Hypomorphic CDHR1 variants may result in retinitis pigmentosa with relative preservation of cone function

ABSTRACT Purpose Retinitis pigmentosa (RP) associated with biallelic variants in CDHR1 has rarely been reported, and detailed phenotyping data are not available. RP implies relative preservation of foveal cones, when compared to cone-rod dystrophy associated with biallelic null variants in CDHR1. We hypothesize that RP may occur in association with one or more hypomorphic CDHR1 alleles. Materials and methods Retrospective report of a 48-year-old patient with CDHR1-associated RP with a hypomorphic missense variant c.562 G>A, p. (Gly188Ser) and a novel, unreported variant affecting a canonical splice acceptor site (c.784–1 G>C). Clinical examination, multimodal retinal imaging, electroretinography, visual field testing, and mesopic microperimetry were undertaken 8 years apart. Scotopic microperimetry was also performed. The DNA sequence context of the variants was examined to identify theoretical CRISPR-Cas9 base-editing strategies. Results The patient presented at 35 years with a 12-year history of nyctalopia. His best corrected visual acuity was 20/20. Clinical presentation, multimodal retinal imaging studies, electroretinography, and mesopic microperimetry were typical of a progressive rod-cone dystrophy (i.e. classic RP). There were no scotomas within the central field as would be expected at this age in CDHR1-associated cone-rod dystrophy. Scotopic microperimetry suggested some preservation of macular cone over rod function, although both were severely impaired. A suitable CRISPR adenine base editor was identified that could theoretically correct the missense variant c.562 G>A, p. (Gly188Ser). Conclusions CDHR1-associated RP shows a relative preservation of cone function in the presence of a presumed hypomorphic allele and may be considered a hypomorphic disease phenotype. Further work is required to identify modifying factors that determine disease phenotype since macular dystrophy, with relative sparing of rods, may also occur with hypomorphic CDHR1 alleles


Introduction
Cadherin Related Family Member 1 (CDHR1) gene (previously known as PCDH21) was identified as a cause of autosomal recessive retinal degeneration in 2010 (1,2).CDHR1 is expressed at the base of the photoreceptor outer segments in both rods and cones at the nascent edge of developing outer segment discs where it forms connections to the periciliary ridge of the inner segment (3).The Cdhr1 -/-knockout mouse demonstrated evidence of early cone and rod dysfunction, associated with shortened and disorganised outer segments with progressive photoreceptor cell death (4,5).Accordingly, patients with biallelic null variants appear to develop cone-rod dystrophy with a phenotype that broadly aligns to that observed in the knockout mouse model (6).
Hypomorphic alleles in CDHR1 have most commonly been associated with late-onset macular dystrophy, which may be misclassified clinically as age-related macular degeneration (AMD) (7,8).In this group, full-field electroretinography is often normal or borderline, suggesting a relative preservation of generalised cone and rod function when compared to conerod dystrophy patients (6).Retinitis pigmentosa (RP) has rarely been reported in association with biallelic variants in CDHR1 (1,2,(9)(10)(11).Although this group has not been phenotyped in detail.Further characterization of CDHR1-associated RP may help to highlight possible mechanisms by which hypomorphic CDHR1 variants differentially affect rod and cone photoreceptors.
Herein we describe a patient with classic RP who harbours biallelic variants in CDHR1: a known hypomorphic missense variant and a novel, unreported intronic variant that is predicted to severely affect a canonical splice acceptor site.Detailed phenotyping was undertaken over 8 years, including scotopic microperimetry, which permits a functional assessment of macular cone and rod function.We highlight the presence of hypomorphic alleles in other published CDHR1-associated RP cases, discuss the variability in hypomorphic CDHR1 phenotypes, possible underlying mechanisms, and their relevance to gene therapy clinical trials expected in the near future.

