RNA interference to reduce Egr1 expression in rods delays retinal degeneration in a model of retinitis pigmentosa

Animals

Animal experimentation was approved by the Cantonal Veterinary Office of Zürich, Switzerland (license numbers: ZH019/2019 and ZH105/2022) and adhered to the ARVO statement for the use of Animals in Ophthalmic and Vision Research. Wild-type (C57BL6/J, 129S6) and transgenic mice (Rho^P23H/P23H, Rho^P23H/+) were housed at the Laboratory Animal Services Centre (LASC) of the University of Zürich with a 14/10 h light/dark cycle and an average light intensity of 60–150 lx at cage level. Rho^P23H/P23H mice (strain # 017,628, Jackson laboratory, Bar Harbor, ME) were bred with C57BL6/J wild-type mice to generate Rho^P23H/+ mice. All mice used were tested negative for the rd8 mutation in the Crumbs 1 (Crb1) gene [53,54].

Lentiviral-mediated generation of shEgr1 expressing stable cell lines, PMA treatment, and transgenic Egr1 overexpression in 661W

The photoreceptor-derived 661W cell line [55] and HEK293T (ATCC CRL-3216) cells were cultured in DMEM, 10% heat-inactivated FBS (Gibco, Thermo Fisher Scientific, Waltham, MA) and 1% penicillin-streptomycin (Gibco, Thermo Fisher Scientific) at 37 °C and 5% CO₂. For gene silencing, lentiviral-pseudotyped particles were used to generate stable cell lines expressing shRNAs against Egr1 (Sigma-Aldrich, St. Louis, MO; Table 1) and a non-targeting scrambled shRNA (SHC002, Sigma-Aldrich; Table 1). To prepare lentiviral particles, HEK293T cells were cultured in 75 cm² flasks and co-transfected with shRNAs in combination with the ViraPower lentiviral expression vector system and Lipofectamin 3000 (Invitrogen, Thermo Fisher Scientific) in serum-free medium. The serum-free medium was replaced the next day with complete medium. The culture supernatants containing the lentivirus particles were collected 48–72 h post-transfection, cleared by centrifugation at 200 g for 5 min, and filtered through a 0.45 µm pore size filter (Merck & Co., Schaffhausen, Switzerland). Filtrates were applied in a 1:1 mixture with medium containing 6 µg/ml polybrene (Sigma-Aldrich) to infect 661W cells. Infected cells were selected with 2 µg/ml puromycin for 2–3 weeks. To induce endogenous expression of Egr1, cells were treated with 10 ng/mL of phorbol 12-myristate 13-acetate (PMA, P1585-P1MG, Sigma-Aldrich) in DMSO (D8418, Sigma-Aldrich) for 1 h. Subsequently, cells were washed once with 1X PBS (Gibco), harvested, and RNA prepared for real-time PCR to investigate gene expression (see below).

To overexpress Egr1, 661W cells were seeded at a density of 50’000 cells/ml in wells of 12-well plates and incubated overnight. Cells were transfected with Lipofectamine 3000 (Invitrogen, Thermo Fischer Scientific) and 1 μg of plasmid DNA containing the overexpression or a control plasmid (see next section) in Opti-MEM (Thermo Fischer Scientific) for 24 or 48 h. Subsequently, cells were washed once with 1X PBS (Gibco), harvested, and RNA was prepared for real-time PCR (see below).

