Protein expression in human trabecular meshwork: downregulation of RhoGDI by dexamethasone in vitro

Tissue procurement

Human eyes from 30 normal donors (aged 20–35 years old) were obtained through the Guangdong Province Eye Bank, China and used in this study. Within 12 h after death, TMT was dissected in a ring shape. The possibility of contamination by small amounts of sclera was noted, in accordance with the report by Bhattacharya et al. [31,34]. Harvested TM samples were collected in tubes and stored at −80 °C until use.

Cell culture

An immortal human trabecular meshwork cell line [35] was cultured in Dulbecco’s modified Eagle’s medium and F12 (DMEM/F12) containing 15% newborn calf serum. TMCs were treated with different levels of DEX (10⁻⁵, 10⁻⁶, and 10⁻⁷ M) added to the medium for one- to three-day durations. Since we hoped to find the early expressed proteins responding to DEX treatment, we used a short-term (i.e. within 3 days) cell culture, as previously reported [36,37]. Meanwhile, besides the 10⁻⁷ M DEX used in most studies [9,37,38], we also used 10⁻⁶ and 10⁻⁵ M DEX to see the more obvious effects on TMCs. At various time points, cell cultures were terminated for further analyses. The cellular changes were measured by phase-contrast microscopy and 3-(4,5-dimethylthiazol-2-yl)-2,5- diphenyltetrazolium bromide (MTT) assay (described below). Soluble proteins extracted from the TMCs were used for subsequent 2-DE and western blot analyses.

Cell viability assays

The proliferation rate of TMCs was measured using an MTT assay [39]. Each group of TMCs was trypsinized and seeded in 96-well plates with an initial density of 5×10³ cells per well. Subsequently, cells were cultured in medium with various concentrations of DEX for 1, 2, or 3 d. Upon color development, the absorbance of each sample was measured using a multiwell scanning spectrophotometer. The absorbance of each experiment was analyzed in five wells, and three independent experiments were performed. Statistic analysis (Student’s t-test) was conducted to compare the differences between DEX-treated and -untreated cells at each time point.

Immunohistochemical assay

TMCs were cultured on a glass coverslip in 12-well chamber dishes, with or without DEX treatment, as described above. Slide-mounted cells were used for immunostaining of RhoGDI by means of an HRP-DAB Cell and Tissue Staining Kit (Maixin-Bio, Fuzhou, China) according to the manufacturer’s instructions. Briefly, TMCs were fixed in ice-cold acetone for 20 min and cell membranes were permeabilized with 0.1% Triton X-100, followed by quenching of the endogenous peroxidase activity with 3% H₂O₂ and blocking non-specific binding with goat serum albumin. Then, cells were incubated with antibodies against RhoGDI (#2564; Cell Signaling Technology, Danvers, MA) at 4 °C, overnight. Cells were then incubated with secondary goat antibodies, and color development was done with DAB reagents. A negative control was run, in which the cells were incubated in PBS without the addition of primary antibodies.

Protein extraction

Pooled TMT was thoroughly cut into small pieces with a scissor. Then an aliquots-of-lysis buffer (8 mol/l urea, 4% CHAPS, 2% Pharmalyte 3–10; GE Healthcare, Ltd., NJ) was added to the samples, followed by homogenization and sonication on ice. The soluble protein was obtained by centrifugation (at 12,000× g for 30 min at 4 °C). For cultured cells, collection was performed by rinsing with PBS, removal from the flask with a rubber policeman, and pelleting at 500× g centrifugation. After adding an aliquot lysis buffer, cell pellets were sonicated. Insoluble materials were removed by centrifugation (at 12,000× g for 30 min 4 °C). The soluble protein portion was quantified by BCA assay using a protein assay kit (Bio-Rad, Hercules, CA) and stored at −80 °C until the next analyses.

Two-dimensional gel electrophoresis

Two-dimensional gel electrophoresis was performed using instruments, kits, and the 2-DE protocol from GE Healthcare (GE Healthcare Bio-Sciences AB, Uppsala, Sweden), and according to our previously reported protocols [40], unless otherwise specified. Briefly, the extracted protein solution was precipitated with a 2-D clean-up kit. The pellets were dissolved in rehydration solution to a certain volume, according to the size of the immobilized pH gradient (IPG) strips, and centrifuged at 15,000× g for 10 min to remove insoluble materials. The supernatant was used as the loading sample on the IPG strips.

First-dimensional isoelectric focusing (IEF) was performed using the Ettan IPGphor II unit. IPG strips (18 cm, pH 3–10, linear gradient) were loaded with protein samples and rehydrated overnight. Isoelectric focusing was run at 15 °C, totally focused over 8 kVh. First dimension strips were subjected to the standard reduction and alkylation steps before second-dimension electrophoresis, as follows. The strips were soaked for 15 min in 50 mM Tris-HCl solution (pH 8.8) containing 6 M urea, 30% glycerol, 2% SDS, 2% DTT, and 0.001% bromophenol (w/v). They were then soaked for an additional 15 min in the same solution, except that 2.5% iodoacetamide was substituted for the 2% DTT.

