Aim: The aim of this study is to identify potential serum biomarkers of Alzheimer’s disease (AD) for early diagnosis and to evaluate these markers on a large cohort. Methods: We performed two-dimensional difference gel electrophoresis to compare the serum of AD patients and normal controls. Western blot or enzyme-linked immunosorbent assay (ELISA) was used to identify the expression levels of proteins. Results: In this study, a total of 13 differentially expressed proteins were identified. Among them, 2 proteins (inter-alpha-trypsin inhibitor heavy chain H4 [ITI-H4], Apolipoprotein A-IV) were validated by Western blot and 4 proteins (Cofilin 2, Tetranectin, Zinc-alpha-2-glycoprotein [AZGP1], Alpha-1-microglobulin/bikunin precursor [AMBP]) were validated by ELISA, respectively. Western blot results showed that the full size of the ITI-H4 protein was increased, while a fragment of ITI-H4 was decreased in AD patients. In contrast, 1 fragment of Apo A-IV was mainly found in control group and rare to be detected in AD patients. On the other hand, ELISA results showed that Cofilin 2, Tetranectin, AZGP1, and AMBP were significantly increased in AD patients, and Cofilin 2 is strongly correlated with the Mini-Mental State Examination scores of the AD patients. Serum Cofilin 2 was unchanged in Parkinson disease patients as compared to the control group, indicating a specific correlation of serum Cofilin 2 with AD. Moreover, Cofilin 2 was increased in both the serum and brain tissue in the APP/PS1 transgenic mice. Conclusion: Our study identified several potential serum biomarkers of AD, including: ITI-H4, ApoA-IV, Cofilin 2, Tetranectin, AZGP1, and AMBP. Cofilin 2 was upregulated in different AD animal models and might play important roles in AD pathology.

Alzheimer’s disease (AD) is the most common form of age-related dementia that affects innumerable people each year and is difficult for diagnosis and effective treatment [1]. Since it was first described by Alzheimer [2] in 1907, the acknowledged history of AD has transcended more than 1 century. Over the past few decades, the sequential events that happen in the pathphysiology of AD have been profoundly studied, yet the precise etiology of the disease still remains mysterious, and the preventative and curative strategies are largely unsuccessful. Due to the complication of AD pathology, it is noticeable that other AD-related proteins also exist in addition to those we already know. Accordingly, the discovery of novel AD-related proteins or biomarkers will provide new insight into the molecular changes in the pathology of AD and that is critical for unraveling the pathophysiology of the disease [3].

Due to the easily availability and less invasive method, serum is a preferred specimen for the early diagnosis of AD. Every day, around 500 mL of cerebrospinal fluid (CSF) are absorbed into the blood, serum may offer a rich source of brain-related disease biomarkers [4]. Serum Aβ is the most widely used AD marker; however, the specificity and sensitivity are still too low to be used in clinical practice [5]. Other markers such as Clusterin, YKL-40 are being studied, but not specifically for the diagnosis of AD [6‒8]. A major problem for using serum for biomarker discovery is that 95% of total serum protein is comprised with high-abundance proteins, like albumin and immunoglobin, which will cover the signals of proteins with extremely low concentration [9]. To address this problem, immunoaffinity subtraction of these highly abundant proteins by using targeted polyclonal antibody columns and spin filters has recently emerged as a promising tool for serum pre-fractionation [10].

As an improved proteomic technology, 2-dimensional differential in-gel electrophoresis (2D-DIGE) provided a new opportunity to identify biomarkers or therapeutic targets for AD treatment [11]. DIGE incorporates 3 types of fluorescent molecules (CyDyes), which are used to prep-label samples prior to separation by 2-DE [10]. Proteins of interest will be identified by tandem mass spectrometry (LC-MS/MS) [12]. To our knowledge, only one paper has been published by using 2D-DIGE platform to analyze pre-fractionated human plasma from patients with AD [13]. However, the sample capacity of this article is too small, which could reduce the credibility of the results. Furthermore, no novel AD-related proteins were found in this article. In the present study, we design our experiment with 3 aims: (i) identification of differentially expressed proteins in a small cohort (n = 12), (ii) validation of these differentially expressed proteins in a large cohort (n = 18 for AD patients, and n = 54 for age-matched normal controls), and (iii) validation the expression of Cofilin 2 in different AD models.

Human Serum Sample Collection

The study was approved by the Medical Ethics Committee of Affiliated Hospital of Medical College of Qingdao University, and all subjects signed informed consent prior to enrolment in the study. The serum of AD patients and normal controls is provided by the Department of Neurology, Affiliated Hospital of Medical College of Qingdao University. Patient selection basis: (1) age ranges 60–80; (2) with typical symptoms of Alzheimer’s disease; (3) no other neurological diseases were found, such as Parkinson’s disease (PD), dementia with Lewy bodies, frontotemporal dementia, and vascular dementia. Normal controls selection basis (1) age ranges 60–80 and; (2) without neurological or other mental diseases. All AD patients met NINCDS-ADRDA Workgroup criteria for the clinical diagnosis of AD. They were medicine free for at least 2 weeks prior to the study and displayed progressive cognitive decline. Detailed demographic information is presented in Table 1.

Table 1.

Detailed demographic information of AD patients and controls

 Detailed demographic information of AD patients and controls
 Detailed demographic information of AD patients and controls

Serum Pre-Fractionation

The process of depletion was performed according to the protocol described previously [14]. The Agilent Human 14 Multiple Affinity Removal Column (Hu-14, 4.6 × 50 mm) is used to remove 14 abundant proteins from the serum, including albumin, IgG, antitrypsin, IgA, transthyretin, IgM, haptoglobin, transferrin, alpha2-macroglobin, fibrinogen, alpha1-acid glycoprotein, apolipoprotein AI, apolipoprotein AII, and complement C3.

