Expression of caveolin in trabecular meshwork cells and its possible implication in pathogenesis of primary open angle glaucoma

Abstract

Purpose: Primary open-angle glaucoma (POAG), which is the most common form of glaucoma, has been associated with a heterogeneous genetic component. A genome-wide association study has identified a common sequence variant at 7q31 (rs4236601 [A]) near the caveolin genes in patients with POAG. Caveolins are a family of integral membrane proteins which participate in many cellular processes, including vesicular transport, cholesterol homeostasis, signal transduction, cell adhesion and migration. The goal of this study was to investigate the expression and regulation of caveolin 1 (CAV-1) and caveolin 2 (CAV-2) in normal and glaucoma trabecular meshwork (TM) cells.

Methods: CAV-1 and CAV-2 protein expression was quantified by immunoblot analysis using lysates isolated from primary and immortalized TM cells or TM tissue dissected from normal and POAG eyes. The localization of caveolins in TM cells was assessed by immunofluorescent microscopy. CAV-1 and CAV-2 protein expression was also investigated in TM cells at various time points after subjecting the cells to known glaucomatous insults like dexamethasone (DEX) and tumor growth factor beta2 (TGF-β2) treatment. Phosphorylation of CAV-1 at tyrosine 14 in normal and glaucoma TM cell lines was evaluated using a specific monoclonal antibody (Ab). The 5' upstream region of the CAV-1 gene was amplified and the sequence variant rs4236601 (A/G polymorphic site) and several putative transcription factor-binding sites were modified by in vitro mutagenesis. The effect of nucleotide sequence modifications in the CAV-1 upstream region on gene expression was assayed in a luciferase-based system in TM and non-TM cells.

Results: CAV-1 and CAV-2 are expressed in TM cells, with localization to the cytoplasm and perinuclear region. DEX increased CAV-1 expression in immortalized glaucoma TM cells by 2.8±0.1 (n=3) fold at 24 h and 2.5±0.1 (n=3) fold at 48 h, compared to 1.3±0.06 (n=3) fold at 24 and 48 h in immortalized normal TM cells. Phosphorylation of CAV-1 at Tyr14 was reduced by 3.2±0.15 (n=3) fold in glaucomatous TM cells when compared to normal TM cells. In POAG and normal TM tissue, CAV-1 expression was found to be uniform. CAV-2, on the other hand, was variable in independent normal and glaucoma TM tissue. Substitution of a G for an A at base pair -2,388 upstream of the start codon of CAV-1, corresponding to the minor allele rs4236601 [A], increased transcriptional activity in TM and non-TM cells when compared to the native sequence. Deletion analysis of putative transcription factor binding sites in the CAV-1 promoter region caused cell-specific effects on gene expression.

Conclusions: CAV-1 and CAV-2 are expressed in normal and glaucoma tissue and TM cell lines. Phosphorylation of Tyr14 in CAV-1 and transcriptional regulation of CAV-1 expression may have a role in glaucomatous alterations in TM cells.

Figure 1

Organization of the human CAV-1 and CAV-2 genes and localization of TF binding sites. A: Exon-intron structure of the caveolins genes and localization of SNP rs4236601 marker (modified from [19]). The sizes of exons (white rectangular boxes) and the distance between them (below Y) are shown. Marker rs4236601 is located 2,388 bp upstream from the start codon of CAV-1. B: localization of SNP and putative TF binding sites (ovals) in the upstream region of CAV-1. Other details of TF-binding sites 1–8 are present in Table 2.The sequence of primers used for the modifications of the fragment are presented in Table 1.

Figure 2

CAV-1 and CAV-2 in TM of POAG patients and controls. Total protein (20 μg) from TM cells were resolved in 12% SDS-polyacrylamide gels, transferred on PVDF membrane and probed with Abs to CAV-1 (A) and CAV-2 (B). The amount of CAV-1 was similar both in POAG patients and controls, whereas the CAV-2 was highly variable among samples. Heterooligomers CAV-1 – CAV-2 are shown by arrowheads.

