Proteomic Analysis of Embryonic and Young Human Vitreous

Abstract

Purpose

The proteomic profile of vitreous from second-trimester human embryos and young adults was characterized using mass spectrometry and analyzed for changes in protein levels that may relate to structural changes occurring during this time. This vitreous proteome was compared to previous reports to confirm proteins already identified and reveal novel ones.

Methods

Vitreous from 17 human embryos aged 14 to 20 weeks gestation (WG) and from a 12-, a 14-, a 15-, and a 28-year-old was individually analyzed using tandem mass spectrometry–based proteomics. Peptide spectral count associations with embryonic age were assessed using a general linear model of fold changes and Spearman's rank correlation. Differences between embryonic and young adult vitreous proteomes were also compared. Immunohistochemistry was used to evaluate three proteins in five additional fetal (10–18 WG) human eyes.

Results

There were 1217 proteins identified in fetal and young adult human vitreous, 206 after quantile normalization and variance filtering. In embryos, the peptide counts of 37 proteins changed significantly from 14 to 20 WG: 75.7% increased, 24.3% decreased. Immunohistochemistry confirmed the absence of clusterin and cadherin in 10 and 14 WG eyes and their presence at 18 WG. Comparing embryonic to young adult vitreous, 47 proteins were significantly higher or lower. A total of 768 proteins not previously identified in the literature are presented.

Conclusions

Proteins previously unreported in the human vitreous were identified. The human vitreous proteome undergoes significant changes during embryogenesis and young adulthood. A number of protein levels change considerably during the second trimester, with the majority decreasing.

Vitreous is an important ocular tissue that undergoes extensive developmental change ultimately resulting in an optically clear tissue.^1,2 During the first trimester (1–12 weeks gestation [WG]) of ocular development, a complex vascular network composed of the hyaloid artery, the vasa hyaloidia propria, and the tunica vasculosa lentis develops within the vitreous body to nourish the developing eye, particularly the anterior segment and lens.^1,3 During the second trimester (approximately 13–26 WG) this hyaloid vasculature regresses, and by birth the vitreous body is a mostly acellular, optically clear gel.^4–7

Proteomic analysis of vitreous has been previously used to investigate hyaloid regression in postnatal mice.^8,9 Proteomics have also been used to characterize vitreous in a variety of human diseases.^10–24 The Human Eye Proteome Project is a recent (2012) review and summary of the current status in proteomic studies of the eye.²⁵ This review identified 460 nonredundant human vitreous proteins in published literature. A more recent study identified 1205 proteins in normal human vitreous, 682 of which were previously unreported.²⁶ To date, proteomic studies of human embryonic vitreous have not been undertaken. The rationale to do so is that protein expression during this period of embryogenesis may be related to the significant structural changes occurring during this stage of vitreous development. The underlying hypothesis is that vitreous protein composition changes from 14 to 20 WG and differs at different fetal stages and in comparison to young adulthood.

Methods

Study Design

This cross-sectional study analyzed the protein profiles of a series of postmortem human embryonic vitreous samples aged 14 to 20 WG (N = 17) and a 12-, a 14-, a 15-, and a 28-year-old. Five additional fetal eyes (10, 14, and 18 WG) were studied by immunohistochemistry.

Vitreous Origin and Storage

Vitreous was obtained from 26 eyes [ages: 10 (n = 2), 14 (n = 3), 15, 16, 17, 18, 18.5 (n = 3), 19 (n = 3), and 20 (n = 4) WG; and 12, 14, 15, and 28 years old]. Human fetal eyes were obtained at surgery to terminate pregnancy with maternal consent and ethical approval from the Human Ethics Committees of the University of Sydney and the University of New South Wales, in accordance with the tenets of the Declaration of Helsinki. Ultrasound and postmortem measurements of foot length were used to determine the gestational age. In all subjects, the vitreous was obtained within 2 to 4 hours post mortem. Only one vitreous body of each subject was employed in this study. The fellow eye was employed for unrelated studies of retinal embryology.

Eyes were examined at the time of dissection to confirm that there was no gross pathology in the anterior and posterior segments. Unfixed eyes were dissected with an initial incision made approximately 1 to 2 mm posterior to the limbus (depending on eye size), and then carefully extended around the limbus using microscissors, so as to remove the anterior eye structures (cornea, iris, and lens). The vitreous including hyaloid vessels was removed with the lens, which was subsequently removed during viewing with a dissecting microscope. For human fetal eyes, the vitreous can be readily removed from the eye cup as a “globule” or gel body. The entire vitreous was immediately placed in sterile Eppendorf tubes, snap-frozen, and stored at −80°C. No maternal or family history was available for those subjects other than the age in WG. The young adult eyes were obtained from the New England Eye Bank.

