Oxidative Stress Induces a VEGF Autocrine Loop in the Retina: Relevance for Diabetic Retinopathy

Abstract

Background: Oxidative stress (OS) plays a central role in diabetic retinopathy (DR), triggering expression and release of vascular endothelial growth factor (VEGF), the increase of which leads to deleterious vascular changes. We tested the hypothesis that OS-stimulated VEGF induces its own expression with an autocrine mechanism.

Methods: MIO-M1 cells and ex vivo mouse retinal explants were treated with OS, with exogenous VEGF or with conditioned media (CM) from OS-stressed cultures.

Results: Both in MIO-M1 cells and in retinal explants, OS or exogenous VEGF induced a significant increase of VEGF mRNA, which was abolished by VEGF receptor 2 (VEGFR-2) inhibition. OS also caused VEGF release. In MIO-M1 cells, CM induced VEGF expression, which was abolished by a VEGFR-2 inhibitor. Moreover, the OS-induced increase of VEGF mRNA was abolished by a nuclear factor erythroid 2-related factor 2 (Nrf2) blocker, while the effect of exo-VEGF resulted Nrf2-independent. Finally, both the exo-VEGF- and the OS-induced increase of VEGF expression were blocked by a hypoxia-inducible factor-1 inhibitor.

Conclusions: These results are consistent with the existence of a retinal VEGF autocrine loop triggered by OS. This mechanism may significantly contribute to the maintenance of elevated VEGF levels and therefore it may be of central importance for the onset and development of DR.

Keywords: HIF-1; MIO-M1; Müller cell; Nrf2; VEGFR2; conditioned medium; diabetic retinopathy; retinal explant.

Conflict of interest statement

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Figures

Figure 1

(A) and (B) represent vascular endothelial growth factor (VEGF) mRNA expression in MIO-M1 cells and in retinal explants, respectively, exposed to oxidative stress (OS; 400 µM H2O2 for MIO-M1 cells and 100 µM H2O2 for retinal explants) for 24 h and effect of a vascular endothelial growth factor receptor 2 (VEGFR2) blocker (0.1 µM Apatinib for MIO-M1 cells and 25 µM SU1498 for retinal explants). * p < 0.05 and ** p < 0.01 relative to controls (Ctrl); §p < 0.05 and §§p < 0.01 relative to OS. (C) and (D) represent VEGF mRNA expression in MIO-M1 cells and in retinal explants, respectively, exposed to different concentrations of exogenous VEGF (exo-VEGF) for 24 h. * p < 0.05 ** p < 0.01 and *** p < 0.001 relative to Ctrl. (E) and (F) represent VEGF mRNA expression in MIO-M1 cells and in retinal explants, respectively, in response to the most effective concentration of exo-VEGF (1 ng/mL in MIO-M1 cells and 10 ng/mL in retinal explants) and the effect of a VEGFR2 blocker (Apatinib for MIO-M1 cells and SU1498 for retinal explants). * p < 0.05 and *** p < 0.001 relative to Ctrl; §p < 0.05 and §§p < 0.01 relative to exo-VEGF. n = 3 in (A–E); n = 5 in (F).

Figure 2

VEGFR2 immunofluorescence in MIO-M1 cells (A,C,E) and in cryostat sections of retinal explants (B,D,F) in Ctrl (A,B), in the presence of 1 ng/mL (MIO-M1 cells) or 10 ng/mL (retinal explants) of exo-VEGF (C,D) and in the presence of exo-VEGF with a VEGR2 blocker (0.1 µM Apatinib for MIO-M1 cells and 25 µM SU1498 for retinal explants: E,F). In the images of MIO-M1 cells (A,C,E), above background, specific VEGFR2 immunolabeling was highlighted using Adobe Photoshop, and it appears as bright, whitish dots. Background, non-specific immunostaining is dark green. Cell nuclei were visualized with Hoechst and actin filaments with rhodamine-conjugated phalloidin. In retinal explants (B,D,F), specific VEGR2 immunostaining is bright green, while cell nuclei were visualized with DAPI counterstain. Putative immunolabeled blood vessels are indicated by arrows; putative immunolabeled Müller cell processes are indicated by arrowheads. GCL, ganglion cell layer; INL, inner nuclear layer; IPL, inner plexiform layer; ONL, outer nuclear layer; OPL, outer plexiform layer. Scale bars, 20 µm. G and H indicate VEGFR2 immunofluorescence levels normalized to Ctrl in MIO-M1 cells and in retinal explants, respectively, in the different experimental conditions. eV, exo-VEGF. * p < 0.05 and ** p < 0.01 relative to Ctrl. n = 4.

