A Description of the Yield of Genetic Reinvestigation in Patients with Inherited Retinal Dystrophies and Previous Inconclusive Genetic Testing

In the present era of evolving gene-based therapies for inherited retinal dystrophies (IRDs), it has become increasingly important to verify the genotype in every case, to identify all subjects eligible for treatment. Moreover, combined insight concerning phenotypes and genotypes is crucial for improved understanding of thevisual impairment, prognosis, and inheritance. The objective of this study was to investigate to what extent renewed comprehensive genetic testing of patients diagnosed with IRD but with previously inconclusive DNA test results can verify the genotype, if confirmation of the genotype has an impact on the understanding of the clinical picture, and, to describe the genetic spectrum encountered in a Swedish IRD cohort. The study included 279 patients from the retinitis pigmentosa research registry (comprising diagnosis within the whole IRD spectrum), hosted at the Department of Ophthalmology, Skåne University hospital, Sweden. The phenotypes had already been evaluated with electrophysiology and other clinical tests, e.g., visual acuity, Goldmann perimetry, and fundus imaging at the first visit, sometime between 1988–2015 and the previous—in many cases, multiple—genetic testing, performed between 1995 and 2020 had been inconclusive. All patients were aged 0–25 years at the time of their first visit. Renewed genetic testing was performed using a next generation sequencing (NGS) IRD panel including 322 genes (Blueprint Genetics). Class 5 and 4 variants, according to ACMG guidelines, were considered pathogenic. Of the 279 samples tested, a confirmed genotype was determined in 182 (65%). The cohort was genetically heterogenous, including 65 different genes. The most prevailing were ABCA4 (16.5%), RPGR (6%), CEP290 (6%), and RS1 (5.5%). Other prevalent genes were CACNA1F (3%), PROM1 (3%), CHM (3%), and NYX (3%). In 7% of the patients there was a discrepancy between the diagnosis made based on phenotypical or genotypical findings alone. To conclude, repeated DNA-analysis was beneficial also in previously tested patients and improved our ability to verify the genotype–phenotype association increasing the understanding of how visual impairment manifests, prognosis, and the inheritance pattern. Moreover, repeated testing using a widely available method could identify additional patients eligible for future gene-based therapies.


Introduction
Inherited retinal dystrophies (IRDs) are one of the most common causes of serious visual impairment in children and young adults in developed countries [1,2]. Until quite recently, IRDs have been untreatable, but during the last decades, extensive research concerning gene-based therapies [3][4][5][6][7] has evolved and the first gene augmentation therapy, Voretigene Neparvovec for treatment of RPE65-associated retinal dystrophies [8,9], was approved in USA in 2017 and in Europe 2018. Since the novel therapies such as gene The study included 279 patients, with inconclusive previous DNA test results, from the retinitis pigmentosa research registry hosted at the Department of Ophthalmology, Skåne University Hospital, Lund, Sweden. Despite the name, the registry includes subjects with the whole spectrum of IRDs. The patients had made their first visit to the department between 1988 and 2015 and the initial appointment included a thorough clinical examination that mapped the phenotype carefully. Among the most prevalent diagnoses (based on the phenotype) were RP (94 subjects), CRD (38 subjects), Stargardt diseases (STGD) (24 subjects), X-linked juvenile retinoschisis (XLRS) (22 subjects), LCA (14 subjects), cone dystrophy (CD) (12 subjects), congenital stationary night blindness (CSNB) (11 subjects), macular dystrophy (11 subjects), and Usher syndrome (9 subjects). Previous DNA analyses were performed between 1995 and 2020 in cooperation with several collaborators, using both research laboratories and commercial facilities. Over time, the available techniques have developed from single gene tests and APEX panels to NGS panels and WES. Of the subjects, 122 had been tested with single-gene analysis, often including a range of genes on several occasions and in many different laboratories, while 157 of the patients that were investigated more recently had been tested with APEX-or NGS panels. A few cases with unsolved genotypes had also been tested with WES in addition to any of the other methods. In many cases, several DNA tests have been carried out over time. In this study, the term, inconclusive test results, means that either no pathogenic variant at all had been identified with previous tests or that only one pathogenic variant had been detected in a gene that is known to cause autosomal recessive disease. The study included 117 females and 162 males. They were all between 0 and 25 years of age at the time of their first visit (median 10 and mean 11 with standard deviation 6). Patients from widely distributed parts of Sweden are represented in the cohort, in which 60% had been referred from areas outside the department's own region, Skåne. Hence, these results provide information about the genetic characteristics of Swedish IRD patients on a national level rather than on a regional level. The study was conducted in accordance with the Tenets of the Declaration of Helsinki and it was approved by the Ethical Committee for Medical Research at Lund University (nr 2015/602). All subjects gave their informed consent concerning the study including the DNA analysis.