Animals

APP/PS1 mice (APP KM670/671NL, PSEN1deltaE9) and the wild-type littermate were purchased from The Jackson Laboratory. The SAMP8 and SAMR1 mice were supplied from the First Affiliated Hospital of Tianjin College of Traditional Chinese Medicine (Tianjin, China). C57BL/6J was purchased from Vital River Laboratory (China). The animals were housed according to the standard mice housing procedures. The process of intra-cerebroventricular-injecting Aβ25–35 and intraperitoneal injection of scopolamine in the mice have been described in the previous report. All experiments were approved and performed in accordance with the institutional guidelines of the Experimental Animal Center of the Affiliated Hospital of Medical College of Qingdao University.

Cell Culture

SK-N-SH/SK-N-SH APPwt cells were grown in DMEM, supplemented with 10% fetal bovine serum, 100 U/mL penicillin/streptomycin. In addition, SK-N-SH APPwt cells were supplemented with 200 μg/mL G418. Cell cultures were maintained at 37°C in a humidified atmosphere containing 5% CO2 and passed every 2–4 days based on 85% confluence. For AD cell model, SK-N-SH cells were treated with different concentration Aβ25–35 (10 μM, 30 μM, and 50 μM) overnight.

2D-DIGE Analysis and Candidate Protein Biomarkers Identification

2D-DIGE analysis and LC-MS/MS were performed according to the protocol described previously [10]. Typically, to increase the signal consistency of each gel, the photomultiplier tube was set to ensure maximum pixel intensity values for all gel images within a range of 40,000–60,000 pixels. DeCyder 7.0 (GE Healthcare) was used to analyze the images. In DIA module, spots were detected, matched, and normalized; in BVA module, spot statistics were reviewed. The spots with average ratio more than +1.5 or less than −1.5 and with statistical difference (p < 0.05) were isolated for identification. LC-MS analysis was carried out using a Surveyor MS Pump Plus HPLC system coupled to a Thermo Fisher Finnigan LTQ linear ion trap mass spectrometer (Thermo Fisher Corporation, San Jose, CA, USA) using nano-electrospray ionization.

Western Blot Analysis

Western blot was performed as described before. For analysis of mice brain samples, the brains were removed and dissected for cortex and hippocampus on ice, and then homogenized thoroughly in the RIPA lysis buffer (150 mM NaCl, 50 mM Tris [pH 7.4], 1% NP40, 0.5% sodium deoxycholate, and 0.1% SDS). For analysis of serum samples, whole serum was used and diluted 20-fold with ddH2O. All the samples were subjected to electrophoresis, transferred onto PVDF membranes and incubated with anti-ITI-H4 (Santa Cruz, sc-515353), anti-Apo A-IV (Santa Cruz, sc-374543), anti-Cofilin 2 (Santa Cruz, sc-166958), anti-Cofilin 1 (Santa Cruz, sc-53934), and anti-β-actin (CST, 3700) overnight at 4°C. After washing with TBST for 5 times, HRP-coupled secondary antibodies were applied at room temperature for 1 h with gentle agitation. The signal was detected with LAS 4000 FujiFilm imaging system (FujiFilm, Tokyo, Japan) and analyzed with Image J software.

Enzyme-Linked Immunosorbent Assay

Enzyme-linked immunosorbent assay (ELISA) was performed according to the instruction of each kit (Uscnk, Wuhan, China). Each sample was performed in duplicate. Briefly, Serum (or diluted serum) at 100 μL was added to plate and incubated overnight at 4°C. The next day, the supernatant was removed and the plates were washed 3 times with PBS. After that, 100 μL of biotin-conjugated antibodies were added to each well and incubated for 1 h at 37°C. After 3 times of washing, the 100 μL avidin-conjugated HRP was added to each well and incubated for 1 h at 37°C, followed by another 3 times of washing. TMB substrate was added to each well and incubated for 15–30 min at 37°C. After that, the stop solution was added to each well. For the signal detection, each well was determined using 450 nm as a primary wavelength and 630 nm as a reference wavelength.

RNA Interference

For Cofilin 2 knockdown, specific siRNAs for human Cofilin 2 (Santa Cruz, sc-37027) and control (Santa Cruz, sc-37007) were mixed with lipofectamine RNAiMAX (Invitrogen) under the previous transfection protocol according to the manufacturer’s specification. The mixture was incubated in a final volume of 500 μL for 20 min and then added to 80% confluent SK-N-SH cells in 6-well plates for a final volume of 3 mL. Five hours later, the medium was changed to the fresh complete medium for an additional 43 h before cells were harvested.

Statistical Analysis

For the 2D-DIGE experiment, DeCyder 7.0 (GE Healthcare) was used to analysis data from DIGE (DIA and BVA model). For Western blot and ELISA data, one-way ANOVA with SPSS version 13.0 (SPSS Inc., Chicago, IL, USA) was used for the analysis of the data. The differences between the groups were analyzed by Bonferroni’s post hoc test. For Western blot analysis, the control group was normalized to 100%. All data were shown as mean ± SEM, and prism software (GraphPad Prism 5, La Jolla, CA, USA) was used to create the pictures. A value of p < 0.05 was considered to be statistically significant.

Clinical Data

We included 54 control subjects with normal cognition and 18 individuals with AD in the present study. Compared to controls, individuals with AD scored significantly worse on the MMSE (p < 0.001). Moreover, AD group endorsed significantly higher ration of the expression of at least 1 copy of the ApoEε4 allele than control group (p < 0.001). No significant difference was found in age, gender, and education between the 2 groups in the DIGE study as well as the validation groups (p > 0.05). After clotting and centrifugation, the serum from each subject was frozen and stored at −80°C in aliquots until use. The information of each subject was summarized in Table 1.

