**Characterization of Bone and Bone-Based Graft Materials Using FTIR Spectroscopy**

M.M. Figueiredo, J.A.F. Gamelas and A.G. Martins *Chemical Engineering Department, University of Coimbra Portugal* 

#### **1. Introduction**

314 Infrared Spectroscopy – Life and Biomedical Sciences

Zhang, L.; Small, G.W.; Haka, A.S.; Kidder, L.H. & Lewis, E.N. (2003) Classification of

Fourier transform infrared microscopic imaging data of human breast cells by cluster analysis and artificial neural networks. *Appl Spectrosc*. 2003 Jan;57(1):14-22.

> Despite its static appearance, bone is a very dynamic living tissue that undergoes constant remodelling throughout life. After blood, bone is the second most commonly transplanted tissue in human medicine, and thus a thorough characterization of this material is of crucial importance. For instance, although it has been accepted that the whole bone strength is related to the bone mineral density (BMD), changes in this parameter often do not correlate with the probability of fracture. To improve the mechanical performance of bone, other material properties related to bone quality, and not only quantity, have to be investigated.

> FTIR spectroscopy constitutes an excellent tool to characterize the bone matrix because its main components (carbonated hydroxyapatite and collagen) absorb infrared radiation at distinct, almost complementary, regions within the 500-4000 cm-1 range. This not only enables the study of the main contributions of each component separately, but also allows a further investigation of relevant parameters that mostly affect the structural and mechanical properties of bone, as well as its active metabolism. Furthermore, FTIR analysis, particularly when combined with microscopic technologies, enables the measurement of spatial variations in bone composition, allowing their correlation with micro-to-macro morphostructural properties. In clinical studies, this allows the comparison between sound and diseased bone, and the analysis of therapeutic effects of drugs, among many other examples. On the other hand, FTIR spectroscopy is also commonly used to characterize bone grafts. Moreover, evidence that optimizing the osteointegration requires a fundamental knowledge of the material properties both of the bone graft and of the host bone tissue, has driven to extensive research on this subject, often supported by FTIR spectroscopy.

#### **2. Bone composition and structure**

Besides providing mechanical support, bone also plays diverse important metabolic functions. Even though bones possess various sizes and shapes, they share general chemical and structural features that, despite a static appearance on adult vertebrates, vary with and within species (Gamsjaeger et al., 2010; Mkukuma et al., 2004). In fact, among similar individuals, bone properties change with age, nutrition, hormonal equilibrium and health condition, in addition to other factors such as biomechanical environment (Aerssens et al., 1997; Fratzl et al., 2004). These variations are relatively restricted due to the constant remodelling processes that re-establish equilibrium. The bone turnover or remodelling, that consists on the resorption,

Characterization of Bone and Bone-Based Graft Materials Using FTIR Spectroscopy 317

Fig. 1. Evolution of the apatite crystal and of the surrounding hydrated layer. During maturation and growth of the mineral, the apatitic domain of the crystal increases whereas the hydrated layer decreases, due to continuous ionic exchange between these domains and the solution bath. Soluble charged groups of proteins (Pr) can also participate in the ionic

This principal component of the organic matrix of bone is a large fibrous protein with a highly repetitive amino acid sequence [Gly (glycine)–X–Y]n (often X is proline and Y is hydroxyproline). Other amino acids may also be found, such as alanine, lysine and hydroxylysine. These monomers are bound together by peptide bonds (between the carboxyl and amino groups of adjacent amino acids) constituting three polypeptide chains (two 1 and one 2 chains) that fold into a unique structure (Fig. 2). This consists of a single uninterrupted triple helix which represents more than 95% of the molecule (tropocollagen) and two non-helical domains (the telopeptides) containing the – COOH and – NH2 terminals

The integrity of the collagen molecule is attributed to the formation of intra- and intermolecular cross-linking which contribute to the mechanical strength of the bone organic matrix (Eyre & Wu, 2005; Ottani et al., 2001; Saito & Marumo, 2009). Inter-molecular collagen cross-links can be divided into two types: lysine hydroxylase/lysyl oxidasecontrolled cross-links (enzymatic cross-links) and glycation/oxidation-induced cross-links. In both cases, cross-linking sites are at specific Lys (lysine) or Hyl (hydroxylysine) residues of the collagen molecule. Concerning the first type, the enzymatic activity leads to the formation of Lys and Hyl aldehyde-derived covalent cross-links, in which a molecular fragment binds a telopeptide and a triple helix of adjacent molecules (divalent cross-link). As a consequence of chemical reactions and rearrangements, divalent cross-links may be converted into trivalent cross-links, in which a helix is bound to two telopeptides. As

equilibrium of the non-apatitic domain (adapted from Farlay et al., 2010).

of the protein.

followed by the replacement of bone with little change in shape, acts as a dynamic response to multiple local and systemic factors (Robling et al., 2006). The bone metabolic functions are regulated by the combined action of two principal cell lineages: the bone forming cells (mainly osteoblasts and osteocytes) and bone resorption cells (osteoclats). Diverse enzymes and many other molecules, such as the important bone morphogenetic proteins (BMPs), participate in those dynamic cellular processes (Katz et al., 2009; Urist, 1965).

The composition of the extracellular matrix of bone, often described as a two-phase composite, consists of about 65 wt. % mineral component (carbonated hydroxyapatite) and 25 wt. % organic component (mainly type I collagen), being the remaining 10 wt. % water (Judas et al., 2005). These components have extremely different mechanical properties: the mineral is stiff and brittle while the (wet) protein is much more elastic and also much tougher than the mineral. Bone combines both the stiffness and the toughness in an unusual arrangement of material properties that result on a remarkable tensile strength and resistance against fracture (Fratzl et al., 2004; Rho et al., 1998).

The bone components are so tightly embedded that require the use of chemical procedures to obtain an effective separation of the mineral from the organic phase. The mineral component may be obtained using thermal treatment at high temperatures (calcination) to eliminate the organic component and conversely, the latter may be isolated using acid demineralization processes (Martins et al., 2008).

#### **Mineral phase**

Bone mineral is a poorly crystalline calcium-deficient apatite, with a Ca:P ratio that differs from 1.67, which is the theoretical value for pure hydroxyapatite [Ca10(PO4)6(OH)2]. The nonstoichiometric biological apatites contain several ion substitutions. For example, Na+ and Mg2+ may substitute Ca2+ ions, HPO42- ions may substitute phosphate ions, and Cl and Fmay replace OH-. Additionally, carbonate ions, the most abundant substitutions (3-8 wt. %), may occupy either the OH- (type A apatite) or PO43- (type B apatite) positions in the crystal lattice. The mineral component of bone is usually closer to B-type apatite (Landi, 2003; Murugan et al., 2006).

The crystalline structure of carbonated hydroxyapatite belongs to the hexagonal system, although a portion of the bone mineral remains amorphous. In fact, the apatite crystals comprise two different environments: a non-apatitic hydrated domain, containing diverse labile and reactive ions, surrounds a relatively inert and more stable apatitic domain (Fig. 1) (Farlay et al., 2010). In the interface between these domains, labile anions (PO43-, HPO42- and CO3 2-) and cations (Ca2+, Mg2+) are easily and reversibly exchangeable. During initial crystal formation, ionic exchange occurs and ions gradually incorporate into the interior unit cell. The maturation of the mineral is associated with reduction of labile non-apatitic environments while stable apatitic domains augment. Moreover, as bone mineral becomes more mature, it contains less structural defects, both the size and number of crystals increase and its composition becomes closer to the stoichiometric hydroxyapatite. For this reason, crystal maturity is often associated with crystallinity (that augments with crystal size and mineral perfection).