Plasmid cloning and AAV generation

shRNAmir DNA fragments containing shEgr1 or shCtrl sequences were designed using the miR-E backbone following the recommendations by Fellmann [46] and synthesized by Genewiz (Azenta Life Science, South Plainfield, NJ). The shRNA cassettes were cloned in the 3′UTR of GFP. For overexpression, the Egr1 coding sequence from mouse (mEgr1; Addgene plasmid #11729) [56] was connected with a P2A cassette to a GFP coding sequence and cloned downstream of the ubiquitous cytomegalovirus (CMV) promoter in an AAV2 backbone for in vitro experiments (AAV2::CMV-mEgr1-P2A-eGFP-pA-WPRE). A pcDNA3-CMV-eGFP plasmid (Clontech, Addgene #2487) was used as control. To achieve cell specificity in vivo, constructs were placed under control of the mOP (mouse opsin) promoter [52] to transduce rods, and an engineered human OPN1WM (opsin medium wavelength) promoter to target middle wavelength (M) cones [51]. All constructs were cloned into an AAV backbone with mutated type 2 inverted terminal repeats (ITRs), and packaged into the AAV2(QuadYF+TV; 7m8)/2 capsid variant containing five point mutations (Y272F, Y444F, T491V, Y500F, and Y730F) [49] to target rods, and into the AAV2/7m8 capsid [50] to target cones. Due to AAV packaging limitations, the cone-specific overexpression vector did not contain eGFP.

Subretinal injections

Injections were performed at 4 weeks of age as described earlier [57]. Briefly, pupils were dilated with 1% cyclogyl (Alcon Pharmaceuticals, Fribourg, Switzerland) and 5% neosynephrin (Ursapharm Schweiz GmbH, Hünenberg, Switzerland). Mice were anesthetized with a subcutaneous injection of ketamin (85 mg/kg, Pfizer AG, Zürich, Switzerland) / xylazine (10 mg/kg, Elanco Animal Health GmbH, Basel, Switzerland) and placed on a heating pad set to 37 °C during the procedure. Lacrinorm® Carbomerum (Bausch & Laumb Swiss AG, Zug, Switzerland) was applied to keep the eyes moist. The head was held by a stereotactic adaptor (Hugo Sachs Elektronik – Harvard Apparatus GmbH, March–Hugstetten, Germany) and the temporal sclera was punctured with a 30G needle beneath the ora serrata. A 5 µl Hamilton syringe with a 34G blunt-end needle (Hamilton Bonaduz AG, Bonaduz, Switzerland) was placed in a micromanipulator (H. Saur Laborbedarf, Reutligen, Germany) and inserted through the pre-punctured site for the transvitreal subretinal injection of 1 µl AAV solution nasally to the optic nerve. All AAVs were applied at a concentration of 1×10⁹ vg/µL. To visually control the procedure, the injection solution was supplemented with 10% fluorescein (1 mg/ml, Akorn Inc., IL). After injections, 1 drop of viscotears (Bausch+Lomb, Vaughan, Canada) was applied, and anesthesia was reversed with atipazemole (2 mg/kg, Graeub, Bern, Switzerland). Mice were placed on a heating pad until fully awake. Two to three weeks after injection, mice were subjected to fluorescent funduscopy and optical coherence tomography (OCT) to test for transgene expression and to detect potential injection-inflicted tissue damages, such as bleeding or persistent retinal detachment. Such eyes were excluded from the study.

White light damage

To induce retinal degeneration, mice were exposed to high levels of white light as described in [58]. Briefly, 129S6 mice were placed in darkness the evening before the experiment. Thirty min before light exposure, pupils were dilated (as described above) in dim red light. The mice were placed in cages without a grid but lined with aluminum foil (to homogeneously distribute light within the cage), which were positioned under the light-exposure device to reach 13’000 lx at cage level. Exposure lasted for 1 h. After this time, mice were returned to their home cages and kept in darkness for 24 h. Afterwards, mice were housed in the normal light/dark cycle until analysis.

In vivo imaging by funduscopy and optical coherence tomography

Anesthesia and pupil dilation were performed as described above. One drop of 2% methocel (OmniVision AG, Neuhausen, Switzerland) was applied to keep the eyes moist. Fundus images and OCT scans were acquired using the Micron IV system (Phoenix Research Labs, Pleasanton, CA) equipped with filters for imaging green fluorescence.

