The P23H mouse line (RHO^P23H/+) employed in this study was generated as previously described (Sakami et al., 2011). All studies on animals were conducted in strict accordance with the institutional guidelines for animal research and approved by the Italian Ministry of Health; Department of Public Health, Animal Health, Nutrition and Food Safety in accordance with the law on animal experimentation (article 31; D.L. 26/2014; protocol number: 0016304-21/07/2020-DGSAF-MDS-P). RHO^P23H/+ mice were maintained on the C57Bl/6J background. In all experiments, we used as controls aged-matched littermates of RHO^P23H/+ mice.

Drug treatments were performed by once-daily intraperitoneal injection of NSC668394, at a dose of 0.226 mg/kg, as previously described (Celik et al., 2016).

Scotopic and photopic electrophysiological recordings were performed as described (Barbato et al., 2017). A National Instruments amplifier with a xenon Ganzfeld stimulator (CSO, Costruzione Strumenti Oftalmici, Florence, Italy) was used to record mice. Briefly, mice were dark-adapted for at least 3 h. Animals were anesthetized and positioned in a stereotaxic apparatus under dim red light. Their pupils were dilated with a drop of 0.5% tropicamide (Visufarma, Rome, Italy), and body temperature was maintained at 37.5°C. The electrophysiological signals were recorded through gold-plate electrodes inserted under the lower eyelids in contact with the cornea. The electrodes in each eye were referred to a needle electrode inserted subcutaneously at the level of the corresponding frontal region. The different electrodes were connected to a two-channel amplifier. For ERG analysis in dark-adapted conditions (scotopic), eyes were stimulated with light flashes. Five different light intensity stimuli were used ranging from 1 × 10⁻⁴ to 20 cd·s/m². Amplitudes of a- and b-waves were plotted as a function of increasing light intensity. After the completion of responses obtained in scotopic conditions, the recording session continued with the purpose of dissecting the cone pathway through the photopic ERG. Photopic cone responses were isolated in light conditions with a constant background illumination of 50 cd·s/m², with 10 flashes, and a light intensity of 20 cd·s/m².

HeLa cells were obtained from American Type Culture Collection (ATCC) and were cultured in Dulbecco's Modified Eagle Medium (DMEM) (Gibco) supplemented with 10% (v/v) FBS and 5% penicillin-streptomycin. Cell lines were maintained at 37°C, 5% CO₂ in a humidified incubator according to the guidelines provided by the vendors. To analyze rhodopsin degradation, cells were plated for 24 h and then treated with 200-nM Bafilomycin A1 (Baf) (Sigma–Aldrich, B1793) for 3 h, 100-μg/ml cycloheximide (CHX) (Sigma–Aldrich, C4859), 100-mM Bortezomib (BTZ) (Sigma–Aldrich, 5043140001) and 10 μM of NSC668394 or DMSO as previously reported (Bulut et al., 2012).

Cells were transfected at 70–80% confluence with 2.5 μg of pCS2-RHO P23H for 48 h. For plasmid transfection Lipofectamine 2000 (Invitrogen, 12566014) was used, following the manufacturer's protocol.