#### **Organic phase**

The organic matrix of bone consists of collagen and a series of non-collagenous proteins and lipids. About 85–90 wt. % of the total bone protein corresponds to Type I collagen fibers.

followed by the replacement of bone with little change in shape, acts as a dynamic response to multiple local and systemic factors (Robling et al., 2006). The bone metabolic functions are regulated by the combined action of two principal cell lineages: the bone forming cells (mainly osteoblasts and osteocytes) and bone resorption cells (osteoclats). Diverse enzymes and many other molecules, such as the important bone morphogenetic proteins (BMPs), participate in

The composition of the extracellular matrix of bone, often described as a two-phase composite, consists of about 65 wt. % mineral component (carbonated hydroxyapatite) and 25 wt. % organic component (mainly type I collagen), being the remaining 10 wt. % water (Judas et al., 2005). These components have extremely different mechanical properties: the mineral is stiff and brittle while the (wet) protein is much more elastic and also much tougher than the mineral. Bone combines both the stiffness and the toughness in an unusual arrangement of material properties that result on a remarkable tensile strength and

The bone components are so tightly embedded that require the use of chemical procedures to obtain an effective separation of the mineral from the organic phase. The mineral component may be obtained using thermal treatment at high temperatures (calcination) to eliminate the organic component and conversely, the latter may be isolated using acid

Bone mineral is a poorly crystalline calcium-deficient apatite, with a Ca:P ratio that differs from 1.67, which is the theoretical value for pure hydroxyapatite [Ca10(PO4)6(OH)2]. The nonstoichiometric biological apatites contain several ion substitutions. For example, Na+ and

may replace OH-. Additionally, carbonate ions, the most abundant substitutions (3-8 wt. %), may occupy either the OH- (type A apatite) or PO43- (type B apatite) positions in the crystal lattice. The mineral component of bone is usually closer to B-type apatite (Landi, 2003;

The crystalline structure of carbonated hydroxyapatite belongs to the hexagonal system, although a portion of the bone mineral remains amorphous. In fact, the apatite crystals comprise two different environments: a non-apatitic hydrated domain, containing diverse labile and reactive ions, surrounds a relatively inert and more stable apatitic domain (Fig. 1) (Farlay et al., 2010). In the interface between these domains, labile anions (PO43-, HPO42- and CO32-) and cations (Ca2+, Mg2+) are easily and reversibly exchangeable. During initial crystal formation, ionic exchange occurs and ions gradually incorporate into the interior unit cell. The maturation of the mineral is associated with reduction of labile non-apatitic environments while stable apatitic domains augment. Moreover, as bone mineral becomes more mature, it contains less structural defects, both the size and number of crystals increase and its composition becomes closer to the stoichiometric hydroxyapatite. For this reason, crystal maturity is often associated with crystallinity (that augments with crystal size and

The organic matrix of bone consists of collagen and a series of non-collagenous proteins and lipids. About 85–90 wt. % of the total bone protein corresponds to Type I collagen fibers.

and F-

Mg2+ may substitute Ca2+ ions, HPO42- ions may substitute phosphate ions, and Cl-

those dynamic cellular processes (Katz et al., 2009; Urist, 1965).

resistance against fracture (Fratzl et al., 2004; Rho et al., 1998).

demineralization processes (Martins et al., 2008).

**Mineral phase** 

Murugan et al., 2006).

mineral perfection).

**Organic phase** 

Fig. 1. Evolution of the apatite crystal and of the surrounding hydrated layer. During maturation and growth of the mineral, the apatitic domain of the crystal increases whereas the hydrated layer decreases, due to continuous ionic exchange between these domains and the solution bath. Soluble charged groups of proteins (Pr) can also participate in the ionic equilibrium of the non-apatitic domain (adapted from Farlay et al., 2010).

This principal component of the organic matrix of bone is a large fibrous protein with a highly repetitive amino acid sequence [Gly (glycine)–X–Y]n (often X is proline and Y is hydroxyproline). Other amino acids may also be found, such as alanine, lysine and hydroxylysine. These monomers are bound together by peptide bonds (between the carboxyl and amino groups of adjacent amino acids) constituting three polypeptide chains (two 1 and one 2 chains) that fold into a unique structure (Fig. 2). This consists of a single uninterrupted triple helix which represents more than 95% of the molecule (tropocollagen) and two non-helical domains (the telopeptides) containing the – COOH and – NH2 terminals of the protein.

The integrity of the collagen molecule is attributed to the formation of intra- and intermolecular cross-linking which contribute to the mechanical strength of the bone organic matrix (Eyre & Wu, 2005; Ottani et al., 2001; Saito & Marumo, 2009). Inter-molecular collagen cross-links can be divided into two types: lysine hydroxylase/lysyl oxidasecontrolled cross-links (enzymatic cross-links) and glycation/oxidation-induced cross-links. In both cases, cross-linking sites are at specific Lys (lysine) or Hyl (hydroxylysine) residues of the collagen molecule. Concerning the first type, the enzymatic activity leads to the formation of Lys and Hyl aldehyde-derived covalent cross-links, in which a molecular fragment binds a telopeptide and a triple helix of adjacent molecules (divalent cross-link). As a consequence of chemical reactions and rearrangements, divalent cross-links may be converted into trivalent cross-links, in which a helix is bound to two telopeptides. As

Characterization of Bone and Bone-Based Graft Materials Using FTIR Spectroscopy 319

Fig. 3. Hierarquical structure of the bone matrix (from Rho et al., 1998).

to cancellous (or trabecular) bone.

**3. FTIR spectroscopy applied to bone characterization** 

adapted to perform".

(Fig. 2) (Nudelman et al., 2010). The association of the mineralized fibrils constitutes larger fibers which are layered in parallel arrangement forming lamellae with an alternate oriented "plywood" appearance (Goldman et al., 2005). Lamellar bone may be disposed concentrically around blood vessels, like rings of a tree, to form osteons (Havers Systems). This organization, from a nano to a macroscopic level, leads essentially to two distinct morphologies: compact bone, composed of densely packed osteons (Silva et al., 2005), and cancellous bone, wherein lamellae are assembled as spaced spicules (trabeculae), originating an interconnected microporous structure (Lozupone & Favia, 1990; Olszta et al., 2007). Compact bone is also called cortical because it always occupies a peripheral position relative

It is interesting to perceive that each hierarchical structure of this anisotropic composite material is optimised to achieve a remarkable mechanical performance, from its basic building block (the mineralized collagen fibril wherein hydroxyapatite crystals have the long axis parallel to the longitudinal axis of the collagen molecules) to the tri-dimensional architecture of compact and cancellous bone. Moreover, each structural level may undergo minor but fundamental modifications, to better adapt and respond to various biomechanical forces (Ottani et al., 2001). As stated in a basic rule of skeletal biology known as Wolff´s law, "the bone´s morphology is a reflection of what function the bone has been built to do or

The Fourier transform infrared (FTIR) spectroscopy is based upon the absorption of IR radiation during vibrational transitions in covalently bound atoms. The frequencies and intensities of the infrared bands provide relevant information on the nature of the molecular

Fig. 2. Structure of the collagen molecule: three intertwined polypeptide chains (two 1 and one 2) constitute a major triple helical domain (tropocollagen) and two non-triple helical domains (telopeptides) at the –COOH and –NH2 terminals of the protein. The typical staggered arrangement found in the collagen fibrils is displayed both in 2D and 3D configurations. The gap regions between the molecules are nucleation sites for hydroxyapatite.

suggested by the sequence of the mechanism of formation, divalent cross-links are classified as immature, whereas trivalent are considered as mature. These cross-links are also commonly named after their reactivity with sodium borohydride: typically, divalent crosslinks are reducible fragments whereas trivalent are non-reducible. As for the second type, glycation cross-links, these are formed between collagen helices (telopeptides are not involved), being frequently associated with the ageing process that is generally deleterious to the function of the bone tissue. In fact, the extent and type of collagen cross-linking are known to change throughout life, leading to less flexible and less elastic properties. Besides the usual bone mass loss associated with age, the accumulation of glycation cross-links and of mature enzymatic cross-links in collagen is also thought to contribute to the fragility and brittleness of bone (Saito & Marumo, 2009; Viguet-Carrin et al., 2010).

#### **Bone matrix structure**

The structure of the mature bone matrix (Fig. 3) is highly organized on several hierarquical levels, starting with nanoscopic crystals of hydroxyapatite that are oriented and aligned within collagen fibrils (Fratzl et al., 2004; Olszta et al., 2007; Rho et al., 1998; Tzaphlidou, 2005; Viguet-Carrin et al., 2005). These result from the self-assembly of collagen molecules and mineral deposition, mainly in the regularly spaced gap regions of the organic structure

1.23 nm

0.24 nm

Tropocollagen NH2 COOH-terminal telopeptides -terminal telopeptides

Fig. 2. Structure of the collagen molecule: three intertwined polypeptide chains (two 1 and one 2) constitute a major triple helical domain (tropocollagen) and two non-triple helical domains (telopeptides) at the –COOH and –NH2 terminals of the protein. The typical staggered arrangement found in the collagen fibrils is displayed both in 2D and 3D

configurations. The gap regions between the molecules are nucleation sites for hydroxyapatite.

suggested by the sequence of the mechanism of formation, divalent cross-links are classified as immature, whereas trivalent are considered as mature. These cross-links are also commonly named after their reactivity with sodium borohydride: typically, divalent crosslinks are reducible fragments whereas trivalent are non-reducible. As for the second type, glycation cross-links, these are formed between collagen helices (telopeptides are not involved), being frequently associated with the ageing process that is generally deleterious to the function of the bone tissue. In fact, the extent and type of collagen cross-linking are known to change throughout life, leading to less flexible and less elastic properties. Besides the usual bone mass loss associated with age, the accumulation of glycation cross-links and of mature enzymatic cross-links in collagen is also thought to contribute to the fragility and

The structure of the mature bone matrix (Fig. 3) is highly organized on several hierarquical levels, starting with nanoscopic crystals of hydroxyapatite that are oriented and aligned within collagen fibrils (Fratzl et al., 2004; Olszta et al., 2007; Rho et al., 1998; Tzaphlidou, 2005; Viguet-Carrin et al., 2005). These result from the self-assembly of collagen molecules and mineral deposition, mainly in the regularly spaced gap regions of the organic structure

brittleness of bone (Saito & Marumo, 2009; Viguet-Carrin et al., 2010).

