It is unclear how genes, such as RPGR (retinitis pigmentosa guanine triphosphatase regulator) that are expressed in both rods and cones, cause variable disease pathogenesis. Using transcriptomic analysis, we show that loss of RPGR in a rod-dominant mouse retina (Rpgrko) results in predominant alterations in genes involved in actin cytoskeletal dynamics, prior to onset of degeneration. We validated these findings and found an increase in activated RhoA-GTP levels and polymerized F-actin in the Rpgrko mouse retina. To assess the effect of the loss of RPGR in the all-cone region of the human retina, we used Nrl−/− (neural retina leucine zipper) mice, to generate Rpgrko::Nrl−/− double-knock-out (Rpgr-DKO) mice. These mice exhibited supranormal cone response to light and substantially retained retinal architecture. Transcriptomic analysis revealed predominant up-regulation of retinal pigmented epithelium (RPE)-specific genes associated with visual cycle, whereas fatty acid analysis showed mild decrease in docosahexaenoic acid in the retina of the Rpgr-DKO mice when compared with the Nrl−/− mice. Our data reveal new insights into distinct intracellular pathways that are involved in RPGR-associated rod and cone dysfunction and provide a platform to design new treatment modalities.
Our vision starts with detection of photons of light by photoreceptors (rods and cones). These neurons form the bulk of the retinal cell types and, along with the overlaying RPE, comprise the light-sensing compartment of the eye (1). Photoreceptors are polarized neurons with a distinct sensory outer segment (OS; also called sensory cilium) and an inner segment involved in protein synthesis and transport (2,3). The OS is loaded with photopigment opsin (rod opsin in rods and cone opsin in cones), which binds to the chromophore to carry out light detection cascade. RPE is involved in regulating visual cycle by participating in the regeneration of the chromophore and OS maintenance (4–6). Vision loss due to degeneration and dysfunction of photoreceptors typically involves malfunction of the OS or the RPE. As cones are responsible for the bulk of our day vision, cone degeneration would eventually lead to legal blindness.
Retinitis pigmentosa (RP) is an inherited form of neurodegenerative retinal disease that affects ∼1 in 3000 individuals (7). While RP is clinically and genetically heterogeneous, it is characterized by progressive loss of rod function and night blindness followed by cone loss and complete blindness, usually by the third decade of life (8,9). RP can be inherited in an autosomal dominant, autosomal recessive or X-linked manner. Of these, X-linked RP (XLRP) is the most severe and prevalent form of RP. Mutations in two genes, retinitis pigmentosa guanine triphosphatase (GTPase) regulator (RPGR) and RP2 account for >90% of XLRP cases (10,11). RPGR mutations are the most common cause of XLRP, accounting for >70% of XLRP cases and ∼15% of simplex (isolated) RP cases (12,13). RPGR-associated disease shows variable degrees of severity with respect to early involvement of either rods or cones in males (14,15); some female carriers are also known to develop a relatively severe form of the disease (16). RPGR mutations are also associated with cone degeneration and macular atrophy (17,18), indicating predominantly cone dysfunction in the cone-rich region of the retina called the macula.
The RPGR gene encodes multiple alternatively spliced isoforms, of which there are two major isoforms: the constitutive isoform, which contains 19 exons (RPGR1–19) and ORF15 isoforms containing 15 exons terminating in intron 15 (RPGRORF15) (19–24). RPGR is implicated in regulating protein trafficking in photoreceptors (22,25,26). The Rpgrko mouse undergoes progressive retinal degeneration and mis-trafficking of opsins in photoreceptors (27). Naturally occurring mouse and canine models of RPGR mutations as well as additional gene-targeted mouse mutants of Rpgr have been reported (28–30). These models exhibit distinct rod and cone alterations as observed in patients.
Given discordant rod and cone involvement in RPGR-XLRP; here, we present our findings of the discordant effect of the loss of RPGR on rod-dominant or cone-only rodent retinas. Our studies not only provide new platforms to assess human cone disease in RP, but will also assist in developing targeted therapeutic strategies.
