Proteomic Analysis of Climatic Keratopathy Droplets

Abstract

purpose. To identify the proteins in the corneal droplets of climatic droplet keratopathy (CDK), a disease that results in the formation of droplets on the cornea. Progressive accumulation of droplets in CDK leads to visual loss.

methods. Proteomic mass spectrometry of the CDK specimens was performed after fractionation of proteins in 4% to 20% SDS-polyacrylamide gels. Droplets were derived from two human donors. Immunohistochemistry with antibodies was performed to confirm the presence of identified proteins on donor tissues from patients with CDK and control subjects.

results. Proteomic analyses revealed identification of 105 proteins in CDK specimens. Immunohistochemical analyses confirmed localization of annexin A2 and glyceraldehyde 3-dehydrogenase (GAPDH), proteins identified by proteomic analyses in CDK specimens. The proteins were subjected to analyses with the Kyoto Encyclopedia of Genes and Genomes (KEGG) Database which showed that a few biochemical pathways were more frequent for the identified proteins.

conclusions. Approximately 105 proteins were identified in CDK specimens, and a subset of them was confirmed by immunohistochemistry. Several of these may play a role in fibril or deposit formation.

Climatic droplet keratopathy (CDK) is an acquired degenerative disease of the cornea that has been found to be highly prevalent in certain rural communities around the world. ^{1 2 3} It is characterized by the haziness and opalescence of the most anterior layers of the cornea. In the initial stages, the cornea is often characterized by multiple tiny and tightly confluent, translucent, subepithelial yellow-gold deposits, localized peripherally close to the temporal and/or nasal limbus. 

In many regions of the world where CDK has been reported, such as Patagonia in Argentina, ^{3 4} North Africa, the Red Sea region, ² and the Punjab in India, ⁵ the climate is characterized by high solar irradiation and reflection, aridity, constant winds, and, in some areas, elevated diurnal temperatures. Thus, there is exposure to high solar ultraviolet radiation because of ground reflection and the lack of cloud cover and shadows. Constant intense winds carrying particles of dust, sand, or ice that can provoke subtle microerosion of the corneal epithelium may play a role in CDK. Although the etiology of CDK is unknown, there is strong epidemiologic evidence that risk factors such as lifetime exposure to those climatic conditions may be involved. ^{6 7} Occurrence of CDK in general does not correlate with any known ocular or other systemic disease. It is primarily a corneal disease; previous studies have rarely found it to be associated with conjunctiva. ³ A higher prevalence of pinguecula, pterygium, cataract, and pseudoexfoliation has been found in patients with CDK compared with age-matched control subjects. ³ Despite the age of patients affected by CDK and the rigorous climatic and environmental features of the region, dry eye is not a common finding among the patients with CDK or control subjects. ⁴ Based on clinical findings, the patients have been divided into three stages or grades depending on the degree of severity. ^{3 7} Grade 1 is characterized by multiple tiny and tightly confluent, translucent, subepithelial deposits, best observed with back-scattered slit illumination and high magnification, with a prelimbal fringe of clear cornea. In grade 2, haziness spreads over the inferior two thirds of the cornea, giving a tarnished appearance. Grade 3 is characterized by the presence of clusters of golden subepithelial droplets of different sizes, some reaching 1 mm in diameter. ³ CDK predominantly occurs in adults, principally in men who work as laborers and are exposed to injurious environments including light throughout their lives, compared with women or children. ^{3 4 6 7 8 9 10} The correlation of corneal aesthesia with CDK has not been systematically evaluated. In many patients with CDK, a moderate to severe corneal hypoesthesia has been observed in advanced cases. ^{3 4} There appears to be a correlation between the grade of CDK and corneal sensitivity. 