**Bone matrix structure**

300 nm

1 1

2

Fig. 3. Hierarquical structure of the bone matrix (from Rho et al., 1998).

(Fig. 2) (Nudelman et al., 2010). The association of the mineralized fibrils constitutes larger fibers which are layered in parallel arrangement forming lamellae with an alternate oriented "plywood" appearance (Goldman et al., 2005). Lamellar bone may be disposed concentrically around blood vessels, like rings of a tree, to form osteons (Havers Systems). This organization, from a nano to a macroscopic level, leads essentially to two distinct morphologies: compact bone, composed of densely packed osteons (Silva et al., 2005), and cancellous bone, wherein lamellae are assembled as spaced spicules (trabeculae), originating an interconnected microporous structure (Lozupone & Favia, 1990; Olszta et al., 2007). Compact bone is also called cortical because it always occupies a peripheral position relative to cancellous (or trabecular) bone.

It is interesting to perceive that each hierarchical structure of this anisotropic composite material is optimised to achieve a remarkable mechanical performance, from its basic building block (the mineralized collagen fibril wherein hydroxyapatite crystals have the long axis parallel to the longitudinal axis of the collagen molecules) to the tri-dimensional architecture of compact and cancellous bone. Moreover, each structural level may undergo minor but fundamental modifications, to better adapt and respond to various biomechanical forces (Ottani et al., 2001). As stated in a basic rule of skeletal biology known as Wolff´s law, "the bone´s morphology is a reflection of what function the bone has been built to do or adapted to perform".

#### **3. FTIR spectroscopy applied to bone characterization**

The Fourier transform infrared (FTIR) spectroscopy is based upon the absorption of IR radiation during vibrational transitions in covalently bound atoms. The frequencies and intensities of the infrared bands provide relevant information on the nature of the molecular

Characterization of Bone and Bone-Based Graft Materials Using FTIR Spectroscopy 321

Fig. 4. Typical FTIR spectra of bone, hydroxyapatite and collagen showing the vibrational

originated from the mineral phase, in accordance to its larger proportion in the composite. In particular, the bands at 557 and 600 cm-1 correspond mainly to 4 PO43- bending vibrations, despite some minor contribution from collagen (amide bands) in that region. Moreover, the absorptions at 961 and 1012 cm-1 correspond to the symmetric (1) and asymmetric (3) stretching of phosphate, respectively. Thus, most of the absorptions from phosphate vibrations are clearly observed both in the spectra of bone and of hydroxyapatite. It should be also mentioned that acidic phosphate (HPO42-), a frequent anionic substitution in the crystal lattice of hydroxyapatite, usually originates a band at *ca.* 1110 cm-1, which is

On the other hand, the collagen moiety of bone originates the typical Amide I and Amide II bands at 1634 and 1548 cm-1, respectively. Furthermore, the bands at 1410 and 1445 cm-1 show a different profile and higher intensity in the spectrum of bone relative to its organic model compound. These bands correspond, in fact, to absorptions from CH2 wagging and bending vibrations superimposed with those from asymmetric stretching (3) vibrations of CO32- groups, present as ionic substitutes in the apatite crystal. Carbonate also originates a band at *ca.* 870 cm-1, which is assigned to the 2 bending vibration. This band is characteristic

assignments of the most significant bands.

of a type B apatite.

normally overlapped with that from 3 PO43- vibration.

bonds, their environment and their relative content in the material being analysed. In some cases, some structural information regarding molecular conformations can also be obtained (Nyquist et al., 1997).

In order to obtain FTIR spectra, solid state samples in powdered form are often mixed with the infrared transparent KBr salt and prepared as pellets. Using FTIR in the transmission mode, the incident infrared radiation passes through the sample pellet and the spectrum corresponds to the transmitted fraction of the radiation. Alternatively, using the attenuated total reflection (ATR) mode, the solid samples can be analysed directly (without mixing with other substances) and the infrared spectrum corresponds mainly to the interaction up to about 2 m depth from the surface of the sample.

In addition to the conventional FTIR spectroscopy, a microscopic analysis coupled to the vibrational spectra is also currently used to characterize bone samples, providing further insight into the correlation between structural and functional features. In fact, the introduction of the technique of infrared microspectroscopy (FTIRM) in 1988, followed by the development of the infrared imaging (FTIRI) from 1998, enabled to combine the advantages of hystomorphometry of the bone tissue with spectroscopy in the infrared range. Firstly, FTIRM allowed acquiring the sequential infrared mapping of a sampling area. Later, using focal plane array detector technologies, FTIRI enabled the simultaneous spectral acquisition in each detector pixel, representing an important progress both quantitative and qualitatively. In fact, FTIR imaging provides all the functionality of mapping but does so faster, with better spatial resolution and most importantly, with considerably better sensitivity than FTIRM. Nowadays, both techniques are regularly applied to the *in vitro* analysis of human tissues, namely bone samples (healthy, diseased and archeological bone), as well as animal models (Isaksson et al., 2010; Lebon et al., 2008; Marcott et al., 1998). Samples are usually prepared from thin cut sections that, for FTIRM/FTIRI analyses and in opposition to conventional histological analysis, do not require specific staining. However, samples may have to be fixed with appropriate solvents and embedded on a rigid support (often polymeric resins) and these procedures may cause some deviations of the infrared absorption bands relative to the direct analysis of the material (Aparicio et al., 2002; Paschalis et al., 2011). To overcome this interference, the thin sections may be analyzed in the ATR mode (that most frequently requires little or no sample preparation) but this option may cause some loss of spatial resolution due to intrinsic limitations of the technique (Agilent Technologies, 2011; Petibois et al., 2009).

#### **3.1 FTIR spectrum of the bone matrix**

The composite nature of the bone matrix becomes very clear when its FTIR spectrum is compared with those of the model compounds of its components, hydroxyapatite and collagen (Fig. 4). In fact, the spectrum of bone exhibits all the most intense bands observed in the spectrum of hydroxyapatite (at 500-700 cm-1 and 900-1200 cm-1) and that of collagen (in the 1200-1700 cm-1 and 2800-3700 cm-1 regions), being nearly coincident with the sum of the respective profiles. Nevertheless, there are some new bands (namely at around 870 cm-1 and 1400-1450 cm-1) originated from carbonate substitutions in the crystal lattice of hydroxyapatite.

From the detailed spectral assignments presented in Table 1, some bands must be discussed, particularly in the 500-1700 cm-1 region of the spectrum of bone. The most intense bands are

bonds, their environment and their relative content in the material being analysed. In some cases, some structural information regarding molecular conformations can also be obtained

In order to obtain FTIR spectra, solid state samples in powdered form are often mixed with the infrared transparent KBr salt and prepared as pellets. Using FTIR in the transmission mode, the incident infrared radiation passes through the sample pellet and the spectrum corresponds to the transmitted fraction of the radiation. Alternatively, using the attenuated total reflection (ATR) mode, the solid samples can be analysed directly (without mixing with other substances) and the infrared spectrum corresponds mainly to the interaction up to

In addition to the conventional FTIR spectroscopy, a microscopic analysis coupled to the vibrational spectra is also currently used to characterize bone samples, providing further insight into the correlation between structural and functional features. In fact, the introduction of the technique of infrared microspectroscopy (FTIRM) in 1988, followed by the development of the infrared imaging (FTIRI) from 1998, enabled to combine the advantages of hystomorphometry of the bone tissue with spectroscopy in the infrared range. Firstly, FTIRM allowed acquiring the sequential infrared mapping of a sampling area. Later, using focal plane array detector technologies, FTIRI enabled the simultaneous spectral acquisition in each detector pixel, representing an important progress both quantitative and qualitatively. In fact, FTIR imaging provides all the functionality of mapping but does so faster, with better spatial resolution and most importantly, with considerably better sensitivity than FTIRM. Nowadays, both techniques are regularly applied to the *in vitro* analysis of human tissues, namely bone samples (healthy, diseased and archeological bone), as well as animal models (Isaksson et al., 2010; Lebon et al., 2008; Marcott et al., 1998). Samples are usually prepared from thin cut sections that, for FTIRM/FTIRI analyses and in opposition to conventional histological analysis, do not require specific staining. However, samples may have to be fixed with appropriate solvents and embedded on a rigid support (often polymeric resins) and these procedures may cause some deviations of the infrared absorption bands relative to the direct analysis of the material (Aparicio et al., 2002; Paschalis et al., 2011). To overcome this interference, the thin sections may be analyzed in the ATR mode (that most frequently requires little or no sample preparation) but this option may cause some loss of spatial resolution due to intrinsic limitations of the technique

The composite nature of the bone matrix becomes very clear when its FTIR spectrum is compared with those of the model compounds of its components, hydroxyapatite and collagen (Fig. 4). In fact, the spectrum of bone exhibits all the most intense bands observed in the spectrum of hydroxyapatite (at 500-700 cm-1 and 900-1200 cm-1) and that of collagen (in the 1200-1700 cm-1 and 2800-3700 cm-1 regions), being nearly coincident with the sum of the respective profiles. Nevertheless, there are some new bands (namely at around 870 cm-1 and 1400-1450 cm-1) originated from carbonate substitutions in the crystal lattice of hydroxyapatite. From the detailed spectral assignments presented in Table 1, some bands must be discussed, particularly in the 500-1700 cm-1 region of the spectrum of bone. The most intense bands are

(Nyquist et al., 1997).

about 2 m depth from the surface of the sample.

