Ref-1 is overexpressed in neovascular eye disease and targetable with a novel inhibitor

Abstract

Reduction-oxidation factor-1 or apurinic/apyrimidinic endonuclease 1 (Ref-1/APE1) is a crucial redox-sensitive activator of transcription factors such as NF-κB, HIF-1α, STAT-3 and others. It could contribute to key features of ocular neovascularization including inflammation and angiogenesis; these underlie diseases like neovascular age-related macular degeneration (nAMD). We previously revealed a role for Ref-1 in the growth of ocular endothelial cells and in choroidal neovascularization (CNV). Here, we set out to further explore Ref-1 in neovascular eye disease. Ref-1 was highly expressed in human nAMD, murine laser-induced CNV and Vldlr^-/- mouse subretinal neovascularization (SRN). Ref-1's interaction with a redox-specific small molecule inhibitor, APX2009, was shown by NMR and docking. This compound blocks crucial angiogenic features in multiple endothelial cell types. APX2009 also ameliorated murine laser-induced choroidal neovascularization (L-CNV) when delivered intravitreally. Moreover, systemic APX2009 reduced murine SRN and downregulated the expression of Ref-1 redox regulated HIF-1α target carbonic anhydrase 9 (CA9) in the Vldlr^-/- mouse model. Our data validate the redox function of Ref-1 as a critical regulator of ocular angiogenesis, indicating that inhibition of Ref-1 holds therapeutic potential for treating nAMD.

Keywords: Angiogenesis; Hypoxia; Inflammation; Neovascular age-related macular degeneration; Redox signaling; Reduction–oxidation factor-1.

Fig. 1

Ref-1 is highly expressed in neovascular age-related macular degeneration (nAMD). Immunostaining of Ref-1 on paraffin sections of (a) human nAMD and healthy control eyes, where DAPI (in blue) shows nuclear staining and magenta indicates Ref-1 expression in multiple retinal layers and in the RPE-choroid. The images shown here are representative of three nAMD patients and non-AMD, healthy controls (n = 3). Scale bars = 20 μm. GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer; IS/OS, inner and outer segments of photoreceptors; RPE, retinal pigment epithelium; CC, choriocapillaris. (b) Mean fluorescence intensity (MFI) quantification of Ref-1 staining in INL and ONL of the retina, and in the RPE layer of nAMD vs controls. Mean ± SEM, n= 3. *p<0.05, unpaired t-test with Welch’s correction

Fig. 2

Association of Ref-1 with cell-specific markers in human nAMD vs healthy control eyes. Representative images of retinal and RPE-choroid sections from nAMD and non-AMD control subjects stained with Ref-1 (magenta), RBPMS and PNA (yellow), calbindin, vimentin, rhodopsin, and RPE-65 (green), and DAPI (blue). The control eye sections were used for IgG control wherein sections were stained with preimmune serum in lieu of primary antibodies or lectin. Ref-1 associates partially with markers of RGCs (RBPMS), horizontal cells (calbindin), Müller glia (vimentin), rod (rhodopsin) and cone (PNA) photoreceptors, and the RPE (RPE65) with intense expression seen in multiple retinal layers and RPE of nAMD vs control samples. Scale bar: 50 μm. Representative images of sections from n = 2 or 3 nAMD and non-AMD controls. GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer; IS/OS, inner and outer segments of photoreceptors; RPE, retinal pigment epithelium; CC, choriocapillaris

Fig. 3

Ref-1 is overexpressed in murine laser-induced choroidal neovascularization (L-CNV). (a) Choroidal flat-mount staining of L-CNV eyes, demonstrating the higher expression of Ref-1 (magenta) in and around the lesions and its association with vasculature (stained by GS-IB4) on day 1 and day 5 post laser compared to the flat-mounts of untouched control eyes. (b, c) Mouse eye cryosections of the (b) retina and (c) RPE/choroid, showing increased Ref-1 expression in the GCL, INL, ONL and IS/OS, and in RPE-choroid complex where neovascularization is seen via GS-IB4 staining (green) in the L-CNV eyes contrasted to non- L-CNV, control eyes. The apparent disorganization of retinal layers is likely a sectioning artifact. Scale bars = 50 μm. GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer; IS/OS, inner and outer segments of photoreceptors; RPE, retinal pigment epithelium

Fig. 4

Ref-1 is highly expressed and associates with subretinal neovascularization (SRN) in Vldlr^−/− mice. (a) Immunoblot of Ref-1 levels in choroidal tissues containing SRN tufts of Vldlr^−/− mice compared to age-matched (C57BL/6J or C57) controls. Pooled choroids from 3 mice per condition. (b) Confocal images of SRN on choroidal flat mounts immunostained for Ref-1 (magenta), hypoxyprobe (cyan) and GS-IB4 (green). Ref-1 colocalizes with markers of vasculature and hypoxia at SRN tufts. Scale bar = 20 μm. (c) Immunoblot quantification of Ref-1 levels (normalized to β-actin); two-way ANOVA with Tukey’s post-hoc tests, *p<0.05 (n=3). Note that the variable Ref-1 expression pattern seen between the immunoblots and flat-mounts is because the immunoblot samples were whole lysates prepared from Vldlr^−/− mouse choroid. (d) Mean fluorescence intensity (MFI) quantification of Ref-1 at SRN tufts. Upregulated Ref-1 MFI at the tufts of SRN related to age-matched controls; two-way ANOVA with Tukey’s post-hoc tests, ****p<0.0001 (n=8 per group)

