PCARE-Associated Retinopathy – Genetics, Clinical Characteristics, and Natural History

Abstract

Purpose: The purpose of this study was to describe the mutational landscape, clinical characteristics, and natural history of PCARE-associated retinopathy.

Methods: Retrospective cohort study including 28 patients (56 eyes) affected by an inherited retinal disease related to PCARE variants. The main outcome measures were best-corrected visual acuity (BCVA) and degree of vision impairment, kinetic visual field (KVF) area delimited with the V4e target, area of macular atrophy (MA) with definitely decreased autofluorescence (DDAF) on short-wavelength autofluorescence, total macular volume (TMV) and foveal sparing (FS) on optical coherence tomography.

Results: The median age at first examination was 40.7 years (Interquartile range [IQR] = 28.8–49.6), whereas the median follow-up time was 5.7 years (IQR = 3.6–7.1). The retinal phenotype was consistent with a severe generalized photoreceptor dystrophy with MA in all patients. DDAF lesions were observed in 85% of the eyes. Loss of FS (occurring at a median age of 45 years) was associated with a mean BCVA (logMAR) worsening by 1.1 (95% confidence interval [CI] = 0.6 to 1.5, P < 0.001). Low vision and blindness in the better-seeing eye occurred at median ages of 50 and 57 years, respectively. Longitudinal analysis revealed the following mean slopes of change: BCVA (logMAR) worsened by 0.06/year (95% CI = 0.03 to 0.09, P < 0.001), KVF area decreased by −23%/year (95% CI = −35% to −12%, P = 0.004), square root-transformed DDAF area expanded by 0.20 mm/year (95% CI = 0.16 to 0.23, P < 0.001), and TMV declined by −0.015 mm³/year (95% CI = −0.023 to −0.007, P = 0.003). Eleven novel PCARE variants were identified.

Conclusions: PCARE-associated retinopathy is a severe generalized photoreceptor dystrophy with MA. Although visual field loss occurs early, useful central vision is often retained into late adulthood because of FS. Based on the age of onset of legal blindness, the optimal therapeutic window appears to be before the fifth decade of life.

Inherited retinal diseases (IRDs) are a clinically heterogeneous group of retinopathies with genetic etiology, which cause visual impairment due to irreversible photoreceptor and/or retinal pigmented epithelium (RPE) degeneration and affect approximately 1 in 3500 individuals, constituting a major cause of visual disability among working age individuals.^1,2 More than 250 genes have been associated with IRDs, reflecting high genetic heterogeneity.³ The photoreceptor cilium actin regulator (PCARE) gene (formerly known as C2ORF71; MIM #613425) was first identified as a cause of non-syndromic autosomal recessive retinitis pigmentosa (RP) type 54 (MIM #613428) in 2010.^4,5 Comprising 2 exons, this gene encodes a 1289-amino acid protein predominantly expressed in the primary cilium of retinal photoreceptors, where it regulates outer segment disk formation and renewal.^6–8 PCARE likely contributes to a substantial proportion of autosomal recessive IRDs, with estimates ranging from 1% in a large French cohort of genetically unresolved cases to 16% in a Swiss cohort.^9,10 Although the phenotype was originally labeled as typical RP,^4,5 a later study by Audo et al. using full-field electroretinogram (ERG) pointed toward a severe photoreceptor dystrophy with early cone involvement and macular atrophy (MA).⁹ Currently, no treatment exists for PCARE-associated retinopathy, yet it appears to be an ideal candidate for gene therapy development. Indeed, most cases result from biallelic predicted null variants, consistent with the phenotype observed in the Pcare^−/− mouse,⁶ implying potential benefit from a gene augmentation approach. Additionally, the short coding sequence (3.8 kb) fits within the maximum cargo capacity of adeno-associated viruses used in previous gene therapies for IRDs (approximately 4.7 kb).¹¹ Moreover, a naturally occurring canine model is already available,¹² which could be pivotal in testing novel promising treatments, as was the case for RPE65 gene therapy.¹³ While preclinical research moves forward, in-depth knowledge of the clinical phenotype and natural history of PCARE-associated retinopathy needed as a basis for forthcoming clinical trials is currently limited.¹⁴ Herein, we aim to inform the audience about the clinical features of the disease and to describe the natural history of its visual and anatomic outcomes. 

