Proteomic analysis and identification of aqueous humor proteins with a pathophysiological role in diabetic retinopathy

J O U RN A L OF P ROT EO M IC S 7 5 ( 2 0 12 ) 29 5 0 –29 5 9

Available online at www.sciencedirect.com

www.elsevier.com/locate/jprot

Proteomic analysis and identification of aqueous humor proteins with a pathophysiological role in diabetic retinopathy☆ Shang-Yi Chianga, b , Ming-Ling Tsaib , Chih-Yuan Wangc , Ann Chend , Yu-Ching Choue , Ching-Wu Hsiaa , Yung-Fu Wuf , Han-Min Cheng, h , Tzu-Hao Huangi , Pei-Hsiu Chenj , Hung-Te Liua , Hao-Ai Shuia,⁎ a

Graduate Institute of Medical Sciences, National Defense Medical Center, Taipei, Taiwan Department of Ophthalmology, Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan c Department of Internal Medicine, National Taiwan University Hospital, College of Medicine, National Taiwan University, Taipei, Taiwan d Department of Pathology, Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan e School of Public Health, National Defense Medical Center, Taipei, Taiwan f Department of Medical Research, Tri-Service General Hospital, Taipei, Taiwan g Department of Life Science, Fu-Jen Catholic University, New Taipei City , Taiwan h Biomedical and photonic interdisciplinary research center, Fu-Jen Catholic University, New Taipei City, Taiwan i Department of Nursing, Chang Gung University of Science and Technology, Taoyuan, Taiwan j Graduate Institute of Life Sciences, National Defense Medical Center, Taipei, Taiwan b

AR TIC LE I N FO Available online 20 December 2011

ABS TR ACT Diabetic retinopathy (DR) can cause irreversible blindness and is the severest microvascular complication in the eyes of patients with diabetic mellitus (DM). The identification of

Keywords:

susceptibility factors contributing to development of DR is helpful for identifying

Proteomics

predisposed patients and improving treatment efficacy. Although proteomics analysis is

Retinopathy

useful for identifying protein markers related to diseases, it has never been used to

Aqueous humor

explore DR-associated susceptibility factors in the aqueous humor (AH). To better under-

Susceptibility factors

stand the pathophysiology of DR and to identify DR-associated risk factors, a gel-based pro-

Biomarkers

teomics analysis was performed to compare AH protein profiles of DM patients with and

Pathophysiology

without development of DR. MALDI-TOF MS was then performed to identify protein spots that were differentially expressed between the two groups and western blot analysis was used to validate the expressional change of protein demonstrated by proteomics. Our proteomics and bioinformatics analysis identified 11 proteins differentially expressed between DR and control groups. These proteins are linked to biological networks associated with nutrition transport, microstructure reorganization, angiogenesis, anti-oxidation, and neuroprotection. The data may provide potential AH biomarkers and susceptibility factors for predicting DR development, and provide an insight into the underlying pathophysiological mechanisms of DR. This article is part of a Special Issue entitled: Proteomics: The clinical link. © 2011 Elsevier B.V. All rights reserved.

☆

This article is part of a Special Issue entitled: Proteomics: The clinical link. ⁎ Corresponding author at: Graduate Institute of Medical Sciences, National Defense Medical Center, 161 Min-Chuan East Road, 6th Section, Taipei, Taiwan (114), ROC. Tel.: +886 2 8792 3100x18905. E-mail address: [email protected] (H.-A. Shui). 1874-3919/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jprot.2011.12.006

J O U RN A L OF P ROT EO M I CS 7 5 ( 2 0 12 ) 29 5 0 –2 95 9

1.

Introduction

Diabetic mellitus (DM) is a worldwide metabolic disorder that affects millions of people in developed countries. Most DM patients develop serious macrovascular complications, such as coronary artery disease, stroke, and peripheral vascular disease, as well as serious microvascular complications, such as nephropathy and retinopathy [1]. In the eye, DM can cause microvascular problems in different tissues (iris, lens, and retina), leading to rubeosis, cataract, glaucoma, vitreous hemorrhage, temporary blurring of vision, and diabetic retinopathy (DR) [2]. DR is the severest microvascular complication and is the main cause of irreversible blindness in people younger than 65 years in the working population in developed countries [2]. Different DM patients show different susceptibilities to development of DR. The identification of susceptibility factors contributing to development of DR might identify predisposed patients and improve treatment efficacy. Ocular fluids, i.e. the tear, vitreous humor (VH) and aqueous humor (AH), exchange substances continuously with the blood through the capillary wall and also with various eye tissues through direct and indirect contact [3]. As a result, levels of proteins in the ocular fluids can be used to indicate the health states of blood vessels and/or eye tissues [3]. In terms of DMinduced microvascular complications, the shift from a healthy state to the complications can influence protein constituents and/or protein abundance in the ocular fluids. For example, levels of various proteins are altered in the tear film in the DM condition [3–6], and are changed in both VH and AH in the DR condition [7–14]. These proteins can serve as biomarkers for predicting the development of DM or DR [3]. In terms of DR, the discovery of more ocular fluid biomarkers could improve the prediction, diagnosis and prognosis of this eye illness. Proteomics analysis is useful for identifying AH proteins associated with various eye diseases [3,15–20], but it has not previously been used to explore AH protein biomarkers in DR patients. Some pathophysiological mechanisms underlying the development of DR have been uncovered [21–24]. Hyperglycemia is the factor that initiates the development of DR, leading to serial pathological changes (i.e. microvascular injury, chronic inflammation, edema, oxidative stress, hypoxia, neurodegeneration and angiogenesis) that eventually cause retinal abnormality [21–24]. It has been demonstrated that, during the development of DR, some proteins that participate in the DR-associated biochemical events leak from the blood and/or eye tissues into the ocular fluids. For example, increased levels of DRassociated inflammatory cytokines and angiogenic factors [23,24], that participate in chronic inflammation, stimulate angiogenesis, and increase vascular permeability, retina edema, and ischemia, can be detected in both VH and AH in DR patients [11,25,26]. Thus, identification of intraocular proteins associated with DR could provide an insight into the pathogenesis of the illness, and discover susceptibility factors contributing to the development of DR. Proteomics analysis is a useful approach to identify proteins in body fluids associated with various physiological or pathological conditions. Proteomics approaches have been applied to the discovery of DR-associated pathogenic factors in the eye, but these studies focused exclusively on the