Materials and methods
A 35-year-old patient was referred to the Oxford Eye Hospital, a tertiary referral centre for inherited retinal degenerations in 2009.Data were collected retrospectively as part of the patient's routine clinical care.At each clinical visit, the patient underwent evaluation of best-corrected visual acuity, slit-lamp biomicroscopy, and fundoscopic examination.Goldmann kinetic visual field testing and Humphrey central, static visual field testing were also performed.Multimodal retinal imaging was undertaken including digital fundus photography (Optos, Dunfermline, Scotland), short-wavelength fundus autofluorescence (FAF), and optical coherence tomography (OCT) imaging (Heidelberg Spectralis, Heidelberg Engineering GmbH, Heidelberg, Germany).Mesopic and scotopic microperimetry were undertaken using the macular integrity assessment microperimeter (MAIA, CentreVue, Padova, Italy), mesopic microperimetry was performed using a test protocol as previously described (12).These visual function tests were performed as part of the visual function in retinal degeneration study (ethics reference 20/WM/0283).Mesopic microperimetry has been extensively used and is an established clinical outcome measure in clinical trials.Scotopic microperimetry is a novel examination technique conducted under scotopic luminance conditions with short wavelength (cyan light at 505 nm) and long-wavelength stimuli (red light at 627 nm) (13).This technique enables the mapping of central retinal sensitivity, distinguishing between rod and cone photoreceptor sensitivities in the macula (14).Full-field electroretinography (ERG) was also undertaken (Diagnosys LLC, Lowell, USA).
The patient underwent molecular genetic testing with targeted next-generation sequencing of a panel of 109 genes associated with RP and RP-like phenotypes, as previously described (15).A customised HaloPlex Target Enrichment system (Agilent Technologies, Santa Clara, United States) was utilised.Sequencing was done using Illumina MiSeq technology, which involves data generated by The Wellcome Trust Centre for Human Genetics at the University of Oxford.Data were analysed, and the variants identified were confirmed by Sanger Sequencing, which has an estimated sensitivity >99%.The findings of the genomic laboratory report of this patient's parents were also retrieved.

Results
A 35-year-old male was referred to a tertiary referral centre for retinal genetics in 2009 with nyctalopia and difficulties with dark adaptation, noted from the age of 23 years.There was no family history of retinal degeneration (Figure 1).His parents were not known to be consanguineous and were of European ancestry.He did not have any other medical conditions or take regular medications.
In the most recent clinical assessment, at 48 years of age, the best corrected visual acuity was 20/20 in both eyes, with low myopia (defined as ≤ −0.5 and > −3.00 D).Intraocular pressures were 19 mmHg in right eye and 17 mmHg in left.He had constricted visual fields to confrontation.Anterior segments were normal with no significant cataract.Pseudocolor fundus imaging identified bilateral peripheral pigment migration (Figure 2A,B).Short-wavelength autofluorescence imaging showed hypoFAF, peripheral mottling with centripetal progression of degeneration over time (Figure 2C,D), but no ring of increased autofluorescence that is sometimes seen in patients with RP.The macula also revealed mottled hypoFAF, indicating relatively early-onset macular degeneration.OCT imaging of the posterior pole showed outer nuclear layer thinning and disruption of the ellipsoid band in the parafoveal area, extending peripherally.The ellipsoid band was present at the fovea although the width of the band reduced over an 8-year period (Figure 2E,F).
These structural findings correlated with longitudinal and functional assessments on mesopic microperimetry which identified progressive constriction of the central visual field over the interval (Figure 3A,C,E).Mesopic microperimetry, which is considered a cone driven test (17), showed overall impaired sensitivity compared with controls (Figure 3B,D,F), with a reduction in retinal sensitivity over 8-years within the central island.Scotopic microperimetry, which examines the sensitivity of dark-adapted macular cones and rods revealed reduced cyan sensitivity, implying severely impaired rod function (Figure 3C,D).Scotopic red, which is theoretically predominantly cone driven, was also severely reduced compared with normal, implying a reduced cone function (Figure 3G,H) (18).Comparatively, scotopic red was marginally better than cyan, in keeping with a rod-cone degeneration (Figure 3C-H).Goldmann kinetic perimetry documented progressive, bilateral peripheral visual field constriction  over an 8-year period (Figure S1) Responses on full-field electroretinography, performed in 2009, were all largely extinguished (Figures S2-4).
CRISPR DNA analysis showed that correction of the c.562 G>A p(Gly188Ser) variant could be achieved with an adenine base editor (ABE) with the introduction of an irreversible A>G edit in genomic DNA (16).For this to occur, a modified CRISPR-associated (Cas-9) endonuclease creates a singlestranded nick in a DNA sequence at a location specified by a guide RNA, that exists a specific number of nucleotides downstream of the protospacer adjacent motif (PAM) site.The PAM site is a recognition sequence to which the Cas9 binds that is specific to the species of bacterium from which it originates.A Cas9 nickase may produce edits within a range of nucleotides downstream of the PAM site in a region called the "editing window."The sequence context of this mutation was conducive to ABE since a suitable PAM site was identified, placing the target nucleotide within the editing window (Table 2).The intronic variant (c.784-1 G>C) would require a glycosylase base editor (GBE) to edit C>G.However, with currently available technologies, no suitable PAM sites or guide RNA sequences were identified to target this variant.KKH-SaCas9 PAMs were found close to mutation (Table 2).