(Agilent Technologies, 2011; Petibois et al., 2009).

**3.1 FTIR spectrum of the bone matrix** 

Fig. 4. Typical FTIR spectra of bone, hydroxyapatite and collagen showing the vibrational assignments of the most significant bands.

originated from the mineral phase, in accordance to its larger proportion in the composite. In particular, the bands at 557 and 600 cm-1 correspond mainly to 4 PO43- bending vibrations, despite some minor contribution from collagen (amide bands) in that region. Moreover, the absorptions at 961 and 1012 cm-1 correspond to the symmetric (1) and asymmetric (3) stretching of phosphate, respectively. Thus, most of the absorptions from phosphate vibrations are clearly observed both in the spectra of bone and of hydroxyapatite.

It should be also mentioned that acidic phosphate (HPO42-), a frequent anionic substitution in the crystal lattice of hydroxyapatite, usually originates a band at *ca.* 1110 cm-1, which is normally overlapped with that from 3 PO43- vibration.

On the other hand, the collagen moiety of bone originates the typical Amide I and Amide II bands at 1634 and 1548 cm-1, respectively. Furthermore, the bands at 1410 and 1445 cm-1 show a different profile and higher intensity in the spectrum of bone relative to its organic model compound. These bands correspond, in fact, to absorptions from CH2 wagging and bending vibrations superimposed with those from asymmetric stretching (3) vibrations of CO3 2- groups, present as ionic substitutes in the apatite crystal. Carbonate also originates a band at *ca.* 870 cm-1, which is assigned to the 2 bending vibration. This band is characteristic of a type B apatite.

Characterization of Bone and Bone-Based Graft Materials Using FTIR Spectroscopy 323

Fig. 5. Vibration modes for phosphate anions in tetrahedral symmetry and carbonate anions

an active vibration in the infrared because it does not change the dipole moment of the molecule (Nyquist et al., 1997). In carbonated apatites, the CO32- characteristic bands

2, as a single band) and at 1400-1450 cm-1 (

band). Depending on the type of carbonate substitution in the crystal lattice, bands may appear at slightly different wavenumbers: type *A* carbonated apatite is characterized by a

configuration has these bands at about 870, 1430 and 1450 cm-1, respectively (Fleet, 2009; Landi, 2003). The 2 band is often used to estimate the CO32- content of the sample, by calculating the respective area underneath. Moreover, this band may also contain a component at *ca.* 865 cm-1, due to the contribution of non-apatitic (labile) carbonate

As for type I collagen, the organic component of bone, many vibrational bands are originated from the amide groups that constitute the peptide bonds of this protein. The characteristic vibrations of amide groups in collagen are usually combined modes designated as Amide I-VII, Amide A and Amide B (Chang & Tanaka, 2002; Garip & Severcan, 2010). The Amide I band, typically observed in the 1600-1700 cm-1 range, is the most intense absorption band in collagen. It results mainly from the C=O stretching vibration (with minor contribution from C-N stretch) and is directly related to the backbone conformation. Amide II, which originates a band at 1500-1600 cm-1 in collagen, is also conformationally sensitive and results mainly from an out-of-phase combination of N-H inplane bending and C-N stretching vibrations of the peptide linkages. The Amide III band (1200-1300 cm-1), just like the Amide II band, mainly arises from C–N stretch and N–H inplane bending. The main difference between Amide II and III lies in the relative contribution of these modes for the respective mixed vibrations: whereas the N–H bending mode makes the larger contribution for the Amide II, the opposite occurs for the Amide III band. Amide IV is a C=O in-plane deformation and Amide V, VI and VII are out-of-plane motions (Roeges, 1994). The Amide A and Amide B bands, near 3300 cm-1 and 3100 cm-1 respectively, result from a Fermi resonance between the N-H stretch fundamental band and the overtone

3 double band around 1450 and 1540 cm-1, whereas type *B*

3, usually as a double

in D3h symmetry (adapted from Nyquist et al., 1997).

appear at 870-880 cm-1 (

(Verdelis et al., 2007).

of amide II (Lee et al., 1999).

2 band at *ca.* 880 cm-1 and a


Table 1. Band assignments for the FTIR spectra of bone, hydroxyapatite and collagen, illustrated in Fig. 4.

As described, the infrared spectrum is a good diagnostic of the presence of phosphate (and also of carbonate) groups in the mineral component of bone, as well as the presence of amide groups in the organic fraction. In fact, these groups originate many typical absorption bands, as predicted theoretically.

Concerning the phosphate anions in tetrahedral geometry, there are four fundamental vibrations distributed as 1 (symmetric stretch), 2 (bend), 3 (asymmetric stretch), and 4 (bend) (Fig. 5) and these vibration modes (except 2) give rise to medium to strong bands (Nyquist et al., 1997). Typically, in hydroxyapatite, 1 (a non degenerate vibration mode) originates one intense band at around 960 cm-1, and 3 and 4 (triply degenerate modes) originate up to three bands each in the 1000-1100 cm-1 and 500-650 cm-1 regions, respectively. Finally, the doubly degenerate mode 2 may be observed at less than 500 cm-1.

Carbonate anions may also show up to four normal modes of vibration: symmetric stretch (1), out-of-phase bending (2), asymmetric stretch (3) and in-plane CO2 bending (4) (Fig. 5). Whereas 2 is a non degenerate mode, both 3 and 4 are doubly degenerate and 1 is not

(cm-1) Bone Hydroxyapatite Collagen

500-750 cm-1

Amide IV-VII

Wavenumber Vibration modes

630 4 PO43- bend

+Amide (organic) 4 PO43- bend

+Amide (organic) 4 PO43- bend

871 2 CO32- bend 961 <sup>1</sup> PO43- sym stretch <sup>1</sup> PO43- sym stretch 1012 3 PO43- asym stretch 3 PO43- asym stretch

1250 Amide III Amide III 1395 CH2 wag

CH2 wag (organic)

1548 Amide II Amide II 1634 Amide I Amide I

3072 Amide B Amide B 3278 Amide A Amide A 3500 OH OH OH Table 1. Band assignments for the FTIR spectra of bone, hydroxyapatite and collagen,

3 CO32- (mineral) CH2 bend

2850 CH2 sym stretch CH2 sym stretch 2930 CH2 asym stretch CH2 asym stretch

As described, the infrared spectrum is a good diagnostic of the presence of phosphate (and also of carbonate) groups in the mineral component of bone, as well as the presence of amide groups in the organic fraction. In fact, these groups originate many typical absorption

Concerning the phosphate anions in tetrahedral geometry, there are four fundamental

originate up to three bands each in the 1000-1100 cm-1 and 500-650 cm-1 regions,

Carbonate anions may also show up to four normal modes of vibration: symmetric stretch

3 and 

2 (bend),

3 and 

3 (asymmetric stretch), and

4 (triply degenerate modes)

1 is not

4) (Fig. 5).

2) give rise to medium to strong bands

2 may be observed at less than 500 cm-1.

3) and in-plane CO2 bending (

4 are doubly degenerate and

1 (a non degenerate vibration mode)

4

1 (symmetric stretch),

2), asymmetric stretch (

<sup>557</sup>4 PO43- bend (mineral)

<sup>600</sup>4 PO43- bend (mineral)

<sup>1410</sup>3 CO32- (mineral)+

<sup>1445</sup>CH2 bend (organic)+

illustrated in Fig. 4.

bands, as predicted theoretically.

(Nyquist et al., 1997). Typically, in hydroxyapatite,

originates one intense band at around 960 cm-1, and

2 is a non degenerate mode, both

(bend) (Fig. 5) and these vibration modes (except

respectively. Finally, the doubly degenerate mode

vibrations distributed as

1), out-of-phase bending (

(

Whereas

Fig. 5. Vibration modes for phosphate anions in tetrahedral symmetry and carbonate anions in D3h symmetry (adapted from Nyquist et al., 1997).

an active vibration in the infrared because it does not change the dipole moment of the molecule (Nyquist et al., 1997). In carbonated apatites, the CO32- characteristic bands appear at 870-880 cm-1 (2, as a single band) and at 1400-1450 cm-1 (3, usually as a double band). Depending on the type of carbonate substitution in the crystal lattice, bands may appear at slightly different wavenumbers: type *A* carbonated apatite is characterized by a 2 band at *ca.* 880 cm-1 and a 3 double band around 1450 and 1540 cm-1, whereas type *B* configuration has these bands at about 870, 1430 and 1450 cm-1, respectively (Fleet, 2009; Landi, 2003). The 2 band is often used to estimate the CO32- content of the sample, by calculating the respective area underneath. Moreover, this band may also contain a component at *ca.* 865 cm-1, due to the contribution of non-apatitic (labile) carbonate (Verdelis et al., 2007).

