CoPP-Induced-Induced HO-1 Overexpression Alleviates Photoreceptor Degeneration With Rapid Dynamics: A Therapeutic Molecular Against Retinopathy

Abstract

Purpose: Retinitis pigmentosa (RP) causes progressive photoreceptor degeneration in the retina. The N-methyl-N-nitrosourea (MNU)–administered mouse is used as a chemically induced RP model with rapid progression rate. This study was designed to study heme oxygenase-1 (HO-1) expression in the MNU-administered mice, and to explore the therapeutic effects of cobalt protoporphyrin (CoPP).

Methods: The HO-1 expression in the retina of MNU-administered mice was analyzed. CoPP was injected intravenously into the MNU-administered mice. Subsequently, the CoPP-treated mice were subjected to functional and morphologic examinations.

Results: HO-1 was involved in the MNU-induced photoreceptor degeneration. CoPP treatment enhanced retinal HO-1 expression in the MNU-administered mice. Electroretinogram (ERG) examination and behavioral tests showed that CoPP treatment improved the retinal responsiveness of MNU-administered mice. Histologic analysis and optical coherence tomography (OCT) examination showed that retinal architecture of the CoPP-treated mice was more intact than that of the MNU+vehicle group. Cone photoreceptors in the MNU-administered mice were rescued efficiently by CoPP treatment. Furthermore, multielectrode array (MEA) recording showed that CoPP treatment mitigated the spontaneous firing response, enhanced the light-induced firing response, and preserved the basic configurations of visual signal pathway in the MNU-administered mice. Mechanism studies suggested that CoPP afforded these therapeutic effects by modulating the apoptosis cascades and alleviating the oxidative stress in degenerative retinas.

Conclusions: CoPP alleviated photoreceptor degeneration and rectified the signaling abnormities in MNU-administered mice. CoPP may serve as a potential medication against degenerative retinopathy.

Retinitis pigmentosa (RP) comprises a heterogeneous group of hereditary retinopathies that features progressive photoreceptor degeneration. RP patients always suffer from severe visual functional impairments including night blindness, visual field constriction, visual acuity loss, and complete blindness.^1,2 Hitherto, the exact molecular basis underlying RP remains unknown. Genetics studies suggest that more than 100 mutations are involved in the RP pathogenesis.^3,4 This tremendous heterogeneity is challenging for any therapeutic strategies that seek to deal with the initial gene defects. Currently, photoreceptor apoptosis has been considered as the shared trait of RP with different etiologic backgrounds.⁵ Photoreceptors are responsible for converting light stimulus into electrical signals, which is defined as the first step of visual processing. Their delicate structure, complex function, and energetic oxygen metabolism make photoreceptors extremely vulnerable to external insults.⁶ After the initial death of genetically affected photoreceptors, cellular oxygen consumption reduces prominently, leading to a drastic elevation of oxygen stress in the retina.^7,8 Retinal hyperoxia is toxic, since it generates a burst of reactive oxygen species (ROS) that would perturb the redox metabolism of photoreceptors. Moreover, excessive ROS increase the permeability of the mitochondrial membrane and trigger cytochrome c leakage.⁹ These disturbances would initiate the apoptosis cascades in photoreceptors. Researchers have developed several antiapoptotic therapies against photoreceptor degeneration on this basis. Animal models are essential for screening potential therapeutic molecules against RP.^10,11 In particular, the N-methyl-N-nitrosourea (MNU)–administered mouse serves as a chemically induced RP model with rapid progression rate. This model is characterized by the extinguished electroretinogram (ERG) waveform, disorganized retinal structure, and secondary neural remodeling. All these pathologic hallmarks are similar to those occurring in RP patients to some extent.^12,13 

Heme oxygenase-1 (HO-1) is a rate-limiting enzyme in heme catabolism. HO-1 decomposes heme to biliverdin with the concurrent release of free iron and carbon oxide (CO). These end products have shown potent antiapoptotic and antioxidant ability under pathologic conditions.¹⁴ HO-1 is reactive to external insults including inflammation, heat shock, oxidative damage, and ischemia-reperfusion injury.^14–16 For instance, HO-1 expression in Müller cells increases significantly during retinal ischemic injury. Conversely, inhibiting HO-1 expression with HO-1 short-interfering RNA (siRNA) causes severe inflammatory reaction and retinal damage.¹⁷ These findings highlight the possibility that enhanced HO-1 expression might act as an endogenous compensatory reaction to confer protective effects. Several pioneering studies^18,19 have used HO-1 to counteract cellular injuries. HO-1 overexpression via gene transfer can ameliorate the retinal ischemia-reperfusion injury and the succeeding retinal ganglion cell (RGC) apoptosis.¹⁹ Another study²⁰ has shown that HO-1 overexpression protects human lens epithelial cells from oxidative insults. However, the exact role of HO-1 in RP pathology remains to be explored. 

This study was designed to evaluate HO-1 expression in the retinas of MNU-administrated mice. Cobalt protoporphyrin (CoPP), a HO-1 inducer, was delivered intravenously into MNU-administered mice. We showed that pharmacologic induction of HO-1 alleviates effectively the photoreceptor degeneration in MNU-administrated mice. We focused not only on the retinal morphology, but also on the viability of photoreceptors across different retinal regions. Furthermore, the multielectrode array (MEA) system was used to analyze visual signal transmission within retinal circuits. These findings would be useful for developing a new medication against degenerative retinopathy. 

Materials and Methods

Animals and Pharmacologic Administration

All animal procedures were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. C57BL/6 mice, 8-week-old, of both sexes, were maintained in an air-conditioned facility (room temperature: 18°C–23°C, humidity: 40%–60%, under 12/12-hour light/dark cycle; standard chow and water ad libitum). Animals were divided randomly into four groups: (1) normal control group: mice were left without any pharmacologic administration; (2) MNU group: mice received an intraperitoneal injection of MNU (684-93-5, 60 mg/kg; Sigma-Aldrich, St. Louis, MO, USA); (3) MNU+CoPP group: mice received an intravenous injection of CoPP via caudal vein (C1900, 5 mg/kg; Sigma-Aldrich) 48 hours before MNU administration; and (4) MNU+vehicle group: mice received an intravenous injection of phosphate-buffered saline (PBS) 48 hours before MNU administration. In the preparation process, CoPP was firstly dissolved in 0.1 M NaOH and then diluted to a final concentration of 0.02 M with PBS. The injection approach and dosage for MNU followed previous pharmacologic studies.^21–23 Figure 1 is a schematic illustration of the experimental protocols. 