DNA Extraction
DNA was extracted from venous blood drawn from the precubital vein. Buffy coats of nucleated cells obtained from anticoagulated blood (EDTA) were resuspended in 15 mL polypropylene centrifugation tubes with 3 mL of nuclei lysis buffer (10 mM Tris-HCl, 400 mM NaCl and 2 mM Na2EDTA, pH 8.2). The cell lysates were digested overnight at 37 • C with 0.2 mL of 10% SDS and 0.5 mL of a protease K solution (1 mg protease K in 1% SDS and 2 mM Na2EDTA). After digestion was complete, 1 mL of saturated NaCl (approximately 6 M) was added to each tube and shaken vigorously for 15 s, followed by centrifugation at 2500 rpm for 15 min. The precipitated protein pellet was left at the bottom of the tube and the supernatant containing the DNA was transferred to another 15 mL polypropylene tube. Exactly 2 volumes of room temperature absolute ethanol were added, and the tubes inverted numerous times until the DNA precipitated. The precipitated DNA strands were removed with a plastic spatula or pipette and transferred to a 1.5 mL microcentrifuge tube containing 100-200 pl TE buffer (10 mM Tris-HCl, 0.2 mM Na2EDTA, pH 7.5). The DNA was allowed to dissolve for 2 h at 37 • C before quantitating.

Ophthalmological Examinations
For assessment of overall retinal function, full-field electroretinograms (ffERG) according to the ISCEV standards at the time [19,20] were recorded in all of the patients. In subjects that had their appointment after the multifocal electroretinography (mfERG) technique had been introduced (from 2002), macular function was measured with mfERG according to the ISCEV standards of the time [21,22]. Best corrected visual acuity (BCVA) was tested monocularly on a decimal letter chart at 5 or 3 m (m) and visual fields were mapped with a Goldmann perimeter, likewise monocularly, with standardized objects V4e, I4e, 04e, 03e, and 02e. For structural analysis, fundus color and red free photographs, and during later years, optical coherence tomography (OCT) and autofluorescence (FAF) images were also obtained. Moreover, slit lamp and fundus examinations were conducted.

Results
Pathogenic class 4 or 5 genetic variants explaining the phenotype were found in 182 of the 279 (65%) samples that were re-analyzed with the NGS retinal dystrophy panel. A description of the pathogenic variants as well as data concerning age at first examination, gender, genotype, and phenotype at first examination are presented in Table 2. The cohort was genetically heterogenous showing disease -causing variants in 65 different genes ( Figure 1 and Table 3). The most frequently mutated gene was the ABCA4 gene with pathogen variants in 30 of the 182 (16.5%) cases with a verified genotype. Other prevalent causative genes in this Swedish cohort were CEP290 (11 out of 182, 6%), RPGR (11 out of 182, 6%), RS1 (10 out of 182, 5.5%), CACNA1F (6 out of 182, 3%), CHM (6 out of 182, 3%), NYX (6 out of 182, 3%), and PROM1 (6 out of 182, 3%).
In 13 out of the 182 (7%) patients, there was a discrepancy between the diagnosis based on phenotypical or genotypical findings alone. The most common error was that CSNB initially was considered to be XLRS or choroideremia or that early choroideremia was mistaken for RP. In two cases, Bardet-Biedl syndrome initially was interpreted as achromatopsia before more general symptoms such as obesity and renal problems were apparent.  Table 3. Showing the spectrum of mutated genes that were found in the Swedish cohort of IRDs patients and the number of patients with pathogen variants in each specific gene.

Number of Patients
Showing the frequency of mutated genes that were found in the study. Table 3. Showing the spectrum of mutated genes that were found in the Swedish cohort of IRDs patients and the number of patients with pathogen variants in each specific gene.