Photoreceptor dysfunction and transcriptomic analysis of Rpgrko in a rod-dominant background
To assess the pathogenic mechanism of the loss of Rpgr, we wanted to select an age of mice that is prior to any detectable photoreceptor dysfunction. Therefore, we analyzed the photoreceptor function in the Rpgrko mice by electroretinography (ERG). Although such studies have been reported (27), we aimed to assess the photoreceptor dysfunction using similar conditions in all our experiments. While 8-month-old Rpgrko mice show decreased rod (scotopic) and cone (photopic) responses (Fig. 1A and B) and decrease in the thickness of the outer nuclear layer (ONL; photoreceptor nuclei) (Fig. 1C) when compared with the wild-type (WT) counterpart, no effect was observed at 1 month of age. Therefore, we selected this age for further analysis.
To examine the effect of the loss of RPGR in a rod-dominant retina, we performed RNAseq analysis from three biological replicates of 1-month-old WT and Rpgrko mouse retinas followed by differential gene expression analysis. Significantly differentially expressed genes were selected based on a cutoff of at least 2-fold change in the expression level between the two groups and false discovery rate (FDR) <0.05. Our analysis identified 132 differentially expressed transcripts that were annotated (Supplementary Material, Table S1). Bioinformatics analysis using ingenuity pathway analysis (IPA) and Database for Annotation, Visualization and Integrated Discovery, v6.7 (DAVID) identified five major categories of disrupted cellular functions: molecular transport, small-molecule biochemistry, cellular function and maintenance, cellular movement and cellular assembly and organization (Supplementary Material, Table S2). Among the cellular functions specific to the retina, we identified: cell death and survival, cellular assembly and organization, cellular function and maintenance, cell-to-cell signaling and interaction and cell morphology (Supplementary Material, Table S3). Further examination of the transcriptomics data revealed differential regulation of genes involved in actin cytoskeletal dynamics and in human diseases. As actin defects were implicated in RPGR-associated ciliary dysfunction in a previous study using in vitro cell culture model (31), we selected this pathway for further analysis. We validated the expression of three genes implicated in actin dynamics: STARD13 (StAR-related lipid-transfer domain protein 13), RTKN2 (Rhotekin 2) and SRPX (sushi-repeat containing protein, X-linked). STARD13 regulates RhoA and is involved in regulating cell adhesion mediated by actin cytoskeleton (32). RTKN2, on the other hand, is an effector of Rho GTPases, which binds to the activated form of Rho GTPases (33,34). SRPX was previously implicated in XLRP (35) and was also found to associate with the actin cytoskeleton (36). We validated the changes in the expression of Rtkn2, Stard13 and Srpx by quantitative polymerase chain reaction (qPCR). Consistent with our RNAseq data, we detected a ∼3.8-folds increase in the expression of Stard13 and >25-folds increase in Rtkn2 transcript levels in the Rpgrko mice (Fig. 2A; Supplementary Material, Fig. S1A). On the other hand, Srpx expression was undetectable in the Rpgrko retina.
Actin cytoskeletal defects in Rpgrko mice
Actin dynamics is associated with syndromic ciliopathies, including Bardet–Biedl syndrome (BBS) and is implicated in regulating cilia length (37,38). Moreover, actin filaments are localized at the base of the OS and are implicated in disc morphogenesis (39). To gain further insight into the status of actin cytoskeleton in the Rpgrko retina, we first validated the changes in the protein level of RTKN2 protein in the retina. Immunoblot analysis of WT and Rpgrko retinal extracts revealed a significant increase in the levels of the RTKN2 protein (Fig. 2B). As RTKN2 is involved in regulating polymerization of actin, we then calculated the ratio between polymerized filamentous (F)-actin and the globular (G)-actin in the 1-month-old WT and Rpgrko retinas. By fractionating the retina followed by immunoblot analysis using anti-actin antibody, we found a 2-fold increase in the F- to G-actin ratio in the Rpgrko retina when compared with the WT retina (Fig. 2C and D). No change in total actin levels was observed.
RhoGTPases are involved in the regulation of actin dynamics and increase in RhoA-GTP results in increased actin polymerization (40,41). We, therefore, hypothesized that increased levels of polymerized actin in the Rpgrko retina are associated with higher RhoA-GTP levels. Effectors of small GTPases recognize the GTP-bound state of the GTPases (42,43). RTKN is an effector for RhoA-GTPase as it preferentially binds to RhoA-GTP (33). To gain insights into the predominant state of RhoA in the Rpgrko retina, we performed a pull-down assay using recombinant glutathione S-transferase (GST)-tagged RTKN (GST-RTKN) and mouse retinal extracts loaded with non-hydrolyzable GTP analog (GTPγS). As shown in Figure 2E and F, we found a ∼1.6-fold increase in the levels of RhoA pulled down by GST-RTKN in the GTPγS-loaded Rpgrko retinal extracts when compared with the WT retinal extracts. No association between RhoA and purified GST protein was detected.
