Oxidative stress-mediated TXNIP loss causes RPE dysfunction

Abstract

The disruption of the retinal pigment epithelium (RPE), for example, through oxidative damage, is a common factor underlying age-related macular degeneration (AMD). Aberrant autophagy also contributes to AMD pathology, as autophagy maintains RPE homeostasis to ensure blood-retinal barrier (BRB) integrity and protect photoreceptors. Thioredoxin-interacting protein (TXNIP) promotes cellular oxidative stress by inhibiting thioredoxin reducing capacity and is in turn inversely regulated by reactive oxygen species levels; however, its role in oxidative stress-induced RPE cell dysfunction and the mechanistic link between TXNIP and autophagy are largely unknown. Here, we observed that TXNIP expression was rapidly downregulated in RPE cells under oxidative stress and that RPE cell proliferation was decreased. TXNIP knockdown demonstrated that the suppression of proliferation resulted from TXNIP depletion-induced autophagic flux, causing increased p53 activation via nuclear localization, which in turn enhanced AMPK phosphorylation and activation. Moreover, TXNIP downregulation further negatively impacted BRB integrity by disrupting RPE cell tight junctions and enhancing cell motility by phosphorylating, and thereby activating, Src kinase. Finally, we also revealed that TXNIP knockdown upregulated HIF-1α, leading to the enhanced secretion of VEGF from RPE cells and the stimulation of angiogenesis in cocultured human retinal microvascular endothelial cells. This suggests that the exposure of RPE cells to sustained oxidative stress may promote choroidal neovascularization, another AMD pathology. Together, these findings reveal three distinct mechanisms by which TXNIP downregulation disrupts RPE cell function and thereby exacerbates AMD pathogenesis. Accordingly, reinforcing or restoring BRB integrity by targeting TXNIP may serve as an effective therapeutic strategy for preventing or attenuating photoreceptor damage in AMD.

Fig. 1. Oxidative stress significantly reduces TXNIP expression and the suppression of RPE cell proliferation.

a ARPE-19 cells were treated with the indicated concentration of H₂O₂ for 4 h. The cells were lysed and subjected to western blotting for the indicated antibodies. b ARPE-19 cells were stimulated with 0.25 mM H₂O₂ for the indicated times. c MTT assay of the viability of ARPE-19 cells after treatment with H₂O₂ compared with that of ARPE-19 cells not treated with H₂O₂. The results shown are representative of three independent experiments. d ARPE-19 cells were treated with the indicated concentration of H₂O₂ for 4 h. The cells were lysed and subjected to western blotting for the indicated antibodies. e ARPE-19 cells were treated with 0.25 mM H₂O₂ for the indicated time periods. Cell proliferation was assessed by the Wst-1 assay. The results shown are representative of three independent experiments. f Immunoblot analysis of cell cycle-related markers in shTXNIP cells. β-Actin was used as an internal loading control.*p < 0.05; **p < 0.01. The error bars indicate the SEM

Fig. 2. Inhibition of autophagy attenuates the TXNIP loss-induced suppression of RPE cell proliferation.

a Immunoblot analysis of microtubule-associated protein 1 light chain 3 (LC3)-I to LC3-II conversion in ARPE-19 cells treated with the indicated concentration of H₂O₂ for 4 h. The cells were lysed and subjected to western blotting for the indicated antibodies. b ARPE-19 cells were stimulated with 0.25 mM H₂O₂ for the indicated times. c TXNIP-depleted cells displayed notably increased LC3-II accumulation (left panel). RPE cells were plated on coverslips coated with 10 μg/ml fibronectin. Immunostaining was performed on fixed cells with anti-α-tubulin and anti-LC3 antibodies (right panel). LC3-positive puncta (%, the number of LC3 positive cells/the number of total cells) in the images were counted. *p < 0.01 versus shCtrl. d The effect of a lysosomal inhibitor, Bafilomycin A1 (Baf A1), on autophagy induction by TXNIP loss in shTXNIP cells. RPE cells were treated with 10 nM Baf A1 for 30 min. The cell lysates were collected and analyzed by western blotting using an anti-LC3 antibody. e Representative images of mRFP-GFP-LC3 puncta. The colocalization of GFP and red fluorescent protein (RFP), as indicated by yellow dots in the overlapped GFP and RFP images, was visible in autophagosomes, whereas only RFP fluorescence, indicated by red puncta, was observed in autolysosomes. Autophagosome-positive and autolysosome-positive mRFP-GFP-LC3-transfected cells were counted. f The depletion of LC3 attenuated the TXNIP loss-mediated suppression of RPE cell proliferation. shTXNIP cells were transfected with siCtrl or siLC3B and cultured for the indicated time periods. Cell proliferation was performed by the Wst-1 assay. *p < 0.01. The error bars indicate the SEM. Scale bars, 50 μm