Discussion
The most common clinical phenotype of CDHR1-associated retinopathy reported in the literature is cone-rod dystrophy (CRD), characterised by severe, early functional deficits in cone and rod function with progressive photoreceptor cell death (6).These observations are consistent with longitudinal functional and structural observations in the Cdhr1 -/-mouse model, both of which are likely to represent the null phenotype (21).There have been rare reports in the literature of RP occurring in association with biallelic variants in CDHR1, although these cases have not been characterised in detail (Table 3).Furthermore, hypomorphic CDHR1 alleles have been associated with late-onset macular dystrophy (MD) (7,8,21).
Here, we present deep phenotyping of a patient with RP who harbours biallelic variants in CDHR1.Rod-cone dystrophy is defined based on recordable ERG responses which our patient and those described in Table 3 do not have at first assessment.Despite the initial symptoms being described as night vision impairment, there may have been a similar early reduction of cone and rod-related full field ERG responses, making this a mixed cone and rod degeneration as opposed to typical RP.However, we have classified this patient to have RP based on his microperimetry results and fundus imaging, not ERG.He has a relatively preserved structure and function of the foveal and parafoveal area compared to a typical patient of this age with CDHR1-associated cone-rod dystrophy.RPE atrophy is present in the mid-peripheral retina, peripheral to the major vascular arcades and more posteriorly (Figure 2B,  D).This would not be expected at this age in the context of CDHR1-associated CRD or MD with associated focal areas of loss of retinal sensitivity.
Scotopic microperimetry testing offers novel data into the relative function of macular rods and cones in CDHR1associated retinal degeneration.Mesopic microperimetry confirms progressive constriction of the field of macular sensitivity associated with a reduction in absolute thresholds over time.Scotopic microperimetry confirms the relative preservation of macular cone function over rod that is more characteristic of RP.In healthy eyes, scotopic microperimetry cyan sensitivity is believed to elicit a rod dominant photoreceptor response and scotopic red driven by a mixture of cone and rod, but cone dominated.However, caution should be exercised regarding this assumption in patients with retinal degeneration, since at very low levels of scotopic cyan sensitivity the stimulus may be sufficiently bright enough to elicit a cone dominant response, especially in the presence of severe rod dysfunction.The severely reduced cyan sensitivity in this case suggests severely impaired rod function.However, it remains unclear whether the residual scotopic cyan sensitivity detected rod or cone driven.One hypothesis, consistent with all reported CDHR1associated retinal degeneration cases where clinical data suggest RP (Table 3), is that one or more hypomorphic CDHR1 alleles may confer relative preservation of cone function that manifests clinically as RP over CRD.Ultrastructural analysis has shown that CDHR1 forms physical connections between the leading edge of nascent outer segment discs and the inner segment (3). Biochemical analysis has identified a cleavage event that presumably releases the ectodomain following which outer segments discs grow towards the RPE (22).However, these and other functions of CDHR1 have not been studied specifically in rods and cones.It is likely that specific hypomorphic CDHR1 variants (i.e.c.783 G>A) may contribute to more than one phenotype (i.e.MD or RP), which may depend on the second CDHR1 allele, and/or other modifying factors.
The mechanisms underlying these variations in CDHR1 phenotypes are unclear since MD patients exhibit relative preservation of rod function, and RP patients show relative preservation of cones.Other genes expressed at the base of the photoreceptor outer segment similarly have variable phenotypes, and indeed different modes of inheritance, depending on the variant.Specific missense variants in PRPH2 underlie RP or a macular phenotype (central areolar choroidal dystrophy, CACD), whereas mono-allelic loss-of-function variants may cause a Stargardt disease-like phenotype.Moreover, biallelic loss-of-function variants in PROM1 result in pan-retinal degeneration, with dominant missense variants producing macula predominant disease (23).
The CDHR1 variants reported in this patient have a low population frequency in genetic databases and are predicted to be pathogenic in silico (Table 1).Moreover, this report confirms the likely pathogenicity of p. (Gly188Ser) which was reported in association with MD in three patients who were heterozygous for the variant (8,24).This variant is highly conserved across species and lies within one of the cadherin tandem repeat domains in the extracellular region of the CDHR1 protein (Figure 1B) (24,25).Screening of this variant using ESEfinder (exonic splice enhancer) did not identify any aberrant splicing signals (26).The novel c.784-1 G>C variant disrupts one of the two invariant nucleotides of the splice acceptor site and is predicted to lead to aberrant splicing, with frameshift and nonsense mediated decay.The pathogenicity of this variant could be confirmed using an in silico splicing assay.
DNA base editing could, in theory, be used to correct the c.562 G>A CDHR1 variant: in silico analysis showed that a SpCas9-utilising adenine base editor (ABE) could correct c.562 G>A (Table 2).There are no predicted bystander edits within the editing window, increasing the safety profile.The canonical splice site variant c.784-1 G>C could be corrected with a glycosylase base editor (GBE), however there were no available PAM sites for the currently available SpCas9-utilising construct.An available KKH-SaCas9 PMA site was found adjacent to the variant (Table 2) and hence could possibly be used, further in vitro and in vivo studies are needed to test and validate such construct.
We conclude that patients with one or more hypomorphic alleles may present with RP, which has not yet been described in any patient with biallelic null variants in CDHR1.These observations, supported by scotopic microperimetry, and consistent with variants associated with other published cases of CDHR1-associated RP, support the clinical management of this patient group and selection for clinical trials expected in the near future.