Actin cytoskeleton has been proposed to regulate basal disc orientation during OS morphogenesis (39). Ultrastructural analysis of the OS of the Rpgrko mice revealed altered basal OS morphology even at 1 month of age, which progressed with age and was more prominent at 3 months of age (Supplementary Material, Fig. S1B).
Expression of Rpgr in cone-dominant mouse retinas
Past work suggests that, in cases where the causative gene is expressed specifically in rods, secondary effects of rod loss manifest in cones as well. This is largely because rods account for the majority (97%) of photoreceptors in mammalian retina. However, in several instances, the causative gene (such as RPGR) is expressed in a variety of cell-types and mutations also manifest as early loss of cone function (44–46). Thus, it is conceivable that RPGR plays distinct roles in rods and cones and that cone death also involves the effect of the loss of RPGR directly on cone function. In primates, the fovea represents a cone-enriched region of the retina and dysfunction in this region results in the loss of daytime vision. To generate a cone-enriched model of RPGR loss, we utilized Nrl−/− mice, which develop cone-only retinas with a complete absence of rod photoreceptors. These mice express cone-specific genes and their cones display morphological and physiological features of WT cones (47,48). Thus, there is a great utility of Nrl−/− mice to understand cone biology and responses of human fovea during disease, as evident from previous studies to assess primary and secondary effects of retinal degeneration on cones (49–51).
We first performed qRT-PCR using RNA from retinas of WT (rod-dominant) and Nrl−/− mice. Compared with the WT mice, we found a ∼2-fold up-regulation of Rpgr transcript expression in the Nrl−/− retina (Fig. 3A). Although our analysis of Rpgr expression in the WT mice is also taking into account its levels in cones, such levels would be minimal compared with 95–97% rod photoreceptors.
Generation and characterization of Rpgr-DKO mice
We, next, generated Rpgr-DKO mice by breeding the Rpgrko mice to the Nrl−/− mice. The loss of RPGR was validated by immunoblot analysis of mouse retinal extracts using a previously reported anti-RPGR antibody (21). As shown in Figure 3B, expression of both major isoforms of RPGR: constitutive RPGR (RPGRconst) and RPGRORF15 is undetectable in the Rpgr-DKO mice. We also tested whether the loss of RPGR results in alterations in the expression of other cone-specific genes. qRT-PCR and immunoblot analyses revealed no differences in the expression of S-cone opsin (Fig 3C and D) and cone alpha subunit of cone phosphodiesterase (data not shown) between Rpgr-DKO and Nrl−/− mouse retinas. The amount of protein loaded in the gel (20 µg) is such that S-opsin expression is undetectable in the WT mouse retina so that higher expression in Nrl−/− and Rpgr-DKO mouse retina is easily detected.
Cone function alterations in Rpgr-DKO
Previous studies using animal models and RPGR patient information revealed a decline in both rod and cone function with age (15,52,53). Using the Rpgr-DKO mice, we conducted an age-dependent functional analysis of light signal processing in the absence of rods using ERG. As controls, we used age-matched Nrl−/− mice. Consistent with previous studies (47,48), Nrl−/− photopic ERG amplitude was relatively higher when compared with the WT mice (data not shown). Interestingly, we detected even higher (supranormal) photopic amplitude (∼2-fold high) in the Rpgr-DKO mice (Fig. 4A). Analysis of Flicker ERG waveforms also showed similar results (Fig. 4B). There were sharp triangular waveforms in Rpgr-DKO mice at all ages tested as opposed to although sharp, but low-amplitude waveforms in Nrl−/− mice. Such sharp responses are typical of functional cone photoreceptors. Although there was progressive decline of cone response in Rpgr-DKO mice, the function was significantly higher even up to 9 months of age when compared with Nrl−/− mice. As expression levels of other cone proteins are similar in the two mouse models, such differences are unlikely to be due to differences in cone cell number.