Histopathologic examination of CDK tissue under a light microscope often demonstrates globular deposits of different sizes under the corneal epithelium, within Bowman’s membrane and the anterior stroma. The coalescence and increased volume of these spherules may cause disruption of Bowman’s membrane and elevation and thinning of the corneal epithelium. ^{1 11 12 13} The droplets generally lack positive staining for fat or calcium. ^{1 12 13} Electron microscopy has shown that the globules are round, electron-dense, sharply demarcated structures, always surrounded by basement membrane material and adjacent disorganized collagen fibrils. ^{1 13} The origin and exact composition of the droplets remain unclear; however, proteins appear to be at least some of the components. ^{11 12 13} Analyses of CDK globules in corneas undergoing penetrating keratoplasty (PKP) showed high protein content, with average protein contents of 72 μg/mg of wet tissue. SDS-PAGE has shown that the proteins of molecular mass 20 to 300 kDa are contained in the droplets, with a major fraction appearing to be of molecular mass 67 kDa. ^{11 14} Recent immunohistochemical studies in surgical specimens of CDK with monoclonal antibodies to N-(carboxymethyl)-l-lysine (CML), N-(carboxyethyl)-l-lysine (CEL), pyrraline, pentosidine, and imidazolone have shown these moieties to be immunoreactive for protein modifications in CDK. ¹⁵ We have described these droplets as superficial, only to indicate that they occur in and just beneath the epithelial layer and do not penetrate the deep stroma and endothelial layers of the cornea. Further investigation and protein identification in the droplets (CDK specimens) is expected to advance our understanding of the pathogenesis of this disease. Given the nature of the disease, of particular interest would be to identify enzymatic proteins or proteins that bear the potential for deposit formation. Toward this goal, the proteomic analyses of superficial corneal specimens obtained during PKP was performed. We expect that targeting pathways most frequently involving the identified proteins early on and simultaneously will be helpful for intervention. Toward this goal, we performed bioinformatic analyses to determine the biochemical pathway of identified proteins. 

Methods

Tissue Procurement and Isolation of CDK Specimens

The CDK specimens (predominantly droplet-containing/droplet-rich tissue samples) were obtained from the cornea of a patient undergoing corneal transplantation. The cornea with CDK droplets was procured from a donor after surgery, according to a protocol approved by institutional review process and the tenets of the Declaration of Helsinki. The corneal specimens: cases 1 and 2 were the corneas from left eyes of male patients (outdoor laborers), case 1 is a 67-year-old, and case 2 is a 72-year-old, both of them with typical CDK grade 2 in the right eye and grade 3 in the left eye. Both patients had had a bilateral, painless, and slowly progressive decrease in visual acuity over the past 10 years. Best corrected visual acuity (BCVA) was 20/200 in the right eye and counting fingers in the left eye. Both eyes in both patients presented moderately prominent nasal pingueculae and moderate nuclear sclerosis of the lens. In both eyes of both patients, no abnormalities of the iris were observed biomicroscopically. Ocular tension was normal in both eyes of both patients, and the fully dilated fundus examination was unremarkable in both eyes of both patients. The CDK specimen was collected by using sharp scissors and fine forceps, capturing an area of the droplets for examination by microscope. 

Protein Analyses

The proteins were extracted from CDK specimens by using 125 mM Tris-HCl (pH 7.0), 100 mM NaCl, 0.1% Triton X-100, 0.1% Genapol C-100, and 0.1% SDS. Insoluble materials were removed by centrifugation (8000g for 5 minutes), and soluble proteins were quantified by the Bradford assay. ¹⁶ Proteins were fractionated over 4% to 20% gradient SDS-polyacrylamide gels (Invitrogen Inc., Carlsbad, CA). The proteins (10 μg) fractionated on the gels were stained with blue gel stain (GelCode; Pierce Biotechnology, Inc., CA). The protein bands were excised, destained, reduced with dithiothreitol (DTT), alkylated with iodoacetamide, and subjected to mass spectrometric analysis. 