Methods

Study Design and Patient Selection

This was an observational cohort study including 28 patients with a molecular diagnosis of PCARE-associated retinopathy. The patients were identified from institutional databases held at 2 national referral centers for IRDs: the rare disease center REFERET of 15 to 20 Hospital, Paris, France (25, 89%) and the Department of Ophthalmology of IRCCS San Raffaele Scientific Institute, Milan, Italy (3, 11%). Findings from four patients (CIC00261, CIC02622, CIC00643, and CIC01571) were previously published.^9,15,16 Genetic diagnosis was achieved by applying a Next-Generation Sequencing approach followed by confirmation and segregation of identified variants by direct Sanger sequencing.^15,17–19 Variants were interpreted according to the American College of Medical Genetics and Genomics guidelines.²⁰ Clinical and imaging data were prospectively acquired as part of standard clinical practice from 2004 to 2024 and analyzed in May 2024. All studies were carried out in accordance with the Declaration of Helsinki and were approved by ethics committees (CPP Ile de France V, Project number 06693, N° EUDRACT 2006-A00347-44, December 11, 2006). Signed informed consent for the genetic testing and permission to use medical data for research purposes was obtained from participants; no financial compensation or incentive was offered to patients. 

Data Collection, Retinal Imaging, and Functional Tests

Clinical notes of each follow-up visit were reviewed to retrieve data on family and medical history, subjective refraction (converted in its spherical equivalent in diopters), best-corrected visual acuity (BCVA) measured on Early Treatment Diabetic Retinopathy Study (ETDRS) charts, slit lamp biomicroscopy (in particular, phakic status), functional tests, and retinal imaging. The ERG recordings adhering to the standards set from the International Society for Clinical Electrophysiology of Vision (ISCEV)²¹ were performed using a Diagnosys Espion system with ColorDome LED full-field stimulator (Diagnosys LLC, Lowell, MA, USA) and DTL electrodes. Kinetic visual fields (KVFs) were performed using either a manual Goldmann perimeter or an Octopus 900 semi-automated perimeter (Haag-Streit AG, Köniz, Switzerland). Short-wavelength fundus autofluorescence (SW-AF) and spectral domain optical coherence tomography (SD-OCT) scans were acquired using a SPECTRALIS HRA + OCT (Heidelberg Engineering, Heidelberg, Germany). 

Outcome Variables

BCVA was recorded in logMAR, and that of eyes with very low vision, expressed semi-quantitatively as counting fingers, hand motion, light perception, or no light perception, converted to 1.98, 2.28, 2.7, or 3 logMAR, respectively.²² Patients were categorized by their degree of vision impairment based on BCVA on the logMAR scale in the better-seeing eye, following the criteria established in the 11th edition of the International Classification of Diseases and Related Health Problems (ICD).²³ Specifically, mild visual impairment was defined as BCVA > 0.3 and ≤ 0.5, moderate impairment as BCVA > 0.5 and ≤ 1.0, severe as BCVA > 1.0 and ≤ 1.3, and blindness as BCVA > 1.3. The KVF area delimited with the V4e target was measured in degrees² either automatically, for tests performed using an Octopus 900 perimeter, or after digitalization and subsequent analysis on ImageJ software (http://imagej.nih.gov/ij/; National Institutes of Health, Bethesda, MD, USA) for tests performed with a conventional Goldmann manual perimeter. A previous study demonstrated minimal differences in KVF area measurements between these two testing methods when considering the V4e isopter.²⁴ The area of the blind spot and any scotomas detected with the V4e target were also measured and subtracted from the KVF area. SW-AF images were used to identify the presence of definitely decreased autofluorescence (DDAF) lesions corresponding to macular RPE atrophy. DDAF was defined as well-demarcated hypoautofluorescence with “darkness” levels at least 90% relative to the optic nerve head.²⁵ The semiautomated Region Finder tool available on the built-in manufacturer software (HEYEX 1.9.14.0; Heidelberg Engineering) was used to measure the DDAF area on 30 degrees or 55 degrees SW-AF images in mm². For each eye, either 30 degrees or 55 degrees SW-AF images were consistently used across all visits, as area measurements between the two fields of view are non-interconvertible.²⁶ The choice was determined retrospectively based on whether the final MA lesion extended beyond the 30 degrees image. Fovea-centered SD-OCT raster scans (at least 19 ART > 10 B-scans covering the central 20 degrees of the macula) were used to annotate the presence of foveal sparing (FS), defined by discernible ellipsoid zone and RPE bands at the central point of the fovea,²⁷ and to measure the total macular volume (TMV) in the central 3 mm of the ETDRS grid, as previously described.^19,28 Inner limiting membrane and Bruch membrane were automatically segmented in all B-scans by the built-in manufacturer software (Heidelberg Eye Explorer 1.9.14.0; Heidelberg Engineering). Manual adjustments were applied when the automatic segmentation was inaccurate. Eyes with low-quality or incomplete scans were excluded from the analysis. All measurements and adjustments were performed by a single grader (author L.B.). 