Rosette formation is a hallmark of Nrl−/− mice (48,54) and has been reported to be associated with the loss of photoreceptor function (50,55,56). We, therefore, examined the effect of the loss of Rpgr on rosette formation by using retinal cross sections and whole mounts. As shown in Figure 4C and Supplementary Material, Figure S2, we observed mild decrease (∼25%) in the number of rosettes in the Rpgr-DKO retina when compared with the Nrl−/− retina.
Retinal morphology and protein trafficking in Rpgr-DKO mice
To assess the effect of the loss of RPGR on the morphology of the cone-only retina, we performed histological analysis of retinal sections at different stages. As shown in Figure 5A, the ONL (composed of photoreceptor nuclei) in the Rpgr-DKO mice was similar in thickness to that of the Nrl−/− retinas at 1 month of age. However, whereas Nrl−/− mice showed a decline in ONL thickness (consistent with previous reports of photoreceptor degeneration in these mice (54) at 3, 6 and 9 months of age, the Rpgr-DKO mouse retina showed relative delay in the decline in ONL thickness, with a thicker ONL even at 9 months of age. Ultrastructural analysis of mouse retinas at early age (1 month; prior to onset of degeneration in Nrl−/− retina) and later stage (6 months) corroborated our histological analysis. We found no apparent difference in cone morphology between the Rpgr-DKO and the Nrl−/− mice at 1 month of age whereas cone OS morphology was altered slightly at 6 months of age in Rpgr-DKO (Fig.5B). Overall, these data suggest that the loss of RPGR in mice spares cone development, but not maintenance.
RPGR is proposed to regulate ciliary protein trafficking in photoreceptors (22,26,57). To test whether Rpgr-DKO mice also exhibit protein trafficking defects, we performed immunofluorescence analysis using antibodies against cone outer segment (COS) proteins in Rpgr-DKO and Nrl−/− mouse retinas. Consistent with the previous observations using the Rpgrko mice (27,28), we found mis-localization of M-opsin in the Rpgr-DKO mice (Fig. 5C). Although the Nrl−/− mice exhibit M-opsin mis-localization, this effect is more pronounced in the Rpgr-DKO mice. As control, no mis-localization was observed in the WT mouse retina. We did not observe any change in S-opsin trafficking in the Rpgr-DKO mice (Fig. 5D). This could either be because there is minor S-opsin mis-localization, which escaped our detection limits or because S-opsin mis-localizes secondary to rod degeneration in the Rpgrko mice.
Fatty acid analysis of the Rpgr-DKO retinas
Previous studies showed that build-up of polyunsaturated fatty acids, such as docosahexaenoic acid (DHA; C22:6n3) and very long-chain fatty acids results in supranormal ERG responses (58). DHA is the most abundant fatty acid in the phospholipids of rod OSs. It has been shown that abnormal build-up of DHA and other long-chain fatty acids leads to supranormal ERG responses in mice. We, therefore, hypothesized that the cone-enriched Rpgr-DKO retinas have altered fatty acid content, which results in a hyperabnormal ERG response. To test this hypothesis, we assessed the fatty acid composition of phospholipids in the Rpgr-DKO retinas and compared it to that in the Nrl−/− retinas. First, we observed that the major fatty acids in the cone-rich Nrl−/− retinas are: stearic acid (C18; ∼28%), palmitic acid (C16; ∼25%), DHA (C22:6n3; ∼20%), oleic acid (C18:1; ∼14%) and arachidonic acid (C20:4n6; ∼10%) (Fig. 6A). As evident, DHA is not the most abundant fatty acid in the cone-rich retinas. Analysis of the Rpgr-DKO mice revealed that the level of DHA was significantly (P < 0.05) reduced when compared with the Nrl−/− mice. No significant change in other fatty acids was detected in this analysis.
Transcriptomic analysis of Rpgr-DKO mice
Our data show that the loss of RPGR does not affect cone ultrastructure and protein trafficking, but there is a significant increase in the cone function. To gain insights into the pathways underlying such an effect, we assessed the transcriptome of the Rpgr-DKO mice when compared with the Nrl−/− mice. We compared the samples at 1 month of age when there is no apparent degeneration in the Nrl−/− and the Rpgr-DKO mice. We identified 110 differentially expressed genes based on a cut-off of at least 2-fold change in the expression level between the two groups (Supplementary Material, Table S4). IPA revealed that these genes belong to five major molecular functions: cell-to-cell signaling and interaction, cellular function and maintenance, cellular movement, cellular development and cellular growth and proliferation (Supplementary Material, Table S5). Of these, 12 genes were identified as parts of four canonical pathways that are altered in the diseased retina: visual cycle, complement system, retinol biosynthesis and phototransduction pathway (Supplementary Material, Table S6). The majority of these pathways including the complement system and visual cycle are implicated in Leber congenital amaurosis (LCA) and age-related macular degeneration (59,60). We, therefore, validated the alteration of two of the complement genes by qPCR and found a 2- to 3-fold down-regulation of C1qa and C1qb.
