Spatially and Temporally Regulated NRF2 Gene Therapy Using Mcp-1 Promoter in Retinal Ganglion Cell Injury

Abstract

Retinal ganglion cell degeneration triggered by axonal injury is believed to underlie many ocular diseases, including glaucoma and optic neuritis. In these diseases, retinal ganglion cells are affected unevenly, both spatially and temporally, such that healthy and unhealthy cells coexist in different patterns at different time points. Herein, we describe a temporally and spatially regulated adeno-associated virus gene therapy aiming to reduce undesired off-target effects on healthy retinal neurons. The Mcp-1 promoter previously shown to be activated in stressed retinal ganglion cells following murine optic nerve injury was combined with the neuroprotective intracellular transcription factor Nrf2. In this model, Mcp-1 promoter-driven NRF2 expression targeting only stressed retinal ganglion cells showed efficacy equivalent to non-selective cytomegalovirus promoter-driven therapy for preventing cell death. However, cytomegalovirus promoter-mediated NRF2 transcription induced cellular stress responses and death of Brn3A-positive uninjured retinal ganglion cells. Such undesired effects were reduced substantially by adopting the Mcp-1 promoter. Combining a stress-responsive promoter and intracellular therapeutic gene is a versatile approach for specifically targeting cells at risk of degeneration. This strategy may be applicable to numerous chronic ocular and non-ocular conditions.

Keywords: AAV gene therapy; Mcp-1; NRF2; glaucoma; optic neuropathy; oxidative stress; retina.

Figure 1

Temporal Profile of Mcp-1 Promoter Transcription Activity and RGC Death following Murine ONC (A) Schematic representation of the AAV2/2 construct. The Mcp-1 promoter was designed to drive EGFP reporter gene expression (pMcp-1.EGFP). (B) Quantification of EGFP-positive cells and dying cells labeled with Sytox Orange each in a different group of mice (N = 8 for each group) before ONC and 5, 10, 15, and 20 days after injury. Images were obtained using in vivo confocal microscopy. The y axis indicates the number of Sytox-positive or EGFP-positive cells relative to that of the same mouse 5 days after ONC, presented in percentages. The x axis designates days after ONC. (C) Retinal flat mount prepared at 1, 5, and 10 days after ONC, showing colocalization of EGFP driven by the Mcp-1 promoter and Sytox Orange following ONC. Colocalization of Sytox Orange and EGFP is not seen at day 10. The arrowheads designate Sytox-positive cells in which Mcp-1 promoter is activated. Vector was injected at 2.0 μL (1.0 × 10¹² gc/mL) per injection. ITR, inverted terminal repeat; hGHpA, human growth hormone polyadenylation signal. EGFP (green), Sytox Orange (red). Scale bar, 20 μm.

Figure 2

Therapeutic Utilities of Mcp-1 Promoter-Driven BDNF and NRF2 Overexpression in Murine ONC (A) Schematic representation of the AAV2/2 constructs. Construct carried BDNF (pMcp-1.BDNF) or NRF2 (pMcp-1.NRF2) as the therapeutic gene driven by the Mcp-1 promoter. (B) qRT-PCR analysis of therapeutic gene expression levels 5 days after injury. The graphs show relative mRNA expression levels in ONC eyes compared to non-treated contralateral eyes (N = 8 per group). (C) Typical western blots using antibodies against human BDNF (left) or NRF2 (right) 5 days after ONC. The 66 kDa band (arrowhead) represents uncleaved NRF2, while the lower band (30 kb) is its cleaved fragment. β-actin and TATA-box binding protein (TBP) were used as internal controls. (D) Representative in vivo confocal images of cell death visualized with Sytox Orange 5 days after ONC. The upper panels are images from control eyes, whereas the lower panels are images from the ONC eyes. Quantification of Sytox-positive RGCs from in vivo images (right; N = 10 per group). (E) qRT-PCR analysis of RGC marker genes. The graphs show levels of mRNA expression in ONC eyes. Values are expressed relative to the non-treated eye 7 days after ONC (N = 8 per group). Each vector was injected at 2.0 μL (1.0 × 10¹² gc/mL) per injection for all experiments. Data represent means ± SEM, *p < 0.05. ITR, inverted terminal repeat; hGHpA, human growth hormone polyadenylation signal; No Inj, no injection; ND, non-detectable; NS, not significant.

