Increased expression of catalase and superoxide dismutase 2 reduces cone cell death in retinitis pigmentosa

Abstract

Oxidative and nitrosative damage are major contributors to cone cell death in retinitis pigmentosa (RP). In this study, we explored the effects of augmenting components of the endogenous antioxidant defense system in models of RP, rd1, and rd10 mice. Unexpectedly, overexpression of superoxide dismutase 1 (SOD1) in rd1 mice increased oxidative damage and accelerated cone cell death. With an elaborate mating scheme, genetically engineered rd10 mice with either inducible expression of SOD2, Catalase, or both in photoreceptor mitochondria were generated. Littermates with the same genetic background that did not have increased expression of SOD2 nor Catalase provided ideal controls. Coexpression of SOD2 and Catalase, but not either alone, significantly reduced oxidative damage in the retinas of postnatal day (P) 50 rd10 mice as measured by protein carbonyl content. Cone density was significantly greater in P50 rd10 mice with coexpression of SOD2 and Catalase together than rd10 mice that expressed SOD2 or Catalase alone, or expressed neither. Coexpression of SOD2 and Catalase in rd10 mice did not slow rod cell death. These data support the concept of bolstering the endogenous antioxidant defense system as a gene-based treatment strategy for RP, and also indicate that coexpression of multiple components may be needed.

Figures

Figure 1

Superoxide dismutase 1 (SOD1) overexpression significantly decreases cone function and cone cell number in rd1+/+ mice. Transgenic mice in which the actin promoter drives expression of human SOD1 were crossed with rd1+/+ mice and offspring were crossed to obtain rd1+/+ mice that carried the Sod1 transgene (Sod1-rd1+/+ mice). (a) At postnatal day (P) 25, rd1+/+, and Sod1-rd1+/+ mice were euthanized and retinal homogenates were run in western blots using an antibody directed against human SOD1. Immunoblots (IBs) showed strong expression of human SOD1 in Sod1-rd1+/+ and no detectable expression in rd1+/+ mice. Stripping and reprobing of IBs with an antibody directed against β-actin showed that loading was equivalent. (b) At P25, the mean (±SEM) number of carbonyl adducts determined by enzyme-linked immunosorbent assay of retinal homogenates showed a significant increase in oxidized proteins in Sod1-rd1+/+ mice (n = 6) compared to rd1 mice (n = 9;*P < 5.0 × 10−4 by unpaired Student's t-test). (c) At P35, compared to rd1 mice, Sod1-rd1+/+ mice appeared to show lower cone density in all four quadrants of the retina by confocal microscopy of peanut agglutinin–stained retinal flat mounts (scale bar = 50 µm) and this was confirmed by image analysis (d; *P < 2.0 × 10−4, **P < 0.02, ***P < 0.002, ****P < 0.01 by unpaired Student's t-test). (e) Representative wave forms from photopic electroretinograms (ERGs) done in low background illumination at P25 showed lower b-waves for Sod1-rd1+/+ mice than rd1+/+ mice and measurements confirmed a significant reduction in mean (±SEM) b-wave amplitude Sod1-rd1+/+ mice (*P < 0.05 by unpaired Welch's t-test).