Among the up-regulated genes associated with visual cycle regulation, we detected a ∼2.5-fold increase in Rgr (retinal G-protein-coupled receptor) and ∼3-fold increase in Lrat (lecithin:retinol acyltransferase) (61–63) (Fig. 6B). We also observed a 3-fold increase in Rpe65 (retinal pigment epithelium 65) (Fig. 6B), which is involved in chromophore recycling (64–66). Importantly, mutations in these three genes are associated with retinal degeneration (59,67,68). Immunoblot analysis of RPE extracts from the Nrl−/− and the Rpgr-DKO mice revealed ∼2.5-fold increase in RPE65 protein levels (Fig. 6C; Supplementary Material, Fig. S3).
RPGR-XLRP is considered one of the most severe forms of RP and is associated with complex phenotypic presentation of differential rod or cone disease (11,12,69–71). With recent advances in gene therapy for RPGR-XLRP in the mouse and canine models (52,72) and the observation of rescue of rpgr-knock-down-associated phenotype in zebrafish studies (73), it is imperative to now design a suitable therapeutic plan for patients by understanding rod and cone disease pathogenesis. Our studies have revealed different effects of the loss of RPGR on rod and cone-rich retinas and identified potential pathways that are altered during disease.
Loss of RPGR in cone-rich retina
Our analysis of the Rpgr-DKO mice revealed new information on unusually preserved cone population and supranormal cone function. Other reports using the Nrl−/− retina have also revealed relatively slow cone loss due to mutations in genes expressed in both rods and cones, such as Cep290, peripherin/Rds, Rpe65, Grk1 and insulin receptor (49–51,55,74–76). On the one hand, the Rpgr-DKO mice showed partial rescue of ONL degeneration of the Nrl−/− mice, while on the other hand, they exhibited even higher ERG responses. Although Nrl−/− mice exhibit higher cone ERGs when compared with WT mice, our results show that these ERGs can be even higher in the absence of RPGR. These observations indicate that the loss of RPGR potentially overcomes a negative effect on cone ERGs in the Nrl−/− mice and suggest a potential genetic interaction between NRL and RPGR. Since NRL is a transcription factor and some RPGR isoforms are also expressed in the nucleus (19), an association between these two proteins cannot be ruled out at this stage.
It has been reported that patients with macular degeneration or RP can present with supranormal photopic response (77). However, clinical testing for such ERG changes may not be feasible in all patients as they may present a late stage of the disease by which time the cone function has already deteriorated. Nonetheless, our results support the idea that patients or individuals with XLRP can be assessed for hyperabnormal cone response at their earliest clinic visit. It should be noted that the Nrl−/− mice developed increased number of S-cones, which form a minority population in human retina. Therefore, if the enhanced response were predominantly S-cone enriched, then a spatial testing using multifocal ERG or visual field test would be beneficial for early diagnosis of patients.
