**4. Another plasma biomarker study**

the early stage of AD, whereas levels of CSF Aβ42, but not TAU or P-TAU181, was positively

It is thought that the production and accumulation of unfavorable Aβ species proceeds over time as the disease progresses. Abnormal activity by the Aβ species is initiated before patho‐ logical change and reaches a plateau before the clinical symptoms appear. Thereafter, elevation of TAU and P-TAU that are the biomarkers for neuronal injury, dysfunction, and degeneration, become apparent in the later stage of the disease and correlate with clinical symptom severity [8]. On the other hand, MRI imaging is valuable as it is the last biomarker to show abnormality. As such, MRI retains a closer relationship with cognitive performance later on in the disease compared with other biomarkers. Moreover, none of the biomarkers is stable; that is, the rate

The revised guideline for AD diagnosis was released by a working group from the National Institute of Aging in 2011, in which both CSF and imaging biomarkers have been implemented. The new guideline provides evidence of an ongoing AD pathophysiological process, and it is also possible to make a preclinical diagnosis of MCI due to AD [21–23]. AD is classified into

Since fluid biomarkers of either CSF or blood plasma can serve as objective criteria for dementia diagnosis, this guideline is aiming at early and reliable diagnosis. However, it is clear that at present no single biomarker plays a sufficient discriminatory role in screening for future

Compared with CSF, blood sampling is a less invasive procedure, more easily accessible, and cost reductive, thus the finding of reliable blood biomarkers for AD is being given the highest priority. There has been an increasing research effort to examine the potential biomarkers of AD in blood plasma. However, for blood-based biomarkers, it has to be noted that blood plasma contains several tens of thousands of different proteins. In addition, the range of protein concentrations are extremely varied (attaining to 12 orders of magnitude), and the lower the concentration, the greater the diversity of proteins [24, 25]. Moreover, none of the current methods allows us to directly detect components in the low concentration region [25]. These conditions make it extremely challenging or almost impossible to directly analyze blood, even though possible biomarker candidates are more likely to be present in the areas of low concentration. The change in concentration of the blood components may often be on a very small scale and cover a wide range of both peripheral and central processes. Additionally, the less abundant proteins may be masked by highly abundant plasma proteins such as albumin and immunoglobulin. Therefore, focusing on concentration change of a particular AD-specific marker, which may be in low concentration, can be the most challenging to discover [26, 27].

It was reported that the BBB is disrupted resulting in increased permeability with aging and in AD [28, 29]. It is also thought that this event occurs in the relatively early stage of the aging

three separate stages: preclinical AD, MCI due to AD, and AD with dementia.

correlated with whole brain volume in nondemented controls [17].

of change for each biomarker is not linear over time [8].

development of late-onset AD or dementia.

**3. Plasma markers**

70 Update on Dementia

The brain-derived proteins present in blood plasma are limited compared with those in CSF due to the presence of the BBB. It is also likely that if potential brain-derived proteins are present in blood plasma, it is conceivable they are considerably diluted in the large volume of plasma and underwent proteolysis and excretion. These possible events make the study more challenging. As mentioned previously, in plasma, there are several tens of thousands of different proteins present at concentrations in the millimolar to femtomolar or lower range. This extremely varied range of protein concentrations in plasma makes it almost impossible to directly analyze low concentration components.

Therefore, in our study, instead of direct examination of plasma, we initially utilized a cell culture model, mouse primary culture neuron. After Aβ-treatment, we identified proteins present outside of the cells (culture supernatant), in which Aβ-dependent secreted proteins are expected to be present, using a proteomic approach, and focused on the proteins that were increased by Aβ-treatment, and discovered a biomarker candidate. Ultimately, we verified the potential candidate with animal model (transgenic mice) and human plasma samples (**Figure 1**).

**Figure 1.** Process of biomarker identification (our study).

The cytotoxicity due to Aβ<sup>42</sup> is thought to be directly linked to neural cell death [1]. Amyloiddependent neurotoxicity is known to perturb Ca2+ homeostasis in neuronal cells [51]. Possibly, Aβ impairs membrane Ca2+ pumps and enhances Ca2+ influx through voltage-dependent channels and ionotropic glutamate receptors (**Figure 2**).