Figure 2

Rd10+/+ mice with inducible increased expression of superoxide dismutase 2 (SOD2) and Catalase in the mitochondria of photoreceptors. (a) Schematic diagram of the TRE/Sod2 and TRE/Catalase transgenes are shown. The tetracycline response element (TRE) was coupled to the full-length cDNA for mouse-Sod2. The ornithine transcarbamylase (OTC) leader sequence, which mediates mitochondrial localization, was ligated to the N terminus cDNA for human Catalase and the peroxisomal localization signal (PLS) was deleted from the C terminus prior to coupling to the TRE. Using these constructs, TRE/Sod2 and TRE/Catalase transgenic mice were generated. (b) Multiple crosses were done to generate TRE/Sod2(+/−)-TRE/Catalase(+/−)-rd10+/+ mice and homozygous interphotoreceptor retinol–binding protein promoter/reverse tetracycline transactivator-rd10+/+ mice (IRBP/rtTA(+/+)-rd10+/+ mice). These two types of mice were crossed to yield four groups of offspring, null-rd10+/+, Sod2-rd10+/+, Catalase-rd10+/+, and Sod2/Catalase-rd10+/+ mice for which the genotypes are shown. (c) Null-rd10+/+, Sod2-rd10+/+, Catalase-rd10+/+, and Sod2/Catalase-rd10+/+ mice were given normal drinking water or water supplemented with 2 mg/ml of doxycycline between postnatal day (P) 10 and P25. Mice were euthanized and the mitochondrial fractions of retinal homogenates were run in immunoblots using antibodies specific for murine SOD2, human Catalase, and murine cyclooxygenase 4 (COX4), which is known to localize to mitochondria. Background levels of murine SOD2 were seen in retinal mitochondria of all mice, but when treated with doxycycline, only Sod2-rd10+/+ and Sod2/Catalase-rd10+/+ mice showed a substantial increase in SOD2. Likewise, when treated with doxycycline Catalase-rd10+/+ and Sod2/Catalase-rd10+/+ showed strong bands for Catalase. Strong bands for COX4 were seen in the retinal mitochondria of all mice.

Figure 3

Cooverexpression of superoxide dismutase 2 (SOD2) and Catalase in mitochondria reduce superoxide radicals in the retinas of rd10+/+ mice. At postnatal day (P) 35, hydroethidine was injected intraperitoneally into wild-type mice (n = 4), null-rd10+/+ mice treated with doxycycline between P10 and P35 as described in Materials and Methods (n = 4), or Sod2/Catalase-rd10+/+ mice treated with doxycycline between P10 and P35 (n = 4) and after 18 hours the mice were euthanized and ocular sections were examined by confocal microscopy. Representative sections showed minimal fluorescence in the retinas of wild-type mice (a–c), strong fluorescence primarily in the remaining outer nuclear layer (ONL) and outer plexiform layer of the retinas of null-rd10+/+ mice (d–f), and minimal fluorescence in the retinas of Sod2/Catalase-rd10+/+ mice (g–i). This demonstrates a marked increase in superoxide radicals in the outer retina of mice after degeneration of rods that is reduced by coexpression of SOD2 and Catalase. Scale bar = 20 µm. GCL, ganglion cell layer; INL, inner nuclear layer.

Figure 4

Increased expression of Catalase and superoxide dismutase 2 (SOD2) significantly reduce carbonyl content in the retinas of postnatal day (P) 50 rd10+/+ mice. Starting at P10, the mothers of null-rd10+/+, Sod2-rd10+/+, Catalase-rd10+/+, and Sod2/Catalase-rd10+/+ mice and after weaning the mice themselves were treated with doxycycline. Mice were euthanized at P35 or P50 and protein carbonyl content was measured by enzyme-linked immunosorbent assay of retinal homogenates. At P35, the mean (±SEM) carbonyl content per mg retinal protein was significantly greater in Sod2-rd10+/+ mice than null-rd10+/+, Catalase-rd10+/+, or Sod2/Catalase-rd10+/+ mice (a; *P < 0.05; **P < 0.01 by Tukey–Kramer test). At P50, the mean (±SEM) carbonyl content per mg retinal protein was significantly less in Sod2/Catalase-rd10+/+ mice compared to null-rd10+/+, Sod2-rd10+/+, or Catalase-rd10+/+ mice (b; **P < 0.01 by Tukey–Kramer test).