Statistical Analysis

The primary objective of the study was to estimate the slope of change in BCVA, the KVF area, the DDAF area, and TMV over the follow-up. Given the nested data structure, which includes both eyes from the same patient and multiple longitudinal measurements from the same eye, linear mixed models with random effects for both the intercept and slope were used. For the KVF area analysis, only eyes with a baseline V4e isopter diameter of 10 degrees or larger were included, and the area was transformed to its natural logarithm to model an exponential decay, as previously performed.^29,30 For the DDAF area analysis, the square root-transformed value in mm was also used to account for a possible effect of baseline area on its growth rate.^31,32 Similar models were used to test associations between structural and functional outcome measures. Additional secondary objectives included estimating the incidence rates of moderate or severe vision impairment (“low vision”), blindness, and foveal atrophy. The Kaplan-Meier method was used to analyze time-to-event outcomes. The median age at first occurrence of each outcome was computed, including the entire cohort, and incidence rates were calculated for patients who had not yet developed the outcome at their first visit but had available longitudinal data. The level of significance was set at α < 0.05. Results of statistical tests are reported with 95% confidence intervals (95% CIs). All analyses were conducted using RStudio version 2023.03.0+386 (R Foundation for Statistical Computing, Vienna, Austria) using the lme4, lmerTest, survival, survminer, and ggplot2 packages. 

Results

A total of 28 patients (17 women, 61%) from 25 different families were identified (Supplementary Table S1). The median age of the cohort at first examination (baseline) was 40.7 years (IQR = 28.8–49.6 years), ranging from 16 to 65 years. Most patients (21/28, 75%) had longitudinal data available, with a median follow-up duration of 5.7 years (IQR = 3.6–7.1 years). 

Clinical Phenotype

Based on retinal electrophysiology and imaging, all included patients exhibited a consistent retinal phenotype characterized by severe generalized photoreceptor dystrophy with MA. Specifically, retinal function was assessed by ERG in 24 patients (86%). The majority (22/24, 92%) displayed no distinguishable waveforms above background noise on both the rod-mediated dim-flash dark-adapted (DA 0.01) ERG and cone-mediated light-adapted (LA) ERGs. Only 2 patients aged 28 to 31 years at the time of examination, had LA ERGs with detectable waveforms, albeit with severely subnormal amplitudes and delayed peak times. On SW-AF, MA was characterized by DDAF corresponding to well-demarcated areas of RPE atrophy in 86% of the eyes (48/56) and by faintly decreased autofluorescence corresponding to hypopigmented macular mottling in 14% of eyes (8/56) (Fig. 1). FS was relatively common at baseline (22/56, 39%), and the incidence rate of foveal atrophy was 6.4 per 100 eye-years (95% CI = 2.4 to 14.0; Fig. 2). Overall, the median age at first detection of foveal atrophy was 45.3 years (Fig. 3). 

Figure 1.

 

Retinal phenotype of PCARE-associated retinopathy on color fundus photography, short-wavelength autofluorescence (SW-AF), and spectral-domain optical coherence tomography (SD-OCT). Five patients of varying ages are shown. The youngest patient (CIC15014) exhibits mottled atrophic changes at the macula and complete loss of outer retinal photoreceptor bands but no lesion with definitely decreased autofluorescence (DDAF) lesion on SW-AF. Three other patients (CIC10790, CIC12632, and CIC10015) show retinal pigment epithelium atrophy at the macula, well-demarcated DDAF lesions on SW-AF, and outer retinal atrophy with a variable degree of foveal sparing on SD-OCT (absent in CIC12632, complete in CIC10790, and partial in CIC10015). The oldest patient (CIC07776) presents with extensive atrophy of the posterior pole accompanied by pigmentary clumps in the macula. On SW-AF, coalescing DDAF lesions can be seen in the macula, along with additional atrophic areas nasally to the optic nerve head. On SD-OCT, complete atrophy of the outer retina and RPE is observed throughout the macula, along with irregularities of the Bruch's membrane and a large hyper-reflective lesion with signal shadowing at the foveal center, corresponding to the pigmented clump observed on color fundus photography.