Several scenarios may account for the unusual cone response that we observed in the Rpgr-DKO mice. (i) An inherent property of the Nrl−/− retina: such a scenario seems unlikely because several previous studies have shown that ablation of a retinal disease gene in the Nrl−/− mice results in a decrease in the cone function and deterioration of retinal morphology (49–51). Therefore, we hypothesize that the effect that we have observed is specifically due to the loss of Rpgr on the Nrl−/− background. (ii) The Nrl−/− mouse retina develops rosettes, which can result in artifacts: it has been hypothesized that such structures are formed due to aberrant photoreceptor-Muller glia interactions, activity of RPE65 or photoreceptor-RPE interactions (50,55,56). However, those studies have not shown association between reductions in rosettes and increased cone function. Intriguingly, reduction in rosettes was shown in the Rpe65ko and Rpe65 hypomorphic mutation on the Nrl−/− background (55,56). However, our studies show that reduction in rosettes is associated with increased levels of Rpe65 and Lrat. We hypothesize that while the loss of Rpe65 more significantly affects rosette formation, the loss of RPGR has a low-to-moderate effect on rosette formation, which is independent of its effect on the levels of Rpe65 and Lrat. Therefore, it is also unlikely that the mild reduction in whorls and rosettes is contributing to the increased photopic amplitude of the Rpgr-DKO mice. (iii) A more likely scenario is that there is a predominant cone visual cycle dysfunction in the absence of RPGR. This is evident from the predominant alterations in the expression of RPE-specific genes in the Rpgr-DKO mice and alterations in the DHA levels of the Rpgr-DKO retinas. Changes in the fatty acid composition are predicted to affect the fluidity of the OS membranes and may result in altered ERGs. Previous studies showed that accumulation of DHA and other long-chain fatty acids is associated with supranormal ERG responses (58). However, the Rpgr-DKO mice showed a mild decrease in DHA levels when compared with the Nrl−/− mice. Additional investigations are necessary to delineate the significance and physiological relevance of reduced the DHA and function of cone-rich retinas.
Our studies suggest that either RPGR plays a critical role in the RPE or that up-regulation of the visual cycle-associated genes in the RPE is a secondary effect of altered light processing capacity of cones in the absence of RPGR. The light-sensitive chromophore 11-cis retinal is bound to the opsin and upon activation by light quanta, isomerizes to all-trans retinol. This initiates the phototransduction cascade (78). The all-trans retinol is then trafficked to the RPE where it is acted upon by LRAT to produce all-trans retinyl esters followed by action of RPE65 to produce 11-cis retinol. These cascades result in the recycling of the chromophore, which is then returned to the OS (51,56). It has been proposed that cones may possess distinct mechanisms of faster regeneration kinetics of the chromophore to better adapt to light conditions that allow uninterrupted daytime vision (79). An increase in Rpe65, Lrat as well as Rgr expression would enhance the rate of synthesis of reaction intermediates, which would increase the chromophore regeneration rate. Further studies are needed to test this hypothesis.
Loss of RPGR in rod-dominant mice
We found predominant alterations in actin dynamics-associated pathways in our analyses of the Rpgrko retina. The actin cytoskeleton is proposed to regulate basal OS disc alignment as well as photoreceptor axon retraction (34,39,80,81). Our analysis of RhoA activation and increased F- to G-actin ratio at 1 month of age indicate an early involvement of cytoskeletal defects in the Rpgrko mice. In support of this hypothesis, we detected altered basal OS morphology in the Rpgrko mice, prior to onset of photoreceptor dysfunction and degeneration. However, we did not detect significant alteration in ERG at early stages of the Rpgrko mice. We reckon that these initial changes in actin dynamics may not significantly alter photoreceptor function at early stages whereas a slow and progressive build-up of the insult due to cytoskeletal alterations eventually lead to ciliary dysfunction and photoreceptor degeneration.
Although molecular mechanisms underlying the activation of RhoA in ciliary dysfunction are still unclear, it was shown that RhoA-GTP levels are also up-regulated in the absence of BBS-associated proteins (37). Thus, interplay between RPGR and BBS proteins to regulate actin cytoskeletal dynamics in the retina cannot be ruled out at this stage. Given an association of actin polymerization with cilia length control (38) and the involvement of RPGR and BBS proteins in the ciliary function, we propose that deregulation of actin cytoskeleton plays a significant role in the pathogenesis of retinal ciliopathies as well as syndromic ciliopathies.
Photoreceptor OS undergoes periodic disc shedding at the distal tips, which involves phagocytosis of the shed discs by the RPE (3). Actin cytoskeleton and specifically Myosin 7A is shown to play crucial roles in the phagocytosis of the shed discs and phagosomes and melanosome transport in the RPE. It would be interesting to examine RPE pathology in RPGR-associated disease.
Overall, our results show (i) an unexpected response of cone photoreceptors and (ii) involvement of distinct pathways associated with rod and cone dysfunction, due to the loss of RPGR. Additional investigations are necessary to evaluate the mechanisms underlying such defects. The deregulated pathways are also implicated in other retinal degenerative diseases, such as LCA, BBS, Usher syndrome and other syndromic ciliopathies. Therefore, our studies may provide crucial insights and identify potential commonalities among multiple retinal diseases that cause degeneration and dysfunction of photoreceptors.