Discussion
Since gene-based treatments like gene augmentation/replacement [8,9,[23][24][25][26][27][28][29][30][31], gene silencing, AONs [32,33], and gene editing using the CRISPR/Cas9 system [3,[5][6][7] may be the future for patients with IRDs, confirmation of the genotype has become even more crucial during the last years. In our department, we have, since the 1990s, strived to both perform careful phenotyping and to verify the genotype in all our IRD patients. However, we have failed to identify the causative genetic background in quite a few of them and therefore, we wanted to investigate if it is beneficial to perform genetic re-testing with a widely available broad NGS panel for IRDs. WES or WGS could possibly have revealed more pathogenic variants, but in this study, we wanted to test a method that is affordable in a clinical setting and for the health care systems in different countries. In these patients, that had previously been investigated with various techniques such as single-gene analysis, APEX panels, NGS panels, and WES with inconclusive results, the renewed testing with a comprehensive NGS panel revealed the presence of the genotype in 182 individuals (65%). Thus, the success rate was approximately the same as the general yield described for first time-testing using NGS (50-71%) [34,35] although our subjects were selected unsolved cases. This means that it is of great value to re-test IRDs patients with unsolved genotypes using a broad NGS panel for IRDs. When it comes to the cases with compound heterozygosity it would, of course, be ideal to perform segregation analyses for all of them, but in this study, it was not possible to make contact with and test relatives of all of the patients. In many cases, NGS data could confirm that the variants were in trans and in all cases we were very careful in the interpretation of the genetic data only considering the genotype as causative if it was completely consistent with the phenotype. It is difficult to set a proper interval for DNA re-testing. In our study, the positive yield of testing for the patients with the shortest re-test interval (previously tested between 2016-2020 with APEX or NGS panels) was 32 out of 49 samples (65%), which means a similar positive success rate as for the whole group, indicating the usefulness of re-testing with quite short intervals.
When it comes to the prevalence of different causative genes in this Swedish cohort, which to our knowledge is the first larger cohort investigated concerning the genetic spectrum in Sweden, the ABCA4 gene was the most common gene, encountered in 16.5% of the patients. This is in line with both an international estimate by Schneider et.al., 2022 called the Global Retinal Inherited Disease (GRID) dataset [13], and with reports from separate countries, although the absolute percentage varies slightly: GRID 25%, USA 14% [36], Canada 20% [36], Brazil 21% [37], Taiwan 15% [38], and Italy 26% [39]. Our secondmost common genes were RPGR and CEP290, which were found in 6% of the patients, respectively. RGPR is also among the most prevalent genes in other studies; fourth-most common in the GRID dataset (3.4%) [13], in USA and Canada it was the third-most common gene (10% and 4%) [36], the fourth-most common in Brazil (5%) [37], fifth-most common in Taiwan (5%) [38], and in the Italian cohort, it was the third-most common gene (5%) [39]. CEP 290, on the other hand, is only represented to the same extent in the Brazilian cohort (5.5%) [37], while it is less common in the other cohorts (1-3%) [13,36,38,39]. Another difference is that USH2A is quite common in the other studies, being the second-most prevalent gene in the GRID dataset (15%) [13] as well as in the Italian (11%) [39] and the Canadian (6%) [36] cohorts, the third-most prevalent gene in Taiwan and Brazil (10% and 5%, respectively) [37,38], and found in 3% of American IRD patients [36], while it was found in only two of our 182 patients (1.1%) with a verified genotype. It is well known that genes have different prevalences in various countries and geographic areas, but most of the difference concerning the USH2A gene in our study can be explained by the fact that the patients with Usher syndrome type 2A are referred to us at an older age (mean age 39 at genotypic diagnosis in our registry) than the investigated group, since their visual decline becomes evident somewhat later in life. EYS was also among the more common genes in some of the other cohorts; e.g., second-most common among the Taiwanese subjects (12%) [38], third-most common in the GRID dataset (4.4%) [13,36], and was found in 4% of Brazilian IRD patients [37], while it was actually absent from our study as well as from the American and Canadian cohorts [36]. Concerning RS1, the setting was the opposite. It was among the more common genes in our cohort, verified in 5.5% of the subjects, but less prevalent in the other studies, in which it was described in only 0.5-2% of the patients [37,39,40] or was not specified at all [13,36]. Thus, these data indicate that to an extent, the same genes are the most prevalent across different cohorts with the exception of certain genes, e.g., RS1, CEP290, and EYS, that show more inconsistent distribution. This is of special interest when it comes to introducing gene-based therapies, since particular genes have a more urgent need to be dealt with in some populations than in others. Figure 2 shows the genotypic pattern of the 201 patients with established genotypes belonging to the same age group in the RP registry. In this group, 45 different genotypes were demonstrated. It can be noted from Figures 1 and 2 that some genes such as CNGB3, RHO, CLN3, BEST1, BCM, and GUCY2D were quite well covered in the former analyses and not many new cases were encountered in the re-analysis. For ABCA4, RPGR, and RS1, new variants were discovered in rather many subjects although these genes were among the most prevalent causative genotypes also in the registry cohort and thus the coverage of those genes has improved. It is also noteworthy that CEP290, CACNA1F, CHM, and NYX are much better covered in the newer genetic work-up identifying many more subjects than in the registry cohort.