2951

proteome of the retina in DR animal models and that of the VH in DR patients and controls [27–33]. To the best of our knowledge, no study has been performed to explore the AH proteome in DR patients, despite the fact that proteomics analysis of AH proteins is feasible and has been used to study non-DR eye problems, such as myopia, uveal melanoma, corneal rejection, and primary open angle glaucoma [3,15–20]. Although the AH does not make direct contact with the retina, proteins released from retinal tissues can enter AH through the disrupted blood-retinal barrier and cilio-retinal circulation in DR [34,35], or by diffusion through the VH and the VH–AH barrier [36–38]. As a result, the AH contains proteins that could be used to indicate the development of DR in DM patients [7–14]. Thus, the aim of the present study were to perform a proteomic comparison of AH proteins in DM patients with and without the development of DR, in order to identify potential AH biomarkers and susceptibility factors for predicting DR development.

2.

Materials and methods

2.1.

Subjects

Twenty-two DM subjects, 11 with DR (DR group) and 11 without DR (control group), were recruited for this study. All 22 were scheduled for phacoemulsification cataract surgery for insertion of an intraocular lens and met the inclusion criteria that they had no history or slit-lamp evidence of ocular trauma, no prior intraocular surgery, no ocular disease other than cataract, and no use of systemic antimetabolites, corticosteroids, or immunosuppressants. The mean age of the DR patients was 70.65± 10.00 years (range 55–89 years) and that of the control subjects 67.22 ± 8.55 years (range 55–88 years). The study protocol was reviewed and approved by the Institutional Review Board of the Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan.

2.2.

Aqueous humor collection

Collection of AH was performed by the same operator. Briefly, after rinsing twice with 5% povidine iodine, 1 or 2 drops of proparacaine hydrochloride 0.5% (Alcaine, Alcon, Ft. Worth, TX, USA) were applied twice to the eye about to undergo surgery. A clear corneal incision was then made with a 1.3 mm MVR blade and the wound irrigated with normal saline to remove blood clots, debris, and epithelial cells. The AH was then collected under the surgical microscope using a 1 ml tuberculin syringe and a 30 gauge blunt needle, which was used to avoid damage to the iris and the anterior lens capsule to prevent protein contamination. The AH samples were rapidly cooled in ice, centrifuged for 15 min at 4 °C to remove cells, and immediately stored in liquid nitrogen.

2.3.

Quantification of total protein in the AH

Protein concentration in the AH samples before desalting was measured to estimate the original protein concentration of the AH in the eyes. Protein concentration was measured using a protein assay kit based on the Bradford's method according to

2952

J O U RN A L OF P ROT EO M IC S 7 5 ( 2 0 12 ) 29 5 0 –29 5 9

the manufacturer's instruction (Bio-Rad, Hercules, CA, USA) using serial dilutions of bovine serum albumin as concentration standards.

2.4.

Desalting and concentration of AH samples

The AH samples (120 μl/sample) collected were desalted and concentrated by ultrafiltration, a method confirmed by our calculation to allow the recovery of a greater amount of AH proteins and confirmed to have good reproducibility by testing on a standard protein mixture. Briefly, the AH was centrifuged at 10,000 g at 4 °C on an AmiconTM spin column (Millipore, MA, USA), then the proteins retained in the upper chamber were washed twice by addition of low-salt Tris buffer (25 mM, pH 7.5) and centrifugation, and the retained AH protein solution collected. After desalting, concentration of the retained AH protein was measured again to estimate the recovery rate of protein to ensure that there is no unacceptable loss of proteins caused by the ultrafiltration procedure. The desalted AH proteins were then mixed with detergentcontaining buffer (9 M urea, 2% CHAPS, 1% DTT, and 0.5% IPG buffer pH 3–10) by gentle shaking for 15 min at room temperature.

2.5.

Two-dimensional gel electrophoresis (2-DE)

Samples from the patients were analyzed by two-dimensional electrophoresis (2-DE) individually. Although loading a fixed amount of proteins onto gels can minimize the variation of protein level caused by varied amount of cells or tissues, it is inappropriate to apply this idea to body fluids which show big difference in both total protein concentration and composition under different conditions [15,18]. To ensure that comparison of protein level is based on the same scale of volume without variation due to sample dilution or condensation, proteins from a fixed sample volume, instead of a fixed protein amount, should be loaded onto 2-DE gels [15,18]. In the present study, proteins desalted from the same volume of original AH samples (120 μl) were loaded onto IPG strips (Immobiline 13-cm DryStrip, pH 3–10, GE Healthcare, NJ, USA) for simultaneous rehydration. Isoelectric focusing was then performed with the voltage–time program of 50 V for 12 h, 500 V for 1 h, 1000 V for 1 h, and 8000 V for a total of 120,000 Vh. The IPG strips were then equilibrated for 15 min at room temperature sequentially in 50 mM Tris, pH 8.4, containing 6 M urea, 2% SDS, 1% DTT, and 30% glycerol, then in 50 mM Tris, pH 8.4, 6 M urea, 2% SDS, 2.5% iodoacetamide, and 30% glycerol before SDS-PAGE. SDS-PAGE was run as the second-dimension separation for 6 h at 15 °C using a vertical electrophoresis system and 12.5% gels at 20 mA/gel. The gels were then fixed, stained using SyproRuby fluorescent dye according to the manufacturer's instructions (Invitrogen, Carlsbad, USA), and scanned using a Typhoon Trio laser scanner (GE Healthcare, NJ, USA).