Materials and Methods
All animal procedures and experiments were performed in accordance with the guidelines of Institutional Animal Care and Use Committee. Mice were housed in animal facility at University of Massachusetts Medical School, where they were maintained on a standard diet in a 12 h light to 12 h dark cycle. Lighting conditions were kept constant in all cages with illumination of 10–15 lux at the level of the cages. The Nrl−/− and Rpgrko mice were procured from Dr Anand Swaroop and Dr Tiansen Li (National Eye Institute), respectively. Homozygous Nrl−/− and Rpgrko mice were bred to generate Rpgr::Nrl−/− DKO mice. Genotypes were confirmed by PCR genotyping and immunoblotting analysis. All mice were also genotyped to exclude the rd1 and rd8 alleles.
Primers and antibodies
List of antibodies and primers used in the present study is provided as Supplementary Material, Tables S7 and S8.
ERG, histology and immunofluorescence
ERGs were recorded as described previously by using the Espion e2 recording system (Diagnosys, Lowell, MA, USA) (82,83). For scotopic response, mice were dark-adapted overnight and all procedures were performed under dim red light. Light-adapted (Photopic) ERGs were recorded after light adaptation with a background illumination of 30 cd/m2 (white 6500 K) for 8 min by two protocols: (i) the stimulus strength of 10 cd.s/m2 was chosen for single flash. Twenty trials were averaged for single-flash responses and (ii) 30 Hz flicker ERG was performed under photopic conditions.
Histological and immunofluorescence analyses of paraffin- and cryopreserved retina and of flat mounts were performed as described (82–84). All images were taken by using a scanning confocal microscope (Leica TCS SP5 II laser; Leica Microsystems).
Transmission electron microscopy: eyes were collected from euthanized mice and fixed in 2.5% glutaraldehyde in 0.1 m sodium cacodylate buffer (pH 7.2) for 20 min at room temperature, as described (82). Specimens were visualized with a transmission electron microscope (Philips CM-10; Philips, Eindhoven, The Netherlands), coupled with a charge-coupled device digital camera (Gatan Erlangshen 785; Gatan, Inc., Warrendale, PA, USA).
The total RNA was extracted from 1-month WT, Rpgrko, Nrl−/− and Rpgr-DKO mice by using RNeasy plus mini kit (Qiagen, USA) according to manufacturer's instructions. The starting sample was consisted of neuroretina + RPE (including choroid). Three biological replicates were included for each genotype. The quantity of RNA was measured by using spectrophotometer (Nanodrop, Thermo Scientific, USA) and quality of all samples was examined on the Agilent 2100 bio-analyzer using the RNA 6000 Nano Chip (Agilent Technologies, Santa Clara, CA, USA). We only included RNA samples with a RNA integrity number value >8. All subsequent steps were performed at Beijing Genomics Institute (Hong Kong). RNA samples were first treated with DNase I, followed by purification of poly (A) RNA using magnetic beads. To prepare libraries, mRNA was fragmented into short fragments, and the first-strand of cDNA was synthesized using random hexamer-primed reverse transcription. Buffers, dNTPs, RNase H and DNA polymerase I were added to synthesize the second-strand. The double-strand cDNA was purified using magnetic beads followed by end repair and Ion Proton adaptors were ligated to the ends of these fragments. Ligated products were selected by size and purified by using Tris-acetate-EDTA-agarose gel. Finally, the fragments were enriched using PCR amplification, purified using magnetic beads and dissolved in appropriate amount of Epstein–Barr solution. Quality of the sample library was determined by using Agilent 2100 Bioanalyzer and these products were sequenced via the Ion Proton platform according to the manufacturer's specifications. Data were collected and bioinformatics analysis was performed.
To estimate gene expression levels, we used RPKM (reads per kilobase per million mapped reads) values. To estimate differential expression of genes, we calculated log2 ratio from the RPKM values. To identify the genes with potential biological significance, we selected based on their log2 values either >1 or <−1, FDR of <0.05 and P-value < 0.05. We selected genes that were significantly altered at least 2-fold difference in all biological replicates. For pathway and network analyses, we used IPA (Ingenuity Systems, Redwood City, CA, USA; www.ingenuity.com) and DAVID software. The analysis was performed in two ways: (i) with all tissues and (ii) with retina only.