The second-dimensional electrophoresis (SDS–PAGE) was conducted on 10% polyacrylamide gels (PROTEAN^® II xi Cell; Bio-Rad) with protein molecular weight markers (Takara Biotech Co. Ltd., Dalian, China). At the end of electrophoresis, gels were stained with hot Coomassie brilliant blue R-350 [41].

After destaining, gels were scanned using an ImageScanner II and analyzed by Melanie software (GE Healthcare). Protein spots were first outlined automatically, and many were further outlined manually, and quantified. The measured molecular weight (Mr) and isoelectric point (pI) values of each protein spot were calculated according to the method introduced by Westermeier and Naven [42].

Mass spectral analysis and protein identification

MS analysis and protein identification were performed as reported previously [40,43]. Following imaging analysis, protein spots on the large gel run with TMT protein were selected for identification. Spots with relatively higher intensities on 2-DE gels were excised, dried, digested with trypsin (sequencing grade; Promega, WI), and mechanically spotted onto MS plates by an Ettan Spot Handling Workstation (GE Healthcare). The molecular masses of peptides were measured by MALDI-TOF-MS using an Ettan MALDI-TOF Pro mass spectrometer (GE Healthcare). On the basis of peptide-mass matching, peptide peaks obtained from mass mapping were identified using the NCBInr database. All spots were examined for methionine oxidation, for iodoacetamide modification, and, with no-limitation, for pI. The confidence of identification was indicated by the number of matching peptides, and the percent of the amino acid-sequence coverage. Proteins were accepted on the basis of four or more peptides having matched [43].

Cluster analysis of trabecular meshwork proteins

The categories of identified proteins were analyzed by inputting their accession codes into a bioinformatic databse, which automatically classifies genes based on gene ontology (GO) terms. Before GO-based binning, any identified protein with a GI code was recoded according to gene/RefSeq accession numbers, using the websites UniProt, NCBI, and BLAST.

Western blotting analysis

Western blotting was conducted according to specifications of the antibody manufacturer and our previous report [40]. Briefly, aliquots of lysates (30 μg of protein equivalent) were subjected to 10% SDS–PAGE under reducing conditions, then blotted onto a PVDF membrane. Antibodies against RhoGDI (#2564) and RhoA (#2117), and a Phototope-HRP western blot kit (all from Cell Signaling Technology) were used for western blotting assay. Densitometric analysis was performed to measure the intensity of the RhoGDI and RhoA bands with the use of the molecular imaging software in the Kodak Image Station 4000MM (Kodak, Rochester, NY). Each experiment was repeated three times.

Protein expression pattern on two-dimensional electrophoresis gels

To date, 2-DE protein-expression, combined gel-spot identification for human TMT has not yet been conducted because of difficulties in obtaining enough proteins from limited tissue samples, although the proteome of TMCs has been extensively studied [29,33,44]. In this study, the 2-DE gel pattern of human TM tissue and a human TMC line was obtained. We first compared the 2-DE protein patterns of human normal TMT to that of TMCs. Figure 1 shows that the protein spot distribution of TMCs is somewhat different from that of the TMT on the 2-DE gel. However, the average number of protein spots on TMT gels (877) was similar to that on TMC gels (963). In the current study, we used 2-DE (18 cm×20 cm, gel size) to separate the soluble protein of human TMT, and only observed 877 protein spots on the gels stained with Coomassie brilliant blue R-350. It should be kept in mind that the limitations of 2-DE proteomics, that is, missing high molecular-weight proteins, proteins outside the pI range of the isoelectric focusing gel, many insoluble proteins, and low-abundance proteins, account for many of the TM proteins not found on the gels.

The distribution of the protein spots detected in the gel was determined by means of their pI and molecular mass (Appendix 1). According to the data measured on each gel, more than 67% of the TMT protein spots had a pI between 5 and 8, and approximately 89% of the identified spots had masses between 20 kDa and 100 kDa. Among TMC protein spots, 74% had a pI between 5 and 8, and 91% had masses between 20 and 100 kDa. In Appendix 1, there were discrepancies between the calculated pI and M_r values, compared to the measured pI and M_r values for many of the identified spots. This is because we obtained the measured pI by pH gradient graphs and the measured M_r values by co-running molecular weight standard proteins [42]. The data accuracy was limited by reference markers and by the accuracy of measurement on the gels. The theoretical pI and M_r values were calculated from the amino acid sequence of a protein in the database, with fewer effects concerning the post-translation modification and the molecular stereostructure. Therefore, we could see that the many identified proteins in Appendix 1 had different measured and theoretical pI or M_r. However, the mapping data we present here could prove to be a useful reference for future studies on differentiating expressed proteins in the TM from various glaucoma disorders.