2.6.

Spot detection, quantification, and comparison

Spot detection, gel matching, and spot quantification were carried out using ImageMaster 2D PlatinumTM (GE Healthcare, NJ, USA) analysis software. To estimate the percentage of a given

protein spot in the total proteins in the AH, the percent spot volume (PSV) of each protein spot on a 2-D gel was calculated by the software using the equation: PSV=(spot volume/total volume of all spots)×100%. Since a fixed volume of sample, instead of a fixed amount of proteins, was loaded onto gels, the PSV for different samples was measured on the same scale of volume without distortion by sample dilution or condensation. Thus, the estimated PSV can be used to approximate the original proportion of protein in the AH without considering the dilution factor. To compare levels of a given protein spot between different AH samples with different protein concentrations, the PSV, which is a ratio of protein level, was converted and normalized to a real protein concentration in AH fluid using the equation: Normalized protein level (NPL) = PSV × total protein concentration measured before desalting. Student's t-test was used to compare the log-transformed NPLs, which meet the normality and variance homogeneity assumption, between control and DR groups [39]. Sequential Goodness of Fit (SGoF) test was then performed to exclude false positive results of t-test caused by multiple hypotheses testing problem [40,41]. Protein spots that showed a significant difference in both t-test (p < 0.05) and SGoF test (p < 0.05) between the two groups were selected for protein identification.

2.7. In-gel digestion, MALDI-TOF MS, and protein identification Spots that showed significant difference for both t-test (p< 0.05) and SGoF test (p< 0.05) between the control and DR groups were manually excised and subjected to in-gel digestion, then the digested samples were extracted and analyzed by MALDI-TOF MS (Autoflex II, Bruker Daltonics, Bremen, Germany). For protein identification by peptide mass fingerprinting (PMF), the mass spectrum was obtained from signals generated from at least 500 laser shots. Mass spectra were processed using FlexanalysisTM software (Bruker Daltonics, Bremen, Germany). Mass (monoisotopic mass) lists were obtained using BiotoolsTM software (Bruker Daltonics, Bremen, Germany) and were used to search the SwissProt database (released 2011; Homo Sapiens taxonomy) using the MS-Fit database search engine (http:// prospector.ucsf.edu/prospector/mshome.htm). For each PMF search, the mass tolerance was set at 150 ppm, instrument was set as MALDI-TOF-TOF, and one missed tryptic cleavage was allowed. Carbamidomethyl cysteine was set as a constant modification, while oxidization of methionine and phosphorylation of serine, threonine and tyrosine were set as variable modifications. The specificity of each protein identity was finally confirmed by checking whether it matches the molecular mass and pI of the protein spot on 2-DE.

2.8.

Bioinformatics analysis of biological networks

The proteomics-identified proteins were analyzed using the database and web-tool STRING (Search Tool for the Retrieval of Interacting Genes/Proteins) (http://string.embl.de/) to obtain functional regulatory networks. The networks were algorithmically generated based on the interaction of the input proteins with other proteins from the STRING database. The biological functions of the networks were categorized into different clusters.

J O U RN A L OF P ROT EO M I CS 7 5 ( 2 0 12 ) 29 5 0 –2 95 9

2.9.

Western blot analysis

Western blotting was performed over the same biological replicates to validate the changes of expression of protein spot demonstrated by 2-DE. To ensure that the comparison of protein levels is based on the same volume scale without variation caused by sample dilution, a fixed volume of AH sample (20 μl) was loaded onto each gel lane. Briefly, an equal volume of 2× reducing SDS sample buffer was added to the AH and the mixture heated at 95 °C for 10 min. The height of samples within tips before and within gel wells after the sample loading were used as loading controls to monitor whether a fixed volume was exactly loaded for different samples. The samples (20 μl AH/gel lane) were subjected to 10% SDS-PAGE and the proteins transferred to a Hybond PVDF membrane (GE Healthcare, NJ, USA), which was then blocked at room temperature for 1 h in 20 ml of blocking buffer [Tris-buffered saline, pH 8.0, containing 0.05% Tween20 (TBST) and 5% skimmed milk]. The membrane was then incubated for 1 h at room temperature with a primary antibody diluted in blocking buffer, washed three times with TBST, and incubated for 1 h at room temperature with an appropriate secondary antibody in blocking buffer. The primary antibodies used were mouse antibodies against human apolipoprotein AI or podocan (1:5000) (both from Abcam, Cambridge, UK) or goat antibodies against human transferrin (1:10,000) (Bethyl Laboratories, Montgomery, TX), while the secondary antibodies used were rabbit anti-mouse (1:10,000, Chemicon) or donkey anti-goat (Abcam, Cambridge, UK) IgG antibodies. Finally, bound antibodies were visualized using an ECL detection kit (Millipore, MA, USA), and the intensity of the stained bands quantified using a densitometer.

2.10.

2953

Statistics

SPSS software (Chicago, IL, USA) was used for checking normality and variance homogeneity and performing student's t-test. Public SGoF+software (v7.2, http://webs.uvigo.es/acraaj/SGoF.htm) was used to perform SGoF test to exclude false positive results of t-test caused by multiple hypotheses testing problem [40,41]. A difference was considered significant at p<0.05 for both t-test and SGoF test. The data are presented as the mean±SD.

3.

Results

3.1. The total AH protein concentration is increased in DR patients The average total AH protein concentration measured before the desalting procedure in the DR group was significantly higher than that in the control group (421.50 ± 104.63 vs. 185.59 ± 51.07 μg/ml, p < 0.05), suggesting that the AH protein levels can be an indicator of DR.

