; a decrease of this ratio (decrease of PME and/or increase of Pi) correlated with prolonged acute tubular necrosis. In regards to the determination of this ratio in the preserved kidney, the difficulty was the presence of a phosphate buffer in the preservation solution, this Pi signal overlapping the intracellular inorganic phosphate. Determination of this ratio can be obtained using Chemical Shift Imaging (CSI). This technique uses spatial resolution to obtain a 2D-CSI spectra, where each spectrum comes from a voxel. The interest is to minimize the Pi signal of the preservation solution in the voxel's content. In these conditions, a better correlation was obtained between PME/Pi ratio and serum creatinine 14 days after transplantation (Niekisch et al., 2004). A decrease of the PME/Pi ratio was shown as a monoexponential time dependent function in pig kidneys (von Elverfeldt et al., 2007). This decrease shows the influence of preservation time on transplantation outcome, and the authors determined a different decay constant with the preservation solution (comparison between two solutions) suggesting a predictive indicator of the preservation quality with time. This PME/Pi ratio indicator can be obtained with the kidney staying in its preservation container, at preservation temperature. No injection or biopsy is necessary, allowing easy integration of this exam in the period between collection and transplantation; the only condition being a non-magnetic container to be introduced in the magnet of a magnetic resonance apparatus

After transplantation it is possible to perform an in vivo 31P MRS of the kidney, which remains fully non invasive as it does not require injection nor anesthesia. More information can be obtained from the 31P spectrum of kidney since in addition to the PME and PDE signals, ATP signals can be observed and intracellular pH can be determined with the chemical shift of the Pi. When spectrum is obtained from the whole kidney, -ATP to Pi ratio measured 2 or 3 years after transplantation has been shown to be an excellent indicator of long term survival of the transplanted kidney (3 years when ratio > 1.2 and 5 years when ration > 2.5) (Seto et al., 2001). When MRS was performed in the 4 weeks after transplantation, the PME to PDE ratio seemed to be an indicator of allograft function (Klemm et al., 1998), showing a good correlation with serum creatinine levels. High PME/PDE ratio reflects the cell membrane regeneration process occurring during the first weeks after ischemia; PME level is considered as a reflection of the membrane phospholipid metabolism with the degree of cell growth and regeneration, while in contrast PDE level correspond to intermediate metabolites formed during cell membrane degradation. A decreased PME/PDE ratio was found in patients with delayed graft function due to acute tubular necrosis, but was not observed in patient with allograft rejection. A better localization of the voxel for 31P MRS could be obtained with CSI technique (Vyhnanovska et al., 2011), allowing the discard of surrounding tissue from the kidney signal in the spectra and the reduction in result dispersion, even if resolution did not permit to distinguish cortex from medulla in the spectra. In these conditions it is possible to distinguish different types of kidney failure after transplantation: (i) acute tubular necrosis, characterized by a decrease of PDE/-ATP, PDE/Pi and PME/Pi compared to controls; (ii) acute rejection episodes (ARE), characterized by an increase of PME/-ATP and PME/Pi ratio compare to controls. This method permitted the demonstration that late kidney graft dysfunction patients exhibited a higher Pi/-ATP

and lower PDE/Pi and PME/Pi ratio compared with ARE patients.

(whole body scanner MR for this exam).

From liquid NMR analysis of urine in IPK or perfusate in an autotransplantation model of large white pig, to hrmas analysis of cell culture or biopsy from Human (abstracted in the figure 5.) through imaging, MRS, combined with metabolomic analysis is a powerful tool for the study of transplantation. Despite all the difficulties in regards to kidney heterogeneity or low detection limit for cell analysis, MRS is still an interesting implement deserving of amelioration. Advances concerning imaging and contrast agents as well as transposition of hrmas to imaging are also the key for better diagnosis, prognosis and transplantation comprehension.

#### **5. Acknowledgment**

We thank all people working in the INSERM U927 unit and allow us to access in vitro or in vivo samples: William Hebrard, Sébastien Giraud, Frédéric Favreau, Nicolas Chatauret and Raphael Thuillier, for his grammatical help especially. Thank to Nadège Boildieu, technician performed almost all the NMR acquisition in the team.

#### **6. Abbreviations**


Magnetic Resonance Spectroscopy (MRS) in Kidney Transplantation: Interest and Perspectives 119

Borel, M., Pastoureau, P., Papon, J., Madelmont, J.C., Moins, N., Maublant, J., & Miot-

Bretan, P.N., Baldwin, N., Novick, A.C., Majors, A., Easley, K., Ng, T., Stowe, N., Rehm, P.,

Desmoulin, F., Bon, D., Martino, R., & Malet-Martino, M. (2008). Étude critique de

NMR analysis of biological tissues: a critical report. CR Chimie *11*, 423-433. Doucet, C., Dutheil, D., Petit, I., Zhang, K., Eugene, M., Touchard, G., Wahl, A., Seguin, F.,

diffusion-weighted MR imaging. European Radiology *20*, 1374-1383. Faure, J.P., Hauet, T., Han, Z., Goujon, J.M., Petit, I., Mauco, G., Eugene, M., Carretier, M., &

Favreau, F., Thuillier, R., Cau, J., Milin, S., Manguy, E., Mauco, G., Zhu, X., Lerman, L.O., &

Feng, J., Li, X., Pei, F., Chen, X., Li, S., & Nie, Y. (2002). 1H NMR analysis for metabolites in

Feng, J., Liu, H., Bhakoo, K.K., Lu, L., & Chen, Z. (2011). A metabonomic analysis of organ

Feng, J., Liu, H., Zhang, L., Bhakoo, K., & Lu, L. (2010). An insight into the metabolic

Foxall, P.J., Mellotte, G.J., Bending, M.R., Lindon, J.C., & Nicholson, J.K. (1993). NMR

Fuller, T.F., Serkova, N., Niemann, C.U., & Freise, C.E. (2004). Influence of donor

Garrod, S., Humpfer, E., Spraul, M., Connor, S.C., Polley, S., Connelly, J., Lindon, J.C.,

Garrod, S., Humpher, E., Connor, S.C., Connelly, J.C., Spraul, M., Nicholson, J.K., & Holmes,

specific response to USPIO administration. Biomaterials.

metabonomic analysis of biofluids. Nanotechnology *21*, 395101.

research *8*, 2594-2600.

recipient patients. Transplantation *48*, 48-53.

experimental therapeutics *302*, 861-870.

Transplant *10*, 30-39.

biochemistry *301*, 1-7.

Med *41*, 1108-1118.

Kidney international *43*, 234-245.

grafts. The Journal of urology *171*, 1296-1300.

hepatic tissue. Magn Reson Med *45*, 781-790.

Noirault, E. (2009). Longitudinal profiling of articular cartilage degradation in osteoarthritis by high-resolution magic angle spinning 1H NMR spectroscopy: experimental study in the meniscectomized guinea pig model. Journal of proteome

Sreem, S.B., & Steinmuller, D.R. (1989). Pretransplant assessment of renal viability by phosphorus-31 magnetic resonance spectroscopy: clinical experience in 40

l'utilisation de la RMN HR-MAS pour l'analyse des tissus biologiques, HR-MAS

Milinkevitch, S., Hauet, T.*, et al.* (2004). Influence of colloid, preservation medium and trimetazidine on renal medulla injury. Biochimica et biophysica acta *1673*, 105-114. Eisenberger, U., Thoeny, H.C., Binser, T., Gugger, M., Frey, F.J., Boesch, C., & Vermathen, P.

(2010). Evaluation of renal allograft function early after transplantation with

Papadopoulos, V. (2002). Polyethylene glycol reduces early and long-term cold ischemia-reperfusion and renal medulla injury. The Journal of pharmacology and

Hauet, T. (2010). Anti-thrombin therapy during warm ischemia and cold preservation prevents chronic kidney graft fibrosis in a DCD model. Am J

serum and urine from rats administrated chronically with La(NO3)3. Analytical

responses of ultra-small superparamagnetic particles of iron oxide using

spectroscopy as a novel approach to the monitoring of renal transplant function.

pretreatment with N-acetylcysteine on ischemia/reperfusion injury in rat kidney

Nicholson, J.K., & Holmes, E. (1999). High-resolution magic angle spinning 1H NMR spectroscopic studies on intact rat renal cortex and medulla. Magn Reson

E. (2001). High-resolution (1)H NMR and magic angle spinning NMR spectroscopic investigation of the biochemical effects of 2-bromoethanamine in intact renal and


#### **7. References**


ARE Acute rejection episode ASL Arterial spin labeling ATP Adenosine triphospate

Fp Perfusion fraction GFR Glomerular filtration rate GPC Glycerophosphocholine GPE Glycerophosphoethanolamine

IPK Isolate perfused kidney MP Machine perfusion

NAC N-acetylcysteine

PDE Phosphodiesters PEG Polyethylene glycol Pi Inorganic phosphate PME Phosphomonoesters TMAO Trimethylamine-N-oxide

**7. References** 

MRI Magnetic resonance imaging MRS Magnetic Resonance Spectroscopy

NMR Nuclear Magnetic Resonance

UW University of Wisconsin solution

Transplantation *68*, 300-303.

*13*, 329-348.

preservation. Transplantation *40*, 659-661.

BOLD Blood oxygen level dependent CSI Chemical Shift Imaging CSP Cold-storage preservation DW Diffusion weighted EC EuroCollins solution

eGFR Estimated glomerular filtration rate

hrmas High resolution magic angle spinning

USPIO Ultra-small superparamagnetic particles of iron oxides

Alijani, M.R., Cutler, J.A., DelValle, C.J., Morres, D.N., Fawzy, A., Pechan, B.W., & Helfrich,

Artz, N.S., Sadowski, E.A., Wentland, A.L., Grist, T.M., Seo, S., Djamali, A., & Fain, S.B.

Baumert, H., Goujon, J.M., Richer, J.P., Lacoste, L., Tillement, J.P., Eugene, M., Carretier, M.,

Beckmann, N., Hof, R.P., & Rudin, M. (2000). The role of magnetic resonance imaging and

Blasco, H., Corcia, P., Moreau, C., Veau, S., Fournier, C., Vourc'h, P., Emond, P., Gordon, P.,

transplanted kidneys. Magnetic resonance imaging *29*, 74-82.

in early amyotrophic lateral sclerosis. PloS one *5*, e13223.

G.B. (1985). Single-donor cold storage versus machine perfusion in cadaver kidney

(2011). Arterial spin labeling MRI for assessment of perfusion in native and

& Hauet, T. (1999). Renoprotective effects of trimetazidine against ischemiareperfusion injury and cold storage preservation: a preliminary study.

spectroscopy in transplantation: from animal models to man. NMR in biomedicine

Pradat, P.F., Praline, J.*, et al.* (2010). 1H-NMR-based metabolomic profiling of CSF


Magnetic Resonance Spectroscopy (MRS) in Kidney Transplantation: Interest and Perspectives 121

Huhn, S.D., Szabo, C.M., Gass, J.H., & Manzi, A.E. (2004). Metabolic profiling of normal and

Klemm, A., Rzanny, R., Fünfstück, R., Werner, W., Schubert, J., Kaiser, W.A., & Stein, G.

Kwiatkowski, A., Wszola, M., Kosieradzki, M., Danielewicz, R., Ostrowski, K., Domagala,

Lenz, E.M., Bright, J., Knight, R., Wilson, I.D., & Major, H. (2004). Cyclosporin A-induced

Liu, Q., Vekemans, K., van Pelt, J., Pirenne, J., Himmelreich, U., Heedfeld, V., Wylin, T.,

Malvezzi, P., Bricault, I., Terrier, N., & Bayle, F. (2009). Evaluation of Intrarenal Oxygenation

Mills, P.H., Hitchens, T.K., Foley, L.M., Link, T., Ye, Q., Weiss, C.R., Thompson, J.D., Gilson,

Moers, C., Varnav, O.C., van Heurn, E., Jochmans, I., Kirste, G.R., Rahmel, A., Leuvenink,

Montet, X., Ivancevic, M.K., Jorge-Costa, M., Pochon, S., Perchère, A., & Vallée, J.P.

enhanced magnetic resonance imaging. Investigative radiology *38*, 584-592. Niekisch, M.B., Von Elverfeldt, D., El Saman, A., Hennig, J., & Kirste, G. (2004). Improved

after normothermic ischemia-reperfusion injury. NMR Biomed *24*, 194-200. Palmucci, S., Mauro, L.A., Veroux, P., Failla, G., Milone, P., Ettorre, G.C., Sinagra, N.,

Findings. Transplantation proceedings *43*, 960-966.

bioanalytical chemistry *378*, 1511-1519.

imaging. Radiology *168*, 497-505.

proceedings *41*, 3383-3386.

perturbations. Magn Reson Med.

Transplantation *90*, 966-973.

641-644.

pharmaceutical and biomedical analysis *35*, 599-608.

3147-3152.

hypertensive rat kidney tissues by hrMAS-NMR spectroscopy. Analytical and

(1998). 31P-magnetic resonance spectroscopy (31P-MRS) of human allografts after renal transplantation. Nephrology, dialysis, transplantation : official publication of the European Dialysis and Transplant Association - European Renal Association *13*,

P., Lisik, W., Nosek, R., Fesolowicz, S., Trzebicki, J.*, et al.* (2007). Machine perfusion preservation improves renal allograft survival. Am J Transplant *7*, 1942-1947. Le Bihan, D., Breton, E., Lallemand, D., Aubin, M.L., Vignaud, J., & Laval-Jeantet, M. (1988).

Separation of diffusuion and perfusion in intravoxel incoherent motion MR

changes in endogenous metabolites in rat urine: a metabonomic investigation using high field 1H NMR spectroscopy, HPLC-TOF/MS and chemometrics. Journal of

Brassil, J., Monbaliu, D., & Dresselaers, T. (2009). Discriminate Liver Warm Ischemic Injury During Hypothermic Machine Perfusion by Proton Magnetic Resonance Spectroscopy : A Study in a Porcine Model. Transplantation

by Blood Oxygen Level-Dependent Magnetic Resonance Imaging in Living Kidney Donors and Their Recipients : Preliminary Results. Transplantation proceedings *41*,

W.D., Arepally, A., Melick, J.A.*, et al.* (2011). Automated detection and characterization of SPIO-labeled cells and capsules using magnetic field

H.G., Squifflet, J.P., Paul, A., Pirenne, J.*, et al.* (2010). The value of machine perfusion perfusate biomarkers for predicting kidney transplant outcome.

(2003). Noninvasive measurement of absolute renal perfusion by contrast medium-

pretransplant assessment of renal quality by means of phosphorus-31 magnetic resonance spectroscopy using chemical shift imaging. Transplantation *77*, 1041-1045. Oostendorp, M., de Vries, E.E., Slenter, J.M., Peutz-Kootstra, C.J., Snoeijs, M.G., Post, M.J.,

van Heurn, L.W., & Backes, W.H. (2011). MRI of renal oxygenation and function

Giuffrida, G., Zerbo, D., & Veroux, M. (2011). Magnetic Resonance With Diffusion-Weighted Imaging in the Evaluation of Transplanted Kidneys : Preliminary


Goujon, J.M., Hauet, T., Menet, E., Levillain, P., Babin, P., & Carretier, M. (1999). Histological

Griffin, J.L., Walker, L.A., Garrod, S., Holmes, E., Shore, R.F., & Nicholson, J.K. (2000). NMR

laboratory rat. Comparative biochemistry and physiology *127*, 357-367. Gronwald, W., Klein, M.S., Zeltner, R., Schulze, B.D., Reinhold, S.W., Deutschmann, M.,

Hauet, T., Baumert, H., Amor, I.B., Gibelin, H., Tallineau, C., Eugene, M., Tillement, J.P., &

Hauet, T., Baumert, H., Gibelin, H., Hameury, F., Goujon, J.M., Carretier, M., & Eugene, M.

Hauet, T., Gibelin, H., Godart, C., Eugene, M., & Carretier, M. (2000c). Kidney retrieval

magnetic resonance (NMR) pectroscopy. Clin Chem Lab Med *38*, 1085-1092. Hauet, T., Gibelin, H., Richer, J.P., Godart, C., Eugene, M., & Carretier, M. (2000d). Influence

Hauet, T., Goujon, J.M., Tallineau, C., Carretier, M., & Eugene, M. (1999). Early evaluation of

Heilmann, M., Neudecker, S., Wolf, I., Gubhaju, L., Sticht, C., Schock-Kusch, D., Kriz, W.,

Hene, R.J., Van der Grond, J., Boer, W.H., Mali, W.P.T., & Koomans, H.A. (1994). Pre-

Holmes, E., Nicholson, J.K., Nicholls, A.W., Lindon, J.C., Connor, S.C., Polley, S., &

Chemometrics and Intelligent Laboratory Systems *44*, 245-255.

spectroscopy. Kidney international *46*, 1694-1699.

to cold ischemia. The Journal of surgical research *82*, 228-233.

fingerprinting of urine. Kidney international *79*, 1244-1253.

and experimental therapeutics *292*, 254-260.

injury. Cryobiology *41*, 280-291.

surgical research *68*, 116-125.

Nephrol Dial Transplant.

research *93*, 1-8.

evaluation of proximal tubule cell injury in isolated perfused pig kidneys exposed

spectroscopy based metabonomic studies on the comparative biochemistry of the kidney and urine of the bank vole (Clethrionomys glareolus), wood mouse (Apodemus sylvaticus), white toothed shrew (Crocidura suaveolens) and the

Immervoll, A.K., Boger, C.A., Banas, B., Eckardt, K.U.*, et al.* (2011). Detection of autosomal dominant polycystic kidney disease by NMR spectroscopic

Carretier, M. (2000a). Pharmacological limitation of damage to renal medulla after cold storage and transplantation by trimetazidine. The Journal of pharmacology

(2000b). Noninvasive monitoring of citrate, acetate, lactate, and renal medullary osmolyte excretion in urine as biomarkers of exposure to ischemic reperfusion

conditions influence damage to renal medulla: evaluation by proton nuclear

of retrieval conditions on renal medulla injury: evaluation by proton NMR spectroscopy in an isolated perfused pig kidney model. The Journal of surgical

renal reperfusion injury after prolonged cold storage using proton nuclear magnetic resonance spectroscopy. The British journal of surgery *86*, 1401-1409. Hauet, T., Mothes, D., Goujon, J.M., Caritez, J.C., Carretier, M., & Eugene, M. (1997a).

Evaluation of injury preservation in pig kidney cold storage by proton nuclear magnetic resonance spectroscopy of urine. The Journal of urology *157*, 1155-1160. Hauet, T., Mothes, D., Goujon, J.M., Caritez, J.C., Le Moyec, L., Carretier, M., & Eugene, M.

(1997b). Evaluation of normothermic ischemia and simple cold preservation injury in pig kidney by proton nuclear magnetic resonance spectroscopy. The Journal of

Bertram, J.F., Schad, L.R., & Gretz, N. (2011). Quantification of glomerular number and size distribution in normal rat kidneys using magnetic resonance imaging.

transpiantation assessment of renal viability with 31P magnetic resonance

Connelly, J. (1998). The identification of novel biomarkers of renal toxicity using automatic data reduction techniques and PCA of proton NMR spectra of urine.


**7** 

*Japan* 

**Acute Effects of Branched-Chain Amino** 

Tomoko Kutsuzawa1, Daisaku Kurita2 and Munetaka Haida3

The use of 31P-magnetic resonance (MR) spectroscopy (31P-MRS) allows noninvasive measurement of high-energy phosphate compounds such as phosphocreatine (PCr) and adenosine triphosphate (ATP), and the low-energy breakdown product, inorganic phosphate (Pi), in exercising muscle. Ratios of measureable high- and low-energy phosphate metabolites (e.g., PCr/Pi or PCr/(PCr+Pi) have been utilized as indices of the overall bioenergetic state of the cell (Sapega et al., 1987). Extracellular Pi does not exist in sufficient quantities to significantly affect 31P-MR spectra (Sapega et al., 1987). Intracellular pH (pHi) can also be measured noninvasively based on the pH-dependent chemical shift of cellular Pi that appears as PCr peaks on 31P-MRS (Sapega et al., 1987; Taylor et al., 1983). A decrease in pHi during exercise suggests lactate accumulation in

We have previously found, using 31P-MRS, that skeletal muscle metabolism in patients with chronic respiratory impairment undergoes specific changes (Kutsuzawa et al., 1992, 1995). Patients with chronic respiratory impairment display significant decreases in PCr and pHi during even mild exercise, suggesting that ATP production is reduced and that lactate levels accumulate rapidly in their muscles, suggesting reduced oxidative capacity. In addition, recent studies have demonstrated that reduced oxidative capacity in skeletal muscles correlates with an accelerated lactate response to exercise in patients with chronic obstructive pulmonary disease (COPD) (Maltais et al., 1996; Saey et al., 2005). Several factors such as inactivity, malnutrition and/or hypoxemia might contribute to altered muscle

Weight loss and muscle wasting are common features in patients with COPD (Bernard et al., 1998) and muscle wasting contributes to muscle weakness and exercise limitations in patients with COPD (Schols et al., 1993). Skeletal muscle is the major protein store that

**1. Introduction** 

muscles (Sapega et al., 1987; Taylor et al., 1983).

metabolism (Mannix et al., 1995; Payen et al., 1993; Sala et al., 1999).

*3Tokai University Junior College of Nursing and Medical Technology* 

**Acid Ingestion on Muscle pH during** 

**Exercise in Patients with Chronic** 

**Obstructive Pulmonary Disease** 

*1School of Health Sciences, Tokai University 2IT Education Center, Tokai University* 


## **Acute Effects of Branched-Chain Amino Acid Ingestion on Muscle pH during Exercise in Patients with Chronic Obstructive Pulmonary Disease**

Tomoko Kutsuzawa1, Daisaku Kurita2 and Munetaka Haida3 *1School of Health Sciences, Tokai University 2IT Education Center, Tokai University 3Tokai University Junior College of Nursing and Medical Technology Japan* 

#### **1. Introduction**

122 Magnetic Resonance Spectroscopy

Pereira, R.S., Gonul, I.I., McLaughlin, K., Yilmaz, S., & Mahallati, H. (2011). Assessment of

Sadowski, E.A., Djamali, A., Wentland, A.L., Muehrer, R., Becker, B.N., Grist, T.M., & Fain,

Serkova, N., Klawitter, J., & Niemann, C.U. (2003). Organ-specific response to inhibition of mitochondrial metabolism by cyclosporine in the rat. Transpl Int *16*, 748-755. Seto, K., Ikehira, H., Obata, T., Sakamoto, K., Yamada, K., Kashiwabara, H., Yokoyama, T., &

Siew, E.D., Ware, L.B., & Ikizler, T.A. (2011). Biological markers of acute kidney injury. J Am

Thoeny, H.C., & De Keyzer, F. (2011). Diffusion-weighted MR Imaging of Native and

Thoeny, H.C., Zumstein, D., Simon-zoula, S., Eisenberger, U., De Keyser, F., Hofmann, L.,

von Elverfeldt, D., Niekisch, M., Quaschning, T., El Saman, A., Kirste, G., Kramer-Guth, A.,

Vyhnanovska, P., Dezortova, M., Herynek, V., Taborsky, P., Viklicky, O., & Hajek, M. (2011).

Wang, Y., Bollard, M.E., Nicholson, J.K., & Holmes, E. (2006). Exploration of the direct

hepato- and nephrotoxicity. Chemical research in toxicology *18*, 639-654. Wight, J.P., Chilcott, J.B., Holmes, M.W., & Brewer, N. (2003). Pulsatile machine perfusion

Wolff, S.D., & Balaban, R.S. (1988). NMR studies of renal phosphate metabolites in vivo:

Yamamoto, A., Zhang, J.L., Rusinek, H., Chandarana, H., Vivier, P.H., Babb, J.S., Diflo, T.,

Zhao, L., Gao, H., Lian, F., Liu, X., Zhao, Y., & Lin, D. (2011). 1H NMR-based metabonomic

streptozotocin. American journal of physiology *300*, F947-F956.

Shift Imaging Method. Transplantation proceedings *43*, 1570-1575.

Study. American Journal of Roentgenology *194*, 407-413.

transplanted kidneys. Magnetic resonance imaging *28*, 56-64.

P magnetic resonance spectroscopy. Transplantation *72*, 627-630.

Soc Nephrol *22*, 810-820.

Biomed *20*, 652-657.

Transplanted kidneys. *259*, 25-38.

Experience. Radiology *241*, 812-821.

Clinical transplantation *17*, 293-307.

Radiology *260*, 781-789.

Chronic Renal Allograft Nephropathy Using Contrast-Enhanced MRI: A Pilot

S.B. (2010). Blood oxygen level-dependent and perfusion magnetic resonance imaging : detecting differences in oxygen bioavailability and blood flow in

Tanada, S. (2001). Long-term assessment of posttransplant renal prognosis with 31

Vock, P., Boesch, C., Frey, F.J., & Vermathen, P. (2006). Functional Evaluation of Transplanted Kidneys with Diffusion-weighted and BOLD MR Imaging : Initial

& Hennig, J. (2007). Kinetics of PME/Pi in pig kidneys during cold ischemia. NMR

In Vivo (31)P MR Spectroscopy of Human Kidney Grafts Using the 2D-Chemical

metabolic effects of mercury II chloride on the kidney of Sprague-Dawley rats using high-resolution magic angle spinning 1H NMR spectroscopy of intact tissue and pattern recognition. Journal of pharmaceutical and biomedical analysis *40*, 375-381. Waters, N.J., Waterfield, C.J., Farrant, R.D., Holmes, E., & Nicholson, J.K. (2005).

Metabonomic deconvolution of embedded toxicity: application to thioacetamide

vs. cold storage of kidneys for transplantation: a rapid and systematic review.

effects of hydration and dehydration. American journal of physiology *255*, 581-589.

John, D.G., Benstein, J.A., Barisoni, L.*, et al.* (2011). Quantitative evaluation of acute renal transplant dysfunction with low-dose three dimensional MR renography.

analysis of metabolic profiling in diabetic nephropathy rats induced by

The use of 31P-magnetic resonance (MR) spectroscopy (31P-MRS) allows noninvasive measurement of high-energy phosphate compounds such as phosphocreatine (PCr) and adenosine triphosphate (ATP), and the low-energy breakdown product, inorganic phosphate (Pi), in exercising muscle. Ratios of measureable high- and low-energy phosphate metabolites (e.g., PCr/Pi or PCr/(PCr+Pi) have been utilized as indices of the overall bioenergetic state of the cell (Sapega et al., 1987). Extracellular Pi does not exist in sufficient quantities to significantly affect 31P-MR spectra (Sapega et al., 1987). Intracellular pH (pHi) can also be measured noninvasively based on the pH-dependent chemical shift of cellular Pi that appears as PCr peaks on 31P-MRS (Sapega et al., 1987; Taylor et al., 1983). A decrease in pHi during exercise suggests lactate accumulation in muscles (Sapega et al., 1987; Taylor et al., 1983).

We have previously found, using 31P-MRS, that skeletal muscle metabolism in patients with chronic respiratory impairment undergoes specific changes (Kutsuzawa et al., 1992, 1995). Patients with chronic respiratory impairment display significant decreases in PCr and pHi during even mild exercise, suggesting that ATP production is reduced and that lactate levels accumulate rapidly in their muscles, suggesting reduced oxidative capacity. In addition, recent studies have demonstrated that reduced oxidative capacity in skeletal muscles correlates with an accelerated lactate response to exercise in patients with chronic obstructive pulmonary disease (COPD) (Maltais et al., 1996; Saey et al., 2005). Several factors such as inactivity, malnutrition and/or hypoxemia might contribute to altered muscle metabolism (Mannix et al., 1995; Payen et al., 1993; Sala et al., 1999).

Weight loss and muscle wasting are common features in patients with COPD (Bernard et al., 1998) and muscle wasting contributes to muscle weakness and exercise limitations in patients with COPD (Schols et al., 1993). Skeletal muscle is the major protein store that

Acute Effects of Branched-Chain Amino Acid Ingestion on Muscle pH

**2. Methods 2.1 Subjects** 

the participants.

**2.2 Study design** 

mass index; TSF, triceps skinfold thickness.

PaCO2, partial pressure of CO2; PaO2, partial pressure of O2.

during Exercise in Patients with Chronic Obstructive Pulmonary Disease 125

Subjects comprised 10 ambulatory male outpatients with stable COPD (mean age, 70.4 ± 8.8 years) diagnosed according to spirometric findings from moderate to very severe airflow limitation (forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC)<70% and FEV1 <80% of the predicted value) (Global Initiative for Chronic Obstructive Lung Disease, 2010). None of the patients had ever received systemic corticosteroid therapy and two of them had been treated with oxygen inhalation at home. None of them had participated in pulmonary rehabilitation. Exclusion criteria were malignancy, cardiac failure, renal failure, liver cirrhosis, diabetes mellitus, and infection. The ethics committee at our institution approved the study protocol, and both the control individuals and patients provided written informed consent to participate. Tables 1 and 2 show the physical characteristics of

 Patients (n=10) Age years 70.4 ± 8.8 Height cm 163.9 ± 3.8 Weight kg 55.6 ± 10.0 BMI kg/m2 20.6 ± 3.1 Forearm circumference cm 22.9 ± 1.7 TSF cm 5.6 ± 3.0 Grip power (left) kg 36.7 ± 5.5 Table 1. Anthropometric data from the patients. Values are given as mean ± SD. BMI, body

VC L 2.75 ± 0.77 VC (%of predicted) % 84.1 ± 21.8 FEV1 L 1.02 ± 0.40 FEV1/FVC % 40.3 ± 9.4 FEV1 (% of predicted) % 41.5 ± 15.4 pH 7.412 ± 0.031 PaCO2 Torr 41.3 ± 4.8 PaO2 Torr 79.8 ± 7.6 Table 2. Spirometric and blood gas analysis data from the patients. Values are given as mean ± SD. VC, vital capacity; FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity;

All patients fasted overnight, then anthropometric parameters and grip strength were measured at our outpatient clinic. Maximal voluntary grip strength of the non-dominant arm was measured using a dynamometer (DM-100N; Yagami, Nagoya, Japan). Fasting

Patients (n=10)

supplies amino acids to other tissues under specific conditions. Plasma concentrations of free amino acids indicate the balance between exogenous uptake and intercurrent metabolites in protein synthesis and breakdown (Wagenmakers, 1998). The branched-chain amino acids (BCAAs) leucine, isoleucine, and valine account for 35% of the essential amino acids contained in muscle proteins (Harper et al., 1984). Although essential amino acids other than BCAAs are mainly catabolized in the liver, BCAAs can be oxidized in skeletal muscle (Wagenmakers, 1998).