Figure 1

 

A schematic illustration of experimental protocols. (A) Photoreceptor degeneration in mouse retinas was induced by intraperitoneal injection of MNU. CoPP was injected via caudal vein to induce HO-1 overexpression in retinas. (B) These CoPP-treated mice were then subjected to morphologic and functional analyses.

Optokinetic Behavioral Test

Optokinetic behavioral test was performed by using a two-alternative forced-choice paradigm.²⁴ Firstly, mice (one at a time) were placed on a pedestal situated in the center of a square array of computer monitors. Mice reflexively responded to rotating vertical gratings by moving their head in the direction of grating rotation. Generally, each cycle of examination lasted for 3 to 5 minutes. The examiner monitored the head movements of mice with an infrared television camera. The camera allowed the examiner to observe only the mouse and not the rotating grating. Once the mouse became accustomed to the pedestal with the monitors displaying a 50% gray uniform field, the observer initiated a 5-second trial of a rotating sine-wave stimulus with the direction of rotation randomly selected by the computer-controlled protocol. The stepwise function for correct track response was used to determine the response thresholds. Based on the observer's choices, the computer changed grating contrast, using a staircase paradigm, and converged on a threshold that we defined as 70% of correct observer responses. Contrast sensitivity was defined as 100 divided by the lowest percentage contrast yielding a threshold response. The initial stimulus in visual acuity measurements was set as 0.200 cyc/deg sinusoidal pattern with a fixed 100% contrast. The initial pattern in contrast sensitivity measurements was set as 100% contrast at a fixed spatial frequency (0.128 cyc/deg). All patterns were presented at a speed of 12 deg/s with the mean luminance of 70 cd/m². 

Electroretinogram Examination

Animals were kept in darkness for 12 hours before recording. Mice were anesthetized with an intraperitoneal injection of ketamine (80 mg/kg) and chlorpromazine (at 15 mg/kg; Shengda Animal Pharmaceutical, Jilin, China), and then transferred to a heating pad. Their pupils were dilated with 1% atropine and 2.5% phenylephrine hydrochloride eye drops (Xing Qi, Shenyang, China). RETIport system (Roland Consult, Havel, Germany) with custom-made chloride silver electrodes were used for ERG recording. ERG responses were recorded through corneal electrodes, with the reference electrode placed in the shaven skin of the cheek, and a ground electrode clipped to the tail. Scotopic ERG responses were recorded at 0.5 log cd-s/m² intensity and the interstimulus interval was set as 30 seconds. Subsequently, animals were then light adapted for 10 minutes in the presence of 30 cd/m² background light. Photopic ERG responses were recorded at 1.4 log cd-s/m² and the interstimulus interval was set as 0.4 seconds. Recorded responses were filtered with high- and low-pass filters at 100 and 300 Hz, respectively. In total 60 photopic responses and 10 scotopic responses were collected and averaged for analysis. For waveform analysis, the amplitude of the a-wave was defined as the distance from baseline to a-wave trough. Amplitudes of the b-waves were measured from the maximum negative to positive peaks of the recorded responses. 

Multielectrode Array Recording

Mice were euthanized under dim red light and their retinas were gently removed from the eyecup. Retina specimens were transferred into the recording chamber with the ganglion cell layer facing the MEA biochip. The optic nerve head (ONH) of retinal patch was fixed to the middle of the electrode array. Retinal specimens were immersed in oxygenated Ringer's solution (containing the following [in mM]: 124 NaCl, 5 KCl, 25 NaHCO₃, 2.5 CaCl₂, 1.15 MgSO₄, 1.15 KH₂PO₄, and 10 D-glucose) for 1 hour to adapt them to the liquid environment. Light responses were elicited by the white light–emitting diodes. White light (at a mean photonic intensity of 850 mcd·s/m²) was projected onto the retinal surface via a lens focus system. MED-64 system (Alpha Med Sciences, Osaka, Japan) was used to detect the analog extracellular response. The electrode array included 64 electrodes, arranged in an 8 × 8 layout with 100 μm for space configuration. Recorded neuronal signals were AC amplified, and then stored in a compatible computer for subsequent off-line software analysis. For spike detection, a band pass filter (100–3000 Hz) was used to wipe off field potentials. Off-line software (Neuroexplorer; Nex Technologies, Madison, AL, USA) was used to sort out the candidate spike waveforms. The value of regional response was averaged across individual responses from each recording channel belonging to that region. Peristimulus time histograms (PSTHs) were smoothed by using a Gaussian kernel to analyze the ON and OFF responses. RGCs were categorized according to the PSTHs and the raster plots of each unit. 

Spectral-Domain Optical Coherence Tomography (SD-OCT)

Anesthetized animals were transferred to the recording plane of an ultrahigh-resolution instrument (Bioptigen, Durham, NC, USA). Methylcellulose lubricant (Allergan, Inc., Dublin, Ireland) was applied on the corneas, and the probe was positioned near the cornea until the retinal image appeared on the screen. A corresponding box was then focused on the ONH for orientation. Eight measurements at the same distance (0.3 mm) from the edge of the ONH on either side were executed. Three hundred linear B-scans were obtained, and 30 averaged images were captured to achieve a better resolution. Subsequently, the SD-OCT cross-sectional images were analyzed with the InVivoVueTM DIVER 2.4 software (Bioptigen). 