3.2. Two-dimensional gel electrophoresis proteome profiles of the AH from controls and DR patients

AH protein concentration was also measured after the desalting procedure to check the recovery of protein to ensure that there was no unacceptable protein loss during desalting. The preliminary data showed insignificant difference in recovery rate between DR and control groups (data not shown), suggesting

Fig. 1 – Representative 2-DE gels for controls and DR patients. Eleven protein spots differentially expressed between DR and control groups are indicated by the arrows and labeled with the same protein name abbreviations used in Table 1. The positions of proteins which were absent in the control gel, but present in the DR gel, are marked by circles. The molecular mass is indicated on the right, while the pI range is shown at the bottom of each gel.

2954

J O U RN A L OF P ROT EO M IC S 7 5 ( 2 0 12 ) 29 5 0 –29 5 9

that a 2× difference in protein concentration did not lead to difference in protein recovery. Comparative 2-DE based proteomic analysis was performed on the AH from individual control subjects and DR patients. As shown in Fig. 1, 11 protein spots (indicated by the arrows) were differentially expressed between control and DR groups (p<0.05 for both t-test and SGoF analysis), and these protein spots were unequivocally identified by peptide mass fingerprinting (PMF); their characteristics are shown in Table 1, and the original MS of the 11 proteins were annotated and shown in the supplementary data (Supplementary Fig. 1 and Supplementary Table 1). Fig. 2 shows magnified spot images of the 11 DR-associated protein spots and their normalized protein levels. The identified proteins were apolipoprotein A-I (APOA1), serotransferrin (TF), keratin type I cytoskeletal 9 (KRT9), keratin type I cytoskeletal 10 (KRT10), podocan (PODN), matrix metalloproteinase 13 (MMP13), growth factor receptorbound protein 10 (GRB10), brain-specific angiogenesis inhibitor 1-associated protein 2 (BAIAP2), selenoprotein P (SEPP1), cystathionine beta-synthase (CBS), and retrotransposon gag domain-containing protein 1 (RGAG1). Levels of SEPP1, PODN and MMP13 were decreased, while levels of other proteins were increased, in the AH of the DR patients.

3.3. Validation of changes in protein expression by western blot analysis To confirm the changes in protein expression demonstrated by 2-DE, western blot analysis was used to check the levels of three secretory AH proteins APOA1, TF and PODN, selected from Table 1. Unlike other intracellular proteins released from damaged or dead cells, secretory proteins are mainly from

living cells and therefore can serve as biomarkers to indicate certain ongoing active biological processes in cells which are alive. The results showed that the expressional difference of these proteins on western blots between the control and DR groups (Fig. 3) was consistent with the proteomics data (Fig. 2).

3.4. Biological networks regulated by the identified DR-associated proteins Bioinformatics analysis was used to estimate the biological networks regulated by the 11 proteomics-identified proteins. The biological network clusters, as shown in Fig. 4, were classified to nutrition transport (cluster A), reorganization of microstructures and extracellular matrix (ECM) in retina and retinal blood vessels (cluster B), angiogenesis (cluster C), anti-oxidation and neuroprotection (cluster D).

4.

Discussion

Proteomics has not previously been applied to the analysis of susceptibility factors in AH contributing to development of DR. By comparing DM patients with DR and those without DR, the present study demonstrated that DR patients not only showed increased levels of total protein in the AH, but also had a AH protein profile which was different to non-DR subjects (Figs. 1–2). Eleven AH protein spots with altered levels in DR (Table 1) were estimated by bioinformatics, literature searching and text mining to play roles in DR development, impacting on nutrition transport, microstructure reorganization, angiogenesis, anti-

Table 1 – Identified DR-associated proteins. The abbreviation of protein name, MOWSE score, Swiss-Prot database accession number and protein name are shown, followed by the theoretical and observed molecular mass (Mr) and isoelectric point (pI) and number of matching peptides, sequence coverage calculated using BiotoolsTM software and p values from t-test and SGoF test. Protein MOWSE Accession name score No. abbreviation APOA1

3268

TF

P02647

Protein name

Theoretical Observed Matching Sequence P p value Mr(kDa) / Mr (kDa) / peptides coverage value (SGoF pI pI (%) (t-test) test)

Apolipoprotein A-I

30.7/5.6

26/5.4

6

22.8

66100000 P02787

Serotransferrin

77.0/6.8

80/6.74

15

27.7

KRT9

14200000 P35527

Keratin, type I cytoskeletal 9

62/5.1

57/7.54

10

23.9

KRT10

1090000

P13645

Keratin, type I cytoskeletal 10

58.8/5.1

50/5.9

13

29.8

PODN

1560

Q7Z5L7

Podocan

68.9/6.5

74/7.86

7

12.4

MMP13

1337

P45452

53.8/5.3

58/5.1

12

35.9

GRB10

1056

Q13322

67.2/8.1

77/9.1

9

23.6

BAIAP2

4291

Q9UQB8

60.8/9.0

61/8.5

21

67.8

SEPP1

6339

P49908

Collagenase 3 (Matrix metalloproteinase 13) Growth factor receptor-bound protein 10 Brain-specific angiogenesis inhibitor 1-associated protein 2 Selenoprotein P

43.1/8.1

51/5.7

10

48.6

CBS

129569

P35520

Cystathionine beta-synthase

60.5/6.2

60/6.9

19

45.4

RGAG1

1570

Q8NET4

Retrotransposon gag domain- 144/5.8 containing protein 1

149/5.9

15

37.1

6.83 E − 09 2.64 E − 10 5.80 E − 16 1.12 E − 16 2.77 E − 12 1.20 E − 13 1.13 E − 11 1.44 E − 15

4.88 E −02 2.70 E −03 1.12 E −10 7.98 E −12 8.12 E −05 1.44 E −06 5.02 E −04 1.40 E −09

1.52 E − 13 1.04 E − 14 1.65 E − 14

1.15 E −05 1.58 E −08 1.60 E −07

J O U RN A L OF P ROT EO M I CS 7 5 ( 2 0 12 ) 29 5 0 –2 95 9

2955

Fig. 2 – Quantification of protein spots. Upper panel: Magnified images of protein spots with significant increase or decrease in level in DR patients compared to controls. Lower panel: Statistical data for quantification of individual protein spots. All differences between the control and DR groups are significant at the p < 0.05 level (*). N = 11. Error bars indicate SD.