Several investigators have reported that the amino acid profile is altered in the plasma and skeletal muscle of patients with COPD (Engelen et al., 2000b; Hoffored et al., 1990; Pouw et al., 1998; Yoneda et al., 2001). Most of these studies have shown that plasma concentrations of the BCAAs leucine, isoleucine, and valine are reduced (Engelen et al., 2000b; Hoffored et al., 1990; Yoneda et al., 2001). Yoneda et al. (Yoneda et al., 2001) demonstrated that decreased concentrations of BCAAs in COPD are specifically related to weight loss and decreased muscle mass.

Energy expenditure greatly increases in skeletal muscle during exercise, and BCAA oxidization then maximally increases two- to three-fold (Knapik et al., 1991; Wolf et al., 1982). In addition, BCAAs might contribute to energy metabolism during exercise as substrates that expand the pool of tricarboxylic acid (TCA) cycle intermediates (Wagenmakers, 1998). Using 31P-MRS, we have previously found that plasma concentrations of BCAAs correlate with muscle pH at the completion of exercise in patients with COPD (Kutsuzawa et al., 2009). Those findings were consistent with a role of BCAAs in muscle energy metabolism during exercise in patients with COPD.

Supplementation with amino acids, particularly BCAAs, should increase exercise capacity in COPD patients, who might be affected by altered muscle and plasma amino acid profiles. Several studies in healthy subjects have examined the effects of BCAA ingestion on lactate metabolism during long-term, exhaustive exercise (De Palo et al., 2001; MacLean et al., 1996; Vukovich et al., 1992). MacLean et al. (MacLean et al., 1996) examined the effects of a large oral dose (308 mg/kg) of BCAAs on muscle amino acid metabolism during 90 min of exercise. They found that lactate release and arterial lactate values were lower in the group given BCAAs compared to a control group. We have studied the effects of BCAA ingestion on muscle pH during repeated bouts of short-term (3-min) exercise in healthy subjects and have found that BCAA supplementation before exercise can cause the attenuation of acidosis in exercising muscle, probably due to a decrease in lactate production (Kutsuzawa et al., 2011). Another study (Doi et al., 2004) of muscle energy metabolism demonstrated that an infusion of glucose and BCAAs before exercise improved acidic pHi during exercise in patients with liver cirrhosis accompanied by a severe amino acid imbalance.

The effects of BCAA ingestion on muscle pH have been investigated in healthy young participants (Kutsuzawa et al., 2011), but whether such supplementation will benefit exercise capacity in patients with COPD remains unclear. We thus used 31P-MRS to investigate the effects of BCAA ingestion on muscle pH during repeated bouts of short-term (3-min) exercise. Our hypothesis was that BCAA ingestion before a second bout of exercise could prevent metabolic acidosis in exercising muscle.

### **2. Methods**

124 Magnetic Resonance Spectroscopy

supplies amino acids to other tissues under specific conditions. Plasma concentrations of free amino acids indicate the balance between exogenous uptake and intercurrent metabolites in protein synthesis and breakdown (Wagenmakers, 1998). The branched-chain amino acids (BCAAs) leucine, isoleucine, and valine account for 35% of the essential amino acids contained in muscle proteins (Harper et al., 1984). Although essential amino acids other than BCAAs are mainly catabolized in the liver, BCAAs can be oxidized in skeletal

Several investigators have reported that the amino acid profile is altered in the plasma and skeletal muscle of patients with COPD (Engelen et al., 2000b; Hoffored et al., 1990; Pouw et al., 1998; Yoneda et al., 2001). Most of these studies have shown that plasma concentrations of the BCAAs leucine, isoleucine, and valine are reduced (Engelen et al., 2000b; Hoffored et al., 1990; Yoneda et al., 2001). Yoneda et al. (Yoneda et al., 2001) demonstrated that decreased concentrations of BCAAs in COPD are specifically related to weight loss and

Energy expenditure greatly increases in skeletal muscle during exercise, and BCAA oxidization then maximally increases two- to three-fold (Knapik et al., 1991; Wolf et al., 1982). In addition, BCAAs might contribute to energy metabolism during exercise as substrates that expand the pool of tricarboxylic acid (TCA) cycle intermediates (Wagenmakers, 1998). Using 31P-MRS, we have previously found that plasma concentrations of BCAAs correlate with muscle pH at the completion of exercise in patients with COPD (Kutsuzawa et al., 2009). Those findings were consistent with a role of BCAAs in muscle

Supplementation with amino acids, particularly BCAAs, should increase exercise capacity in COPD patients, who might be affected by altered muscle and plasma amino acid profiles. Several studies in healthy subjects have examined the effects of BCAA ingestion on lactate metabolism during long-term, exhaustive exercise (De Palo et al., 2001; MacLean et al., 1996; Vukovich et al., 1992). MacLean et al. (MacLean et al., 1996) examined the effects of a large oral dose (308 mg/kg) of BCAAs on muscle amino acid metabolism during 90 min of exercise. They found that lactate release and arterial lactate values were lower in the group given BCAAs compared to a control group. We have studied the effects of BCAA ingestion on muscle pH during repeated bouts of short-term (3-min) exercise in healthy subjects and have found that BCAA supplementation before exercise can cause the attenuation of acidosis in exercising muscle, probably due to a decrease in lactate production (Kutsuzawa et al., 2011). Another study (Doi et al., 2004) of muscle energy metabolism demonstrated that an infusion of glucose and BCAAs before exercise improved acidic pHi during exercise in patients with liver cirrhosis accompanied

The effects of BCAA ingestion on muscle pH have been investigated in healthy young participants (Kutsuzawa et al., 2011), but whether such supplementation will benefit exercise capacity in patients with COPD remains unclear. We thus used 31P-MRS to investigate the effects of BCAA ingestion on muscle pH during repeated bouts of short-term (3-min) exercise. Our hypothesis was that BCAA ingestion before a second bout of exercise

energy metabolism during exercise in patients with COPD.

muscle (Wagenmakers, 1998).

decreased muscle mass.

by a severe amino acid imbalance.

could prevent metabolic acidosis in exercising muscle.

#### **2.1 Subjects**

Subjects comprised 10 ambulatory male outpatients with stable COPD (mean age, 70.4 ± 8.8 years) diagnosed according to spirometric findings from moderate to very severe airflow limitation (forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC)<70% and FEV1 <80% of the predicted value) (Global Initiative for Chronic Obstructive Lung Disease, 2010). None of the patients had ever received systemic corticosteroid therapy and two of them had been treated with oxygen inhalation at home. None of them had participated in pulmonary rehabilitation. Exclusion criteria were malignancy, cardiac failure, renal failure, liver cirrhosis, diabetes mellitus, and infection. The ethics committee at our institution approved the study protocol, and both the control individuals and patients provided written informed consent to participate. Tables 1 and 2 show the physical characteristics of the participants.


Table 1. Anthropometric data from the patients. Values are given as mean ± SD. BMI, body mass index; TSF, triceps skinfold thickness.


Table 2. Spirometric and blood gas analysis data from the patients. Values are given as mean ± SD. VC, vital capacity; FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity; PaCO2, partial pressure of CO2; PaO2, partial pressure of O2.

#### **2.2 Study design**

All patients fasted overnight, then anthropometric parameters and grip strength were measured at our outpatient clinic. Maximal voluntary grip strength of the non-dominant arm was measured using a dynamometer (DM-100N; Yagami, Nagoya, Japan). Fasting

Acute Effects of Branched-Chain Amino Acid Ingestion on Muscle pH

routine methods in other portions of the samples.

normalized units of PCr/(PCr+Pi).

used as a pH indicator, as follows:

peaks if the peak was split.

**2.6 Statistical analysis** 

**3. Results** 

the end of the exercise period.

indicative of a significant difference.

**3.1 Physical characteristics and nutritional assessment** 

**2.5 31P-MRS** 

during Exercise in Patients with Chronic Obstructive Pulmonary Disease 127

leucine + isoleucine + valine, and total amino acids included all measured amino acids. Serum albumin and prealbumin, as indices of nutritional status, were determined by

Unlocalized MR spectra were obtained using a 2.0-T, 31-cm-bore BEM 250/80 superconducting magnet (Otsuka Electronics Co., Osaka, Japan). The spectrometer was operated at 85 MHz for 1H and at 34.5 MHz for 31P. A 4-cm surface coil was placed on the proximal third of the non-dominant forearm. We accumulated 31P-MR spectra every 3 s for 1 min using a single 90 pulse (50 μs) (Kutsuzawa et al., 1992). We then analyzed the 31P-MR

The signal area for Pi and PCr was determined from each spectrum by Gaussian curve fitting (Kutsuzawa et al., 1992). Relative concentrations of PCr and Pi were evaluated using

Muscle pH (i.e., pHi) was calculated as a difference in the chemical shift between Pi and the PCr peak (Stevens, 1987; Taylor et al., 1983). At physiological pH, Pi exists as HPO42- and H2PO4- ions. The signal from HPO42- is shifted downfield by 2.25 ppm from that of H2PO4-. However, the two forms are in rapid exchange, so a single Pi peak is formed with a chemical shift reflecting the relative proportions of the two ions. The chemical shift in Pi can thus be

pH = 6.75 + log (δ – 3.27)/(5.69 – δ), where δ is the difference of chemical shift in parts per million between the Pi and PCr signals (Taylor et al., 1983). The pH was determined from the weighted average of both Pi

We determined ∆pHi or ∆PCr/(PCr+Pi) as the difference between values just before and at

All data are presented as means standard deviation (SD). MRS variables between the control and BCAA bouts of exercise and between resting and the end of exercise were compared using Student's paired *t*-test. The ∆pHi and ∆PCr/(PCr+Pi) between control and BCAA bouts were compared to determine the effects of BCAA supplementation using the paired Student's *t*-test. We used linear regression analysis to evaluate correlations between concentration of BCAA and BMI, between BCAA and spirometric data, and between BCAA and the MRS variable (∆pHi of the first exercise bout) using the least-squares method. Goodness of fit was examined using the χ2-test. A value of p < 0.05 was considered

Table 1 shows the physical characteristics of the 10 patients. Mean body mass index of patients was 20.6 ± 3.1 kg/m2 and 3 of the patients were considered malnourished (BMI <

spectrum at rest (1 min before onset of exercise) and at the end of 3 min of exercise.

venous blood was obtained from an antecubital vein. Arterial blood was taken from the brachial artery for blood gas analysis while breathing room air.

All patients then performed two consecutive, 3-min constant work-rate exercises (a control bout and a BCAA bout), using the non-dominant forearm. Differences in MRS variables among individuals might be smaller for the non-dominant arm because routine activities might not vary as much as those performed with the dominant arm. Ten minutes after the first bout (control bout) of exercise, the patients consumed 8.0 g of powdered BCAAs (2 packs of LIVACT®; Ajinomoto, Tokyo, Japan) with 100 ml of water. The second bout (BCAA bout) of exercise started at 60 min after BCAA ingestion. Patients remained seated during the interval between first and second bouts of exercise and were allowed water. Muscle metabolism was measured using 31P-MRS during 1.5 min of rest, 3 min of exercise, and 4 min of recovery.

The exercise consisted of repetitively gripping a lever attached to a weight via a pulley system at a rate of 20 grips/min for 3 min while supine and breathing room air. The weight was lifted 5 cm by gripping the lever. To normalize the exercise intensity, the weight was adjusted to 7% of the maximum grip strength, which was suitable for this study because all patients could complete the exercise. Moreover, repetitive handgripping exercise in which the same weight (7% of maximal hang grip) is lifted 5 cm 20 times per min can differentiate healthy from altered muscle metabolism (as a decrease in pHi) (Kutsuzawa et al., 1992, 1995).

#### **2.3 BCAA supplementation**

LIVACT® is a BCAA powder that was developed to treat imbalances in amino acids among patients with liver cirrhosis. One pack of LIVACT (4.75 g) contains 4.0 g of BCAA (3808 mg of leucine, 1904 mg isoleucine and 2288 mg of valine). The composition ratio of leucine, isoleucine and valine at 2:1:1.2 effectively balances nitrogen as well as the plasma amino acid profile (Ohashi et al., 1989). The plasma BCAA level at 2 h after ingesting 8 g of BCAA increases 2-fold (Hamada et al., 2005).

#### **2.4 Nutritional assessment**

Nutritional status was evaluated by biochemical blood testing and anthropometric measurements including height, weight, non-dominant forearm circumference and triceps skin fold thickness (TSF). Body mass index (BMI) was calculated based on height and weight. The circumference of the non-dominant forearm was measured at the proximal third of the forearm, where the MRS surface coil was positioned. TSF of the non-dominant arm was measured using an EIYOKEN-TYPE skinfold caliper (Yagami).

Fasting blood samples from all patients were obtained by venipuncture before the control bout. In six of the 10 patients, second blood samples were obtained before the BCAA bout. Portions of samples were immediately cooled on ice and plasma obtained by centrifugation at 4C was stored at -80C for later amino acid analysis. Samples were deproteinized using 5% sulfosalicylic acid, then plasma levels of amino acids were measured by ion-exchange, high-pressure liquid chromatography with fluorometric detection (Model 8500; Hitachi, Tokyo, Japan) (Dyel et al., 1986). Total BCAAs comprised leucine + isoleucine + valine, and total amino acids included all measured amino acids. Serum albumin and prealbumin, as indices of nutritional status, were determined by routine methods in other portions of the samples.

### **2.5 31P-MRS**

126 Magnetic Resonance Spectroscopy

venous blood was obtained from an antecubital vein. Arterial blood was taken from the

All patients then performed two consecutive, 3-min constant work-rate exercises (a control bout and a BCAA bout), using the non-dominant forearm. Differences in MRS variables among individuals might be smaller for the non-dominant arm because routine activities might not vary as much as those performed with the dominant arm. Ten minutes after the first bout (control bout) of exercise, the patients consumed 8.0 g of powdered BCAAs (2 packs of LIVACT®; Ajinomoto, Tokyo, Japan) with 100 ml of water. The second bout (BCAA bout) of exercise started at 60 min after BCAA ingestion. Patients remained seated during the interval between first and second bouts of exercise and were allowed water. Muscle metabolism was measured using 31P-MRS during 1.5 min of rest, 3 min of exercise, and 4

The exercise consisted of repetitively gripping a lever attached to a weight via a pulley system at a rate of 20 grips/min for 3 min while supine and breathing room air. The weight was lifted 5 cm by gripping the lever. To normalize the exercise intensity, the weight was adjusted to 7% of the maximum grip strength, which was suitable for this study because all patients could complete the exercise. Moreover, repetitive handgripping exercise in which the same weight (7% of maximal hang grip) is lifted 5 cm 20 times per min can differentiate healthy from altered muscle metabolism (as a decrease in

LIVACT® is a BCAA powder that was developed to treat imbalances in amino acids among patients with liver cirrhosis. One pack of LIVACT (4.75 g) contains 4.0 g of BCAA (3808 mg of leucine, 1904 mg isoleucine and 2288 mg of valine). The composition ratio of leucine, isoleucine and valine at 2:1:1.2 effectively balances nitrogen as well as the plasma amino acid profile (Ohashi et al., 1989). The plasma BCAA level at 2 h after ingesting 8 g of BCAA

Nutritional status was evaluated by biochemical blood testing and anthropometric measurements including height, weight, non-dominant forearm circumference and triceps skin fold thickness (TSF). Body mass index (BMI) was calculated based on height and weight. The circumference of the non-dominant forearm was measured at the proximal third of the forearm, where the MRS surface coil was positioned. TSF of the non-dominant

Fasting blood samples from all patients were obtained by venipuncture before the control bout. In six of the 10 patients, second blood samples were obtained before the BCAA bout. Portions of samples were immediately cooled on ice and plasma obtained by centrifugation at 4C was stored at -80C for later amino acid analysis. Samples were deproteinized using 5% sulfosalicylic acid, then plasma levels of amino acids were measured by ion-exchange, high-pressure liquid chromatography with fluorometric detection (Model 8500; Hitachi, Tokyo, Japan) (Dyel et al., 1986). Total BCAAs comprised

arm was measured using an EIYOKEN-TYPE skinfold caliper (Yagami).

brachial artery for blood gas analysis while breathing room air.

min of recovery.

pHi) (Kutsuzawa et al., 1992, 1995).

increases 2-fold (Hamada et al., 2005).

**2.3 BCAA supplementation** 

**2.4 Nutritional assessment** 

Unlocalized MR spectra were obtained using a 2.0-T, 31-cm-bore BEM 250/80 superconducting magnet (Otsuka Electronics Co., Osaka, Japan). The spectrometer was operated at 85 MHz for 1H and at 34.5 MHz for 31P. A 4-cm surface coil was placed on the proximal third of the non-dominant forearm. We accumulated 31P-MR spectra every 3 s for 1 min using a single 90 pulse (50 μs) (Kutsuzawa et al., 1992). We then analyzed the 31P-MR spectrum at rest (1 min before onset of exercise) and at the end of 3 min of exercise.

The signal area for Pi and PCr was determined from each spectrum by Gaussian curve fitting (Kutsuzawa et al., 1992). Relative concentrations of PCr and Pi were evaluated using normalized units of PCr/(PCr+Pi).

Muscle pH (i.e., pHi) was calculated as a difference in the chemical shift between Pi and the PCr peak (Stevens, 1987; Taylor et al., 1983). At physiological pH, Pi exists as HPO42- and H2PO4- ions. The signal from HPO42- is shifted downfield by 2.25 ppm from that of H2PO4-. However, the two forms are in rapid exchange, so a single Pi peak is formed with a chemical shift reflecting the relative proportions of the two ions. The chemical shift in Pi can thus be used as a pH indicator, as follows:

$$\text{pH} = 6.75 + \log \text{ (6 - 3.27)} / \text{(5.69 - 6)},$$

where δ is the difference of chemical shift in parts per million between the Pi and PCr signals (Taylor et al., 1983). The pH was determined from the weighted average of both Pi peaks if the peak was split.

We determined ∆pHi or ∆PCr/(PCr+Pi) as the difference between values just before and at the end of the exercise period.

#### **2.6 Statistical analysis**

All data are presented as means standard deviation (SD). MRS variables between the control and BCAA bouts of exercise and between resting and the end of exercise were compared using Student's paired *t*-test. The ∆pHi and ∆PCr/(PCr+Pi) between control and BCAA bouts were compared to determine the effects of BCAA supplementation using the paired Student's *t*-test. We used linear regression analysis to evaluate correlations between concentration of BCAA and BMI, between BCAA and spirometric data, and between BCAA and the MRS variable (∆pHi of the first exercise bout) using the least-squares method. Goodness of fit was examined using the χ2-test. A value of p < 0.05 was considered indicative of a significant difference.

#### **3. Results**

#### **3.1 Physical characteristics and nutritional assessment**

Table 1 shows the physical characteristics of the 10 patients. Mean body mass index of patients was 20.6 ± 3.1 kg/m2 and 3 of the patients were considered malnourished (BMI <

Acute Effects of Branched-Chain Amino Acid Ingestion on Muscle pH

**3.3 Muscle energy metabolism** 

during Exercise in Patients with Chronic Obstructive Pulmonary Disease 129

Table 4 shows values for individual plasma amino acids. Total amino acids and total BCAAs were not reduced from the normal range. The mean glutamine concentration was approximately at the upper limit of the normal range. Concentrations of total BCAAs tended to correlate with BMI (r = 0.6099, p < 0.1) in patients. Concentrations of total BCAAs did not

Plasma amino acid profiles at 1 h after BCAA ingestion were examined in 6 of 10 patients (Table 5). Valine, leucine, and isoleucine significantly increased by more than 3- to 5-times after BCAA ingestion. Glutamine and arginine were also slightly but significantly increased.

Amino acid (nmol/L) Fasting After BCAA ingestion

Taurine 51.7 ± 6.0 52.6 ± 3.8 Threonine 112.4 ± 15.6 110.8 ± 19.4 Serine 101.7 ± 18.1 104.7 ± 20.5 Asparagine 45.5 ± 8.6 43.5 ± 7.3 Glutamic acid 32.6 ± 11.5 31.3 ± 14.2 Glutamine 749.2 ± 165.0 889.4 ± 167.4\*\* Proline 133.8 ± 20.1 116.9 ± 23.2\* Glycine 210.8 ± 47.7 190.7 ± 48.6\* Alanine 351.1 ± 67.1 315.5 ± 42.7 Citrulline 40.8 ± 16.6 39.6 ± 18.0 Aminobutyric acid 14.6 ± 3.7 15.3 ± 3.4 Valine 203.3 ± 22.9 679.8 ± 160.1\*\* Cystine 51.0 ± 6.4 51.0 ± 5.6 Methionine 25.1 ± 4.6 22.9 ± 4.2 Isoleucine 59.2 ± 12.2 322.6 ± 83.6\*\* Leucine 103.6 ± 19.0 606.5 ± 163.1\*\* Tyrosine 66.0 ± 14.9 61.5 ± 12.0 Phenylalanine 56.6 ± 9.8 53.0 ± 6.7 Histidine 74.2 ± 10.8 77.7 ± 10.4 Tryptophan 60.0 ± 9.1 53.4 ± 8.8\*\* Ornithine 59.0 ± 17.1 59.3 ± 17.7 Lysine 216.8 ± 29.2 232.4 ± 38.6 Arginine 111.1 ± 29.7 128.1 ± 36.3\* Total amino acids 2918.4 ± 332.0 4246.0 ± 320.3\*\* Total BCAAs 366.0 ± 51.5 1609.0 ± 386.9\*\* Table 5. Plasma amino acid concentration before and after ingestion of BCAAs. Valine, leucine, and isoleucine were increased more than 300% by 1 h after BCAA ingestion. Values are given as mean ± SD. \*: p < 0.05 vs. before ingestion; \*\*: p < 0.01 vs. before ingestion.

Representative 31P-MR spectra obtained from a 75-year-old patient with COPD at the end of exercise are shown in Figure 1. The peak for Pi was higher than that for PCr, and shifted

(n=6)

correlate with FEV1(%of the predicted value), which is an index of COPD severity.

20). Table 2 shows spirometric data from the patients with COPD. Mean FEV1 was 1.02 ± 0.40 L. Based on the Global Initiative for Chronic Obstructive Lung Disease (GOLD**)** criteria (Global Initiative for Chronic Obstructive Lung Disease, 2010), 3 of the 10 patients were categorized as stage II, 4 as stage III and 3 as stage IV, but none showed severe hypoxemia (partial pressure of O2 (PaO2), 79.8 ± 7.6 Torr).

#### **3.2 Nutritional assessment**

Values of serum albumin did not decrease in the patients. Although the mean level of prealbumin was within normal range (22-40 mg/dl), two of the patients showed a prealbumin level below the lower limit of the normal range (Table 3).


Table 3. Albumin and prealbumin levels in fasting venous blood from patients. Values are given as mean ± SD.


Table 4. Plasma amino acid concentrations before the first exercise bout. Values are given as mean ± SD. BCAA, branched-chain amino acid.

20). Table 2 shows spirometric data from the patients with COPD. Mean FEV1 was 1.02 ± 0.40 L. Based on the Global Initiative for Chronic Obstructive Lung Disease (GOLD**)** criteria (Global Initiative for Chronic Obstructive Lung Disease, 2010), 3 of the 10 patients were categorized as stage II, 4 as stage III and 3 as stage IV, but none showed severe hypoxemia

Values of serum albumin did not decrease in the patients. Although the mean level of prealbumin was within normal range (22-40 mg/dl), two of the patients showed a

 Patients (n=10) Albumin g/dl 4.5 ± 0.4 Prealbumin mg/dl 25.9 ± 6.5 Table 3. Albumin and prealbumin levels in fasting venous blood from patients. Values are

Amino acid (nmol/L) Range Patients (n=10) Taurine 40 - 93 56.4 ± 14.8 Threonine 67 - 190 101.5 ± 21.0 Serine 72 - 190 92.1 ± 20.8 Asparagine 45 - 97 43.1 ± 8.5 Glutamic acid 12 - 63 38.1 ± 12.0 Glutamine 420 - 700 699.5 ± 146.1 Proline 78 - 270 132.2 ± 35.1 Glycine 150 - 350 167.9 ± 40.6 Alanine 210 - 520 339.5 ± 55.0 Citrulline 17 - 43 36.5 ± 14.7 Aminobutyric acid 7.9 - 27 13.2 ± 4.1 Valine 150 - 310 199.9 ± 34.6 Cystine 29 - 49 49.4 ± 8.2 Methionine 19 - 40 23.7 ± 5.2 Isoleucine 40 - 110 59.4 ± 12.8 Leucine 48 - 180 103.2 ± 21.6 Tyrosine 40 - 90 65.2 ± 13.8 Phenylalanine 43 - 76 56.8 ± 11.2 Histidine 59 - 92 70.7 ± 12.7 Tryptophan 37 - 75 58.6 ± 9.3 Ornithine 30 - 100 53.1 ± 16.3 Lysine 110 - 240 191.5 ± 47.8 Arginine 54 - 130 100.1 ± 28.7 Total amino acids 2100 - 3500 2770.7 ± 391.8 Total BCAAs 270 - 600 362.4 ± 66.4 Table 4. Plasma amino acid concentrations before the first exercise bout. Values are given as

prealbumin level below the lower limit of the normal range (Table 3).

(partial pressure of O2 (PaO2), 79.8 ± 7.6 Torr).

mean ± SD. BCAA, branched-chain amino acid.

**3.2 Nutritional assessment** 

given as mean ± SD.

Table 4 shows values for individual plasma amino acids. Total amino acids and total BCAAs were not reduced from the normal range. The mean glutamine concentration was approximately at the upper limit of the normal range. Concentrations of total BCAAs tended to correlate with BMI (r = 0.6099, p < 0.1) in patients. Concentrations of total BCAAs did not correlate with FEV1(%of the predicted value), which is an index of COPD severity.

Plasma amino acid profiles at 1 h after BCAA ingestion were examined in 6 of 10 patients (Table 5). Valine, leucine, and isoleucine significantly increased by more than 3- to 5-times after BCAA ingestion. Glutamine and arginine were also slightly but significantly increased.


Table 5. Plasma amino acid concentration before and after ingestion of BCAAs. Valine, leucine, and isoleucine were increased more than 300% by 1 h after BCAA ingestion. Values are given as mean ± SD. \*: p < 0.05 vs. before ingestion; \*\*: p < 0.01 vs. before ingestion.

#### **3.3 Muscle energy metabolism**

Representative 31P-MR spectra obtained from a 75-year-old patient with COPD at the end of exercise are shown in Figure 1. The peak for Pi was higher than that for PCr, and shifted

Acute Effects of Branched-Chain Amino Acid Ingestion on Muscle pH

**Control bout BCAA bout**

**∆ ∆**


**4. Discussion** 

production.

**4.1 Muscle pH** 

buffer capacity remains unclear.

reduced in the BCAA bout.




0.0 **? pHi**

during Exercise in Patients with Chronic Obstructive Pulmonary Disease 131

The pHi of all participants in the control bout of exercise correlated with leucine concentration (r=0.6713) and tended to correlate with BCAA concentration (r=0.6015).

Fig. 2. Differences in pHi (∆pHi) and PCr/(PCr+Pi) (∆PCr/(PCr+Pi)) of each patient in control and BCAA bout. Six of 10 patients showed that ∆pHi and ∆PCr/(PCr+Pi) were

We investigated the acute effects of a single BCAA ingestion on muscle energy metabolism during hand-grip exercise in patients with COPD and found a decrease in ∆pHi and ∆PCr/(PCr+Pi) for the BCAA bout. The finding suggests that BCAA supplementation before exercise can cause an attenuation of acidosis in exercising muscle among COPD patients, probably due to an increase in oxidative capacity and a decrease in lactate

Values of pHi can be measured using 31P-MRS as the pH-dependence of a chemical shift between cellular Pi and PCr peaks (Sapega et al., 1987; Taylor et al., 1983). In addition, pHi can be determined as the H+ balance of consumption (PCr hydrolysis, buffering capacity and H+ efflux) and lactate production processes (Kemp, 2004). A decrease in pHi during exercise suggests that lactic acid accumulates in exercising muscle cells due to anaerobic glycolysis (Kemp, 2004; Sapega et al., 1987; Taylor et al., 1983). The line width frequently increased and the Pi peak often split during exercise in patients, suggesting that the pH among muscle fibers is heterogeneous, and supporting the suggestion that this phenomenon reflects metabolic heterogeneity among muscle fiber types (Vandenborne et al., 1993).