As for type I collagen, the organic component of bone, many vibrational bands are originated from the amide groups that constitute the peptide bonds of this protein. The characteristic vibrations of amide groups in collagen are usually combined modes designated as Amide I-VII, Amide A and Amide B (Chang & Tanaka, 2002; Garip & Severcan, 2010). The Amide I band, typically observed in the 1600-1700 cm-1 range, is the most intense absorption band in collagen. It results mainly from the C=O stretching vibration (with minor contribution from C-N stretch) and is directly related to the backbone conformation. Amide II, which originates a band at 1500-1600 cm-1 in collagen, is also conformationally sensitive and results mainly from an out-of-phase combination of N-H inplane bending and C-N stretching vibrations of the peptide linkages. The Amide III band (1200-1300 cm-1), just like the Amide II band, mainly arises from C–N stretch and N–H inplane bending. The main difference between Amide II and III lies in the relative contribution of these modes for the respective mixed vibrations: whereas the N–H bending mode makes the larger contribution for the Amide II, the opposite occurs for the Amide III band. Amide IV is a C=O in-plane deformation and Amide V, VI and VII are out-of-plane motions (Roeges, 1994). The Amide A and Amide B bands, near 3300 cm-1 and 3100 cm-1 respectively, result from a Fermi resonance between the N-H stretch fundamental band and the overtone of amide II (Lee et al., 1999).

Characterization of Bone and Bone-Based Graft Materials Using FTIR Spectroscopy 325

mineral per volume whereas the ratio of the mentioned areas represents the amount of

vibrations in the 900-1200 cm-1 region (Fig. 6). From deconvolution in this spectral range

cm-1 regarded as representative of specific chemical environments: typically, the sub-band at

that at 1030 cm-1, to stoichiometric apatites. Given that the ratio of the areas of the sub-bands at 1020 and 1030 cm-1 has shown to decrease as mineral maturation proceeds, this ratio is often applied to evaluate the maturity of the bone apatite (Miller et al., 2001; Paschalis et al.,

by curve-fitting. Sub-bands associated with the stoichiometry of the apatite (generally used for determination of mineral maturity) and the sub-band from HPO42- (a common substitute

It should be mentioned that since they evolve concomitantly, mineral maturity is normally associated with crystallinity. Thus, the simultaneous expression mineral maturity/ crystallinity is commonly employed, regardless these parameters do not refer to the same mineral properties (Farlay et al. 2011). Although the crystallinity of a mineral is typically calculated from X-ray diffraction, it has been suggested that in the particular case of hydroxyapatite, this parameter may also be determined from the well resolved FTIR band at

half maximum (1/FWHM) of this peak was found to be highly correlated with the state of crystallinity of the apatite (the narrower the peak, the higher the crystallinity), this ratio may be used as a crystallinity index. (Farlay et al. 2011). Moreover, it has been suggested that the

1, 

2- and/or CO3

3 band, being those at *ca.* 1020 and *ca.* 1030

3 phosphate bands with underlying components revealed

4 phosphate vibration. Since the inverse of the full width at

3 PO43- peak at *ca.* 1030 cm-1, most

3 phosphate

2-) and

The mineral maturity is normally estimated based on sub-bands of the

1020 cm-1 is associated to nonstoichiometric apatites (containing HPO4

mineral per amount of collagen per volume analysed).

several individual peaks arise underneath the

2011; Verdelis et al., 2007).

Fig. 6. Spectral region of the

*ca.* 600 cm-1 corresponding to a

1, 

in the apatite crystal) are highlighted (adapted from Verdelis, 2005).

4 PO43- peak at *ca.* 600 cm-1 is more reliable than the

As demonstrated, the spectral contribution of the mineral and organic components to the overall FTIR spectrum of bone may be analyzed almost independently. This enables to estimate, for instance, the relative amount of mineral to organic matter and the proportion of carbonate in the bone apatite. Moreover, since high carbonate (and other ions) substitution correlates with a less mature mineral, FTIR data may also provide a measure of the crystal maturity based on the crystal stoichiometry. On the other hand, the maturity of the organic component of bone may also be evaluated through the analysis of specific amide bands from collagen.

Next section addresses the methodology usually applied to determine these and other parameters important for bone quality, demonstrating the extraordinary potential of FTIR analysis.

### **3.2 Assessing bone quality using FTIR**

Bone quality is a broad term that encompasses factors affecting the geometric and material properties that contribute to fracture resistance. Geometric properties include the macroscopic geometry of the whole bone and the microscopic architecture of the trabeculae. On the other hand, material properties include the composition and arrangement of the primary constituents of bone tissue (collagen and mineral), as well as microdamage and microstructural discontinuities (Boskey, 2006, 2011; Paschalis et al., 1997). As mentioned earlier, the recent interest in bone quality arises from the fact that the traditional measurement of bone strength in clinical practice, namely bone densitometry, does not always reliably predict fracture risk. In fact, the bone mineral density (BMD) measurements primarily show the quantity of bone in the skeleton, overlooking more subtle aspects of bone's properties that may also contribute to its fragility. It is therefore important to understand alterations in bone that occur at the macro-, micro- and nanoscopic levels to determine what parameters affect bone quality and how they change with age or health condition.

FTIR spectroscopy, including FTIRM and FTIRI, providing information on all bone constituents at molecular level, represents an excellent tool to explore bone quality. The most frequently reported outcomes of FTIR related to bone material properties are the relative mineral and organic content, mineral maturity and crystallinity, carbonate substitution into the apatite lattice and collagen crosslinking. Because most of the bands of the bone spectrum are composed of several spectral components, deconvolution methods based on second derivative-spectroscopy and curve fitting have to be applied, being the results normally reported as a percentage of the area of specific underlying bands (Donnely 2010; Petibois et al., 2009). Hence, FTIR data are most commonly used for comparison purposes (between regions of the same sample, among samples or throughout time) rather than in absolute terms.

The proportion between the mineral and organic content in bone expresses its degree of mineralization and can be calculated from the ratio of the integrated area due to phosphate bands (usually 1, 3 in the 900-1200 cm-1 region) to that of a amide band (normally Amide I in the 1600-1700 cm-1 region). This ratio is an important indicator of the maturation or aging of bone and is linearly related to the bone mineral density (BMD) routinely measured in clinical medicine by dual energy X-ray absorptiometry. However, it should be pointed out that these parameters are not expressed in the same basis (BMD quantifies the amount of

As demonstrated, the spectral contribution of the mineral and organic components to the overall FTIR spectrum of bone may be analyzed almost independently. This enables to estimate, for instance, the relative amount of mineral to organic matter and the proportion of carbonate in the bone apatite. Moreover, since high carbonate (and other ions) substitution correlates with a less mature mineral, FTIR data may also provide a measure of the crystal maturity based on the crystal stoichiometry. On the other hand, the maturity of the organic component of bone may also be evaluated through the analysis of specific amide bands from

Next section addresses the methodology usually applied to determine these and other parameters important for bone quality, demonstrating the extraordinary potential of FTIR

Bone quality is a broad term that encompasses factors affecting the geometric and material properties that contribute to fracture resistance. Geometric properties include the macroscopic geometry of the whole bone and the microscopic architecture of the trabeculae. On the other hand, material properties include the composition and arrangement of the primary constituents of bone tissue (collagen and mineral), as well as microdamage and microstructural discontinuities (Boskey, 2006, 2011; Paschalis et al., 1997). As mentioned earlier, the recent interest in bone quality arises from the fact that the traditional measurement of bone strength in clinical practice, namely bone densitometry, does not always reliably predict fracture risk. In fact, the bone mineral density (BMD) measurements primarily show the quantity of bone in the skeleton, overlooking more subtle aspects of bone's properties that may also contribute to its fragility. It is therefore important to understand alterations in bone that occur at the macro-, micro- and nanoscopic levels to determine what

parameters affect bone quality and how they change with age or health condition.