Fig. 5

Interaction of APX2009 results in chemical shift perturbations (CSPs) in Ref-1. (a) 2D ¹⁵N-HSQC spectrum with the superimposition of the free protein spectrum (blue) and spectrum with APX2009 (10-fold molar excess) in red. Each ¹H/¹⁵N peak corresponds to one backbone or side chain amide. Specific chemical shift perturbations are shown in small boxes. (b) CSPs vs residue number reveal interactions with Ref-1. Interacting residues with CSPs higher than 0.02 include Asp70, Val131, Gln137, Ser164, Val172, and Asn212. (c) These interacting residues are shown as stick renderings (N, blue, O, red, and C green, for residues 70 and 212, cyan for 131 and 172, and magenta for 137 and 164) in a cartoon rendering of Ref-1/APE1 (4QHD). In (d) and (e), a semi-transparent surface of Ref-1 is shown in light gray, cartoon rendering in gray, and interacting residues in stick renderings (colors as indicated in (c)). Surface residues appear in color on the surface. Residues 70 and 212 are located within the endonuclease active site of the protein, while 137 and 164 define a small pocket on the opposite face of the protein. The other two interacting residues, 131 and 172, are internal residues. (f) A closeup of APX2009 docked using AutoDock 4.2 in the small pocket defined by Q137 and S164 (shown as stick models, C magenta, O red, and N blue) in Ref-1/APE1 (PDB ID 4QHD, shown as a cartoon with a semi-transparent surface). Two consensus poses for APX2009 are shown, one with “ring-in” (C yellow, O red, and N blue) and one with ring out (C, cyan). (g) A closeup view of the consensus “ring-in” and “ring-out” poses from AutoDock Vina. Coloring of the APX2009 stick models are the same as in (f)

Fig. 6

APX2009 inhibits proliferation of iCECs, BMECs and HUVECs in culture. (a, b, c) APX2009 is antiproliferative on (a) iCECs, (b) BMECs, and (c) HUVECs after treatment for 48 h. Mean ± SEM (n = 3 technical replicates). The data are representative of three biological replicates. (d-l) APX2009 (at 24 h of treatment) dose-dependently blocks the cell cycle of (d, g, j) iCECs, (e, h, k) BMECs and (f, i, l) HUVECs at G₁/S-phase with an accompanying cell population increase in G₀/G₁-phase. Mean ± SEM of percentage of cells in each phase; two-way ANOVA with Tukey’s post-hoc tests, *p<0.05, ***p<0.001, ****p<0.0001 vs DMSO controls (n = 3 biological replicates)

Fig. 7

APX2009 hampers tube formation and migration of iCECs, BMECs and HUVECs in culture. (a, b, c) Representative images of cell migration in the scratch wound assay by (a) iCECs, (b) BMECs and (c) HUVECs with indicated doses of APX2009. (d, e, f) Quantitative analyses of cell migration in (d) iCECs, (e) BMECs and (f) HUVECs depict that APX2009 obstructs migration in a concentration-dependent manner. Mean ± SEM (n=6 images). *p<0.05, ***p<0.001 vs. DMSO control; one-way ANOVA with Dunnett’s post-hoc test. (g, h, i) Representative images of tube formation on a layer of Matrigel by (g) iCECs, (h) BMECs and (i) HUVECs with indicated doses of APX2009. (j, k, l) Quantitative analyses of tube formation illustrate that APX2009 significantly reduced the tube formation ability of (j) iCECs, (k) BMECs and (l) HUVECs. Mean ± SEM (n=3 per concentration). *p<0.05, ***p<0.001 vs. DMSO control; one-way ANOVA with Dunnett’s post-hoc test. The data from both assays are representative of three biological replicates

Fig. 8

APX2009 is non-toxic to hPSC-RGCs. (a) Representative images of BRN3b:tdTomato signal in RGCs cultured over 7 days with no treatment (untreated), APX2009, DMSO (vehicle control) and staurosporine (SP; positive control). APX2009 treated RGCs show no toxicity and remained similar to untreated cells, while DMSO alone treated RGCs show negligible, non-significant levels of cell death and SP treatment results in profound RGC cell death. Scale bars = 100 μm (b-h) Quantitative analysis of BRN3b:tdTomato fluorescence intensity on cells treated with APX2009, DMSO and SP vs. untreated cells at different days: (b) day 1, (c) day 2, (d) day 3, (e) day 4, (f) day 5, (g) day 6 and (h) day 7. Mean ± SEM (n=20 images per treatment). ns, non-significant, *p<0.05, **p<0.01, ****p<0.0001 vs. untreated cells; one-way ANOVA with Dunnett’s post-hoc tests. Representative data from four independent experiments