The H+ and lactate produced can either be buffered and removed intracellularly or released to the interstitium. Several transport systems as well as muscle buffering remove H+ during intense skeletal muscle contraction. Buffers include free Pi, PCr and histidine residues in both proteins and some dipeptides, and the capacity of such buffers is decreased by highintensity exercise via a decrease in protein buffering (Bishop et al., 2009). Supplementation with β-alanine, which increases muscle protein buffer capacity, attenuates the fall in blood pH during exercise (Baguet et al., 2010). Whether BCAA supplementation affects muscle




0 **? PCr/(PCr+Pi)**

**Control bout BCAA bout**

toward that of PCr during the control bout of exercise (Fig. 1A). In contrast, the peak of Pi was lower than that of PCr, and the Pi peak went back toward the normal chemical shift (-4.9 ppm). The values of pHi were calculated as 6.39 (Fig. 1A) and 6.84 (Fig. 1B) during the control and BCAA bouts of exercise, respectively. Values of PCr/(PCr+Pi) during the control (Fig. 1A) and BCAA bout (Fig. 1B) of exercise were 0.467 (Fig. 1A) and 0.630 (Fig. 1B), respectively.

Fig. 1. 31P-MR spectra after 3 min of exercise, obtained from a 75-year-old patient with COPD. A and B, First (control) and second (after BCAA ingestion) bouts of exercise, respectively; Pi, inorganic phosphate; PCr, phosphocreatine.

Calculated values of pHi and PCr/(PCr+Pi) in the control bout of exercise were 6.39 and 0.467, respectively, and those in the BCAA bout were 6.84 and 0.630, respectively.

Mean values of pHi and PCr/(PCr+Pi) at rest and at the end of each exercise are shown in Tables 6 and 7, respectively. Both pHi and PCr/(PCr+Pi) were significantly decreased at the end of each exercise bout. Mean values for pHi and PCr/(PCr+Pi) were significantly higher at the end of the BCAA bout than at the end of the control bout. Thus, ∆pHi and ∆PCr/(PCr+Pi) were significantly smaller in the BCAA bout. In addition, acidic changes in pHi were attenuated in 6 of the 10 patients in the BCAA bout (Fig. 2).


Table 6. Changes in muscle pH. \*, p < 0.05 vs. control bout; †, p < 0.05 vs. resting; #, p < 0.01 vs. resting. ∆pH, difference in pHi between resting and end of exercise.


Table 7. Changes in PCr/(PCr+Pi). \*, p < 0.05 vs. control bout; #, p < 0.01 vs. resting. ∆PCr/(PCr+Pi), difference in PCr/(PCr+Pi) between resting and end of exercise.

The pHi of all participants in the control bout of exercise correlated with leucine concentration (r=0.6713) and tended to correlate with BCAA concentration (r=0.6015).

Fig. 2. Differences in pHi (∆pHi) and PCr/(PCr+Pi) (∆PCr/(PCr+Pi)) of each patient in control and BCAA bout. Six of 10 patients showed that ∆pHi and ∆PCr/(PCr+Pi) were reduced in the BCAA bout.

#### **4. Discussion**

130 Magnetic Resonance Spectroscopy

toward that of PCr during the control bout of exercise (Fig. 1A). In contrast, the peak of Pi was lower than that of PCr, and the Pi peak went back toward the normal chemical shift (-4.9 ppm). The values of pHi were calculated as 6.39 (Fig. 1A) and 6.84 (Fig. 1B) during the control and BCAA bouts of exercise, respectively. Values of PCr/(PCr+Pi) during the control (Fig. 1A) and

Pi


**B**

PCr

BCAA bout (Fig. 1B) of exercise were 0.467 (Fig. 1A) and 0.630 (Fig. 1B), respectively.

 Chemical Shift (ppm) Chemical Shift (ppm) Fig. 1. 31P-MR spectra after 3 min of exercise, obtained from a 75-year-old patient with COPD. A and B, First (control) and second (after BCAA ingestion) bouts of exercise,

Calculated values of pHi and PCr/(PCr+Pi) in the control bout of exercise were 6.39 and

Mean values of pHi and PCr/(PCr+Pi) at rest and at the end of each exercise are shown in Tables 6 and 7, respectively. Both pHi and PCr/(PCr+Pi) were significantly decreased at the end of each exercise bout. Mean values for pHi and PCr/(PCr+Pi) were significantly higher at the end of the BCAA bout than at the end of the control bout. Thus, ∆pHi and ∆PCr/(PCr+Pi) were significantly smaller in the BCAA bout. In addition, acidic changes in

At rest 7.00 ± 0.08 7.03 ± 0.09 End of exercise 6.70 ± 0.19 # 6.86 ± 0.14 \* † ∆pH -0.31 ± 0.19 -0.17 ± 0.17 \* Table 6. Changes in muscle pH. \*, p < 0.05 vs. control bout; †, p < 0.05 vs. resting; #, p < 0.01

At rest 0.870 ± 0.023 0.876 ± 0.032 End of exercise 0.423 ± 0.056 0.502 ± 0.103 \* # ∆PCr/(PCr+Pi) -0.447 ± 0.059 # -0.373 ± 0.094 \*

Table 7. Changes in PCr/(PCr+Pi). \*, p < 0.05 vs. control bout; #, p < 0.01 vs. resting. ∆PCr/(PCr+Pi), difference in PCr/(PCr+Pi) between resting and end of exercise.

Control bout BCAA bout

Control bout BCAA bout

0.467, respectively, and those in the BCAA bout were 6.84 and 0.630, respectively.


respectively; Pi, inorganic phosphate; PCr, phosphocreatine.

pHi were attenuated in 6 of the 10 patients in the BCAA bout (Fig. 2).

vs. resting. ∆pH, difference in pHi between resting and end of exercise.

PCr

**A**

Pi

We investigated the acute effects of a single BCAA ingestion on muscle energy metabolism during hand-grip exercise in patients with COPD and found a decrease in ∆pHi and ∆PCr/(PCr+Pi) for the BCAA bout. The finding suggests that BCAA supplementation before exercise can cause an attenuation of acidosis in exercising muscle among COPD patients, probably due to an increase in oxidative capacity and a decrease in lactate production.

#### **4.1 Muscle pH**

Values of pHi can be measured using 31P-MRS as the pH-dependence of a chemical shift between cellular Pi and PCr peaks (Sapega et al., 1987; Taylor et al., 1983). In addition, pHi can be determined as the H+ balance of consumption (PCr hydrolysis, buffering capacity and H+ efflux) and lactate production processes (Kemp, 2004). A decrease in pHi during exercise suggests that lactic acid accumulates in exercising muscle cells due to anaerobic glycolysis (Kemp, 2004; Sapega et al., 1987; Taylor et al., 1983). The line width frequently increased and the Pi peak often split during exercise in patients, suggesting that the pH among muscle fibers is heterogeneous, and supporting the suggestion that this phenomenon reflects metabolic heterogeneity among muscle fiber types (Vandenborne et al., 1993).

The H+ and lactate produced can either be buffered and removed intracellularly or released to the interstitium. Several transport systems as well as muscle buffering remove H+ during intense skeletal muscle contraction. Buffers include free Pi, PCr and histidine residues in both proteins and some dipeptides, and the capacity of such buffers is decreased by highintensity exercise via a decrease in protein buffering (Bishop et al., 2009). Supplementation with β-alanine, which increases muscle protein buffer capacity, attenuates the fall in blood pH during exercise (Baguet et al., 2010). Whether BCAA supplementation affects muscle buffer capacity remains unclear.

Acute Effects of Branched-Chain Amino Acid Ingestion on Muscle pH

correlation between pHi and total BCAA after exercise.

**4.4 Effects of amino acid supplementation on patients with COPD** 

exercise test when compared to a control group (Blomstrand et al., 1996).

lactic threshold (Engelen et al., 2000a).

muscle damage.

during Exercise in Patients with Chronic Obstructive Pulmonary Disease 133

pyruvate + glutamic acid → alanine + α-ketoglutarate The accumulated pyruvate can generate alanine through this reaction instead of lactate during exercise. Engelen et al. discovered that glutamic acid concentrations in the quadriceps femoris muscles of COPD patients are reduced and might relate to a reduced

Secondly, BCAAs can be oxidized during moderate to heavy exercise. Several studies have reported that endurance exercise activates the branched-chain α-keto acid dehydrogenase (BCKDH) complex, which catalyzes branched-chain α-keto acid to form coenzyme A compounds in human and rat skeletal muscle (Shimomura et al., 1995; Wagenmakers et al., 1989). Shimomura et al. (Shimomura et al., 1993) reported that activity of the BCKDH complex increases in rat hindlimb muscles after 5 min of electrically induced contraction. Rates of leucine oxidation were almost double from rest to moderate intensity in healthy individuals performing steady-state exercise (Lamont et al., 2001). The oxidation of BCAAs does not relate to glycolysis or produce lactate. This mechanism might also explain the

The effects of BCAA ingestion before and/or after exercise on muscle damage and on muscle metabolism have been investigated in athletes since the 1990s. Several studies have demonstrated that BCAA supplementation attenuates the increase in blood lactate dehydrogenase (Coombes et al., 2000; Koba et al., 2007) and creatine kinase (Coombes et al., 2000) after prolonged exercise. Another study has shown that supplementation with amino acids including BCAAs attenuates delayed-onset muscle soreness (Nosaka et al., 2006). These findings suggested that BCAAs attenuate the degree of exercise-induced

Intense muscle activity generates intracellular protons and lactate accumulation. Several investigators have studied the effects of BCAA supplementation on lactate accumulation during exercise (Blomstrand et al., 1996; De Palo et al., 2001; MacLean et al., 1996; Matsumoto et a., 2009; Vukovich et al., 1992) in healthy subjects and athletes. MacLean et al. (MacLean et al., 1996) examined the effects of a large oral dose (308 mg/kg) of BCAAs on muscle amino acid metabolism during 90 min of dynamic knee extensor exercise. They reported that long-term exercise after BCAA administration resulted in significantly greater muscle NH3, alanine and glutamine levels, as well as lower lactate production. Matsumoto et al. (Matsumoto et al., 2009) studied the effects of BCAA supplementation on the lactate threshold during incremental exercise. They demonstrated that BCAA ingestion immediately before an incremental exercise test after 6 days of BCAA supplementation increased oxygen uptake and workload at the lactate threshold, the onset of blood lactate accumulation. Blomstrand et al. reported a smaller decrease in muscle glutamic acid levels in healthy volunteers who consumed BCAAs and then performed a prolonged submaximal

Supplementation with amino acids, particularly BCAAs and/or glutamic acid, should increase exercise capacity in COPD patients, as this capacity may be affected by altered muscle and plasma amino acid profiles. A small number of studies have examined

#### **4.2 Plasma BCAA profile in COPD**

Several studies have investigated plasma amino acid profiles in COPD patients (Engelen et al., 2000b; Hoffored et al., 1990; Pouw et al., 1998; Yoneda et al., 2001). No specific plasma amino acid profile has been identified, probably due to the heterogeneity of disease stages in COPD (for example, severity of airflow limitation, hypoxemia, and malnutrition). One study of Japanese patients with COPD (Yoneda et al., 2001) found that concentrations of glutamic acid and glutamine were elevated, whereas those of BCAAs were decreased except for leucine. We have previously investigated plasma amino acid levels in fasting venous blood among COPD patients and age-matched healthy subjects, revealing that plasma glutamine was elevated without overall changes in BCAAs (Kutsuzawa et al., 2009). The present study also showed that levels of plasma BCAA were within normal limits. One previous study (Yoneda et al., 2001) demonstrated that decreased concentrations of BCAAs in COPD are specifically related to weight loss and decreased muscle mass. Our previous (Kutsuzawa et al., 2009) and present studies have shown that the concentration of BCAAs correlated or tended to correlate with BMI (r=0.54488, r=0.6099), respectively, so nutritional status relates to a low BCAA concentration. However, severity of COPD did not correlate with the concentration of BCAAs, since no correlation was seen with FEV1/FEV1pred.

We measured BCAA levels immediately before the BCAA bout in 6 of 10 patients. Each BCAA concentration after BCAA ingestion increased significantly, by more than 3- to 5-fold compared to fasting values and to a consistent 2- to 3-fold more than the upper limit of the normal range. Another study demonstrated that plasma levels of BCAAs at 2 h after BCAA supplementation increase 2-fold after ingesting > 8 g of BCAAs (Hamada et al., 2005). Plasma levels of BCAA by the BCAA bout have thus been sufficiently increased. Increased plasma levels of BCAAs after administration subsequently enhance BCAA uptake by muscle during exercise (MacLean et al., 1996).

#### **4.3 Plasma BCAAs and muscle energy metabolism**

We have previously reported that plasma BCAA (leucine, isoleucine, valine) concentrations correlate with both intracellular pH and the PCr index during exercise (Kutsuzawa et al., 2009). The present study also showed the relationship between muscle pH and leucine (r=0.6713, p<0.05). Total BCAAs tended to correlate with muscle pH (r=0.6015, p<0.1). A previous study (Doi et al., 2004) investigated muscle energy metabolism in patients with liver cirrhosis accompanied by low plasma BCAA concentrations using 31P-MRS and found an acidic pHi during exercise.

Low plasma BCAA concentrations might affect lactate metabolism through two mechanisms. Firstly, BCAAs expand the pool of TCA cycle intermediates (Rutten et al., 2005; Wagenmakers, 1998) and react with α-ketoglutarate to produce branched-chain α-keto acids and glutamic acid in the presence of BCAA aminotransferase (Vandenborne et al., 1993; Wagenmakers, 1998):

$$\text{BCAA} + \mathfrak{a}\text{-ketoglutarate} \leftrightarrow \text{branched-chain } \mathfrak{a}\text{-keto acids} + \mathfrak{g}\text{lutamic acid}$$

BCAAs would therefore affect intracellular glutamic acid concentrations. During the first minutes of exercise, glutamic acid generates TCA cycle intermediates via the alanine aminotransferase reaction in skeletal muscle (Rutten et al., 2005):

Several studies have investigated plasma amino acid profiles in COPD patients (Engelen et al., 2000b; Hoffored et al., 1990; Pouw et al., 1998; Yoneda et al., 2001). No specific plasma amino acid profile has been identified, probably due to the heterogeneity of disease stages in COPD (for example, severity of airflow limitation, hypoxemia, and malnutrition). One study of Japanese patients with COPD (Yoneda et al., 2001) found that concentrations of glutamic acid and glutamine were elevated, whereas those of BCAAs were decreased except for leucine. We have previously investigated plasma amino acid levels in fasting venous blood among COPD patients and age-matched healthy subjects, revealing that plasma glutamine was elevated without overall changes in BCAAs (Kutsuzawa et al., 2009). The present study also showed that levels of plasma BCAA were within normal limits. One previous study (Yoneda et al., 2001) demonstrated that decreased concentrations of BCAAs in COPD are specifically related to weight loss and decreased muscle mass. Our previous (Kutsuzawa et al., 2009) and present studies have shown that the concentration of BCAAs correlated or tended to correlate with BMI (r=0.54488, r=0.6099), respectively, so nutritional status relates to a low BCAA concentration. However, severity of COPD did not correlate with the

concentration of BCAAs, since no correlation was seen with FEV1/FEV1pred.

We measured BCAA levels immediately before the BCAA bout in 6 of 10 patients. Each BCAA concentration after BCAA ingestion increased significantly, by more than 3- to 5-fold compared to fasting values and to a consistent 2- to 3-fold more than the upper limit of the normal range. Another study demonstrated that plasma levels of BCAAs at 2 h after BCAA supplementation increase 2-fold after ingesting > 8 g of BCAAs (Hamada et al., 2005). Plasma levels of BCAA by the BCAA bout have thus been sufficiently increased. Increased plasma levels of BCAAs after administration subsequently enhance BCAA uptake by muscle

We have previously reported that plasma BCAA (leucine, isoleucine, valine) concentrations correlate with both intracellular pH and the PCr index during exercise (Kutsuzawa et al., 2009). The present study also showed the relationship between muscle pH and leucine (r=0.6713, p<0.05). Total BCAAs tended to correlate with muscle pH (r=0.6015, p<0.1). A previous study (Doi et al., 2004) investigated muscle energy metabolism in patients with liver cirrhosis accompanied by low plasma BCAA concentrations using 31P-MRS and found

Low plasma BCAA concentrations might affect lactate metabolism through two mechanisms. Firstly, BCAAs expand the pool of TCA cycle intermediates (Rutten et al., 2005; Wagenmakers, 1998) and react with α-ketoglutarate to produce branched-chain α-keto acids and glutamic acid in the presence of BCAA aminotransferase (Vandenborne et al.,

BCAA + α-ketoglutarate ↔ branched-chain α-keto acids + glutamic acid BCAAs would therefore affect intracellular glutamic acid concentrations. During the first minutes of exercise, glutamic acid generates TCA cycle intermediates via the alanine

**4.2 Plasma BCAA profile in COPD** 

during exercise (MacLean et al., 1996).

an acidic pHi during exercise.

1993; Wagenmakers, 1998):

**4.3 Plasma BCAAs and muscle energy metabolism** 

aminotransferase reaction in skeletal muscle (Rutten et al., 2005):

#### pyruvate + glutamic acid → alanine + α-ketoglutarate

The accumulated pyruvate can generate alanine through this reaction instead of lactate during exercise. Engelen et al. discovered that glutamic acid concentrations in the quadriceps femoris muscles of COPD patients are reduced and might relate to a reduced lactic threshold (Engelen et al., 2000a).

Secondly, BCAAs can be oxidized during moderate to heavy exercise. Several studies have reported that endurance exercise activates the branched-chain α-keto acid dehydrogenase (BCKDH) complex, which catalyzes branched-chain α-keto acid to form coenzyme A compounds in human and rat skeletal muscle (Shimomura et al., 1995; Wagenmakers et al., 1989). Shimomura et al. (Shimomura et al., 1993) reported that activity of the BCKDH complex increases in rat hindlimb muscles after 5 min of electrically induced contraction. Rates of leucine oxidation were almost double from rest to moderate intensity in healthy individuals performing steady-state exercise (Lamont et al., 2001). The oxidation of BCAAs does not relate to glycolysis or produce lactate. This mechanism might also explain the correlation between pHi and total BCAA after exercise.

#### **4.4 Effects of amino acid supplementation on patients with COPD**

The effects of BCAA ingestion before and/or after exercise on muscle damage and on muscle metabolism have been investigated in athletes since the 1990s. Several studies have demonstrated that BCAA supplementation attenuates the increase in blood lactate dehydrogenase (Coombes et al., 2000; Koba et al., 2007) and creatine kinase (Coombes et al., 2000) after prolonged exercise. Another study has shown that supplementation with amino acids including BCAAs attenuates delayed-onset muscle soreness (Nosaka et al., 2006). These findings suggested that BCAAs attenuate the degree of exercise-induced muscle damage.

Intense muscle activity generates intracellular protons and lactate accumulation. Several investigators have studied the effects of BCAA supplementation on lactate accumulation during exercise (Blomstrand et al., 1996; De Palo et al., 2001; MacLean et al., 1996; Matsumoto et a., 2009; Vukovich et al., 1992) in healthy subjects and athletes. MacLean et al. (MacLean et al., 1996) examined the effects of a large oral dose (308 mg/kg) of BCAAs on muscle amino acid metabolism during 90 min of dynamic knee extensor exercise. They reported that long-term exercise after BCAA administration resulted in significantly greater muscle NH3, alanine and glutamine levels, as well as lower lactate production. Matsumoto et al. (Matsumoto et al., 2009) studied the effects of BCAA supplementation on the lactate threshold during incremental exercise. They demonstrated that BCAA ingestion immediately before an incremental exercise test after 6 days of BCAA supplementation increased oxygen uptake and workload at the lactate threshold, the onset of blood lactate accumulation. Blomstrand et al. reported a smaller decrease in muscle glutamic acid levels in healthy volunteers who consumed BCAAs and then performed a prolonged submaximal exercise test when compared to a control group (Blomstrand et al., 1996).

Supplementation with amino acids, particularly BCAAs and/or glutamic acid, should increase exercise capacity in COPD patients, as this capacity may be affected by altered muscle and plasma amino acid profiles. A small number of studies have examined

Acute Effects of Branched-Chain Amino Acid Ingestion on Muscle pH

plasma BCAA concentrations.

lactate concentrations.

**5. Conclusions** 

COPD patients.

**7. References** 

**6. Acknowledgments** 

Tokai University School of Medicine.

ISSN 14396319

(August 1998), pp. 629-634, ISSN 1073449X

during Exercise in Patients with Chronic Obstructive Pulmonary Disease 135

This study did not measure concentrations of amino acids in muscle. Glutamic acid and BCAAs are taken up by skeletal muscle after consuming meals containing protein and the carbon skeletons are used for *de novo* glutamine synthesis (Wagenmakers, 1998). A few studies (Pouw et al., 1998; Engelen et al., 2000b) have measured concentrations of muscle amino acids in patients with COPD. Pouw et al. (Pouw et al., 1998) reported that levels of BCAAs in plasma and muscle do not differ between patients with COPD and healthy individuals. However, Engelen et al. (Engelen et al., 2000b) showed that plasma and muscle levels of leucine and isoleucine are significantly decreased in COPD subtypes with emphysema, indicating that leucine metabolism is altered in COPD patients. Both studies demonstrated a decrease in glutamic acid in muscles in such patients. Muscle concentrations of glutamic acid and BCAAs might thus be decreased in patients with COPD and low

Furthermore, we did not measure plasma concentrations of lactate, because the magnet in our MR equipment was too narrow to collect venous blood from the exercising forearm during exercise. We also did not determine whether attenuation of metabolic acidosis in muscle resulted in a decrease in plasma lactate levels. However, the present exercise model involved a small muscle mass, which produced only limited changes in arterial H+ and

BCAA ingestion before a second bout of hand-grip exercise in patients with COPD resulted in a significant decrease in ∆pHi and ∆PCr/(PCr+Pi) at the end of the second bout of exercise. These findings suggest that BCAAs can help to prevent metabolic acidosis in exercising muscle among COPD patients, probably by increasing oxidative capacity and by decreasing lactate production. BCAA ingestion should thus improve exercise capacity in

This study was supported in part by a grant from the Department of Home Care Service,

Baguet, A., Koppo, K., Pottier, A., & Derave, W. (2010). β-Alanine supplementation reduces

Bernard, S., LeBlanc, P., Whittom, F., Carrier, G., Jobin, J., Belleau, R., & Maltais, F. (1998).

Bishop, D., Edge, J., Mendez-Villanueva, A., Thomas, C., & Schneiker, K. (2009) High-

acidosis but not oxygen uptake response during high-intensity cycling exercise. *European Journal of Applied Physiology,* Vol. 108, No. 3, (February 2010), pp. 495-503,

Peripheral muscle weakness in patients with chronic obstructive pulmonary disease. *American Journal of Respiratory and Critical Care Medicine,* Vol. 158, No. 2,

intensity exercise decreases muscle buffer capacity via a decrease in protein

supplementation with BCAAs or glutamic acid during exercise in COPD patients. Menier et al. studied the effects of BCAA administration to COPD patients during rehabilitation and showed no effects of BCAA supplementation on maximal oxygen uptake or maximal work (Menier et al., 2001). Rutten et al. studied the metabolic and functional effects of glutamic acid intake in COPD patients. Oral glutamic acid ingestion (30 mg/kg body weight every 20 min for 80 min) did not increase muscle glutamic acid concentration or reduce plasma lactate levels in COPD patients or healthy controls (Rutten et al., 2008).

The effects of a single administration of BCAAs on muscle energy metabolism in COPD patients have not been investigated. The present study showed that BCAA ingestion attenuated ∆pHi and ∆PCr/(PCr+Pi) at the end of exercise compared with the control bout. The findings suggested that BCAA supplementation before exercise could increase oxidative phosphorylation and decrease lactate production, then attenuate acidosis in exercising muscle in COPD patients. We have previously investigated the effects of BCAA ingestion on muscle pH during similar repeated bouts of short-term (3-min) exercise in healthy young participants (Kutsuzawa et al., 2011). The pHi at the second exercise bout without BCAA ingestion was significantly decreased compared with the first bout. In contrast, neither pHi after 3 min of exercise nor ∆pHi differed significantly between the first and second exercise bouts after BCAA ingestion in healthy participants. Effects of BCAA ingestion on muscle pH may be intense in COPD patients, probably due to low BCAA concentrations. Muscle energy metabolism might be affected by BCAAs during exercise acting as an energy source and as substrates that expand the pool of TCA cycle intermediates.

Patients with COPD frequently develop hypoxemia at rest and/or during exercise, which can influence muscle energy metabolism. PaO2 during exercise was not evaluated in the present study, although PaO2 at rest ranged from 67 to 90 mmHg. Our previous study evaluated PaO2 during the same hand-grip exercise in COPD patients, finding no significant decrease during exercise (Kutsuzawa et al., 1992). This suggests that reduced oxygen availability might not contribute to lactate accumulation in a control bout in COPD patients.

#### **4.5 Limitations of study**

The present study did not investigate muscle energy metabolism during two consecutive short-term bouts of exercise with a 1-h interval without BCAA supplementation. We previously studied the effects of a single administration of BCAAs on muscle pH during repeated short-term exercise (with or without BCAAs) in healthy young males (Kutsuzawa et al., 2011). In that study, healthy individuals performed repeated hand-grip exercises that consisted of the same weight, same distance, and double the frequency of the present study to induce low muscle pH. All participants who did not take BCAAs showed a pHi at completion of the second bout that was significantly decreased compared to that of the first bout. These findings suggested that the protocol resulted in decreased pHi at the end of the second bout of exercise as compared with the first bout. Therefore, pHi and PCr/(PCr+Pi) at the end of the second bout of hand-grip exercise in the present study might have been decreased.

This study did not measure concentrations of amino acids in muscle. Glutamic acid and BCAAs are taken up by skeletal muscle after consuming meals containing protein and the carbon skeletons are used for *de novo* glutamine synthesis (Wagenmakers, 1998). A few studies (Pouw et al., 1998; Engelen et al., 2000b) have measured concentrations of muscle amino acids in patients with COPD. Pouw et al. (Pouw et al., 1998) reported that levels of BCAAs in plasma and muscle do not differ between patients with COPD and healthy individuals. However, Engelen et al. (Engelen et al., 2000b) showed that plasma and muscle levels of leucine and isoleucine are significantly decreased in COPD subtypes with emphysema, indicating that leucine metabolism is altered in COPD patients. Both studies demonstrated a decrease in glutamic acid in muscles in such patients. Muscle concentrations of glutamic acid and BCAAs might thus be decreased in patients with COPD and low plasma BCAA concentrations.

Furthermore, we did not measure plasma concentrations of lactate, because the magnet in our MR equipment was too narrow to collect venous blood from the exercising forearm during exercise. We also did not determine whether attenuation of metabolic acidosis in muscle resulted in a decrease in plasma lactate levels. However, the present exercise model involved a small muscle mass, which produced only limited changes in arterial H+ and lactate concentrations.

#### **5. Conclusions**

134 Magnetic Resonance Spectroscopy

supplementation with BCAAs or glutamic acid during exercise in COPD patients. Menier et al. studied the effects of BCAA administration to COPD patients during rehabilitation and showed no effects of BCAA supplementation on maximal oxygen uptake or maximal work (Menier et al., 2001). Rutten et al. studied the metabolic and functional effects of glutamic acid intake in COPD patients. Oral glutamic acid ingestion (30 mg/kg body weight every 20 min for 80 min) did not increase muscle glutamic acid concentration or reduce plasma

The effects of a single administration of BCAAs on muscle energy metabolism in COPD patients have not been investigated. The present study showed that BCAA ingestion attenuated ∆pHi and ∆PCr/(PCr+Pi) at the end of exercise compared with the control bout. The findings suggested that BCAA supplementation before exercise could increase oxidative phosphorylation and decrease lactate production, then attenuate acidosis in exercising muscle in COPD patients. We have previously investigated the effects of BCAA ingestion on muscle pH during similar repeated bouts of short-term (3-min) exercise in healthy young participants (Kutsuzawa et al., 2011). The pHi at the second exercise bout without BCAA ingestion was significantly decreased compared with the first bout. In contrast, neither pHi after 3 min of exercise nor ∆pHi differed significantly between the first and second exercise bouts after BCAA ingestion in healthy participants. Effects of BCAA ingestion on muscle pH may be intense in COPD patients, probably due to low BCAA concentrations. Muscle energy metabolism might be affected by BCAAs during exercise acting as an energy source and as substrates that expand the pool of TCA cycle

Patients with COPD frequently develop hypoxemia at rest and/or during exercise, which can influence muscle energy metabolism. PaO2 during exercise was not evaluated in the present study, although PaO2 at rest ranged from 67 to 90 mmHg. Our previous study evaluated PaO2 during the same hand-grip exercise in COPD patients, finding no significant decrease during exercise (Kutsuzawa et al., 1992). This suggests that reduced oxygen availability might not contribute to lactate accumulation in a control bout in

The present study did not investigate muscle energy metabolism during two consecutive short-term bouts of exercise with a 1-h interval without BCAA supplementation. We previously studied the effects of a single administration of BCAAs on muscle pH during repeated short-term exercise (with or without BCAAs) in healthy young males (Kutsuzawa et al., 2011). In that study, healthy individuals performed repeated hand-grip exercises that consisted of the same weight, same distance, and double the frequency of the present study to induce low muscle pH. All participants who did not take BCAAs showed a pHi at completion of the second bout that was significantly decreased compared to that of the first bout. These findings suggested that the protocol resulted in decreased pHi at the end of the second bout of exercise as compared with the first bout. Therefore, pHi and PCr/(PCr+Pi) at the end of the second bout of hand-grip exercise in

lactate levels in COPD patients or healthy controls (Rutten et al., 2008).

intermediates.