Mouse eyes were enucleated and the RPE was separated from the retina. Cells were collected after transfections or treatments to extract total protein. Both mice and cell samples were lysed by using RIPA buffer (150-mM sodium chloride, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 50-mM Tris, pH 8.0) with inhibitors cocktail (Thermo Fischer Scientific, 78420). The concentration of total protein was determined by Bradford analysis and quantified by using a NanoDrop ND-8000 spectrophotometer (NanoDrop Technologies). Proteins were fractionated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and transferred to PVDF membranes (EMD Millipore, IPVH00010), then blocked in Tween 0.1%-Tris-buffered saline containing 5% non-fat milk or 5% bovine serum albumin (Sigma–Aldrich, 9048-46-8) for 1 h at room temperature and subsequently incubated overnight at 4°C with primary antibodies. For the Western blot analysis, the following antibodies were used: Mouse anti-Lamp1 (1:500, Sigma–Aldrich, L1418), mouse anti-Ezrin (1:1000, Novex, 357300), rabbit anti-phospho-Ezrin (Thr567) (1:700, Sigma-Aldrich, PA5-37763), rabbit anti-SQSTM1/p62 (1:1000, Sigma–Aldrich, P0067), rabbit anti-LC3 (1:1,000, Novus LC3B/MAP1LC3B), mouse anti-GAPDH (1:1000, Santa Cruz, SC-32233), rabbit anti-Perk (1:1,000, Cell Signaling, 3192), rabbit anti-BiP (1:1,000, Cell Signaling, 3177), rabbit anti-Xbp1s (1:1,000, Cell Signaling, 12782), mouse anti-Chop (1:1,000, Cell Signaling, 2895), mouse anti-RHO (1:500, Santa Cruz, sc-57432). After washing with 1% TBS, the membranes were incubated for 1 h at room temperature with the following secondary antibodies: Goat anti-rabbit IgG antibody, HPR conjugate and goat anti-mouse IgG antibody, HPR conjugate (1:10,000 EMD Millipore, 12–348; 12–349). Western blot detection was done with a GE detector (GE Healthcare Life Sciences) and quantified using ImageJ software.

Mouse eyes were fixed overnight in 4% paraformaldehyde in PBS at 4°C and then cryopreserved by treatment, first with 15% and then with 30% sucrose in phosphate-buffered saline and embedded in OCT. Twenty-micrometer cryosections were collected on slides (Superfrost Plus; Fisher Scientific, Pittsburgh, PA). Cells were fixed with 4% formaldehyde (Sigma–Aldrich) for 10 min at room temperature followed by washing with 1% PBS. After the fixation, the cells were permeated with blocking buffer (0.5% BSA, 0.005% saponin, 0.02% NaN₃) for 1 h at room temperature. The following primary antibodies were used: rabbit anti-LAMP1 (1:400, abcam, ab24170), mouse anti-RHO (1:100, abcam, ab5417), rabbit anti-c-Arrestin (1:500, EMD Millipore, AB15282), rabbit anti-GFAP (1:400, Dako GA52461-2), rabbit anti-Iba-1 (1:1,000, Wako 019-19741). All incubations were performed overnight at 4°C. After washing with 1% PBS, slides were incubated with the following secondary antibodies: Alexa 594 goat anti-rabbit/mouse/rat (1:1,000, Invitrogen A-11037 rabbit, A-11032 mouse) or Alexa 488 goat anti-rabbit/mouse/rat (1:1,000, Invitrogen A-11008 rabbit, A-11001 mouse) and DAPI (1:500, Vector Laboratories H-1200) for 1 h at room temperature; then, the slides were washed with 1% PBS and mounted with PBS/glycerol and imaged with a Zeiss LSM700 microscope. All samples were imaged, and images were identically processed.

The number of apoptotic cells was analyzed by TdT-mediated dUTP nick end labeling (TUNEL), using the in situ Cell Death Detection Kit, Fluorescein (Roche 11684795910) following the manufacturer's directions. Twelve-micrometer cryosections were collected on slides and subjected to TUNEL assay. To consider the presence of unspecific results, some retina sections were incubated with the reaction mix without TUNEL reaction enzyme. Sections were observed with a Leica DM-5500 microscope and then confocal images were acquired using the LSM700 Zeiss Confocal Microscopy system. The number of TUNEL-positive cells was evaluated in the dorsal and ventral part of the retina by manual counts with a Leica DM-6000 microscope, with the objective Leica ∞/0.17/D, HCX PL FLUOTAR, 40X/0.75 that has an area of 0.31 mm².