Histologic and Immunohistochemical Analysis

Mice were euthanized and their eyes were quickly removed for fixation in buffered 4% paraformaldehyde (Dulbecco's PBS; Mediatech, Inc., Herndon, VA, USA) for 6 hours. A hole was made in the nasal ora serrata with a 25-G needle to facilitate entry of fixative into the eyeball and for orientation purposes. Eyecups were then rinsed in 0.01 M PBS (pH 7.4), dehydrated in a graded ethanol series, and embedded in paraffin wax (polyethylene glycol 400 disterate; Polysciences, Warrington, PA, USA). Six retinal sections (thickness, 5 μm) cut from the ONH were processed with hematoxylin-eosin staining and were evaluated under the microscope. Thickness analysis was performed at 250-μm intervals along the vertically superior-inferior axis by using the Image-Pro Plus software (Media Cybernetics, Silver Spring, MD, USA) Three sections from each eye were randomly chosen for morphometric analysis. Retinal flat mount preparations were generated by first removing the ONH and then carefully separating the sensory retina from the eyecup. For immunohistochemistry, retinal specimens were rinsed in 0.01 M PBS, permeabilized in 0.3% Triton X-100, and blocked in 3% bovine serum albumin for 1 hour at room temperature. Subsequently, they were incubated with primary antibodies directed against S-cone opsin (1:400, AB5407; Millipore, Bedford, MA, USA), or M-cone opsin (1:400, AB5405; Millipore) overnight at 4°C. After washing with PBS, the retinal specimens were incubated with the Cy3-conjugated anti-rabbit IgG (1:400, BA1032; Jackson ImmunoResearch Laboratories, West Grove, PA, USA) for 1 hour. The inner segments of cone photoreceptors were stained with a peanut agglutinin (PNA) conjugated to Alexa Fluor 488 (1:200, L21409; Invitrogen, Carlsbad, CA, USA) overnight at 4°C. PNA binds preferentially to domains of the interphotoreceptor matrix sheath associated with cone inner segments.^25,26 Cell nuclei were labeled by incubating the retinal specimen with 10 μg/mL 4′,6 -diamidino-2-phenylindole-dihydrochloride (DAPI, D9542; Sigma-Aldrich, Natick, MA, USA) for 1 hour at room temperature. Retinal specimens were rinsed and mounted with antifade Vectashield Mounting Medium (H-1200; Vector Labs, Burlingame, CA, USA). Immunofluorescence images were taken with a fluorescence microscope (LSM 510 META; Zeiss, Thornwood, NY, USA) under identical digital exposure settings. Labeled cone cells were counted from four bins (0.1764 mm²) located 1.0 mm superiorly, temporally, inferiorly, and nasally from the ONH via the Axiovision Rel 4.6 software (Carl Zeiss, Oberkochen, Germany). 

TdT–Deoxyuridine Triphosphate Terminal Nick-End Labeling (TUNEL) Assay

Retinal sections were stained with an in situ cell death detection POD Kit (Roche, Mannheim, Germany) according to manufacturer's protocol.²⁷ TUNEL-labeled sections were counterstained with DAPI, mounted on slides, and then visualized with confocal microscopy (LSM510, Zeiss). Apoptotic index (AI) of the outer nuclear layer (ONL) was calculated on the basis of cell numbers (number of TUNEL-positive nuclei/total number of photoreceptor cell nuclei × 100). 

Western Blot Analysis

Retinal tissue was cut into pieces and homogenized in buffer containing 0.23 mol sucrose, 2 mmol EDTA, 5 mmol Tris-HCl (pH 7.5), and 0.1 mmol phenylmethylsulfonyl fluoride. After centrifugation, aliquot extracts containing equal amounts of protein (20 μg) were electrophoresed, transferred, and probed with primary antibody against HO-1 (1:1000, ab13248; Abcam, Cambridge, UK). The membrane was washed with PBS and then incubated with horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG antibody (1:1000, BA1054; Santa Cruz Biotechnology, Santa Cruz, CA, USA). Bands were visualized by using an enhanced chemiluminescence detection system (Super Signal ECL kit West Pico; Pierce, Rockford, IL, USA). 

Quantitative Reverse Transcription–Polymerase Chain Reaction (qRT-PCR)

Total RNA was extracted from retinal tissues with TRIzol reagent (Invitrogen, Gaithersburg, MD, USA) and cDNA was synthesized by using μMACS DNA Synthesis kit (Miltenyi Biotech GmbH, Bergisch-Gladbach, Germany). All quantitative PCR reactions were performed via a real-time CFX96 Touch PCR detection system (Bio-Rad Laboratories, Reinach, Switzerland). The primers used in qRT-PCR were as follows: Bax: 5′-AGCTCTGAACAGATCATGAAGACA-3′ (forward) and 5′-CTCCATGTTGTTGTCCAGTTCATC3′ (reverse); Bcl-2: 5′-GGACAACATCGCTCTGTGGATGA-3′ (forward) and 5′-CAGAGACAGCCAGGAGAAATCAA-3′ (reverse); caspase-3: 5′-TGTCGATGCAGCTAACC-3′ (forward) and 5′-GGCCTCCACTGGTATCTTCTG-3′ (reverse); Calpain-2: 5′-CCCCAGTTCATTATTGGAGG-3′ (forward) and 5′-GCCAGGATTTCCTCATTCAA-3′ (reverse). The relative expression levels were normalized and quantified to obtain the ΔΔCT values (DATA assist Software v2.2; Applied Biosystems, Foster City, CA, USA). 

Intravitreal Injection of HO-1 siRNA

HO-1 siRNA was designed and synthesized according to a standard commercial protocol (Boao Biotechnology, Beijing, China). Primer strands used for HO-1 were as follows: 5′-TCTTGGCTGGCTTCCTTACC-3′ (forward) and 5′-GGATGTGCTTTTCGTTGGGG-3′ (reverse). Intravitreal injection of HO-1 siRNA followed the previously described method.^18,28 Briefly, HO-1 siRNA was centrifuged at 825g for 5 minutes before injection and then dissolved in 5 μL PBS. Mice were anesthetized and their pupils were dilated with 1% atropine and 2.5% phenylephrine hydrochloride eye drops (Xing Qi). The syringe needle of a Hamilton microinjector (Hamilton Company, Reno, NV, USA) was inserted 3 mm above the temporal corneoscleral limbus into the vitreous cavity and was stopped before the retinal surface. Two microliters of siRNA or vehicle (PBS) was pushed into the vitreous cavity slowly. The ofloxacin eye ointment (Xing Qi) was applied immediately after intravitreal injection to avoid infection. Intravitreal injection was performed 2 hours after CoPP administration. 