RNA isolation and real-time PCR

Mice were euthanized with CO₂ inhalation followed by decapitation. Retinas were isolated through a slit in the cornea and snap-frozen in liquid nitrogen. To separate the transduced from the untransduced retinal area, the retina was flattened out and cut in half using the GFP signal as guidance. The two halves were snap-frozen in liquid nitrogen. RNA from tissue or cells was isolated using an RNA isolation kit (NucleoSpin RNA, Macherey-Nagel GmbH & co.KG, Düren, Germany) according to the manufacturer’s instructions, including an on-column DNase digestion step (740,963, Macherey-Nagel GmbH & co.KG). First-strand cDNA synthesis was performed using M-MLV reverse transcriptase (Promega, Dübendorf, Switzerland), 650 ng of total RNA and 20 pmol of oligo-dT primers. Gene expression was analyzed via semiquantitative real-time PCR (QuantStudio 3, Thermo Fisher Scientific) using 10 ng of cDNA template and PowerUp SYBR green Master Mix (Thermo Fisher Scientific). Primer pairs for Egr1 (F: 5′-ACA ACC CTA TGA GCA CCT GAC C-3′; R: 5′-GGC AGA GGA AGA CGA TGA AG-3′), Actb (F: 5′-CAA CGG CTC CGG CAT GTG C-3′; R: 5′-CTC TTG CTC TGG GCC TCG-3′), and Atf3 (F: 5′-ACC TCC TGG GTC ACT GGT ATT TG-3′; R: 5′-TTC TTT CTC GCC GCC TCC TTT TCC-3′) were designed to span large intronic regions and avoid known single nucleotide polymorphisms in the mouse sequence. Normalization was performed with Actb as housekeeping gene, and relative expression was calculated using the comparative threshold cycle method (2^-ΔΔCT) [59].

Determination of ONL thickness and immunofluorescence

To determine the thickness of the ONL, dorsally marked eyes were enucleated and fixed in 2.5% glutaraldehyde at 4 °C overnight. Dissected eyes were post-fixed in 1% osmium tetroxide for 1 h and embedded in Epon812 (Sigma-Aldrich). Nasal-temporal sections (0.5 µm) were cut through the optic nerve head, counterstained with toluidine blue, and analyzed by light microscopy (AxioImager Z2, Carl Zeiss AG, Feldbach, Switzerland). Retinal panoramas were reconstructed from single 10X images using Adobe Photoshop CS6 (Adobe Systems, Inc.). The ONL thickness (from the outer limiting membrane to the outer plexiform layer) was measured every 200 µm starting from the optic nerve head with the Adobe Photoshop CS6 ruler tool.

Eyes for immunofluorescence analysis were marked and fixed in 4% PFA in 0.1 M PBS for 1.5 h at 4 °C. After removal of the lens, eyes were cryo-protected in 30% sucrose (Sigma-Aldrich) in 0.1 M PBS for at least 2 h, embedded and frozen in freezing medium (O.C.T., Leica Biosystems, Nussloch, Germany). Embedded eyes were stored at −80 °C until cut into 12 µm thin sections. Immunolabelling was performed overnight at 4 °C with the following primary antibodies: anti-EGR1 (1:500, MA5–15008, Thermo Fisher Scientific), anti-RHO (1:100, O4886, Sigma-Aldrich), anti-OPN1SW (1:500, sc-14363, Santa Cruz, Dallas, TX) and anti-ARR3 (AB15282, 1:1000, Merck & Co.). Peanut Agglutinin (PNA) lectin from Arachis hypogaea (peanut) conjugated with Alexa Fluor™ 647 (L32460, Thermo Fischer Scientific) was used to label the cone interphotoreceptor matrix [60]. Incubation of primary antibodies and PNA was done in blocking solution containing 3% normal goat serum (Sigma-Aldrich) and 1% Triton X-100 (Sigma-Aldrich) in PBS. The next day, sections were washed, incubated with the appropriate secondary antibodies in blocking solution for 1 h, and with 4’,6-diamidino-2-phenylindole (DAPI, 1:6000) in 0.1 M PBS for 3 min. Sections were mounted with MOWIOL (Sigma-Aldrich) antifade medium and imaged using a fluorescence microscope (AxioImager Z2, Carl Zeiss AG). To count cone cells, retinal panoramas from immunofluorescence images were reconstructed with Adobe Photoshop CS6, and 600 µm-wide sections were selected within the transduced and the untransduced areas of the same retina. Cones (ARR3-positive cells) were counted in these areas using the multi-point tool in ImageJ (Version 1.52a).

