NRF2 promotes neuronal survival in neurodegeneration and acute nerve damage

Abstract

Oxidative stress contributes to the loss of neurons in many disease conditions as well as during normal aging; however, small-molecule agents that reduce oxidation have not been successful in preventing neurodegeneration. Moreover, even if an efficacious systemic reduction of reactive oxygen and/or nitrogen species (ROS/NOS) could be achieved, detrimental side effects are likely, as these molecules regulate normal physiological processes. A more effective and targeted approach might be to augment the endogenous antioxidant defense mechanism only in the cells that suffer from oxidation. Here, we created several adeno-associated virus (AAV) vectors to deliver genes that combat oxidation. These vectors encode the transcription factors NRF2 and/or PGC1a, which regulate hundreds of genes that combat oxidation and other forms of stress, or enzymes such as superoxide dismutase 2 (SOD2) and catalase, which directly detoxify ROS. We tested the effectiveness of this approach in 3 models of photoreceptor degeneration and in a nerve crush model. AAV-mediated delivery of NRF2 was more effective than SOD2 and catalase, while expression of PGC1a accelerated photoreceptor death. Since the NRF2-mediated neuroprotective effects extended to photoreceptors and retinal ganglion cells, which are 2 very different types of neurons, these results suggest that this targeted approach may be broadly applicable to many diseases in which cells suffer from oxidative damage.

Figures

Figure 8. RGC survival at 2 and 4 weeks after ONC.

(A–C) Immunostaining of RGC cells in retinal flat mounts. Two weeks before ONC, the eyes of adult C57Bl/6J mice received an intravitreal injection of AAV2 vectors expressing GFP (n = 4) (A), SOD2-2A-CAT (n = 6) (B), or NRF2 (n = 6) (C). Retinae were harvested 2 weeks after ONC and stained for β3-tubulin (Tuj1) (green), a specific RGC marker. (D) Quantification of RGC survival at 2 weeks and 4 weeks after ONC. Numbers were normalized to the WT C57Bl/6J retinae without ONC. The number of retinae examined in each group is shown in each bar. Error bars represent the mean ± SEM. *P < 0.05 and **P < 0.01 by 2-way ANOVA.

Figure 7. Evaluation of function following infection with AAV-NRF2.

(A) The optomotor assay was used to test the visual acuity in rd10 mice that were uninjected or injected with AAV-GFP alone or AAV-GFP plus AAV-NRF2. The left and right eyes are shown separately. (B) Ratio of R/L eye visual acuity in rd10 mice at P40 and P50. (C and D) Photopic ERG was performed in rd10 mice at P40. Representative waveforms (C) and average R/L eye ratios of b-wave amplitude (D) for control (n = 17) and AAV-NRF2–treated (n = 23) mice are shown. The number of mice tested in each group is shown in each bar. Error bars represent the mean ± SEM. *P < 0.05 and ***P < 0.0001 by unpaired 2-tailed Student’s t test.

Figure 6. Evaluation of oxidation in AAV-NRF2–treated retinae.

(A–F) Hydroethidine emitted red fluorescence after reacting with superoxide species in the peripheral cross-sections of P40 rd10 retinae infected with AAV-GFP (A–C) and AAV-GFP plus AAV-NRF2 (D–F). (G–L) Anti-acrolein staining of P30 rd1 retinae infected with AAV-GFP (G, H, and K) and with AAV-GFP plus AAV-NRF2 (I, J, and L). (K and L) High-magnification images of the mid-dorsal region in the retinae (highlighted by red frames) shown in H and J. Scale bars: 20 μm (A–F, K, and L); 1.5 mm (G–J).

Figure 5. Effect of NRF2 on cone morphology.

(A and B) PNA staining of surviving cones (red) in P50 rd1 retinae. All images were taken from the mid-dorsal retinal region (~1.5 mm dorsal from the optic nerve head). Insets are high-magnification images of individual cells (original magnification, ×40 for A, B, D, and E insets). (C) Quantification of the percentages of GFP+ cones with PNA staining in the mid-dorsal region of P50 rd1 retinae. Approximately 300 cells per retina were counted for each group. Error bars represent the mean ± SEM. The number of retinae quantified is shown in each bar. *P < 0.05 by unpaired 2-tailed Student’s t test. (D and E) PNA staining of P30 rd1 retinae. (F) PNA (red) and S+M opsin (white) staining in a WT retina. Opsin proteins localized to cone OSs. (G–L) PNA and opsin staining in P160 Rho–/– retinae. Lack of PNA staining and mislocalization of opsins to cell soma were observed in AAV-GFP–infected retinae (n = 9) (G–I), while PNA and opsin colocalized to the remaining cone membrane structures in the retinae infected with AAV-GFP plus AAV-NRF2 (n = 6) (J–L). Scale bars: 20 μm.