Figure 2.

 

Natural history in a case of PCARE-associated retinopathy, with loss of foveal sparing and development of macular atrophy with definitely decreased autofluorescence (DDAF). At baseline, a 22-year-old patient with the homozygous PCARE c.1541del [p.(Pro514Hisfs*27)] variant had macular mottling on color fundus photography, faintly hypoautofluorescent changes on short-wavelength autofluorescence (SW-AF), and preservation of the outer retinal and retinal pigment epithelium bands at the foveal on spectral-domain optical coherence tomography (i.e. foveal sparing). After 12 years of follow-up, a well demarcated area of retinal pigment epithelium atrophy corresponding to a definitely decreased autofluorescence lesion on SW-AF is visible at the center of the macula, whereas spectral-domain optical coherence tomography demonstrates the loss of foveal sparing with disruption or absence of the ellipsoid zone and retinal pigment epithelium bands.

Figure 3.

 

Time-to-event curves for development of foveal atrophy, low vision, and blindness in PCARE-associated retinopathy. The median age for the loss of foveal sparing is 45 years, while the median ages for the development of low vision (best-corrected visual acuity in the better-seeing eye worse than 0.5 [logMAR] or 20/70 [Snellen]) and blindness (best-corrected visual acuity in the better-seeing eye worse than 1.3 [logMAR] or 20/400 [Snellen]) are 50 and 57 years, respectively.

Visual Impairment, Acuity, and Field

BCVA and refractive error data were collected for 52 of 56 eyes (93%) and 26 of 28 patients (93%) (Fig. 4A). Longitudinal data over a median follow-up time of 5.6 years (IQR = 3.5–6.9 years) were available for 73% (19/26) of the patients (Fig. 4B). Myopia was the most prevalent refractive error, with an estimated mean spherical equivalent of −4.1 diopters (95% CI = −5.4 to −2.8, P < 0.001) in phakic eyes (39/52 at last follow-up visit, 75%). At baseline, the median BCVA (logMAR) was 0.6 (IQR = 0.3–1.5), with 37% of eyes (19/52) having a BCVA (logMAR) worse than 1.00. Twenty-three percent of the patients (6/26) met the criteria for low vision in the better-seeing eye, whereas another 23% (6/26) met the criteria for blindness. Among patients with mild or no vision impairment at baseline (14/26, 54%), the incidence rate of low vision and blindness was 3.9 per 100 person-years (95% CI = 0.8 to 11.5) and 2.1 per 100 person-years (95% CI = 0.3 to 7.7), respectively. Overall, the median ages at first diagnosis of low vision and blindness were 49.6 and 57.1 years, respectively (see Fig. 3). The mean BCVA (logMAR) in the presence of FS was 0.23 (95% CI = 0.03 to 0.43, P = 0.04), whereas the development of foveal atrophy was associated with a mean BCVA (logMAR) worsening by 1.1 (95% CI = 0.6 to 1.5, P < 0.001). Longitudinally, the estimated mean slope of BCVA (logMAR) decline was 0.06/year (95% CI = 0.03 to 0.09, P < 0.001), corresponding to a mean loss of 3 ETDRS letters per year of follow-up (see the Table). 

Figure 4.

 

Longitudinal change of outcome variables in PCARE-associated retinopathy. Scatterplots illustrating changes in best-corrected visual acuity (BCVA), kinetic visual field V4e isopter area, definitely decreased autofluorescence (DDAF) area, and total macular volume (TMV) over age or follow-up time. In each graph, individual measurements from all patients, eyes, and visits are represented by dots. Solid black lines connecting dots represent serial measurements from the same eye over time. In the graphs illustrating changes over follow-up time (right panels), the colored solid line represents the best-fit linear regression line (slope) for the entire cohort (shaded gray areas mark the standard error). The natural logarithm was used to model the longitudinal change in the KVF area, whereas the square root transformation (sqrt) was used to model the longitudinal change in DDAF area. CF = counting finger (1.98 logMAR); HM = hand motion (2.28 logMAR); LP = light perception (2.7 logMAR).

