Autosomal recessive inherited retinal dystrophies (IRDs) are genetically heterogeneous and often associated with loss of function of cone and rod photoreceptors, potentially leading to blindness (1). Cadherin-related family member 1 (CDHR1) is a causative gene for retinal dystrophy, with an autosomal recessive form of inheritance (2, 3). The cadherin-related family member 1 (CDHR1), also known as protocadherin 21, or photoreceptor-specific cadherin, is a structural transmembrane photoreceptor protein that localizes at the base of cone and rod photoreceptor outer segments and plays an essential role in the maintenance of the structure and survival of photoreceptors (4). It is encoded by CDHR1, a gene located on chromosome 10q23.1 (5).

Cdhr1-deficient mice exhibit progressive outer retinal degeneration and early electroretinographic abnormalities involving both rod and cone pathways.

The Cdhr1-deficient mouse exhibits structural and functional evidence of progressive outer retinal degeneration at a slow rate (6). And, electroretinograms (ERGs) showed early functional deficits affecting both rod and cone photoreceptors. In humans, CDHR1 variants have been associated with a range of phenotypes, including cone–rod dystrophies (CRD) (7, 8), rod–cone dystrophies (RCD), retinitis pigmentosa (RP) (9), and late-onset macular degeneration (LOMD) (10, 11).

In this study, we report a Japanese family harboring a novel CDHR1 variant and exhibiting marked intrafamilial phenotypic variability, highlighting the complexity of genotype–phenotype correlations in CDHR1-associated retinal dystrophies and the importance of comprehensive clinical and genetic evaluation.

This was a retrospective case series of patients presenting to Juntendo University hospital with a diagnosis of IRDs. Three male siblings from the same Japanese family were included. Detailed medical history and comprehensive ophthalmic examinations were performed, including visual acuity, visual field test, fundus photography (CLARUS 500; Carl Zeiss Meditec Inc., Dublin, CA, USA), fundus autofluorescence (CLARUS 500; Carl Zeiss Meditec Inc., Dublin, CA, USA), optical coherence tomography (OCT, Spectralis; Heidelberg Engineering, Inc., Heidelberg, Germany), full-field electroretinography (full-field ERG, LE-4000, Tomey, Nagoya, Japan), were performed in accordance with ISCEV standards (12). Genetic analysis was conducted using next-generation system (NGS) with a gene panel. Pathogenicity of identified variants was interpreted with reference to established databases and ACMG guidelines. All CDHR1 variants were annotated based on the RefSeq transcript NM_033100.4. This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Juntendo University (M08-0468). Written informed consent was obtained from all participants.

We report three brothers with IRD in whom CDHR1 variants were identified, exhibiting marked phenotypic variability within the same family. Although the affected siblings showed overlapping clinical features, their genetic findings were not identical, and their clinical presentations differed substantially. Cases 1 and 2 showed similar findings characterized by severe macular atrophy, decreased visual acuity, and central scotoma, consistent with LOMD. However, the presence of night blindness and electroretinographic evidence of combined rod–cone dysfunction was atypical for isolated macular dystrophy. In contrast, Case 3 presented with fundus features resembling RP, including widespread retinal degeneration, bone-spicule pigmentation, subjective night blindness, and non-recordable full-field ERG responses, suggesting a rod–cone dystrophy phenotype.

Both CDHR1 variants identified in this study are currently classified as VUS according to ACMG/AMP guidelines. The c.2027T>A (p.Ile676Asn) variant was initially assigned PP3 based on in silico prediction; however, upon re-evaluation, the REVEL score (0.565) does not meet the threshold recommended by the ClinGen Sequence Variant Interpretation Working Group for supporting pathogenicity. Therefore, PP3 should not be applied. This variant shows relatively higher allele frequencies in East Asian populations, including the Japanese cohort, with reported homozygous individuals in public databases, supporting the application of BS1. Nevertheless, in Cases 2 and 3, c.2027T>A was detected together with a rare CDHR1 variant, c.748C>A (p.Pro250Thr), which is absent from population databases. Although parental segregation analysis was not available, the recurrent identification of these two heterozygous variants in affected siblings suggests a possible trans configuration. In this context, and consistent with the Japanese guideline of IRDs (13), PM3 supporting may be cautiously considered. Importantly, the available evidence remains insufficient to reclassify either variant as pathogenic or likely pathogenic, and both variants should be interpreted as VUS. Their contribution to disease pathogenesis remains hypothetical.