group, since their visual decline becomes evident somewhat later in life. EYS was also among the more common genes in some of the other cohorts; e.g., second-most common among the Taiwanese subjects (12%) [38], third-most common in the GRID dataset (4.4%) [13,36], and was found in 4% of Brazilian IRD patients [37], while it was actually absent from our study as well as from the American and Canadian cohorts [36]. Concerning RS1, the setting was the opposite. It was among the more common genes in our cohort, verified in 5.5% of the subjects, but less prevalent in the other studies, in which it was described in only 0.5-2% of the patients [37,39,40] or was not specified at all [13,36]. Thus, these data indicate that to an extent, the same genes are the most prevalent across different cohorts with the exception of certain genes, e.g., RS1, CEP290, and EYS, that show more inconsistent distribution. This is of special interest when it comes to introducing genebased therapies, since particular genes have a more urgent need to be dealt with in some populations than in others. Figure 2 shows the genotypic pattern of the 201 patients with established genotypes belonging to the same age group in the RP registry. In this group, 45 different genotypes were demonstrated. It can be noted from Figures 1 and 2 that some genes such as CNGB3, RHO, CLN3, BEST1, BCM, and GUCY2D were quite well covered in the former analyses and not many new cases were encountered in the re-analysis. For ABCA4, RPGR, and RS1, new variants were discovered in rather many subjects although these genes were among the most prevalent causative genotypes also in the registry cohort and thus the coverage of those genes has improved. It is also noteworthy that CEP290, CACNA1F, CHM, and NYX are much better covered in the newer genetic work-up identifying many more subjects than in the registry cohort. In the re-analyzed material, the gender distribution was slightly skewed, which can be explained by the occurrence of X-linked disorders that were encountered in 40 of the subjects (22% of the subjects with a verified genotype).
The basis for the choice of age range of 25 years or younger in the study was that younger patients are more suitable for future treatments, since many of the IRDs are progressive and thus, early detection is essential for enough viable retinal cells to be left for decent treatment results. Moreover, it is very important for young patients to obtain a correct diagnosis as early as possible, in order to enable adequate visual habilitation including visual aids, as well as fair expectations concerning the course of the visual impairment. In line with this, we can confirm the importance of a combined phenotypic and genotypic work-up, since in 7% of the patients, the result of genetic testing or clinical examinations alone led to different diagnoses, delaying correct counselling. For instance, In the re-analyzed material, the gender distribution was slightly skewed, which can be explained by the occurrence of X-linked disorders that were encountered in 40 of the subjects (22% of the subjects with a verified genotype).
The basis for the choice of age range of 25 years or younger in the study was that younger patients are more suitable for future treatments, since many of the IRDs are progressive and thus, early detection is essential for enough viable retinal cells to be left for decent treatment results. Moreover, it is very important for young patients to obtain a correct diagnosis as early as possible, in order to enable adequate visual habilitation including visual aids, as well as fair expectations concerning the course of the visual impairment. In line with this, we can confirm the importance of a combined phenotypic and genotypic work-up, since in 7% of the patients, the result of genetic testing or clinical examinations alone led to different diagnoses, delaying correct counselling. For instance, in some early cases, X-linked congenital stationary night blindness (CSNB) due to CACNA1F variants was diagnosed as X-linked RP with the risk of giving the family incorrect information concerning the progression of the disease over time, since CSNB is a stationary and XLRP a progressive disorder. In some cases, the genetic result was also important for the confirmation of the inheritance pattern.
To conclude, renewed DNA-analysis was also beneficial in previously tested patients with inconclusive genetic test results, and it improved our ability to verify the genotype-phenotype association increasing the understanding of visual impairment, disease prognosis, and sometimes the inheritance pattern. Thus, repeated testing using a widely available method may identify additional patients eligible for future gene-based therapies.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The authors have full control of all primary data and agree to allow the journal to review the data on request.