Fig. 3. TXNIP loss-mediated p53 activation regulates autophagy.

a LC3 and p53 expression levels were significantly increased in shTXNIP cells. Cells were lysed and subjected to western blotting using the indicated antibodies. b RPE cells were plated on coverslips coated with 10 μg/ml fibronectin. Immunostaining was performed on fixed cells with an anti-p53 antibody. Scale bars, 50 μm. c shTXNIP cells were transfected with eGFP or eGFP-TXNIP constructs. The cells were lysed and blotted with the indicated antibodies. d The phosphorylation of AMPK at Thr¹⁷² and ULK1 at Ser⁵⁵⁵ was increased in shTXNIP cells, but the phosphorylation of mTOR at Ser²⁴⁴⁸ and ULK1 at Ser⁷⁵⁷ was decreased. e RPE cells were transfected with 10 μM p53 siRNA for 48 h. Cells were lysed and blotted with the indicated antibodies. f RPE cells were transfected with siCtrl or siLC3B and cultured for the indicated time periods. Cell proliferation was analyzed by the Wst-1 assay. g Cell lysates were immunoprecipitated with an anti-MDM2 antibody, blotted, and probed with anti-p53 and anti-TXNIP antibodies. h, i The activation of p53 at Ser¹⁵ was notably increased in shTXNIP cells. Cells were lysed and subjected to western blotting using the indicated antibodies (h). Immunostaining was performed on fixed cells with anti-p53 and phospho-p53 Ser¹⁵ antibodies. Scale bars, 30 μm (i). ^#p < 0.01. The error bars indicate the SEM

Fig. 4. TXNIP is involved in BRB integrity through the regulation of Src kinase activation.

a, b RPE cells were plated on coverslips coated with 10 μg/ml fibronectin. Immunostaining was performed on fixed cells with an anti-ZO1 antibody (a) and TRITC-labeled phalloidin (b). The fluorescence intensity of ZO-1 along the cell membrane was quantified from the images and is shown in the graphs. c Rose plots tracking the movement of six single shCtrl and shTXNIP cells. Each color represents the track of an individual cell. The quantification of velocity and distance from five independent experiments are shown. d RPE cell lysates were collected and analyzed by western blotting using the indicated antibodies. e Immunostaining was performed on fixed cells with mouse an anti-paxillin antibody and a rabbit anti-FAK antibody. *p < 0.01 versus shCtrl. The error bars indicate the SEM. Scale bars, 30 μm (a); and 50 μm (b and e)

Fig. 5. Loss of TXNIP increases HIF-1α and VEGF expression.

a RPE cells were plated on coverslips coated with 10 μg/ml fibronectin. Immunostaining was performed on fixed cells with an anti-HIF-1α antibody. b RPE cells were lysed and blotted with the indicated antibodies. c The expression level of VEGF in RPE cells was analyzed by reverse transcription PCR (RT-PCR, left panel) and qPCR (right panel). d The secretion level of VEGF in RPE cells was determined by ELISA. Each experiment was performed in triplicate and repeated three times to assess the reproducibility of the results. e Representative images of capillary-like tube formation in HRMECs induced by coculture with RPE cells after 18 h. The relative wound and tube areas were calculated using ImageJ software. *p < 0.01 versus shCtrl. The error bars indicate the SEM. Scale bars, 30 μm (a) and 100 μm (e)