Table.

 

Summary of Mixed-Effect Linear Regression Models Describing the Natural History of PCARE-Associated Retinopathy

 

KVFs were performed in 41 of 56 eyes (73%) and 21 of 28 patients (75%; Fig. 4C). At baseline, when considering the field delimited by the largest and brightest target V4e, 22% (9/41) of the eyes exhibited only a peripheral constriction, 20% (8/41) also presented with a large annular scotoma, 54% (22/41) showed severe constriction with residual peripheral perceptive islands, whereas 5% (2/41) were not able to perceive the target. The smaller and dimmer III1e target was seen by only 34% (14/41) of the eyes. At baseline, the median residual KVF area was 5799 degrees² (IQR = 642–8695 degrees²). After excluding eyes with a residual field diameter of less than 10 degrees at baseline, data from 20 of 41 eyes (49%) having serial KVFs repeated over a median follow-up time of 5.9 years (IQR = 4.0–7.1 years). Linear mixed-model analysis of the logarithm-transformed KVF areas considering the V4e isopter revealed a mean decay equivalent to 23% of the residual field per each year of follow-up (95% CI = 35% to 12%, P = 0.004) (see the Table). 

SW-AF and SD-OCT Analysis of Macular Atrophy

Among eyes with at least one DDAF lesion on SW-AF images, the area could not be quantified in 17% (8/50) due to either extensive lesions with unclear borders exceeding the field-of-view or poor image quality. Among the remaining 40 eyes (83%), the median DDAF area at baseline was 3.9 mm² (IQR = 1.3–9.8; Fig. 4E). Longitudinal data were available for 30 of 40 eyes (75%) over a median follow-up duration of 5.8 years (IQR = 3.6–6.6 years; Fig. 4F). The mean DDAF area growth rate was estimated at 1.18 mm²/year (95% CI = 0.69 to 1.65, P < 0.001). However, a strong positive correlation was observed between the random effects for the intercept and the slope across patients, suggesting that those with larger baseline DDAF areas tended to exhibit a faster growth rate (see the Table). After square root-transformation of the DDAF areas, the mean growth rate resulted 0.20 mm/year (95% CI = 0.16 to 0.23, P < 0.001), now without a correlation between the random intercept and slope (see the Table). A larger DDAF area was associated with a mean BCVA (logMAR) worsening by 0.06/mm² (95% CI = 0.03 to 0.10, P = 0.002) and a mean reduction in the KVF area by 12.4% (95% CI = 20.2% to 4.4%, P = 0.003; Supplementary Fig. S1). 

Volumetric data from macular OCT scans could be computed for 40 of 56 eyes (71%), with the median TMV at baseline being 1.69 mm³ (IQR = 1.53–1.82) at baseline (Fig. 4G). Longitudinal data were available for 25 of 40 eyes (63%) over a median follow-up of 5.5 years (IQR = 3.4–6.6 years; Fig. 4H). The mean slope of TMV decline over the follow-up was estimated at −0.015 mm³/year (95% CI = −0.023 to −0.007, P = 0.003; see the Table). Each 0.1 mm³ reduction in TMV was associated with a mean worsening of BCVA (logMAR) by 0.18 (95% CI = 0.01 to 0.25, P < 0001), but not with a significant reduction in the KVF area (P = 0.38; see Supplementary Fig. S1). 

