Metabolic plasticity in a Pde6bSTOP/STOP retinitis pigmentosa mouse model following rescue

Abstract

Objective: Retinitis pigmentosa (RP) is a hereditary retinal disease characterized by progressive photoreceptor degeneration, leading to vision loss. The best hope for a cure for RP lies in gene therapy. However, given that RP patients are most often diagnosed in the midst of ongoing photoreceptor degeneration, it is unknown how the retinal proteome changes as RP disease progresses, and which changes can be prevented, halted, or reversed by gene therapy.

Methods: Here, we used a Pde6b-deficient RP gene therapy mouse model and performed untargeted proteomic analysis to identify changes in protein expression during degeneration and after treatment.

Results: We demonstrated that Pde6b gene restoration led to a novel form of homeostatic plasticity in rod phototransduction which functionally compensates for the decreased number of rods. By profiling protein levels of metabolic genes and measuring metabolites, we observed an upregulation of proteins associated with oxidative phosphorylation in mutant and treated photoreceptors.

Conclusion: In conclusion, the metabolic demands of the retina differ in our Pde6b-deficient RP mouse model and are not rescued by gene therapy treatment. These findings provide novel insights into features of both RP disease progression and long-term rescue with gene therapy.

Keywords: Gene therapy; Inflammation; Metabolism; OXPHOS; Phototransduction; Proteomics; Retina; Retinal plasticity; Retinitis pigmentosa.

Figure 1

Pde6b restoration rescued PDE6 expression and most but not all dysregulated proteins. (A) Schematic representation of experiment. Tamoxifen injection (at 4 weeks) activates CreERT2 recombinase, which splices out the stop cassette, leading to PDE6B expression. (B–J) WT, mutant, and treated retinas were analyzed at 4 and/or 8 weeks of age. Treated mice were tamoxifen-injected at 4 weeks of age. (B-D, G-J) Retinas were analyzed by label-free liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based proteomics. (B–D) Expression of PDE6 subunits was reduced in mutant and restored in treated mice. (E) Representative PDE6B immunoblot of retinal lysates. β-Actin was used as a loading control. (F) Representative images of retinal sections immunostained for PDE6B and counterstained with Hoechst 33342. Scale bar, 15 μm. (G, H) Venn diagrams representing the number of unique or overlapping proteins that were differentially expressed in 4-week-old-mutant and treated mice (FDR<0.1) (G) and in 8-week-old mutant and treated mice (H) in comparison to 8-week-old WT mice. (I, J) Volcano plots showing differentially expressed proteins between 8-week-old WT and treated (I) and 8-week-old mutant and treated (J) retinas. Proteins with fold change >5, and FDR<0.1 are highlighted. (K) Scotopic (−3 and −2 log cd × s/m²) and mesopic a-wave amplitudes. (L) Scotopic (−3 and −2 log cd × s/m²) and mesopic b-wave amplitudes. (M) Photopic b-wave amplitudes. (K–M) N = 5 for WT and treated, N = 7 for mutant. (B-D, K-M) Data, presented as mean ± SEM, were compared by ANOVA. ∗P ≤ 0.05; ∗∗P ≤ 0.01; ∗∗∗P ≤ 0.001.

Figure 2

Pde6b gene restoration increased expression of proteins involved in phototransduction. Retinas from WT, mutant, and treated mice were analyzed at 4 and/or 8 weeks of age by label-free liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based proteomics. Treated mice were tamoxifen-injected at 4 weeks of age. (A–F) RHO (A), GNAT1 (B), RCVRN (C), CNGA1 (D), ABCA4 (E), and GRK1 (F) are essential for the phototransduction cascade and were significantly downregulated in mutant compared to WT retinas. Their expression could be restored in treated mice. (G) Heat map representation of proteins involved in sensory perception of light stimulus. (A–F) Data, presented as mean ± SEM, were compared by ANOVA. ∗P ≤ 0.05; ∗∗P ≤ 0.01; ∗∗∗P ≤ 0.001.

