A Single, Abbreviated RPGR-ORF15 Variant Reconstitutes RPGR Function In Vivo

Abstract

purpose. The retinitis pigmentosa GTPase regulator (RPGR) is essential for the maintenance of photoreceptor viability. RPGR is expressed as constitutive and ORF15 variants because of alternative splicing. This study was designed to examine whether the retina-specific ORF15 variant alone could substantially substitute for RPGR function. A further objective was to test whether the highly repetitive purine-rich region of ORF15 could be abbreviated without ablating the function, so as to accommodate RPGR replacement genes in adenoassociated virus (AAV) vectors.

methods. A cDNA representing RPGR-ORF15 but shortened by 654 bp in the repetitive region was placed under the control of a chicken β-actin (CBA) hybrid promoter. The resultant construct was transfected into mouse embryonic stem cells. Clones expressing the transgene were selected and injected into mouse blastocysts. Transgenic chimeras were crossed with RPGR knockout (KO) mice. Mice expressing the transgene but null for endogenous RPGR (Tg/KO) were studied from 1 month to 18 months of age by light and electron microscopy, immunofluorescence, and electroretinography (ERG). The results were compared with those of wild-type (WT) and RPGR-null control mice.

results. Transgenic RPGR-ORF15 was found in the connecting cilia of rod and cone photoreceptors, at approximately 20% of the WT level. Photoreceptor morphology, cone opsin localization, expression of GFAP (a marker for retinal degeneration) and ERGs were consistent with the transgene exerting substantial rescue of retinal degeneration due to loss of endogenous RPGR.

conclusions. RPGR-ORF15 is the functionally significant variant in photoreceptors. The length of its repetitive region can be reduced while preserving its function. The current findings should facilitate the design of gene replacement therapy for RPGR-null mutations.

Mutations in retinitis pigmentosa GTPase regulator (RPGR) account for >70% of X-linked RP and approximately 10% of all RP cases. ^{1 2 3} Ablation of the RPGR gene in mice ⁴ and RPGR mutations in Siberian huskies ⁵ also lead to photoreceptor degeneration, suggesting that RPGR is essential for mammalian photoreceptor survival. In both patients with RPGR mutations ^{6 7} and mice without RPGR, ⁴ early cone photoreceptor defects, in addition to rod degeneration, have been noted, indicating that RPGR is necessary for normal function of both rods and cones. 

RPGR transcripts undergo a complex splicing process and generate constitutive and ORF15 variants by using alternative polyadenylation sites and splicing sites. ^{2 8 9 10} Both variants have the same N-terminal domain that shares sequence homology with the regulator of chromatin condensation, ¹ a nuclear protein that catalyzes guanine nucleotide exchange for the small GTPase Ran. However, their remaining C-terminal domains are entirely different. The ORF15 variant includes a long stretch of purine-rich region encoding alternating glycine and glutamic acid residues. RPGR constitutive variants are found in most tissues, whereas ORF15 variants are highly expressed in the retina. ^{2 11} In mice, the RPGR-ORF15 protein is found primarily in photoreceptors. ¹¹ A large number of disease-causing mutations in the ORF15 exon, but none in the exons unique to the constitutive variant, have been found in X-linked patients with RP, ^{7 12} implying that the ORF15 variants are functionally significant in photoreceptors. The function of RPGR is not fully understood. In cone and rod photoreceptors, RPGR is concentrated in the connecting cilia, ^{4 11} thin bridges that join the biosynthetic inner segments (IS) and light-sensing outer segments (OS). A primary defect in mice without RPGR is the mislocalization of cone opsins in the photoreceptors. ⁴ These observations suggest that RPGR may be involved in the regulation of protein trafficking through the connecting cilia. 

The multitude of variants expressed from the RPGR gene poses challenges to its functional studies as well as to the design of gene-replacement therapies. The present study was performed to test the hypothesis that the retina-specific RPGR-ORF15 variant alone may be sufficient for normal photoreceptor function. Toward this objective, rescue of the RPGR-null phenotype was attempted by expressing an ORF15 transgene in an RPGR KO background. A secondary question was whether the repetitive region of the ORF15 variant can be shortened without ablating the protein function. This question arose because the predicted full-length RPGR-ORF15 variant is >4.5 kb and approaches the packaging size limit for currently available AAV vectors. We therefore designed our transgene construct so that it included an in-frame deletion in the purine-rich repetitive region. 

Materials and Methods

Generation of Transgenic Mice

A cDNA clone (T12) representing the mouse RPGR-ORF15 variant was obtained by reverse transcription and PCR (RT-PCR). ¹⁰ Clone T12 is among the largest cDNAs that can be amplified from the C57BL6 mouse retina and carries an in-frame deletion of 654 bp in the purine-rich region compared with the published sequence determined from the 129/Sv mouse strain. ² Clone T12 was subcloned into the pCBA vector between the cytomegalovirus (CMV) enhancer β-actin promoter (CBA) and a bovine growth hormone (BGH) polyadenylation sequence. To generate the transgenic lines, the construct was transfected into embryonic stem (ES) cells by electroporation. Neomycin-resistant clones were picked and screened by immunoblotting for RPGR-ORF15 expression. Three ES clones expressing an RPGR-ORF15 variant were injected into C57BL6 blastocysts to generate chimeras. Male chimeric founders were crossed with RPGR knockout (KO) females. Offspring mice carrying the transgene in an RPGR KO background were genotyped by PCR, with the primer pair RPGR50 (5′-CGAGTCGCCCCTGCTATCTTTCCGA) and RPGR8R (5′-TGGTTCCTCCACAGGCAGC). One ES clone propagated the transgene through the germ line. Transgenic mice established from this clone were analyzed in detail. All experiments involving animals were performed in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Histology and ERG Analyses

Histology and ERG analyses were performed as previously described. ⁴ For morphometric analyses of the photoreceptor outer segment (OS) length and the outer nuclear layer (ONL) thickness, measurements were taken at five consecutive points, each separated by 100 μm, beginning at 300 to 400 μm from either side of the optic nerve head. To perform ERG testing, we dark adapted the mice overnight. Dark-adapted ERGs to generate rod-dominated responses were elicited with 10-μs flashes of white light (4.3 log foot lambert) in the dark presented in a Ganzfeld dome. Light-adapted ERGs to isolate cone function were elicited with the same white light (4.3 log foot lambert) presented at 2 Hz in the presence of a background light of 12 foot lambert. All responses represent the average of four traces. 

Immunoblotting and Immunofluorescence

For detection of RPGR and RPGR interacting protein (RPGRIP), eyes were embedded in optimal cutting temperature (OCT) compound without fixation and quick frozen in liquid nitrogen. For detection of glial fibrillary acidic protein (GFAP) and cone opsins, eyes were fixed in 4% formaldehyde/PBS overnight, embedded in OCT, and frozen. Cryosections at 10-μm-thick were cut and collected on pretreated glass slides. Immunofluorescence procedures and primary antibodies have been described. ¹¹ For immunoblot analysis, retinas were homogenized in an SDS-protein sample buffer. Total retinal homogenates were used for detection of RPGR variants on immunoblots. 

Statistical Analyses

Differences in mean ERG amplitude, implicit time, ONL thickness, and OS length by genotype were evaluated by Student’s t-test, with correction for unequal group variances where relevant. Analyses were performed on computer (JMP, ver. 3.2; SAS Institute, Cary, NC). 

Results

Expression of an Abbreviated RPGR-ORF15 in Transgenic Mice

The mouse RPGR-ORF15 cDNA was cloned into a transgenic vector under the transcriptional control of a hybrid CBA ¹³ promoter (Fig. 1) . This promoter was non–tissue specific and was expected to drive expression in most cells including both rods and cones. Indeed, the transgene was expressed not only in ES cells but was expressed in multiple tissues including the retina (described later) and the brain (data not shown). Use of nonspecific promoter allowed us to take an unconventional approach to the generation of transgenic mice. Instead of performing pronuclear injections followed by screening multiple lines of mice for evidence of transgene expression, we prescreened neomycin-resistant embryonic stem (ES) cell clones for transgene expression before proceeding with the generation of transgenic mice. We were thus able to achieve transgene expression with less effort and with fewer animals involved. 

The level of RPGR-ORF15 transgene expressed in the retina was quantified by immunoblot analysis in mice 2 months of age by using an anti-RPGR antibody (Fig. 2A) . A band migrating at approximately 210 kDa was detected in transgenic mice in an RPGR-null background (Tg/KO). This protein was slightly smaller than the endogenous WT RPGR-ORF15 and was present in a lower amount. Quantification of the transgenic RPGR protein indicated that it was expressed at approximately 20% of the endogenous RPGR level in WT mice (Fig. 2B) . As expected, the constitutive RPGR protein was not detected in the transgenic mice, indicating that RPGR-ORF15 was the sole RPGR variant in the Tg/KO animals (Fig. 2A) . 

Localization of Transgenic RPGR-ORF15 to the Connecting Cilia

Endogenous RPGR, both of the ORF15 and constitutive isoforms, is highly concentrated in the connecting cilia of photoreceptors. ^{4 10 11} To determine whether the transgenic RPGR exhibits the same characteristics as the endogenous WT RPGR proteins, we evaluated its subcellular localization by immunofluorescence. Retinal sections from transgenic and WT control mice were stained with the RPGR S1 antibody (targeting epitopes common to all RPGR variants) or the ORF15 antibody (targeting epitopes unique to the ORF15 variants). Specific staining was detected at the junction between the IS and OS of photoreceptors, corresponding to the connecting cilia (Figs. 3A 3B) . Intensity of staining in the transgenic retina appeared weaker than in WT retina, consistent with a lower level of the transgenic protein than that of the endogenous RPGR, as shown by immunoblot analysis. Compared with the uniform expression of RPGRIP, a resident protein in the connecting cilia, ^{14 15} there appeared to be variability in the signal intensities of transgenic RPGR-ORF15 among individual photoreceptors (Fig. 3C) . Furthermore, colocalization with RPGRIP (Fig. 3C)provided confirmation that the transgenic RPGR-ORF15 was indeed localized in the connecting cilia. 

Structural and Functional Rescue of the RPGR-Null Phenotype by the Transgene

To evaluate the effects of transgene expression on photoreceptor cell survival, we examined both retinal morphology and retinal function in RPGR KO and Tg/KO mice between 1 and 18 months of age. Rod OS of Tg/KO retinas at 1 month of age appeared very similar to those of the WT mice, which differed from the loosely organized nascent disc structures that we saw by means of electron microscopy in the KO retina (data not shown). ⁴ By 14 months of age, both the IS and OS of the KO mice became shortened, and more than half of the photoreceptors cells were lost (Fig. 4A) . Tg/KO retinas at this age showed little evidence of cell loss and had approximately normal lengths of the IS and OS (Fig. 4A) . By morphometric analyses, we quantified ONL thickness and OS length at 14 months of age in Tg/KO mice, their KO littermates, and age-matched WT control mice (Fig. 4B) . Mean ONL thickness was 73% greater and mean OS length was 99% longer in Tg/KO retinas than in KO retinas (P = 0.008 and 0.002, respectively). Compared with WT mice at 14 months of age, transgenic mice showed a nearly complete restoration of normal retinal morphology at the light microscopy level (Fig. 4B) . 

Retinal morphology was also assessed by electron microscopy between 14 and 18 months of age (Fig. 5) . At both ages, the lengths of the IS and OS in the KO retinas were markedly shorter than those in the Tg/KO and WT retinas. The KO retinas had highly disorganized OS and contained membranous whirls derived from disc materials. In contrast, many photoreceptors in the Tg/KO retinas maintained relatively well aligned and tightly packed disks, though some also showed disorganization in the OS, such as membrane vesiculation and irregular disc spacing. The latter observation may be attributable to variable expression of the transgene in individual photoreceptors noted earlier and hence insufficient rescue in the lower-expressing photoreceptors. 

Mislocalization of cone opsins in cone photoreceptor cell bodies and synapses was a prominent phenotype in the KO mouse retinas from an early age, as previously described. ⁴ To observe the phenotype rescue in cone photoreceptors, we examined the Tg/KO, KO, and WT control retinas by immunofluorescence for cone opsins from 1 through 14 months of age (Fig. 6A) . At all ages examined, opsins in the KO cone photoreceptors showed mislocalization in the IS, perinuclear regions, and synaptic terminals. In contrast, cone opsin staining in Tg/KO, as in the WT, was confined to the cone OS, indicating restoration of RPGR function in cone cells. The number of cones in the KO retina was considerably reduced by 14 months of age, with the remaining cones showing severely shortened OS (Fig. 6A) . In contrast, the number and integrity of cone cells was well maintained in the transgenic mice, demonstrating the rescue effect of RPGR-ORF15 in these cells. 

Upregulation of glial fibrillary acidic protein (GFAP) expression in the retina is a nonspecific marker of retinal degeneration. In the RPGR KO mouse retinas, GFAP upregulation was apparent from 3 months of age onward. ⁴ As an additional outcome measure for phenotype rescue by the transgene, we examined GFAP expression in Tg/KO and control animals at different ages. As expected, GFAP was markedly upregulated in the KO retinas in mice from 3 through 14 months of age. Virtually no GFAP signal was detected in the Tg/KO retinas at all ages tested (Fig 6Band data not shown). 

To determine the extent of retinal function restoration by the ORF 15 transgene, we recorded ERGs at 14 to 16 months of age in Tg/KO, littermate KO, and age-matched WT mice. Results of dark-adapted (rod) and light-adapted (cone) ERGs are shown in Figure 7 . The mean rod b-wave amplitude was 52% larger and the mean cone b-wave amplitude was 80% larger in Tg/KO mice than in littermate KO mice (P = 0.02 and P = 0.001, respectively). The mean rod ERG a-wave amplitude and ERG implicit times also showed significant improvements in the Tg/KO mice over the KO mice (data not shown). Although ERGs showed a marked improvement in retinal function, the mean ERG amplitude in the Tg/KO mice did not reach the WT level (Fig. 7) . This finding may be related to the morphologic examinations wherein incomplete restoration of photoreceptor morphology was seen at the ultrastructural level and was probably caused by the much lower and variegated transgene expression compared with that of the endogenous WT RPGR. Despite this observation, overall analyses in the transgenic and control mice clearly demonstrated that gene replacement intervention with an abbreviated RPGR-ORF15 provided substantial structural and functional rescue of mutant photoreceptors that lacked RPGR. 

Discussion

In this study, we tested the hypothesis that the retina-specific ORF15 variant is the functionally important isoform of RPGR in photoreceptors. We introduced an ORF15 transgene into the RPGR-null background and observed the transgenic mice for evidence of phenotypic rescue with a variety of assays. Characteristics of the photoreceptor phenotype in the RPGR mutant mice, as previously reported, ⁴ include mislocalization of cone opsin from a very early age, histopathological changes in photoreceptor OS, revealed by electron microscopy, and photoreceptor degeneration and ensuing decline of retinal function, as shown by ERG. Introduction of the ORF15 transgene substantially reversed all these phenotypic features in the RPGR KO retinas. Considering that the transgenic RPGR-ORF15 is expressed at only one fifth the level of WT endogenous RPGR, the ORF15 variant alone appears sufficient to reconstitute most, if not all, of the RPGR function in cone and rod photoreceptors. 

The transgenic RPGR-ORF15 construct has an in-frame deletion within the purine-rich repetitive region of ORF15. By immunofluorescence, we observed a photoreceptor-to-photoreceptor variation in the level of transgene expression. Such variegated expression may be related to the site of transgene integration and the inherent properties of the promoter. Overall, transgene RPGR-ORF15 was expressed at approximately 20% of the endogenous WT level and was found to localize correctly in the connecting cilia of photoreceptors, indicating that the abbreviated ORF15 behaved as the WT protein. This relatively low level of expression was unexpectedly shown to rescue substantially the RPGR KO phenotype both functionally and morphologically. This suggests that a complete restoration of photoreceptor function and viability by ORF15 alone is likely if its expression level could be increased and made uniform. This is an encouraging finding for future viral-mediated gene therapy studies wherein expression of moderate levels of an abbreviated RPGR-ORF15 could be an effective treatment for patients with an RPGR-null mutation. It is also notable that in this study transgene expression was stable, and the effect of rescue was long lasting. 

Gene transfer vectors derived from recombinant AAV are currently the most effective means for gene delivery to retinal photoreceptors. ^{16 17} AAV is a nonpathogenic Dependovirus that cannot reproduce without the presence of a helper virus and is not known to be associated with any human disease. AAV-based vectors are able to provide long-term transgene expression, mostly by forming stable, transcriptionally active monomeric and concatameric episomes. In several gene-therapy studies, the efficacy of AAV vectors has been demonstrated by successful delivery of transgenes to the photoreceptor cells followed by anticipated therapeutic outcomes. However, one downside of AAV-based vectors is their relatively small packaging capacity of approximately 4.7 kb. Consequently, the application of an AAV system may depend on the required minimum size of functional RPGR expression cassettes. The reported full-length RPGR-ORF15 in combination with commonly used promoter would exceed this size limit. In this study, we were able to rescue the RPGR KO phenotype by transgenic expression of an internally truncated RPGR-ORF15. This suggests a lack of tight functional constraint on the length of the repetitive region in the ORF15 protein. Our data thus make it possible to package efficiently a functional RPGR expression cassette in an AAV vector. 