Identification of Differentially Expressed Spots Using 2D-DIGE

Immuno-depleted serum proteome profiles of 12 AD patients and 12 age-matched controls were analyzed by using DIGE. For each gel, 3 images were generated to the 3 samples (AD patients, age-matched controls, and internal standard). A representative DIGE gel showing the overlay of Cy3, Cy5, and Cy2 images from one such gel is shown in Figure 1a. In DIA workspaces, approximate 600 spots were detected in each gel by Decyder software. In BVA module, Cy2 image from gel number 3 was chosen as master gel as it had the maximum number of spots. Thirteen spots were found to be differentially expressed with the criteria. Spots of proteins differentially expressed in the serum of AD patients and age-matched controls are marked with black circles in pick gel (Fig. 1b). Thirteen protein spots were found differentially expressed in the serum from AD patients as compared with controls (Table 2). Among them, 10 protein spots (spots 125, 143, 156, 279, 358, 364, 384, 388, 390, and 550) were upregulated, 3 spots (235, 453 and 457) were down-regulated. The quantitative changes in these proteins are shown in details in Table 2.

Table 2.

Quantitative changes in differently expressed proteins in serum of AD patients compared with age-matched controls

 Quantitative changes in differently expressed proteins in serum of AD patients compared with age-matched controls
 Quantitative changes in differently expressed proteins in serum of AD patients compared with age-matched controls
Fig. 1.

Identification of differentially expressed spots using 2D-DIGE. Abundant protein depleted serum was subjected to 2D-DIGE quantitative analysis to identify proteins with differing abundance between AD patient and age-matched controls. a Over-lay of cy2-marked loading control, cy3-marked and cy5-marked serum sample from age-matched controls or AD patient (n= 12). b Distribution of differentially expressed protein spots. The samples were separated using IPG gel (pH 3–10, 18 cm) in the first phase and 12.5% SDS-PAGE; 150 μg of protein was used in each gel. The spots showing significant differences between AD patients and controls (see Table 2) were labeled in a 2D-DIGE gel.

Fig. 1.

Identification of differentially expressed spots using 2D-DIGE. Abundant protein depleted serum was subjected to 2D-DIGE quantitative analysis to identify proteins with differing abundance between AD patient and age-matched controls. a Over-lay of cy2-marked loading control, cy3-marked and cy5-marked serum sample from age-matched controls or AD patient (n= 12). b Distribution of differentially expressed protein spots. The samples were separated using IPG gel (pH 3–10, 18 cm) in the first phase and 12.5% SDS-PAGE; 150 μg of protein was used in each gel. The spots showing significant differences between AD patients and controls (see Table 2) were labeled in a 2D-DIGE gel.

Close modal

Identification of Differentially Expressed Proteins in AD Patients

The 13 spots of interest were manually excised from colloidal coomassie stained preparative gels of pooled AD patients and age-matched controls depleted serum for in-gel trypsin proteolysis and subsequently LC-MS/MS (LTQ) analysis, and the results are shown in Table 3. All these 13 differentially expressed protein spots corresponded to 13 different proteins. The 10 spots with increased expression in AD patients were identified as Cofilin 2 (spots 125), Superoxide dismutase (Cu-Zn) (spot 143), Lipocalin-1 (spot 156), Tetranectin (spot 279), Dermcidin (spot 358), Zinc-alpha-2-glycoprotein (AZGP1) (spot 364), Clusterin (spot 384), Alpha-1-microglobulin/bikunin precursor (AMBP) (spot 388), Ficolin-2 (spot 390) and Inter-alpha-trypsin inhibitor heavy chain H4 (ITI-H4) (spots 550). The 3 spots with decreased expression in AD patients were identified as Apolipoprotein A-IV (Apo A-IV) (spots 235), Hemopexin (spots 453) and Alpha-amylase 1 (spots 457). All of these proteins were consistent with theoretical MWs and pI ranges based on the positions of spots on the gel.

Table 3.

Detailed identification information of differentially expressed proteins in serum of AD patients compared with age-matched controls

 Detailed identification information of differentially expressed proteins in serum of AD patients compared with age-matched controls
 Detailed identification information of differentially expressed proteins in serum of AD patients compared with age-matched controls

Protein Validation Using the Whole Serum

In order to in line with the clinical practice, whole serum was used in the validation by means of Western blot analysis. Our previous proteomics study of serum samples from PD patients showed that ITI-H4 and Apo AIV existed 2 forms (full length and fragmented bands) in the serum [14], respectively, ITI-H4 and Apo AIV (n = 16) were carried out in this experiment. Equal volume of whole serum was loaded, and the blots were repeated 3 times. As shown in Figure 2, the level of ITI-H4 fragment (35 kDa) was reduced by 31.3%, while the level of ITI-H4 full size (120 kDa) was increased by 43.3% in AD serum samples compared with control serum samples (Fig. 2a–c). Most importantly, the ratio of ITI-H4 (120 kDa) to ITI-H4 (35 kDa) in AD serum samples was 3.2 fold higher as compared to serum samples from control group (Fig. 2d). As for Apo A-IV, it was found the level of Apo A-IV full size (46 kDa) was relatively unchanged between 2 groups (Fig. 2e); however, the level Apo A-IV fragment (26 kDa) was reduced by 67.3% in serum samples from AD patients as compared to those from control group (Fig. 2f).

Fig. 2.

Expression of ITI-H4 and Apo A-IV in individual serum samples from 16 AD patients and 16 control subjects. a Representative panel of Western blots. It shows that full size (46 kDa) and 26 kDa fragment of Apo A-IV and full size (120 kDa) and 35 kDa fragment of ITI-H4 existed in AD and control serum. b–d Quantitative comparison of the Western blot of ITI-H4. The level of full size ITI-H4 was increased, whereas fragmented ITI-H4 was decreased in AD patients compared with controls. e, f Quantitative comparison of the Western blot of Apo A-IV. The level of full size Apo A-IV was unchanged, whereas fragmented Apo A-IV was decreased in AD patients compared with controls. Data represent mean ± SEM for 16 individual subjects per group. **p< 0.01 compared with control, Student’s ttest.