Statistical analysis

Statistical analysis was performed using GraphPad Prism (GraphPad Software, San Diego, CA). Significance was tested by an unpaired t test when two independent groups were compared, or a paired t test when the two measurements came from the same animal. When more than 2 groups were compared, a one-way ANOVA followed by a Tukey’s (when all groups were compared to each other) or Dunnett’s (when all groups were compared to a specific group) multiple comparison test was applied. When assessing the influence of two independent variables and their interaction (2×2 factorial design), a two-way ANOVA was applied. Data were considered significantly different with p≤0.05. Only p values <0.05 are shown in the figures. All tests and number of animals/samples are indicated in the figure legends.

Targeting Egr1 for knockdown and transgenic overexpression

To modulate Egr1 expression in vivo, we chose shRNA-based RNA interference (RNAi) and transgenic overexpression methods to decrease or increase endogenous EGR1 levels, respectively. To identify an effective sequence for RNAi, we generated six 661W cell lines that stably expressed different Egr1 shRNAs (Table 1) and determined endogenous Egr1 mRNA levels by real-time PCR. Five out of the six shRNA sequences significantly reduced Egr1 mRNA levels in the cell lines, with sequence 1 being the most effective (78% downregulation; Figure 2A). This cell line was selected for investigating shRNA-mediated Egr1 silencing also under stress conditions. For this purpose, control 661W and 661W-shEgr1 cells were treated with 10 ng/mL PMA, a well known Egr1 activator via the stress-activated MAPK pathways [61], for 1 h and tested for endogenous Egr1 mRNA expression. PMA increased Egr1 levels 3.8-fold in both control and 661W-shEgr1 cells. Importantly, Egr1 levels after PMA treatment were 55% lower in 661W-shEgr1 cells than in control cells (Figure 2B), demonstrating that the shRNA repressed endogenous Egr1 also in stressed cells, which is essential for using this approach in degenerating photoreceptors. Transfection of 661W cells with the overexpression plasmid resulted in a 22.6- and 15.6-fold increase in Egr1 mRNA levels at 24 and 48 h post-transfection, respectively (Figure 2C). The overexpression resulted in functional EGR1 proteins, as evidenced by the 13.6-fold upregulation of the EGR1 target gene Atf3 (activating transcription factor 3 [62,63]) 48 h after transfection (Figure 2D). The small induction of Atf3 expression in control-transfected cells may have been caused by a slight stress induced by the serum-reduced OptiMEM transfection medium [64].

To generate AAVs for the in vivo experiments, we cloned the sequence for knocking down Egr1 (shEgr1) and the control sequence (shCtrl, Table 1) into the miR-E scaffold [46] and placed the cassettes in the 3′-untranslated region (UTR) of GFP (Figure 2E). The AAV construct for overexpression contained the mouse Egr1 (mEgr1) coding sequence connected to eGFP via a P2A site. All constructs were cloned downstream of the rod-specific (mOP [52]) and the cone-specific (OPN1MW [51]) promoter. Due to the limited AAV packaging capacity, the cone-specific overexpression construct did not contain eGFP (Figure 2E, Table 2).