FTIR spectroscopy, including FTIRM and FTIRI, providing information on all bone constituents at molecular level, represents an excellent tool to explore bone quality. The most frequently reported outcomes of FTIR related to bone material properties are the relative mineral and organic content, mineral maturity and crystallinity, carbonate substitution into the apatite lattice and collagen crosslinking. Because most of the bands of the bone spectrum are composed of several spectral components, deconvolution methods based on second derivative-spectroscopy and curve fitting have to be applied, being the results normally reported as a percentage of the area of specific underlying bands (Donnely 2010; Petibois et al., 2009). Hence, FTIR data are most commonly used for comparison purposes (between regions of the same sample, among samples or throughout time) rather

The proportion between the mineral and organic content in bone expresses its degree of mineralization and can be calculated from the ratio of the integrated area due to phosphate

in the 1600-1700 cm-1 region). This ratio is an important indicator of the maturation or aging of bone and is linearly related to the bone mineral density (BMD) routinely measured in clinical medicine by dual energy X-ray absorptiometry. However, it should be pointed out that these parameters are not expressed in the same basis (BMD quantifies the amount of

3 in the 900-1200 cm-1 region) to that of a amide band (normally Amide I

collagen.

analysis.

than in absolute terms.

1, 

bands (usually

**3.2 Assessing bone quality using FTIR** 

mineral per volume whereas the ratio of the mentioned areas represents the amount of mineral per amount of collagen per volume analysed).

The mineral maturity is normally estimated based on sub-bands of the 1, 3 phosphate vibrations in the 900-1200 cm-1 region (Fig. 6). From deconvolution in this spectral range several individual peaks arise underneath the 3 band, being those at *ca.* 1020 and *ca.* 1030 cm-1 regarded as representative of specific chemical environments: typically, the sub-band at 1020 cm-1 is associated to nonstoichiometric apatites (containing HPO4 2- and/or CO3 2-) and that at 1030 cm-1, to stoichiometric apatites. Given that the ratio of the areas of the sub-bands at 1020 and 1030 cm-1 has shown to decrease as mineral maturation proceeds, this ratio is often applied to evaluate the maturity of the bone apatite (Miller et al., 2001; Paschalis et al., 2011; Verdelis et al., 2007).

Fig. 6. Spectral region of the 1, 3 phosphate bands with underlying components revealed by curve-fitting. Sub-bands associated with the stoichiometry of the apatite (generally used for determination of mineral maturity) and the sub-band from HPO42- (a common substitute in the apatite crystal) are highlighted (adapted from Verdelis, 2005).

It should be mentioned that since they evolve concomitantly, mineral maturity is normally associated with crystallinity. Thus, the simultaneous expression mineral maturity/ crystallinity is commonly employed, regardless these parameters do not refer to the same mineral properties (Farlay et al. 2011). Although the crystallinity of a mineral is typically calculated from X-ray diffraction, it has been suggested that in the particular case of hydroxyapatite, this parameter may also be determined from the well resolved FTIR band at *ca.* 600 cm-1 corresponding to a 4 phosphate vibration. Since the inverse of the full width at half maximum (1/FWHM) of this peak was found to be highly correlated with the state of crystallinity of the apatite (the narrower the peak, the higher the crystallinity), this ratio may be used as a crystallinity index. (Farlay et al. 2011). Moreover, it has been suggested that the 4 PO4 3- peak at *ca.* 600 cm-1 is more reliable than the 3 PO4 3- peak at *ca.* 1030 cm-1, most

Characterization of Bone and Bone-Based Graft Materials Using FTIR Spectroscopy 327

Fig. 8. Typical FTIR spectrum of collagen obtained from demineralized bovine bone in the region of Amide I and Amide II, resolved to its underlying components. Sub-bands at 1660 and 1690 cm-1 may be used to determine the maturity of collagen crosslinks (adapted from

Parameter Ratio of the integrated areas or peaks intensities Mineral/organic content Phosphate bands (900-1200 cm-1)/amide I band (1600-1700 cm-1) Mineral maturity 1030 cm-1 sub-band (stochiometric apatites)/1020 cm-1 sub-band

CO32-/PO43- ratio Carbonate bands (850-890 cm-1)/phosphate bands (900-1200 cm-1)

Collagen maturity 1660 cm-1 sub-band (non-reducible/mature cross-links)/1690

Table 2 summarizes the most important parameters generally used to evaluate bone quality derived from FTIR spectroscopy data. In addition, Figure 9 illustrates the spatial distribution of some of these parameters in thin sections of osteoporotic bone samples, as

Bone tissue posses a remarkable capacity to self-repair microdamages. Nevertheless, when a bone defect reaches a critical size, healing usually requires the introduction of grafting materials. Ideally, these should be able to provide a porous scaffold to enable cell invasion

cm-1 sub-band (reducible/immature cross-links)

Ratio of integrated area or intensity of sub-bands at *ca.* 870 cm-1 (B-type carbonate), 880 cm-1 (A-type carbonate) and 865 cm-1

(non-stochiometric)

(labile carbonate)

Table 2. Relevant bone quality parameters derived from the FTIR data.

and attachment, and the subsequent formation of new bone.

**4. Characterization of bone-based graft materials using FTIR spectroscopy** 

Paschalis et al., 2001).

Type of CO32substitution

observed by FTIRI.

frequently associated with crystallinity (Shemesh, 1990). Finally, it should be emphasized that mineral maturity and crystallinity seem to be strongly correlated in synthetic apatites, but poorly correlated in human bone (especially in pathological bone). As expected in such a dynamic remodelling tissue, both parameters depend upon numerous factors including age, nutrition and health condition.

On the other hand, the maturity of the mineral phase is also related to the amount of substitutions in the crystal lattice of hydroxyapatite, being carbonate the most abundant in bone. Carbonate to phosphate ratio indicates the level of carbonate substitution and is calculated as the ratio of the carbonate band integrated area (850-890 cm-1) to that of 1, 3 phosphate bands. This parameter seems a good indicator of the bone turnover and remodeling activity (Isaksson et al 2010). A combination of second derivative spectroscopy and curve fitting of the carbonate band (Fig. 7) reveals whether the carbonate has replaced hydroxide (A-type) or phosphate (B-type) in the apatite crystal, or is loosely attached to its surface (labile CO32-). Considering that acidic phosphate is another eventual substitute in bone apatite, mineral maturity may also be estimated from the ratio between the overall 1, 3 phosphate band and the HPO42- sub-band at *ca.* 1110 cm-1.

Fig. 7. Spectral region of the 2 carbonate band with components determined by curvefitting, indicating the different types of carbonate substitution (adapted from Verdelis, 2005).

As previously mentioned, one of the most distinct features of type I collagen in mineralized tissues is its cross-linking chemistry. The intermolecular cross-linking provides high tensile strength and viscoelasticity to the fibrillar matrices. FTIR has been used to analyse the secondary structure of collagen, namely through the analysis of Amide I band. Information on protein structure may be obtained from the underlying bands of Amide I spectral region by resolution enhancement techniques (Fig. 8) (Paschalis et al., 2003). Of particular interest are two sub-bands: 1660 and 1690 cm-1, related to the presence of non-reducible (mature) and reducible (immature) collagen cross-links, respectively. Thus, the ratio of the integrated areas of these bands (1660/1690) can be used as an indicator of the collagen maturity (Paschalis et al., 2011; Paschalis et al., 2001; Saito & Marumo, 2009).

frequently associated with crystallinity (Shemesh, 1990). Finally, it should be emphasized that mineral maturity and crystallinity seem to be strongly correlated in synthetic apatites, but poorly correlated in human bone (especially in pathological bone). As expected in such a dynamic remodelling tissue, both parameters depend upon numerous factors including age,

On the other hand, the maturity of the mineral phase is also related to the amount of substitutions in the crystal lattice of hydroxyapatite, being carbonate the most abundant in bone. Carbonate to phosphate ratio indicates the level of carbonate substitution and is calculated as the ratio of the carbonate band integrated area (850-890 cm-1) to that of

phosphate bands. This parameter seems a good indicator of the bone turnover and remodeling activity (Isaksson et al 2010). A combination of second derivative spectroscopy and curve fitting of the carbonate band (Fig. 7) reveals whether the carbonate has replaced hydroxide (A-type) or phosphate (B-type) in the apatite crystal, or is loosely attached to its surface (labile CO32-). Considering that acidic phosphate is another eventual substitute in bone apatite, mineral maturity may also be estimated from the ratio between the overall 1,

2 carbonate band with components determined by curve-

fitting, indicating the different types of carbonate substitution (adapted from Verdelis, 2005).

As previously mentioned, one of the most distinct features of type I collagen in mineralized tissues is its cross-linking chemistry. The intermolecular cross-linking provides high tensile strength and viscoelasticity to the fibrillar matrices. FTIR has been used to analyse the secondary structure of collagen, namely through the analysis of Amide I band. Information on protein structure may be obtained from the underlying bands of Amide I spectral region by resolution enhancement techniques (Fig. 8) (Paschalis et al., 2003). Of particular interest are two sub-bands: 1660 and 1690 cm-1, related to the presence of non-reducible (mature) and reducible (immature) collagen cross-links, respectively. Thus, the ratio of the integrated areas of these bands (1660/1690) can be used as an indicator of the collagen maturity

3 phosphate band and the HPO42- sub-band at *ca.* 1110 cm-1.