COPD patients.

**4.5 Limitations of study** 

the present study might have been decreased.

BCAA ingestion before a second bout of hand-grip exercise in patients with COPD resulted in a significant decrease in ∆pHi and ∆PCr/(PCr+Pi) at the end of the second bout of exercise. These findings suggest that BCAAs can help to prevent metabolic acidosis in exercising muscle among COPD patients, probably by increasing oxidative capacity and by decreasing lactate production. BCAA ingestion should thus improve exercise capacity in COPD patients.

#### **6. Acknowledgments**

This study was supported in part by a grant from the Department of Home Care Service, Tokai University School of Medicine.

#### **7. References**


Acute Effects of Branched-Chain Amino Acid Ingestion on Muscle pH

(January 1991), pp. 43-47, ISSN 87507587

ISSN 15677249

ISSN 00224707

1073449X

02615614

01959131

00223751

293, ISSN 1073449X

pp. 1019-1024, ISSN 00030805

during Exercise in Patients with Chronic Obstructive Pulmonary Disease 137

Knapik, J., Meredith, C., Jones, B., Fieldry, R., Young, V., & Evans, W. (1991). Leucine

Koba, T., Hamada, K., Sakurai, M., Matsumoto, K., Hayase, H., Imaizumi, K., Tsujimoto, H.,

Kutsuzawa, T., Shioya, S., Kurita, D., Haida, M., Ohta, Y., & Yamabayashi, H. (1992). 31P-

Kutsuzawa, T., Shioya, S., Kurita, D., Haida, M., Ohta, Y., & Yamabayashi, H. (1995). Muscle

Kutsuzawa, T., Shioya, S., Kurita, D., & Haida, H. (2009). Plasma branched-chain amino acid

Kutsuzawa, T., Kurita, D., & Haida, M. (2011). Acute effects of branched-chain amino acids

Lamont, L. S., McCullough, A. J., & Kalhan, S. (2001). Relationship between leucine

MacLean, D. A., Graham, T. E., & Saltin, B. (1996) Stimulation of muscle ammonia

Maltais, F., Simard, A., Simard, C., Jobin, J., Desgagnes, P., & LeBlanc, P. (1996).

Mannix, E. T., Boska, M. D., Galassetti, P., Burton, G., Manfredi, F., & Farber, M. O. (1995).

No. 4, (March 2011), pp. 101-107, ISSN 13403141

*Disease,* Vol. 141, No. 4 pt. 1, (April 1990), pp. 902-908, ISSN 00030805 Kemp, G. J. (2004). Mitochondrial dysfunction in chronic ischemia and peripheral

study of amino acid and catecholamine levels. *American Review of Respiratory* 

vascular disease. *Mitochondrion* Vol. 4, No. 5-6, (September 2004), pp. 629-640,

metabolism during fasting and exercise. *Journal of Applied Physiology,* Vol. 70, No. 1,

& Mitsuzono, R. (2007). Branched-chain amino acids supplementation attenuates the accumulation of blood lactate dehydrogenase during distance running. *Journal of Sports Medicine and Physical Fitness,* Vol. 47, No. 3, (September 2007), pp. 316-322,

NMR study of skeletal muscle metabolism in patients with chronic respiratory impairment. *American Review of Respiratory Disease,* Vol. 146, No. 4, (October 1992),

energy metabolism and nutritional status in patients with chronic obstructive pulmonary disease. A 31P magnetic resonance study. *American Journal of Respiratory and Critical Care Medicine,* Vol. 152, No. 2, (August 1995), pp. 647-652, ISSN

levels and muscle energy metabolism in patients with chronic obstructive pulmonary disease. *Clinical Nutrition,* Vol. 28, No. 2, (April 2009), pp. 203-208, ISSN

on muscle pH during exercise. *Advances in Exercise and Sports Physiology,* Vol. 16,

oxidation and oxygen consumption during steady-state exercise. *Medicine and Science in Sports and Exercise,* Vol. 33, No. 2, (February 2001), pp. 237-241, ISSN

production during exercise following branched-chain amino acid supplementation in humans. *Journal of Physiology,* Vol. 493, No. 3, (June 1996), pp. 909-922, ISSN

Oxidative capacity of the skeletal muscle and lactic acid kinetics during exercise in normal subjects and in patients with COPD. *American Journal of Respiratory and Critical Care Medicine*, Vol. 153, No. 1, (January 1996), pp. 288-

Modulation of ATP production by oxygen in obstructive lung disease as assessed

buffering in human skeletal muscle. *European Journal of Physiology,* Vol. 458, No. 5, (September 2009), pp. 929-936, ISSN 00316768


http://www.goldcopd.org/uploads/users/files/GOLDReport\_April112011.pdf


Blomstrand, E., Ek, S., & Newsholme, E. A. (1996). Influence of ingesting a solution of

Coombes, J. S., & McNaughton, L. R. (2000). Effects of branched-chain amino acid

De Palo, E. F., Gatti, R., Cappellin, E., Schiraldi, C., De Palo, C. B., & Spinella, P. (2001).

Deyl, Z., Hyanek, J., & Horakova, M. (1986). Profiling of amino acids in body fluids and

Doi, J., Shiraishi, K., Haida, M., & Matsuzaki, S. (2004). Abnormality of energy metabolism

Engelen, M. P. K. J., Schols, A. M. W. J., Does, J. D., Gosker, H. R., Deutz, N. E. P., Wouters,

Engelen, M. P. K. J., Wouters, E. F. M., Deutz, N. E. P., Menheere, P. P. C. A., & Schols, A. M.

Global Initiative for Chronic Obstructive Lung Disease. (2010). Definition, In: *Global strategy* 

 http://www.goldcopd.org/uploads/users/files/GOLDReport\_April112011.pdf Hamada, K., Koba, T., Sakurai, M., Matsumoto, K., Higuchi, T., Imaizumi, K., Hayase, H., &

Harper, A. E., Miller, R. H., & Block, K. P. (1984). Branched-chain amino acid metabolism. *Annual Review of Nutrition,* Vol. 4, (July 1984), pp. 409-454, ISSN 01999885 Hofford, J. M., Milakofsky, L., Vogel, W. H., Sacher, R. S., Savage, G. J., & Pell, S. (1990). The

(September 2009), pp. 929-936, ISSN 00316768

(September 2000), pp. 240-246, ISSN 00224707

(June 1986), pp. 177-250, ISSN 03784347

485-490, ISSN 08999007

ISSN 09394451

ISSN 03850005

1073449X

ISSN 00029165

August 10 2011, available from

2005), pp. 1-10 (In Japanese), ISSN 02868202

buffering in human skeletal muscle. *European Journal of Physiology,* Vol. 458, No. 5,

branched-chain amino acids on plasma and muscle concentrations of amino acids during prolonged submaximal exercise. *Nutrition,* Vol. 12, No. 7-8, (July 1996), pp.

supplementation on serum creatine kinase and lactate dehydrogenase after prolonged exercise. *Journal of Sports Medicine and Physical Fitness,* Vol. 40, No. 3,

Plasma lactate, GH and GH-binding protein levels in exercise following BCAA supplementation in athletes. *Amino Acids* Vol. 20, No. 1, (February 2001), pp. 1-11,

tissues by means of liquid chromatography. *Journal of Chromatography* Vol. 379,

in the skeletal muscle of patients with liver cirrhosis and changes under administration of glucose and branched-chain amino acids. *Tokai Journal of Experimental and Clinical Medicine.* Vol. 29, No. 4, (December 2004), pp. 191-198,

E. F. M. (2000a). Exercise-induced lactate increase in relation to muscle substrates in patients with chronic obstructive pulmonary disease. *American Journal of Respiratory and Critical Care Medicine,* Vol. 162, No. 5, (November 2000), pp. 1697-1704, ISSN

W. J. (2000b). Factors contribution to alterations in skeletal muscle and plasma amino acid profiles in patients with chronic obstructive pulmonary disease. *American Journal of Clinical Nutrition,* Vol. 72, No. 6, (December 2000), pp. 1480-1487,

*for the diagnosis, management, and prevention of chronic obstructive pulmonary disease,* 

Ueno, H. (2005). Effective dose of branched-chain amino acids on blood response in healthy men. *Journal of Japanese Society of Clinical Nutrition,* Vol. 27, No. 1, (January

nutritional status in advanced emphysema associated with chronic bronchitis. A

study of amino acid and catecholamine levels. *American Review of Respiratory Disease,* Vol. 141, No. 4 pt. 1, (April 1990), pp. 902-908, ISSN 00030805


Acute Effects of Branched-Chain Amino Acid Ingestion on Muscle pH

159, No. 6, (June 1999), pp. 1726-1734, ISSN 1073449X

No. 4, (August 1987), pp. 410-420, ISSN 01959131

(June 1993), pp. 290-296, ISSN 03044165

Verlag, ISBN 3-540-16213-5, Berlin Heidelberg

*and Medicine,* Vol. 1, No. 1, (1983), pp. 77-94, ISSN 07351313

00030805

00223166

ISSN 00029513

00652598

307-316, ISSN 10501606

pp. 159-167, ISSN 14396319

during Exercise in Patients with Chronic Obstructive Pulmonary Disease 139

Sapega, A. A., Sokolow, D. P., Graham, T. J., & Chance, B. (1987). Phosphorus nuclear

Schols, A. M. W. J., Soeters, P. B., Dingemans, A. M. C., Mostert, R., Franzen, P. J., &

Shimomura, Y., Fujii, H., Suzuki, M., Fujitsuka, N., Naoi, M., Sugiyama, S., & Harris, R. A.

Shimomura, Y., Fujii, H., Suzuki, M., Murakami, T., Fujitsuka, N., & Nakai, N. (1995).

Stevens, A. N. (1987). NMR Spectroscopy: Application to metabolic research. In: *Functional* 

Taylor, D. J., Bore, P. J., Styles, P., Gadian, D. G., & Radda, G. K. (1983). Bioenergetics of

Vandenborne, K., Walter, G., Leigh, J. S., & Goelman, G. (1993). pH heterogeneity during

Vukovich, M. D., Sharp, R. L., King, D. S., & Kershishnik, K. (1992). The effect of protein

Wagenmakers, A. J. M., Brookes, J. H., Coakley, J. H., Reilly, T., & Edwards, R. H. (1989).

Wagenmakers, A. J. M. (1998). Protein and amino acid metabolism in human muscle.

Wolfe, R. R., Goodenough, R. D., Wolfe, M. H., Royle, G. T., & Nodel, E. R. (1982).

pulmonary disease. *American Journal of Respiratory and Critical Care Medicine,* Vol.

magnetic resonance: A non-invasive technique for the study of muscle bioenergetics during exercise. *Medicine and Science in Sports and Exercise,* Vol. 19,

Wouters, E. F. M. (1993). Prevalence and characteristics of nutritional depletion in patients with stable COPD eligible for pulmonary rehabilitation. *American Review of Respiratory Disease,* Vol. 147, No. 5, (May 1993), pp. 1151-1156, ISSN

(1993). Branched-chain 2-oxo acid dehydrogenase complex activation by tetanic contractions in rat skeletal muscle. *Biochimica et Biophysica Acta.* Vol. 1157, No. 3,

Branched-chain α-keto acid dehydrogenase complex in rat skeletal muscle: Regulation of the activity and gene expression by nutrition and physical exercise *Journal of Nutrition,* Vol. 125, No. 6 SUPPL, (June 1995), pp. 1762S-1765S, ISSN

*studies using NMR*, McCready, V. R., Leach, M., & Ell, P. J., pp. 61-84, Springer-

intact human muscle. A 31P nuclear magnetic resonance study. *Molecular Biology* 

exercise in localized spectra from single human muscles. *American Journal of Physiology,* Vol. 265, No. 5 (Cell Physiology 34), (November 1993), pp. C1332-1339,

supplementation on lactate accumulation during submaximal and maximal exercise. *International Journal of Sport Nutrition,* Vol. 2, No. 4, (December, 1992), pp.

Exercise-induced activation of the branched-chain 2-oxo acid dehydrogenase in human muscle. *European Journal of Applied Physiology,* Vol. 59, No. 3, (October 1989),

*Advances in Experimental Medicine and Biology,* Vol. 441, (1998), pp. 307-319, ISSN

Isotopic analysis of leucine and urea metabolism in exercising humans. *Journal* 

by 31P-MRS. *Journal of Applied Physiology,* Vol. 78, No. 6, (June 1995), pp. 2218-2227, ISSN 87507587


Matsumoto, K., Koba, T., Hamada, K., Tsujimoto, H., & Mitsuzono, R. (2009). Branched-

Menier, R., Talmud, J., Laplaud, D., Bernard, M-P. (2001). Branched-chain aminoacids and

Nosaka, K., Sacco, P., & Mawatari, K. (2006). Effects of amino acid supplementation on

Payen, J. F., Wuyam, B., Reutenauer, P. L. H., Stiegliz, P., Paramelle, B., & Le Bas, J. F. (1993).

Pouw, M., Schols, A. M. W. J., Deutz, N. E. P., Wouters, E. F. M. (1998). Plasma and

Rutten, E. P. A., Engelen, M. P. K. J., Schols, A. M. W. J., & Deutz, N. E. P. (2005). Skeletal

Rutten, E. P. A., Engelen, M. P. K. J., Gosker, H., Bast, A., Cosemans, K., Vissers, Y.L.J.,

Saey, D., Michaud, A., Couillard, A., Cote, C. H., Mador, M. J., LeBlanc, P., Jobin, J, &

Sala, E., Roca, J., Marrades, R. M., Alonso, J., Gonzales De Suso, J. M., Moreno, A., Barbera, J.

*Metabolism,* Vol. 16, No. 6, (December 2006), pp. 620-635, ISSN 1526484X Ohashi, H., Sukegawa, E., Takami, T., Yoshida, T., & Muto, Y. (1989). Effects of

*Vitaminology.* Vol. 55, No. 1, (April 2009), pp. 52-58, ISSN 03014800

ISSN 87507587

00224707

ISSN 04466586

592-598, ISSN 00030805

801, ISSN 1073449X

408-415, ISSN 02615614

13631950

1073449X

by 31P-MRS. *Journal of Applied Physiology,* Vol. 78, No. 6, (June 1995), pp. 2218-2227,

chain amino acid supplementation increases the lactate threshold during an incremental exercise test in trained individuals. *Journal of Nutritional Science and* 

retraining of patients with chronic obstructive lung disease. *Journal of Sports Medicine and Physical Fitness,* Vol. 41, No. 4, (December 2001), pp. 500-504, ISSN

muscle soreness and damage. *International Journal of Sport Nutrition and Exercise* 

supplementation with branched-chain amino acids on protein-nutritional status in rats treated by carbon tetrachloride. *Japanese Journal of Gastroenterology,* Vol. 86, No. 8, (August 1989), pp. 1645-1653 (In Japanese with English Abstract),

Muscular metabolism during oxygen supplementation in patients with chronic hypoxemia. *American Review of Respiratory Disease,*Vol. 147, No. 3, (March 1993), pp.

muscle amino acid levels in relation to resting energy expenditure and inflammation in stable chronic obstructive pulmonary disease. *American Journal of Respiratory and Critical Care Medicine* Vol. 157, No. 2, (February 1998), pp. 797-

muscle glutamate metabolism in health and disease: State of art. *Current Opinion in Clinical Nutrition and Metabolic Care,* Vol. 8, No. 1, (January 2005), pp. 41-51, ISSN

Wouters, E. F. M., Deutz, N. E. P., & Schols, A. M. W. J. Metabolic and functional effects of glutamate intake in patients with chronic obstructive pulmonary disease (COPD). *Clinical Nutrition,* Vol. 27, No. 3, (June 2008), pp.

Maltais, F. (2005). Contractile fatigue, muscle morphometry, and blood lactate in chronic obstructive pulmonary disease. *American Journal of Respiratory and Critical Care Medicine,* Vol. 171, No. 10, (May 2005), pp. 1109-1115, ISSN

A., Nadal, J., DeJover, L., Rodriguez-Roisin, R., & Wagner, P. D. (1999) Effects of endurance training on skeletal muscle bioenergetics in chronic obstructive pulmonary disease. *American Journal of Respiratory and Critical Care Medicine,* Vol. 159, No. 6, (June 1999), pp. 1726-1734, ISSN 1073449X


**Part 2** 

**MRS Beyond the Clinic** 

*of Applied Physiology,* Vol. 52, No. 2, (February 1982), pp. 458-466, ISSN 87507587

Yoneda, T., Yoshikawa, M., Fu, A., Tsukaguchi, K., Okamoto, Y., & Takenaka, H. (2001). Plasma levels of amino acids and hypermetabolism in patients with chronic obstructive pulmonary disease. *Nutrition,* Vol. 17, No. 2, (February 2001), pp. 95-99, ISSN 08999007

**Part 2** 

**MRS Beyond the Clinic** 

140 Magnetic Resonance Spectroscopy

Yoneda, T., Yoshikawa, M., Fu, A., Tsukaguchi, K., Okamoto, Y., & Takenaka, H. (2001).

87507587

ISSN 08999007

*of Applied Physiology,* Vol. 52, No. 2, (February 1982), pp. 458-466, ISSN

Plasma levels of amino acids and hypermetabolism in patients with chronic obstructive pulmonary disease. *Nutrition,* Vol. 17, No. 2, (February 2001), pp. 95-99,

**8** 

*1France 2USA* 

**NMR Spectroscopy as a Tool to Provide** 

Benjamin Bourgeois1, Howard J. Worman2 and Sophie Zinn-Justin1

Nuclear magnetic resonance (NMR) has become an important technique for determining the three-dimensional (3D) structure of biological macromolecules. Since 1985, it has been used to determine the structures of approximately 8,000 proteins, 1,000 DNA/RNA complexes and 180 protein/nucleic acid complexes (http://www.pdb.org/pdb/ statistics/holdings.do). At first, NMR was limited to relatively small and soluble proteins or protein domains. The study of large proteins was hindered by the presence of overlapping large peaks in the NMR spectra. This has been in part alleviated by the introduction of isotope (2H, 15N, 13C) labeling and multidimensional (3D, 4D) experiments (Sattler et al., 1999). By using these techniques, it is now possible to study proteins up to 40 kDa. For instance, the 3D solution structure of the maltodextrin-binding protein (41kDa) has been recently solved using NMR (Madl et al., 2009; Figure 1A). Comprehensive NMR studies of integral membrane proteins in solution have long been impaired by substantial problems of sample preparation, including the inability to produce sufficient quantities of isotopically labelled protein as well as the difficulties associated with the limited thermal stability, sample heterogeneity and short lifetimes of such proteins. However, NMR allows investigation of the very conformational mobility that to a large extent interferes with the process of crystallization of membrane proteins. Thus, by focusing on proteins with a sufficient expression yield and screening sample and detergent conditions in a microtiter-plate format, it has been possible to determine 71 structures of integral membrane proteins corresponding to 49 unique proteins (http://www.drorlist.com/ nmr/MPNMR.html). The first NMR structure determination of a detergent-solubilized seven-helix transmembrane (7TM) protein, the phototaxis receptor sensory rhodopsin II (pSRII) from *Natronomonas pharaonis*, was recently reported to illustrate that NMR can provide structures of large membrane proteins (Gautier et al., 2010; Figure 1B). The challenge is now to apply similar techniques to the study of other 7TM proteins, including G-protein coupled receptors (GPCRs) that represent the most important class of targets for current therapeutic agents. These proteins can only be

**1. Introduction** 

*2Department of Medicine and Department of Pathology and Cell Biology, College of Physicians and Surgeons, Columbia University, New-York* 

**Mechanistic Clues About Protein** 

*1Laboratoire de Biologie Structurale et Radiobiologie, CEA Saclay & URA CNRS 2096, Gif-sur-Yvette* 

**Function and Disease Pathogenesis** 

### **NMR Spectroscopy as a Tool to Provide Mechanistic Clues About Protein Function and Disease Pathogenesis**

Benjamin Bourgeois1, Howard J. Worman2 and Sophie Zinn-Justin1 *1Laboratoire de Biologie Structurale et Radiobiologie, CEA Saclay & URA CNRS 2096, Gif-sur-Yvette 2Department of Medicine and Department of Pathology and Cell Biology, College of Physicians and Surgeons, Columbia University, New-York 1France 2USA* 

#### **1. Introduction**

Nuclear magnetic resonance (NMR) has become an important technique for determining the three-dimensional (3D) structure of biological macromolecules. Since 1985, it has been used to determine the structures of approximately 8,000 proteins, 1,000 DNA/RNA complexes and 180 protein/nucleic acid complexes (http://www.pdb.org/pdb/ statistics/holdings.do). At first, NMR was limited to relatively small and soluble proteins or protein domains. The study of large proteins was hindered by the presence of overlapping large peaks in the NMR spectra. This has been in part alleviated by the introduction of isotope (2H, 15N, 13C) labeling and multidimensional (3D, 4D) experiments (Sattler et al., 1999). By using these techniques, it is now possible to study proteins up to 40 kDa. For instance, the 3D solution structure of the maltodextrin-binding protein (41kDa) has been recently solved using NMR (Madl et al., 2009; Figure 1A). Comprehensive NMR studies of integral membrane proteins in solution have long been impaired by substantial problems of sample preparation, including the inability to produce sufficient quantities of isotopically labelled protein as well as the difficulties associated with the limited thermal stability, sample heterogeneity and short lifetimes of such proteins. However, NMR allows investigation of the very conformational mobility that to a large extent interferes with the process of crystallization of membrane proteins. Thus, by focusing on proteins with a sufficient expression yield and screening sample and detergent conditions in a microtiter-plate format, it has been possible to determine 71 structures of integral membrane proteins corresponding to 49 unique proteins (http://www.drorlist.com/ nmr/MPNMR.html). The first NMR structure determination of a detergent-solubilized seven-helix transmembrane (7TM) protein, the phototaxis receptor sensory rhodopsin II (pSRII) from *Natronomonas pharaonis*, was recently reported to illustrate that NMR can provide structures of large membrane proteins (Gautier et al., 2010; Figure 1B). The challenge is now to apply similar techniques to the study of other 7TM proteins, including G-protein coupled receptors (GPCRs) that represent the most important class of targets for current therapeutic agents. These proteins can only be

NMR Spectroscopy as a Tool to Provide Mechanistic

2KLR; Jehle et al., 2010).

A B C

Fig. 2. Ribbon representation of the solid-state NMR structures of **(A)** the M2

transmembrane peptide of the influenza A virus in DMPC lipid bilayers (PDB ID 2KQT; Cady et al., 2010), (B) the HET-s(218-289) prion in its amyloid form (PDB ID 2KJ2; Van Melckebeke et al., 2010) and (C) the -crystallin domain in B-crystallin oligomers (PDB ID

NMR can be used to study a large panel of molecular complexes, including those characterized by a low affinity binding interaction. At the present time, NMR has been used to solve about 1,500 3D structures of protein complexes. The majority of these structural studies have been performed on complexes characterized by submicromolar affinities. In cases of weak binding, characterized by a millimolar to micromolar affinity, it is possible to map and characterize details of the interface. The simplest and most popular approach is the so-called "chemical shift mapping" technique wherein changes in backbone chemical shifts of one molecule are monitored during titration with another molecule by recording a series of heteronuclear single-quantum correlation (HSQC) spectra (Figure 3). Upon addition of a ligand, chemical shift, line width and intensity changes of specific HSQC peaks indicate that the corresponding residues experience a modification of their chemical environment. This modification is due to the complex formation, and may result either from the proximity between the observed residue and the ligand or from binding-induced conformational changes in the first molecule close to the observed residue. The exact effects of the complex formation on the HSQC spectrum depend on the relative values of the NMR time scale and the exchange rate of the complex. The NMR time scale refers to the chemical shift scale: differences in chemical shifts between species in exchange. If this difference is larger than the exchange rate, then the exchange is classified as SLOW. If this difference is equal to the exchange rate, then the exchange is INTERMEDIATE. If this difference is smaller than the exchange rate, the exchange is FAST. Protein-protein interactions with Kd < 1μM (tight binding) are generally in slow exchange. In this case, two NMR signals are seen for the free and bound forms. During the titration, the free form signal gradually disappears and the bound form signal appears. Protein-protein interactions with Kd > 1mM (weak binding) are generally in fast exchange. In such cases, only one averaged signal is seen with a chemical shift fractionally weighted according to the populations and chemical shifts of the free and bound forms. Protein-protein interactions with a Kd between 1 μM and 1 mM are generally in intermediate exchange. This gives rise to broadened signals from both the free and bound forms. During addition of a ligand, identification of the NMR peak changes and mapping of these changes onto the protein structure provides information on the binding interface and the conformational changes of the protein due to the interaction (Figure 3). Analysis of these data opens up great opportunities for understanding the dynamics of interaction networks

Clues About Protein Function and Disease Pathogenesis 145

expressed in small quantities and are of limited stability, particularly in their unbound state. Samples must be conformationally and chemically homogeneous and stable at high concentrations. Despite these problems, there are encouraging indications suggesting that, in the near future, heterologous expression systems will provide labelled GPCRs in sufficient quantities for NMR analysis to characterize the details of the interactions between GPCRs and their ligands.

Fig. 1. Ribbon representation of the solution NMR structures of **(A)** the maltodextrinbinding protein (PDB ID 2KLF; Madl et al., 2009) and **(B)** the phototaxis receptor sensory rhodopsin II (pSRII) from *Natronomonas pharaonis* (PDB ID 2KSY; Gautier et al., 2010).

Most NMR studies of biological macromolecules have been carried out in solution. Since about 10 years ago, developments in NMR technology, sample preparation, pulse-sequence methodology and structure calculation protocols have allowed the structure determination of several proteins by solid-state NMR spectroscopy. Thus, 17 structures corresponding to 11 unique proteins have been solved using oriented samples of protein/lipids mixture (http://www.drorlist.com/ nmr/SPNMR.html). In particular, the M2 protein, a small membrane protein that enables hydrogen ions to enter the influenza A viral particle, was characterized by NMR to better understand its 3D structure and its interaction with drugs used against flu (Hu et al., 2010; Sharma et al., 2010; Figure 2A). Solid-state NMR was also used to solve 18 structures corresponding to 5 proteins forming fibrils (Van Melckebeke et al., 2010; Figure 2B) and 8 proteins yielding microcrystalline samples (Jehle et al., 2010; Figure 2C). Like liquid-state NMR, solid-state NMR is based on local structure information. As a consequence, large amounts of data on several differently labelled samples are typically required to obtain sufficient long-range information. However, NMR is particularly useful to study fibrils and microcrystals because these preparations are highly ordered on a local scale and consequently show narrow resonance lines. In recent years, considerable progress has been made in using solid-state NMR spectroscopy to determine atomic-resolution structures of amyloid fibrils associated with serious disorders such as Alzheimer disease, Parkinson disease, prion diseases and type 2 diabetes mellitus (for review: Tycko, 2011). Although the architecture of fibrils is considered to be a continuous stack of β-sheet ladders termed a cross-β structure, there may be significant variations in the supramolecular organization of the peptides within the fibrils. NMR provides detailed insight on the structural organization of these fibrils, highlighting their assembly process and inspiring the design of fibril-binding compounds and inhibitors.

expressed in small quantities and are of limited stability, particularly in their unbound state. Samples must be conformationally and chemically homogeneous and stable at high concentrations. Despite these problems, there are encouraging indications suggesting that, in the near future, heterologous expression systems will provide labelled GPCRs in sufficient quantities for NMR analysis to characterize the details of the interactions between

GPCRs and their ligands.

A B

and inspiring the design of fibril-binding compounds and inhibitors.

Fig. 1. Ribbon representation of the solution NMR structures of **(A)** the maltodextrinbinding protein (PDB ID 2KLF; Madl et al., 2009) and **(B)** the phototaxis receptor sensory rhodopsin II (pSRII) from *Natronomonas pharaonis* (PDB ID 2KSY; Gautier et al., 2010).

Most NMR studies of biological macromolecules have been carried out in solution. Since about 10 years ago, developments in NMR technology, sample preparation, pulse-sequence methodology and structure calculation protocols have allowed the structure determination of several proteins by solid-state NMR spectroscopy. Thus, 17 structures corresponding to 11 unique proteins have been solved using oriented samples of protein/lipids mixture (http://www.drorlist.com/ nmr/SPNMR.html). In particular, the M2 protein, a small membrane protein that enables hydrogen ions to enter the influenza A viral particle, was characterized by NMR to better understand its 3D structure and its interaction with drugs used against flu (Hu et al., 2010; Sharma et al., 2010; Figure 2A). Solid-state NMR was also used to solve 18 structures corresponding to 5 proteins forming fibrils (Van Melckebeke et al., 2010; Figure 2B) and 8 proteins yielding microcrystalline samples (Jehle et al., 2010; Figure 2C). Like liquid-state NMR, solid-state NMR is based on local structure information. As a consequence, large amounts of data on several differently labelled samples are typically required to obtain sufficient long-range information. However, NMR is particularly useful to study fibrils and microcrystals because these preparations are highly ordered on a local scale and consequently show narrow resonance lines. In recent years, considerable progress has been made in using solid-state NMR spectroscopy to determine atomic-resolution structures of amyloid fibrils associated with serious disorders such as Alzheimer disease, Parkinson disease, prion diseases and type 2 diabetes mellitus (for review: Tycko, 2011). Although the architecture of fibrils is considered to be a continuous stack of β-sheet ladders termed a cross-β structure, there may be significant variations in the supramolecular organization of the peptides within the fibrils. NMR provides detailed insight on the structural organization of these fibrils, highlighting their assembly process

Fig. 2. Ribbon representation of the solid-state NMR structures of **(A)** the M2 transmembrane peptide of the influenza A virus in DMPC lipid bilayers (PDB ID 2KQT; Cady et al., 2010), (B) the HET-s(218-289) prion in its amyloid form (PDB ID 2KJ2; Van Melckebeke et al., 2010) and (C) the -crystallin domain in B-crystallin oligomers (PDB ID 2KLR; Jehle et al., 2010).