Fig. 2.

Expression of ITI-H4 and Apo A-IV in individual serum samples from 16 AD patients and 16 control subjects. a Representative panel of Western blots. It shows that full size (46 kDa) and 26 kDa fragment of Apo A-IV and full size (120 kDa) and 35 kDa fragment of ITI-H4 existed in AD and control serum. b–d Quantitative comparison of the Western blot of ITI-H4. The level of full size ITI-H4 was increased, whereas fragmented ITI-H4 was decreased in AD patients compared with controls. e, f Quantitative comparison of the Western blot of Apo A-IV. The level of full size Apo A-IV was unchanged, whereas fragmented Apo A-IV was decreased in AD patients compared with controls. Data represent mean ± SEM for 16 individual subjects per group. **p< 0.01 compared with control, Student’s ttest.

Close modal

Validation Using ELISA Analysis

Serum proteins testing by ELISA have been representing a mainstay of clinically based assays. With the use of the commercially available Cofilin 2, Tetranectin, AZGP1, and AMBP ELISA kit, these candidate biomarker levels were quantified for the original and validation cohort and using whole serum. In line with our 2D-DIGE analysis, significantly higher serum concentrations of Cofilin 2 (p < 0.01), Tetranectin (p < 0.05), AZGP1 (p < 0.01), and AMBP (p < 0.01) were observed in the AD patients and compared to healthy controls (Fig. 3a, c).

Fig. 3.

Serum levels of Cofilin 2, Tetranetin, AZGP1, and AMBP in AD patients and healthy controls. a, b The differentially expressed proteins were detected and analyzed by ELISA in serum samples from AD patients and healthy controls. c, d ROC curve of Cofilin 2, Tetranetin, AZGP1, and AMBP. Data represent means ± SEM (n= 18 for AD group and 54 for controls). *p< 0.05; **p< 0.01 versus the control group.

Fig. 3.

Serum levels of Cofilin 2, Tetranetin, AZGP1, and AMBP in AD patients and healthy controls. a, b The differentially expressed proteins were detected and analyzed by ELISA in serum samples from AD patients and healthy controls. c, d ROC curve of Cofilin 2, Tetranetin, AZGP1, and AMBP. Data represent means ± SEM (n= 18 for AD group and 54 for controls). *p< 0.05; **p< 0.01 versus the control group.

Close modal

To determine the potential impact of the use of these 4 proteins as markers for discriminating between AD patients and age-matched controls, the ELISA results for these 4 proteins for the patients (n = 18) and age-matched controls (n = 54) were used to generate receiver operating characteristic (ROC) curves (Fig. 3b). The area under the curve (AUC) was measured for each marker to determine which protein(s) yielded the highest discriminatory power within our patient set. The data showed that Cofilin 2 has the highest sensitivity and specificity among these 4 proteins (Fig. 3d).

Validation of Cofilin in AD and PD Patients’ Serum

Since Cofilin 2 level showed the highest of AUC, we were wondering whether this change only happened in AD patient, or it is a common change in neurodegenerative disease like PD. Equal volume of whole serum from AD patients, PD patients and age-matched controls was loaded on the gel, and Western blot analyses confirmed a robust increase of Cofilin 2 level in the serum of AD patients (Fig. 4a, b), whereas relatively no changes of Cofilin 2 level detected in PD patients as compared to age-matched controls (Fig. 4d, e).

Fig. 4.

Expression of Cofilin 2 and Cofilin 1 in individual serum samples from 16 AD patients, 16 PD patients, and 16 control subjects. a, b Representative panel and Quantitative comparison of the Western blot of Cofilin 2 in AD patients. c The level of Cofilin 2 was increased in AD patients compared with controls and highly correlated with MMSE of AD patients. d, e Representative panel and Quantitative comparison of the Western blot of Cofilin 2 in PD patients. The level of Cofilin 2 was unchanged in PD patients compared with controls. f Representative panel of the Western blot of Cofilin 1 in AD patients. No Cofilin 1 was detected in the human serum. Data represent mean ± SEM for 16 individual subjects per group. **p< 0.01 compared with control, Student’s ttest.

Fig. 4.

Expression of Cofilin 2 and Cofilin 1 in individual serum samples from 16 AD patients, 16 PD patients, and 16 control subjects. a, b Representative panel and Quantitative comparison of the Western blot of Cofilin 2 in AD patients. c The level of Cofilin 2 was increased in AD patients compared with controls and highly correlated with MMSE of AD patients. d, e Representative panel and Quantitative comparison of the Western blot of Cofilin 2 in PD patients. The level of Cofilin 2 was unchanged in PD patients compared with controls. f Representative panel of the Western blot of Cofilin 1 in AD patients. No Cofilin 1 was detected in the human serum. Data represent mean ± SEM for 16 individual subjects per group. **p< 0.01 compared with control, Student’s ttest.

Close modal

The Mini-Mental State Examination (MMSE) score is an important indicator of the cognitive level of patients. Patients with scores of 30–29 are considered cognitively normal and scores decline during progression from mild cognitive impairment (28–26) to early, mid, and late AD. The correlation between the serum levels of Cofilin 2 and the MMSE scores of the patients is shown in Figure 4c. The results show good correlation between MMSE scores and amounts of Cofilin 2 (r = −0.872, p < 0.001). Up to now, only one paper has been published relating Cofilin to the pathology of AD, in which the author describes actin-depolymerizing factor/Cofilin-actin rods in the brain of AD patients without discriminating Cofilin 1 or Cofilin 2, the 2 isoforms of Cofilin [15]. Our data showed no expression of Cofilin 1 in the human serum (Fig. 4f), which confirmed the result that Cofilin 2 not Cofilin 1 has a different expression pattern between AD patients and age-matched controls.