Modulation of Egr1 expression in rods of wild-type mice

To assess the potential of our constructs to modulate Egr1 expression in rods, we injected the AAVs carrying the shEgr1, the mEgr1 or the control sequence under control of the mOP promoter into the subretinal space of 4-week-old wt mice and assessed their transduction and expression profiles. At 8 weeks post-injection (wpi, 12 weeks of age), the injection sites were visible as hyperreflective spots in the fundus images (Figure 3A-C, arrows) and minor disturbances to the retinal layering in OCT scans (blue lines and frames). However, OCT scans through the optic nerve head (red lines and frames) demonstrated that the retinal layers were overall preserved, excluding toxicity of the vectors or the transgenic sequences in normal retinas. Fluorescent funduscopy indicated that approximately 20%–30% of the retinal area was transduced. Immunostaining of retinal sections for rhodopsin (RHO) and imaging of (e)GFP fluorescence (surrogate marker for AAVs) showed that transgene expression was restricted to photoreceptors (Figure 3A-C, lower panels). Since EGR1 is activated upon damage or stress [40,44], it is important to note that injections did not cause widespread EGR1 expression per se, as only few rods transduced with AAV::mOP-ctrl (Figure 3A) co-expressed endogenous EGR1 and GFP. Similarly, only very few EGR1-positive photoreceptor nuclei were detected after injection of AAV::mOP-shEgr1 (Figure 3B). Whether this was due to an efficient downregulation of Egr1 expression by the shRNA or to the absence of stress after the injection, as suggested by the control virus, could not be determined. In contrast, injection of AAV::mOP-mEgr1 resulted in a large number of EGR1-positive cells in the ONL, indicating a strong transgenic expression of Egr1 in rods (Figure 3C). To assess the strength of the overexpression, we used real-time PCR to amplify Egr1 from total retinal RNA isolated from dark-adapted mice two weeks after injection. Even though we used total retinal RNA that was isolated not only from the 20%–30% transduced rods but also from all non-transduced cells in the outer and inner retina, we detected a significant increase in Egr1 expression in retinas injected with AAV::mOP-mEgr1 (Figure 3D). Due to the presence of RNA from untransduced cells in the analysis, the measured twofold increase underestimates the strength of upregulation in transduced cells. Since Egr1 is not, or only very weakly, expressed in healthy rods, we could not evaluate the in vivo efficacy of our shEgr1 sequence in wild-type mice. Therefore, we assessed retinal Egr1 levels in Rho^P23H/+ mice at 12 wpi (16 weeks of age). To increase the specificity of our analysis, we separated the untransduced and transduced halves of the retina using the GFP signal as guidance, then isolated total RNA from each half separately. While Egr1 mRNA levels in the retinas of control-injected mice did not differ between the untransduced and transduced areas, Egr1 was significantly downregulated after injection of AAV::mOP-shEgr1 (Figure 3E). Lastly, we measured the ONL thickness in wild-type mice at 8 wpi to directly address potential retinal toxicity of our AAVs (Figure 3F). The similar ONL thickness between treatments indicated that modulation of Egr1 expression in healthy rods did not cause cellular toxicity. Although significant, the slight increase in ONL thickness (approximately 3 µm) in AAV::mOP-mEgr1 injected mice suggests a minor swelling of the ONL for unknown reasons.

These results demonstrated an effective and specific transduction of rod photoreceptors by the mOP-driven transgenes after AAV-mediated delivery, as shown before [65]. The constructs were able to successfully modify Egr1 expression, with no adverse effects in wt mice.

Egr1 downregulation delays rod degeneration in Rho^P23H/+ mice

To analyze if the rod-specific modulation of Egr1 expression affects photoreceptor degeneration we injected the three viruses into the retina of 4-week-old Rho^P23H/+ mice. Immunofluorescence of retinal cross sections at 12 wpi showed that the degenerating retina expressed the transgenes specifically in the ONL (Figure 4A), similarly to the observation made in wt retinas (Figure 3A-C). Since GFP-positive cells only rarely co-localized with EGR1 in AAV::mOP-shEgr1 injected mice, downregulation of EGR1 in transduced rods of Rho^P23H/+ mice may have been effective. In AAV::mOP-mEgr1 injected retinas, however, GFP and EGR1 showed strong co-localization indicating upregulation of EGR1 (Figure 4A). While fundus imaging and OCT scans did not show obvious differences between the treatments (Figure 4B), we measured the ONL thickness on retinal cross sections at 4, 8 and 12 wpi (Figure 4C). At 4 wpi, the ONL in mice injected with AAV::mOP-mEgr1 was slightly thicker, reflecting the observation already made in wt mice (Figure 3F), while AAV::mOP-shEgr1 was without effect on the ONL thickness. By 8 wpi, the ONL thickness decreased as degeneration progressed in controls and AAV::mOP-mEgr1 but not in AAV::mOP-shEgr1 injected mice. In these latter mice, the ONL thickness was comparable to the thickness at 4 wpi suggesting a protective effect and attenuation of the degeneration by the downregulation of Egr1 in Rho^P23H/+ rods by RNAi. Interestingly, the overexpression of EGR1 did not accelerate the degenerative process. At 12 wpi, the preservation of rods by the interference with Egr1 expression was still evident, as the ONL of AAV::mOP-shEgr1 injected mice was still significantly thicker than the ONL in the controls. The attenuation of degeneration by the AAV::mOP-shEgr1 injections was also evident by the longitudinal ONL analysis of the different treatments. In contrast to the control and AAV::mOP-mEgr1 injected mice, thinning of the ONL was not significant in AAV::mOP-shEgr1 treated mice (Figure 4D). Altogether, these results suggest that Egr1 moderately contributes to rod degeneration in Rho^P23H/+ mice and that reducing Egr1 expression delays but does not prevent photoreceptor loss.