Few early works used 2-DE to display the protein expression of human TM. Steely et al. [45] reported a human TM protein pattern obtained from 2-DE gels (16×21 cm) using radio-labeled samples visualized by X-ray film exposure. More than 400 spots were observed by this technique, but no further identification was pursued. In an earlier report, Russell and colleagues [34] compared the age-related differences of human TM protein maps on 2-DE gels. Their studies obtained approximately 160 spots from TM proteins of donors aged 9 to 91 years old. Clearly, our technique resulted in more protein spots being detected upon 2-DE gel separation. Additionally, 2-DE has been previously used to display the protein expression of TMCs with or without DEX treatment [33,46]. In the present study, we compared the 2-DE pattern of human TMT and a TMC line. We reported the pattern for 877 proteins from human normal TMT, 235 of which were identified and localized on a 2-DE gel. Since Steely et al. [33] have displayed a 2-DE protein pattern for the GTM3 cell line, and effectively identified 87 primary proteins and 93 variants of these proteins, we did not pursue detailed protein identification for those spots detected from our TMC analysis.

Proteins identified from human trabecular meshwork

The protein spots separated on the gel (Figure 2) were identified. A total of 444 visible spots from the 2-DE gel were mechanically treated by the spot-handling workstation; 235 of these were subsequently identified by MALDI-TOF-MS. Among the 235 identified spots, 72 were determined to be identical by protein accession number. Therefore, Appendix 1 is only composed of 163 proteins with varying parameters, including protein name and accession number, theoretical and measured M_r and pI values, matched the peptides and amino acid sequence coverage. The spot number marked on the gel (Figure 2) for localizing a particular protein has also been indicated in Appendix 1. It was surprising to find that about one-third of the protein spots (72/235) belonged to isomers or were products of post-translational modifications. The spots representing variants of the same protein on the 2-DE of TM has previously been reported for the GTM3 cell proteome [33]. These dynamic proteins may serve as focused targets for future studies, to determine their functional significance in TM and glaucoma.

Several of the proteins we identified in TM have been reported to be associated with glaucoma. For example, myocilin was initially identified as a 55 kDa protein secreted into the media of TM cultures and induced by glucocorticoid [12,47,48]. Myocilin affects the flexibility and resilience of TM, which, if disrupted, may lead to alteration(s) in the aqueous outflow [38]. In the current study of normal human TM, myocilin was identified in two individual spots, with different molecular weights on 2-DE (spots 147 and 154 shared the same protein code, Figure 2, Appendix 1), which suggests the possibility of a post-translational modification. In addition, Steely et al. [33] suggested that calreticulin, located on the acidic side at 62 kDa (spot 26), might be modified by glycosylation, since its expected mass is 48.1 kDa. Glycosylation may influence both the calcium-signaling role and chaperonin activity of calreticulin in TMCs. Other studies have demonstrated that a mutation of cytochrome P450 (spot 54) could result in primary congenital glaucoma in some patients [17,49]. Expression of the angiopoietin-like factor CTD6, one of the 15 most abundant types of cDNA in TM [50], was believed to stimulate the deposition of specific ECM in the cornea [51]. In our present study, there were at least four spots (spot 63, 170, 181, and 182) identified as angiopoietins, which suggested this protein may play a key role in normal TM, and possibly in glaucoma. Meanwhile, several identified proteins in this study belonged to cytoskeletal protein families or extracellular matrix proteins, such as vimentin, lamin, actin, and annexin. These proteins were also the major components involved in TMCs in response to DEX or transforming growth factor β (TGF-β) treatment [7-19,29], and regulated by Rho GTPase [4]. The above-mentioned proteins are expressed in normal TMT, and they have also been reported to be changed in studies on TMCs in vitro or in vivo. We anticipate a similar scenario for other yet-uncharacterized proteins that merit future study. For example, RhoGDI (spot 225) was downregulated in our DEX-treated TMCs, which will be discussed in later sections of this paper.

The identified proteins were clustered according to gene ontology (Table 1). Based on GO biological terms, the major categories of identified proteins were metabolic process, cell adhesion, anti-apoptosis, cell motility, carbohydrate metabolic process, signal transduction, and regulation of transcription. In Table 1, we also presented the categories of identified proteins from human TMT from Bhattacharya et al. [31] and from TMCs from Steely et al. [33]. Bhattacharya et al. [31] identified 316 proteins from human TMT by SDS–PAGE and liquid chromatography (LC) tandem MS. Steely et al. [33] revealed the 2-DE based proteome of TMCs and identified fewer spots compared to our study. Generally, the major categories of identified proteins in our study were very similar to those of Bhattacharya et al. [31]. However, several kinds of proteins were absent or less abundant in the Bhattacharya et al. [31] TMT proteome. It is possible that this discrepancy is due to the the two studies’ different protein-separation techniques. We took the visible spots from 2-DE gels for the subsequent tryptic digestion and MALDI-TOF-MS analysis. In Bhattacharya et al., SDS–PAGE combined LC/MS/MS might have been able to identify some proteins present only in trace amounts.