Mass Spectrometry

For protein identification, gel slices were excised and digested in situ with sequencing-grade trypsin (Promega Biosciences, Inc., Madison, WI). Digestion mixtures were loaded onto precolumns (360 mm OD × 100 mm ID fused silica; Polymicro Technologies, Phoenix, AZ), packed with 3-cm irregular C18 (5–15 μm nonspherical; YMC, Inc., Wilmington, NC), and washed with 0.1 M AcOH for 5 minutes before switching in-line to the resolving column (7-cm spherical C18, 360 × 100). Once the columns were in line, the peptides were gradient eluted with a gradient of 0% to 100% A in 30 minutes, where A was 0.1 M AcOH in nanopure H₂O, and B was 0.1 M AcOH in 80% MeCN. All samples were analyzed on a mass spectrometer (model LTQ; Thermo Electron Finnigan, San Jose, CA). Electrospray was accomplished with a spray system (TriVersa Nanomate; Advion Biosystems, Ithaca, NY) with a voltage of 1.7 kV and a flow rate of approximately 250 nL/min. The mass spectrometer was operated in a data-dependent mode with the top five most abundant ions in each spectrum being selected for sequential tandem (MS/MS) experiments. The exclusion list was used (1 repeat, 180 seconds return time) to increase the dynamic range. All MS/MS spectra were searched (Sequest, ver. 2.7; Sequest Technologies, Inc., Lisle, IL, and Mascot; Matrix Science, Boston, MA) using NCBI nonredundant, Ensemble ¹⁷ and Swiss Prot ¹⁸ databases. Searches of database entries were restricted to Metazoa (animals) and allowed a maximum of two missed cleavages. The protonated molecule ions MH⁺ and monoisotopic were defined for the peak mass data input. For protonated MH⁺ peptides, Sequest Sp and Xcorr cutoff scores were 500 and 1.8, respectively. For Mascot searches, the score was greater than 78. All spectra were visually inspected for determination of correct database assignment. The potential chemical modifications of a peptide such as the alkylation of a cysteine, carbamidomethyl (C), and the oxidation of a methionine residue (M) and acetylation of lysine (K) were also considered in the search. For all mass lists, the peptide tolerance/error was at most 50 ppm, and no restriction was applied for both the protein isoelectric point and molecular weight. Database search results were tabulated and visually inspected (Scaffold; Proteome Software, Portland, OR). These methods are routinely used in our laboratory. A combined list of all identified proteins was prepared. The Swiss Prot and GenBank accession numbers for proteins were obtained, and the latter was used for search against the Dragon database (http://pevsnerlab.kennedykrieger.org/annotate.htm) to obtain information from the Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway Database (http://www.genome.jp/kegg/pathway.html). The identified proteins were also subjected to comparison with a normal human soluble corneal proteomic dataset, ¹⁹ by using suitable modification of an available in-house written macro program (Excel; Microsoft, Redmond, WA). ²⁰  

Immunohistochemistry

Histologic evaluations were made according to published protocols for human ocular tissues used in our laboratory. ^{21 22 23} The 4% paraformaldehyde in phosphate-buffer-fixed and subsequently paraffin-embedded anterior eye sections (∼8 μm) were prepared and stained with commercially procured antibodies to annexin and glyceraldehyde 3-phosphate dehydrogenase (GAPDH). The presence of annexin and GAPDH proteins was verified by florescence analyses with Alexa 488- and Alexa 594-coupled secondary antibodies (Invitrogen-Molecular Probes, Eugene, OR). To ensure identical processing and uniform exposure, control samples (without antibody and primary antibody treated sections) were examined side by side on the same slide. Fluorescence images were obtained with a laser scanning confocal microscope (model TCS-SP5; Leica Microsystems, Exton, PA). A series of 1-μm x–y (en face) images were collected and stitched together for an image representing a three-dimensional projection of the entire 8-μm section. Confocal microscopic panels were composed with image-analysis software (Photoshop 8.0; Adobe Systems, San Jose, CA). 

Results

Isolation and Fractionation of CDK Specimens

Anterior segment images of affected human eyes were taken with the informed consent of the patients (Fig. 1) . The CDK that affects people from the Patagonia region of Argentina ^{3 4} is characterized by the presence of characteristic droplets in the cornea that are more prominent in advanced cases (Figs. 1A 1B) . Characteristic droplet-containing specimens, as indicated by the arrow (Fig. 1A) , were excised and subjected to fractionation on SDS-PAGE. The fractionated CDK specimen proteins were subjected to staining with Coomassie blue (Fig. 2)and subsequently subjected to mass spectrometric analyses. 