Figure 4. Effects of antioxidative TFs on cone survival.

(A) Schematics of AAV-PGC1a and AAV-NRF2 constructs. (B–D) Flat-mounted P50 rd1 retinae infected with AAV vectors expressing GFP, NRF2, and PGC1a (n = 10) (B), GFP and NRF2 (n = 8) (C), and GFP and PGC1a (n = 15) (D). (E–G) Persisting cones in the mid-dorsal retina. (H) Cone density during the course of rd1 degeneration following infection with the indicated constructs. Average cone density and number of retinae examined for each group are summarized in Supplemental Table 1. Error bars represent the mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.0001 by 2-way ANOVA. (I and J) Flat-mounted P110 rd10 retinae infected with AAV vectors expressing GFP (n = 6) (I) and GFP and NRF2 (n = 13) (J). (K and L) Higher-magnification images of the mid-dorsal retinae. (M–P) Flat-mounted P160 Rho–/– retinae infected with AAV vectors expressing GFP (n = 9) (M) and GFP and NRF2 (n = 6) (N), with high-magnification images (O and P). (Q) Quantification of cone density in P110 rd10 and in P160 Rho–/– retinae. The number of retinae quantified for each group is shown at the bottom of each bar. Error bars represent the mean ± SEM. **P < 0.01 by unpaired 2-tailed Student’s t test. Scale bars: 1.5 mm (B–D, I, J, M, and N); 50 μm (E–G, K, L, O, and P).

Figure 3. SOD2 and catalase effects on cone survival.

(A) Schematics of AAV-GFP and AAV-SOD2-2A-CAT constructs. (B–I) Ectopic SOD2 and catalase expression by AAV vectors was visualized by IHC in flat-mounted P60 rd1 retinae. (J and L) Flat-mounted P50 rd1 retinae infected with AAV vectors expressing GFP (n = 29) (J) and GFP, SOD2, and catalase (n = 7) (L). D, dorsal; N, nasal; V, ventral; T, temporal. (K and M) High-magnification images of persisting cones in a 250 μm × 250 μm square from the mid-dorsal retina (1.5 mm dorsal to the optic nerve head) are shown for each group. (N) Illustration of the scheme for quantifying cone density. One 250 μm × 250 μm square located 1.5 mm dorsally, nasally, ventrally, and temporally to the optic nerve head was chosen to represent cone survival in each leaflet of the flat-mounted retinae. GFP+ cones were counted in each square, and the average number of 4 squares was used for each retina. (O) Cone density during the course of rd1 degeneration. On average, 17 retinae per time point were quantified for each group. Error bars represent the mean ± SEM. **P < 0.01 by unpaired 2-tailed Student’s t test. Scale bars: 20 μm (B–I); 1.5 mm (J and L); 50 μm (K and M).

Figure 2. Cones express a high level of GFP from an AAV8 with a CMV promoter.

(A) Dissected retinae were imaged for GFP fluorescence (green) driven by AAV-GFP, following infection at P0 and imaging at P30. (B) Cross-section of an AAV-GFP–infected retina. The ONL shows a high degree of infection (green). (C) High-magnification image of a retinal cross-section showing bright GFP in nearly all cones (colabeled with PNA in red) and weak GFP expression in some rods. (D–F) Short- and medium-wavelength opsin staining (S+M opsin, shown in white) of a 5-month-old Rho–/– mouse retina infected with AAV-GFP. Opsin staining coincided with GFP expression in cones. Scale bars: 0.5 mm (A and B); 20 μm (C–F).

Figure 1. Expression of endogenous antioxidative enzymes in mouse photoreceptors.

(A) Schematic illustration of rod and cone death kinetics of the 3 RP mouse models used in this study. (B–E) Immunostaining of endogenous SOD2 (B and C) and GPX1 (D and E) proteins in WT retinal cross-sections. Cone OSs and ISs are highlighted by PNA (green). Both enzymes were enriched in photoreceptor ISs. Arrowheads point to cones, which expressed a higher level of SOD2/GPX1 than did surrounding rods. (F–I) SOD2 (F and G) and GPX1 (H and I) expression in P30 rd1 retina. OSs, ISs, and ONLs had collapsed to a thin layer, and SOD2/GPX1 expression was mostly lost. Scale bars: 5 μm (B–I).