Case 1 carries only a single heterozygous CDHR1 c.748C>A (p.Pro250Thr) variant.

This monoallelic VUS is insufficient to establish CDHR1-associated autosomal recessive disease. Alternative explanations include an undetected second pathogenic allele (e.g., deep intronic variant, regulatory variant, or copy-number variant) or other genetic contributors not captured by our panel-based NGS. Therefore, Case 1 is considered genetically inconclusive despite its clinical resemblance to Cases 2 and 3.

In addition, beyond the REVEL score shown in Table 1, no additional in silico prediction tools were applied in this study. Given the absence of segregation analysis and comprehensive in silico evidence, the pathogenic relevance of the identified CDHR1 variants cannot be definitively established and should be interpreted with caution.

Variants in CDHR1 have been previously associated with a spectrum of autosomal recessive retinal dystrophies, including CRD (7, 8); RCD and RP (9, 14), and more recently, LOMD (10, 11). Intrafamilial phenotypic heterogeneity among individuals carrying identical CDHR1 variants has also been documented (7, 15). However, the mechanisms underlying these phenotypic differences remain poorly understood.

All three cases in this study shared several features: severe macular atrophy, late-onset symptoms, and self-reported night blindness, with ERG confirming dysfunction of both photoreceptor systems. These findings suggest that CDHR1-associated retinal dystrophy may present along a phenotypic continuum ranging from macular-predominant cone–rod dystrophy to generalized rod–cone dystrophy. Even in patients initially diagnosed with macular dystrophy, careful longitudinal follow-up is warranted to monitor for potential peripheral retinal involvement as the disease progresses.

An additional CDHR1 variant, c.2027T>A (p.Ile676Asn), was identified in Cases 2 and 3; however, this variant is currently classified as VUS, and its pathogenic role remains unclear. Other variants identified in this family, including those in PROM1 and RP1, may be relevant to retinal disease in general, but there is no direct evidence that they contribute to the observed phenotypic variability in the present cases.

Although non-Mendelian mechanisms such as genetic modifiers, oligogenic inheritance, or allelic expression imbalance have been proposed in IRD (16), these mechanisms remain speculative in the present family and are not supported by direct genetic or functional evidence. Accordingly, the phenotypic variability observed in this family should be interpreted with caution, and the possibility of additional undetected genetic or environmental factors cannot be excluded.

Although segregation analysis and functional validation were not performed, these findings expand the understanding of CDHR1 genotype–phenotype correlations. We propose that CDHR1 variants can manifest across a broad clinical spectrum, and that modifier variants or transcript-level regulation may underlie intrafamilial phenotypic diversity. Clinicians should consider CDHR1 variants in patients with macular dystrophy accompanied by night blindness or ERG evidence of rod–cone dysfunction.

This study has several limitations, including the absence of segregation analysis in the parents and the lack of in vitro or in vivo functional studies. Nevertheless, our findings contribute valuable insights into the expanding genotype–phenotype correlations in CDHR1-related retinal dystrophy.

Furthermore, even in cases with fundus findings suggestive of macular dystrophy, it is crucial to consider CDHR1 gene variants in clinical management when patients report night blindness or show evidence of rod-cone dysfunction on ERG. Taken together, our findings suggest that CDHR1-associated retinal dystrophy may manifest along a phenotypic continuum, ranging from macular-predominant disease to generalized retinal degeneration. However, the interpretation of genotype–phenotype correlations must be approached cautiously, particularly in the absence of segregation analysis and functional validation.

This study has several limitations, including the lack of segregation analysis, comprehensive structural variant detection, and in vitro or in vivo functional studies. Further investigations incorporating these approaches will be necessary to clarify the pathogenic relevance of the identified variants and to better understand the mechanisms underlying intrafamilial phenotypic variability in CDHR1-associated retinal dystrophy.

In conclusion, our findings illustrate marked intrafamilial phenotypic variability in a Japanese family with IRD in which CDHR1 variants were identified. These observations highlight the importance of integrating molecular, electrophysiological, and imaging analyses, as well as careful interpretation of genetic findings, for accurate diagnosis, prognostic assessment, and genetic counseling of patients with IRDs.