Molecular Genetics

In our cohort, we identified 25 distinct PCARE (NM_001029883.3) variants were found in our cohort, comprising 13 (52%) frameshift, 10 nonsense (40%), and 2 (8%) missense variants. A detailed assessment of identified variants is provided in a Supplementary Table S2, whereas a schematic representation of the variant localization across the PCARE protein, along with the tolerance landscape of the protein residues to missense changes, is depicted in Supplementary Figure S2.³³ Eleven of 25 variants (44%) were previously not reported in the literature, including the following: c.398_401del [p.(Glu133Valfs*48)], c.469C>T [p.(Gln157*)], c.589A>G [p.(Lys197Glu)], c.814C>T [p.(Gln272*)], c.1139G>A [p.(Trp380*)], c.1302C>A [p.(Cys434*)], c.1786_1789del [p.(Glu597Argfs*147)], c.2802del [p.(Glu935Argfs*2)], c.3058del [p.(Gln1020Serfs*15)], c.3424_3425del [p.(Pro1142Thrfs*54)], and c.3717_3718del [p.(Cys1240Phefs*3)]. Frameshift or nonsense variants with a premature stop codon leading to nonsense mediated mRNA decay or truncated protein (null variants) accounted for the majority of alleles (54/56, 96%), and most patients (26/28, 93%) had a genotype consisting of two null variants. Therefore, genotype-phenotype correlation analyses were not conducted. Nearly all those variants were located in the first of the two exons of the gene, except for the c.3717_3718del [p.(Cys1240Phefs*3)] variant. This variant was found in two unrelated subjects and caused a frameshift resulting in a premature termination codon in the second exon of the transcript. This is predicted to lead to a truncated protein, where the last 49 amino acids are lost and replaced with 2 incorrect amino acids [p.(Cys1240Phefs*3)]. Sixteen patients (54%) were homozygous, and 12 (43%) were presumably compound heterozygous. Familial segregation of compound heterozygous variants could be confirmed in 5 of 12 patients (42%; see Supplementary Table S1). According to the American College of Medical Genetics and Genomics (ACMG) guidelines, 9 of 25 variants (36%) were classified as pathogenic, 15 (60%) as likely pathogenic, and 1 (4%) as variant of uncertain significance (VUS). The only VUS was the p.(Lys197Glu) missense variant, found in presumed compound heterozygosity (familial co-segregation was not available) with a nonsense pathogenic variant in a patient presenting with a disease phenotype consistent with PCARE-associated retinopathy. This variant was never found on the gnomAD version 4.0.0 database and affected an amino acid residue located just four positions upstream of the p.(Ile201Thr) variant (Supplementary Fig. S2), whose pathogenicity has already been demonstrated experimentally.^4,6 The c.3002G>A [p.(Trp1001*)] was the most prevalent allele in our cohort (11/56, 20%), observed in 7 unrelated patients originating from various countries in the Mediterranean area (France, Italy, Greece, Turkey, Romania, and Algeria). The c.2950C>T [p.(Arg984*)] allele was the second most frequent (10/56, 18%), present in the homozygous state in 4 patients from a large consanguineous family of Ashkenazi Jews from Egypt, and in the heterozygous state in 2 additional unrelated patients. 

Discussion

Since the approval of the first gene therapy for RPE65-associated IRD, interest in the other genes causing IRDs has grown exponentially. Gene augmentation remains the most suitable strategy for recessive conditions.¹¹ Among these, PCARE-associated retinopathy should be amenable to such a therapeutic approach and appears to be an ideal candidate for the development of a gene therapy product. This study characterizes the largest cohort of patients with PCARE-associated retinopathy as of this writing, providing a detailed analysis of the retinal phenotype using electrophysiology and imaging, and reporting on the natural history with longitudinal data over an average follow-up of almost 7 years, and exploring potential functional and structural outcome measures for upcoming clinical trials. 