Hela cells were fixed with a mixture of 4% paraformaldehyde (PFA) and 0.05% glutaraldehyde (GA) for 10 min at room temperature, then washed with 4% PFA once to remove the residual GA and fixed again with 4% PFA for 30 min at room temperature. Next, the cells were incubated with a blocking/permeabilizing mixture (0.5% BSA, 0.1% saponin, 50-mM NH₄Cl) for 30 min and subsequently with the primary monoclonal antibody anti-RHO, diluted 1:100 in blocking/permeabilizing solution. The following day, the cells were washed and incubated with the secondary antibody, the anti-mouse Fab fragment coupled to 1.4-nm gold particles (diluted 1:50 in blocking/permeabilizing solution) for 2 h at room temperature. The cells were then post-fixed as described in Polishchuk and Polishchuk (2019). After dehydration, the specimens were embedded in epoxy resin and polymerized at 60°C for 72 h. Thin 60-nm sections were cut on a Leica EM UC7 microtome. The EM images were acquired from thin sections using a FEI Tecnai-12 electron microscope equipped with a VELETTA CCD digital camera (FEI, Eindhoven, the Netherlands).

Total RNA was extracted from murine eyes using the miRNeasy Kit (QIAGEN 217004) according to the manufacturer's instructions. Ribonucleic acid was quantified using a NanoDrop ND-8000 spectrophotometer (NanoDrop Technologies). The cDNAs were generated using the QuantiTect Reverse Transcription Kit (Qiagen) for the qRT-PCR analysis. The qRT-PCR reactions were performed with nested primers (Table 1) and carried out with the Roche Light Cycler 480 system. The qRT-PCR reaction was performed using cDNA (200–500 ng), 10 μl of the SYBR Green Master Mix (ROCHE), and 400-nM primers, in a total volume of 20 μl. The PCR conditions for all the genes were as follows: Preheating, 95°C for 60 s; cycling, 45 cycles of 95°C for 10 s, 60°C for 10 s, and 72°C for 15 s. Quantified results were expressed in terms of cycle threshold (Ct). The Ct values were averaged for each triplicate. The HPRT gene was used as the endogenous control for the experiments. Differences between the mean Ct values of the tested genes and those of the reference gene were calculated as DCtgene = Ctgene – Ctreference. Relative expression was analyzed as 2-DCt. Relative fold changes in expression levels were determined as 2-DDCt.

All data are expressed as mean ± standard error unless indicated otherwise. The number of samples in each experiment is provided in appropriate methods description or figure legends. The statistical analysis was carried out at least from three independent experiments for each genotype and experimental conditions. In all experiments, the significance of differences between groups was evaluated and made using t-test. The p-values of significance are indicated where appropriate in the figure legends.