Determination of Antioxidant and Oxidative Marker Levels

Superoxide dismutase (SOD) activity and malondialdehyde (MDA) concentration were measured by following a previously described method.²⁹ SOD activity was analyzed with the SOD Assay Kit-WST (Jiancheng Biotech Ltd., Nanjing, China). One unit (U) of SOD activity was defined as the amount of enzyme causing half inhibition in the nitroblue tetrazolium reduction rate. SOD activity was expressed as U/mg protein. The MDA concentration was assessed by using a total bile acid colorimetric assay under the guidance of the manufacturer's protocol (Jiancheng Biotech). The manganese-superoxide dismutase (Mn-SOD) activity was measured by using commercially available kits according to the manufacturer's instructions (Beyotime Biotechnology, Shanghai, China). Mn-SOD activity was expressed as U/mg protein. The concentration of 8-OHdG was assessed by competitive ELISA assay (R&D Systems, Minneapolis, MN, USA) and the concentration was expressed as μg/mg DNA. 

Statistical Analysis

Statistical differences between animal groups were processed by using ANOVA analysis followed by Bonferroni's post hoc analysis. All the values are presented as mean ± standard deviation (SD). P < 0.05 was considered statistically significant. 

Results

HO-1 Expression in the Retinas of MNU-Administered Mice

Three days after MNU administration (at P3), the HO-1 protein level in the retinas of the MNU group was significantly higher than in the normal controls (P < 0.01; n = 10; Fig. 2). On the other hand, the HO-1 protein level in the retinas of the MNU+CoPP group was significantly higher than in the MNU group (P < 0.01; n = 10). The HO-1 protein level in the retinas of the MNU+CoPP group was also significantly higher than in normal controls (P < 0.01; n = 10). Subsequently at P7, the HO-1 protein level in the retinas of the MNU group was prominently reduced. The HO-1 protein level in the retinas of the MNU group was significantly lower than in normal controls (P < 0.01; n = 10). On the other hand, the HO-1 protein level in the retinas of the MNU+CoPP group was significantly higher than in normal controls (P < 0.01; n = 10). These results suggested that intraperitoneal injection of CoPP could enhance HO-1 expression in the retinas of MNU-administered mice. 

Figure 2

 

The HO-1 protein level was significantly higher in the retinas of MNU group than in the normal control at P3. The HO-1 protein level was significantly higher in the retinas of MNU+CoPP group than that in the MNU group. Subsequently, the HO-1 protein level in the retinas of MNU group reduced prominently at P7. The HO-1 protein level in the MNU+CoPP group was significantly higher than that of the normal controls, suggesting that CoPP could induce HO-1 overexpression in mice retinas (ANOVA analysis followed by Bonferroni's post hoc analysis was performed, ^#P < 0.01, for differences between groups; n = 10).

CoPP Treatment Alleviated Photoreceptor Loss in Degenerative Retinas

OCT images showed that the normal controls had thicker retinas than mice in the MNU group. The ONL in the retinas of the MNU group was indiscernible. Moreover, the mice in the MNU+CoPP group had relatively intact retinal architecture. Quantification analysis showed that the retinal thickness averaged 0.152 ± 0.012 mm in the MNU group versus 0.156 ± 0.013 mm in the MNU+vehicle group (P > 0.05; n = 10). The average retinal thickness of the MNU+CoPP group was 0.181 ± 0.013 mm, which was significantly larger than that of the MNU+vehicle group (P < 0.01; n = 10). These findings suggested that CoPP was effective in ameliorating the retinal damages in MNU-administered mice. On closer inspection, the retinal sections were visualized microscopically. The retinas of the normal controls were highly organized, whereas the retinal structures of the MNU group were severely disrupted (Fig. 3B). In agreement with the OCT data, the ONL in the retinal sections of the MNU group was not discernible. On the other hand, the ONL in the retinal sections of the MNU+CoPP group was efficiently preserved. The average ONL thickness in the MNU+CoPP group of mice was 20.92 ± 5.66 vs. 1.69 ± 1.06 mm for the MNU+ vehicle group (P < 0.01; n = 10). The mice in the MNU+CoPP group maintained a substantial proportion (∼56.9%) of ONL as compared with the normal controls. 

Figure 3

 

(A) OCT images showed that the retinas of normal controls were significantly thicker than those of the MNU group. Moreover, the mice in the MNU+CoPP group had more intact retinal architecture than did the MNU group, suggesting that the CoPP was effective to preserve the morphology of degenerative retinas. (B) The ONL in the retinal sections of the MNU group was not discernible, while the ONL in the retinal sections of the MNU+CoPP group was efficiently preserved. The average ONL thickness was significantly larger in mice in the MNU+CoPP group than in the MNU group. (C) Numerous TUNEL-labeled cells were found in the ONL of the MNU group. The AI of the MNU+CoPP group was significantly smaller than that of the MNU group, suggesting that CoPP could alleviate photoreceptor apoptosis in MNU-administered mice (ANOVA analysis followed by Bonferroni's post hoc analysis was performed, ^#P < 0.01, for differences between groups; n = 10).

TUNEL labeling was performed in the retinal sections to evaluate the apoptosis status. TUNEL-labeled cells were rarely detected in the retinal section of normal controls (Fig. 3C). However, numerous TUNEL-labeled cells were found in the ONL of the MNU group, suggesting that MNU toxicity induced pervasive photoreceptor apoptosis. Massive TUNEL-labeled cells were also found in the ONL of the MNU+vehicle group. Conversely, the TUNEL-labeled cells in the MNU+ CoPP group were markedly fewer than in the MNU+vehicle group. Quantification analysis showed that the AI of the MNU group was significantly larger than that of the normal controls (P < 0.01; n = 10). The AI in the MNU+vehicle group was not significantly different from that in the MNU group (P > 0.05; n = 10). On the other hand, the AI of the MNU+CoPP group was significantly smaller than that of the MNU group (P < 0.01; n = 10). These findings suggested that CoPP could alleviate photoreceptor apoptosis in MNU-administered mice. Furthermore, the CoPP-induced beneficial effects on the retinal morphology were reversed by adjunctive HO-1 siRNA administration, as evidenced by OCT examination (Fig. 4A) and histologic analysis (Fig. 4B). Moreover, the CoPP-induced inhibitive effects on photoreceptor apoptosis were abolished by adjunctive HO-1 siRNA administration. These findings suggest that HO-1 should be a molecular mediator underlying the antiapoptotic effects (Fig. 4C). 