Vitreous Prefractionation and Proteomics

Fifty microliters of undiluted vitreous was separated by one-dimensional SDS-PAGE, and proteins were visualized using Coomassie Brilliant Blue G-250 stain (Bio-Rad, Hercules, CA, USA). Each sample/gel lane was cut into 40 slices, and proteins were subjected to in-gel digestion using trypsin (Promega, Madison, WI, USA). Tryptic peptides were analyzed by nanospray liquid chromatography-tandem mass spectrometry (LC-MS/MS) using a LTQ linear ion trap mass spectrometer (Thermo Scientific, San Jose, CA, USA) as described previously.^16,27 Proteins that were identified in at least two independent vitreous samples by a minimum of two peptides matched in the same or adjacent gel slices were reported. The total peptide-spectral matches for each of these reported proteins were compiled as a semiquantitative measure of protein abundance. Proteins with peptide counts below this detection threshold were considered zero concentrations.

Database Searching

All MS/MS samples were analyzed using X! Tandem (version X! Tandem Piledriver [2015.04.01.1]; The GPM, thegpm.org, in the public domain). X! Tandem was set up to search the uni.HUMANRevConcat.2015_03.fasta.pro database (March 2015; 179,818 entries) assuming the digestion enzyme trypsin. X! Tandem was searched with a fragment ion mass tolerance of 0.50 Da and a parent ion tolerance of 1.00 Da. Glu->pyro-Glu of the n-terminus, ammonia loss of the n-terminus, gln->pyro-Glu of the n-terminus, oxidation of methionine, and propionamide of cysteine were specified in X! Tandem as variable modifications.

Criteria for Protein Identification

Scaffold (version Scaffold_4.4.1.1; Proteome Software, Inc., Portland, OR, USA) was used to validate MS/MS-based peptide and protein identifications. Peptide identifications were accepted if they could be established at greater than 89.0% probability to achieve a false discovery rate (FDR) less than 0.5% by the Scaffold Local FDR algorithm. Protein identifications were accepted if they could be established at greater than 99.0% probability to achieve an FDR less than 1.0% and contained at least two identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm.²⁸ Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. Proteins were annotated with GO terms from the National Center for Biotechnology Information (downloaded April 24, 2015).²⁹

Statistical Analyses

Spectral peptide counts were normalized using quantile normalization. A variance filtering procedure was applied such that proteins with zero values for >50% of their data within an age group were excluded from analysis. Age groups [early embryonic (14–18 WG, N = 7), late embryonic (19–20 WG, N = 7), and young adult (12, 14, 15 and 28 years old, N = 4)] were compared using a general linear model of the log values from the normalized data. Fold changes are presented. P values were adjusted for FDR using the Benjamini-Hochberg (BH) correction.³⁰ Spearman rank-order correlations were additionally conducted between the embryonic age and proteins. Analyses were conducted in Stata Version 13.1 (StataCorp LP; College Station, TX, USA).

Immunohistochemistry (IHC)

Dystroglycan and cadherin were found to undergo statistically significant change in peptide numbers from 14 to 20 WG, and clusterin similarly changed from the embryo to young adulthood. The commercial availability of human antibodies to these proteins enabled IHC evaluation in embryonic human eyes as previously described by our group.^31–33

Five formalin-fixed, paraffin-embedded whole human eyes from embryos aged 10 (n = 2), 14 (n = 2), and 18 WG were used for histologic examination of the hyaloid vasculature. Immunohistochemistry experiments were performed in duplicate in the same eye and, when tissue was available, in multiple eyes of the same gestational age. Whole eye cups were processed and embedded in paraffin and serially sectioned at the horizontal meridian from cornea to optic nerve at 5 μm. Commercially available primary antibodies (Novus Biologicals, LLC, Littleton, CO, USA) directed against dystroglycan, cadherin, and clusterin were employed. Deparaffinization and rehydration were performed using a series of baths beginning with xylene, 100%, 95%, and 70% ethanol, with a final wash in distilled water. Antigen retrieval was performed using a sodium citrate buffer solution at pH 6.0 (Dako, Carpinteria, CA, USA) and heating under pressure in a microwave oven. The slides were rinsed and washed in a buffer (0.05 M Tris/HCl, 0.15 M NaCl, 0.05% Tween 20, pH 7.6) between each step of the IHC protocol. Three percent hydrogen peroxide was used for 10 minutes to block endogenous peroxidase. Rabbit anti-human primary polyclonal antibody for each protein was diluted to the manufacturer's recommendation using an antibody diluent with background reducer (Dako) and incubated on the sections for 1 hour. Horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (Dako) was applied to the sections and allowed to incubate for one-half hour. The brown-colored enzyme product was developed using a 3,3′-diaminobenzidine (DAB) chromogen (Dako) for 5 minutes. The sections were counterstained using Mayer's hematoxylin (Dako) for 30 seconds, washed in distilled water, then dehydrated in ascending graded ethanol baths, cleared in xylene, and coverslipped using a permanent mounting media. The immunostained slides were examined on a Zeiss Axioskop bright-field microscope (Carl Zeiss, Inc., Thornwood, NY, USA). The images were acquired using a Spot II digital camera (Diagnostics Instruments, Inc., Sterling Heights, MI, USA).