Figure 3

(A) VEGF release in the culture medium of MIO-M1 cells exposed to OS for 24 h, as measured with ELISA. (B) VEGF mRNA expression in MIO-M1 cells incubated for 24 h in the conditioned medium of untreated MIO-M1 cells (CM-Ctrl) or of MIO-M1 cells exposed to OS for 24 h (CM-OS), with or without the VEGFR2 blocker Apatinib at 0.1 µM. *** p < 0.001 relative to control MIO-M1 cell cultures (Ctrl) (A) or to untreated MIO-M1 cell cultures (that is incubated in non-conditioned medium, Unt) (B), §§§p < 0.001 relative to CM-OS, and ###p < 0.001 relative to CM-Ctrl as evaluated with one-way ANOVA followed by Newman–Keuls post-hoc test. op < 0.005 with respect to Unt as evaluated with one-way ANOVA followed by uncorrected Fisher’s LSD post-hoc test. n = 4 in (A); n = 6 in (B). (C–G): Images of VEGFR2 immunofluorescence in MIO-M1 cells in the experimental conditions as in (B). Other details as in Figure 2. Scale bar, 20 µm.

Figure 4

(A) VEGF release in the culture medium of retinal explants exposed to OS for 24 h, as measured with ELISA. (B) VEGF mRNA expression in retinal explants incubated for 24 h in non-conditioned medium (Ctrl) or in explants incubated for 24 h in the conditioned medium of explants exposed to OS for 24 h (CM-24h). (C) VEGF release in the culture medium of retinal explants exposed to OS for 5 days, as measured with ELISA. (D) VEGF mRNA expression in retinal explants incubated for 24 h in the conditioned medium of untreated explants (CM-Ctrl) or in explants incubated for 24 h in the conditioned medium of explants exposed to OS for 5 days (CM-5D). ** p < 0.01 and *** p < 0.001 relative to Ctrl. n = 3 in all graphs.

Figure 5

Immunofluorescence images documenting nuclear factor erythroid 2-related factor 2 (Nrf2) production and nuclear translocation in Ctrl, OS-treated (OS) and in OS-treated MIO-M1 cells incubated in the presence of the Nrf2 inhibitor ML385 at 5 µM (OS + ML385). (A1), (B1) and (C1) show the overall Nrf2 immunostaining of MIO-M1 cells in the three different conditions; (A2), (B2) and (C2) show the Hoechst-stained cell nuclei; (A3), (B3) and (C3) are merged images of the previous two; (A4), (B4) and (C4) show Nrf2 immunostaining limited to the cell nucleus. Scale bar, 50 µm. The histograms in (D) represent the quantification of immunofluorescence intensity within the nuclei of MIO-M1 cells in the different experimental conditions. The values are relative to that corresponding to 0 µM ML385. ** p < 0.01 and *** p < 0.001 relative to 0 µM ML385; §§p < 0.01 relative to OS + 0 µM ML385; ###p < 0.001 relative to OS + 1 µM ML385. The histograms in (E) represent VEGF mRNA expression in MIO-M1 cells exposed to OS for 24 h and effect of the Nrf2 inhibitor ML385. ** p < 0.01 relative to Ctrl; §§p < 0.01 relative to OS. n = 8 in (D); n = 3 in (E).

Figure 6

Immunofluorescence images documenting Nrf2 production and nuclear translocation in Ctrl, in exo-VEGF-treated and in exo-VEGF-treated MIO-M1 cells incubated in the presence of the Nrf2 inhibitor ML385 (exo-VEGF + ML385). (A1), (B1) and (C1) show the overall Nrf2 immunostaining of MIO-M1 cells in the three different conditions; (A2), (B2) and (C2) show the Hoechst-stained cell nuclei; (A3), (B3) and (C3) are merged images of the previous two; (A4), (B4) and (C4) show Nrf2 immunostaining limited to the cell nucleus. Scale bar, 50 µm. The histograms in (D) represent the quantification of immunofluorescence intensity within the nuclei of MIO-M1 cells in the different experimental conditions. The values are relative to that corresponding to 0 µM ML385; eV, exo-VEGF. The histograms in (E) represent VEGF mRNA expression in MIO-M1 cells exposed to exo-VEGF (eV) for 24 h and effect of the Nrf2 inhibitor ML385. *** p < 0.001 relative to Ctrl. n = 4 in (D); n = 3 in (E).

Figure 7

VEGF mRNA expression in MIO-M1 cells exposed to OS or to exo-VEGF (eV) for 24 h and effect of the hypoxia inducible factor-1 inhibitor acriflavine (ACF) at 5 µM. * p < 0.05, ** p < 0.01 and *** p < 0.001 relative to Ctrl; §§§p < 0.001 relative to OS; ###p < 0.001 relative to eV. n = 4.

Figure 8

Schematic interpretation of the results of the present study and of other data in the literature showing the possible mechanism of an OS-induced VEGF autocrine loop in the retina. This mechanism can be divided into two parts (pathways). According to our observations, in “pathway 1”, OS triggers Nrf2 activation and nuclear translocation. In the nucleus, Nrf2 induces the expression of antioxidant genes, and in particular the one coding the HO-1 enzyme, which is reported in the literature to be able to stabilize HIF-1α. For the sake of simplicity, the diagram does not indicate HIF-1α or STAT3 nuclear translocation, induction of VEGF transcription, VEGF translation, or VEGF release, but simply indicates that HIF-1α stabilization results in VEGF release. This activates “pathway 2”, in which the released VEGF would bind to VEGFR2, which in turn (likely through STAT3 activation, according to the literature) would again stimulate HIF-1α stabilization and nuclear translocation for further VEGF expression and release.