Differentially expressed genes were chosen for further validation by qRT-PCR analysis. The total RNA from all the mice by using an RNeasy plus mini kit (Qiagen), with gDNA eliminator columns to remove any contaminating genomic DNA. Primers were designed to amplify fragments of 120–200 bp, with shorter fragments preferred. Approximately 1 µg of total RNA was reverse transcribed to cDNA by using a Verso™ cDNA Kit (Thermo Fisher Scientific) and qPCR was performed on Bio-Rad CFX96 Real-time PCR Detection System (Bio-Rad) by using SsoFast EvaGreen Supermix (Bio-Rad). Reactions were performed in duplicates for at least three biological replicates. The Ct values were normalized to a loading control (glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and β-tubulin). Gene-expression fold changes were calculated using the ΔΔCT method where the relative expression is calculated as 2−ΔΔCt and Ct represents the threshold cycle. The data are presented based on the published guidelines (85). Fold change was calculated considering relative gene expression of control sample as ‘1’.
Retinal fractionation to obtain the F- and G-actin fractions was performed as described (37,86,87). Briefly, fresh retinas from WT and Rpgrko (six retinas from each genotype) were homogenized in F-actin stabilization buffer (50 mm piperazine-N,N′-bis(2-ethanesulfonic acid at pH 6.9, 50 mm NaCl, 5 mm MgCl2, 5 mm ethylene glycol tetraacetic acid, 5% glycerol, 0.1% NP40, 0.1% Triton X-100, 0.1% Tween 20, 0.1% β-mercaptoethanol, 1 mm adenosine triphosphate, protease inhibitor cocktail) and passed several times through a 29-G syringe. The lysate was incubated at 37°C for 10 min and then centrifuged at 350 × g for 5 min to remove cell debris. Supernatant was collected and centrifuged again at 100 000 × g for 1 h. The resultant supernatant was transferred immediately into another tube and was used as the G-actin fraction. The pellet was resuspended in equal volume of the buffer containing 100 µm cytochalsin D and kept on ice for 60 min with gentle mixing for every 15 min. Samples were then sonicated and centrifuged and the supernatant was collected for F-actin fraction. The G- and F-actin fractions were analyzed by immunoblotting. Fold change was calculated considering relative band intensity of control sample as ‘1’.
RhoA activation assay
RhoA activity assay was performed with the use of a Rho activation assay kit (Cytoskeleton) in accordance with the manufacturer's instructions. Briefly, fresh retinas (six retinas each genotype) were dissected and were lysed in ice-cold lysis buffer. The resultant homogenate was centrifuged at 1700 × g for 10 min at 4°C. Supernatant was collected into fresh tube and 20 µg of GST or GST-RTKN conjugated beads were added in the presence of GTPγS. These samples were incubated on a nutator for 1 h at 4°C. Samples were then centrifuged at 4000 rpm and the beads were washed three times with wash solution. Final bead pellet was resuspended in sample buffer, boiled for 5 min and analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and immunoblotting. Fold change was calculated considering relative band intensity of control sample as ‘1’.
Fatty acid profiles were determined for retina. Total lipids were extracted following the method of Bligh and Dyer (88) with modifications (89). To each lipid extract were added 15:0 and 17:0 as internal standards. The lipid extracts were subjected to acid hydrolysis/methanolysis to generate fatty acid methyl esters (FAMEs) (90). FAMEs were quantified using an Agilent Technologies 6890N gas chromatograph with flame ionization detector (GC) (91). The results were also confirmed by GC-MS.
All data are presented as means ± SEM. Data groups were compared by non-parametric t-tests (two groups) or multiple t-tests (more than two groups) using the GraphPad software. Differences between groups were considered statistically significant if P < 0.05. The statistical significance is denoted with asterisks (*P = 0.01–0.05; **P = 0.001–0.01; ***P = 0.0001–0.001 and ****P ≤ 0.00001).
This work is supported by grants from National Eye Institute (EY022372); University of Massachusetts Center for Clinical and Translational Sciences (UMCCTS); Foundation Fighting Blindness; Massachusetts Lions Eye Research Foundation (MLERF) and University of Massachusetts Cell Biology Confocal Core and Electron Microscopy Core (award no. S10RR027897).
We thank Dr John Heckenlively and Dr Visvanathan Ramamurthy for critical discussions about the data.
Conflict of Interest statement. None declared.