Figure 3

Immunofluorescence localization of CAV-1 and CAV-2 in TM cells. CAV-1 (A, D, green). CAV-2 (B, E, red). NTM cells (A-C). GTM cells (D-F). Polyclonal rabbit anti-CAV-1 and monoclonal mouse anti-CAV-2 were used for immunostaining. C and F: merged; blue – DAPI staining. CAV-1 and CAV-2 immunoreactivity is identified as punctate staining with a predominant cytoplasmic localization (arrows). In NTM cells CAV-1 was mainly found in the cytoplasm (A, arrows), while CAV-2 was present in cytoplasm, the perinuclear area as well as in the cell membranes (B, arrowheads). In GTM cells CAV-1 and CAV-2 are colocalized in dot-like structures (arrows, D-F).

Figure 4

CAV-1 expression and secretion from NTM and GTM cells. A: DEX differentially induces CAV-1 expression in NTM-5 and GTM-3 cells after 24 and 48 h. Western blot probed with CAV-1 Ab. Below – the same blot was reprobed for actin. B: The first four lanes (1–4) for NTM samples shown in A were scanned using KODAK MI Software. C: The next four lanes (5–8) for GTM samples were scanned. The values indicated represent the means±SEM from four independent experiments. D and E: CAV-1 in cell extracts (CE; lanes 1–3 and 7–9) and conditioned media (CM; lanes 4–6 and 10–12) of NTM-5 and GTM-3 cells. Lanes 1 and 7 – control cells; lanes 2 and 8 – cells were incubated with 100 μM of DEX for 96 h; lanes 3 and 9 – cells incubated with 5 ng/ml of TGFβ2 for 48 h. TGFβ2 reduced CAV-1 expression to 85% in NTM-5 cells (lane 3) and to 68% in GTM-3 cells (lane 9), whereas DEX did not affect CAV-1 expression after 96 h. Low level of CAV-1 secretion was observed from NTM cells (D, lanes 4–6) which was increased by DEX (lane 5). No expression was observed from GTM cells (E, lanes10–12). F: Bands for NTM samples shown in D were scanned. G: Bands for GTM samples shown in E were scanned. The values indicated represent the means±SEM from four independent experiments.

Figure 5

Phosphorylation of Tyr14 in CAV-1 in NTM-5 and GTM-3 cells. A: NTM and GTM cell lysates were immunoprecipitated by polyclonal antirabbit Ab against CAV-1 and the precipitate was suspended in a 2× loading buffer (ABCam Immunoprecipitation Protocol) and western blotted with monoclonal antimouse CAV-1 Ab. B: The same blot was stripped and reprobed with monoclonal antimouse pTyr14-CAV-1 antibody. C: Lanes shown in A and B were scanned using KODAK MI Software. The amount of non-phosphorylated CAV-1 is similar in NTM-5 and GTM-3 samples (A), while the amount of CAV-1 phosphorylated at Tyr14 (P-Tyr14-CAV-1) is 2.3 times lower in GTM-3 cells. Arrowhead indicates heavy chain immunoglobulins. The values indicated represent mean±SEM from four independent experiments.

Figure 6

Effect of CAV-1 promoter modification on the expression level in LUC system. Cells were transfected with 1 µg pGL3 promoterless LUC reporter plasmids inserted with intact or modified 2,465 bases CAV-1 promoter region (pGL3-CAV-1 promoterA or pGL3-CAV-1promoterG). LUC activity was averaged. The SEM is indicated by the error bars. One – ratio of activity of construct with A to activity of construct with G in position −2,388 of CAV-1 promoter; 2 – 8 putative cis-elements were deleted from CAV-1 promoter and the activity of the construct with deleted sequence was expressed in percentage to the unmodified construct. The following putative TF-binding sites were deleted: 2– NFκB (position −1537 −1527); 3 - HNF-3/Fkh homolog (position - 889 −882); 4 - POU-IV position (−807 −800); 5 – Sp-1 (position −236 −226); 6 – Ets (−182–174); 7 - SP-1 (position −124–115); 8– NFkB (position −40 −28). Other details are presented in Table 2. A: The effect of CAV-1 promoter modification on LUC activity in HEK-293, SH-SY5Y and Y79 cells. B: The effect of CAV-1 promoter modification on LUC activity in NTM-5 and GTM-3 cells.