Results

Mass spectrometry identified 1217 proteins (Supplementary Table S1) in at least two independent vitreous samples when all subjects (from 14 WG to 28 years of age) were considered. After quantile normalization and variance filtering, 206 proteins remained.

Human Fetal Vitreous Compared to Young Adult

Samples were divided into dichotomous age groups, embryonic (14–20 WG) and young adult (12, 14, 15, and 28 years old), and the proteomes were compared. Forty-seven proteins (23%) were differentially expressed between the embryonic and young adult vitreous. All of these identified proteins had peptide count values that were significantly lower in the embryonic vitreous, with the fold change ranging from 1.5 to 11 times lower in the embryonic vitreous; the average fold change was 3 times lower (Table 1).

Table 1.

Comparison of Embryonic Versus Young Adult Vitreous Proteins With FDR Adjusted P Values < 0.05

Fetal Vitreous Changes From 14 to 20 WG

Early (14–18 WG) and late (19–20 WG) second-trimester age groups were also compared using a general linear model of the log values from the normalized data; 37/206 (18.0%) proteins were statistically different between early versus late age groups using a BH-corrected P value (P < 0.05) and had a fold change that ranged from 0.27 to 5.35. Table 2 lists those proteins that increased and Table 3 lists those that decreased when comparing early (14–18 WG) to late (19–20 WG) second trimester.

Table 2.

Vitreous Proteins That Increased When Comparing Early (14–18 WG) Versus Late (19–20 WG) Second-Trimester Proteins [FDR Adjusted P Values < 0.05]

Table 3.

Vitreous Proteins That Decreased When Comparing Early (14–18 WG) Versus Late (19–20 WG) Second-Trimester Proteins [FDR Adjusted P Values < 0.05]

Spearman's rank correlation of the total peptide counts (all proteins) showed that there was no statistically significant (R = 0.37, P < 0.15) increase or decrease in overall proteins from 14 to 20 WG. Analysis of changes in individual proteins showed that there was a significant increase or decrease in quantile-normalized peptide spectral counts from 14 to 20 WG (BH FDR P < 0.05) in 30.1% (62/206) of proteins (Table 4). Of note, each protein found to decrease from 14 to 20 WG by Spearman's rank correlation analysis was also found to be decreased by the grouped (14–18 vs. 19–20 WG) analysis described above. The same was true for each protein that increased during 14 to 20 WG, confirming these results.

Table 4.

Spearman's Correlation of Vitreous Protein Counts With Embryonic Ages (14–20 WG, n = 17)

Of these 62, 45 (72.6%) proteins had increased expression during this period of gestation (R > 0.58, BH FDR P < 0.05). Notable in this group is cadherin (cadh2) (R = 0.745, P ≤ 0.008), which is important for cell adhesion.³⁴ Pigment epithelium-derived factor (PEDF), which has been shown to be important in antiangiogenesis,^14,35–38 was also found to increase significantly with fetal age from 14 to 20 WG (R = 0.677, P < 0.015). Cytochrome C (cytc), a known mediator of developmental apoptosis,³⁹ had increased expression (R = 0.683, P < 0.014), as did dystroglycan (dag1), the laminin binding component of the dystrophin-glycoprotein complex⁴⁰ (R = 0.753, P < 0.007).

Of the 62 proteins with levels that changed significantly from 14 to 20 WG, 17 (27.4%) had decreased expression (R < −0.580, BH FDR P < 0.05). Notable proteins that decreased were cofilin-1 (cof1; R = −0.638, BH FDR P < 0.026), which has been shown to promote/mediate angiogenesis,⁴¹ peroxiredoxin (prdx5; R = −0.799, P < 0.005), which may play an antioxidant protective role,⁴² and glycolytic enzyme enolase⁴³ (enoa; R = −0.688, P < 0.013).