PCARE-associated retinopathy manifests as a severe generalized photoreceptor dystrophy. In our cohort with a median age of 41 years, most patients exhibited almost no distinguishable waveform from background noise on both rod- and cone-mediated ERGs. However, a few younger patients had residual responses arising from cone photoreceptors, possibly indicating an earlier involvement of rods. This contrasts with the findings of Gerth-Kahlert et al., who studied a cohort of 13 patients with a mean age of 32 years and noted a more pronounced involvement of cones than rods in younger patients.¹⁴ Nonetheless, early degeneration of cone photoreceptors is evident from the extensive – yet heterogeneous – KFV alterations ensuing at a relatively young age, including severe peripheral restriction, and annular and central scotomas when considering the V4e isopter. Furthermore, some patients were unable to perceive smaller and dimmer targets. Additionally, the mean 23% annual rate of the KVF area decline outpaces estimates reported for typical RP, which range from 5% to 15%.^29,34–37 Along with severe visual field constriction, all patients exhibited a progressive form of MA, as previously noted.^14,15 In 86% of eyes, it appeared as a well-demarcated area of RPE atrophy corresponding to DDAF on SW-AF images, eventually associated with coarse pigment plaques and/or large choroidal excavations in older individuals. Despite this, low vision and blindness in the better-seeing eye were reached at median ages of 50 and 57 years, respectively, most probably because of a long-lasting preservation of the foveal anatomy. Furthermore, none of the patients experienced blindness in both eyes before the age of 40 years. These findings suggest that the optimal therapeutic window for future gene augmentation approaches might be before the fifth decade of life. However, the mean slope of BCVA (logMAR) loss was estimated at 0.06 per year, corresponding to approximately 3 ETDRS letters, with a wide standard deviation of 0.05, possibly reflecting a floor effect after the onset of severe vision impairment or blindness (i.e. after the loss of FS). Indeed, among eyes with available longitudinal data, 34% already had a baseline BCVA (logMAR) worse than 1.0. This limits the suitability of BCVA measurements as an endpoint for clinical trials, as already noted for other IRDs.³⁸ Consequently, we evaluated two structural outcome measures as potential surrogate endpoints for future clinical trials. In detail, MA could be reliably quantified as the DDAF area on SW-AF using semi-automated tools (such as the Region Finder incorporated in Heidelberg SPECTRALIS instruments) and enlarged progressively during the follow-up evaluated in this study, seemingly without a floor effect. Meanwhile, TMV reflects the anatomic status of the fovea and parafovea without the need for a binary categorization (i.e. presence/absence of FS). We observed a significant mean growth rate for square root-transformed DDAF areas, estimated at 0.20 mm/year, a figure identical to that estimated in a recent meta-analysis for ABCA4-associated retinopathy,³⁹ for which it already serves as a surrogate endpoint accepted by regulatory authorities.^25,32,40,41 Furthermore, a larger DDAF area demonstrated significant associations with worse visual function parameters, including BCVA and KVF area. On the other hand, TMV showed a significant mean linear decrease of approximately 0.02 mm³/year. As expected, a lower TMV was associated with a worsening of BCVA, but not with a reduced KVF area, limiting its ability to fully reflect the full range of declining visual function in patients affected with PCARE-associated retinopathy. 

Last, from the genetic perspective, we found that over 90% of patients had 2 variants leading to the formation of a premature termination codon, presumably resulting in null alleles because of non-sense mediated decay, consistent with existing literature.¹⁴ Thus, our results confirm the hypothesis that the disease mechanism stems from a complete loss-of-function of the PCARE protein in photoreceptor cilia. Furthermore, we identified 11 novel PCARE variants, expanding the array of about 80 published variants already associated with a clinical phenotype (http://lovd.nl/c2orf71). 

Limitations

The findings of our study must be interpreted considering its retrospective design, which inherently introduces differences in follow-up, available data, and testing methods among patients. Particularly, the KFV testing devices (manual Goldman and semi-automatic Octopus 900) varied over the years during which the data included in the study were acquired. In addition, the age of disease onset and initial symptoms were unknown for many patients. Furthermore, other tests with potential relevance as outcome measures in the field of IRDs, such as full-field stimulus threshold, mobility testing, chromatic pupil campimetry, and cone cell imaging were not routinely performed in clinical practice and represent areas for future research.^38,42,43 Last, phasing of variants was unavailable for many patients presumed to be compound heterozygous, implying the need for further reports to confirm the pathogenicity of the novel variants reported in this study, as well as functional assays for the p.(Lys197Glu) missense VUS. 

Conclusions

PCARE-associated retinopathy is characterized by a severe generalized photoreceptor dystrophy with MA. Despite early visual field loss, useful central vision is often retained into late adulthood because of FS. Based on the age of onset of legal blindness in the better-seeing eye, the optimal window for future interventions appears to be before the fifth decade of life. These findings could prove beneficial for patient counseling and designing clinical trials. 

Acknowledgments

DNA samples included in this study originate from NeuroSensCol DNA bank, dedicated for research in neurosensory disorders (PI: I Audo, partner with CHNO des Quinze-Vingts, Inserm and CNRS). 

Supported by the IHU FOReSIGHT [ANR‐18‐IAHU‐0001] by French state funds managed by the Agence Nationale de la Recherche within the Investissements d'Avenir program. The sponsor or funding organization had no role in the design or conduct of this research. 

Disclosure: L. Bianco, None; A. Antropoli, None; A. Benadji, None; R. Atia, None; O. Palacci, None; C. Condroyer, None; A. Antonio, None; J. Navarro, None; M. Battaglia Parodi, None; J.-A. Sahel, None; C. Zeitz, None; I. Audo, None