(Paschalis et al., 2011; Paschalis et al., 2001; Saito & Marumo, 2009).

1, 3

nutrition and health condition.

Fig. 7. Spectral region of the

Fig. 8. Typical FTIR spectrum of collagen obtained from demineralized bovine bone in the region of Amide I and Amide II, resolved to its underlying components. Sub-bands at 1660 and 1690 cm-1 may be used to determine the maturity of collagen crosslinks (adapted from Paschalis et al., 2001).


Table 2. Relevant bone quality parameters derived from the FTIR data.

Table 2 summarizes the most important parameters generally used to evaluate bone quality derived from FTIR spectroscopy data. In addition, Figure 9 illustrates the spatial distribution of some of these parameters in thin sections of osteoporotic bone samples, as observed by FTIRI.

#### **4. Characterization of bone-based graft materials using FTIR spectroscopy**

Bone tissue posses a remarkable capacity to self-repair microdamages. Nevertheless, when a bone defect reaches a critical size, healing usually requires the introduction of grafting materials. Ideally, these should be able to provide a porous scaffold to enable cell invasion and attachment, and the subsequent formation of new bone.

Characterization of Bone and Bone-Based Graft Materials Using FTIR Spectroscopy 329

Currently, the most common alternatives to bone-based grafts are synthetic biomaterials (often composites made of Ca/P ceramics and polymers) (Kikuchi et al., 2001). Although the raw materials of these composites are selected to fulfil many requirements (such as biocompatibility, osteoconductivity and non-toxicity), being often produced/modified to mimic the characteristics of natural bone, the best clinical results still derive from the application of autografts. Nonetheless, synthetic materials not only overcome most disadvantages of the natural bone grafts, but also provide more consistent and controllable properties. This is the case of diverse ceramic materials, namely hydroxyapatite, that besides being applied as components of composites, are also frequently used alone. In fact, synthetic hydroxyapatite is the most commonly used graft material mostly due to its resemblance with the bone mineral and similar osteocondutive potential. Synthetic hydroxyapatite, however, possesses some limitations including poor mechanical properties (specially when exposed to wet environments) and lack of osteoinductive properties. In addition, for the time being, synthetic apatites are not yet able to duplicate the composition and structure of the mineral component of bone, known to influence the grafts osteointegration. In fact, natural hydroxyapatite is nonstoichiometric and contains carbonate and other ions built into its structure. Although the presence of these ions is very low, they play a vital role in the biological reactions associated with bone metabolism (Barrere et al., 2006; Joschek et al., 2000). It has been shown that carbonate incorporation tends to decrease the crystallinity and to augment the solubility of hydroxyapatite, enhancing its biodegradation rate. Thus, synthetic hydroxyapatite is frequently modified chemically in order to include additional

A given type of graft can function by more than one mechanism. For instance, autografts are osteoconductive, osteoinductive and osteogenic whereas allografts and xenografts are mainly osteoconductive. Evidences that natural apatites also possess osteocondutive properties has stimulated their application as grafting materials, alone or combined with other materials. Currently, this mineral component of bone is usually obtained from calcined xenografts, being commercially available from different animal origins, under various formulations. On the other hand, in addition to osteocondutive properties, demineralized bone grafts have demonstrated to also possess osteoinductive capacities, which are mainly attributed to the exposure of bone morphogenetic proteins (BMPs) caused by the demineralization procedure. Thus, not only natural bone grafts, but also calcined and demineralized bone grafts constitute very important options that require extensive research regarding materials characterization, processing methods and application conditions, in

properties favorable to bone grafting.

order to help the prediction of their clinical outcome.

**4.1 Bone-derived hydoxyapatite: Influence of the calcination temperature** 

One strategy to improve non-autograft materials while maintaining their advantages regarding their chemical and physical properties is to process natural bone, for example by heat treatment (Haberko et al., 2006; Joschek et al., 2000; Murugan et al., 2006; Ooi et al., 2007). Natural hydroxyapatite obtained in this way has the advantage of inheriting the chemical composition and structure of the raw material (Catanese et al., 1990; Etok et al., 2007; Murugan et al., 2003), being therefore an alternative solution for numerous applications based on its analogous synthetic products. As compared to allogeneic bone, hydroxyapatite derived from xenogeneic bone (usually of bovine origin) is considered a good option because it is easier to obtain at lower cost and is available in unlimited supply.

Fig. 9. Typical images obtained by FTIRI from an osteoporotic human iliac crest showing the spatial distribution of bone quality parameters in regions of cortical and cancellous bone: mineral/organic ratio (A), mineral maturity (B), and collagen maturity (C) (from Boskey, 2011).

A standard strategy applied when a bone loss occurs is using bone grafts which include autografts, allografts and xenografts (Bauer & Muschler, 2000; Torroni, 2009). An autograft consists of a tissue that is transferred from one site of the organism of an individual to a different location, an allograft represents a tissue transfer between individuals of the same species, and xenografts, between different species. Bone autografts, the "gold standard" in bone grafting procedures, have the advantage of no adverse immunological response and thus represent the best for inducing new bone formation. However, they are usually available in limited quantity (depending on donor site anatomy) and harvesting requires additional surgery, resulting in increased morbidity for the patient. Human allografts, which can be appropriately preserved in bone banks being thus available in greater abundance, represent a valuable alternative to autografts (Judas et al., 2005). Even so, the application of allografts is often limited due to uncertainty on compatibility and risk of disease transmission. Xenografts, most often of bovine or porcine origin, have no limitations regarding availability but, similarly to allogenous bone, can introduce risk of rejection and disease transmission.

Fig. 9. Typical images obtained by FTIRI from an osteoporotic human iliac crest showing the spatial distribution of bone quality parameters in regions of cortical and cancellous bone: mineral/organic ratio (A), mineral maturity (B), and collagen maturity (C) (from Boskey,

A standard strategy applied when a bone loss occurs is using bone grafts which include autografts, allografts and xenografts (Bauer & Muschler, 2000; Torroni, 2009). An autograft consists of a tissue that is transferred from one site of the organism of an individual to a different location, an allograft represents a tissue transfer between individuals of the same species, and xenografts, between different species. Bone autografts, the "gold standard" in bone grafting procedures, have the advantage of no adverse immunological response and thus represent the best for inducing new bone formation. However, they are usually available in limited quantity (depending on donor site anatomy) and harvesting requires additional surgery, resulting in increased morbidity for the patient. Human allografts, which can be appropriately preserved in bone banks being thus available in greater abundance, represent a valuable alternative to autografts (Judas et al., 2005). Even so, the application of allografts is often limited due to uncertainty on compatibility and risk of disease transmission. Xenografts, most often of bovine or porcine origin, have no limitations regarding availability but, similarly to allogenous bone, can introduce risk of rejection and

2011).

disease transmission.

Currently, the most common alternatives to bone-based grafts are synthetic biomaterials (often composites made of Ca/P ceramics and polymers) (Kikuchi et al., 2001). Although the raw materials of these composites are selected to fulfil many requirements (such as biocompatibility, osteoconductivity and non-toxicity), being often produced/modified to mimic the characteristics of natural bone, the best clinical results still derive from the application of autografts. Nonetheless, synthetic materials not only overcome most disadvantages of the natural bone grafts, but also provide more consistent and controllable properties. This is the case of diverse ceramic materials, namely hydroxyapatite, that besides being applied as components of composites, are also frequently used alone. In fact, synthetic hydroxyapatite is the most commonly used graft material mostly due to its resemblance with the bone mineral and similar osteocondutive potential. Synthetic hydroxyapatite, however, possesses some limitations including poor mechanical properties (specially when exposed to wet environments) and lack of osteoinductive properties. In addition, for the time being, synthetic apatites are not yet able to duplicate the composition and structure of the mineral component of bone, known to influence the grafts osteointegration. In fact, natural hydroxyapatite is nonstoichiometric and contains carbonate and other ions built into its structure. Although the presence of these ions is very low, they play a vital role in the biological reactions associated with bone metabolism (Barrere et al., 2006; Joschek et al., 2000). It has been shown that carbonate incorporation tends to decrease the crystallinity and to augment the solubility of hydroxyapatite, enhancing its biodegradation rate. Thus, synthetic hydroxyapatite is frequently modified chemically in order to include additional properties favorable to bone grafting.

A given type of graft can function by more than one mechanism. For instance, autografts are osteoconductive, osteoinductive and osteogenic whereas allografts and xenografts are mainly osteoconductive. Evidences that natural apatites also possess osteocondutive properties has stimulated their application as grafting materials, alone or combined with other materials. Currently, this mineral component of bone is usually obtained from calcined xenografts, being commercially available from different animal origins, under various formulations. On the other hand, in addition to osteocondutive properties, demineralized bone grafts have demonstrated to also possess osteoinductive capacities, which are mainly attributed to the exposure of bone morphogenetic proteins (BMPs) caused by the demineralization procedure. Thus, not only natural bone grafts, but also calcined and demineralized bone grafts constitute very important options that require extensive research regarding materials characterization, processing methods and application conditions, in order to help the prediction of their clinical outcome.