Expression of Cofilin 2 in Different AD Animal Models

Our previous proteomics study has demonstrated that Cofilin 2 level was significantly upregulated in the serum of AD patients. To determine whether this change in AD patient serum paralleled with those in AD animal model, we compared Cofilin 2 levels in sera and brain tissue in different AD animal models by Western blot analysis. In accord with our proteomics result, a robust increase of Cofilin 2 level was observed in the cortex, hippocampus and serum of 12-month-old SAMP8 mice as compared to 12-month-old SAMR1 mice (Fig. 5a, b). Moreover, we found Cofilin 2 level was significantly enhanced in the cortex and serum of Aβ25–35-injected rats (Fig. 5c, d). At the meantime, a notable increase of Cofilin 2 level was detected in the hippocampus and serum of APP/PS1 mice as compared to the wild-type mice (Fig. 5e, f).

Fig. 5.

Confirmation of differential expression of Cofilin 2 in different AD animal models. a Representative panel of the Western blot of Cofilin 2 in the cortex, hippocampus, and serum of 12-month-old SAMP8 and SAMR1 mice. b Quantitative comparison of the Western blot of Cofilin 2 in the cortex, hippocampus, and serum of 12-month-old SAMP8 and SAMR1 mice. c Representative panel of the Western blot of Cofilin 2 in the cortex, hippocampus, and serum of Aβ25–35-injected mice. d Quantitative comparison of the Western blot of Cofilin 2 in the cortex, hippocampus, and serum of Aβ25–35-injected mice. e Representative panel of the Western blot of Cofilin 2 in the cortex, hippocampus, and serum of 12-month old APP/PS1 and wild-type mice. f Quantitative comparison of the Western blot of Cofilin 2 in the cortex, hippocampus, and serum of 12-month-old APP/PS1 and wild-type mice.

Fig. 5.

Confirmation of differential expression of Cofilin 2 in different AD animal models. a Representative panel of the Western blot of Cofilin 2 in the cortex, hippocampus, and serum of 12-month-old SAMP8 and SAMR1 mice. b Quantitative comparison of the Western blot of Cofilin 2 in the cortex, hippocampus, and serum of 12-month-old SAMP8 and SAMR1 mice. c Representative panel of the Western blot of Cofilin 2 in the cortex, hippocampus, and serum of Aβ25–35-injected mice. d Quantitative comparison of the Western blot of Cofilin 2 in the cortex, hippocampus, and serum of Aβ25–35-injected mice. e Representative panel of the Western blot of Cofilin 2 in the cortex, hippocampus, and serum of 12-month old APP/PS1 and wild-type mice. f Quantitative comparison of the Western blot of Cofilin 2 in the cortex, hippocampus, and serum of 12-month-old APP/PS1 and wild-type mice.

Close modal

Expression of Cofilin 2 in Different AD Animal Models

It has been proposed that Cofilin 2 may result in the hyper-phosphorylation of Tau in vulnerable brain regions and initiate the events in pathophysiology of AD [16]. Therefore, we employed RNA interference to assess the effect of Cofilin 2 knockdown on the phosphorylated state of Tau using neuroblastoma SK-N-SH cells. As showed in our in vivo experiment, Cofilin 2 level was found to be significantly increased in 30 μM Aβ25–35 treated neuroblastoma SK-N-SH cells (Fig. 6a, b), whereas Cofilin 2 expression was reduced by 84.6% in SK-N-SH cells when it was knocked down with siCofilin 2 compared to the cells transfected with siScramble (Fig. 6c). Knockdown of Cofilin 2 engendered a dramatic reduction in tau phosphorylation in SK-N-SH cells. Quantitative analysis showed that the level of tau phosphorylation at Ser199, Thr205, Thr231, Ser396, and Ser404 reduced by 62.8%, 34.2%, 15.6%, 74.5%, and 67.6%, respectively, compared to siScramble group (Fig. 6d).

Fig. 6.

Knockdown of Cofilin 2 reduces the hyperphosphorylation of tau in SK-N-SH cells. a Representative panel of the Western blot of Cofilin 2 in 10 μM, 30 μM, 100 μM Aβ25–35 treated SK-N-SH cells. b Quantitative comparison of the Western blot of Cofilin 2 in 10 μM, 30 μM, 100 μM Aβ25–35 treated SK-N-SH cells. c Representative panel of the Western blot of Cofilin 2, p-Tau (S199), p-Tau (T205), p-Tau (T231), p-Tau (S396), and p-Tau (S404) after SK-N-SH cells were transfected with siScramble or siCofilin 2. d Quantitative comparison of the Western blot of Cofilin 2, p-Tau (S199), p-Tau (T205), p-Tau (T231), p-Tau (S396), and p-Tau (S404) after SK-N-SH cells were transfected with siScramble or siCofilin 2. Quantified results were normalized to β-actin. For all the results above, a representative experiment of 3 performed is shown. Values were expressed as percentages compared to the control group (set to 100%) and represented as group mean ± SEM. n= 6 per group. *p< 0.05, **p< 0.01 versus control group.

Fig. 6.

Knockdown of Cofilin 2 reduces the hyperphosphorylation of tau in SK-N-SH cells. a Representative panel of the Western blot of Cofilin 2 in 10 μM, 30 μM, 100 μM Aβ25–35 treated SK-N-SH cells. b Quantitative comparison of the Western blot of Cofilin 2 in 10 μM, 30 μM, 100 μM Aβ25–35 treated SK-N-SH cells. c Representative panel of the Western blot of Cofilin 2, p-Tau (S199), p-Tau (T205), p-Tau (T231), p-Tau (S396), and p-Tau (S404) after SK-N-SH cells were transfected with siScramble or siCofilin 2. d Quantitative comparison of the Western blot of Cofilin 2, p-Tau (S199), p-Tau (T205), p-Tau (T231), p-Tau (S396), and p-Tau (S404) after SK-N-SH cells were transfected with siScramble or siCofilin 2. Quantified results were normalized to β-actin. For all the results above, a representative experiment of 3 performed is shown. Values were expressed as percentages compared to the control group (set to 100%) and represented as group mean ± SEM. n= 6 per group. *p< 0.05, **p< 0.01 versus control group.