By focusing on this mechanism, we identified the Ca2+-related protein as a potential biomarker for AD using primary neurons as a cell culture model [52]. Since phosphatidylserine (PS) is flip-flopped and appears in the outer layer of the plasma membrane during the apoptotic process, we focused on PS-binding proteins in the culture supernatant and used a unique method to identify a potential biomarker candidate.

Thermoresponsive magnetic nanoparticles disperse well in an aqueous solution at a temper‐ ature below 10°C and are aggregated and become responsive to magnets at 20°C or higher. In this study, we coated magnetic beads with thermoresponsive polymers (polyethyleneimine) together with myristate and then coated them with PS [52]. We mixed these particles with a culture supernatant in the presence of Ca2+ and collected the PS binding fraction with ethylene glycol-bis(β-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA). After running SDS-PAGE, we performed in-gel digestion with trypsin and analyzed the tryptic peptides by reverse-phase liquid chromatography coupled with MALDI TOF/TOF MS spectrometry and performed database analysis for peptide sequencing. From this proteomic approach, about 240 types of proteins were indicated to be increased in the Aβ42-treated sample, compared with the control, suggesting that they were upregulated by Aβ42. From among these proteins, we focused on annexin A5, one of the annexin family proteins that commonly bind Ca2+ and phospholipid. It was shown that annexin A5 was augmented in both the brain and blood plasma in an ADmodel mouse (Tg2576 transgenic mice), overexpressing mutant human APP [52]. Technetiumlabeled annexin A5 was detectable in the brain after intravenous injection in humans, showing that annexin A5 crosses the BBB [53].

**Figure 2.** Aβ-dependent perturbation of calcium homeostasis in AD.

#### **4.1. Methods**

This extremely varied range of protein concentrations in plasma makes it almost impossible

Therefore, in our study, instead of direct examination of plasma, we initially utilized a cell culture model, mouse primary culture neuron. After Aβ-treatment, we identified proteins present outside of the cells (culture supernatant), in which Aβ-dependent secreted proteins are expected to be present, using a proteomic approach, and focused on the proteins that were increased by Aβ-treatment, and discovered a biomarker candidate. Ultimately, we verified the potential candidate with animal model (transgenic mice) and human plasma samples

The cytotoxicity due to Aβ<sup>42</sup> is thought to be directly linked to neural cell death [1]. Amyloiddependent neurotoxicity is known to perturb Ca2+ homeostasis in neuronal cells [51]. Possibly, Aβ impairs membrane Ca2+ pumps and enhances Ca2+ influx through voltage-dependent

By focusing on this mechanism, we identified the Ca2+-related protein as a potential biomarker for AD using primary neurons as a cell culture model [52]. Since phosphatidylserine (PS) is flip-flopped and appears in the outer layer of the plasma membrane during the apoptotic process, we focused on PS-binding proteins in the culture supernatant and used a unique

Thermoresponsive magnetic nanoparticles disperse well in an aqueous solution at a temper‐ ature below 10°C and are aggregated and become responsive to magnets at 20°C or higher. In this study, we coated magnetic beads with thermoresponsive polymers (polyethyleneimine) together with myristate and then coated them with PS [52]. We mixed these particles with a culture supernatant in the presence of Ca2+ and collected the PS binding fraction with ethylene glycol-bis(β-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA). After running SDS-PAGE, we performed in-gel digestion with trypsin and analyzed the tryptic peptides by reverse-phase

to directly analyze low concentration components.

**Figure 1.** Process of biomarker identification (our study).

channels and ionotropic glutamate receptors (**Figure 2**).

method to identify a potential biomarker candidate.

(**Figure 1**).