NMR can be used to study a large panel of molecular complexes, including those characterized by a low affinity binding interaction. At the present time, NMR has been used to solve about 1,500 3D structures of protein complexes. The majority of these structural studies have been performed on complexes characterized by submicromolar affinities. In cases of weak binding, characterized by a millimolar to micromolar affinity, it is possible to map and characterize details of the interface. The simplest and most popular approach is the so-called "chemical shift mapping" technique wherein changes in backbone chemical shifts of one molecule are monitored during titration with another molecule by recording a series of heteronuclear single-quantum correlation (HSQC) spectra (Figure 3). Upon addition of a ligand, chemical shift, line width and intensity changes of specific HSQC peaks indicate that the corresponding residues experience a modification of their chemical environment. This modification is due to the complex formation, and may result either from the proximity between the observed residue and the ligand or from binding-induced conformational changes in the first molecule close to the observed residue. The exact effects of the complex formation on the HSQC spectrum depend on the relative values of the NMR time scale and the exchange rate of the complex. The NMR time scale refers to the chemical shift scale: differences in chemical shifts between species in exchange. If this difference is larger than the exchange rate, then the exchange is classified as SLOW. If this difference is equal to the exchange rate, then the exchange is INTERMEDIATE. If this difference is smaller than the exchange rate, the exchange is FAST. Protein-protein interactions with Kd < 1μM (tight binding) are generally in slow exchange. In this case, two NMR signals are seen for the free and bound forms. During the titration, the free form signal gradually disappears and the bound form signal appears. Protein-protein interactions with Kd > 1mM (weak binding) are generally in fast exchange. In such cases, only one averaged signal is seen with a chemical shift fractionally weighted according to the populations and chemical shifts of the free and bound forms. Protein-protein interactions with a Kd between 1 μM and 1 mM are generally in intermediate exchange. This gives rise to broadened signals from both the free and bound forms. During addition of a ligand, identification of the NMR peak changes and mapping of these changes onto the protein structure provides information on the binding interface and the conformational changes of the protein due to the interaction (Figure 3). Analysis of these data opens up great opportunities for understanding the dynamics of interaction networks

NMR Spectroscopy as a Tool to Provide Mechanistic

further exploited to follow biological events.

several proteins located at the nuclear envelope.

disease pathogenesis.

**2. NMR and genetic diseases involving the nuclear envelope** 

Clues About Protein Function and Disease Pathogenesis 147

(Dedmon et al., 2002). The 3D structure of the putative heavy-metal binding protein TTHA1718 from *Thermus thermophilus HB8* was solved from data acquired on living *E. coli*  samples (Sakakibara et al., 2009). In eukaryotic cells, the first in-cell NMR studies were performed by injecting proteins into *Xenopus laevis* oocytes (Selenko et al., 2006). Cellpenetrating peptides have also been used to deliver labelled proteins that can be observed in living cells (Inomata et al., 2009). However, the low sensitivity of the method and the short lifetime of the samples have so far prevented the acquisition of structural data in most of these eukaryotic systems. Development of ultra-fast methods for multidimensional experiments has already contributed to the first successes of in-cell NMR and should be

Solid-state and liquid-state NMR techniques are now essential tools for understanding biological processes from a mechanistic point of view at the atomic level. They can describe critical steps of a disease mechanism, highlight the mechanism of action of a particular drug and inspire the design of binding compounds and complex assembly inhibitors. We will now focus on the use of NMR for the elucidation of the molecular mechanisms of specific genetic diseases caused by either missense mutations or amino acid deletions/insertions in

We have used NMR to study the molecular mechanisms of genetic diseases caused by mutations in genes encoding inner nuclear membrane proteins (see for examples Krimm et al., 2002 and Caputo et al., 2006). These proteins play critical roles in nuclear structure and positioning (Figure 4). They also influence genome spatial and temporal organization and genome functional properties. Because of their many biological functions, the pathological mechanisms resulting from mutations are complex and not well understood (Dauer & Worman, 2009). A wide range of diseases affecting different organs system have been linked to mutations in the *LMNA* gene, encoding the A-type nuclear lamins, which are intermediate filament proteins mainly located at the inner nuclear membrane of differentiated somatic cells. These diseases include dilated cardiomyopathy with variable muscular dystrophy, Dunnigan-type familial partial lipodystrophy, a Charcot-Marie-Tooth type 2 disease, mandibuloacral dysplasia and Hutchinson-Gilford progeria syndrome. Adult-onset autosomal dominant leukodystrophy is caused by duplication of the gene encoding lamin B1, which is an intermediate filament protein located at the inner nuclear membrane of all nucleated somatic mammalian cells. In addition, several diseases are linked to mutations in genes encoding integral proteins of the inner nuclear membrane that are associated with nuclear lamins, such as emerin, LBR and MAN1. Structural studies of the proteins involved in these so-called "laminopathies" or "nuclear envelopathies" have provided mechanistic clues about the functions of the nuclear envelope and insights into the

Some mutations causing laminopathies are frameshifts or premature termination codons resulting in truncated protein. In these cases, the protein is generally rapidly degraded and its function is lost. NMR can however be used to structurally characterize the native protein, thus providing molecular details on its lost function (Caputo et al., 2006). Others mutations generate amino acid substitutions that can destabilize the protein leading to its premature degradation, interfere with its degradation pathway or modify its interaction with other

(Mackereth et al., 2011). Which molecules are in competition for binding to the studied protein? Which molecules bind simultaneously? Is a conformational change necessary for the complex formation? Is this conformation change involved in the regulation of other interactions? How multiple domain proteins exploit their flexibility in order to recognize their targets?

Fig. 3. Mapping of binding sites by NMR. **(A)** In the presence of free and ligand-bound proteins, the NMR signal varies as a function of the exchange rate. Its chemical shifts and linewidths evolve from a FAST exchange to a SLOW exchange condition**.** When the exchange is FAST, only one averaged chemical shift is observed. When it is SLOW, the chemical shifts corresponding to the free and bound forms are observed. **(B)** The chemical shift changes measured during the titration of a labelled protein with its binding partner are displayed in the case of FAST exchange (upper panel) and SLOW exchange (lower panel). **(C)** Residues whose NMR chemical shifts and/or linewidths change during the titration with ligands A and B are colored in red and orange, respectively, on the surface representation of the 3D structure of the labelled protein. This "chemical shift mapping" study shows that the labelled protein recognizes its two ligands by different surfaces.

NMR allows for direct observations of NMR-active nuclei (1H, 15N, 13C, 31P) within any NMR-inactive environment and can therefore be employed to investigate the structures of appropriately labelled biomolecules either in cell extracts or inside live cells. It has been used to structurally characterize biomolecules in the presence of cell extracts, in order to investigate post-translational modifications, conformational changes and binding events at the residue level within macromolecules in conditions mimicking the cellular environment (Liotakis et al., 2010). To accomplish this, NMR-active labelled recombinant proteins are injected into cell extracts and their chemical shift evolution is followed in the extracts after addition of specific enzyme co-factors or binding partners. These experiments are particularly powerful at describing the molecular binding events regulated by posttranslational modifications. Furthermore, NMR has been applied to the observation of protein-protein interactions and the determination of protein structures in *E. coli*. The socalled STINT-NMR approach entails sequentially expressing two (or more) proteins within a single bacterial cell in a time-controlled manner and monitoring the protein interactions using in-cell NMR spectroscopy (Burz et al., 2006). NMR has also been used to follow the behavior of intrinsically disordered proteins in the crowded environment of bacterial cells

(Mackereth et al., 2011). Which molecules are in competition for binding to the studied protein? Which molecules bind simultaneously? Is a conformational change necessary for the complex formation? Is this conformation change involved in the regulation of other interactions? How multiple domain proteins exploit their flexibility in order to recognize

A B C

with ligands A and B are colored in red and orange, respectively, on the surface

representation of the 3D structure of the labelled protein. This "chemical shift mapping" study shows that the labelled protein recognizes its two ligands by different surfaces.

NMR allows for direct observations of NMR-active nuclei (1H, 15N, 13C, 31P) within any NMR-inactive environment and can therefore be employed to investigate the structures of appropriately labelled biomolecules either in cell extracts or inside live cells. It has been used to structurally characterize biomolecules in the presence of cell extracts, in order to investigate post-translational modifications, conformational changes and binding events at the residue level within macromolecules in conditions mimicking the cellular environment (Liotakis et al., 2010). To accomplish this, NMR-active labelled recombinant proteins are injected into cell extracts and their chemical shift evolution is followed in the extracts after addition of specific enzyme co-factors or binding partners. These experiments are particularly powerful at describing the molecular binding events regulated by posttranslational modifications. Furthermore, NMR has been applied to the observation of protein-protein interactions and the determination of protein structures in *E. coli*. The socalled STINT-NMR approach entails sequentially expressing two (or more) proteins within a single bacterial cell in a time-controlled manner and monitoring the protein interactions using in-cell NMR spectroscopy (Burz et al., 2006). NMR has also been used to follow the behavior of intrinsically disordered proteins in the crowded environment of bacterial cells

Fig. 3. Mapping of binding sites by NMR. **(A)** In the presence of free and ligand-bound proteins, the NMR signal varies as a function of the exchange rate. Its chemical shifts and linewidths evolve from a FAST exchange to a SLOW exchange condition**.** When the exchange is FAST, only one averaged chemical shift is observed. When it is SLOW, the chemical shifts corresponding to the free and bound forms are observed. **(B)** The chemical shift changes measured during the titration of a labelled protein with its binding partner are displayed in the case of FAST exchange (upper panel) and SLOW exchange (lower panel). **(C)** Residues whose NMR chemical shifts and/or linewidths change during the titration

their targets?

(Dedmon et al., 2002). The 3D structure of the putative heavy-metal binding protein TTHA1718 from *Thermus thermophilus HB8* was solved from data acquired on living *E. coli*  samples (Sakakibara et al., 2009). In eukaryotic cells, the first in-cell NMR studies were performed by injecting proteins into *Xenopus laevis* oocytes (Selenko et al., 2006). Cellpenetrating peptides have also been used to deliver labelled proteins that can be observed in living cells (Inomata et al., 2009). However, the low sensitivity of the method and the short lifetime of the samples have so far prevented the acquisition of structural data in most of these eukaryotic systems. Development of ultra-fast methods for multidimensional experiments has already contributed to the first successes of in-cell NMR and should be further exploited to follow biological events.

Solid-state and liquid-state NMR techniques are now essential tools for understanding biological processes from a mechanistic point of view at the atomic level. They can describe critical steps of a disease mechanism, highlight the mechanism of action of a particular drug and inspire the design of binding compounds and complex assembly inhibitors. We will now focus on the use of NMR for the elucidation of the molecular mechanisms of specific genetic diseases caused by either missense mutations or amino acid deletions/insertions in several proteins located at the nuclear envelope.

#### **2. NMR and genetic diseases involving the nuclear envelope**

We have used NMR to study the molecular mechanisms of genetic diseases caused by mutations in genes encoding inner nuclear membrane proteins (see for examples Krimm et al., 2002 and Caputo et al., 2006). These proteins play critical roles in nuclear structure and positioning (Figure 4). They also influence genome spatial and temporal organization and genome functional properties. Because of their many biological functions, the pathological mechanisms resulting from mutations are complex and not well understood (Dauer & Worman, 2009). A wide range of diseases affecting different organs system have been linked to mutations in the *LMNA* gene, encoding the A-type nuclear lamins, which are intermediate filament proteins mainly located at the inner nuclear membrane of differentiated somatic cells. These diseases include dilated cardiomyopathy with variable muscular dystrophy, Dunnigan-type familial partial lipodystrophy, a Charcot-Marie-Tooth type 2 disease, mandibuloacral dysplasia and Hutchinson-Gilford progeria syndrome. Adult-onset autosomal dominant leukodystrophy is caused by duplication of the gene encoding lamin B1, which is an intermediate filament protein located at the inner nuclear membrane of all nucleated somatic mammalian cells. In addition, several diseases are linked to mutations in genes encoding integral proteins of the inner nuclear membrane that are associated with nuclear lamins, such as emerin, LBR and MAN1. Structural studies of the proteins involved in these so-called "laminopathies" or "nuclear envelopathies" have provided mechanistic clues about the functions of the nuclear envelope and insights into the disease pathogenesis.

Some mutations causing laminopathies are frameshifts or premature termination codons resulting in truncated protein. In these cases, the protein is generally rapidly degraded and its function is lost. NMR can however be used to structurally characterize the native protein, thus providing molecular details on its lost function (Caputo et al., 2006). Others mutations generate amino acid substitutions that can destabilize the protein leading to its premature degradation, interfere with its degradation pathway or modify its interaction with other

NMR Spectroscopy as a Tool to Provide Mechanistic

complexes involving these proteins.

Clues About Protein Function and Disease Pathogenesis 149

conformations and fold into different structures on binding to different target proteins. The pathological consequences of mutations located in the poorly-folded regions would then not be observed on the 3D structure of the mutated protein but instead linked to modifications of its binding properties. To tackle the structural description of these proteins, we chose to (1) solve the 3D solution structures of their globular domains and (2) characterize the relative positioning of domains within sub-regions. This "divide-and-conquer" approach is based on the combination of NMR spectroscopy and small-angle X-ray scattering (SAXS) and is a powerful method to characterize the structural ensemble of partially disordered proteins (Madl et al., 2011). From the description of the three-dimensional structures of subregions of A-type lamins and LEM domain proteins, we then (1) determined the impact of disease-linked mutations on the structure and stability of the proteins, (2) mapped interaction surfaces of proteins with their biological partners, (3) positioned the mutations relatively to these functionally important surfaces and (4) characterized the 3D structure of

**3. How do mutations in A-type lamins cause different diseases?** 

precursor that is processed to lamin A, and from amino acids 567-572 for lamin C.

assemblies associate through lateral interactions to form intermediate filaments.

X-ray crystallography has been used to determine the atomic structure of various fragments of the lamin intermediate filament rod domain (Herrmann et al., 2009). In particular, it has been used to solve the crystal structure of a human lamin A fragment comprising residues 305 – 387, which corresponds to the second half of coil2 (PDB ID 1X8Y; Strelkov et al., 2004). This structure revealed a left-handed parallel coiled-coil extending to the predicted end of the rod domain (Figure 5). Essentially the same fragment of lamin B1 has also been resolved and a similar structure was observed (PDB ID 3MOV). Interestingly, the crystal structure of

Nuclear lamins are classified as either A-type or B-type according to homology in sequence, biochemical properties and localization during the cell cycle. In humans, the *LMNB1* and *LMNB2* genes encode lamin B1 and B2, respectively, which are expressed in most or all nucleated somatic cells. The *LMNA* gene encodes the A-type lamins and the major isoforms, lamin A and lamin C, are expressed in most differentiated somatic cells. Lamins A and C are produced by alternative RNA splicing. They share a common region from amino acids 1-566 and differ at their C-terminus, from amino acids 567-664 for prelamin A, the lamin A

Like all intermediate filament proteins, lamins are fibrous in nature and share a tripartite structural organization. A non-α-helical N-terminal region (head) and a C-terminal region (tail) flank a central α-helical coiled-coil domain (rod). Lamins differ from the cytoplasmic intermediate filament proteins in that they have an extended rod domain (42 amino acid longer), that they have a nuclear localization signal (NLS) between their rod and tail regions and that they display a typical tertiary structure in this tail region (Herrmann et al., 2009; Figure 5). In contrast to most cytoplasmic intermediate filament proteins, lamins also contain sites for phosphorylation by mitotic kinases and most contain carboxy-terminal CAAX motifs that signal posttranslational modification by farnesylation. The lamin -helical central rod domain is divided into three coiled regions (termed 1A, 1B and 2), which mediate lamin assembly into filaments. *In vitro* experiments have revealed that elementary lamin dimers first assemble through longitudinal N–C interactions and then these

**3.1 Architecture of the lamin intermediate filaments** 

proteins or nucleic acids. In these cases, NMR can provide information on the structural consequences of the mutation (Krimm et al., 2002). Does the mutation affect a hydrophobic residue in the core of the protein and thus destabilize its 3D structure? Does it affect a residue at the surface of the protein that is critical for a specific interaction with a biological partner? Structural studies utilizing NMR, by identifying the structural consequences of the missense mutations, can point out specific disease mechanisms. Such information could be used to develop therapies targeted to the defects.

Fig. 4. Architecture of the nuclear envelope (reproduced with permission from Chi et al., 2009). The nuclear lamina, comprising the A- and B-type lamins, interacts with several proteins anchored at the inner nuclear membrane such as LBR, emerin, MAN1 and SUN1/2. SUN1/2 proteins also interact with outer nuclear membrane proteins that in turn bind to microtubules and actin. This network of protein-protein interactions is essential for cell structural organization.

We have used NMR to study A-type lamins, showing how different types of mutations that cause different diseases affect protein structure, hence providing clues about pathogenesis. We have also used NMR to study LEM domain proteins, which mediate interactions between the inner nuclear membrane and DNA and are involved in bone and muscle diseases. We have further analysed in detail MAN1, a LEM domain protein that regulates a key signal transduction pathway and whose heterozygous loss of function causes sclerosing bone dysplasias.

Many proteins associated with cellular control and signalling mechanisms such as A-type lamins and LEM domain proteins are modular in structure in that they contain both wellfolded domains and poorly-structured regions (Wright & Dyson, 2009). The presence of poorly-folded regions facilitates protein accessibility to modifying enzymes and interaction with a wide variety of targets. However, this also hinders protein structural characterization, as the poorly-structured regions can adopt a large number of

proteins or nucleic acids. In these cases, NMR can provide information on the structural consequences of the mutation (Krimm et al., 2002). Does the mutation affect a hydrophobic residue in the core of the protein and thus destabilize its 3D structure? Does it affect a residue at the surface of the protein that is critical for a specific interaction with a biological partner? Structural studies utilizing NMR, by identifying the structural consequences of the missense mutations, can point out specific disease mechanisms. Such information could be

Fig. 4. Architecture of the nuclear envelope (reproduced with permission from Chi et al., 2009). The nuclear lamina, comprising the A- and B-type lamins, interacts with several proteins anchored at the inner nuclear membrane such as LBR, emerin, MAN1 and SUN1/2. SUN1/2 proteins also interact with outer nuclear membrane proteins that in turn bind to microtubules and actin. This network of protein-protein interactions is essential for cell

We have used NMR to study A-type lamins, showing how different types of mutations that cause different diseases affect protein structure, hence providing clues about pathogenesis. We have also used NMR to study LEM domain proteins, which mediate interactions between the inner nuclear membrane and DNA and are involved in bone and muscle diseases. We have further analysed in detail MAN1, a LEM domain protein that regulates a key signal transduction pathway and whose heterozygous loss of function causes sclerosing

Many proteins associated with cellular control and signalling mechanisms such as A-type lamins and LEM domain proteins are modular in structure in that they contain both wellfolded domains and poorly-structured regions (Wright & Dyson, 2009). The presence of poorly-folded regions facilitates protein accessibility to modifying enzymes and interaction with a wide variety of targets. However, this also hinders protein structural characterization, as the poorly-structured regions can adopt a large number of

used to develop therapies targeted to the defects.

structural organization.

bone dysplasias.

conformations and fold into different structures on binding to different target proteins. The pathological consequences of mutations located in the poorly-folded regions would then not be observed on the 3D structure of the mutated protein but instead linked to modifications of its binding properties. To tackle the structural description of these proteins, we chose to (1) solve the 3D solution structures of their globular domains and (2) characterize the relative positioning of domains within sub-regions. This "divide-and-conquer" approach is based on the combination of NMR spectroscopy and small-angle X-ray scattering (SAXS) and is a powerful method to characterize the structural ensemble of partially disordered proteins (Madl et al., 2011). From the description of the three-dimensional structures of subregions of A-type lamins and LEM domain proteins, we then (1) determined the impact of disease-linked mutations on the structure and stability of the proteins, (2) mapped interaction surfaces of proteins with their biological partners, (3) positioned the mutations relatively to these functionally important surfaces and (4) characterized the 3D structure of complexes involving these proteins.

### **3. How do mutations in A-type lamins cause different diseases?**

#### **3.1 Architecture of the lamin intermediate filaments**

Nuclear lamins are classified as either A-type or B-type according to homology in sequence, biochemical properties and localization during the cell cycle. In humans, the *LMNB1* and *LMNB2* genes encode lamin B1 and B2, respectively, which are expressed in most or all nucleated somatic cells. The *LMNA* gene encodes the A-type lamins and the major isoforms, lamin A and lamin C, are expressed in most differentiated somatic cells. Lamins A and C are produced by alternative RNA splicing. They share a common region from amino acids 1-566 and differ at their C-terminus, from amino acids 567-664 for prelamin A, the lamin A precursor that is processed to lamin A, and from amino acids 567-572 for lamin C.

Like all intermediate filament proteins, lamins are fibrous in nature and share a tripartite structural organization. A non-α-helical N-terminal region (head) and a C-terminal region (tail) flank a central α-helical coiled-coil domain (rod). Lamins differ from the cytoplasmic intermediate filament proteins in that they have an extended rod domain (42 amino acid longer), that they have a nuclear localization signal (NLS) between their rod and tail regions and that they display a typical tertiary structure in this tail region (Herrmann et al., 2009; Figure 5). In contrast to most cytoplasmic intermediate filament proteins, lamins also contain sites for phosphorylation by mitotic kinases and most contain carboxy-terminal CAAX motifs that signal posttranslational modification by farnesylation. The lamin -helical central rod domain is divided into three coiled regions (termed 1A, 1B and 2), which mediate lamin assembly into filaments. *In vitro* experiments have revealed that elementary lamin dimers first assemble through longitudinal N–C interactions and then these assemblies associate through lateral interactions to form intermediate filaments.

X-ray crystallography has been used to determine the atomic structure of various fragments of the lamin intermediate filament rod domain (Herrmann et al., 2009). In particular, it has been used to solve the crystal structure of a human lamin A fragment comprising residues 305 – 387, which corresponds to the second half of coil2 (PDB ID 1X8Y; Strelkov et al., 2004). This structure revealed a left-handed parallel coiled-coil extending to the predicted end of the rod domain (Figure 5). Essentially the same fragment of lamin B1 has also been resolved and a similar structure was observed (PDB ID 3MOV). Interestingly, the crystal structure of

NMR Spectroscopy as a Tool to Provide Mechanistic

2002; PDB IDs 1IFR and 1UFG).

from NMR experimental data.

**3.2 Post-translational modification of A-type lamins** 

Clues About Protein Function and Disease Pathogenesis 151

essential to assign the inter-proton distances obtained by 1H NMR to specific proton pairs. Molecular modeling calculations were used to calculate a family of 3D structures, all very close one to the other, compatible with the distances measured by NMR (Figure 6B). These structures, which reflect the structure of the domain in solution, have an immunoglobulin fold between residues 430 and 545. Crystal structures of this C-terminal domain of A-type lamins have also been solved and show similar structural organization (Dhe-Paganon et al.,

Fig. 6. Structural characterization of the globular domain found in the tail region of lamins. (A) 1H-15N HSQC spectrum of the domain from human A-type lamin. This spectrum was acquired on a 40 M protein sample in 250 l phosphate buffer pH 7.4 over 1 hour using a 700 MHz Bruker spectrometer. Each residue of the domain is revealed by a peak whose position depends on the residue chemical environment. Well-folded domains exhibit welldispersed peaks, as shown for the lamin tail domain. (B) Cartoon representation of the 3D structure of the fragment of amino acids 428-549 from human A-type lamin, as calculated

Lamins undergo post-translational modifications that alter their localization, assembly and binding properties. Prelamin A is converted into lamin A by four enzymatic posttranslational processing steps: farnesylation of a cysteine in a CAAX motif at the carboxyl terminus of the protein, endoproteolytic cleavage of the last three amino acids of the protein, carboxyl methylation of the newly exposed farnesylcysteine and subsequent endoproteolytic release of the last 15 amino acids, including the farnesylcysteine methyl ester. B-type lamins also contain a CAAX motif and undergo the first three of these enzymatic reactions but not the last. A-type and B-type lamins are also phosphorylated by cyclin dependent kinases, which regulate nuclear lamina assembly and disassembly during mitosis. Other phosphorylation, acetylation and sumoylation events have also been reported whose physiological importance is largely unknown. According to the PhosphoSite database, more than 60 phosphorylation sites are present in human A-type

another human lamin A fragment (residues 328–398), which is largely overlapping with fragment 305–387 but harbours a short segment of the tail domain, was also solved. Unexpectedly, there is no parallel coiled-coil form within the crystal. Instead, the α-helices are arranged such that two antiparallel coiled-coil interfaces are formed (PDB ID 2XV5). The most significant interface has a right-handed geometry, which results from a characteristic 15-residue repeat pattern that overlays the canonical heptad repeat pattern (Figure 5). Analysis of these different modes of coil association gives some clues about the mechanisms of lamin polymerization in cells (Kapinos et al., 2011). Several polymerization mechanisms have been proposed that could co-exist and be modulated by specific chaperone-type molecules as well as post-translational modifications.

Fig. 5. Architecture of lamins, adapted with permission from Kapinos et al., 2011. **(A)** General structural organization of lamin monomers. Two A-type lamin coil 2 fragments were studied at atomic resolution (PDB IDs 1X8Y and 2XV5). Their X-ray structures are displayed in the cartoon, in which the sequence of amino acids 328-382, common to both fragments, is colored in gray and the remaining residues colored blue (N-terminus) or red (C-terminus). **(B)** Two models of lamin filament structural organization proposed from the analysis of the rod fragment X-ray structures. The C-terminal rod sections of two dimers labelled 1 and 3 are colored blue and the N-terminal rod section of dimer 2 is colored green. Broken lines represent coiled-coil interfaces that have been observed in crystal structures and dotted lines indicate other theoretically possible coiled coil interactions.

We have solved the 3D solution structure of the C-terminal globular domain of A-type lamins, which participates to the recognition of various lamin binding proteins (PDB ID 1IVT; Krimm et al., 2002). Determination of this structure was made from experimental measured values of short inter-proton distances (<6 Å) obtained by 1H, 15N and 13C NMR on a 13C and 15N labelled sample of the domain. Figure 6A shows the well-dispersed 1H-15N HSQC spectrum of the C-terminal globular domain of A-type lamins. The labeling was

another human lamin A fragment (residues 328–398), which is largely overlapping with fragment 305–387 but harbours a short segment of the tail domain, was also solved. Unexpectedly, there is no parallel coiled-coil form within the crystal. Instead, the α-helices are arranged such that two antiparallel coiled-coil interfaces are formed (PDB ID 2XV5). The most significant interface has a right-handed geometry, which results from a characteristic 15-residue repeat pattern that overlays the canonical heptad repeat pattern (Figure 5). Analysis of these different modes of coil association gives some clues about the mechanisms of lamin polymerization in cells (Kapinos et al., 2011). Several polymerization mechanisms have been proposed that could co-exist and be modulated by specific chaperone-type

Fig. 5. Architecture of lamins, adapted with permission from Kapinos et al., 2011. **(A)** General structural organization of lamin monomers. Two A-type lamin coil 2 fragments were studied at atomic resolution (PDB IDs 1X8Y and 2XV5). Their X-ray structures are displayed in the cartoon, in which the sequence of amino acids 328-382, common to both fragments, is colored in gray and the remaining residues colored blue (N-terminus) or red (C-terminus). **(B)** Two models of lamin filament structural organization proposed from the analysis of the rod fragment X-ray structures. The C-terminal rod sections of two dimers labelled 1 and 3 are colored blue and the N-terminal rod section of dimer 2 is colored green. Broken lines represent coiled-coil interfaces that have been observed in crystal structures

**Coil 2** assembly with **coil 1**

**Nter**

**Nter** 

and dotted lines indicate other theoretically possible coiled coil interactions.

We have solved the 3D solution structure of the C-terminal globular domain of A-type lamins, which participates to the recognition of various lamin binding proteins (PDB ID 1IVT; Krimm et al., 2002). Determination of this structure was made from experimental measured values of short inter-proton distances (<6 Å) obtained by 1H, 15N and 13C NMR on a 13C and 15N labelled sample of the domain. Figure 6A shows the well-dispersed 1H-15N HSQC spectrum of the C-terminal globular domain of A-type lamins. The labeling was

molecules as well as post-translational modifications.

**Model 1**

**Model 2** 

**Nter** 

**A** 

**Nter** 

**B**

essential to assign the inter-proton distances obtained by 1H NMR to specific proton pairs. Molecular modeling calculations were used to calculate a family of 3D structures, all very close one to the other, compatible with the distances measured by NMR (Figure 6B). These structures, which reflect the structure of the domain in solution, have an immunoglobulin fold between residues 430 and 545. Crystal structures of this C-terminal domain of A-type lamins have also been solved and show similar structural organization (Dhe-Paganon et al., 2002; PDB IDs 1IFR and 1UFG).