Close modal

Nowadays, as there is still no effective treatment or curable drugs for AD, early and accurate diagnosis has become a prime requirement for the management of AD [17]. Although noninvasive imaging techniques such as MRI have already been applied for AD diagnosis clinically, it is difficult to measure the minor difference in hippocampus size and behavior with these imaging approaches. Nevertheless, the development of (MS)-based proteomic technologies has offered us a practicality approach for the study of AD biomarker research over the last decades. A panel of novel candidate biomarkers has been identified in urine, saliva, CSF, and serum AD patients by using a classic proteomics platform, 2D-DIGE in combination with MS or tandem MS [18‒21]. Of note, rare studies on salivary proteomics of AD were reported because of low protein content in the saliva [22]. Moreover, due to the long distance between brain and the organs that produce urine, it is difficult for proteins from the central nervous system to be secreted into urine [23]. CSF diagnosis (by invasive means) may not always be feasible, and better noninvasive diagnostic techniques are needed. This is especially important since many AD patients are elder people, whose body is relatively poor and cannot bear the impact of spinal puncture. In order to circumvent these limitations, novel detection techniques based on molecular biomarkers using serum are emerging at the forefront for AD diagnosis [17]. 2D-DIGE pilot study on affinity depleted serum allowed for comparative analysis of protein expression between AD patients and age-matched controls [24].

In this study, 2D-DIGE proteomic study was carried out and a number of different expressed spots were identified, 13 of which were detected as significant (Table 2). The 13 proteins were then identified from these 13 spots (Table 3). Some of those identified proteins, like Superoxide dismutase (Cu-Zn), Lipocalin-1, Clusterin, and Apolipoprotein A-IV have been reported previously in AD and other neurodegenerative diseases, which supports our findings and enhances confidence of this study, while some of the identified proteins including Dermcidin and Ficolin-2, were not reported directly in context of AD [25‒28].

In our previous proteomics study, it has been shown that ITI-H4 and Apo A-IV both existed in 2 forms in the serum [14]. We suggested that the ratio between full size ITI-H4 (120 kDa) and fragmented ITI-H4 (35 kDa) and fragmented Apo A-IV (∼26 kDa) in the serum may serve as a diagnostic indicator for PD. In this study, we report the same phenomena in the serum of AD patients which showed elevated serum level of ITI-H4 (120 kDa/35 kDa) and reduced serum level of Apo A-IV (∼26 kDa) in the serum level of AD patients. Our result along with the previous study suggested the ITI-H4 and Apo A-IV might be a common pathological feature for neurodegenerative disease. Though the roles of ITI-H4 and Apo A-IV in develop of AD and PD are not clear, it indicates that AD and PD might have a similar pathogenesis and this pathogenesis might initiate the neuronal degenerative procedures.

In addition, some other proteins showed potential functional relevance with AD, Cofilin 2, Tetranectin, AZGP1, and AMBP were brought forward through a series of validation steps. In line with the proteomics result, the ELISA data showed a robust increase of Cofilin 2, Tetranectin, AZGP1, and AMBP in the serum of AD patients as compared to controls (50.7%, 22.8%, 33.1%, and 45.5%, respectively). ROC curves were generated from the ELISA data for these 4 proteins in a large cohort (Fig. 3). The curves were generated using logistic regression, with the ability to identify which potential biomarker can identify disease easily. When the area under the curve is larger than 0.7 for the protein, people believe the protein to have the high ability in discrimination of patients from healthy people [29]. The AUC for these 4 proteins is listed in Figure 3d and only Cofilin 2 can fulfill the above requirement. The other 3 proteins were not high enough as markers for AD diagnosis. Western blot analysis of Cofilin 2 showed that the serum level of Cofilin 2 was increased by 68.2% in AD patients, which confirms our ELISA data. The correlation between the serum level of Cofilin 2 and the MMSE scores of the AD patients is relatively high which suggests that the presence of these proteins in serum may reflect disease progression.

With our previous experience, we are wondering to know the expression pattern of Cofilin 2 in other neurodegenerative disease such as PD. It has been demonstrated that the serum level of Cofilin 2 is relatively unchanged as compared to the control (Fig. 4d, e). This is really important because neurodegenerative disease may share a common biomarker in most cases due to the similar pathogenesis signal pathway. Although this result needs further validation in other neurodegenerative disease, it is no doubt that Cofilin 2 is a more accurate potential biomarker of AD.

Cofilin is an actin-binding protein that belongs to the actin depolymerization factor family and is ubiquitous in eukaryotic cells. It binds to actin filaments (F-actin) and regulates the depolymerization and remodeling of F-actin. Cofilin has a variety of biological functions, such as involvement in actin reorganization, nuclear translocation, cytokinesis, and angiogenesis [30‒32]. There are 2 isoforms of Cofilin found in humans and mice, Cofilin 1 and Cofilin 2, respectively. Recent studies have demonstrated that actin-depolymerizing factor/Cofilin-actin rods expressed in the brains of AD patients [15]. However, the author did not clarify the isoform of Cofilin involved. Our result shows no expression of Cofilin 1 in the serum of human. It indicates that Cofilin-2 but not Cofilin 1 plays some roles in the pathology of AD.