oxidation, and neuroprotection. Among them, APOA1, TF and PODN, are secretory proteins which can serve as biomarkers to indicate certain ongoing active biological processes in living cells. By western blot analysis, expressional changes of protein level of the three secretory proteins in DR patients were validated (Fig. 3). The roles of the identified 11 proteins in the development of DR are summarized in Fig. 4 and are discussed in detail below. Some issues for proteomics analysis in this study should be considered and discussed. First, log transformation of spot intensities/volumes is needed to fit normality and homogeneity of variance before t-test [39]. However, original non-transformed data should be used for data presentation, as log transformation distorts the original biological information, such as the values of

protein concentration. Second, false positive results of t-test might be obtained because of “multiple hypotheses testing problem” caused by interaction between different protein factors. These false positive results can be ruled out by SGoF analysis [40,41]. Third, although it is a routine procedure to use western blot analysis to validate the expressional changes of protein demonstrated by 2-DE analysis, there is still a significant disparity in protein level displayed by these two technologies (Fig. 2 vs. Fig. 3). This disparity might arise from the difference in detection methods, the electrotransfer step, types of gel and normalization of data. DM patients usually develop microvascular complications in certain organs, such as the kidney and eye, which cause breakdown of the blood-tissue barrier, leading to leakage of

2956

J O U RN A L OF P ROT EO M IC S 7 5 ( 2 0 12 ) 29 5 0 –29 5 9

Fig. 3 – Western blot confirmation of the expressional change of some proteomics-identified proteins. Upper panel: representative blots for APOA1, TF, and PODN, showing differences in level in a fixed volume (20 ul) of AH between the two groups (control vs. DR). The names of the proteins are indicated on the left. Lower panel: quantitative data for the indicated proteins. n = 11, * p < 0.05. Error bars indicate SD.

proteins from the blood into tissues and body fluids, such as urine in the kidney and intraocular fluids in the eye. This leakage can cause an increase in total protein level in the body fluids and the total protein concentration in the fluids can

be used as a marker for the severity of the microvascular complications. For example, proteinuria, an abnormal increase in urine protein level, is a hallmark of diabetic nephropathy in DM patients and has been used as a standard diagnostic and prognostic marker for DM-induced kidney complications in the clinic [42]. Similarly, an increase in protein level is observed in both the VH and AH in DM patients [43,44], and is probably caused by loss of endothelial barrier function in the eye [43,44]. Our data showing that total protein levels in the AH were increased in DR patients confirmed this observation, suggesting that the total protein levels in the AH could serve as a marker of DR. Although the AH does not come into direct contact with the retina, proteins released from the damaged retina and retinal blood vessels in DR can diffuse through the VH to the AH, as large molecules can be exchanged between these fluids through the VH–AH barrier (pathway I in Fig. 4) [36–38]. Some studies have also shown that proteins involved in DR pathogenesis simultaneously appear in the VH and AH [7–14], suggesting that, besides VH proteins, AH proteins can also serve as biomarkers to indicate pathological changes in the retina and/or retinal blood vessels in DM patients. Besides VH–AH diffusion, the retinal proteins can also enter the cilia–retina circulation through the disrupted blood–retinal barrier and spread to the AH via the blood– aqueous barrier (pathway II in Fig. 4) [34,45], so the overall retinal microvascular state can be estimated by examining the AH protein profile. Levels of two nutrition transport proteins, APOA1 and TF, were found to be altered in the AH in DR patients. APOA1 participates in the transport of cholesterol, while TF is responsible for iron transport. The homeostasis of either nutrient is important for maintaining the health of a tissue and an imbalance of these nutrients caused by dysregulation of their transport proteins is a factor leading to disease states [46,47]. In

Fig. 4 – Schematic diagrams of eye and protein-interaction networks respectively showing diffusion of retinal proteins to the AH through (I) the VH–AH barrier and (II) the cilio-retinal circulation and showing the roles of the identified AH proteins in the development of DR, including potential pathogenic roles in (A) nutrition transport, (B) reorganization of microstructures and extracellular matrix in retina and retinal blood vessels (C) angiogenesis, and (D) anti-oxidation and neuroprotection.

J O U RN A L OF P ROT EO M I CS 7 5 ( 2 0 12 ) 29 5 0 –2 95 9

terms of the eye, APOA1 overexpression is an early event in the retina of diabetic patients and is involved in the pathophysiology of DR [33,48]. A previous report demonstrated that APOA1 levels are increased in the VH in DR patients [33], while the present study showed a similar increase in the AH (Figs. 2–3), suggesting that disturbance of cholesterol homeostasis plays a role in DR development. The present study also showed that TF levels were elevated in the AH of DR patients. TF is actively synthesized in the eye and its concentration is relatively high in the intraocular fluids [49]. Since TF is a survival factor for eye tissues [50], DRassociated dysregulation of TF might have an impact on iron homeostasis in the retina and contribute to DR development. The intracellular and extracellular microstructures formed by cytoskeletal and ECM proteins play important roles in maintaining retina integrity and vision function [51–53]. In the present study, three microstructural proteins, i.e. KRT9, KRT10 and PODN, were identified in the AH of DR patients. KRT9 and KRT10 are intermediate filament proteins expressed in retinal epithelial cells [54], while PODN is a microvascular ECM protein responsible for ECM remodeling and the structural integrity of the blood vessel wall [55,56]. The appearance of these three structural proteins in the AH may indicate the loss of integrity of photoreceptors, retinal pigment epithelium and retinal blood vessels. Moreover, MMP13, a microstructure regulating protein, was also identified in the AH of DR patients. MMP13 belongs to a large family of proteases known to play an important role in degrading and remodeling the structural proteins in tissues associated with some diabetic states [57]. The changed levels in MMP in the AH in DR patients may indicate an impact of DR on the microstructure remodeling systems which control the integrity of the retina and retinal blood vessels. Hypoxia-induced angiogenesis, which leads to the growth of new blood vessels in the retina and other eye tissues during DR development, is a key factor that exacerbates DR and causes severe vision loss. Although some angiogenic factors, such as VEGF and erythropoietin, play well known roles in DR pathogenesis, the angiogenesis process is very complicated and can involve more protein factors. In the present study, two of the identified proteins (i.e. GRB10 and BAIAP2) have clearly documented functions that are directly linked to angiogenesis. For example, GRB10 can influence angiogenesis signal transduction by preventing degradation of VEGF receptors [58], and BAIAP2 can inhibit angiogenesis by blocking αvβ5 integrin in endothelial cells [59,60]. Thus, these two proteins could serve as intraocular fluid markers for estimating angiogenesis states during the development of DR. During DR development, retina tissues and intraocular fluids are driven to a more oxidative state, leading to oxidative stress which causes death of neuronal cells in retina [21]. In the present study, an antioxidant protein (SEPP1) and a neuroprotective protein (CBS) were identified. SEPP1 is a protein that contains 10 selenocysteine residues that protect host cells and the ECM against oxidative damage at sites of inflammation [61,62]; the decrease in SEPP1 levels in the AH in DR patients suggests that the DR-associated oxidative stress in eye tissues is caused, at least in part, by reduced power of the anti-oxidation system. CBS is an enzyme

2957

responsible for catalyzing the formation of hydrogen sulfide, which is a signaling neuromodulator with cytoprotective effects against apoptotic injury in retina [63]. The appearance of CBS in the AH in DR patients suggests that the hydrogen sulfide-mediated cytoprotection mechanism can be activated during the development of DR. In conclusion, some proteins involved in multiple biochemical events in the pathogenesis of DR were identified in the AH by proteomics analysis. Using bioinformatics tools, the biological networks impacted by these AH proteins were displayed (Fig. 4). These play roles in nutrition transport, microstructure reorganization, angiogenesis, anti-oxidation, and neuroprotection. Our data may shed light on the pathophysiological roles of these proteins in DR, and these proteins could have potential as AH biomarkers for prediction of DR development. Further studies for validation of biological and biomarker roles of the 11 proteins are needed to be performed in other independent patients with a bigger sample size using high-throughput methods, such as protein microarray. In addition, it is also needed to extend the data analysis to study the correlation between the 11 proteins and other known clinical parameters. Supplementary materials related to this article can be found online at doi:10.1016/j.jprot.2011.12.006.

REFERENCES

[1] Priem HA. Diabetes mellitus and the eye. Acta Clin Belg 2004;59:237–40. [2] Negi A, Vernon SA. An overview of the eye in diabetes. J R Soc Med 2003;96:266–72. [3] Grus FH, Joachim SC, Pfeiffer N. Proteomics in ocular fluids. Proteomics Clin Appl 2007;1:876–88. [4] Herber S, Grus FH, Sabuncuo P, Augustin AJ. Two-dimensional analysis of tear protein patterns of diabetic patients. Electrophoresis 2001;22:1838–44. [5] Grus FH, Sabuncuo P, Dick HB, Augustin AJ, Pfeiffer N. Changes in the tear proteins of diabetic patients. BMC Ophthalmol 2002;2:4. [6] Grus FH, Sabuncuo P, Augustin AJ. Analysis of tear protein patterns of dry-eye patients using fluorescent staining dyes and two-dimensional quantification algorithms. Electrophoresis 2001;22:1845–50. [7] Er H, Doganay S, Ozerol E, Yurekli M. Adrenomedullin and leptin levels in diabetic retinopathy and retinal diseases. Ophthalmologica 2005;219:107–11. [8] Murugeswari P, Shukla D, Rajendran A, Kim R, Namperumalsamy P, Muthukkaruppan V. Proinflammatory cytokines and angiogenic and anti-angiogenic factors in vitreous of patients with proliferative diabetic retinopathy and eales' disease. Retina 2008;28:817–24. [9] Roldan-Pallares M, Rollin R, Martinez-Montero JC, Fernandez-Cruz A, Bravo-Llata C, Fernandez-Durango R. Immunoreactive endothelin-1 in the vitreous humor and epiretinal membranes of patients with proliferative diabetic retinopathy. Retina 2007;27:222–35. [10] Funatsu H, Yamashita H, Sakata K, Noma H, Mimura T, Suzuki M, et al. Vitreous levels of vascular endothelial growth factor and intercellular adhesion molecule 1 are related to diabetic macular edema. Ophthalmology 2005;112:806–16. [11] Funatsu H, Yamashita H, Noma H, Mimura T, Nakamura S, Sakata K, et al. Aqueous humor levels of cytokines are related

2958

[12]

[13]

[14]

[15]

[16] [17]

[18]

[19]

[20]

[21]

[22] [23]

[24]

[25]

[26]

[27]

[28]

[29]