Figure 4

 

(A) OCT examination showed that the CoPP-induced beneficial effects on the retinal morphology could be abolished by adjunctive HO-1 siRNA administration. (B) Histologic analysis showed that the CoPP-induced beneficial effects on the photoreceptors were mitigated by HO-1 siRNA administration. (C) Massive TUNEL-labeled cells were found in the retinal sections of the MNU+CoPP+siRNA group, suggesting that HO-1 should be responsible for the antiapoptotic effects of CoPP (ANOVA analysis followed by Bonferroni's post hoc analysis was performed, ^#P < 0.01, for differences between groups; n = 10).

CoPP Treatment Enhanced Photoreceptor Responsiveness in Degenerative Retinas

Representative ERG waveforms of each animal group are shown in Figure 5A. Both the scotopic or photopic ERG waveforms were undetectable in the MNU group, suggesting that the visual function of these mice was damaged severely. By contrast, the ERG responsiveness of the MNU+CoPP group showed less deterioration. The scotopic and photopic b-wave amplitudes in the MNU+Copp group were significantly larger than those of the MNU group (P < 0.01; n = 10; Figs. 5B, 5C). Comparison analysis showed that the scotopic and photopic b-wave amplitudes in the MNU+CoPP group were ∼53.5% and ∼56.3% of the normal controls, respectively. The scotopic and photopic a-wave amplitudes in the MNU+CoPP group were significantly larger than those of the MNU+vehicle group (P < 0.01; n = 10; Figs. 5D, 5E). The b/a-ratio represents the efficiency of signal transmission from the photoreceptors to bipolar cells. As shown in Figures 5F and 5G, the b/a ratio of the MNU+CoPP group was significantly larger than that of the MNU group (P < 0.01; n = 10). However, the mice in the MNU+vehicle group showed no significant improvement in the b/a ratio (P > 0.05 versus MNU group; n = 10). These data suggested that ERG impairments in the MNU-administered mice could be partially ameliorated by CoPP treatment. Furthermore, the CoPP-induced protective effects on the electrophysiological response were abolished by adjunctive HO-1 siRNA administration (Fig. 6). 

Figure 5

 

(A) Representative ERG waveforms of each animal group. The visual function of the MNU group was severely impaired, since their scotopic or photopic ERG waveforms were undetectable. (B, C) The scotopic and photopic b-wave amplitudes were significantly smaller in the MNU group than in the normal controls. The scotopic and photopic b-wave amplitudes were significantly larger in the MNU+Copp group than in the MNU+vehicle group. (D, E) The scotopic and photopic a-wave amplitudes in the MNU+CoPP group were significantly larger than those in the MNU+vehicle group. (F, G) The b/a ratio of the MNU+CoPP group was significantly larger than that of the MNU group. These data suggested that CoPP treatment could partially ameliorate the ERG impairments in MNU-administered mice (ANOVA analysis followed by Bonferroni's post hoc an analysis was performed, ^#P < 0.01, for differences between groups; n = 10). The b/a ratio of the MNU+CoPP group was significantly larger than that of the MNU group (P < 0.01; n = 10).

Figure 6

 

The CoPP-induced protective effects on the ERGs were abolished by adjunctive HO-1 siRNA administration, suggesting that the mechanism underlying the CoPP-induced protection should be ascribed to HO-1 activity. (A) Representative ERG waveforms, the scotopic (B) and photopic (C) b-wave amplitudes, the scotopic (D) and photopic (E) a-wave amplitudes, the scotopic (F) and photopic (G) b/a ratio (ANOVA analysis followed by Bonferroni's post hoc analysis was performed, #P < 0.01, for differences between groups; n = 10).

To determine whether the improved electrophysiological responsiveness could lead to vision improvements in a behavioral sense, we tested these experimental animals by the vision-guided optokinetic tests (Fig. 7A). Virtual cylinders were projected on the wall of the box. Figure 7B shows a typical track response of the mouse when the virtual cylinders are turning in clockwise direction. The mice in the MNU group responded poorly to raster stimulus. Quantification analysis showed that the visual acuity and contrast sensitivity in the MNU group were both significantly smaller than for the normal controls (P < 0.01, n = 10; Figs. 7C, 7D). The mice in the MNU+vehicle group showed no significant improvement in the optokinetic tests (P > 0.05 versus MNU group; n = 10). Conversely, the visual acuity and contrast sensitivity were both significantly larger in the MNU+CoPP group than in the MNU group (P < 0.01; n = 10). Furthermore, HO-1 siRNA administration blocked the CoPP-induced protective effects on the optokinetic performance. The visual acuity and contrast sensitivity of the MNU+Copp+siRNA group were not significantly different from those in the MNU group (P > 0.05; n = 10; Figs. 7E, 7F). 

Figure 7

 

(A) Animals were subjected to vision-guided optokinetic tests. (B) Mice showed a typical track response when the virtual cylinders are turning in a clockwise direction. The mice in the MNU group responded poorly to raster stimulus. The visual acuity (C) and contrast sensitivity (D) in the MNU group were both significantly smaller than in the normal controls. The mice in the MNU+vehicle group showed no significant improvement in the optokinetic tests. Conversely, the visual acuity and contrast sensitivity were both significantly larger in the MNU+CoPP group than in the MNU group. Adjunctive HO-1 siRNA administration abolished the CoPP-induced protective effects on optokinetic performance. Visual acuity (E) and contrast sensitivity (F) of the MNU+CoPP+siRNA group were not significantly different from those in the MNU group (ANOVA analysis followed by Bonferroni's post hoc analysis was performed, ^#P < 0.01, for differences between groups; n = 10).