Immunohistochemistry

Immunohistochemistry studies confirmed the presence of dystroglycan, clusterin, and cadherin, all shown to have a significant difference in expression in human embryonic vitreous. Positive staining was localized in hyalocytes and hyaloid vessels. Clusterin and cadherin were not detected in 14 WG vitreous cells or structures (Figs. 1A, 2A) but did appear in the hyaloid vasculature by 18 WG (Figs. 1B, 2B). This is consistent with the increase in protein concentrations detected in the proteomic analyses (see above). Dystroglycan was detected in the hyaloid vessels of all five embryonic eyes (10, 14, 18 WG; Fig. 3). Immunohistochemical detection of these proteins in hyalocytes and in hyaloid vessels of human embryos suggests that at least these particular proteins that were detected by proteomics originated from vitreous or structures within the vitreous body. This same approach of confirming proteomics with IHC has been previously been employed by others to study hyaloid vessel regression in the mouse.⁸

Figure 1.

Immunohistochemistry in a 14 WG human vitreous demonstrates that clusterin was not detected in the hyaloid vessels (A) and cadherin was not detected in the endothelial cells of the hyaloid vessels (B). Scale bar: 10 μm.

Figure 2.

Immunohistochemistry in an 18 WG human vitreous demonstrates that positive staining for clusterin in the hyaloid vessels (A) and cadherin was detected in the endothelial cells of the hyaloid artery (B). This immunostaining pattern confirmed the results of proteomics analysis, which showed that clusterin and cadherin increased from 14 to 20 WG. Scale bar: 10 μm.

Figure 3.

Immunohistochemistry for dystroglycan demonstrates positive staining localized to the cell junctions of the hyaloid vascular endothelium in a 14 WG (A) and an 18 WG (B) human embryo. Scale bar: 10 μm.

Discussion

Proteomic investigation of vitreous composition in 14 to 20 WG human embryos and young adult human eyes revealed a complex tissue with 1217 detectable proteins, of which 206 were quantile normalized and variance filtered. While many of these are likely native to vitreous, especially in the embryo, many may also arise from adjacent tissues such as the vascular and neuronal elements of the retina and ciliary body. A review of previously reported human vitreous proteins identified 545 nonredundant proteins.²⁵ A recent high-resolution Fourier transform study of the normal human vitreous proteome identified 1205 proteins, 682 (57%) of which had not been previously described in vitreous.²⁶ The study reported herein identified 1217 vitreous proteins, 744 (61%) of which had not been previously reported in either of the aforementioned studies (Fig. 4).

Figure 4.

Venn diagram comparing the proteomes of three human vitreous studies (present study versus Murthy et al.²⁶ versus Semba et al.²⁵).

The protein composition of human embryonic vitreous differs from young adult vitreous and changes significantly from 14 to 20 WG, during the second trimester of gestation. This occurs concurrently with regression of embryonic vitreous vessels as well as other developmental changes in the eye. Of the 206 quantile-normalized/variance-filtered proteins identified in human fetal vitreous, 37 (18%) showed a significant change in expression from 14 to 20 WG (early to late second trimester). Some of these, such as PEDF and cofilin-1, are known to be important in angiogenesis and may be critically involved in this and one or more other processes of vitreous or ocular embryogenesis. Thus, it is intriguing to postulate that some of the changes detected in the proteomic profile of human fetal vitreous during the second trimester are related to the regression of the embryonic vitreous vasculature, although other developmental events are also occurring during this period. In this context, a better understanding of hyaloid vasculature regression may prove useful in developing new therapies for vasoproliferative eye diseases, metastatic carcinomas, and others. To this end, the proteins identified in this study, especially those shown to change in concentration and those corroborated in previous studies, may be worth exploring as potential avenues for new therapeutic approaches to both stimulate endogenous inhibitory pathways and suppress stimulatory pathways in the management of neovascular pathologies.

A similar proteomic study was undertaken in the developing mouse vitreous using two-dimensional gel electrophoresis (2DE), a distinctly different methodology than that employed in the present study.⁸ Furthermore, the mouse specimens consisted of lens with pupillary membrane, tunica vasculosa lentis, and vasa hyaloidea propria, while the specimens in the present study were purely vitreous. This may explain why only a small number of proteins were found in both mouse and human proteomes. This and contradictory findings regarding changes from early to late fetal stages are possibly due to different methodologies as well as species differences.

Future studies should explore the fetal human proteome in search of “missing proteins”⁴⁴ with the findings validated by multiple reaction monitoring (MRM) methodologies, as has been previously described.⁴⁵