72 Update on Dementia

To quantify plasma annexin A5, we previously established the chemiluminescent enzyme immunoassay (CLEIA) system with two clones of monoclonal antibodies against human annexin A5: one clone was conjugated to a glass bead and used for trapping annexin A5 present in the blood plasma; the other clone was labeled with horseradish peroxidase (HRP) and used for quantification of the trapped annexin A5 [52]. The HRP catalyzes the oxidation of a luminol solution that includes a phenol-derivative acting as an enhancer, and produces light. This system was useful to quantify plasma annexin A5 in the range from 0.16 to 20.0 ng/ml [52]. We obtained blood samples from 150 AD, 50 DLB, 14 mild cognitive impairment (MCI), and six depression patients, and 298 healthy elderly individuals from the senior citizen's clubs. AD patients met NINCDS-ADRDA [54] and DLB patients diagnosed as probable DLB according to the latest consensus diagnostic criteria [55]. Statistical analysis was done using JMP version 9.0.0 (SAS Institute Inc., Cary, NC, USA). The mean response of each experimental group was compared with its simultaneous control by unpaired Student's *t*-test, setting a significant difference at *P* < 0.05. To examine the plasma annexin A5 levels in diagnoses of AD, DLB, and MCI, logistic regression modeling was employed to construct receiver operator (ROC) curves.

The plasma level of annexin A5 was significantly increased in AD patients compared to that of a control group (*P*-value < 0.0001 in the logistic regression analysis), suggesting that annexin A5 is a potentially positive biomarker for AD (**Figure 3**) [52].

**Figure 3.** Comparison of plasma levels of annexin A5 in AD, DLB, MCI, depression, and age-matched healthy control.

For quantification of plasma annexin A5, we used a previously established chemiluminescent enzyme immunoassay system with monoclonal antibodies against human annexin A5 [52] (see Section 4.1). Individual plasma annexin A5 concentration is plotted in (A). The probability of either AD, DLB, or MCI can be predicted by a logistic regression model with the plasma level of annexin A5. Receiver operating characteristic (ROC) curves are shown in (B)–(D). The areas under the curve are 86.3%, 83.8%, and 91.6% for AD (B), DLB (C), and MCI (D), respectively. AD, Alzheimer's disease; DLB, dementia with Lewy bodies; MCI, mild cognitive impairment.

As annexin A5 binds not only phospholipids but also Ca2+, it might have a role in protecting against Ca2+-induced damage by chelating elevated intracellular Ca2+. A defensive role against apoptosis induced by the participation of annexin A5 was also reported, in that annexin A5 plays a role in reducing the toxicity of the amyloidogenic proteins through interaction with them, such as amyloid polypeptides and α-synuclein [56].

On the other hand, dementia with Lewy bodies (DLB) shares clinical and pathological features with other dementia subtypes such as AD, vascular dementia, and Parkinson's disease (PD), which makes it difficult to distinguish in clinical practice. Lewy bodies are often found in the brains of AD patients. Also, the lack of valid and reliable methods for assessing the core clinical symptoms of both AD and DLB makes its identification even more difficult. We analyzed plasma level of annexin A5 in DLB. When average concentrations of plasma annexin A5 are compared among AD, DLB, and control groups, the values of AD and DLB were significantly higher than healthy control subjects (**Figure 3A**). Moreover, the ROC analyses showed good separation of patients with either AD or DLB from the control group (healthy volunteers) (**Figure 3B** and **C**) [57]. These suggest that annexin A5 is a potential biomarker for both AD and DLB. There is a similarity between AD and DLB. Lewy bodies are often found in the brains of AD patients. The therapeutic agent, acetylcholinesterase inhibitor, is effective not only in AD but also in DLB. From these results, annexin A5 reflects the above-mentioned similarity of AD and DLB.

To examine when annexin A5 becomes elevated during the course of disease development, we analyzed plasma samples from MCI patients (early stage of dementia). Average concen‐ tration was significantly higher than for the control group and the level was comparable with that of AD (**Figure 3A**). The areas under the ROC curve was 91.6% (*P* < 0.001) (**Figure 3D**), suggesting that annexin A5 is also a potential biomarker for MCI. Therefore, it is presumable that elevation of annexin A5 is likely to take place from the early developmental stage of AD. Although a sample number is very limited, plasma annexin A5 level in depression was comparable with the control (**Figure 3A**).

We next tracked plasma level of annexin A5 over a 3-year period in late stage AD patients. The plasma level of annexin A5 tended to be unchanged or slightly decreased, which indicates that biosynthesis of annexin A5 might be downregulated during the late stage, due to the progres‐ sion of neuronal cell damage (data not shown).