Proteomic Identification and Bioinformatic Analyses

Each band in both samples (Fig. 2)was excised, reduced, alkylated, and trypsin digested, and the resultant peptides were subjected to sequence identification by mass spectrometry and to protein identification by database searches. Using this approach, we were able to identify 105 proteins in the CDK specimens (Table 1) . Only the proteins that presented good mass spectra and at least two sequenced peptides were included in the identification list shown in the table. All identified proteins were examined to determine their position in the cellular biochemical pathways by using Dragon annotation tools and the KEGG database. As shown in Figure 3 , this resulted in the determination of the most frequent biochemical pathways to which the identified CDK specimen proteins belong. Most of the proteins identified in CDK specimens belong to cell junctions (Fig. 3)followed by glycolysis, suggesting that cell-junction proteins are likely to be involved in CDK’s etiology. Several identified proteins such as lysozyme, transketolase, pyruvate kinase, aldehyde dehydrogenase, and retinal dehydrogenase 1 (Table 1)are enzymatic. We observed differences in the number of proteins identified between the two samples (Table 1) , and because of the identical outcome in two technical repeats (not shown), we attribute these differences to the nature of these samples. It is likely that the samples posed differences in their digestibility and ionization due to the inherent differences leading to identification in the different number of proteins in them. CDK droplets are often yellow and have not been found to stain positively for fat ^{12 13} ; however, the catalysis of the natural small molecule substrates for several of these enzymes, such as retinal dehydrogenase and aldehyde dehydrogenase, may lead to local accumulation of their products in the droplets, which would cause the yellow color. Several of the proteins (Table 1)such as the isoform of plectin-1 ^{24 25} , desmoplakin, ^{26 27} tenascin, ²⁸ and α-actinin-4 ²⁹ are either secreted extracellular matrix (ECM) proteins or have the potential to form deposits and, under certain conditions, to initiate fibril formation, leading to formation of isolated deposits. Many of these proteins bind to proteoglycans, ²⁸ which may provide the nucleation necessary to form deposits. Several proteoglycans, such as gelsolin and decorin, were also identified in the CDK specimens. Several of the proteins or their mutant variants have also been implicated in disease processes such as keratosis ^{26 27} or glomerulosclerosis, ²⁹ which are characterized by deposits, plaques, and lesions. Since surgical techniques are far from perfect, the possibility that samples of droplets in the cornea are contaminated with surrounding cells from normal cornea cannot be ruled out. We have performed a comparison of proteins identified in the current analyses with a previously reported dataset of soluble human cornea proteins that were identified by peptide mass fingerprinting and tandem mass spectrometry. ¹⁹ In the previous investigation normal cornea, 1-D and 2-D gel fractionation, and several different detergent extractions were used, and the total protein data from these analyses were compared with the current proteomic data from droplets (Supplementary Table S1). A previous study, due to its utilization of normal cornea, was not limited by tissue availability. Many cell junction proteins, a category that appeared most frequently among the identified proteins in our bioinformatic analysis (Fig. 3)also has high potential to initiate formation of deposits. Epiplakin, which has been uniquely observed in droplets (Supplementary Table S1), and desmoplakin are proteins ³⁰ that play roles in the formation of ceramide deposits. Previous proteomic and bioinformatic analyses have also suggested that the importation of plasma proteins into the human cornea ³¹ and the presence of hemoglobin, endoplasmin, and serum albumin in the droplets (Table 1 , Supplementary Table S1) are consistent with this hypothesis. Inactivation of proteolytic activities suggested by the presence of calpain-1 and serpin among the identified proteins and uniquely present in the droplets (Supplementary Table S1) could be a mechanism for formation of deposits. It may be that these proteases are posttranslationally modified due to oxidative modification or to UV cross-linking or both, leading to nonclearance of some initial surface deposition or a submicroscopic nuclear deposition. 

Immunohistochemical Analyses

Annexin and GAPDH are among the proteins that were identified in CDK specimens (Table 1) . The entire length of cornea from a normal donor was subjected to immunohistochemical analyses by using antibodies to annexin and GAPDH and confocal microscopy (Fig. 4) . Nuclear staining was performed to show corneal cellularity (Fig. 4A) . As shown in Figure 4B , the presence of annexin in the normal cornea was rather toward the limbus region at the edge of the cornea, not in the central part. The absence or weak presence of annexin in the central cornea is further shown in Figure 4C , and a relative increase in annexin is observed along the edges of the cornea (Fig. 4D); however, in multiple control specimens (corneas obtained from different healthy donors) very weak staining was observed for annexin in the cornea, with the exception of the presence in the limbus region. 