Figure 3

Pde6b gene restoration halted activation of Müller cells. WT, mutant, and treated retinas were analyzed at 4 and/or 8 weeks of age. Treated mice were tamoxifen-injected at 4 weeks of age. (A-C, G) Retinas were analyzed by label-free liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based proteomics (A) Quantitative analysis of the gliosis marker glial fibrillary acidic protein (GFAP) expression revealed significantly higher levels in mutant compared to treated and WT retinas at 8 weeks of age. (B, C) The quantitative analysis of CD44 and S100A6 revealed significantly higher levels of CD44 (B) and S100A6 (C) in mutant compared to treated mice at 8 weeks of age. (A–C) Data, presented as mean ± SEM, were compared by ANOVA. ∗P ≤ 0.05; ∗∗∗P ≤ 0.001. (D, E) Representative images of retinal sections immunostained for CD44 (D) and S100A6 (E). Both proteins are exclusively expressed in Müller cells. Scale bar, 15 μm. (F) Representative GFAP, CD44, and S100A6 immunoblots of retinal lysates. β-Actin was used as a loading control. (G) Heat map representation of gliosis-associated proteins. ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer.

Figure 4

Pde6b gene restoration reversed/halted the activation of the JAK-STAT and MAPK pathways. WT, mutant, and treated retinas were analyzed at 4 and/or 8 weeks of age. Treated mice were tamoxifen-injected at 4 weeks of age. (A-B, D-I) Retinas were analyzed by label-free liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based proteomics. (A–B) Quantitative analysis of STAT1 (A) and STAT3 (B). (C) Representative STAT3 and pSTAT3 immunoblot of retinal lysates. β-Actin was used as a loading control. (D–E) Quantitative analysis of ERK1 (D), and ERK2 (E). (F–G) Quantitative analysis of MAP2K1 (F) and MAP2K2 (G). (H) Quantitative analysis of complement component 3 (C3) expression. (I) Heat map representation of proteins involved in positive regulation of ERK1/2 cascade (GO 0070374, GO 0050727). (A-B, D-H) Data, presented as mean ± SEM, were compared by ANOVA. ∗P ≤ 0.05; ∗∗P ≤ 0.01; ∗∗∗P ≤ 0.001.

Figure 5

Upregulation of OXPHOS related proteins in treated and mutant retinas. (A, B, D–I) WT, mutant, and treated retinas were analyzed at 4 and/or 8 weeks of age. Treated mice were tamoxifen-injected at 4 weeks of age. Retinas were analyzed by label-free liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based proteomics. (A) Quantitative analysis of LDHB. (B) Quantitative analysis of LDHA expression. (C) Representative LDHA immunoblot of retinal lysates. β-Actin was used as a loading control. (D–I) Quantitative analysis of the OXPHOS proteins MPC1 (D), CHDH (E), TIMMDC1 (F), and cytochrome C oxidase (COX) nuclear-encoded subunits COX6C (G), COX7A1 (H), COX7B (I). (A, B, D–I) Data, presented as mean ± SEM, were compared by ANOVA. ∗P ≤ 0.05; ∗∗P ≤ 0.01; ∗∗∗P ≤ 0.001. (J) Representative immunoblot of the OXPHOS complexes I–V of retinal lysates. β-Actin was used as a loading control. (K) Lactate secretion from retinal explants after 15, 30, and 60 min from WT (week 10) and mutant (week 8) mice. (L) Glucose consumption from retinal explants after 60 min from WT (week 10) and mutant (week 8) mice. (M) ATP analysis showed increased levels in mutant retinas compared to WT at 8 weeks of age. (L–M) Data, presented as mean ± SEM, were compared by unpaired t-test. ∗P ≤ 0.05.

Figure 6

Upregulation of OXPHOS-related proteins in neurons from mutant retinas.(A) Simplified scheme of the electron transport chain located within the inner mitochondrial membrane. (B–J) MACS enriched Müller cell and neuronal cell fractions from 8-week-old WT and mutant retinas were subjected to label-free liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based proteomics. Quantitative analysis of NDUFS3 (B), NDUFB1 (C), CYC1 (D), COX6C (E), COX7A1 (F), COX7B (G), COX7C (H), VDAC1 (I) and mt-CO3 (J). (B–J) Data, presented as mean ± SEM, were compared by two-way ANOVA followed by Bonferroni post hoc test. ∗P ≤ 0.05; ∗∗P ≤ 0.01; ∗∗∗P ≤ 0.001.

Figure 7

Graphical summary.