Fig. 6. Structural characterization of the globular domain found in the tail region of lamins. (A) 1H-15N HSQC spectrum of the domain from human A-type lamin. This spectrum was acquired on a 40 M protein sample in 250 l phosphate buffer pH 7.4 over 1 hour using a 700 MHz Bruker spectrometer. Each residue of the domain is revealed by a peak whose position depends on the residue chemical environment. Well-folded domains exhibit welldispersed peaks, as shown for the lamin tail domain. (B) Cartoon representation of the 3D structure of the fragment of amino acids 428-549 from human A-type lamin, as calculated from NMR experimental data.

#### **3.2 Post-translational modification of A-type lamins**

Lamins undergo post-translational modifications that alter their localization, assembly and binding properties. Prelamin A is converted into lamin A by four enzymatic posttranslational processing steps: farnesylation of a cysteine in a CAAX motif at the carboxyl terminus of the protein, endoproteolytic cleavage of the last three amino acids of the protein, carboxyl methylation of the newly exposed farnesylcysteine and subsequent endoproteolytic release of the last 15 amino acids, including the farnesylcysteine methyl ester. B-type lamins also contain a CAAX motif and undergo the first three of these enzymatic reactions but not the last. A-type and B-type lamins are also phosphorylated by cyclin dependent kinases, which regulate nuclear lamina assembly and disassembly during mitosis. Other phosphorylation, acetylation and sumoylation events have also been reported whose physiological importance is largely unknown. According to the PhosphoSite database, more than 60 phosphorylation sites are present in human A-type

NMR Spectroscopy as a Tool to Provide Mechanistic

Clues About Protein Function and Disease Pathogenesis 153

adipose tissue diseases do not destabilize the domain structure but may hinder the interaction of A-type lamin with a binding partner that plays a critical role in adipose tissue. Recessive mutations in the immunoglobulin-like domain of A-type lamins causing progeroid disorders also affect residues located at the surface of this domain and cluster at a specific site (Figure 8C; Verstraeten et al., 2006). This hot spot is distinct from the surface affected in partial lipodystrophy and is conserved in A-type lamins. It may correspond to another protein binding site specific to A-type lamins, with mutations causing progeroid disorders hindering this putative interaction. Experiments to identify the proteins interacting with A-type lamins at the different identified hot spots are now underway.

A B C

residues mutated in progeria-like diseases are displayed in green sticks.

**protein structure and interactions** 

complexes (Gonzales et al., 2008).

Fig. 8. Localization of the residues identified as mutated in laminopathies in the 3D structure of the tail globular domain of A-type lamins (PDB ID 1IVT; Krimm et al., 2002; Verstraeten et al., 2006). In **(A)**, residues mutated in striated muscle diseases are displayed in blue sticks.

In **(B),** residues mutated in adipose tissue diseases are displayed in red sticks. In **(C),**

**3.4 Perspectives: NMR study of how lamin post-translational modifications affect** 

Several studies have shown that the lamin tail globular domain and the flexible lamin regions are involved in binding to other proteins. We know that flexible regions of lamins are highly modified in cells (Figure 7). This suggests that post-translational modifications could be involved in regulating lamin recognition. In fact, post-translational modification of partners has already been shown to be involved in regulating lamin binding. For example, lamin A is involved in the sequestration of c-Fos, a member of the dimeric activator protein 1 (AP-1) transcription factor family that regulates different cellular pathways including cell proliferation, death survival differentiation and oncogenic transformation. Phosphorylation of c-Fos by ERK1/2 protein kinase reduces lamin A/c-Fos interaction, thus activating AP1

In-cell NMR is a useful way to analyse these regulation pathways involving posttranslational modifications of lamins or lamin partners. Using NMR chemical shift mapping as presented in part 1, modified residues can be identified onto the lamin tail or onto its binding partners. Next, interaction between modified or unmodified lamins and their binding partners, and conversely between lamins and modified or unmodified partners, can

lamins. Most of these are located in unstructured regions of the proteins (Figure 7). These phosphoresidues could be important for lamin functions by regulating interactions with other nuclear proteins.

Fig. 7. Phosphorylation sites in human A-type lamins that have been reported in the PhosphoSite database (http://www.phosphosite.org/)**.**

Post-translational modifications can be followed by NMR, either by adding the appropriate enzyme to the labelled protein sample or by diluting the labelled protein into cell extracts and adding the appropriate co-factor (Liotakis et al., JACS 2010). Series of NMR spectra are then recorded to follow the protein chemical shift evolution with time. Residues whose chemical shifts are modified because of the presence of enzymes in the sample are identified. These residues are either the targets of the enzyme or are close to the modified residues. NMR is particularly efficient in dissecting the modification kinetics of protein regions rich in inter-dependant modification sites.

#### **3.3 Localization and structural impact of disease-linked missense mutations identified in A-type lamins**

Several different diseases are caused by mutations in the *LMNA* gene encoding lamin A and C (Worman et al., 2010). More than 1,500 mutations have been identified within this gene. Amino acid substitutions in various regions of A-type lamins cause striated muscle diseases, partial lipodystrophy or progeroid disorders but in most cases a correlation between the localization of the mutation and the disease phenotype is not readily apparent. We searched for a correlation between the localization of the mutated residue in the lamin 3D structure and the disease phenotype. We focused on the lamin tail globular domain, because amino acid substitutions in this region cause either striated muscle diseases, lipodystrophy or progeroid disorders. Analysis of the position of the mutated residues in the NMR structure of the immunoglobulin-like domain and experimental measurement of the impact of several selected mutations on this structure were carried out (Krimm et al., 2002). We observed that most dominantly inherited mutations causing diseases of striated muscle affect residues of the hydrophobic core of the immunoglobulin-like domain (Figure 8A). These mutations probably destabilize the 3D structure of this domain, leading to an overall loss of protein function. Such destabilization has been experimentally verified for the Arg453Trp mutant. In contrast, dominantly inherited mutations causing Dunnigan-type familial partial lipodystrophy, a disease affecting adipose tissues, affect residues located at the surface of the immunoglobulin-like domain and cluster at a specific positively charged site (Figure 8B). This site is conserved in A-type and B-type lamins. This suggests that mutations causing

lamins. Most of these are located in unstructured regions of the proteins (Figure 7). These phosphoresidues could be important for lamin functions by regulating interactions with

Fig. 7. Phosphorylation sites in human A-type lamins that have been reported in the

Post-translational modifications can be followed by NMR, either by adding the appropriate enzyme to the labelled protein sample or by diluting the labelled protein into cell extracts and adding the appropriate co-factor (Liotakis et al., JACS 2010). Series of NMR spectra are then recorded to follow the protein chemical shift evolution with time. Residues whose chemical shifts are modified because of the presence of enzymes in the sample are identified. These residues are either the targets of the enzyme or are close to the modified residues. NMR is particularly efficient in dissecting the modification kinetics of protein

**3.3 Localization and structural impact of disease-linked missense mutations identified** 

Several different diseases are caused by mutations in the *LMNA* gene encoding lamin A and C (Worman et al., 2010). More than 1,500 mutations have been identified within this gene. Amino acid substitutions in various regions of A-type lamins cause striated muscle diseases, partial lipodystrophy or progeroid disorders but in most cases a correlation between the localization of the mutation and the disease phenotype is not readily apparent. We searched for a correlation between the localization of the mutated residue in the lamin 3D structure and the disease phenotype. We focused on the lamin tail globular domain, because amino acid substitutions in this region cause either striated muscle diseases, lipodystrophy or progeroid disorders. Analysis of the position of the mutated residues in the NMR structure of the immunoglobulin-like domain and experimental measurement of the impact of several selected mutations on this structure were carried out (Krimm et al., 2002). We observed that most dominantly inherited mutations causing diseases of striated muscle affect residues of the hydrophobic core of the immunoglobulin-like domain (Figure 8A). These mutations probably destabilize the 3D structure of this domain, leading to an overall loss of protein function. Such destabilization has been experimentally verified for the Arg453Trp mutant. In contrast, dominantly inherited mutations causing Dunnigan-type familial partial lipodystrophy, a disease affecting adipose tissues, affect residues located at the surface of the immunoglobulin-like domain and cluster at a specific positively charged site (Figure 8B). This site is conserved in A-type and B-type lamins. This suggests that mutations causing

PhosphoSite database (http://www.phosphosite.org/)**.**

regions rich in inter-dependant modification sites.

other nuclear proteins.

**in A-type lamins** 

adipose tissue diseases do not destabilize the domain structure but may hinder the interaction of A-type lamin with a binding partner that plays a critical role in adipose tissue. Recessive mutations in the immunoglobulin-like domain of A-type lamins causing progeroid disorders also affect residues located at the surface of this domain and cluster at a specific site (Figure 8C; Verstraeten et al., 2006). This hot spot is distinct from the surface affected in partial lipodystrophy and is conserved in A-type lamins. It may correspond to another protein binding site specific to A-type lamins, with mutations causing progeroid disorders hindering this putative interaction. Experiments to identify the proteins interacting with A-type lamins at the different identified hot spots are now underway.

Fig. 8. Localization of the residues identified as mutated in laminopathies in the 3D structure of the tail globular domain of A-type lamins (PDB ID 1IVT; Krimm et al., 2002; Verstraeten et al., 2006). In **(A)**, residues mutated in striated muscle diseases are displayed in blue sticks. In **(B),** residues mutated in adipose tissue diseases are displayed in red sticks. In **(C),** residues mutated in progeria-like diseases are displayed in green sticks.

#### **3.4 Perspectives: NMR study of how lamin post-translational modifications affect protein structure and interactions**

Several studies have shown that the lamin tail globular domain and the flexible lamin regions are involved in binding to other proteins. We know that flexible regions of lamins are highly modified in cells (Figure 7). This suggests that post-translational modifications could be involved in regulating lamin recognition. In fact, post-translational modification of partners has already been shown to be involved in regulating lamin binding. For example, lamin A is involved in the sequestration of c-Fos, a member of the dimeric activator protein 1 (AP-1) transcription factor family that regulates different cellular pathways including cell proliferation, death survival differentiation and oncogenic transformation. Phosphorylation of c-Fos by ERK1/2 protein kinase reduces lamin A/c-Fos interaction, thus activating AP1 complexes (Gonzales et al., 2008).

In-cell NMR is a useful way to analyse these regulation pathways involving posttranslational modifications of lamins or lamin partners. Using NMR chemical shift mapping as presented in part 1, modified residues can be identified onto the lamin tail or onto its binding partners. Next, interaction between modified or unmodified lamins and their binding partners, and conversely between lamins and modified or unmodified partners, can

NMR Spectroscopy as a Tool to Provide Mechanistic

residue flexible linker (Cai et al., 2001).

al., 2001).

interphase, or both.

Clues About Protein Function and Disease Pathogenesis 155

residues (Figure 10). NMR characterization of the fragment from amino acids 1-168 of LAP2 containing both the LEM-like and LEM domains has shown that the two domains are structurally independent, non-interacting domains connected by an approximately 60-

Nter Cter

A B

**4.2 LEM domains recognize the DNA binding BAF protein** 

Fig. 10. Schematic organization of nuclear envelope proteins containing a LEM domain. **(A)** Three human inner nuclear membrane proteins are displayed: LAP2 β, MAN1 and emerin. Dark blue boxes represent LEM domains, the light blue box represents a LEM-like domain, green rectangles represent transmembrane segments. **(B)** Ribbon representation of the average solution structure of the LEM domain of LAP2 (PDB ID 1H9F; Laguri et

Barrier-to-autointegration factor (BAF) is a highly conserved metazoan protein, which functions in nuclear assembly, chromatin organization and gene transcription. BAF dimers bind to double-stranded DNA non-specifically and thereby bridge DNA molecules to form a large, discrete nucleoprotein complex. BAF is able to interact simultaneously with both LAP2 and DNA *in vitro* (Shumaker et al., 2001). This suggested that LEM proteins and BAF mediate chromatin attachment to the nuclear envelope during nuclear assembly or

The interaction between the fragment containing amino acids 1-168 of LAP2 and the BAF– DNA complex was characterized by NMR (Cai et al., 2001). It was proposed on the basis of chemical shift mapping experiments that the LEM-like domain recognizes DNA (Figure 11A) while the LEM domain recognizes the DNA binding protein BAF (Figure 11B). However, when the whole BAF-DNA complex was added to the labelled LAP2 fragment, only the peaks corresponding to the LEM domain shifted, indicating that only the LEM

Subsequently, the solution structure of the BAF-emerin LEM domain complex was solved on the basis of NMR data (Cai et al., 2007). This structure revealed that one BAF dimer interacts with one emerin LEM domain (Figure 12A). The stoechiometry of the complex prevents association between one molecule of BAF dimer and multiple LEM domain proteins. Moreover, the interaction surface between BAF and emerin does not overlap with the two symmetry related DNA binding sites on the BAF dimer (Figure 12B). Thus, BAF can

domain interacts with the BAF–DNA nucleoprotein complex (Figure 11C).

simultaneously interact with emerin and DNA.

be tested. Our strategy is to produce recombinant lamins and partners and inject them into human cell extracts with or without addition of cofactors that are essential for posttranslational-modifications. Figure 9 shows our first 1H-15N HSQC spectra of the tail region of lamin C incubated into 293T cell extracts during 1 to 24 hours. This experiment shows that it is possible to follow the fate of a partially unfolded protein in cell extracts during one day.

Fig. 9. Following modification and binding events involving the lamin C tail in 293T cell extracts: feasibility experiments. 1H-15N HSQC spectra were recorded on a Bruker 700 MHz spectrometer after 1 hour (red) and 24 hours (blue) of incubation of a 50 M sample of labelled protein in 250 l of cell extracts at 5mg/ml. The protein is stable in these conditions. Co-factors should then be added in order to allow modification of the labelled protein.

#### **4. How do mutations in LEM domain proteins affect interactions between the nuclear envelope and chromatin?**

#### **4.1 Architecture of LEM domain proteins**

LAP2 isoform β (LAP2 β), emerin and MAN1 are human nuclear proteins that share a highly conserved structural motif called the LEM domain (Lin et al., 2000). The LEM domain contains approximately 50 amino acids. It is located at the N-terminus of these proteins and is often separated from the rest of the protein by predicted unstructured regions (Figure 10). Both MAN1 and emerin possess one copy of the LEM domain in their N-terminal nucleoplasmic region whereas LAP2 contains a LEM-like domain and a LEM domain connected by a predicted unstructured linker.

The 3D solution structures of emerin and LAP2 LEM domains and LAP2 LEM-like domain have been solved using NMR (Wolff et al., 2001; Laguri et al., 2001; Cai et al., 2001). These LEM and LEM-like domains share a similar 3D structure, mainly composed of a short N-terminal helix and two large α-helices interacting through a set of conserved amino acids

be tested. Our strategy is to produce recombinant lamins and partners and inject them into human cell extracts with or without addition of cofactors that are essential for posttranslational-modifications. Figure 9 shows our first 1H-15N HSQC spectra of the tail region of lamin C incubated into 293T cell extracts during 1 to 24 hours. This experiment shows that it is possible to follow the fate of a partially unfolded protein in cell extracts during one day.

Fig. 9. Following modification and binding events involving the lamin C tail in 293T cell extracts: feasibility experiments. 1H-15N HSQC spectra were recorded on a Bruker 700 MHz spectrometer after 1 hour (red) and 24 hours (blue) of incubation of a 50 M sample of labelled protein in 250 l of cell extracts at 5mg/ml. The protein is stable in these conditions. Co-factors should then be added in order to allow modification of the labelled protein.

**4. How do mutations in LEM domain proteins affect interactions between the** 

LAP2 isoform β (LAP2 β), emerin and MAN1 are human nuclear proteins that share a highly conserved structural motif called the LEM domain (Lin et al., 2000). The LEM domain contains approximately 50 amino acids. It is located at the N-terminus of these proteins and is often separated from the rest of the protein by predicted unstructured regions (Figure 10). Both MAN1 and emerin possess one copy of the LEM domain in their N-terminal nucleoplasmic region whereas LAP2 contains a LEM-like domain and a LEM domain

The 3D solution structures of emerin and LAP2 LEM domains and LAP2 LEM-like domain have been solved using NMR (Wolff et al., 2001; Laguri et al., 2001; Cai et al., 2001). These LEM and LEM-like domains share a similar 3D structure, mainly composed of a short N-terminal helix and two large α-helices interacting through a set of conserved amino acids

**nuclear envelope and chromatin? 4.1 Architecture of LEM domain proteins** 

connected by a predicted unstructured linker.

residues (Figure 10). NMR characterization of the fragment from amino acids 1-168 of LAP2 containing both the LEM-like and LEM domains has shown that the two domains are structurally independent, non-interacting domains connected by an approximately 60 residue flexible linker (Cai et al., 2001).

Fig. 10. Schematic organization of nuclear envelope proteins containing a LEM domain. **(A)** Three human inner nuclear membrane proteins are displayed: LAP2 β, MAN1 and emerin. Dark blue boxes represent LEM domains, the light blue box represents a LEM-like domain, green rectangles represent transmembrane segments. **(B)** Ribbon representation of the average solution structure of the LEM domain of LAP2 (PDB ID 1H9F; Laguri et al., 2001).

#### **4.2 LEM domains recognize the DNA binding BAF protein**

Barrier-to-autointegration factor (BAF) is a highly conserved metazoan protein, which functions in nuclear assembly, chromatin organization and gene transcription. BAF dimers bind to double-stranded DNA non-specifically and thereby bridge DNA molecules to form a large, discrete nucleoprotein complex. BAF is able to interact simultaneously with both LAP2 and DNA *in vitro* (Shumaker et al., 2001). This suggested that LEM proteins and BAF mediate chromatin attachment to the nuclear envelope during nuclear assembly or interphase, or both.

The interaction between the fragment containing amino acids 1-168 of LAP2 and the BAF– DNA complex was characterized by NMR (Cai et al., 2001). It was proposed on the basis of chemical shift mapping experiments that the LEM-like domain recognizes DNA (Figure 11A) while the LEM domain recognizes the DNA binding protein BAF (Figure 11B). However, when the whole BAF-DNA complex was added to the labelled LAP2 fragment, only the peaks corresponding to the LEM domain shifted, indicating that only the LEM domain interacts with the BAF–DNA nucleoprotein complex (Figure 11C).

Subsequently, the solution structure of the BAF-emerin LEM domain complex was solved on the basis of NMR data (Cai et al., 2007). This structure revealed that one BAF dimer interacts with one emerin LEM domain (Figure 12A). The stoechiometry of the complex prevents association between one molecule of BAF dimer and multiple LEM domain proteins. Moreover, the interaction surface between BAF and emerin does not overlap with the two symmetry related DNA binding sites on the BAF dimer (Figure 12B). Thus, BAF can simultaneously interact with emerin and DNA.

NMR Spectroscopy as a Tool to Provide Mechanistic

A B

**and BAF** 

2003; Frock el al., 2006).

Clues About Protein Function and Disease Pathogenesis 157

Fig. 12. Ribbon representation of **(A)** the solution structure of the BAF2 dimer (orange) bound to the emerin LEM domain (red) (PDB ID 2ODG; Cai et al., 2007) and **(B)** the X-ray structure of

**4.3 Localization and structural impact of disease-linked mutations identified in emerin** 

X-linked Emery-Dreyfus muscular dystrophy (EDMD) is characterized by early contractures of the spine, Achilles tendon and elbows, muscle wasting in a humerperoneal distribution and dilated cardiomyopathy with conduction system abnormalities (Muchir & Worman, 2007). It is caused by mutations in the *EMD* gene encoding emerin (Bione et al., 1994), an integral protein of the inner nuclear membrane. At present, 130 mutations homogeneously distributed along the *EMD* gene have been reported (Brown et al., 2011). Approximately 90% of these mutations result in the complete absence of protein as characterized by immunohistochemical staining. The few mutations allowing modified emerin protein expression yield protein that is expressed in reduced amounts and is frequently mislocalized. The majority of patients with residual emerin expression display clinical phenotypes indistinguishable from their null counterparts. However, it is difficult to argue that there is no discernable genotype-phenotype correlation, as clinical data collection on EDMD patients is often incomplete. Moreover, the relationship between emerin absence or decreased expression and muscle abnormalities is poorly understood. As emerin has been hypothesized to regulate muscle specific gene expression and nuclear architecture, some have suggested that absence or reduced levels affect cellular susceptibility to mechanical stress, cellular proliferation and differentiation (Fidziańska & Hausmanowa-Petrusewicz,

Recently, a mutation (A12T) in the *BANF1* gene encoding BAF has been shown to induce a progeroid syndrome that shares physiopathological characteristics with Hutchinson-Gilford progeria syndrome (HGPS) (Puente et al., 2011). HPGS has received considerable attention because of its striking premature aging phenotype, including alopecia, diminished subcutaneous fat, premature atherosclerosis, and skeletal abnormalities. The vast majority of HGPS cases are associated with a de novo nucleotide substitution at position 1824 (C->T) in the *LMNA* gene encoding A-type lamins (Eriksson et al., 2003; De Sandre-Giovannoli et al., 2003). This mutation does not affect the coded amino acid (and is thus generally referred to as G608G), but partially activates a cryptic splice donor site in exon 11 of *LMNA*, leading to

the BAF2 dimer (orange) bound to DNA (blue) (PDB ID 2BZF; Bradley et al., 2005).

Fig. 11. Mapping the LAP2 residues involved in BAF–DNA nucleoprotein recognition (reproduced with permission from Cai et al., 2001). (**A**) Comparison of the 1H-15N HSQC spectra of LAP2 region containing amino acids 1-168 in the absence (red) and presence of 1.5 equivalents (blue) of a duplex DNA dodecamer. Only cross-peaks of the LEM-like domain shift upon DNA binding. (**B**) Comparison of 1H-15N HSQC spectra of LAP2 region 1-168 in the absence (red) and presence (blue) of one equivalent of BAF. Only the cross-peaks in the LEM domain change significantly upon binding to BAF. (**C**) Comparison of the 1H-15N HSQC spectra of LAP2 region containing amino acids 1-168 in the absence (red) and presence (blue) of the BAF–DNA nucleoprotein complex. Only the peaks originating from the LEM domain disappear upon binding the BAF–DNA nucleoprotein complex.

Fig. 11. Mapping the LAP2 residues involved in BAF–DNA nucleoprotein recognition (reproduced with permission from Cai et al., 2001). (**A**) Comparison of the 1H-15N HSQC spectra of LAP2 region containing amino acids 1-168 in the absence (red) and presence of 1.5 equivalents (blue) of a duplex DNA dodecamer. Only cross-peaks of the LEM-like domain shift upon DNA binding. (**B**) Comparison of 1H-15N HSQC spectra of LAP2 region 1-168 in the absence (red) and presence (blue) of one equivalent of BAF. Only the

Comparison of the 1H-15N HSQC spectra of LAP2 region containing amino acids 1-168 in the absence (red) and presence (blue) of the BAF–DNA nucleoprotein complex. Only the peaks originating from the LEM domain disappear upon binding the BAF–DNA

cross-peaks in the LEM domain change significantly upon binding to BAF. (**C**)

nucleoprotein complex.

Fig. 12. Ribbon representation of **(A)** the solution structure of the BAF2 dimer (orange) bound to the emerin LEM domain (red) (PDB ID 2ODG; Cai et al., 2007) and **(B)** the X-ray structure of the BAF2 dimer (orange) bound to DNA (blue) (PDB ID 2BZF; Bradley et al., 2005).

#### **4.3 Localization and structural impact of disease-linked mutations identified in emerin and BAF**

X-linked Emery-Dreyfus muscular dystrophy (EDMD) is characterized by early contractures of the spine, Achilles tendon and elbows, muscle wasting in a humerperoneal distribution and dilated cardiomyopathy with conduction system abnormalities (Muchir & Worman, 2007). It is caused by mutations in the *EMD* gene encoding emerin (Bione et al., 1994), an integral protein of the inner nuclear membrane. At present, 130 mutations homogeneously distributed along the *EMD* gene have been reported (Brown et al., 2011). Approximately 90% of these mutations result in the complete absence of protein as characterized by immunohistochemical staining. The few mutations allowing modified emerin protein expression yield protein that is expressed in reduced amounts and is frequently mislocalized. The majority of patients with residual emerin expression display clinical phenotypes indistinguishable from their null counterparts. However, it is difficult to argue that there is no discernable genotype-phenotype correlation, as clinical data collection on EDMD patients is often incomplete. Moreover, the relationship between emerin absence or decreased expression and muscle abnormalities is poorly understood. As emerin has been hypothesized to regulate muscle specific gene expression and nuclear architecture, some have suggested that absence or reduced levels affect cellular susceptibility to mechanical stress, cellular proliferation and differentiation (Fidziańska & Hausmanowa-Petrusewicz, 2003; Frock el al., 2006).

Recently, a mutation (A12T) in the *BANF1* gene encoding BAF has been shown to induce a progeroid syndrome that shares physiopathological characteristics with Hutchinson-Gilford progeria syndrome (HGPS) (Puente et al., 2011). HPGS has received considerable attention because of its striking premature aging phenotype, including alopecia, diminished subcutaneous fat, premature atherosclerosis, and skeletal abnormalities. The vast majority of HGPS cases are associated with a de novo nucleotide substitution at position 1824 (C->T) in the *LMNA* gene encoding A-type lamins (Eriksson et al., 2003; De Sandre-Giovannoli et al., 2003). This mutation does not affect the coded amino acid (and is thus generally referred to as G608G), but partially activates a cryptic splice donor site in exon 11 of *LMNA*, leading to

NMR Spectroscopy as a Tool to Provide Mechanistic

(Mansharamani & Wilson, 2005).

Clues About Protein Function and Disease Pathogenesis 159

(Figure 14A). Its N-terminal and the C-terminal regions are located on the nucleoplasmic side of the membrane. The N-terminal nucleoplasmic region of MAN1 (amino acids 1-471) contains an N-terminal LEM (Figure 14B). This domain is essential for emerin, lamin and BAF binding (Liu et al., 2003; Mansharamani & Wilson, 2005). We have used 1H, 15N and 13C NMR and SAXS to study the structural organization of the C-terminal nucleoplasmic region of MAN1 (Caputo et al., 2006; Kondé et al., 2010). This region is composed of a globular winged helix (WH) domain (Figure 14C), a linker region, a globular U2AF Homology Motif (UHM) domain (Figure 14D) and a C-terminal unstructured region. The WH domain presents a fold found in several transcription factors and it recognizes DNA (Caputo et al., 2006). The UHM domain is essential for MAN1 binding to Smad2 and Smad3 that are transcriptional regulators of the TGF pathway (Lin et al., 2005; Pan et al., 2005). The Cterminal region also recognizes other transcription regulators as the death-promoting transcriptional repressor Btf and the transcriptional repressor germ-cell-less GCL

Fig. 14. Structural organization of the MAN1 protein. **(A)** Domain composition of MAN1 (the two transmembrane regions are colored in green) and localization of the mutations causing bone genetics diseases (Hellemans et al., 2004). **(B)** Model of the LEM domain of MAN1, built by homology on the basis of the LAP2 LEM domain structure (PDB ID 1H9F; Laguri et al., 2001). **(C)** NMR structure of the WH domain (PDB ID 1CH0; Caputo et al., 2006). **(D)** Model calculated for the UHM domain on the basis of the secondary structure information deduced from the NMR chemical shift analysis and the 3D structure of the

To understand the mechanisms of regulation of the TGFβ pathway by MAN1, we characterized the molecular determinants of its interaction with Smad2 and Smad3. We showed that the first WH domain is not essential for Smad2 binding. However, the linker region as well as the UHM domain and the C-terminus are all involved in Smad2 binding. NMR and SAXS data were used to reconstitute the fluctuating 3D structure of the Smad2

UHM domain of U2AF65 (PDB ID 1O0P; Kondé et al., 2010).

binding region of MAN1 (Kondé et al., 2010).

the production of a prelamin A mRNA that contains an internal deletion of 150 base pairs. This transcript is then translated into a protein known as progerin, which lacks 50 amino acids near the C terminus. The new progeroid syndrome caused by a mutation in BAF differs from HGPS by the fact that affected patients have not cardiovascular deficiencies and live longer. A 3D model of BAF shows that the mutated residue is located on the surface of the protein (Figure 13). Nevertheless, the A12T mutation is not predicted to affect BAF dimerization or binding to either DNA or emerin, raising the possibility that this amino acid substitution could impair the interaction with other proteins, its subcellular localization or its stability. Examination of the effect of the BAF A12T mutation in the fibroblasts of affected patients showed a reduction in protein levels, indicating that this mutation affects the stability of BAF. Surprisingly, the decrease in BAF appears to be linked to a delocalization of emerin in the endoplasmic reticulum and abnormalities in the nuclear lamina. In HeLa cells that overexpressed a phosphomimetic BAF missense mutant S4E, but not S4A, emerin also mislocalized from the nuclear envelope (Bengtsson & Wilson, 2005). Furthermore, phosphorylation of serine 4 inhibited BAF binding to emerin and lamin A in blot overlay assays. These results suggest that the N-terminal region of BAF is normally involved in emerin localization at the nuclear envelope and that the A12T mutation could destabilize a lamin-emerin-BAF complex. Structural data on this complex are lacking so further structural analysis of the consequences of the BAF A12T mutation is not currently possible.