Most of proteomics studies tended to focus on patients or just animals. However, from the translational study of view, we are wondering whether change of Cofilin 2 may also happen in AD animal models. Our present study showed that Cofilin 2 level was notably increased in the serum and hippocampus of APP/PS mice as compared to wild-type mice. At the meantime, we found Cofilin 2 level was also increased in the cortex of Aβ25–35 intra-cerebroventricular-injected AD rat model. Therefore, it is suggested that, due to the expression patterns of Cofilin 2 in human and in animals, Cofilin 2 might also serve as a biomarker for treatment of AD in drug evaluation using APP/PS1 mice.

In conclusion, we have confirmed the differential expression of ITI-H4 and ApoA-IV not only in the PD samples but also in AD samples. They might be also the common biomarkers for neurodegenerative disease. In addition, we showed a 9.4-fold increase in Cofilin 2 in serum of AD patients. This result was confirmed at the serum level by ELISA and Western blot analysis. Above proteins might be used as biomarkers combined with other clinical tests for AD diagnosis and they are also the potential indicators for treatment of the disease in clinic or for AD drug development in preclinical study.

This research was conducted ethically in accordance with the World Medical Association Declaration of Helsinki. The subjects have given their written informed consent. The human and animal studies were approved by the Affiliated Hospital of Qingdao University’s committee on human research and the reference number is QYFY-WZLL-25923.

The authors declare no conflict of interest.

The data collection was supported by the Affiliated Hospital of Qingdao University X Research Fund (QDFY + X2021050).

Xuezhi Zhang collected patients’ serum. Wenwen Yu performed proteomic experiments. Xuezhi Zhang, Xuelei Cao, and Yongbin Wang performed cell experiments, Western blot, and ELISA assays. Xuezhi Zhang and Wenwen Yu analyzed the data and drafted the manuscript. Chao Zhu and Jialiang Guan designed and supervised the project and finalized the manuscript.

All data generated or analyzed during this study are included in this article. Further enquiries can be directed to the corresponding author.