Fig. 9

Intravitreal (IVT) APX2009 ameliorates laser-induced choroidal neovascularization (L-CNV) in a mouse model. (a) IVT pharmacokinetics (PK) of APX2009 (10 μM) in C57BL/6J mice. Linear regression slopes of male (blue) and female (red) mice demonstrate the log concentration of APX2009 (y-axis) at different time points (x-axis). No significant difference in the slopes (that demonstrate the transform of individual data points) of male vs. female mice (n=3 mice/sex); simple linear regression analysis. (b) Representative OCT and FA images obtained 7-days post L-CNV, and confocal microscopy images of GS-IB4 and agglutinin stained L-CNV lesions 7 days post-treatment with a single IVT injection including vehicle (PKT), mouse anti-VEGF₁₆₄ antibody (5 ng/eye), or APX2009 (25 and 50 μM). (c) Quantification of lesion volumes measured from OCT images on day 7. (d, e) Bar graphs representing (d) percentage of FA lesion grades and (e) number of grade 3 lesions. (f, g) Quantification of CNV lesion volumes from confocal images for vasculature markers, (f) GS-IB4 and (g) agglutinin. Anti-VEGF and APX2009 (25 and 50 μM) treatments show markedly reduced lesion volumes (yellow brackets) and fluorescein leakage compared to vehicle control as observed via OCT, FA, and confocal images. Mean ± SEM (n=6 mice/treatment). *p<0.05, **p<0.01, ***p<0.001 vs. vehicle treatment; one-way ANOVA with Dunnett’s post-hoc tests. Scale bar (OCT images) = 100 μm, scale bar (GS-IB4 and agglutinin stained confocal images) = 20 μm. GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer

Fig. 10

IVT APX2009 diminishes microglia and macrophage cells in mouse retinas of the eyes that underwent L-CNV. (a) Representative Z-stack confocal images (of deepest retinal layer) of retinal flat-mounts stained with microglial cell marker IBA1 (red), macrophage marker CD11b (magenta) and vasculature marker GS-IB4 (green) of eyes that had received IVT treatment of vehicle (PKT), mouse anti-VEGF₁₆₄ antibody (5 ng/eye), or APX2009 (25 and 50 μM) upon laser challenge. (b, c) Quantification of (b) IBA1- and (c) CD11b-positive cells. Seven days post-laser challenge, the eyes that received anti-VEGF and APX2009 (25 and 50 μM) treatments showed significantly diminished number of activated IBA1- and CD11b-positive cells in the retina compared to vehicle. Mean ± SEM (n=13-18 images/treatment captured right above the L-CNV lesion area). **p<0.01, ***p<0.001, ****p<0.0001 vs. vehicle treatment; one-way ANOVA with Dunnett’s post-hoc tests. Scale bars = 20 μm

Fig. 11

Intraperitoneal (i.p.) APX2009 reduces subretinal neovascularization (SRN) and CA9 expression in Vldlr^−/− mouse eyes. (a) Representative OCT, fundus, and FA images obtained 7-days (on P22) post i.p. treatments of vehicle (PKT) and APX2009 (12.5 and 25 mg/kg). (b) Representative confocal microscopy images of CA9 (magenta), GS-IB4 (green), and DAPI (blue)-stained SRN lesions on choroidal flat-mounts. (c) GS-IB4 stained choroidal flat-mounts with SRN lesion area highlighted in white, and the total area of the flat-mounts are outlined as dashed white lines. (d) Quantification of the number of lesions counted from FA images. Mean ± SEM (n=10-12 eyes), *p<0.05 vs. vehicle treatment; Kruskal-Wallis test. (e) Mean fluorescence intensity (MFI) quantification of CA9 expression in the SRN lesions. Mean ± SEM (n=32-48 lesions), *p<0.05, **p<0.01 vs. vehicle treatment; one-way ANOVA with Dunnett’s post-hoc test. (f) Percentage quantification of SRN area performed using Adobe Photoshop. Mean ± SEM (n=9-12 flat-mounts per treatment). *p<0.05, **p<0.01 vs. vehicle treatment; one-way ANOVA with Dunnett’s post-hoc test. APX2009 noticeably reduced subretinal alterations (yellow arrows) on the OCT, and the number of SRN lesions (fundus and FA), SRN area and substantially downregulated CA9 expression in these lesions compared to vehicle treatment. Scale bars = 100 μm (OCT images), 20 μm (CA9 and GS-IB4-stained confocal images), 500 μm (GS-IB4 stained choroidal flat-mounts with SRN lesion area). GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer