|Home | About | Journals | Submit | Contact Us | Français|
Leber congenital amaurosis (LCA), a severe autosomal recessive childhood blindness, is caused by mutations in at least 15 genes. The most common molecular form is a ciliopathy due to NPHP6 (CEP290) mutations and subjects have profound loss of vision. A similarly severe phenotype occurs in the related ciliopathy NPHP5 (IQCB1)-LCA. Recent success of retinal gene therapy in one form of LCA prompted the question whether we know enough about human NPHP5 and NPHP6 disease to plan such treatment. We determined that there was early-onset rapid degeneration of rod photoreceptors in young subjects with these ciliopathies. Rod outer segment (OS) lamination, when detectable, was disorganized. Retinal pigment epithelium lipofuscin accumulation indicated that rods had existed in the past in most subjects. In contrast to early rod losses, the all-cone human fovea in NPHP5- and NPHP6-LCA of all ages retained cone nuclei, albeit with abnormal inner segments and OS. The rd16 mouse, carrying a hypomorphic Nphp6 allele, was a good model of the rod-dominant human extra-foveal retina. Rd16 mice showed normal genesis of photoreceptors, including the formation of cilia, followed by abnormal elaboration of OS and rapid degeneration. To produce a model of the all-cone human fovea in NPHP6-LCA, we generated rd16;Nrl−/− double-mutant mice. They showed substantially retained cone photoreceptors with disproportionate cone function loss, such as in the human disease. NPHP5- and NPHP6-LCA across a wide age spectrum are thus excellent candidates for cone-directed gene augmentation therapy, and the rd16;Nrl−/− mouse is an appropriate model for pre-clinical proof-of-concept studies.
Leber congenital amaurosis (LCA) is a genetically heterogeneous retinal blindness detected in infancy (1,2). Whereas one molecular form of LCA (due to mutations in the RPE65 gene) has been treated safely and successfully using retinal gene therapy (3), there are at least 14 more genetic subtypes. The challenge is now to determine which of these other forms are treatable and how. Among the many scientific and medical hurdles to clear en route to a clinical trial, a very basic one demands early attention. When available animal models of a human disease exist, the question should be asked whether there is sufficient similarity between the expressions of the retinal disease in the animals and humans to warrant proof-of-concept experiments in the models.
An LCA disease of strong interest for gene therapy is the most common and a severe molecular form—that associated with Nephrocystin-6 (NPHP6, also known as CEP290) mutations (1,4,5). The NPHP6 gene product is localized primarily to the (connecting) cilium (CC) adjoining the outer segment (OS) to the photoreceptor cell body (6,7), and is thought to be involved in the formation of microtubule-membrane linkers in this region (8). The loss of NPHP6 function may affect cilia biogenesis (9,10), or protein composition of the photoreceptor OS (7,8). NPHP6 interacts with other cilia proteins, including RPGR (7), which, when mutated, causes retinal degeneration (2). As NPHP6 is expressed in all primary cilia, many mutations cause syndromic diseases involving the central nervous system, kidney and liver, in addition to retinal degeneration (6,11–13). Mutations in NPHP5 (also known as IQCB1) are associated with similar ciliopathies and the encoded nephrocystin also interacts with RPGR (14–16). Mutations in NPHP5 cause LCA (14,17,18).
What pre-clinical models are available for proof of concept in NPHP5- and NPHP6-LCA? An alga mutant with a large deletion in the NPHP6 homologue shows very short cilia; the phenotype can be corrected when supplied with the wild-type (WT) protein (8). The Nphp6-deficient Abyssinian cat (19) is a large animal model with a disease course that is not reminiscent of human NPHP6-LCA. The cat has a relatively slow time course of degeneration, substantial retinal physiological responses and rod visual pigment content until late in the disease (20,21). The rd16 mutant mouse with a hypomorphic Nphp6 mutation (7), however, has a rapid retinal degeneration that appears more like the severe and early blindness of the human disease.
We addressed whether human NPHP5 and NPHP6-LCA could be moved closer to being treatable diseases by exploring the fate of rod and cone photoreceptors in humans. Humans with little or no vision still showed some rod cells with abnormalities in the rod OS, but rods disappeared early in life; details of the rod ciliopathy were studied in rd16 mice. The all-cone human fovea was retained in humans at most ages despite the disproportionate vision loss. In order to model the retained foveal cones in patients, we generated mice with all-cone retinas and Nphp6 dysfunction by crossing rd16 mice (7) with Nrl−/− mice (22). An extreme dissociation of cone function and structure in the rd16;Nrl−/− retina, like that in the human ciliopathies, was identified, as well as salient features of the cone pathogenesis. We conclude that the next and most expeditious experimental steps leading to a human clinical trial would best be performed on a mammalian model, such as the rd16;Nrl−/− double-mutant mouse, with cone function as the metric of importance.
NPHP5- or NPHP6-LCA cause congenital blindness due to photoreceptor malfunction, but details of the disease in the rods, representing ~94% of human retinal photoreceptors (23), are unknown. Whether rod photoreceptors even develop in the human diseases is unknown—a critical fact for designing realistic therapies for these forms of LCA. Most of the human mutant retinas we studied showed no detectable photoreceptors outside the central region. Surprisingly, in the mid-peripheral retina of two young patients (NPHP6-P2, age 8, and NPHP5-P16, age 7, Table 1), the outer nuclear layer (ONL), the retinal layer comprised of photoreceptor nuclei, was clearly visible starting at 5–6 mm eccentricity from the fovea (Fig. 1A). A region of major photoreceptor loss, with barely detectable ONL and little or no discernible inner segment (IS) and OS, was present between the retained center and mid-peripheral ONL (Fig. 1A). In NPHP5-P16, spatial topography of ONL thickness across a large expanse of the retina showed a wide annular region lacking photoreceptors and encircled by a midperipheral region of retained ONL (Fig. 1B). In contrast, NPHP6-P3 at age 10 had barely detectable ONL in the mid-peripheral regions (Fig. 1A).
Quantitative signal processing of longitudinal reflectivity profiles (LRPs) in the superior and inferior regions with ONL between eccentricities of 7–9 mm (Fig. 1A, white rectangles) were performed. Unexpectedly, several retinal laminae in P2 and P16 were thicker than normal. These thicker layers included the outer plexiform layer (OPL) where photoreceptors synapse to secondary neurons, and the inner nuclear layer (INL) where bipolar, Müller cell and other nuclei normally reside (Table 2). Also, ONL thickness in both patients was at the upper limit of normal or greater than normal (Table 2). At this retinal eccentricity, rods outnumber cones 28:1 in normal human retinas (23); thus, the dominant cell in the ONL would be rod photoreceptor nuclei. Thickening of layers has been suggested to represent the earliest signs of photoreceptor stress, and a phase that precedes rapid photoreceptor degeneration (24–27).
The structural abnormalities distal to the rod photoreceptor nuclei between the ONL and retinal pigment epithelium (RPE) were studied with quantitative LRP analysis in superior and inferior retinal regions (Fig. 1C). In normal retinas, there are distinct local backscatter peaks associated with the outer limiting membrane (OLM) forming the intercellular adherens junctions between photoreceptors and Müller glia; the junction at the CC between IS and OS; and reflections originating near the apical boundary of the RPE and between RPE and Bruch membrane (Fig. 1C). In P2 and P16, however, these peaks are replaced with a region of similar thickness but with diffusely increased light scatter (S+) likely originating from abnormal rod IS/OS (Fig. 1C).
In retinas devoid of rod photoreceptors, we searched for their ‘tombstones' within RPE cells which can survive for decades even after the loss of photoreceptors. Lipofuscin, an RPE pigment (28), continuously accumulates with age through phagocytosis of OS tips that contain the photoisomerization product all-trans retinal generated during the retinoid cycle of vision (29). In normal eyes, RPE lipofuscin distribution revealed by short-wavelength autofluorescence (SW-AF) imaging shows a low signal at the fovea, gradually increasing to a maximum at ~2–6 mm eccentricity, and then decreasing towards the periphery (Fig. 2A, left). This topography is similar to the spatial density of rod photoreceptors in human retinas (23).
Intraretinal variations of RPE lipofuscin signal in NPHP5-P16 (Fig. 2A, right) can be interpreted in terms of its co-localization with photoreceptor topography (Fig. 1B). Centrally, there is an ~3 mm wide elliptical region of low SW-AF intensity similar to that in normal individuals (Fig. 2A); this region corresponds to the retained central island of photoreceptors dominated by cones (Fig. 1B). Peripherally, a distinct elliptical hyperautofluorescent boundary ~15 mm wide and ~11 mm high (Fig. 2A) co-localizes with the beginning of retained photoreceptors in an area dominated by rods. Further eccentric to the boundary, SW-AF intensity is lower but still higher than normal. Unexpectedly, between the central and peripheral regions, there is an intervening annulus with a detectable lipofuscin signal (Fig. 2A) but no detectable photoreceptors (Fig. 1B). This region likely contained photoreceptors that have since degenerated (Supplementary Material, Fig. S1; Supplementary Text). The accumulated lipofuscin provides support for previous synthesis of its diretinal bisretinoid precursors within photoreceptors. These precursors, in turn, would have required the generation of all-trans retinal from photoisomerization of the retinal chromophore after absorption of light within the mutant photoreceptor OS.
Three other patients illustrate RPE lipofuscin distribution at different stages of NPHP5- and NPHP6-LCA disease. NPHP6-P10 (age 26) shows RPE lipofuscin topography with a broad annulus of higher signal at eccentricities beyond ~1 mm; reduction of signal near the fovea centralis is expected from the absorption of macular pigment and RPE melanin (Fig. 2B). Quantitatively, lipofuscin autofluorescence intensity along the horizontal meridian is near the upper limits of normal (Fig. 2C); most of the central retina has retained detectable photoreceptors (nuclei and a layer of abnormal IS and OS) even though the patient has no perception of light (Table 1). RPE lipofuscin topography in P10 differs substantially from that in other genetic retinal degenerative conditions associated with severe visual loss (Supplementary Material, Fig. S2; Supplementary Text).
NPHP6-P7 (age 19) and NPHP5-P19 (age 23) demonstrate more advanced disease stages having smaller extents of detectable photoreceptors (Fig. 2C). Large regions devoid of photoreceptors surrounding the central elliptical islands have surprisingly high lipofuscin signaling (Fig. 2B). Quantitatively, lipofuscin intensity is mostly within normal limits (Fig. 2C). Along a subset of meridians, both P7 and P19 show a narrow intermittent hyperautofluorescent band near the edge of the central region of retained photoreceptors and the surrounding region of neurodegeneration (Fig. 2B). In P19, a further disease boundary is apparent peripherally where spatially heterogenous RPE atrophy starts with concordant lipofuscin loss (Fig. 2B). Melanin topography in all three patients also shows retained RPE with partial demelanization corresponding to the regions devoid of photoreceptors but retained lipofuscin; this validates the interpretation from lipofuscin topography (Supplementary Text). Of note, the pattern of centrally retained photoreceptor and RPE cells would not be surprising in typical retinitis pigmentosa (RP) in which patients have loss of photoreceptors in adult life after decades of useful vision. Retinal regions in RP, not unexpectedly, also show historical evidence of photoisomerizations underlying the vision (Supplementary Material, Fig. S3). In NPHP5- and NPHP6-LCA, on the other hand, evidence of lipofuscin is unexpected not only because of the profound visual impairment at the time of imaging but also considering a lack of vision even in the first few months of life.
The results in NPHP5- and NPHP6-LCA indicate that there is differentiation of rod photoreceptors during early development. The distal segments of photoreceptors, however, either do not elaborate normally or they become structurally unstable after maturation causing early and severe loss of visual function. These abnormal rods do not survive and degenerate rapidly within the first decade of life.
To examine further the rod abnormalities in NPHP6 deficiency, we took advantage of the rapid rod photoreceptor degeneration in rd16 mice, which recapitulates NPHP6-LCA and probably NPHP5-LCA (7). Early photoreceptor development [postnatal day (p) 2 to p8] proceeds normally in the rd16 mouse retina. Although at p10 (and p12), all layers of the retina were present and grossly normal with IS and OS lengths comparable with the WT (Fig. 3A and B), OS elongation was clearly stalled at p14 and began to regress (Fig. 3B). Simultaneously, pyknotic photoreceptor nuclei appeared in the ONL which started to thin (Fig. 3A and B, arrows). Only about four layers of nuclei remained at p18 (Fig. 3A) with almost complete loss of IS and OS (Fig. 3B). Only a single row of ONL nuclei remained at p28, and these nuclei belonged exclusively to cones (Fig. 3F).
As photoreceptor OS are considered modified primary cilia (30), we examined the CC, basal body (BB) and related structures at the ultrastructural level (Fig. 3C–E). The BB, derived from the mother centriole, anchors and gives rise to the primary cilium, with two distinct parts—CC and ciliary axoneme that provides the backbone for membranous OS discs. The CC is intact and positioned appropriately in rd16 compared with the WT retina; however, the axoneme is reduced or absent in rd16 (Fig. 3C, asterisk), a finding consistent with the failure of OS elongation because of a cilial transport defect in rd16 photoreceptors. Thus, OS remain rudimentary and fail to extend in the rd16 retina, even though the process of cilia biogenesis is initiated (Fig. 3D and E).
High-resolution imaging at the cone-rich foveal region of human retinas confirmed and extended the puzzling finding of profound blindness in patients with NPHP6-LCA despite retained cone photoreceptor nuclei (31). We therefore defined the regions with healthy RPE cells that could support photoreceptors (32) by mapping the melanin pigment they contain (28) using in vivo en face near-infrared autofluorescence (NIR-AF) imaging (33). In a normal control subject, thicker foveal RPE cells (34) have the highest melanin optical density and NIR-AF shows the highest intensity at the foveal region with a gradual decline to ~1 mm eccentricity surrounded by homogenous signal; retinal blood vessels and the optic nerve head show low signal (Fig. 4A, left panel). NPHP6-LCA patients with profound blindness (Table 1) show melanized and structurally sound RPE within central elliptical regions, which are abruptly delineated and surrounded by abnormally reduced melanin signal as in NPHP6-P3 at age 10 (Fig. 4A). We then studied cross-sectional retinal scans to understand the architecture of retinal layers. In the normal eye, the ONL (highlighted blue, Fig. 4B) contains cone photoreceptor nuclei at the fovea and both rod and cone nuclei surrounding the fovea; the ratio of cone to rod densities decreases with eccentricity (23). NPHP6-P3 shows normal ONL thickness at the fovea (Fig. 4B), but gradual thinning of ONL with increasing eccentricity, as would be expected from a retina that had lost rod nuclei but retained cone nuclei. Most NPHP6-LCA patients showed a detectable ONL at the fovea and gradual thinning with eccentricity (31). The retinal phenotype of NPHP5-LCA with en face NIR-AF and cross-sectional imaging was similar to that in patients with NPHP6 mutations (14–18). NPHP5-17 (age 8), for example, had profound blindness (Table 1), retained melanized RPE within a central elliptical region surrounded by abnormally reduced melanin (Fig. 4A) and partially retained ONL thickness at the fovea that gradually thinned with eccentricity (Fig. 4B). All NPHP5-LCA patients showed a similar pattern.
To better understand the relationship between the severe visual loss despite retained cone photoreceptor nuclei, we performed quantitative signal processing (25). In normal foveas, local backscatter characteristics (peaks and troughs) occur distal to the ONL in association with the OLM, CC and reflections originating near the apical and basal boundaries of the RPE (Fig. 4C). Foveal LRPs from NPHP5- and NPHP6-LCA patients (P3, P15, P17 and P18) show the replacement of the distinct OLM and CC signals with a wide band of high backscatter (S+) and a narrower band of low backscatter (S−) between the ONL and the RPE, likely representing remnants of photoreceptor IS and/or OS (Fig. 4C). These results suggest that structural abnormalities occurring distal to the cone photoreceptor nuclei between the ONL and the RPE may contribute to the profound blindness experienced by the patients despite retained photoreceptor nuclei (Fig. 4C).
Our finding of early rod loss but decades-long retention of central cone nuclei in NPHP5- and NPHP6-LCA strongly argues the need for an animal model with a similar phenotype for the design and testing of human treatments. Rd16 mice show only a single row of cone nuclei remaining after rod degeneration (Fig. 3F) and at 6 weeks of age show minimal S- and M-cone-mediated electroretinography (ERG) responses that are near recording noise level and, at best, are 0.9–1.3 log units smaller than WT responses (Fig. 4D and E). To generate an improved mouse model with an enriched cone population for testing treatment efficacy, we generated rd16;Nrl−/− double-mutant mice by breeding rd16 mice with the cone-only Nrl−/− mice (22,35). Rd16;Nrl−/− double-mutant mice (ages 2.5–4 months) showed sizeable S- and M-cone-driven ERG signals, although the amplitudes were reduced on average by 1 log unit compared with Nrl−/− mice (Fig. 4D and E).
We detected significant survival of cone photoreceptors in the rd16;Nrl−/− retina; histology of the eyes shows four to five rows of ONL nuclei (Fig. 4F) at 3–4 months of age (equivalent to a young adult human) when all photoreceptors have degenerated in the rd16 retina. Electron microscopy of rd16;Nrl−/− photoreceptors shows IS with a distinct OLM (Fig. 5A, D, D′), and rare OS with scant OS-like membranes (Fig. 5B, E, E′). BB and occasional CC, however, were observed (Fig. 5C, F, F′). In 3-month-old rd16;Nrl−/− photoreceptors, S-opsin localizes to the cell body and the OPL region instead of the OS as in the WT retina. Notably, no S-opsin is detected in the rd16 retina at this age, consistent with the absence of photoreceptors (Fig. 5G–I).
Here we report finding rod structure in human NPHP5- and NPHP6-mutant retinas. The ONL could be retained in a region of high rod density in some young patients, indicating that rod photoreceptors did develop. These data demonstrate that rod cell loss is the ultimate fate due to a rapid yet progressive degeneration. An alternative hypothesis of rod agenesis due to a possible centrosomal role of NPHP6 (7,11) is not supported. Lipofuscin detected in the RPE in the absence of photoreceptors provides further evidence that rods had produced OS and photoisomerization occurred at an early disease stage. The hyper-reflective layer defined by in vivo microscopy between the ONL and the RPE suggests remnants of IS and OS material. The rd16 mouse histopathology was consistent with the human data in showing disorganized and truncated rod OS in the second postnatal week. Rapid loss of rods at later ages may be due to the strong dependence of rod cell survival on normal OS development (36).
Abnormal rod photoreceptor OS are detected before rod cell death in both human NPHP5- or NPHP6-LCA and mouse Nphp6-mutant (rd16) retinas. rd16 rod cells have mislocalization and accumulation of opsin, a frequent finding in animal models of retinal degeneration as well as human retinopathies with or without ciliopathy as a suspected cause (2,37–39). The result in rd16, however, suggests a ciliary transport abnormality due to altered Nphp6 function. Recently, knockout mice for the Ahi1 gene, which also encodes a cilium-localized protein, were shown to have no rod OS formation and a rapid retinal degeneration with abnormal distribution of opsin throughout photoreceptors (40). Genetically engineered reduction of the opsin dosage in the Ahi1−/− mice led to the same OS formation defect but delayed retinal degeneration, suggesting that the secondary opsin accumulation contributed to cell death. Hence, disrupted intraflagellar transport and defective localization of OS components, key to the normal function and structure of photoreceptors, may be triggers for apoptotic rod cell death (41) in the NPHP5- and NPHP6-mutant retinas.
Secondary to the relatively rapid rod cell death, we noted abnormal thickening of retinal layers in the human retinas with NPHP5 or NPHP6 mutations. Neuronal damage in the mature central nervous system elicits complex remodeling responses of neurons and glia, as documented in animal retinae with genetic or traumatic disease (42,43) and in post-mortem human donor retinas affected with retinal degeneration (38). Using in vivo microscopy in several human retinal diseases, we found unexpected and dramatic retinal thickening, primarily of the INL, that we attribute to Müller glial cell activation following photoreceptor stress (24,25,27,31). This retinal thickening appears to be a marker of an early disease stage of remodeling. A signaling pathway involving injured photoreceptors and Müller cells has been elucidated (44). Our studies indicate that photoreceptor stress or loss in NPHP5- or NPHP6-mutant retinas leads to Müller cell activation, hypertrophy or possibly proliferation, and this is detectable as retinal thickening by in vivo cross-sectional imaging. Intracellular edema of neuronal or glial cells may also contribute to retinal thickening, as suggested by studies of neurotrophic factors in retinal degenerations (45).
We previously reported that INL thickening in NPHP6-LCA has a correlate in rd16 retina (31). Increased ONL thickening in retinal regions with high rod:cone ratios, however, has not been observed in the human retina until the present study. ONL thickening was found in the canine retina with RPGR mutations treated with ciliary neurotrophic factor and an abnormal increase in rods was documented (45). We speculate that the possible thickening of ONL in the human retinas with NPHP5 or NPHP6 defects results from endogenous release of neurotrophic factors after photoreceptor stress, dysfunction or death. Alternatively or additionally, co-localized thickening of ONL, OPL and INL could be secondary to Muller glial cell activation, hypertrophy or possibly proliferation (see above). Observed hyperautofluorescence, which could signal oxidation of the RPE lipofuscin that has accumulated (29), is also consistent with the existence of outer retinal stress in these regions.
Photoreceptors detected in most of the NPHP5- and NPHP6-mutant human retinas were at and near the cone-rich human fovea. For many human retinal degenerations in the RP spectrum and their animal models, it is not uncommon for relatively rapid rod photoreceptor degeneration to be followed by a much slower cone loss (38). What is unusual in the NPHP5- and NPHP6-associated ciliopathies is the profound dysfunction of the cone photoreceptors, the nuclei of which continue to be structurally detectable for extended periods. We must point out that not all molecular defects at the cone cilium lead to this unusual phenotype (46). However, to pinpoint the abnormality leading to this unusual phenotype will not be trivial considering more than 600 proteins comprise the cilium, and there is evidence of extensive epistatic interactions among many ciliary proteins (41,47–49).
The rd16 mouse provides evidence of detectable cone photoreceptors at least for a period after rods had degenerated. However, the mouse retina with a 30:1 rod:cone ratio could not provide a satisfactory comparison to a primate cone-only fovea (50). Some of the cone cell losses observed in both the rd16 retinas and extra-central human retinas may be due to non-autonomous processes (51,52). This hypothesis suggests that the degeneration of the majority photoreceptor, the rods, has a deleterious effect on the survival of the minority cone population. When the cone-rich double-mutant rd16;Nrl−/− mouse was studied, it was clearer that cone cell loss resulting from the Nphp6-defect was slow even though cone dysfunction was quite profound. Similar approaches taking advantage of an all-cone retina of the Nrl−/− mice have been previously used to decipher the primary and secondary disease effects on cone function and survival (36,53–55).
We propose that the present series of human and animal experiments serve as a model for how research can be moved more expeditiously toward human clinical trials of retinal degenerative diseases. Our strategy was to define the detailed disease expression in molecularly clarified humans with blindness and, in parallel, query greater details of mechanism from relevant animal models. Crossing back and forth between human and animal diseases led to a consensus about the most appropriate target retinal cell and mutant animal to begin proof-of-concept studies. Human retinal degenerative diseases have in the past been thought of as clinically- and now molecularly-diagnosed ‘black boxes’. It was previously assumed that the only choice was to discover an agent that works in genetically-similar animals and eventually hand it over to clinicians to administer to the blind patients. The very different strategy we have taken in patients with NPHP5- or NPHP6-LCA has been made possible by sophistication of non-invasive high resolution imaging in human retinal disease (reviewed in 56).
If a gene replacement treatment strategy for NPHP5- or NPHP6-LCA was being planned, for example, we can now state that to target rods, therapy would have to be administered within the first few years of life. In contrast, targeting the cones of the foveal-macular region could be of value across a wide spectrum of disease stages and ages (spanning at least the first four decades of life). The safety outcome for any future clinical trial should include in vivo imaging using optical coherence tomography (OCT) scans of the retina, in addition to more routine clinical ocular and systemic examinations. The low level of visual acuity would not be a reliable safety parameter for most of the members of these LCA subgroups. An efficacy outcome for a therapy that targeted rods would have to quantify a change in peripheral rod-mediated visual function or the rate of peripheral retinal degeneration—exceedingly difficult measurements to make in very young and severely visually impaired patients. An efficacy outcome for cone-targeted therapy, on the other hand, could include cone function, preferably measured under observation of the fundus, and in the treated region of the central retina in older subjects who would be the candidates for early phase trials. In addition, traditional measures of visual acuity would be available. Increased cone sensitivity or improved visual acuity would signal that cone photoreceptor function is restored; a subfoveal change to more normal IS/OS laminar architecture would be an objective efficacy outcome.
And which animal model would be most valuable for proof-of-concept experiments? For rod-directed therapies, the rd16 mutant mouse with the rapid rod degeneration and the Nphp6-deficient Abyssinian cat with the slow disease course could be used for proof-of-concept experiments. Limitations for the rd16 mice would include the urgency for gene delivery to very young mice and use of certain vector pseudo-types that show faster transgene expression. Limitations for the mutant cats would include extended experimental durations awaiting the efficacy of gene therapy to modify the slow natural progression of retinal degeneration over years. For cone-directed therapies, the ideal model should show a dissociation of cone structure and function and a substantial dynamic range of ERG cone function. The all-cone retina of the rd16;Nrl−/− double-mutant mice, with measurable cone cell numbers but severely dysfunctional cones, would fulfill the criterion for a model worth considering for proof-of-concept studies attempting to develop treatment for most ages of patients with NPHP5- or NPHP6-LCA. The difference of hundreds of microvolts between Nrl−/− S- or M-cones and the rd16:Nrl−/− double-mutant cones would provide a useful dynamic range for testing efficacy of the therapy over the short term.
Patients with LCA and disease-causing mutations in NPHP6 (n = 14) or in NPHP5 (n = 5) were included (Table 1). Informed consent was obtained. Procedures followed the Declaration of Helsinki guidelines and were approved by the institutional review board. Fourier-domain OCT was used (RTVue-100; Optovue Inc., Fremont, CA, USA) with published recording and analysis techniques to perform retinal cross-sectional imaging (24,25,31,57). RPE lipofuscin and melanin imaging was performed as previously described (24,31,33,58). Details are provided (Supplementary Material).
Mice were kept in standard conditions; all procedures conformed to NIH Animal Care and Use Committee approved protocols. ERGs were performed using a computer-based system in mice after dark-adaptation (>6 h) and anesthesia (ketamine, 65 mg/kg, xylazine, 5 mg/kg, i.p.) as described (59–61). Microscopy was performed as previously described (6,7,13,31,61). Details are provided (Supplementary Material).
Supported in part by Grousbeck Foundation, intramural funds from the National Eye Institute, Macula Vision Research Foundation, Foundation Fighting Blindness, and Hope for Vision. A.V.C. is a RPB Senior Scientific Investigator.
We acknowledge Jun Zhang and Kunio Nagashima for expert assistance with electron microscopy.
Conflict of Interest statement. None declared.