CoPP Rescued the Cone Photoreceptors in Degenerative Retinas

Typically, cone photoreceptors only account for ∼3% of the ONL in mouse retina, and the remainder of this layer is composed of rod photoreceptors.³⁰ To determine the vitality of cone photoreceptors, we performed immunostaining experiments on the retinal sections and flat mounts (Fig. 8). Intense PNA fluorescence was found at the inner segments of the normal controls. However, the PNA fluorescence in the retinal sections of the MNU-administered group was extremely faint. The MNU+vehicle group showed no significant improvement in the PNA fluorescence. Conversely, a substantial proportion of PNA fluorescence was retained in the retinal sections of the MNU+CoPP group. On closer inspection, the PNA fluorescence in the retinal flat mounts was examined under a fluorescence microscope. The PNA-positive cell count of the MNU group was significantly smaller than that of normal controls (P < 0.01; n = 10; Table). The retinal flat mounts of the MNU+CoPP group showed fairly well-preserved PNA fluorescence. The PNA-positive cell count of the MNU+CoPP group was significantly larger than that of the MNU+vehicle group (P < 0.01; n = 10), suggesting that CoPP treatment led to a significant improvement of cone vitality. In the retinal flat mounts of the MNU+CoPP group, the PNA-positive cell count of the inferior-nasal (IN) quadrant was the smallest among all the retinal quadrants. Conversely, the PNA-positive cell count of the superior-temporal (ST) quadrant was the largest, suggesting that cone photoreceptors in this region were preferentially preserved. 

Figure 8

 

Immunostaining assay was performed on the retinal specimens. PNA fluorescence in the retinal sections of the MNU-administered group was extremely faint. On the other hand, a substantial proportion of PNA fluorescence was retained in the retinal sections of the MNU+CoPP group. The retinal flat mounts of the MNU+CoPP group showed fairly well-preserved PNA fluorescence. Both S-opsin and M-opsin staining were lost in the retinas of the MNU group. By contrast, evident S-opsin and M-opsin staining were preserved in the ONL of the MNU+CoPP group (green: PNA, red: cone opsin, blue: DAPI; ANOVA analysis followed by Bonferroni's post hoc analysis was performed, ^#P < 0.01, for differences between groups; n = 10).

Table

 

Cell Counts in Different Quadrants of Retina Whole Mounts

 

Subsequently, the vitality of different cone populations was examined by using opsin-specific antibodies. Neither S-opsin nor M-opsin staining was found in the retinal sections of the MNU group. By contrast, evident S-opsin and M-opsin staining was preserved in the MNU+CoPP group, although in decayed manner relative to the normal controls. The S- and M-cone opsin-positive cell counts of the MNU+CoPP group were both significantly larger than those of the MNU group (P < 0.01; n = 10). Comparison analysis showed that the S-opsin–positive cell count in the MNU+CoPP group was 41.9% of the normal controls, while the M-opsin–positive cell count was 48.5% of the normal controls. In the MNU+CoPP group, the S- and M-cone opsin-positive cells were distributed throughout the retinal flat mount. However, the distribution was not homogeneous: most of the M-opsin–positive cells were found in the ST quadrant and the fewest in the IN quadrant. On the other hand, most of the S-opsin–positive cells were found in the IN quadrants and the fewest in the ST quadrant. These findings suggested that M-cone photoreceptors in the ST quadrant and S-cone photoreceptors in the IN quadrant benefited most from the CoPP treatment. 

CoPP Induced Effects on the Field Potentials of Degenerative Retinas

In the MEA recording, the electrodes were classified into three categories according to their distances to ONH: the central, the midperipheral, and the peripheral channels (Fig. 9A). Typical waveforms of field potential were recorded in the normal control mice (Fig. 9B). The field potential waveforms of the MNU group were undetectable. Conversely, the field potential waveforms of the MNU+CoPP group were effectively preserved. Quantization analysis showed that the mean amplitude of the MNU group decreased significantly as compared with normal controls (P < 0.01; n = 10; Fig. 9C). The mean amplitude of the MNU+CoPP group was significantly larger than that of the MNU group (P < 0.01; n = 10). In particular, the field potential responses in the MNU+CoPP group were not homogeneously equal: the amplitudes of the peripheral region were significantly larger than those of the midperipheral and central regions (P < 0.01; n = 10). Additionally, the mean amplitude of the midperipheral region was significantly larger than that of the central region (P < 0.01; n = 10). 

Figure 9

 

(A) The electrodes of the MEA were classified into three categories according to their distance to the optic nerve head. (B) Typical waveforms of field potential were recorded in the normal controls. The field potential waveforms of the MNU group were undetectable. Conversely, the field potential waveforms in the MNU+CoPP group were effectively preserved. (C) The mean amplitude of field potentials in the MNU+CoPP group was significantly larger than that of the MNU group. In the MNU+CoPP group, the field potential amplitude of the peripheral region was significantly larger than that of the midperipheral and central regions. (D) The spontaneous firing spikes of RGCs were detected by the MEA system. (E) The spontaneous firing rate of the MNU+CoPP group was significantly higher than that of the normal controls, but was significantly lower than that of the MNU group. (F) Six categories of RGCs were isolated according to their responsive PSTHs. (G) The total firing rate of the MNU group decreased significantly when compared with normal controls. The total firing rate was significantly higher in the MNU+CoPP group than in the MNU group. In the MNU+CoPP group, OFF pathway was more efficiently preserved than the ON pathway (ANOVA analysis followed by Bonferroni's post hoc analysis was performed, ^#P < 0.01, for differences between groups; n = 10).

CoPP-Induced Effects on the Visual Signal Pathways

Spontaneous firing spikes from RGCs were detected by the MEA system (Fig. 9D). The spontaneous firing rate in the MNU was significantly higher than that of normal controls (P < 0.01; n = 10; Fig. 9E). The spontaneous firing rate in the MNU+CoPP group was significantly lower than that of the MNU group (P < 0.01; n = 10). Moreover, the light-induced firing spikes were extracted by off-line sorter. Six categories of RGCs were isolated according to their responsive PSTHs (Fig. 9F). The total firing rate of the MNU group decreased significantly as compared with normal controls (P < 0.01; n = 10; Fig. 9G). The total firing rate was significantly higher in the MNU+CoPP group than that in the MNU group. Both light-induced ON and OFF responses were retained in the MNU+CoPP group, indicating that the basic configurations of visual signal pathways were efficiently preserved in these mice. On closer inspection, the OFF pathway was more efficiently preserved than the ON pathway. The ON response intensity was 39.5% of the normal controls, while the OFF response intensity was 69.3% of the normal controls. 