To assess whether pharmacological induction of autophagy exerts a protective effect when delivered in a fully differentiated retina before the molecular and clinical signs of PRs degeneration, we daily administered the Ezrin's inhibitor “NSC668394” in RHO^P23H/+ mice from PN6 to PN60 using intraperitoneal injection of NSC668394, at a dose of 0.226 mg/kg as previously described (Celik et al., 2016). The starting point of intervention better represents the pre-symptomatic stage of the disease in human patients. Following the daily injection from PN6 onwards, we recorded an improved ERG responses at PN30 and PN60 in NSC668394-treated eyes vs. controls (Figure 1). Importantly, we recorded an amelioration in both scotopic a-wave and b-wave amplitudes at PN30 (Figures 1A,C), which persisted until PN60 (Figures 1D,F). Moreover, photopic ERG records, in the presence of background light, appeared higher amplitudes in NSC668394-treated animals compared to control (Figures 1C,F). Specifically, the a-wave amplitude of the RHO^P23H/+ treated mice showed a trend of increase both at PN30 and PN60, albeit statistical significance was only observed at PN60 (Figures 1A,D). Notably, a significant increase in the b-wave amplitude was observed at both PN30 and PN60 (Figures 1B,E). Thus, we next sought to determine whether the pharmacological NSC668394 treatment was indeed related to rescue of morphological and molecular detrimental cascade underlying PRs' ER-stress and cell death. Importantly, daily injections of NSC668394 to RHO^P23H/+ mice over two consecutive months (beginning at PN6) statistically ameliorated both outer nuclear layer (ONL) thickness and density in both dorsal and ventral portions of retina of the injected eyes, as detected by immunofluorescence assays (Figures 2A–H). First, we performed a morphometric analysis in PN18 retinas of NSC668394- and DMSO-treated RHO^P23H/+ mice when the first evidence of alteration of the retinal structure appears (Supplementary Figures S1, S2). Importantly, NSC668394-treated RHO^P23H/+ mice showed a relevant increase of the thickness of the ONL compared to DMSO-treated mice both in the dorsal and ventral retinal areas (Supplementary Figures S1A,G). Consistent with this finding, treatment of RHO^P23H/+ mice with NSC668394 resulted in a marked preservation of thickness of rod outer segments (OS), detected by immunostaining with anti-Rhodopsin antibody (Supplementary Figures S1A,C,H). Moreover, we measured a complete rescue of cone morphology in NSC668394-treated RHO^P23H/+ that were like those recorded in DMSO-wild-type (WT) mice, as demonstrated by quantifying cones with cone segments (CS) and dendrites clearly labeled by immunostaining with anti-c-Arrestin antibody (Supplementary Figures S1D–F,I). Most importantly, we obtained similar beneficial effects also from NSC668394-treated RHO^P23H/+ animals injected at PN6 and analyzed at PN60 (Figure 2), when retinal degeneration is advanced. Treatment of RHO^P23H/+ mice with NSC668394 led to an increased ONL thickness and density in both dorsal and ventral retina, as measured by DAPI staining and compared with DMSO-treated control mice (Figures 2A,H). Consistent with photoreceptor function assessments, significant recovery was also detected on the thickness of rod OS (Figures 2A–C,I), cones' morphology in NSC668394-treated RHO^P23H/+ that were like those recorded in DMSO-WT mice (Figures 2D,F,J). Besides, we sought to determine whether the pharmacological NSC668394 administration had an impact on PRs viability. Notably, treatment with NSC668394 resulted in a significant decreased number of TUNEL-positive photoreceptor cells compared to DMSO-treated mice, supporting a reduction in PRs' cell death (Figure 3, Supplementary Figure S2). This outcome was significant in Ezrin inhibited- RHO^P23H/+ mice at PN18 (Supplementary Figure S2), but it became even more consistent in treated retina at PN60 (Figures 3A,D). Next, we investigated possible inflammatory effects of Ezrin inhibition in the RHO^P23H/+ retinae. For this purpose, we performed immunofluorescence staining for glial fibrillary acidic protein (GFAP) and ionized calcium-binding adapter molecule 1 (Iba-1) of RHO^P23H/+ mice at PN60 (Figures 3E,M). Firstly, to measure activation of retinal gliosis, we quantified the fluorescent integrated density (IntDen) and the area of GFAP immunoreactivity in dorsal and ventral region of the retina. In RHO^P23H/+ mice GFAP immunoreactivity was significantly higher than in the age-matched WT control and it extended to the ONL, indicating Müller cells activation and retinal gliosis. Remarkably, NSC668394 treatment rescued gliosis as demonstrated by reduction of area of the GFAP immunoreactivity and IntDen (Figures 3E,I). In support of beneficial effect of the NSC668394 administration in RHO^P23H/+ mice, we also observed a reduction of the number of microglia cells in the RHO^P23H/+ retinal cryosections. We quantified the number of microglia cells by Iba-1 immunostaining, hallmark of activated microglial cells and retinal inflammatory response. Under DMSO treatment, RHO^P23H/+ mice showed over twice as many Iba-1-positive cells compared to WT retinal cryosections, confirming that the RHO^P23H accumulation in rods induced cell death and secondarily strong activation of retinal microglia, which was localized in the ONL, inner nuclear layer (INL) and ganglion cell layer (GCL). Consistent with protective effects, NSC668394 decreased Iba1-positive cell number in NSC668394-RHO^P23H/+ treated animals compared to DMSO controls (Figures 3J,M). This result supports that the protective effect of NSC668394 in RHO^P23H/+ mice may be mediated by the clearance of unfolded RHO^P23H aggregates via targeting Ezrin activity and induction of autophagy pathways.