1.
Hardy
JA
,
Higgins
GA
.
Alzheimer’s disease: the amyloid cascade hypothesis
.
Science
.
1992 Apr 10
;
256
(
5054
):
184
5
.
2.
Alzheimer
A
,
Stelzmann
RA
,
Schnitzlein
HN
,
Murtagh
FR
.
An English translation of Alzheimer’s 1907 paper, “uber eine eigenartige erkankung der hirnrinde”
.
Clin Anat
.
1995
;
8
(
6
):
429
31
.
3.
Rong
XF
,
Sun
YN
,
Liu
DM
,
Yin
HJ
,
Peng
Y
,
Xu
SF
,
.
The pathological roles of NDRG2 in Alzheimer’s disease, a study using animal models and APPwt-overexpressed cells
.
CNS Neurosci Ther
.
2017 Aug
;
23
(
8
):
667
79
.
4.
Chung
YC
,
Ko
HW
,
Bok
E
,
Park
ES
,
Huh
SH
,
Nam
JH
,
.
The role of neuroinflammation on the pathogenesis of Parkinson’s disease
.
BMB Rep
.
2010 Apr
;
43
(
4
):
225
32
.
5.
Abdullah
L
,
Paris
D
,
Luis
C
,
Quadros
A
,
Parrish
J
,
Valdes
L
,
.
The influence of diagnosis, intra- and inter-person variability on serum and plasma Abeta levels
.
Neurosci Lett
.
2007 Nov 27
;
428
(
2–3
):
53
8
.
6.
Ijsselstijn
L
,
Dekker
LJ
,
Koudstaal
PJ
,
Hofman
A
,
Sillevis Smitt
PA
,
Breteler
MM
,
.
Serum clusterin levels are not increased in presymptomatic Alzheimer’s disease
.
J Proteome Res
.
2011 Apr 1
;
10
(
4
):
2006
10
.
7.
Prikrylova Vranova
H
,
Henykova
E
,
Mares
J
,
Kaiserova
M
,
Mensikova
K
,
Vastik
M
,
.
Clusterin CSF levels in differential diagnosis of neurodegenerative disorders
.
J Neurol Sci
.
2016 Feb 15
;
361
:
117
21
.
8.
Baldacci
F
,
Lista
S
,
Cavedo
E
,
Bonuccelli
U
,
Hampel
H
.
Diagnostic function of the neuroinflammatory biomarker YKL-40 in Alzheimer’s disease and other neurodegenerative diseases
.
Expert Rev Proteomics
.
2017 Apr
;
14
(
4
):
285
99
.
9.
Tanaka
Y
,
Akiyama
H
,
Kuroda
T
,
Jung
G
,
Tanahashi
K
,
Sugaya
H
,
.
A novel approach and protocol for discovering extremely low-abundance proteins in serum
.
Proteomics
.
2006 Sep
;
6
(
17
):
4845
55
.
10.
Byrne
JC
,
Downes
MR
,
O'Donoghue
N
,
O'Keane
C
,
O'Neill
A
,
Fan
Y
,
.
2D-DIGE as a strategy to identify serum markers for the progression of prostate cancer
.
J Proteome Res
.
2009 Feb
;
8
(
2
):
942
57
.
11.
Dowling
P
,
Ohlendieck
K
.
DIGE analysis of immunodepleted plasma
.
Methods Mol Biol
.
2018
;
1664
:
245
57
.
12.
Hagner-McWhirter
A
,
Winkvist
M
,
Bourin
S
,
Marouga
R
.
Selective labelling of cell-surface proteins using CyDye DIGE Fluor minimal dyes
.
J Vis Exp
.
2008 Nov
;
26
(
21
);
945
.
13.
Kitamura
Y
,
Usami
R
,
Ichihara
S
,
Kida
H
,
Satoh
M
,
Tomimoto
H
,
.
Plasma protein profiling for potential biomarkers in the early diagnosis of Alzheimer’s disease
.
Neurol Res
.
2017 Mar
;
39
(
3
):
231
8
.
14.
Lu
W
,
Wan
X
,
Liu
B
,
Rong
X
,
Zhu
L
,
Li
P
,
.
Specific changes of serum proteins in Parkinson’s disease patients
.
PLoS One
.
2014
;
9
(
4
):
e95684
.
15.
Bamburg
JR
,
Bernstein
BW
,
Davis
RC
,
Flynn
KC
,
Goldsbury
C
,
Jensen
JR
,
.
ADF/Cofilin-actin rods in neurodegenerative diseases
.
Curr Alzheimer Res
.
2010 May
;
7
(
3
):
241
50
.
16.
Woo
JA
,
Liu
T
,
Fang
CC
,
Cazzaro
S
,
Kee
T
,
LePochat
P
,
.
Activated cofilin exacerbates tau pathology by impairing tau-mediated microtubule dynamics
.
Commun Biol
.
2019
;
2
:
112
.
17.
Gollapalli
K
,
Ray
S
,
Srivastava
R
,
Renu
D
,
Singh
P
,
Dhali
S
,
.
Investigation of serum proteome alterations in human glioblastoma multiforme
.
Proteomics
.
2012 Aug
;
12
(
14
):
2378
90
.
18.
Davidsson
P
,
Sjögren
M
.
Proteome studies of CSF in AD patients
.
Mech Ageing Dev
.
2006 Feb
;
127
(
2
):
133
7
.
19.
Ryu
OH
,
Atkinson
JC
,
Hoehn
GT
,
Illei
GG
,
Hart
TC
.
Identification of parotid salivary biomarkers in Sjogren’s syndrome by surface-enhanced laser desorption/ionization time-of-flight mass spectrometry and two-dimensional difference gel electrophoresis
.
Rheumatology
.
2006 Sep
;
45
(
9
):
1077
86
.
20.
Li
F
,
Chen
DN
,
He
CW
,
Zhou
Y
,
Olkkonen
VM
,
He
N
,
.
Identification of urinary Gc-globulin as a novel biomarker for bladder cancer by two-dimensional fluorescent differential gel electrophoresis (2D-DIGE)
.
J Proteomics
.
2012 Dec 21
;
77
:
225
36
.
21.
Conti
E
,
Andreoni
S
,
Tomaselli
D
,
Storti
B
,
Brovelli
F
,
Acampora
R
,
.
Serum DBI and biomarkers of neuroinflammation in Alzheimer’s disease and delirium
.
Neurol Sci
.
2021 Mar
;
42
(
3
):
1003
7
.
22.
Liang
D
,
Lu
H
.
Salivary biological biomarkers for Alzheimer’s disease
.
Arch Oral Biol
.
2019 Sep
;
105
:
5
12
.
23.
Zhang
F
,
Wei
J
,
Li
X
,
Ma
C
,
Gao
Y
.
Early candidate urine biomarkers for detecting Alzheimer’s disease before amyloid-beta plaque deposition in an APP (swe)/PSEN1dE9 transgenic mouse model
.
J Alzheimers Dis
.
2018
;
66
(
2
):
613
37
.
24.
Graf
A
,
Oehler
R
.
Study design in DIGE-based biomarker discovery
.
Methods Mol Biol
.
2012
;
854
:
195
206
.
25.
Csaszar
A
,
Kalman
J
,
Szalai
C
,
Janka
Z
,
Romics
L
.
Association of the apolipoprotein A-IV codon 360 mutation in patients with Alzheimer’s disease
.
Neurosci Lett
.
1997 Jul 25
;
230
(
3
):
151
4
.
26.
Lambert
JC
,
Heath
S
,
Even
G
,
Campion
D
,
Sleegers
K
,
Hiltunen
M
,
.
Genome-wide association study identifies variants at CLU and CR1 associated with Alzheimer’s disease
.
Nat Genet
.
2009 Oct
;
41
(
10
):
1094
9
.
27.
Kallo
G
,
Emri
M
,
Varga
Z
,
Ujhelyi
B
,
Tozser
J
,
Csutak
A
,
.
Changes in the chemical barrier composition of tears in Alzheimer’s disease reveal potential tear diagnostic biomarkers
.
PLoS One
.
2016
;
11
(
6
):
e0158000
.
28.
Muresan
V
,
Ladescu Muresan
Z
.
Shared molecular mechanisms in Alzheimer’s disease and amyotrophic lateral sclerosis: neurofilament-dependent transport of sAPP, FUS, TDP-43 and SOD1, with endoplasmic reticulum-like tubules
.
Neurodegener Dis
.
2016
;
16
(
1–2
):
55
61
.
29.
Zhao
X
,
Lejnine
S
,
Spond
J
,
Zhang
C
,
Ramaraj
TC
,
Holder
DJ
,
.
A candidate plasma protein classifier to identify Alzheimer’s disease
.
J Alzheimers Dis
.
2015
;
43
(
2
):
549
63
.
30.
Elam
WA
,
Kang
H
,
De la Cruz
EM
.
Biophysics of actin filament severing by cofilin
.
FEBS Lett
.
2013 Apr 17
;
587
(
8
):
1215
9
.
31.
Mahaffey
JP
,
Grego-Bessa
J
,
Liem
KF
Jr
,
Anderson
KV
.
Cofilin and Vangl2 cooperate in the initiation of planar cell polarity in the mouse embryo
.
Development
.
2013 Mar
;
140
(
6
):
1262
71
.
32.
Nomura
K
,
Ono
S
.
ATP-dependent regulation of actin monomer-filament equilibrium by cyclase-associated protein and ADF/cofilin
.
Biochem J
.
2013 Jul 15
;
453
(
2
):
249
59
.

Additional information

Xuezhi Zhang and Wenwen Yu contributed equally to this paper.