Mechanisms Underlying the CoPP-Induced Protective Effects

The qRT-PCR was performed to analyze mRNA levels of apoptotic factors in retinal tissues. The mRNA levels of caspase-3, calpain-2, and Bax in the MNU group were not significantly different from those in the MNU+vehicle group (P > 0.05; n = 10;). On the other hand, the mRNA levels of caspase-3, calpain-2, and Bax in the MNU+CoPP group were significantly downregulated as compared with the MNU group (P < 0.01; n = 10; Fig. 10). The mRNA level of Bcl-2 in the MNU+CoPP group was significantly upregulated as compared with the MNU group (P < 0.05; n = 10). These findings suggested that the antiapoptotic mechanism was, at least partly, responsible for the CoPP-induced protection. Moreover, the retinal levels of MDA (a stable metabolite of lipid peroxidation) and 8-OHdG (an indicator of oxidative DNA damage) in the MNU group were significantly higher than those in the normal control group (P < 0.01; n = 10). The retinal MDA concentration was 2.016 ± 0.299 nmol/mg in the MNU+CoPP group versus 4.723 ± 0683 nmol/mg in the MNU+vehicle group (P < 0.01; n = 10). The retinal 8-OHdG concentration was 99.71 ±7.29 μg/mg in the MNU+CoPP group versus 150.26 ±12.36 μg/mg in the MNU+vehicle group (P < 0.01; n = 10). The retinal level of SOD, an endogenous antioxidant, was 169.210 ± 20.006 U/mg in the MNU+CoPP group versus 70.879 ± 13.385 U/mg in the MNU+vehicle group (P < 0.01; n = 10). The retinal level of Mn-SOD, a mitochondrial protein with ROS-scavenging potency, was 33.08 ± 4.45 U/mg in the MNU+CoPP group versus 18.02 ± 4.26 U/mg in the MNU+vehicle group (P < 0.01; n = 10). These findings suggested that the CoPP treatment could alter the retinal oxidation status of degenerative retinas. 

Figure 10

 

The mRNA levels of caspase-3, calpain-2, and Bax in the MNU+CoPP group decreased significantly when compared with the MNU group. The mRNA level of Bcl-2 in the MNU+CoPP group was significantly upregulated when compared with the MNU group. Moreover, the retinal levels of MDA and 8-OHdG in the MNU group were significantly higher than those of the MNU+CoPP group. The retinal level of SOD in the MNU group was significantly lower than that of the MNU+CoPP group. The retinal level of Mn-SOD in the MNU group was also significantly lower than that of the MNU+CoPP group (ANOVA analysis followed by Bonferroni's post hoc analysis was performed, ^#P < 0.01, for differences between groups; n = 10).

Discussion

In this study, we showed that HO-1 is involved in the retinal degeneration of MNU-administered mice. The retinal HO-1 expression increases after MNU administration (at P3). Subsequently, retinal HO-1 expression decreases significantly at P7. The ERG waveform of MNU-administered mice is undetectable, and the retinal structure is severely disrupted. These characteristics are consistent in broad terms with the clinical settings of RP.¹³ Moreover, we demonstrated that a single intraperitoneal injection of CoPP can enhance the HO-1 expression in the retinas of MNU-administered mice. The CoPP-treated mice had more intact ONL and better ERG responsiveness than did the vehicle-treated controls. CoPP treatment also alleviated the photoreceptor apoptosis, mitigated the level of oxidative stress, and enhanced the activities of antioxidant enzymes. In particular, adjunctive HO-1 siRNA administration can abolish the CoPP-induced protective effects. These findings suggest that HO-1 expression is downstream of MNU administration and may be sufficient to provide some protection, but do not show that it is necessary. 

Intraperitoneal injection of MNU is a safe and stable approach to induce photoreceptor degeneration. MNU is a lethal toxicant with carcinogenic effects. A dramatic increase in plasmatic level would cause tumor formation in several organs and even animal death.^31,32 In ophthalmologic experiments, MNU is delivered via the intraperitoneal route to minimize the number of animals and their suffering.^13,33 On the other hand, CoPP is delivered via intravenous injection to ensure relatively a higher concentration of therapeutic molecules in blood plasma. Typically, MNU-induced rod photoreceptor degeneration is achieved within 7 days.³⁴ However, cone photoreceptors in the MNU-administered mice die secondarily to the rod photoreceptors. Some cone photoreceptors in the ST quadrant are resistant to MNU toxicity, and the damage in this region is relatively slight.^23,35 It has been shown that the scotopic ERG response is more sensitive to the MNU toxicity than the photopic ERG response: the scotopic ERG waveform of the MNU-treated mice disappears at P7, while a minimum of the photopic ERG waveform is still retained. The residual photopic waveform eventually disappears at P9, and MNU eliminates all the ERGs of the administered mice.²³ Although the pathologic outcome of the MNU-induced photoreceptor death resembles RP, the mechanistic underpinnings are very different. The MNU-induced photoreceptor death results from alkylation of DNA, depending on the action of alkyladenine DNA glycosylase.¹² On the other hand, RP is caused by the mutations in rod-related genes, which are crucial for retinal function.⁵ Additionally, the progression rate between the MNU-induced photoreceptor degeneration and that of hereditary RP is very different. MNU-induced photoreceptor degeneration is rapidly progressing, while RP is a relatively chronic retinopathy with a much longer course.^1,36 Therefore, the progression rate should be considered as an instinctive disadvantage of the MNU-induced models, as it deviates somewhat from the natural history of human RP. 

Cones make up the minority of photoreceptors in the mammalian retina. M- and S-cones are responsible for high visual acuity and color discrimination, respectively.³⁷ In RP patients, loss of rods merely impairs night vision under dim illumination, whereas the loss of cones causes devastation of visual acuity and eventual blindness.^38,39 Herein, we showed that the M- and S-opsin cone photoreceptors in MNU-administered mice are both rescued by CoPP treatment. However, it is not clear whether CoPP treatment saves preferentially cones, or whether cones and rods are protected to a similar extent. Further studies are necessary to quantify the number of cones and rods to distinguish between effects on each photoreceptor populations. Of note, MNU is equally toxic to both rod and cone photoreceptors.³⁵ The observation that cones are also killed by MNU toxicity does not correspond to cone degeneration in RP patients, since demise of this population occurs in a secondary wave of cell death, long after rods have degenerated.^39,40 Regarding the situation in RP patients, it is still disputed whether isolated protection of cones would work, or whether the cones are destined to die anyhow once the rods are lost. These considerations are crucial in any study that proposes the clinical testing of candidate molecules. 