In CDK specimens, annexin was captured in abundance from the central location, prominently associated with droplets, which is in contrast to its presence in normal cornea (Fig. 5A) . Further immunohistochemical analyses of corneas with CDK showed a strong presence of annexin and GAPDH in the droplets and a rather diffused presence in the rest of the cornea. In certain instances, the droplet-like deposits were present in the cornea (Supplementary Fig. S1). The antibody control (without primary antibody), in identical settings, did not show any fluorescence (Fig. 5C) . 

Discussion

In the present analysis, we identified approximately 105 proteins in CDK specimens. As shown in Figure 3 , the most frequent pathways for which the proteins were identified were cell junction, focal adhesion, and regulation of cytoskeleton in addition to energy metabolism-associated proteins. 

Annexin A1 and A2, both were identified in our proteomic analyses, and immunohistochemical analysis has shown the presence of annexin A2 in CDK specimens. It is important to note that extensively positive staining for annexin II has been shown in corneal epithelial basal cells in rat cornea ³² in contrast to very weak staining in the epithelial cell layer in the present investigation (Fig. 5) . However, similar results were observed with different commercial antibodies to annexin II (rabbit polyclonal [ab19415] and two mouse mAbs [ab8146;Abcam Inc., Cambridge, MA, and H00000302-M02; Novus Biologicals, Littleton, CO]). The observed differences between human and rat corneas are probably due to epitope masking in different immunohistochemical procedures, or human cornea may be inherently slightly different from rat cornea, or the translocation of annexin II from cytoplasm to cell surface is less effective in humans, giving a diffuse, less-intense staining. Annexin A1, ³³ and to a lesser extent annexin A2, ³⁴ as well as GAPDH, ³⁵ are known for membrane fusion. If the CDK deposits form due to the formation of fused membrane that eventually becomes lipid-containing fusion of proteins, then the membrane fusion property of annexin A1 and -A2 and GAPDH would help catalyze some of the stages of the fusion process. Annexin A1 has been implicated in the formation of fibrotic deposits in the lungs, in cultured lung cells after irradiation, ³⁶ and in the aggregation and fusion of complex liposomes. ³⁷ It has been found to be associated with granular material in the skin as well. ³⁸ Annexins bind to lipids, are key mediators of cellular dynamic membrane-cytoskeleton interactions, ³⁹ and have been implicated in a variety of inflammatory pathways, including the regulation of cell death signaling and in phagocytic clearance of apoptotic cells. ^{39 40 41 42} These processes that are likely to be altered in corneas predisposed to form CDK droplets and annexins as well GAPDH overexpression have the potential to mediate some of the early or intermediate steps. It is also important to note that GAPDH, which is a housekeeping protein has been noted in the normal cornea (Fig. 5) , however, the staining is rather diffuse compared with that in the droplets. The possibility that the droplets have substances that release spontaneous fluorescence that may augment secondary antibody-mediated fluorescence cannot be ruled out. A consecutive section without antibody or stained with a nonspecific mouse monoclonal MOPC-21 antibody, ^{43 44} however, did not show such fluorescence (Fig. 5C) . Proteomic identification and immunohistochemical observation of the droplets only suggest an elevated presence of GAPDH and annexin in the droplets and not a complete lack of them in the normal cornea. It is likely that these observations suggest increased accumulation of GAPDH and annexin in CDK compared with normal cornea. The proteomic identification and immunohistochemical confirmation presented herein will pave the way for further mechanistic investigation into the role of these proteins in the initial stages of change in the cornea leading to droplet formation. 

Research on CDK is limited by the lack of a suitable animal model, but perturbation of these pathways in organ or tissue culture models as well as in animal models may provide some clue about the initiation of protein deposition. Analysis of the presence of polymorphisms of identified enzymatic proteins will also provide clues as to whether these activities are responsible for the initiation of reactions that lead to deposit formation. Continued investigation will also provide further details about the proteins, particularly secreted ECM proteins with the potential to form deposits. 