Targeted modulation of Egr1 in cones of wild-type mice

To investigate the transduction and expression profile of the cone-specific AAVs driven by the human OPN1MW gene promoter, the vectors were injected into the subretinal space of 4-week-old wt mice. In vivo analysis was performed via funduscopy and OCT imaging at 8 wpi. As observed with the rod-specific vectors, minor disturbances to retinal layers at the injection sites were visible by OCT (Figure 5A-C, arrows and blue lines and squares), but the overall structure of the retina was preserved, as indicated by OCT scans across the optic nerve (red lines and squares). Widespread viral transduction was detected via fluorescent funduscopy. GFP fluorescence in retinal sections confirmed expression in cells almost exclusively at the outer rim of the ONL, suggesting specificity for cones, which was verified by co-staining with PNA. Only sparse EGR1 staining was detected in AAV::OPN1MW-shEgr1 injected retinas (Figure 5B, bottom row), while increased EGR1 signals were evident in PNA-positive cells at the outer rim of the ONL in AAV::OPN1MW-mEgr1 injected mice (Figure 5C, bottom row). Since a potential toxic effect on cone photoreceptors cannot be detected by measuring the thickness of the ONL, we compared the number of cones across treatments and between transduced and untransduced retinal areas by counting ARR3-positive cells (Figure 5D, left). Results showed that neither up- nor downregulation of Egr1 in cones of wt mice affected the number of cones at 8 wpi (Figure 5D, right).

Cones of Rho^P23H/+ mice are not affected by the modulation of Egr1 expression

Although cones are not directly affected in RP models, they degenerate secondarily for reasons still not well understood. To investigate a potential role of EGR1 in cone survival in the context of primary rod degeneration, we injected the cone-specific vectors in 4-week-old Rho^P23H/+ mice. Fundus and OCT images at 8 wpi did not reveal notable differences across treatments (Figure 6A-C, top rows). Immunofluorescence showed targeted expression of the vectors in Rho^P23H/+ cones and apparently reduced EGR1 signals in AAV-OPN1MW-shEgr1 retinas (Figure 6B, bottom row). Transduction with AAV::OPN1MW-mEgr1 did not notably increase EGR1 expression over control, as most (if not all) cones in the Rho^P23H/+ retina already expressed the transcription factor (Figure 6A,C; bottom rows). To directly assess the effects of EGR1 on cones in the degenerative model, we quantified surviving ARR3-positive cones (Figure 6D, left) and compared their number across treatments and in transduced versus untransduced areas. Data showed that modulation of Egr1 expression in cones of the Rho^P23H/+ mouse did not influence their survival (Figure 6D, right).

CONCLUSIONS

The aim of this work was to define the function of Egr1 in the context of retinal degeneration, addressing the current scarcity of in vivo data despite its implication in various retinal disorders. Our data indicate that Egr1 moderately promotes the development and/or progression of rod degeneration but may not be the primary driver of the disease. However, the precise function of EGR1 in photoreceptor degeneration remains unclear until the activating signal(s) and downstream targets are identified.