Figure 1 .
Figure 1.(a) Pedigree diagram, (b) CDHR1 protein structure.The extracellular portion includes six cadherin-like repeats.The location of the two mutations identified in the proband are indicated by red arrows.The numbers indicate amino acid position.

Figure 2 .
Figure 2. Retinal imaging of the right eye.(A&B): Widefield optos imaging of the right fundus in 2014.(A) and at 8 year follow up (B) where there is increased bone spiculation peripherally, typical of retinitis pigmentosa (C&D): Blue fundus autofluorescence imaging of the right fundus in 2014 (C) and at a 8 year follow up (D) where there is peripheral mottling, small amounts of reduced vasculature (E&F): Optical coherence imaging (30 degrees) of the right eye taken at presentation in 2013 (E) and at 8 year follow up (F) where there is increased thinning of the outer layer, reduction in the subfoveal ellipsoid zone line, blurring between ellipsoid zone to retinal pigment epithelium in the parafoveal area.

Figure 3 .
Figure 3. Microperimetry of the left eye.(A): Mesopic microperimetry in 2014 showing there is detectable sensitivity (B), but this is reduced compared to normal limits implying impaired cone response.(C) Mesopic microperimetry at 8 year follow up showing a number of points with no detectable sensitivity forming a ring scotoma and marked constriction of central retinal sensitivity (D).(E): Scotopic microperimetry in 2022 -cyan sensitivity with histogram, showing reduced sensitivity, implying severely impaired rod function (F), red sensitivity (G) with histogram (H) which appear markedly reduced and almost equally impaired.In healthy eyes, scotopic cyan sensitivity is believed to elicit a rod photoreceptor response and scotopic red driven by a mixture of cone and rod, but cone dominated.

Table 1 .
Summary of the characteristics of the CDHR1 missense variants.

Table 2 .
Suitable base editor and PAM site sequences for the known CDHR1 variants of the patient in this report.

Table 3 .
Summary of retinitis pigmentosa occurring in association with biallelic variants in CDHR1.