Figure 3

Comparison of Therapeutic Efficacy between CMV Promoter- and Mcp-1 Promoter-Driven NRF2 Expression in Murine ONC (A) Schematic representation of the AAV2/2 constructs. Each construct carried NRF2 as the therapeutic gene driven by either the CMV promoter (pCMV.NRF2) or the Mcp-1 promoter (pMcp-1.NRF2). (B) qRT-PCR analysis of human NRF2 gene expression 5 days after injury. The graphs show mRNA expression levels in the ONC eyes relative to contralateral control (No ONC) eyes (N = 8 per group). (C) Representative images of cell death visualized by Sytox Orange 5 days after ONC using in vivo confocal microscopy (left panel). The upper panels are images from control eyes, and the lower panels are from ONC eyes. Quantification of Sytox-positive RGCs (right panel; N = 8 per group). (D) qRT-PCR analysis of RGC marker genes. The graphs show levels of mRNA expression in ONC eyes relative to contralateral eyes 7 days after ONC (N = 8 per group). (E) Measurement of visual acuity (left panel) and contrast sensitivity (right panel) 1 month after ONC (N = 6 per group). Each vector was injected at 2.0 μL (1.0 × 10¹² gc/mL) per injection for all experiments. Data represent mean ± SEM, *p < 0.05. ITR, inverted terminal repeat; hGHpA, human growth hormone polyadenylation signal; No Inj, no injection; NS, not significant.

Figure 4

Off-Target Effects of NRF2 Expression Driven by the Mcp-1 Promoter versus CMV Promoter in Healthy RGCs of Wild-Type Mice (A) qRT-PCR analysis of Ho-1 gene expression in wild-type mice treated or untreated with a therapeutic vector (AAV2/2.pCMV.NRF2 or AAV2/2.pMcp-1.NRF2). Ho-1 expression is regulated by NRF2 (N = 8 per group). (B) qRT-PCR analysis of RGC marker genes in wild-type mice treated or untreated with therapeutic vector. The graphs show levels of mRNA expression in the retina 2 months after AAV injection relative to the contralateral untreated eyes (N = 8 per group). (C) Representative images of retinal section immunoreactivity for Chop (left) and p53 (right) 2 months after AAV injection. Each vector was injected at 2.0 μL (1.0 × 10¹² gc/mL) per injection for all experiments. All data represent means ± SEM, *p < 0.05. GCL, ganglion cell layer; INL, inner nuclear layer; No Inj, no injection; NS, not significant. Scale bar, 50 μm.

Figure 5

Comparative Evaluation of Toxicity from NRF2 Expression Driven by the CMV or Mcp-1 Promoter in Healthy RGCs in Wild-Type Mice (A) Schematic representation of the AAV2/2 vectors used in (B) and (C). Bicistronic constructs were composed of the Mcp-1 promoter (pMcp-1.NRF2-2A-EGFP) or CMV promoter (pCMV.NRF2-2A-EGFP) driving NRF2 connected to EGFP via the self-cleaving peptide (2A). (B) Representative in vivo images of NRF2-expressing EGFP-positive cells (RGCs) in wild-type mice 1, 4, and 8 months after AAV injection obtained using confocal microscopy. (C) Quantification of NRF2-expressing EGFP-positive RGCs from in vivo images (N = 5 each). (D) Schematic representation of the AAV2/2 constructs used in (E) and (F). Each construct carried NRF2 as the therapeutic gene driven by either the CMV promoter (pCMV.NRF2) or the Mcp-1 promoter (pMcp-1.NRF2). (E) qRT-PCR analysis of RGC marker genes in wild-type mice treated or untreated with a therapeutic vector. The graphs show mRNA expression levels in the retina 2 months after AAV injection relative to the contralateral untreated eyes (N = 8 each). (F) Immunohistochemical analysis of retinal sections stained with anti-Brn3a antibodies (green) 4 months after AAV injection (left), with highly magnified images shown in insets. Graphs show quantification of Brn3a-positive cells per DAPI-positive cells (center; N = 6 each) and DAPI-positive cells (right; N = 6 each) in the GCL from images. Each vector was injected at 2.0 μL (1.0 × 10¹² gc/mL) per injection for all experiments. All data represent means ± SEM. *p < 0.05 with Cochran-Armitage test for trend. GCL, ganglion cell layer; INL, inner nuclear layer; 2A, self-cleaving peptide; ITR, inverted terminal repeat; hGHpA, human growth hormone polyadenylation signal. Scale bar, 50 μm.