 Supported in part by a career development award from Research to Prevent Blindness (SKB), National Institutes of Health Grant S10 RR019382 and National Eye Institute Grant P30 EY014801 (BPEI center core grants), an unrestricted grant to the University of Miami from Research to Prevent Blindness, and the HHMI (Howard Hughes Medical Institute) internship program (MM, DL).

 Submitted for publication November 7, 2007; revised February 6, 2008; accepted May 14, 2008.

 Disclosure: M. Menegay, None; D. Lee, None; K.F. Tabbara, None; T.A. Cafaro, None; J.A. Urrets-Zavalía, None; H.M. Serra, None; S.K. Bhattacharya, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Image of the anterior part of the eye. (A) Image of an eye with grade 3 climatic droplet keratopathy with typical subepithelial golden-yellow confluent vesicles or drops. (B) Same case examined biomicroscopically with higher magnification and retroillumination better delineating the vesicles.

Figure 2.

 

Fractionation of proteins from human climatic keratopathy droplets. The isolated droplets from cornea removed during transplantation was extracted and separated on a 4% to 20% SDS-polyacrylamide. Samples 1 and 2 are from cases 1 and 2, respectively. The gel bands from top to bottom were excised and subjected to LC MS/MS analysis.

Table 1.

 

Proteins from Climatic Keratopathy Droplets on Cornea

Protein Name	Swiss-Prot ID	Genbank ID or IPI Acc. No.	KEGG Pathway*	MW (kDa)	Peptide Matches
										Sample 1	Sample 2
14-3-3 Protein gamma	P61981	NM_012479	4110	28	2	0
14-3-3 Protein sigma	P31947	NM_006142	4110	28	4	6
14-3-3 Protein zeta/delta	Q6P3U9	NM_003406	4110, 5130, 5131	28	2	0
24 kDa Protein		IPI00514244		24	5	0
60S Ribosomal protein L3	P39023	NM_001033853		46	3	0
Actin, alpha skeletal muscle	P62736	NM_001613		42	4	3
Actin, cytoplasmic 1	P60709	NM_001101	1430, 4510, 4520, 4530, 4670, 4810, 5130, 5131	42	16	16
Actin-like protein 2	P61160	NM_005722		45	2	0
Actin-related protein 2/3 complex subunit 1B	O15143	NM_005720	4810	41	2	0
Aldehyde dehydrogenase	P30838	NM_000691	10, 53, 71, 120, 220, 280, 310, 340, 350, 360, 380, 410, 561, 620, 640, 650, 903, 980	50	12	4
Alpha 3 type VI collagen isoform 1	P12111	NM_057166	1430, 4510, 4520	344	32	8
Alpha 3 type VI collagen isoform 3	P12111	NM_057166	1430, 4510, 4520	322	0	39
Alpha-actinin-4	O43707	IPI00013808		105	5	0
Annexin A1	P04083	NM_000700		39	3	4
Annexin A2 isoform 1	P07355	NM_001002858		40	20	6
Apolipoprotein D precursor	P05090	NM_001647		21	3	0
ATP synthase alpha chain, mitochondrial	P25705	NM_001001937	190	60	4	0
ATP synthase beta chain, mitochondrial	P06576	NM_001686	190	57	4	0
Calpain-1 catalytic subunit	P07384	NM_005186	4210	82	4	0
Clathrin heavy chain 1	Q00610	NM_004859	5040	192	13	0
Collagen, type VI, alpha 1		IPI00719088		109	7	11
Elongation factor 1-alpha 1	P68104	NM_001402		50	7	6
Endoplasmin	P14625	NM_003299	5215	92	7	0
Enolase 1	Q8WU71	NM_001428	10	47	9	12
Epiptakin		IPI00010951		553	13	7
Fructose-bisphosphate aldolase A	P04075	NM_000034	10, 51, 30, 710	39	3	0
Glyceraldehyde-3-phosphate dehydrogenase	P04406	NM_002046	10, 1510, 5010, 5040, 5050	36	13	0
Growth-inhibiting protein 12	Q8TCD2	NM_002343		78	3	6
Heat shock 70 kDa protein 1	P08107	NM_005345		70	7	0
Heat-shock protein beta-1	P04792	NM_001540		23	12	5
Hemoglobin subunit beta	P02042	NM_000519		16	0	2
Histone H4	P62805	NM_003543		11	5	3
Hypothetical protein	P11021	NM_005347	1510, 4612, 5060	72	2	0
Hypothetical protein DKFZp686I04196		IPI00399007		46	3	3
Hypothetical protein DKFZp761K0511	P08238	NM_007355	4612	85	2	2
IGHG1 protein		IPI00448925		60	2	0
IGHM protein		IPI00472610		53	3	4
IGKV1–5 protein		IPI00419424		26	2	2
IGLC1 protein		IPI00154742		25	2	0
Isoform 1 of gelsolin	P06396	NM_198252	4810	86	2	0
Isoform 1 of heat shock cognate 71 kDa protein	P11142	NM_153201		71	9	3
Isoform 1 of keratin, type I cytoskeletal 13	P13646	NM_153490	1430	50	7	5