J O U RN A L OF P ROT EO M IC S 7 5 ( 2 0 12 ) 29 5 0 –29 5 9

to vitreous levels and progression of diabetic retinopathy in diabetic patients. Graefes Arch Clin Exp Ophthalmol 2005;243: 3–8. Jonas JB, Neumaier M. Erythropoietin levels in aqueous humour in eyes with exudative age-related macular degeneration and diabetic retinopathy. Clin Experiment Ophthalmol 2007;35: 186–7. Shinoda K, Ishida S, Kawashima S, Wakabayashi T, Matsuzaki T, Takayama M, et al. Comparison of the levels of hepatocyte growth factor and vascular endothelial growth factor in aqueous fluid and serum with grades of retinopathy in patients with diabetes mellitus. Br J Ophthalmol 1999;83: 834–7. Tashimo A, Mitamura Y, Nagai S, Nakamura Y, Ohtsuka K, Mizue Y, et al. Aqueous levels of macrophage migration inhibitory factor and monocyte chemotactic protein-1 in patients with diabetic retinopathy. Diabet Med 2004;21: 1292–7. Duan X, Xue P, Wang N, Dong Z, Lu Q, Yang F. Proteomic analysis of aqueous humor from patients with primary open angle glaucoma. Mol Vis 2010;16:2839–46. Missotten GS, Beijnen JH, Keunen JE, Bonfrer JM. Proteomics in uveal melanoma. Melanoma Res 2003;13:627–9. Izzotti A, Longobardi MG, Cartiglia C, Sacca SC. Proteome alterations in primary open angle glaucoma aqueous humour. J Proteome Res 2010;9:4831–8. Duan X, Lu Q, Xue P, Zhang H, Dong Z, Yang F, et al. Proteomic analysis of aqueous humor from patients with myopia. Mol Vis 2008;14:370–7. Grus FH, Joachim SC, Sandmann S, Thiel U, Bruns K, Lackner KJ, et al. Transthyretin and complex protein pattern in aqueous humor of patients with primary open-angle glaucoma. Mol Vis 2008;14:1437–45. Funding M, Vorum H, Honore B, Nexo E, Ehlers N. Proteomic analysis of aqueous humour from patients with acute corneal rejection. Acta Ophthalmol Scand 2005;83:31–9. Madsen-Bouterse SA, Kowluru RA. Oxidative stress and diabetic retinopathy: pathophysiological mechanisms and treatment perspectives. Rev Endocr Metab Disord 2008;9: 315–27. Cheung N, Mitchell P, Wong TY. Diabetic retinopathy. Lancet 2010;376:124–36. Adamis AP, Berman AJ. Immunological mechanisms in the pathogenesis of diabetic retinopathy. Semin Immunopathol 2008;30:65–84. Zhang X, Bao S, Hambly BD, Gillies MC. Vascular endothelial growth factor-A: a multifunctional molecular player in diabetic retinopathy. Int J Biochem Cell Biol 2009;41: 2368–71. Endo H, Naito T, Asahara T, Kajima M, Shiota H. Cytokines in the vitreous fluid of patients with proliferative diabetic retinopathy–vascular endothelial growth factor and platelet-derived growth factor are elevated in proliferative diabetic retinopathy. Nippon Ganka Gakkai Zasshi 2000;104:711–6. Grierson I, Heathcote L, Hiscott P, Hogg P, Briggs M, Hagan S. Hepatocyte growth factor/scatter factor in the eye. Prog Retin Eye Res 2000;19:779–802. Fort PE, Freeman WM, Losiewicz MK, Singh RS, Gardner TW. The retinal proteome in experimental diabetic retinopathy: up-regulation of crystallins and reversal by systemic and periocular insulin. Mol Cell Proteomics 2009;8:767–79. Gao BB, Phipps JA, Bursell D, Clermont AC, Feener EP. Angiotensin AT1 receptor antagonism ameliorates murine retinal proteome changes induced by diabetes. J Proteome Res 2009;8:5541–9. Liu S, Zhang Y, Xie X, Hu W, Cai R, Kang J, et al. Application of two-dimensional electrophoresis in the research of retinal proteins of diabetic rat. Cell Mol Immunol 2007;4:65–70.

[30] Gao BB, Chen X, Timothy N, Aiello LP, Feener EP. Characterization of the vitreous proteome in diabetes without diabetic retinopathy and diabetes with proliferative diabetic retinopathy. J Proteome Res 2008;7: 2516–25. [31] Garcia-Ramirez M, Canals F, Hernandez C, Colome N, Ferrer C, Carrasco E, et al. Proteomic analysis of human vitreous fluid by fluorescence-based difference gel electrophoresis (DIGE): a new strategy for identifying potential candidates in the pathogenesis of proliferative diabetic retinopathy. Diabetologia 2007;50:1294–303. [32] Ouchi M, West K, Crabb JW, Kinoshita S, Kamei M. Proteomic analysis of vitreous from diabetic macular edema. Exp Eye Res 2005;81:176–82. [33] Simo R, Higuera M, Garcia-Ramirez M, Canals F, Garcia-Arumi J, Hernandez C. Elevation of apolipoprotein A-I and apolipoprotein H levels in the vitreous fluid and overexpression in the retina of diabetic patients. Arch Ophthalmol 2008;126: 1076–81. [34] Thornit DN, Vinten CM, Sander B, Lund-Andersen H, la Cour M. Blood-retinal barrier glycerol permeability in diabetic macular edema and healthy eyes: estimations from macular volume changes after peroral glycerol. Invest Ophthalmol Vis Sci 2010;51:2827–34. [35] Hayreh SS. The Cilio-Retinal Arteries. Br J Ophthalmol 1963;47:71–89. [36] Barsotti MF, Bartels SP, Freddo TF, Kamm RD. The source of protein in the aqueous humor of the normal monkey eye. Invest Ophthalmol Vis Sci 1992;33:581–95. [37] Maurice DM. Flow of water between aqueous and vitreous compartments in the rabbit eye. Am J Physiol 1987;252: F104–8. [38] Fatt I. Flow and diffusion in the vitreous body of the eye. Bull Math Biol 1975;37:85–90. [39] Gustafsson JS, Ceasar R, Glasbey CA, Blomberg A, Rudemo M. Statistical exploration of variation in quantitative two-dimensional gel electrophoresis data. Proteomics 2004;4: 3791–9. [40] Carvajal-Rodriguez A, de Una-Alvarez J, Rolan-Alvarez E. A new multitest correction (SGoF) that increases its statistical power when increasing the number of tests. BMC Bioinformatics 2009;10:209. [41] Diz AP, Carvajal-Rodriguez A, Skibinski DO. Multiple hypothesis testing in proteomics: a strategy for experimental work. Mol Cell Proteomics 2011, doi:10.1074/mcp.M110. 004374. [42] Jha P, Das BK, Shrestha S, Majhi S, Chandra L, Sharma S, et al. Glycemic status, lipid profile and proteinuria in diabetic nephropathy. JNMA J Nepal Med Assoc 2010;49:143–6. [43] Fosmark DS, Bragadottir R, Stene-Johansen I, Berg JP, Berg TJ, Lund T, et al. Increased vitreous levels of hydroimidazolone in type 2 diabetes patients are associated with retinopathy: a case–control study. Acta Ophthalmol Scand 2007;85: 618–22. [44] Kato S, Oshika T, Funatsu H, Kokutei K, Yamashita H, Sawa M. Aqueous protein concentration in diabetics. Report 5. Relationship between aqueous protein concentration and pathological findings of iridial vessels. Nippon Ganka Gakkai Zasshi 1992;96:1000–6. [45] Schalnus R, Ohrloff C, Jungmann E, Maass K, Rinke S, Wagner A. Permeability of the blood-retinal barrier and the blood-aqueous barrier in type I diabetes without diabetic retinopathy: simultaneous evaluation with fluorophotometry. Ger J Ophthalmol 1993;2:202–6. [46] Khuseyinova N, Koenig W. Apolipoprotein A-I and risk for cardiovascular diseases. Curr Atheroscler Rep 2006;8: 365–73. [47] Ponka P, Lok CN. The transferrin receptor: role in health and disease. Int J Biochem Cell Biol 1999;31:1111–37.