The recent discoveries have proven that lysosomes may reduce misfolded protein accumulation via activation of autophagy pathway (Mendes and Cheetham, 2008; Sakami et al., 2011). We sought to assess whether the protective effect of NSC668394 is mediated by activation of lysosomal clearance of RHO^P23H, we analyzed NSC668394 administration on Hela cells overexpressing RHO^P23H (HeLa^RHO−P23H). Interestingly, culturing HeLa^RHO−P23H cells in the presence of NSC668394 led to a significant decrease of RHO^P23H in both insoluble and soluble fraction, as demonstrated by Western blot analyses, suggesting that there was a major change in rhodopsin degradation (Figure 4A). To further support that NSC668394 treatment induced lysosomal clearance of RHO^P23H, we assessed lysosomal cargo degradation on Ezrin-inhibited Hela RHO^P23H cells followed by a treatment with Bafilomycin A1 (Baf), a known inhibitor of vacuolar-type H(+)-ATPase that promotes alkalinization of the lysosomal lumen and inhibition of lysosomal cargo degradation, leading to an accumulation of undigested material within the lumen of already fused autolysosomes, and prevents new autophagosomes and lysosomes fusion (Yoshimori et al., 1991). As expected, treatment with Baf on Hela^RHO−P23H cells showed a slight increase of RHO^P23H accumulation in the lumen of autolysosomes (LAMP1-positive organelles) compared with DMSO-treated cells (Figures 4B–C”). Treatment with NSC668394 on Hela^RHO−P23H cells increased the number of autolysosomes structures and showed a moderate increase of RHO^P23H accumulation in the lumen of autolysosomes (LAMP1-positive organelles) compared with DMSO-treated cells (Figures 4B–B”,D–D”). Notably, administration of Baf after 3 h of NSC668394 treatment (Baf/NSC668394-treated Hela^RHO−P23H) induced a high increase of autolysosomes structures containing RHO^P23H in the lumen, in accordance with a block of cargo degradation in already increased number of autolysosomes (Figures 4E–E”). These data were also confirmed in the 661W photoreceptor cell line (Supplementary Figure S3) (Wheway et al., 2019), supporting the notion that NSC668394 treatment increases RHO^P23H degradation via autophagy pathway. Furthermore, to better investigate on rhodopsin sub-cellular localization, we performed an IEM analysis (Figure 5) and carried out morphometric issues. Hela RHO^P23H cells showed an increase of autolysosome-like structures, compared to normal levels, presumably due to the presence of the misfolded protein. Interestingly, the number and the volume of the autolysosome-like structures were even more increased in Baf/NSC668394-treated Hela^RHO−P23H cells (Figures 5A–D), confirming the treatment effectiveness. Most importantly, we found a significant increase in the number of rhodopsin-positive autolysosome-like structures in Baf/NSC668394-treated Hela^RHO−P23H cells (Figures 5A–E). Subsequently, we tested whether the reduction of RHO^P23H protein was also due to a block of protein synthesis or increase of proteasome activity. Remarkably, administration of NSC668394 on Hela^RHO−P23H in the presence of cycloheximide (CHX) treatment, a well-known inhibitor of protein synthesis, further showed a faster degradation of RHO^P23H protein in NSC668394-treated cells compared with DMSO-treated control cells (Supplementary Figure S4A). Taken together, these data indicate that reduction of RHO^P23H protein in NSC668394-treated Hela^RHO−P23H is not due to an altered protein synthesis. To avoid the possibility that NSC668394 might induce proteasome activity, we provide insights into how NSC668394 regulates RHO^P23H clearance. To this end, NSC668394-treated HeLa^RHO−P23H cells were cultured in presence of the Bortezomib (BTZ), a proteasome inhibitor. Notably, following NSC668394 treatment, a significant decrease of RHO^P23H was also detected upon BTZ conditions, as demonstrated by Western blot analysis (Supplementary Figure S4B). On the contrary, the levels of RHO^P23H protein were markedly increased in DMSO-treated HeLa^RHO−P23H cells upon BTZ conditions (Supplementary Figure S4B). Taken together, these data indicate that pharmacological NSC668394 treatment specifically results in activation of lysosomal RHO^P23H degradation.