Isoform 1 of myosin-14	Q9H882	NM_001077186	4530, 4810	228	3	0
Isoform 1 of plectin-1	Q15149	NM_201382		532	11	7
Isoform 1 of protein-glutamine gamma-glutamyltransferase 2	P21980	NM_198951	5040	77	2	0
Isoform 1 of tenascin	P24821	NM_002160	1430, 4510, 4512	241	0	7
Isoform 2 of Bcl-2-like 13 protein	Q9BXK5	NM_015367		53	2	0
Isoform 2C2 of collagen alpha-2(VI) chain	Q6P0Q1	NM_001849	1430, 4510, 4512	109	8	11
Isoform A of decorin	P07585	NM_133507	4350	40	6	0
Isoform A of lamin-A/C	P02545	NM_170708	1430	74	9	6
Isoform alpha of tripartite motif-containing protein 29	Q14134	NM_012101		66	5	0
Isoform alpha-6X1X2B of integrin alpha-6	P23229	NM_001079816	1430, 4510, 4512, 4514, 4640, 4810, 5222	127	2	0
Isoform B1 of heterogeneous nuclear ribonucleoproteins A2/B1	P22626	NM_031243		37	5	0
Isoform C1 of heterogeneous nuclear ribonucleoproteins C1/C2	Q96HM4	NM_001077442		32	6	0
Isoform DPI of desmoplakin	P15924	NM_004415		332	13	3
Isoform long of collagen alpha-1(XII) chain	Q99715	NM_080645		333	4	18
Junction plakoglobin	P14923	NM_021991	5221	81	2	0
Keratin 4	P19013	NM_002272	1430	64	3	2
Keratin, type I cytoskeletal 10	P13645	NM_000421	1430	60	8	4
Keratin, type I cytoskeletal 12	Q99456	NM_000223	1430	53	27	21
Keratin, type I cytoskeletal 14	P02533	NM_000526	1430	51	33	29
Keratin, type I cytoskeletal 16	P08779	NM_005557	1430	51	0	9
Keratin, type I cytoskeletal 17	Q04695	NM_000422	1430	48	3	11
Keratin, type I cytoskeletal 18		IPI00554788		49	6	0
Keratin, type I cytoskeletal 19	P08727	NM_002276	1430	44	11	5
Keratin, type I cytoskeletal 9	P35527	NM_000226	1430	62	6	2
Keratin, type II cytoskeletal 1	P04264	NM_006121	1430	66	12	7
Keratin, type II cytoskeletal 2 epidermal	P35908	NM_000423	1430	66	9	2
Keratin, type II cytoskeletal 3	P12035	NM_057088	1430	64	28	19
Keratin, type II cytoskeletal 5	P13647	NM_000424	1430	62	40	32
Keratin, type II cytoskeletal 6A	P02538	NM_005554	1430	60	18	20
Keratin, type II cytoskeletal 6B	P04259	NM_005555	1430	60	9	7
Keratin, type II cytoskeletal 6C	P02538	NM_005554	1430	60	5	12
Keratin, type II cytoskeletal 6E	P48668	NM_173086	1430	60	7	11
Keratin, type II cytoskeletal 8	P05787	NM_002273	1430	54	3	2
Lactate dehydrogenase A	P00338	NM_005566	10, 272, 620, 640	37	4	3
Laminin receptor-like protein LAMRL5	P08865	NM_002295		33	2	0
Leukocyte elastase inhibitor	P30740	NM_030666		43	2	0
L-lactate dehydrogenase B chain	P07195	NM_002300	10, 272, 620, 640	36	6	0
Lysozyme C precursor	P61626	NM_000239		17	6	5
Macrophage capping protein	P40121	NM_001747		38	3	2
Major vault protein	Q14764	NM_017458		99	3	0
Peroxiredoxin-1	Q06830	NM_002574		22	3	0
Phosphoglycerate kinase 1	P00558	NM_000291	10, 710	44	7	4
Similar to epiplakin 1		IPI00735278		323	0	3
Prohibitin-2		IPI00027252		33	2	0
Pyruvate kinase 3 isoform 1	P14618	NM_182471	10, 230, 620, 710, 4910, 4930	58	6	6
Ras-related protein Rab-11B	P62491	NM_004663		24	2	0
Ras-related protein Rab-14	P61106	NM_016322		24	4	0
Ras-related protein Rab-25	P57735	NM_020387		23	2	0
Retinal dehydrogenase 1	P00352	NM_000689	830	55	2	0
Ribosomal protein L11	P62913	NM_000975	3010	20	2	0
RNA binding motif protein, X-linked-like 1	P38159	NM_002139		42	4	0
Serotransferrin	P02787	NM_001063		77	0	3
Serpin B5	P36952	NM_002639		42	6	0
Serum albumin	Q9P1I7	NM_000477		69	16	23
Transforming growth factor-beta-induced protein ig-h3	Q15582	NM_000358		75	14	13
Transketolase	P29401	NM_001064	30, 710	68	11	6
Transmembrane emp24 domain-containing protein 2	Q15363	NM_006815		23	2	0
Triosephosphate isomerase	P60174	NM_000365	10, 51, 31, 710	31	2	0
Tubulin alpha-1 chain	P68366	NM_006000	4540, 5130, 5131	50	0	3
Tubulin alpha-6 chain	P68363	NM_006082	4540, 5130, 5131	50	5	0
Tubulin beta-2C chain	P68371	NM_006088	4540, 5130, 5131	50	5	0
Ubiquitin and ribosomal protein S27a	P62979	NM_002954		18	3	0
Villin 2	P15311	NM_003379	4670, 4810, 5130, 5131	69	3	0
Vimentin	P08670	NM_003380	1430	54	3	0