Dexmethasone effects on trabecular meshwork cells in vitro

We next used DEX-treated TMCs, an approach widely used to study the mechanism of glaucoma, to search for differential expression patterns of proteins. It has been reported that treating cultured TMCs for weeks with DEX causes a significant increase in average TMC size and significantly increased nuclear size [8,9]. In the present study, we could not detect significant differences between normal and DEX-treated TMCs using phase-contrast microscopy throughout the three-day study period. However, we did observe that the proliferation rate of TMCs was inhibited by 10⁻⁵–10⁻⁷ M DEX, as evidenced by MTT analysis (Figure 3). The inhibition appeared after 48 h and 72 h of DEX treatment. Generally, 10⁻⁷ M DEX has been widely used in previous studies [9,13,37]. Since short-term cultured TMCs were used in the current study, we used three concentrations, including two higher concentrations (10⁻⁶ and 10⁻⁵ M) of DEX, to see the more obvious effects on the cell proliferation and protein expression of TMCs. A stronger inhibition of TMC proliferation occurred at the highest DEX level. The effect of DEX occurred with the highest inhibition of cell proliferation at 10⁻⁵ M at 72 h of treatment.

RhoGDI downregulated in dexamethasone-treated trabecular meshwork cells

We used proteomic analysis to characterize the differentially expressed proteins in cultured TMCs. We found that RhoGDI was the changed protein spot, as evidenced by immunoassays (Figure 4), which revealed that, in DEX-treated TMCs, RhoGDI expression decreased in a dose-dependent manner. By comparing the 2-DE gel patterns of the lysates from TMC lines treated with or without DEX, the RhoGDI spot exhibited lower-intensity staining (Figure 4A). By western blot analysis, the expression of the RhoGDI molecule appeared to be reduced in 24 h DEX-treated cells (Figure 4B). Higher levels of DEX displayed a general pattern of more obvious decreases, as compared to 10⁻⁷ M of DEX. The decrease of RhoGDI expression was also evidenced by immunocytochemical staining (Figure 4C). 10⁻⁶ M DEX reduced the RhoGDI staining in TM cytoplasm, compared to that in mock and control cells. Furthermore, RhoA, a target of RhoGDI, is known to function in TM as a cytoskeleton and ECM regulator. Therefore, RhoA inhibition may be the possible mechanism leading to TMC or tissue damage [4,23,52]. In the present study, we assayed the RhoA expression in our cultured cells. The results revealed that the expression of RhoA was opposite to that detected for RhoGDI by western blot assay (Figure 4D-F). It was also a dose dependent effect, with higher expression corresponding to higher DEX levels. There were a variety of reports that showed that RhoA increased the levels of promoted production of fibronectin, tenascin C, laminin, alpha-smooth muscle actin, collagen IV, laminin, and phosphorylated myosin light chains, as well as increased the expression of genes encoding ECM proteins, cytokines, integrins, cytoskeletal proteins, and signaling proteins in TMCs. The protein kinase inhibitors, including Rho kinase inhibitors, could alter TMC shape and adherence, disrupt the actin bundles, and increase aqueous outflow, thereby lowering IOP [4,53-56]. Meanwhile, RhoGDI forms a complex with the GDP-bound form of the Rho family proteins (e.g., RhoA), to regulate the activation state of RhoGTPase [26]. Thus, RhoGDI, by repressing RhoA activity, may regulate organization of the actin cytoskeleton and cell–cell adhesions in TMCs. In the present study, our results revealed that opposite effects of DEX treatment on the expression of RhoGDI and RhoA in TMCs, suggesting that DEX increases the RhoA activity, at least in part, by antagonizing the RhoGDI function. However, the detailed relationships and interactions of RhoGDI-related proteins, in glaucomatous TM in particular, remain unclear.

In summary, the current study provides 2-DE gel patterns of human TMT and a TMC line. We detected more than 850 spots for both lysates, and 235 spots were successfully identified using 2-DE gel analysis of human normal TM. These data will establish the foundation for future studies into the comprehensive protein expression patterns and effects in normal and glaucomatous TM. Furthermore, by using proteomic methods combined with immunoassays, we have provided evidence of decreasing RhoGDI expression in DEX-treated TMCs. There were distinct effects of DEX on the expression of RhoGDI and RhoA in TMCs. These results may prove useful to begin a focused study of glaucoma pathogenesis.