Figure 5

Increased expression of superoxide dismutase 2 (SOD2) and Catalase in mitochondria of photoreceptors decreases cone cell death in rd10+/+ mice. (a) Fluorescence confocal microscopy of peanut agglutinin (PNA)-stained retinal flat mounts showed little difference in cone cell density in 0.0529 mm2 bins 0.5 mm superior to the center of the optic nerve in rd10+/+ mice at postnatal day (P) 18 or 35 compared to wild-type mice at P18, but by P50 there was an obvious reduction in cone density in rd10+/+ mice. At P18, outer segments were seen in wild-type and rd10+/+ mice, but at P35 and P50, rd10+/+ mice had flattened inner segments and no outer segments. Scale bar = 50 µm. (b) Starting at P10, the mothers of null-rd10+/+, Sod2-rd10+/+, Catalase-rd10+/+, and Sod2/Catalase-rd10+/+ mice were treated with doxycycline in their drinking water. After weaning, the mice themselves were treated with doxycycline. At P50, mice were euthanized and fluorescence microscopy of PNA-stained retinal flat mounts in 0.0529 mm2 bins 0.5 mm superior, inferior, temporal, and nasal to the center of the optic nerve are shown. Sod2/Catalase-rd10+/+ mice appeared to have greater cone density in all four regions of the retina compared to null-rd10+/+, Sod2-rd10+/+, and Catalase-rd10+/+ mice. Sod2-rd10+/+ mice appeared to have the lowest cone density. Scale bar = 50 µm. (c) Quantification of cone density by image analysis in each of the four 0.0529 mm2 bins showed that Sod2/Catalase-rd10+/+ mice had significantly greater mean (±SEM) cone density than Sod2-rd10+/+ mice in the superior, inferior, and nasal quadrants of the retina (*P < 0.05, **P < 0.01 by Tukey–Kramer test). Sod2/Catalase-rd10+/+ mice had significantly greater cone density than null-rd10+/+ mice in the inferior and nasal quadrants. Sod2/Catalase-rd10+/+ mice had significantly greater cone density than Catalase-rd10 mice in the nasal quadrant. Scale bar = 50 µm. (d) Cone density measurements from each of the four quadrants in each mouse were consolidated to provide a single cone density measurement per retina. The mean (±SEM) cone density per retina was significantly greater in P50 Sod2/Catalase-rd10+/+ mice compared to null-rd10+/+, Sod2-rd10+/+, or Catalase-rd10+/+ mice (**P < 0.01 by Tukey–Kramer test).

Figure 6

Overexpression of superoxide dismutase 2 (SOD2) and/or Catalase does not prevent rod cell death in rd10+/+ mice. Rod cell death leads to progressive thinning of the outer nuclear layer (ONL) in rd10+/+ mice. Measurement of ONL thickness of doxycycline-treated mice showed no significant differences by Tukey–Kramer test between null-rd10+/+, Sod2-rd10+/+, Catalase-rd10+/+, and Sod2/Catalase-rd10+/+ mice at P25 (a) and P35 (b). The bars show the mean (±SD).

Figure 7

Increased expression of superoxide dismutase 2 (SOD2) and Catalase preserves some cone cell function at postnatal day (P) 50 in rd10+/+ mice. (a) Scotopic electroretinograms (ERGs) were done at P35 in null-rd10+/+, Sod2-rd10+/+, Catalase-rd10+/+, and Sod2/Catalase-rd10+/+ mice treated with doxycycline. The mean (±SEM) b-wave amplitude for four different stimulus intensities is plotted for each of four groups of mice and there were no significant differences. (b) Low background photopic ERGs were done as described in Materials and Methods at P50. Representative waveforms are shown for each of the four groups and illustrate a substantially better waveform in Sod2/Catalase-rd10+/+ mice compared to null-rd10+/+, Sod2-rd10+/+, or Catalase-rd10+/+ mice. The bars show mean (±SEM) photopic b-wave amplitude, which was significantly higher (**P < 0.01 by Tukey–Kramer test) for Sod2/Catalase-rd10+/+ mice compared to the other three types of mice.