ERG can evaluate the function of different components in visual transmission, with a-wave reflecting the response of photoreceptors and b-wave reflecting the electrophysiological activity of bipolar cells.⁴¹ Our ERG data show that CoPP treatment can preserve partially the photoreceptor responsiveness in MNU-administered mice and enhance the efficiency of signal transmission from the photoreceptors to bipolar cells. MEA technology detects simultaneously the electrophysiological activity of photoreceptor populations and acts as an objective approach to evaluate therapeutic efficiency.⁴² The MEA recording showed that photoreceptors in peripheral retina are rescued more efficiently by CoPP treatment. Moreover, spontaneous RGC hyperactivity occurs in the MNU-administered mice. Spontaneous RGC hyperactivity is detrimental, since it would add undesirable noise to the retinal circuits.^43–45 As the output neurons of the retina, RGCs collect field potentials from presynaptic inputs and convert them into firing spikes.⁴⁶ The success of any therapeutic strategy ultimately depends on the functional recovery of RGCs.⁴⁷ CoPP treatment can alter the electrophysiological properties of RGCs, and alleviate the RGC hyperactivity in MNU-administered mice, thereby improving the signaling efficiency.^48,49 Furthermore, the intrinsic balance between ON and OFF pathways is disturbed, and the OFF pathway would dominate the visual signal transmission in the MNU-administered mice. Since the inner retinal morphology of MNU-administered mice is relatively intact, the basis underlying this pathway reorganization might be attributed to the plasticity of retinal circuits.^50,51 

Photoreceptor apoptosis is considered as the shared trait of different RP phenotypes.^34,52 Our TUNEL results suggest that CoPP is sufficient to alleviate the photoreceptor apoptosis in MNU-administered mice. A previous study⁵³ also has shown that HO-1 overexpression inhibits RGC apoptosis in ischemic retinas. CoPP treatment can enhance Bcl-2 expression and adjust the Bcl/Bax ratio toward a net “antiapoptotic” effect.⁵⁴ CoPP treatment also reduces the mRNA level of caspase-3, a classic mediator in apoptosis cascades. Photoreceptors are metabolically active neurons that tightly depend on the mitochondria for survival. Mitochondrial impairment would result in the release of intermembrane space proteins, and the subsequent activation of mitochondrial-dependent apoptosis pathways. Mn-SOD is a mitochondrial protein that can defend against superoxide radicals.⁵⁵ Our findings showed that CoPP treatment can enhance the retinal Mn-SOD level, indicating that HO-1 overexpression conferred beneficial effects on the mitochondria of photoreceptors. The homeostasis between free radical and antioxidant systems is delicately tuned in photoreceptors.^56–58 Photoreceptor loss would mitigate oxygen consumption and exacerbate the oxidative stress in retina. Generally, the retinal antioxidant enzymes can withstand a variety of insults such as hyperoxia, pH disturbances, excitotoxic reactions, and excessive free radicals.⁵⁹ In this study, the retinal SOD level in the MNU-administered mice was enhanced by CoPP, while the retinal levels of MDA and 8-OHdG were downregulated. These findings suggest that CoPP can bolster the endogenous antioxidants and ameliorate oxidative damage in retina. Recently, there is growing consensus that HO-1–mediated protection is correlated with its antioxidant activity. HO-1 can protect the retinal endothelial cells from oxidative/nitrosative stress.⁶⁰ An in vitro study²⁰ has shown that HO-1 overexpression can enhance the antioxidant enzyme activity and suppress ROS generation in human lens epithelial cells. Another in vivo study⁶¹ also has shown that the HO-1 overexpression inhibits ROS production in the ischemic retinas. Researchers propose that the antioxidative capacity of HO-1 should be ascribed to its end products such as ferrous free iron, CO, and biliverdin.^14,62–64 Biliverdin is metabolized further by biliverdin reductase to the bile pigment (bilirubin), a potent antioxidant molecule.⁶⁵ Biological effects of these products are complex, and they may act cooperatively to alleviate photoreceptors from oxidative insults. A previous study⁶⁶ has shown that the intravenous injection of hemin, a porphyrin-containing ferric iron, is also sufficient to alleviate MNU-induced retinal degeneration. Several studies with different cellular injury models have shown that hemin can suppress ROS production via enhancing HO-1 expression.^67,68 Accordingly, it is highly possible that other potent HO-1 inducers might also be effective in counteracting oxidative stress in degenerative retinas. 

Although the CoPP-treated mice showed improvements over those in the MNU group, their visual function and retinal morphology were significantly worse than those of the normal controls. These findings suggest that CoPP treatment slows the progression of retinal degeneration but does not block completely the photoreceptor loss. CoPP treatment can be used in conjunction with other therapeutic strategies (e.g., gene therapy) to produce satisfactory protective effects on the retina. CoPP would act as a temporizing measure until a genetic defect can be found and a specific therapy devised. The natural course of RP patients is highly variable, as the tremendous heterogeneity lies in the initiating mutation.^1,4,38 These characteristics make RP inherently intractable. To have clinical relevance, testing the CoPP-induced benefits over time in other well-characterized RP models would be a logical next step. 

In summary, HO-1 is involved in MNU-induced photoreceptor degeneration. An intravenous injection of CoPP can enhance retinal HO-1 expression, thereby alleviating the photoreceptor degeneration in MNU-administered mice. CoPP affords these protective effects by inhibiting apoptosis and mitigating oxidative stress. These data highlight the possibility that CoPP may act as a therapeutic molecular agent against photoreceptor degeneration. Further refinements of these findings may lead to the discovery of a new therapeutic strategy for RP. 

Acknowledgments

Supported by the National Key Research and Development Plan of China (No. 2018YFA01073003) and the National Natural Science Foundation of China (No. 81600767). 

Disclosure: Y. Tao, None; L. Cai, None; D. Zhou, None; C. Wang, None; Z. Ma, None; X. Dong, None; G. Peng, None