#### **4.1 Bone-derived hydoxyapatite: Influence of the calcination temperature**

One strategy to improve non-autograft materials while maintaining their advantages regarding their chemical and physical properties is to process natural bone, for example by heat treatment (Haberko et al., 2006; Joschek et al., 2000; Murugan et al., 2006; Ooi et al., 2007). Natural hydroxyapatite obtained in this way has the advantage of inheriting the chemical composition and structure of the raw material (Catanese et al., 1990; Etok et al., 2007; Murugan et al., 2003), being therefore an alternative solution for numerous applications based on its analogous synthetic products. As compared to allogeneic bone, hydroxyapatite derived from xenogeneic bone (usually of bovine origin) is considered a good option because it is easier to obtain at lower cost and is available in unlimited supply.

Characterization of Bone and Bone-Based Graft Materials Using FTIR Spectroscopy 331

structure may be partially degraded by heating at temperatures higher than 800 ºC, originating -tricalcium phosphate (Ca3(PO4)2) and/or CaO. However, the presence of these products in the mineral is better detected by X-ray diffraction than by FTIR spectroscopy (Etok et al., 2007; Haberko et al., 2006; Mkukuma et al., 2004; Ooi et al., 2007; Rogers &

SEM pictures (Fig. 11) of the human bone sample calcined at various temperatures showed that the samples surfaces no longer present the characteristic concentric lamellae around the Havers Canals, due to the elimination of collagen with heat treatment. Nonetheless, the basic microstructure of cortical bone (Havers Canals and osteocyte lacunar spaces, in particular) was preserved after calcination. Additionally, it was also apparent that the size of the apatite crystals increased with temperature, resulting on increased crystalinity (as

Fig. 11. Scanning Electron Microscopy images of human cortical bone calcined at 600, 900 and 1200 ºC (from left to right) showing the increase in crystal size with temperature

Moreover, heat treatment also resulted in different porosity characteristics. Regarding porosity and pore size distribution, assessed by mercury intrusion (Fig. 12), it was clear that samples calcined at 600 ºC exhibit the highest porosity, around 50%, which, for a compact bone, is quite relevant. However, as the heating temperature increases, the porous structure condenses, sintering at very high temperatures (1200 ºC) and originating porosity values

Demineralized bone matrix (DBM) is often applied in orthopedics, periodontics, oral and maxillofacial surgery because of its inherent osteoconductive and osteoinductive properties, generally related, as mentioned, to bone morphogenetic proteins (BMPs) (Bauer & Muschler, 2000; Eppley et al., 2005; Katz et al., 2009). In fact, as mineral is removed, the matrix associated BMPs become available rendering DBM grafts osteoinductive (Pietrzak et al., 2009). These grafts can be used either alone (Libin et al., 1975; Morone & Boden, 1998; Pietrzak et al., 2005) or in combination with bone marrow, autogenous bone graft, or other materials (Kim et al., 2002; Kucukkolbasi et al., 2009; Nade & Burwell, 1977). Additionally, DBM exhibits elastic features, being easily shaped to fill osteochondral lesions with different

Despite the extensive use of DBM, conflicting results have been published in the literature regarding its bone-inducing abilities. This may be a consequence of following different

Daniels 2002).

confirmed by X-ray diffraction).

(Figueiredo et al. 2010).

comparable to those of the non-calcined samples.

shapes and sizes (Costa et al., 2001).

**4.2 Demineralized bone matrix: Effect of the acid concentration** 

The problem that arises with heat treatment at elevated temperatures is that the biogenic composition and structure of bone mineral can change and this may affect the efficacy of this material (Etok et al., 2007; Hillera et al., 2003; Ooi et al., 2007). Although there is some controversy regarding the onset of chemical and structural changes as a consequence of heat treatment, it has been reported that changes in the mineral phase of bone are not significant until degradation and combustion of most of the more labile organic components occurs (around 500 ºC) (Etok et al., 2007; Mkukuma et al., 2004; Murugan et al., 2006; Ooi et al., 2007). Regarding sample mineralogy, it is generally accepted that heat treatment promotes the crystallinity of bone derived hydroxyapatite and increases the crystallite size (Etok et al., 2007; Hillera et al., 2003; Ooi et al., 2007).

A recent study concerning cortical bone samples of different origins (human and animal) subjected to different calcination temperatures (600, 900 and 1200 ºC) revealed that the calcination temperature highly affects the properties of the bone samples (Figueiredo et al., 2010). As expected, higher temperatures led to more pure forms of hydroxyapatite, with higher crystallinity degrees and larger crystallite sizes and a less porous structure. Furthermore, samples heated to the same temperature exhibited similar characteristics, regardless their origin.

FTIR spectra (Fig. 10) indicated that the organic constituents were no longer present in the samples calcined at 600 ºC, suggesting that this temperature is adequate to obtain proteinfree samples. Moreover, these spectra have also revealed that, at this temperature, a carbonated apatite was obtained, being the carbonate removed from the mineral at higher temperatures. Furthermore, no new mineral phases were evident at higher temperatures, in good agreement with the results of thermal analysis, X-ray diffraction and additional confirmation by chemical analysis. Nevertheless, it has been reported that the apatite base

Fig. 10. FTIR spectra of human cortical bone before (control) and after calcination at 600, 900 and 1200 ºC. \*Bands attributed to lattice carbonate vibrations (Figueiredo et al., 2010).

The problem that arises with heat treatment at elevated temperatures is that the biogenic composition and structure of bone mineral can change and this may affect the efficacy of this material (Etok et al., 2007; Hillera et al., 2003; Ooi et al., 2007). Although there is some controversy regarding the onset of chemical and structural changes as a consequence of heat treatment, it has been reported that changes in the mineral phase of bone are not significant until degradation and combustion of most of the more labile organic components occurs (around 500 ºC) (Etok et al., 2007; Mkukuma et al., 2004; Murugan et al., 2006; Ooi et al., 2007). Regarding sample mineralogy, it is generally accepted that heat treatment promotes the crystallinity of bone derived hydroxyapatite and increases the crystallite size (Etok et al.,

A recent study concerning cortical bone samples of different origins (human and animal) subjected to different calcination temperatures (600, 900 and 1200 ºC) revealed that the calcination temperature highly affects the properties of the bone samples (Figueiredo et al., 2010). As expected, higher temperatures led to more pure forms of hydroxyapatite, with higher crystallinity degrees and larger crystallite sizes and a less porous structure. Furthermore, samples heated to the same temperature exhibited similar characteristics,

FTIR spectra (Fig. 10) indicated that the organic constituents were no longer present in the samples calcined at 600 ºC, suggesting that this temperature is adequate to obtain proteinfree samples. Moreover, these spectra have also revealed that, at this temperature, a carbonated apatite was obtained, being the carbonate removed from the mineral at higher temperatures. Furthermore, no new mineral phases were evident at higher temperatures, in good agreement with the results of thermal analysis, X-ray diffraction and additional confirmation by chemical analysis. Nevertheless, it has been reported that the apatite base

Fig. 10. FTIR spectra of human cortical bone before (control) and after calcination at 600, 900 and 1200 ºC. \*Bands attributed to lattice carbonate vibrations (Figueiredo et al., 2010).

2007; Hillera et al., 2003; Ooi et al., 2007).

regardless their origin.

structure may be partially degraded by heating at temperatures higher than 800 ºC, originating -tricalcium phosphate (Ca3(PO4)2) and/or CaO. However, the presence of these products in the mineral is better detected by X-ray diffraction than by FTIR spectroscopy (Etok et al., 2007; Haberko et al., 2006; Mkukuma et al., 2004; Ooi et al., 2007; Rogers & Daniels 2002).

SEM pictures (Fig. 11) of the human bone sample calcined at various temperatures showed that the samples surfaces no longer present the characteristic concentric lamellae around the Havers Canals, due to the elimination of collagen with heat treatment. Nonetheless, the basic microstructure of cortical bone (Havers Canals and osteocyte lacunar spaces, in particular) was preserved after calcination. Additionally, it was also apparent that the size of the apatite crystals increased with temperature, resulting on increased crystalinity (as confirmed by X-ray diffraction).

Fig. 11. Scanning Electron Microscopy images of human cortical bone calcined at 600, 900 and 1200 ºC (from left to right) showing the increase in crystal size with temperature (Figueiredo et al. 2010).

Moreover, heat treatment also resulted in different porosity characteristics. Regarding porosity and pore size distribution, assessed by mercury intrusion (Fig. 12), it was clear that samples calcined at 600 ºC exhibit the highest porosity, around 50%, which, for a compact bone, is quite relevant. However, as the heating temperature increases, the porous structure condenses, sintering at very high temperatures (1200 ºC) and originating porosity values comparable to those of the non-calcined samples.