Fig. 13. 3D model of a BAF dimer (orange and light orange in the figure) in complex with emerin (red) and DNA (cyan) built from the PDB files 2ODG and 2BZF. The position of the T12 residue in mutant BAF is indicated in pink sticks (Puente et al., 2011).

#### **5. How do mutations in the inner nuclear envelope MAN1 deregulate TGF signaling?**

#### **5.1 Architecture of MAN1, a LEM domain protein with an R-Smad binding C-terminal region**

The *LEMD3* gene encodes MAN1, an integral protein of the inner nuclear membrane (Lin et al., 2000). MAN1 consists of 911 amino acids and exhibits two transmembrane domains

the production of a prelamin A mRNA that contains an internal deletion of 150 base pairs. This transcript is then translated into a protein known as progerin, which lacks 50 amino acids near the C terminus. The new progeroid syndrome caused by a mutation in BAF differs from HGPS by the fact that affected patients have not cardiovascular deficiencies and live longer. A 3D model of BAF shows that the mutated residue is located on the surface of the protein (Figure 13). Nevertheless, the A12T mutation is not predicted to affect BAF dimerization or binding to either DNA or emerin, raising the possibility that this amino acid substitution could impair the interaction with other proteins, its subcellular localization or its stability. Examination of the effect of the BAF A12T mutation in the fibroblasts of affected patients showed a reduction in protein levels, indicating that this mutation affects the stability of BAF. Surprisingly, the decrease in BAF appears to be linked to a delocalization of emerin in the endoplasmic reticulum and abnormalities in the nuclear lamina. In HeLa cells that overexpressed a phosphomimetic BAF missense mutant S4E, but not S4A, emerin also mislocalized from the nuclear envelope (Bengtsson & Wilson, 2005). Furthermore, phosphorylation of serine 4 inhibited BAF binding to emerin and lamin A in blot overlay assays. These results suggest that the N-terminal region of BAF is normally involved in emerin localization at the nuclear envelope and that the A12T mutation could destabilize a lamin-emerin-BAF complex. Structural data on this complex are lacking so further structural

analysis of the consequences of the BAF A12T mutation is not currently possible.

Fig. 13. 3D model of a BAF dimer (orange and light orange in the figure) in complex with emerin (red) and DNA (cyan) built from the PDB files 2ODG and 2BZF. The position of the

**5. How do mutations in the inner nuclear envelope MAN1 deregulate TGF**

**5.1 Architecture of MAN1, a LEM domain protein with an R-Smad binding C-terminal** 

The *LEMD3* gene encodes MAN1, an integral protein of the inner nuclear membrane (Lin et al., 2000). MAN1 consists of 911 amino acids and exhibits two transmembrane domains

T12 residue in mutant BAF is indicated in pink sticks (Puente et al., 2011).

**signaling?** 

**region** 

(Figure 14A). Its N-terminal and the C-terminal regions are located on the nucleoplasmic side of the membrane. The N-terminal nucleoplasmic region of MAN1 (amino acids 1-471) contains an N-terminal LEM (Figure 14B). This domain is essential for emerin, lamin and BAF binding (Liu et al., 2003; Mansharamani & Wilson, 2005). We have used 1H, 15N and 13C NMR and SAXS to study the structural organization of the C-terminal nucleoplasmic region of MAN1 (Caputo et al., 2006; Kondé et al., 2010). This region is composed of a globular winged helix (WH) domain (Figure 14C), a linker region, a globular U2AF Homology Motif (UHM) domain (Figure 14D) and a C-terminal unstructured region. The WH domain presents a fold found in several transcription factors and it recognizes DNA (Caputo et al., 2006). The UHM domain is essential for MAN1 binding to Smad2 and Smad3 that are transcriptional regulators of the TGF pathway (Lin et al., 2005; Pan et al., 2005). The Cterminal region also recognizes other transcription regulators as the death-promoting transcriptional repressor Btf and the transcriptional repressor germ-cell-less GCL (Mansharamani & Wilson, 2005).

Fig. 14. Structural organization of the MAN1 protein. **(A)** Domain composition of MAN1 (the two transmembrane regions are colored in green) and localization of the mutations causing bone genetics diseases (Hellemans et al., 2004). **(B)** Model of the LEM domain of MAN1, built by homology on the basis of the LAP2 LEM domain structure (PDB ID 1H9F; Laguri et al., 2001). **(C)** NMR structure of the WH domain (PDB ID 1CH0; Caputo et al., 2006). **(D)** Model calculated for the UHM domain on the basis of the secondary structure information deduced from the NMR chemical shift analysis and the 3D structure of the UHM domain of U2AF65 (PDB ID 1O0P; Kondé et al., 2010).

To understand the mechanisms of regulation of the TGFβ pathway by MAN1, we characterized the molecular determinants of its interaction with Smad2 and Smad3. We showed that the first WH domain is not essential for Smad2 binding. However, the linker region as well as the UHM domain and the C-terminus are all involved in Smad2 binding. NMR and SAXS data were used to reconstitute the fluctuating 3D structure of the Smad2 binding region of MAN1 (Kondé et al., 2010).

NMR Spectroscopy as a Tool to Provide Mechanistic

Clues About Protein Function and Disease Pathogenesis 161

We mutated the highly conserved W765 and Q766 of the linker region and observed the consequences of these mutations on the NMR 1H-15N spectrum of the Smad2 binding region (Figure 15A). Residues with NMR signals modified by the mutations were mapped onto the UHM domain (Figure 15B). They all clustered on one face of the domain and delineated a hydrophobic pocket in which the tryptophane residue binds. We performed SAXS experiments on samples of wild type and mutated Smad2 binding region. SAXS provides accurate information on folding and conformation in solution for both rigid and flexible macromolecules and the indirect Fourier transform of the SAXS intensity gives access to the real space electron pair distance distribution of molecules in solution. Figure 15C (upper panel) shows that the SAXS curves of the wild type and mutated protein fragments are similar, reflecting their common global fold. Calculation of the electron pair distance distributions using GNOM gave access to the radius of gyration Rg, the maximal distance and the I0 value that cannot be measured experimentally and is proportional to the mass. From these values, a Kratky representation was displayed (Figure 15C, lower panel) that highlighted structural differences between wild-type and mutated proteins: the maximum of the curve corresponding to (SRg)2 I/Io is shifted towards higher SRg values, reflecting the more elongated/less structured state of the mutant. Moreover, the wild type protein shows a gyration radius between 20 and 21 Å and a maximal distance between 68 and 72 Å, whereas the mutant shows a gyration radius between 23.3 and 24 Å and a maximal distance is close to 90 Å. This result validates our previous NMR based analysis that strongly suggests a role of W765 and Q766 in anchoring the linker onto the UHM domain. In contrast, the C-terminus of the Smad2 binding region of MAN1 is partially unfolded. From these observations, the positions of the linker and C-terminus regions relatively to the UHM domain were calculated, yielding to a model of the whole Smad2 binding region of MAN1 (Figure 15D). The functional importance of the positioning of the linker region onto the UHM domain was shown using fluorescence binding experiments: mutations of the W765 and Q766 residues into alanine

yielded a 15-fold decrease of the affinity of MAN1 for Smad2 (Figure 15E).

result from enhanced TGF-ß signalling in cells.

**MAN1** 

**5.2 Mechanism of bone diseases caused by the absence of the C-terminal region of** 

Heterozygous loss-of-function mutations in *LEMD3* encoding MAN1 cause the sclerosing bone dysplasias osteopoikilosis, Buschke-Ollendorff syndrome and non-sporadic melorheostosis (Hellemans et al., 2004). Several arguments suggest that the bone diseases

First, all the disease-causing mutations including nonsense or frameshift mutations induce loss of the C-terminal nucleoplasmic fragment of MAN1, which is essential for Smad2 recognition (Figure 14A). Second, the TGF-β pathway is enhanced in cells from patients (Hellemans et al., 2004). Loss of the MAN1-Smad2 interaction may lead to the release of activated Smad2 and Smad3, thus causing an increase of Smad2 dependant transcription (Lin et al., 2005; Pan et al., 2005). It has been proposed that MAN1-Smad2 interaction not only sequestrates Smad2 and Smad3 at the inner nuclear envelope, but also favours their nuclear export (Pan et al., 2005; Figure 16). However, the mechanism by which the nucleoplasmic MAN1 C-terminal region would stimulate this nuclear export remains unclear. MAN1 could simply act as a "nuclear envelope sink" for activated nuclear Smad proteins, sequestering them at the inner nuclear membrane (Figure 16-1) and competitively inhibiting interactions with other nuclear factors involved in gene regulation (Figure 16-2).

Fig. 15. Structural organization of the Smad2 binding motif of MAN1 (adapted with permission from Kondé et al., 2010). **(A)** NMR 1H-15N HSQC spectra of the wild-type (yellow) and mutated (W765A-Q766A, green) Smad2 binding motif of MAN1, representing the NMR finger print of the fold adopted by these protein fragments. Each NMR peak is assigned to the backbone NH bond of a specific amino acid in MAN1 sequence, and its position in the spectrum reflects its chemical environment in the protein domain. **(B)** NMR chemical shift mapping of the UHM region interacting with the linker. On the black ribbon representation of the MAN1 UHM model, residues whose 1H-15N HSQC peak is largely shifted and could not be assigned due to the mutations W765A and Q766A are colored in red, residues whose HSQC peak is shifted by more than 1.5 times the root mean square deviation of the chemical shift difference distribution but could still be assigned are coloured in orange, and residues whose HSQC peak could not be assigned in the MAN1LuhmWQ spectra are colored in cyan. The surface of the opposite face of the UHM domain is showed as an additional view. **(C)** Comparison of the SAXS data recorded on the wild-type (blue) and mutated (W765A-Q766A, red) Smad2 binding motif of MAN1. The upper view displays the superposition of the logarithm of the SAXS intensities as a function of the amplitude of the diffusion vector q = 4\*π\*sin(θ)/λ. The lower view corresponds to the superposition of the Kratky plots: the diffusion vector amplitude multiplied by the gyration radius is plotted on the abscissa and the square of the abscissa multiplied by the SAXS intensity divided by the intensity at 0 is plotted on the ordinate. This representation reveals the type of structure: compact, partially folded or unfolded. **(D)** Model of the Smad2 binding domain of MAN1 consistent with the NMR and SAXS data. The backbone is represented as a ribbon; the side chains of W765 and Q766, in the linker, and of several residues lining the hydrophobic cavity of the UHM domain are displayed as sticks. **(E)** Fluorescence polarization curves of the labelled wild-type (left) and mutated (W765A-Q766A, right) Smad2 binding fragment of MAN1 as a function of the concentration in Smad2.

[Smad2] (M) [Smad2] (M)

A B C

15N chemical shift (in ppm)

Fig. 15. Structural organization of the Smad2 binding motif of MAN1 (adapted with permission from Kondé et al., 2010). **(A)** NMR 1H-15N HSQC spectra of the wild-type (yellow) and mutated (W765A-Q766A, green) Smad2 binding motif of MAN1, representing the NMR finger print of the fold adopted by these protein fragments. Each NMR peak is assigned to the backbone NH bond of a specific amino acid in MAN1 sequence, and its position in the spectrum reflects its chemical environment in the protein domain. **(B)** NMR chemical shift mapping of the UHM region interacting with the linker. On the black ribbon representation of the MAN1 UHM model, residues whose 1H-15N HSQC peak is largely shifted and could not be assigned due to the mutations W765A and Q766A are colored in red, residues whose HSQC peak is shifted by more than 1.5 times the root mean square deviation of the chemical shift difference distribution but could still be assigned are coloured in orange, and residues whose HSQC peak could not be assigned in the

MAN1LuhmWQ spectra are colored in cyan. The surface of the opposite face of the UHM domain is showed as an additional view. **(C)** Comparison of the SAXS data recorded on the wild-type (blue) and mutated (W765A-Q766A, red) Smad2 binding motif of MAN1. The upper view displays the superposition of the logarithm of the SAXS intensities as a function of the amplitude of the diffusion vector q = 4\*π\*sin(θ)/λ. The lower view corresponds to the superposition of the Kratky plots: the diffusion vector amplitude multiplied by the gyration radius is plotted on the abscissa and the square of the abscissa multiplied by the SAXS

intensity divided by the intensity at 0 is plotted on the ordinate. This representation reveals the type of structure: compact, partially folded or unfolded. **(D)** Model of the Smad2 binding domain of MAN1 consistent with the NMR and SAXS data. The backbone is represented as a ribbon; the side chains of W765 and Q766, in the linker, and of several residues lining the hydrophobic cavity of the UHM domain are displayed as sticks. **(E)** Fluorescence polarization curves of the labelled wild-type (left) and mutated (W765A-Q766A, right) Smad2 binding

fragment of MAN1 as a function of the concentration in Smad2.

1H chemical shift (in ppm)

D E

We mutated the highly conserved W765 and Q766 of the linker region and observed the consequences of these mutations on the NMR 1H-15N spectrum of the Smad2 binding region (Figure 15A). Residues with NMR signals modified by the mutations were mapped onto the UHM domain (Figure 15B). They all clustered on one face of the domain and delineated a hydrophobic pocket in which the tryptophane residue binds. We performed SAXS experiments on samples of wild type and mutated Smad2 binding region. SAXS provides accurate information on folding and conformation in solution for both rigid and flexible macromolecules and the indirect Fourier transform of the SAXS intensity gives access to the real space electron pair distance distribution of molecules in solution. Figure 15C (upper panel) shows that the SAXS curves of the wild type and mutated protein fragments are similar, reflecting their common global fold. Calculation of the electron pair distance distributions using GNOM gave access to the radius of gyration Rg, the maximal distance and the I0 value that cannot be measured experimentally and is proportional to the mass. From these values, a Kratky representation was displayed (Figure 15C, lower panel) that highlighted structural differences between wild-type and mutated proteins: the maximum of the curve corresponding to (SRg)2 I/Io is shifted towards higher SRg values, reflecting the more elongated/less structured state of the mutant. Moreover, the wild type protein shows a gyration radius between 20 and 21 Å and a maximal distance between 68 and 72 Å, whereas the mutant shows a gyration radius between 23.3 and 24 Å and a maximal distance is close to 90 Å. This result validates our previous NMR based analysis that strongly suggests a role of W765 and Q766 in anchoring the linker onto the UHM domain. In contrast, the C-terminus of the Smad2 binding region of MAN1 is partially unfolded. From these observations, the positions of the linker and C-terminus regions relatively to the UHM domain were calculated, yielding to a model of the whole Smad2 binding region of MAN1 (Figure 15D). The functional importance of the positioning of the linker region onto the UHM domain was shown using fluorescence binding experiments: mutations of the W765 and Q766 residues into alanine yielded a 15-fold decrease of the affinity of MAN1 for Smad2 (Figure 15E).

#### **5.2 Mechanism of bone diseases caused by the absence of the C-terminal region of MAN1**

Heterozygous loss-of-function mutations in *LEMD3* encoding MAN1 cause the sclerosing bone dysplasias osteopoikilosis, Buschke-Ollendorff syndrome and non-sporadic melorheostosis (Hellemans et al., 2004). Several arguments suggest that the bone diseases result from enhanced TGF-ß signalling in cells.

First, all the disease-causing mutations including nonsense or frameshift mutations induce loss of the C-terminal nucleoplasmic fragment of MAN1, which is essential for Smad2 recognition (Figure 14A). Second, the TGF-β pathway is enhanced in cells from patients (Hellemans et al., 2004). Loss of the MAN1-Smad2 interaction may lead to the release of activated Smad2 and Smad3, thus causing an increase of Smad2 dependant transcription (Lin et al., 2005; Pan et al., 2005). It has been proposed that MAN1-Smad2 interaction not only sequestrates Smad2 and Smad3 at the inner nuclear envelope, but also favours their nuclear export (Pan et al., 2005; Figure 16). However, the mechanism by which the nucleoplasmic MAN1 C-terminal region would stimulate this nuclear export remains unclear. MAN1 could simply act as a "nuclear envelope sink" for activated nuclear Smad proteins, sequestering them at the inner nuclear membrane (Figure 16-1) and competitively inhibiting interactions with other nuclear factors involved in gene regulation (Figure 16-2).

NMR Spectroscopy as a Tool to Provide Mechanistic

properties upon modification (Liokatis et al., 2010).

mutations perturb this functional network.

*Cell Biol*, Vol. 17, No.3, pp. 1154-63

dystrophy. *Nat Genet*, Vol. 8, No.4, pp. 323-7

therapies targeted to the defects.

**7. References** 

MAN1 binding.

**6. Conclusion** 

Clues About Protein Function and Disease Pathogenesis 163

We used protein binding assays and isothermal calorimetry experiments to map the Smad2 binding site onto the MAN1 structure (Kondé et al., 2010). These experiments suggested that MAN1 recognizes Smad2 through a motif similar to the Smad Interaction Motif found in transcription factors. MAN1 could thus compete with transcription factors for Smad2 and Smad3 binding. Further experiments are needed to characterize the fate of Smads after

1H, 15N and 13C NMR spectroscopy provide structural data which, when combined to SAXS data, allow the determination of the 3D structures of partially unfolded proteins. It is now possible to describe not only the structural consequences of disease-linked mutations located in well-folded domains but also their impact on the structural organisation of modular proteins (Mackereth et al., 2011). Such results provide mechanistic clues about how partially folded proteins use conformational switches in order to regulate their functions and highlight the connection between conformational variability, interaction and disease pathogenesis. Regulation of the function of modular proteins is also based on protein modifications. Recent studies propose that NMR characterization of a 15N labelled protein in cell extracts can be a unique tool to examine changes in protein structure and interaction

Characterization of interaction properties of disease-linked proteins is critical to understanding pathogenic mechanisms and may identify other genes to be screened in similar pathologies. NMR is a powerful tool to map low affinity (mM to M) interaction surfaces and determine high affinity (nM to pM) complex structures, and must be used in conjunction with other methods such as isothermal calorimetry or fluorescence to deduce the nature, affinity and stochiometry of the interaction. Analysis of these parameters aims at highlighting the competitive or synergic role of the studied interaction within the large cellular protein-protein interaction network, and at identifying how disease-linked

All theses different approaches highlight NMR spectroscopy as a technique adapted to the description of the consequences of disease-linked mutations in partially unfolded and multipartner proteins, as those associated with cellular control and signalling mechanisms. It has become an essential tool for understanding such biological processes from a mechanistic point of view at the atomic level. It can describe critical steps of disease mechanisms, highlight the mechanism of action of a particular drug and inspire the design of binding compounds and complex assembly inhibitors. NMR could be used in the future to develop

Bengtsson L & Wilson KL (2005). Barrier-to-autointegration factor phosphorylation on Ser-4

Bione S, Maestrini E, Rivella S, Mancini M, Regis S, Romeo G & Toniolo D (1994).

regulates emerin binding to lamin A in vitro and emerin localization in vivo. *Mol* 

Identification of a novel X-linked gene responsible for Emery-Dreifuss muscular

Another hypothesis, also proposed for the MAN1 paralog NET25 (Huber et al., 2009)*,* is that MAN1 functions as a scaffold protein recruiting Smad2, Smad3 and kinases/phosphatases to the inner nuclear membrane. Phosphorylation of Smad linker residues by kinases such as MAPKs or dephosphorylation of Smad C-terminal residues by phosphatases would then lead to enhanced nuclear export (Figure 16-3).

Fig. 16. The role of MAN1 in the TGF-β signaling pathway. The TGF-ß family of cytokines comprises key regulators of metazoan embryonic development and adult tissue homeostasis. In the canonical pathway, ligands of both the TGF-ß and the bone morphogenetic protein members of this family bind to heteromeric serine/threonine kinase receptor complexes, which in turn phosphorylate R-Smad transcription factors at their C-terminal tails. Upon phosphorylation, the R-Smads trimerize. This trimerization allows the R-Smads to interact with the common mediator Smad4. Then, these complexes translocate to the nucleus where they interact with various transcription factors, bind to DNA, and regulate transcription of targeted genes. MAN1 is able to influence the TGF-ß signalling by directly interacting with Smad2 and Smad3 (Hellemans et al., 2004; Lin et al., 2005). This leads to an inhibition of the TGF-ß signalling pathway (Pan et al., 2005). We suggest several putative mechanisms for the TGF-β pathway regulation by MAN1. **(1)** MAN1 could sequestrate the activated complexes R-Smads-Smad4 at the inner nuclear membrane, thus decreasing R-Smads accessibility to its targeted genes. **(2)** MAN1 could also compete with transcription factors for R-Smads binding resulting in a decrease of the transcriptional activity of R-Smads proteins. **(3)** MAN1 could recruit enzymes that modify R-Smads resulting in dissociation of active R-Smads-Smad4 complexes and nuclear export of R-Smads.

We used protein binding assays and isothermal calorimetry experiments to map the Smad2 binding site onto the MAN1 structure (Kondé et al., 2010). These experiments suggested that MAN1 recognizes Smad2 through a motif similar to the Smad Interaction Motif found in transcription factors. MAN1 could thus compete with transcription factors for Smad2 and Smad3 binding. Further experiments are needed to characterize the fate of Smads after MAN1 binding.

### **6. Conclusion**

162 Magnetic Resonance Spectroscopy

Another hypothesis, also proposed for the MAN1 paralog NET25 (Huber et al., 2009)*,* is that MAN1 functions as a scaffold protein recruiting Smad2, Smad3 and kinases/phosphatases to the inner nuclear membrane. Phosphorylation of Smad linker residues by kinases such as MAPKs or dephosphorylation of Smad C-terminal residues by phosphatases would then

Fig. 16. The role of MAN1 in the TGF-β signaling pathway. The TGF-ß family of cytokines comprises key regulators of metazoan embryonic development and adult tissue homeostasis. In the canonical pathway, ligands of both the TGF-ß and the bone morphogenetic protein members of this family bind to heteromeric serine/threonine kinase receptor complexes, which in turn phosphorylate R-Smad transcription factors at their C-terminal tails. Upon phosphorylation, the R-Smads trimerize. This trimerization allows the R-Smads to interact with the common mediator Smad4. Then, these complexes translocate to the nucleus where they interact with various transcription factors, bind to DNA, and regulate transcription of targeted genes. MAN1 is able to influence the TGF-ß signalling by directly interacting with Smad2 and Smad3 (Hellemans et al., 2004; Lin et al., 2005). This leads to an inhibition of the TGF-ß signalling pathway (Pan et al., 2005). We suggest several putative mechanisms for the TGF-β pathway regulation by MAN1. **(1)** MAN1 could sequestrate the activated complexes R-Smads-Smad4 at the inner nuclear membrane, thus decreasing R-Smads accessibility to its targeted genes. **(2)** MAN1 could also compete with transcription factors for R-Smads binding resulting in a decrease of the transcriptional activity of R-Smads proteins. **(3)** MAN1 could recruit enzymes that modify R-Smads resulting in dissociation of active R-Smads-Smad4

lead to enhanced nuclear export (Figure 16-3).

complexes and nuclear export of R-Smads.

1H, 15N and 13C NMR spectroscopy provide structural data which, when combined to SAXS data, allow the determination of the 3D structures of partially unfolded proteins. It is now possible to describe not only the structural consequences of disease-linked mutations located in well-folded domains but also their impact on the structural organisation of modular proteins (Mackereth et al., 2011). Such results provide mechanistic clues about how partially folded proteins use conformational switches in order to regulate their functions and highlight the connection between conformational variability, interaction and disease pathogenesis. Regulation of the function of modular proteins is also based on protein modifications. Recent studies propose that NMR characterization of a 15N labelled protein in cell extracts can be a unique tool to examine changes in protein structure and interaction properties upon modification (Liokatis et al., 2010).

Characterization of interaction properties of disease-linked proteins is critical to understanding pathogenic mechanisms and may identify other genes to be screened in similar pathologies. NMR is a powerful tool to map low affinity (mM to M) interaction surfaces and determine high affinity (nM to pM) complex structures, and must be used in conjunction with other methods such as isothermal calorimetry or fluorescence to deduce the nature, affinity and stochiometry of the interaction. Analysis of these parameters aims at highlighting the competitive or synergic role of the studied interaction within the large cellular protein-protein interaction network, and at identifying how disease-linked mutations perturb this functional network.

All theses different approaches highlight NMR spectroscopy as a technique adapted to the description of the consequences of disease-linked mutations in partially unfolded and multipartner proteins, as those associated with cellular control and signalling mechanisms. It has become an essential tool for understanding such biological processes from a mechanistic point of view at the atomic level. It can describe critical steps of disease mechanisms, highlight the mechanism of action of a particular drug and inspire the design of binding compounds and complex assembly inhibitors. NMR could be used in the future to develop therapies targeted to the defects.

#### **7. References**


NMR Spectroscopy as a Tool to Provide Mechanistic

Vol.17, No.9, pp. 1037-42

Vol.275, No.7, pp. 4840-7

*Soc*, Vol.132, No.42, pp. 14704-5

Clues About Protein Function and Disease Pathogenesis 165

Herrmann H, Strelkov SV, Burkhard P & Aebi U (2009). Intermediate filaments: primary determinants of cell architecture and plasticity. *J Clin Invest*, Vol.119, No.7, pp. 1772-83 Hu F, Luo W & Hong M (2010). Mechanisms of Proton Conduction and Gating in Influenza M2 Proton Channels from Solid-State NMR. *Science,* Vol.330, No.6003, pp. 505-508. Huber MD, Guan T & Gerace L (2009). Overlapping functions of nuclear envelope proteins

spectroscopy of proteins in human cells. *Nature,* Vol.458, pp. 106-109 Jehle S, Rajagopal P, Bardiaux B, Markovic S, Kühne R, Stout JR, Higman VA, Klevit RE, van

fragment of human lamin A. *J Mol Biol*, Vol.408, No.1, pp. 135-46

partial lipodystrophy. *Structure*, Vol.10, No.6, pp. 811-23

Kapinos LE, Burkhard P, Herrmann H, Aebi U & Strelkov SV (2011). Simultaneous

Kondé E, Bourgeois B, Tellier-Lebegue C, Wu W, Pérez J, Caputo S, Attanda W, Gasparini S,

Laguri C, Gilquin B, Wolff N, Romi-Lebrun R, Courchay K, Callebault I, Worman HJ &

human inner nuclear membrane proteins. *Structure*, Vol.9, No.6, pp. 503-11 Lin F, Blake DL, Callebault I, Skerjanc IS, Holmer L, McBurney MW, Paulin-Levasseur M &

Lin F, Morrison JM, Wu W & Worman HJ (2005). MAN1, an integral protein of the inner

Liotakis S, Dose A, Schwarzer D & Selenko P (2010). Simultaneous detection of protein

Liu J, Lee KK, Segura-Totten M, Neufeld E, Wilson KL & Gruenbaum Y (2003). MAN1 and

Mackereth C, Madl T, Bonnal S, Simon B, Zanier K, Gasch A, Rybin V, Valcarcel J & Sattler

Madl T, Bermel W & Zangger K (2009). Use of relaxation enhancements in a paramagnetic

Madl T, Gabel F & Sattler M (2011). NMR and small-angle scattering-based structural analysis of protein complexes in solution. *J Struct Biol,* Vol.173, No.3, pp. 472-82

factor-beta signaling. *Hum Mol Genet,* Vol.14, No.3, pp. 437-45

regulation by U2AF. *Nature,* Vol.475, pp. 408-411

*Angew Chem Int Ed Engl,* Vol.48, No.44, pp. 8259-8262

NET25 (Lem2) and emerin in regulation of extracellular signal-regulated kinase signalling in myoblast differentiation. *Mol Cell Biol*, Vol.29, No.21, pp. 5715-28 Inomata K, Ohno A, Tochio H, Isogai S, Tenno T, Nakase I, Takeuchi T, Futaki S, Ito Y,

Hiroaki H & Shirakawa M (2009). High resolution multi-dimensional NMR

Rossum BJ & Oschkinat H (2010). Solid-state NMR and SAXS studies provide a structural basis for the activation of alphaB-crystallin oligomers. *Nat Struct Mol Biol,*

formation of right- and left-handed anti-parallel coiled-coil interfaces by a coil2

Charbonnier JB, Gilquin B, Worman HJ & Zinn-Justin S (2010). Structural analysis of the Smad2-MAN1 interaction that regulates transforming growth factor-β signaling at the inner nuclear membrane. *Biochemistry,* Vol.49, No.37, pp. 8020-32 Krimm I, Ostlund C, Gilquin B, Couprie J, Hossenlopp P, Mornon JP, Bonne G, Courvalin

JC, Worman HJ & Zinn-Justin S (2002). The Ig-like structure of the C-terminal domain of lamin A/C, mutated in muscular dystrophies, cardiomyopathy, and

Zinn-Justin S (2001). Structural characterization of the LEM motif common to three

Worman HJ (2000). MAN1, an inner nuclear membrane protein that shares the LEM domain with lamina-associated polypeptide 2 and emerin. *J Biol Chem*,

nuclear membrane, binds Smad2 and Smad3 and antagonizes transforming growth

phosphorylation and acetylation by high-resolution NMR spectroscopy. *J Am Chem* 

emerin have overlapping function(s) essential for chromosome segregation and cell division in Caenorhabditis elegans. *Proc Natl Acad Sci U S A*, Vol.100, No.8, pp. 4598-603

M (2011). Multi-domain conformational selection underlies pre-mRNA splicing

environment for the structure determination of proteins using NMR spectroscopy.