In summary, the histopathological, immunocytochemical, and electrophysiological data in this study clearly demonstrate a substantial, long-term structural and functional rescue of photoreceptors in RPGR KO mice expressing a shortened RPGR-ORF15 transgene. This study also shows that the purine-rich repetitive region can be reduced while maintaining its function. Our findings will aid in the design of replacement gene therapy for RPGR-null mutations. 

 Supported by National Eye Institute Grants EY10581 and EY14188 and the Foundation Fighting Blindness.

 Submitted for publication September 8, 2004; accepted October 25, 2004.

 Disclosure: D.-H. Hong, None; B.S. Pawlyk, None; M. Adamian, None; M.A. Sandberg, None; T. Li, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Schematic representation of the RPGR gene structure and RPGR-ORF15 transgene construct. CBA, chicken β-actin hybrid promoter; PolyA, polyadenylation signal derived from the bovine growth hormone gene.

Figure 2.

 

Expression analysis of RPGR-ORF15 transgene in the transgenic mice. (A) Immunoblotting analysis for RPGR in Tg/KO, KO, and WT mouse retinas at 2 months of age, using an RPGR antibody. Immunoblotting for α-tubulin was included as a loading control. Both endogenous and transgenic RPGR-ORF15 variants are seen as migrating just below the 250-kDa marker. In addition, a second band migrating at 150 kDa was found in the Tg/KO retinas presumed to result from alternative splicing. ¹⁰ The constitutive variant of RPGR was absent in the Tg/KO or the KO retinas. (B) Quantification of transgene expression. Levels in WT mice are set as 100 arbitrary units. The data are expressed as means ± SEM.

Figure 3.

 

Immunofluorescence staining for RPGR protein variants. Frozen retinal sections from WT, KO, and Tg/KO mice were stained with antibodies, as indicated. The S1 antibody (A) was common to all RPGR variants, whereas the ORF antibody (B) was specific for the ORF15 variant. RPGR staining is shown in red. Sections were counterstained with Hoechst dye 33342 to highlight cell nuclei (blue). (C) Colocalization of RPGR (red) and RPGRIP (green) in the connecting cilia (arrows). RPE, retinal pigment epithelium; OS, outer segments; IS, inner segments; ONL, outer nuclear layer; INL, inner nuclear layer.

Figure 4.

 

Rescue of photoreceptor degeneration by transgenic RPGR-ORF15, as shown by light microscopy. (A) Light micrographs of representative retinal sections from WT, Tg/KO, and KO littermates at 14 months of age. Note the preservation of OS length and the ONL thickness in the Tg/KO retina compared with those of the KO. (B) Quantification (mean ± SE) of ONL thickness and OS length in WT (n = 3), Tg/KO (n = 3), and KO mice (n = 3) at 14 months of age.

Figure 5.

 

Transmission electron micrographs showing the ultrastructure of the IS and OS in the WT, Tg/KO, and KO retinas at 14 and at 18 months of age. Bar, 5 μm.

Figure 6.

 

Immunofluorescence analyses of the retinal phenotype. (A) Frozen sections from Tg/KO and KO littermates at 1 month and 14 months of age were stained with cone opsin antibodies. Mislocalization of the cone opsins was evident in the KO retinas at both ages, whereas this phenotype was absent in the Tg/KO retinas. (B) Upregulation of GFAP, a marker of photoreceptor degeneration, was apparent in the KO retinas at all ages after 3 months but was absent in the Tg/KO retinas in all ages tested. Shown are images at 14 months of age.

Figure 7.

 

Functional rescue assessed by ERG analyses. Rod ERG b-wave amplitudes from KO (n = 9), Tg/KO (n = 11), and WT (n = 13) mice and cone ERG b-wave amplitudes from KO (n = 9), Tg/KO (n = 11), and WT (n = 9) mice. Each symbol and error bars represent the mean ± SEM.

The authors thank Eliot L. Berson for helpful discussions.