*  The KEGG pathway analysis was performed using the web-based Dragon database (http://pevsnerlab.kennedykrieger.org/annotate.htm) annotate tools. A few accession numbers lacked KEGG pathway entries.

Figure 3.

 

Analyses identified proteins using the KEGG Database for biochemical pathways. Total identified proteins were subjected to analyses of pathways. (A) The pathways (pathway numbers are as in the KEGG Database) that had a frequency higher than two in proteins from CDK specimens are presented. Pathways related to infection have been omitted. (B) The names of pathways identified by number in (A).

Figure 4.

 

Immunohistochemical analyses of identified proteins. (A) DAPI-stained anterior human cadaveric eye segments. (B) The same sections were stained with anti-annexin A2 secondary antibody coupled with Alexa 488. The boxed regions were imaged for Alexa 488 fluorescence and are shown at higher magnification in (C), the central region of the cornea, and (D), the peripheral region/edge of the cornea.

Figure 5.

 

Immunohistochemical analyses of identified proteins. (A) Cornea section from a surgical patient with CDK (62-year-old man from Saudi Arabia) and from (B) a human cadaveric eye (67-year-old Asian man) with no known eye disease. The sections were stained with DAPI, mouse monoclonal anti-annexin A2, and rabbit polyclonal GAPDH, as indicated. The secondary antibodies were coupled with Alexa 488 and Alexa 594, respectively. (C) Cornea section from the patient in (A) subjected to staining with DAPI, with mouse monoclonal MOPC-21 antibody, and without any primary antibody, as indicated.

Supplementary Materials

Supplementary Table S1 

Supplementary Figure S1 

The authors thank Gabriel S. Gaidosh for assistance with confocal microscopy and Victor Perez and Valery Shestopalov for advice and comments on the manuscript.