The mutation P23H rhodopsin was reported to activate ubiquitin/proteasome system (Chapple et al., 2001) and trigger the ER-stress response, which, in turn, lead to cell death (Griciuc et al., 2010). We therefore hypothesized that if NSC668394 directly induces the clearance of RHO^P23H aggregates in the photoreceptor cells, the clearance of RHO^P23H misfolded protein should reduce ER-stress response in vivo. Notably, NSC668394 treatment reduced the levels of Perk, Xbp1, and Chop proteins, commonly used as UPR markers, compared with vehicle-treated RHO^P23H/+ mice at PN18 (Figure 6A). Consistent with this observation, we detected a significant increase in Bip protein level in NSC668394-RHO^P23H/+ retina, as analyzed by Western blot analysis (Figure 6A). In addition, we observed reduced ER-stress and UPR markers, as demonstrated by BiP, Chop, and Xbp1 gene expression levels in NSC668394-RHO^P23H/+ compared with DMSO-WT and DMSO-RHO^P23H/+ mice. Notably, BiP expression level was increased following NSC668394 treatment, suggesting a possible BiP-mediated protective effect. Accordingly, Chop and Xbp1 gene expression levels resulted significatively reduced in NSC668394-RHO^P23H/+ mice compared with DMSO-RHO^P23H/+ mice at PN18. Consistently, the expression levels of both Atf4 and Atf6 genes in NSC668394-RHO^P23H/+ mice resulted similar to DMSO-WT mice, whereas DMSO-RHO^P23H/+ mice showed higher Atf4 and Atf6 expression levels compared to DMSO-WT mice (Figure 6C). Taken together, these results strongly support that pharmacological NSC663894 administration reduces both UPR and ER-stress pathways. Interestingly, RHO^P23H/+ mice showed impairment of autophagy pathway as a consequence of RHO^P23H-mediated ER-stress at the onset and during the progression of retinal degeneration in mice (Sizova et al., 2014; Qiu et al., 2019; Kakavand et al., 2020). We therefore hypothesized that if NSC668394 treatment reduces ER-stress should re-establish and normalize autophagy pathway.

As expected, daily injection of NSC668394 over 2 consecutive weeks reduced the active pT567-Ezrin protein level, without affecting Ezrin total protein level, and was efficient in rescuing imbalance of autophagy pathway in the retina of RHO^P23H/+ mice when compared with vehicle-injected (DMSO) control animals (Figure 6B). Western blot analysis showed a normalization of Lamp1, SQSTM1/p62, and LC3-II protein levels, consistent with a rescue of autophagy pathway (Figure 6B). Interestingly, we also observed a reduction of LC3I. Accordingly, this was due to an increase of autophagic flux, leading to a rapid conversion of LC3-I in LC3-II, which is rapidly degraded.

Accordingly, retinas from NSC668394-RHO^P23H/+ treated mice exhibited Lamp1 and Ctsd expression levels similar to those detected in the retinas from DMSO-WT mice, supporting a normalization of autophagy pathway. Opposite results were observed in the retinas from RHO^P23H/+ mice treated with DMSO, which exhibited increase of both Lamp1 and Ctsd expression levels compared to DMSO-WT mice (Figure 6C), in line with previous results (Sizova et al., 2014). The ability of NSC668394to regulate differently autophagic flux was consistent with autophagic flux on RHO^P23H/+ mice, compared with WT mice, that showed a rapid increase at PN6-PN12, followed by a reduction PN18-PN30 of autophagy, as showed by Western blot analysis (Supplementary Figure S4C).

Altogether, these data show that NSC668394-mediated activation of cell clearance exerts a protective effect in IRD model and may represent an innovative strategy to treat retinal dystrophies due to accumulation of misfolded protein.