Bradley CM, Ronning DR, Ghirlando R, Craigie R & Dyda F (2005). Structural basis for DNA

Burz DS, Dutta K, Cowburn D & Shekhtman A (2006). Mapping structural interactions using in-cell NMR spectroscopy (STINT-NMR). *Nature Methods,* Vol.3, pp. 91-93 Cady SD, Schmidt-Rohr K, Wang J, Soto CS, Degrado WF & Hong M (2010). Structure of the

Cai M, Huang Y, Ghirlando R, Wilson KL, Craigie R & Clore GM (2001). Solution structure of the

one binds BAF and the other binds DNA. *EMBO J*, Vol.20, No.16, pp. 4399-407 Cai M, Huang Y, Suh JY, Louis JM, Ghirlando R, Craigie R & Clore GM (2007). Solution

Caputo S, Couprie J, Duband-Goulet I, Kondé E, Lin F, Braud S, Gondry M, Gilquin B, Worman

Chi YH, Chen ZJ & Jeang KT (2009). The nuclear envelopathies and human diseases. *J* 

Dauer WT & Worman HJ (2009). The nuclear envelope as a signaling node in development

Dedmon MM, Patel CN, Young GB & Pielak GJ (2002). FlgM gains structure in living cells.

De Sandre-Giovannoli A, Bernard R, Cau P, Navarro C, Amiel J, Boccaccio I, Lyonnet S,

Dhe-paganon S, Werner ED, Chi YI & Shoelson SE (2002). Structure of the globular tail of

Eriksson M, Brown WT, Gordon LB, Glynn MW, Singer J, Scott L, Erdos MR, Robbins CM,

Hutchinson-Gilford progeria syndrome. *Nature*, Vol.423, No.6937, pp. 293-8 Fidziańska A & Hausmanowa-Petrusewicz I (2003). Architectural abnormalities in muscle

Frock RL, Kudlow BA, Evans AM, Jameson SA, Hauschka SD & Kennedy BK (2006). Lamin

Gautier A, Mott HR, Bostock MJ, Kirkpatrick JP & Nietlispach D (2010). Structure

solution NMR spectroscopy. *Nat Struct Mol Biol,* Vol.17, No.6, pp. 768–774 Gonzáles JM, Navarro-Puche A, Casar B, Crespo P & Andrés V (2008). Fast regulation of

Hellemans J, Preobrazhenska O, Willaert A, Debeer P, Verdonk PC, Costa T, Janssens K,

Hutchinson-Gilford progeria. *Science*, Vol. 300, No. 5628, pp. 2055.

Vol.443, No.7281, pp. 689-92

Vol.282, No.19, pp. 14525-35

*Biomed Sci*, Vol.16, pp. 96

*Dev*, Vol.20, No.4, pp. 486-500

and disease. *Dev Cell*, Vol.17, No.5, pp. 626-38

*Proc. Natl Acad. Sci. USA,* Vol.99, pp. 12681-12684

nuclear lamin. *J Biol Chem*, Vol.277, No.20, pp. 17381-4

forms of EDMD. *J Neurol Sci*, Vol.210, No.1-2, pp. 47-51

envelope. *J Cell Biol*, Vol.183, No.4, pp. 653-66

melorheostosis. *Nat Genet*, Vol.36, No.11, pp. 1213-8

bridging by barrier-to-autointegration factor. *Nat Struct Mol Biol*, Vol.12, No.10, pp.935-6

amantadine binding site of influenza M2 proton channels in lipid bilayers. *Nature,* 

constant region of nuclear envelope protein LAP2 reveals two LEM-domain structures:

NMR structure of the barrier-to-autointegration factor-Emerin complex. *J Biol Chem*,

HJ & Zinn-Justin S (2006). The carboxyl-terminal nucleoplasmic region of MAN1 exhibits a DNA binding winged helix domain. *J Biol Chem,* Vol.281, No.26, pp. 18208-15

Stewart CL, Munnich A, Le Merrer M & Lévy N. (2003). Lamin a truncation in

Moses TY, Berglund P, Dutra A, Pak E, Durkin S, Csoka AB, Boehnke M, Glover TW & Collins FS (2003). Recurrent de novo point mutations in lamin A cause

nuclei. Ultrastructural differences between X-linked and autosomal dominant

A/C and emerin are critical for skeletal muscle satellite cell differentiation. *Genes* 

determination of the seven-helix transmembrane receptor sensory rhodopsin II by

AP-1 activity throught interaction of lamin A/C, ERK1/2, and c-Fos at the nuclear

Menten B, Van Roy N, Vermeulen SJ, Savarirayan R, Van Hul W, Vanhoenacker F, Huylebroeck D, De Paepe A, Naeyaert JM, Vandesompele J, Speleman F, Verschueren K, Coucke PJ & Mortier GR (2004). Loss-of-function mutations in LEMD3 result in osteopoikilosis, Buschke-Ollendorff syndrome and


**9** 

**Structural and Vibrational Properties** 

**and NMR Characterization of** 

Ana E. Ledesma1, Juan Zinczuk2,

**(2'-furyl)-Imidazole Compounds** 

Juan J. López González3 and Silvia A. Brandán4\*

*Ciencias Experimentales, Universidad de Jaén, Jaén* 

*San Miguel de Tucumán, Tucumán, República* 

*1,2,4Argentina* 

*3Spain* 

*1INQUINOA-UNSE, Facultad de Agronomía y Agroindustrias, Universidad Nacional de Santiago del Estero, Santiago del Estero 2Instituto de Química Rosario (CONICET-UNR), Facultad de Ciencias Bioquímicas y Farmacéuticas, Rosario, Santa Fé 3Departamento de Química Física y Analítica, Facultad de* 

*4Cátedra de Química general, Instituto de Química Inorgánica, Facultad de Bioquímica, Química y Farmacia, Universidad Nacional de Tucumán,* 

The furylimidazoles compounds have a great importance in biochemistry and pharmacology because many of them present interesting properties1,2. They are also found in artificial compounds3, such as agrochemicals,4 pharmaceuticals5, dyes6, plastics, solvents, photographic chemicals, electronics, corrosion inhibitors,7 preservatives, and polymers8. They can be used in synthetic organic chemistry as building blocks, due to their presence as key structural units in many natural products and in important pharmaceuticals. For example, the nature of the N- atoms in the imidazole molecule makes possible an extraordinary variety of reactions and this is the main reason for the great biological

**1. Introduction** 

 Corresponding Author 1 Salerno Perillo, 2005

7 Rohwerder Michalik, 2007. 8 Stroganova, et al., 2000.

2 Szabo, 2002. 3 Gould et. al., 2006. 4 Higashio Shoji, 2004. 5 Monterrey et al., 2004. 6 Ooyama et al., 2007.

 \*

importance of the amino acid, histidine.


### **Structural and Vibrational Properties and NMR Characterization of (2'-furyl)-Imidazole Compounds**

Ana E. Ledesma1, Juan Zinczuk2,

Juan J. López González3 and Silvia A. Brandán4\* *1INQUINOA-UNSE, Facultad de Agronomía y Agroindustrias, Universidad Nacional de Santiago del Estero, Santiago del Estero 2Instituto de Química Rosario (CONICET-UNR), Facultad de Ciencias Bioquímicas y Farmacéuticas, Rosario, Santa Fé 3Departamento de Química Física y Analítica, Facultad de Ciencias Experimentales, Universidad de Jaén, Jaén 4Cátedra de Química general, Instituto de Química Inorgánica, Facultad de Bioquímica, Química y Farmacia, Universidad Nacional de Tucumán, San Miguel de Tucumán, Tucumán, República 1,2,4Argentina 3Spain* 

#### **1. Introduction**

166 Magnetic Resonance Spectroscopy

Mansharamani M & Wilson KL (2005). Direct binding of nuclear membrane protein MAN1

Muchir A & Worman HJ (2007). Emery-Dreifuss muscular dystrophy. *Curr Neurol Neurosci* 

Pan D, Estévez-Salmerón LD, Stroschein SL, Zhu X, He J, Zhou S & Luo K (2005). The inner

Puente XS, Quesada V, Osorio FG, Cabanillas R, Cadiñanos J, Fraile JM, Ordóñez GR, Puente

Sattler M, Schleucher J & Griesinger C (1999). Heteronuclear multidimensional NMR

Selenko P, Serber Z, Gade B, Ruderman J & Wagner G (2006). Quantitative NMR analysis of

Sharma M, Yi M, Dong H, Qin H, Peterson E, Busath DD, Zhou HX & Cross TA (2010).

Shumaker DK, Lee KK, Tanhehco YC, Craigie R & Wilson KL (2001). LAP2 binds to

Strelkov SV, Schumacher J, Burkhard P, Aebi U & Herrmann H (2004). Crystal structure of

Tycko R (2011). Solid-State NMR studies of amyloid fibril structure. *Annu Rev Phys Chem,*

Van Melckebeke H, Wasmer C, Lange A, Ab E, Loquet A, Böckmann A & Meier BH (2010).

solid-state NMR spectroscopy. *J Am Chem Soc,* Vol.132, No.39, pp. 13765-75 Verstraeten VL, Broers JL, van Steensel MA, Zinn-Justin S, Ramaekers FC, Steijlen PM,

Wolff N, Gilquin B, Courchay K, Callebault I, Worman HJ & Zinn-Justin S (2001). Structural

Worman HJ, Ostlund C & Wang Y (2010). Diseases of the nuclear envelope. *Cold Spring Harb* 

Wright PE & Dyson HJ (2009). Linking folding and binding. *Curr Opin Struct Biol,* Vol.19,

Dreifuss muscular dystrophy. *FEBS let*, Vol.501, No.2-3, pp. 171-6

pulsed field gradients *Progr Magn Res Spect,* Vol.34, pp. 93-158

Natl Acad. Sci. USA, Vol.103, pp. 11904-11909

lipid bilayer. *Science,* Vol.330, No.6003, pp. 509-12

variable regions. *EMBO J*, Vol.20, No.7, pp. 1754-64

nuclear lamins. *J Mol Biol*, Vol.343, No.4, pp. 1067-80

*Biol Chem*, Vol.280, No.14, pp. 13863-70

cytokines. *J Biol Chem*, Vol.280, No.16, pp. 15992-6001

*Rep*, Vol.7, No.1, pp. 78-83

*Nature,* Vol.458, pp. 102-105

Vol.62, pp. 279-299

No.16, pp. 2509-22

No.1, pp. 31-8

*Perspect Biol*, Vol.2, No.2

to emerin in vitro and two modes of binding to barrier-to-autointegration factor. *J* 

nuclear membrane protein MAN1 physically interacts with R-Smad proteins to repress signaling by the transforming growth factor-{beta} superfamily of

DA, Gutiérrez-Fernández A, Fanjul- Fernández M, Lévy N, Freije JM & López-Otín C (2011). Exome sequencing and functional analysis identifies BANF1 mutation as the cause of a hereditary progeroid syndrome. *Am J Hum Genet*, Vol.88, No.5, pp. 650-6 Sakakibara D, Sasaki A, Ikeya T, Hamatsu J, Hanashima T, Mishima M, Yoshimasu M,

Hayashi N, Mikawa T, Wälchli M, Smith BO, Shirakawa M, Güntert P & Ito Y (2009). Protein structure determination in living cells by in-cell NMR spectroscopy.

experiments for the structure determination of proteins in solution employing

the protein G B1 domain in Xenopus laevis egg extracts and intact oocytes. Proc.

Insight into the mechanism of the influenza A proton channel from a structure in a

BAF.DNA complexes: requirement for the LEM domain and modulation by

the human lamin A coil 2B dimer: implications for the head-to-tail association of

Atomic-resolution three-dimensional structure of HET-s(218-289) amyloid fibrils by

Kamps M, Kuijpers HJ, Merckx D, Smeets HJ, Hennekam RC, Marcelis CL & Van den Wijngaard A (2006). Compound heterozygosity for mutations in LMNA causes a progeria syndrome without prelamin A accumulation. *Hum Mol Genet,* Vol.15,

analysis of emerin, an inner nuclear membrane protein mutated in X-linked Emery-

The furylimidazoles compounds have a great importance in biochemistry and pharmacology because many of them present interesting properties1,2. They are also found in artificial compounds3, such as agrochemicals,4 pharmaceuticals5, dyes6, plastics, solvents, photographic chemicals, electronics, corrosion inhibitors,7 preservatives, and polymers8. They can be used in synthetic organic chemistry as building blocks, due to their presence as key structural units in many natural products and in important pharmaceuticals. For example, the nature of the N- atoms in the imidazole molecule makes possible an extraordinary variety of reactions and this is the main reason for the great biological importance of the amino acid, histidine.


<sup>\*</sup> Corresponding Author

<sup>1</sup> Salerno Perillo, 2005

<sup>2</sup> Szabo, 2002.

<sup>3</sup> Gould et. al., 2006.

<sup>6</sup> Ooyama et al., 2007.

<sup>7</sup> Rohwerder Michalik, 2007.

<sup>8</sup> Stroganova, et al., 2000.

Structural and Vibrational Properties and NMR Characterization of (2'-furyl)-Imidazole Compounds 169

structures and vibrational properties of 2-(2'-furyl)-imidazole, 4-(2'-furyl)-imidazole, 5-(2'-

All compounds of this series are present in two different stable conformations according to the syn and anti position of the oxygen atom with respect to the N—H bond, named syn and

furyl)-imidazole and N-(2'-furyl)-imidazole are compared and analyzed.

**2. Structure and properties of the (2'-furyl)-imidazole series** 

anti conformers, respectively, as show in Figure 1.

Fig. 1. Molecular structures of the (2'-furyl)-imidazole series

**2.1 Structural analysis** 

Polymers of these compounds, which have potential as semi-conducting material are of great importance, because of their good thermal and chemical stability and relative ease of functionalization, which potentially permit the fine tuning of their physical and electronic properties9

In fact, these compounds provide one of the best studied examples of annular tautomerism, such as the 4(5)-substituted imidazoles10,11,12. Recently, the theoretical and experimental studies on structure, electronic and vibrational properties for the (2'-furyl)-imidazole series were reported by us13,14,15,16,17.

Examination of the spectroscopic data for the each furanics molecules indicated that Nuclear Magnetic resonance spectroscopy, coupled with IR spectroscopy, was potentially the best analytical technique to provide confirmatory structural evidence for their formation. On the other hand, in general, the chemical shifts for the substituted imidazole ring, relative to the unsubstituted rings, suggest that the mesomeric effect of the five-membered rings contributes less to the interaction between two rings than does their inductive effects. It would indicate which rings is more electron-withdrawing in its overall electronic effect than the furan ring, as well as observed in the pyridine-thiophenes and furans18. The downfield shifts of the 13C NMR signals for the carbon atom at the point of substitution on the imidazole ring would be explicate of the significant inductive electron-withdrawing effects of the 2-heteroaryl rings.

For this, reason for predicting the reactivity of this molecules due to the different position of the substituent, in this chapter, a comparison of the structural and vibrational properties of the four molecules belong to the (2'-furyl)-imidazole series was performed in order to evaluate the experimental and theoretical results based on B3LYP calculations by using 6-31G\* and 6- 311++G\*\* basis sets. The optimized geometries and frequencies for the normal modes of vibration were calculated using both theory levels. For a complete assignment of the compounds, the DFT (Density functional Theory) calculations were combined with the scaled quantum mechanical force field (SQMFF) methodology19,20,21 in order to fit the theoretical wavenumbers to the experimental ones. In all molecules of this series, the changes in the chemical shifts of the hydrogen and carbon atoms were also studied. In addition, the electronic properties of those molecules were evaluated by means of natural bond orbital (NBO) and atoms in molecules (AIM) studies in order to analyze the nature and magnitude of the intramolecular interactions. In this chapter, the experimental and theoretical studies of


<sup>9</sup> Nevin et al., 2008.

<sup>20</sup> Rauhut Pulay, 1995b.

<sup>21</sup> Kalincsak Pongor, 2002.

structures and vibrational properties of 2-(2'-furyl)-imidazole, 4-(2'-furyl)-imidazole, 5-(2' furyl)-imidazole and N-(2'-furyl)-imidazole are compared and analyzed.

#### **2. Structure and properties of the (2'-furyl)-imidazole series**

#### **2.1 Structural analysis**

168 Magnetic Resonance Spectroscopy

Polymers of these compounds, which have potential as semi-conducting material are of great importance, because of their good thermal and chemical stability and relative ease of functionalization, which potentially permit the fine tuning of their physical and electronic

In fact, these compounds provide one of the best studied examples of annular tautomerism, such as the 4(5)-substituted imidazoles10,11,12. Recently, the theoretical and experimental studies on structure, electronic and vibrational properties for the (2'-furyl)-imidazole series

Examination of the spectroscopic data for the each furanics molecules indicated that Nuclear Magnetic resonance spectroscopy, coupled with IR spectroscopy, was potentially the best analytical technique to provide confirmatory structural evidence for their formation. On the other hand, in general, the chemical shifts for the substituted imidazole ring, relative to the unsubstituted rings, suggest that the mesomeric effect of the five-membered rings contributes less to the interaction between two rings than does their inductive effects. It would indicate which rings is more electron-withdrawing in its overall electronic effect than the furan ring, as well as observed in the pyridine-thiophenes and furans18. The downfield shifts of the 13C NMR signals for the carbon atom at the point of substitution on the imidazole ring would be explicate of the significant inductive electron-withdrawing effects

For this, reason for predicting the reactivity of this molecules due to the different position of the substituent, in this chapter, a comparison of the structural and vibrational properties of the four molecules belong to the (2'-furyl)-imidazole series was performed in order to evaluate the experimental and theoretical results based on B3LYP calculations by using 6-31G\* and 6- 311++G\*\* basis sets. The optimized geometries and frequencies for the normal modes of vibration were calculated using both theory levels. For a complete assignment of the compounds, the DFT (Density functional Theory) calculations were combined with the scaled quantum mechanical force field (SQMFF) methodology19,20,21 in order to fit the theoretical wavenumbers to the experimental ones. In all molecules of this series, the changes in the chemical shifts of the hydrogen and carbon atoms were also studied. In addition, the electronic properties of those molecules were evaluated by means of natural bond orbital (NBO) and atoms in molecules (AIM) studies in order to analyze the nature and magnitude of the intramolecular interactions. In this chapter, the experimental and theoretical studies of

properties9

were reported by us13,14,15,16,17.

of the 2-heteroaryl rings.

9 Nevin et al., 2008. 10 Bellina et al., 2007. 11 Popov et al., 2004. 12 Sztanke et al., 2005. 13 Ledesma et al., 2008. 14 Ledesma et al., 2010. 15 Ledesma et al., 2009a. 16 Ledesma et al., 2009b. 17 Zinczuk et al., 2009. 18 Jones Civcir, 1997. 19 Rauhut Pulay,1995a. 20 Rauhut Pulay, 1995b. 21 Kalincsak Pongor, 2002. All compounds of this series are present in two different stable conformations according to the syn and anti position of the oxygen atom with respect to the N—H bond, named syn and anti conformers, respectively, as show in Figure 1.

Fig. 1. Molecular structures of the (2'-furyl)-imidazole series

Structural and Vibrational Properties and NMR Characterization of (2'-furyl)-Imidazole Compounds 171

are in beta and gamma positions, respectively. This effect also explains the observed

N-(2'-furyl)-imidazole

*anti* -455.03903057 0 3.63 -455.16243953 0 3.82 *syn* -455.03834901 1.8 3.80 -455.16160715 2.2 3.99 2-(2'-furyl)-imidazole *anti* -455.04955562 15.30 3.90 -455.17597057 15.42 4.08 *syn* -455.05531501 0.0 2.97 -455.18185872 0.0 2.92 4-(2'-furyl)-imidazole *anti* -455.0490477 12.25 3.94 -455.1754201 12.54 4.10 *syn* -455.0537122 0.0 3.23 -455.1802136 0.0 3.24 5-(2'-furyl)-imidazole *anti* -455.0512980 5.73 3.90 - 455.1772971 5.68 4.17 *syn* -455.0534867 0.0 3.37 - 455.1794579 0.0 3.42

isomer those values are between 30-40° and 160-170° respectively.

Another very important difference between all series members is related to the dihedral angles. Thus, the angles values are 180° and 0° for the 2 and 4(5) isomers, while for the first

The stability of the different conformers were investigated using the Electrostatic Surface Potential, ESP maps22,23 because these surfaces are important for describing overall molecular charge distributions as well as anticipating sites of electrophilic addition23. The molecular ESP values for the anti and syn conformers of each member's series by using 6- 31G\* basis set are given in Table 3 while Figure 3 shows the ESP maps for all conformers of this series. The red color represents negatively charged areas of the surface (i.e., those areas where accepting an electrophile is most favorable). The atomic charges derived from the ESPs (MK) and natural atomic charges were also analyzed and the corresponding values are given in Table 4. Note that the most important differences in all cases are observed on the imidazole ring. The important factor responsible for the lesser stability of same conformer is

E (B3LYP/6- 311++G\*\*) (Hartrees)

*E*) calculates (in Hartree), dipolar moments (

*E*  kJ/mol

) for

*E*  kJ/mol

tendency for the C-N bond next to furan ring.

*E* (B3LYP/6- 31G\*) (Hartrees)

Table 1. Total Energies (*E*) and relatives (

Conf.

a In Debyes

22 Sadlej-Sosnowska, 2007. 23 Vektariene at al., 2007.

all structures

Experimentally, all molecules of this series have different characteristics or they are solid of different nature. Thus, the first series's member, N-(2'-furyl)-imidazole, is a liquid at room temperature and probably the two conformers are present in it. For the second series's member, 2-(2'-furyl)-imidazole, the crystallography analysis shows that both conformers are present in the crystalline lattice with equal occupancy and alternated arrangement in the N-H-- -N bonded polymeric chains along the crystal [101] direction. Only three units of the polymeric structure are shown in Figure 2. Finally, the 4-(2'-furyl)-imidazole and 5-(2'-furyl)-imidazole compounds are present in a tautomeric mixture in the solid state as a white power.

The calculation indicates that both conformers of N-(2'-furyl)-imidazole have C1 symmetries, in the 2-(2'-furyl)-imidazole they have C1 and Cs symmetries, while in the remaining molecules all conformers present Cs symmetries.

Fig. 2. View of a N—H---N bonded polymer chain in solid 2-(2'-furyl)-imidazole. The central monomer is related with the right and left side monomers through a crystallographic inversion center and a two fold axis rotation, respectively. The H-bonding is indicated by dashed lines

Table 1 shows a comparison of the total energy and dipole moment values for all molecules by using the 6-31 and 6-311++G\*\* basis sets at B3LYP theory level. Note that all predicts values are similar between them and, only for the N-(2'-furyl)-imidazole, the calculation predicts that the anti conformation is the most stable, while that for the remaining compounds the syn conformation is the most stable.

The comparison between the experimental geometrical parameters (obtained for 2-(2'-furyl) imidazole) with the corresponding theoretical values are shown in Table 2. The analysis shows that the length of C=C bond of the furan ring, does not change for all molecules, while the length of C=C bond of the imidazole ring varies according the following order: (4)5-(2'-furyl)-imidazole > N-(2'-furyl)-imidazole > 2-(2'-furyl)-imidazole. This trend is explained taking into account the next of this bond respect to the furan ring, since in the first molecule that bond is in alpha position respect to C=C bond corresponding to the furan ring, favoring the electronic delocalization, while in the remaining structures those bonds


are in beta and gamma positions, respectively. This effect also explains the observed tendency for the C-N bond next to furan ring.

a In Debyes

170 Magnetic Resonance Spectroscopy

Experimentally, all molecules of this series have different characteristics or they are solid of different nature. Thus, the first series's member, N-(2'-furyl)-imidazole, is a liquid at room temperature and probably the two conformers are present in it. For the second series's member, 2-(2'-furyl)-imidazole, the crystallography analysis shows that both conformers are present in the crystalline lattice with equal occupancy and alternated arrangement in the N-H-- -N bonded polymeric chains along the crystal [101] direction. Only three units of the polymeric structure are shown in Figure 2. Finally, the 4-(2'-furyl)-imidazole and 5-(2'-furyl)-imidazole

The calculation indicates that both conformers of N-(2'-furyl)-imidazole have C1 symmetries, in the 2-(2'-furyl)-imidazole they have C1 and Cs symmetries, while in the

Fig. 2. View of a N—H---N bonded polymer chain in solid 2-(2'-furyl)-imidazole. The central monomer is related with the right and left side monomers through a crystallographic inversion center and a two fold axis rotation, respectively. The H-bonding is indicated by

Unit 2 Unit 1 Unit 3

Table 1 shows a comparison of the total energy and dipole moment values for all molecules by using the 6-31 and 6-311++G\*\* basis sets at B3LYP theory level. Note that all predicts values are similar between them and, only for the N-(2'-furyl)-imidazole, the calculation predicts that the anti conformation is the most stable, while that for the remaining

The comparison between the experimental geometrical parameters (obtained for 2-(2'-furyl) imidazole) with the corresponding theoretical values are shown in Table 2. The analysis shows that the length of C=C bond of the furan ring, does not change for all molecules, while the length of C=C bond of the imidazole ring varies according the following order: (4)5-(2'-furyl)-imidazole > N-(2'-furyl)-imidazole > 2-(2'-furyl)-imidazole. This trend is explained taking into account the next of this bond respect to the furan ring, since in the first molecule that bond is in alpha position respect to C=C bond corresponding to the furan ring, favoring the electronic delocalization, while in the remaining structures those bonds

compounds are present in a tautomeric mixture in the solid state as a white power.

remaining molecules all conformers present Cs symmetries.

compounds the syn conformation is the most stable.

dashed lines

Table 1. Total Energies (*E*) and relatives (*E*) calculates (in Hartree), dipolar moments () for all structures

Another very important difference between all series members is related to the dihedral angles. Thus, the angles values are 180° and 0° for the 2 and 4(5) isomers, while for the first isomer those values are between 30-40° and 160-170° respectively.

The stability of the different conformers were investigated using the Electrostatic Surface Potential, ESP maps22,23 because these surfaces are important for describing overall molecular charge distributions as well as anticipating sites of electrophilic addition23. The molecular ESP values for the anti and syn conformers of each member's series by using 6- 31G\* basis set are given in Table 3 while Figure 3 shows the ESP maps for all conformers of this series. The red color represents negatively charged areas of the surface (i.e., those areas where accepting an electrophile is most favorable). The atomic charges derived from the ESPs (MK) and natural atomic charges were also analyzed and the corresponding values are given in Table 4. Note that the most important differences in all cases are observed on the imidazole ring. The important factor responsible for the lesser stability of same conformer is

 22 Sadlej-Sosnowska, 2007.

<sup>23</sup> Vektariene at al., 2007.

Structural and Vibrational Properties and NMR Characterization of (2'-furyl)-Imidazole Compounds 173

N 1 -18.27 -18.26 -18.28 -18.28 -18.28 -18.28 -18.28 -18.28

C 2 -14.68 -14.68 -14.68 -14.67 -14.68 -14.68 -14.68 -14.68

N 3 -18.37 -18.37 -18.38 -18.38 -18.38 -18.37 -18.37 -18.37

C 4 -14.72 -14.72 -14.71 -14.72 -14.71 -14.71 -14.72 -14.72

C 5 -14.70 -14.70 -14.70 -14.71 -14.70 -14.70 -14.69 -14.69

H 6 -1.10 -1.10 -0.97 -0.98 -0.97 -0.97 -0.97 -0.98

C 7 -1.07 -1.08 -1.07 -1.10 -1.07 -1.07 -1.10 -1.10

H 8 -14.63 -14.63 -14.69 -14.68 -14.69 -14.69 -14.67 -14.67

H 9 -1.07 -1.07 -1.07 -1.07 -1.07 -1.07 -1.07 -1.07

C 10 -14.72 -14.72 -14.75 -14.74 -14.75 -14.75 -14.73 -14.73

C 11 -14.72 -14.72 -14.75 -14.73 -14.75 -14.75 -14.73 -14.73

C 12 -14.67 -14.67 -14.70 -14.68 -14.70 -14.70 -14.68 -14.68

O 13 -22.22 -22.22 -22.26 -22.24 -22.26 -22.25 -22.24 -22.23

H 14 -1.07 -1.07 -1.10 -1.09 -1.10 -1.11 -1.08 -1.08

H 15 -1.07 -1.07 -1.10 -1.09 -1.10 -1.10 -1.08 -1.08

H 16 -1.06 -1.06 -1.09 -1.07 -1.09 -1.08 -1.08 -1.07

For the N-(2'-furyl)-imidazole a region can be observed where the delocalization of charge decrease (blue region between two rings) due to the breaking-off planarity of the both rings, while the remain molecules show π-electron delocalizated surfaces that evidence stables

These results can be related with the energies of the HOMO and LUMO for all molecules. Table 4 shows the comparison of the HOMO and LUMO orbitals energies values as well as

The corresponding values for the N-(2'-furyl)-imidazole show that this compound is the most stable of the series, as expected, because it presents the longer GAP values. On the

Table 3. Molecular electrostatic potentials for all members of the series calculated at

aromatic systems, being smaller for the N-(2'-furyl)-imidazole compound.

*anti syn anti syn anti syn anti syn* 

4-(2'-furyl) imidazole

5-(2'-furyl) imidazole

2-(2'-furyl) imidazole

Átom

N-(2'-furyl) imidazole

B3LYP/6-31G\* level of theory

the GAP energies values of them.


the electrostatic repulsion between the two lone pairs on N and O. Hence, a strong red color on both atoms is observed in the corresponding Figure 3.