J O U RN A L OF P ROT EO M I CS 7 5 ( 2 0 12 ) 29 5 0 –2 95 9

[48] Simo R, Garcia-Ramirez M, Higuera M, Hernandez C. Apolipoprotein A1 is overexpressed in the retina of diabetic patients. Am J Ophthalmol 2009;147:319–25 e1. [49] Laicine EM, Haddad A. Transferrin, one of the major vitreous proteins, is produced within the eye. Exp Eye Res 1994;59: 441–5. [50] Davidson MG, Harned J, Grimes AM, Duncan G, Wormstone IM, McGahan MC. Transferrin in after-cataract and as a survival factor for lens epithelium. Exp Eye Res 1998;66: 207–15. [51] Chauhan BK, Disanza A, Choi SY, Faber SC, Lou M, Beggs HE, et al. Cdc42- and IRSp53-dependent contractile filopodia tether presumptive lens and retina to coordinate epithelial invagination. Development 2009;136:3657–67. [52] Boycott BB, Hopkins JM. A neurofibrillar method stains solitary (primary) cilia in the mammalian retina: their distribution and age-related changes. J Cell Sci 1984;66: 95–118. [53] Evans T, Deng DX, Chen S, Chakrabarti S. Endothelin receptor blockade prevents augmented extracellular matrix component mRNA expression and capillary basement membrane thickening in the retina of diabetic and galactose-fed rats. Diabetes 2000;49:662–6. [54] Vinores SA, Van Niel E, Kim HJ, Campochiaro PA. Simultaneous expression of keratin and glial fibrillary acidic protein by the same cells in epiretinal membranes. Invest Ophthalmol Vis Sci 1992;33:3361–6. [55] Didangelos A, Yin X, Mandal K, Baumert M, Jahangiri M, Mayr M. Proteomics characterization of extracellular space components in the human aorta. Mol Cell Proteomics 2010;9:2048–62.

2959

[56] Shimizu-Hirota R, Sasamura H, Kuroda M, Kobayashi E, Saruta T. Functional characterization of podocan, a member of a new class in the small leucine-rich repeat protein family. FEBS Lett 2004;563:69–74. [57] Kardesler L, Biyikoglu B, Cetinkalp S, Pitkala M, Sorsa T, Buduneli N. Crevicular fluid matrix metalloproteinase-8, -13, and TIMP-1 levels in type 2 diabetics. Oral Dis 2010;16:476–81. [58] Murdaca J, Treins C, Monthouel-Kartmann MN, Pontier-Bres R, Kumar S, Van Obberghen E, et al. Grb10 prevents Nedd4-mediated vascular endothelial growth factor receptor-2 degradation. J Biol Chem 2004;279:26754–61. [59] Koh JT, Kook H, Kee HJ, Seo YW, Jeong BC, Lee JH, et al. Extracellular fragment of brain-specific angiogenesis inhibitor 1 suppresses endothelial cell proliferation by blocking alphavbeta5 integrin. Exp Cell Res 2004;294:172–84. [60] Yoon KC, Ahn KY, Lee JH, Chun BJ, Park SW, Seo MS, et al. Lipid-mediated delivery of brain-specific angiogenesis inhibitor 1 gene reduces corneal neovascularization in an in vivo rabbit model. Gene Ther 2005;12:617–24. [61] Higuchi A, Takahashi K, Hirashima M, Kawakita T, Tsubota K. Selenoprotein P controls oxidative stress in cornea. PLoS One 2010;5:e9911. [62] Steinbrenner H, Bilgic E, Alili L, Sies H, Brenneisen P. Selenoprotein P protects endothelial cells from oxidative damage by stimulation of glutathione peroxidase expression and activity. Free Radic Res 2006;40:936–43. [63] Biermann J, Lagreze WA, Schallner N, Schwer CI, Goebel U. Inhalative preconditioning with hydrogen sulfide attenuated apoptosis after retinal ischemia/reperfusion injury. Mol Vis 2011;17:1275–86.