Molecular Vision 2022; 28:359-368
Received 28 February 2022 | Accepted 04 October 2022 | Published 06 October 2022
1Ruth & Bruce Rappaport Faculty of Medicine, Technion-Israel Institute of Technology, Haifa, Israel; 2Department of Ophthalmology, Rambam Health Care Campus, Haifa, Israel; 3Clinical Research Institute at Rambam, Rambam Health Care Campus, Haifa, Israel; 4Department of Ophthalmology, Hadassah-Hebrew University Medical Center, Jerusalem, Israel
Correspondence to: Tamar Ben-Yosef, Rappaport Faculty of Medicine, Technion, P.O. Box 9649, Bat Galim, Haifa 31096, Israel; Phone: 972-4-829-5228; FAX: 972-4-829-5225; email: email@example.com
Purpose: To identify the molecular mechanisms of the development of autosomal dominant retinitis pigmentosa (adRP) with incomplete penetrance in an Israeli Muslim Arab family.
Methods: Two patients with adRP underwent a detailed ophthalmic evaluation, including funduscopic examination, visual field testing, optical coherence tomography, and electroretinography. Genetic analysis was performed using a combination of whole exome sequencing (WES) and Sanger sequencing. The pathogenicity of the identified intronic variant was evaluated in silico using several web-based tools, in vitro using a minigene-based assay, and in vivo using reverse transcription PCR analysis of lymphocyte-derived RNA. The relative abundance of alternatively spliced transcripts was evaluated using amplicon-based next-generation sequencing. The relative expression levels of PRPF31 and CNOT3 were measured using quantitative PCR (qPCR) analysis.
Results: The two patients recruited in this study had childhood-onset RP, with night blindness as the initial symptom, followed by concentric restriction of the visual field. The funduscopic findings included narrowed retinal blood vessels and peripheral bone spicule pigmentation. By the third decade of life, the full-field electroretinography findings had been remarkably attenuated. In these patients, we identified a novel heterozygous intronic variant at position +5 of PRPF31 intron 11 (c.1146+5G>T). The same variant was also detected in one asymptomatic family member. Through in silico analysis, the variant was predicted to alter the splicing of intron 11. An in vitro splicing assay and a reverse transcription PCR analysis of lymphocyte-derived RNA revealed that the mutant allele yielded mainly a shorter transcript in which exon 11 was skipped. The skipping of exon 11 was expected to cause a frameshift and an aberrant truncated protein (p.Tyr359Serfs*29). The qPCR analysis revealed reduced PRPF31 expression levels in the mutation carriers, without a significant difference between the affected patient and his asymptomatic brother. We evaluated several factors that have been suggested to correlate with non-penetrance of PRPF31 mutations, including the number of cis-acting MSR1 elements adjacent to the PRPF31 core promoter, CNOT3 expression level, and CNOT3 rs4806718 single-nucleotide polymorphism. None of these factors correlated with non-penetrance in the family in this study.
Conclusions: We report a novel intronic mutation in PRPF31 underlying adRP. This report expands the spectrum of pathogenic mutations in PRPF31 and further demonstrates the importance of intronic mutations. Moreover, it demonstrates the phenomenon of incomplete penetrance previously associated with PRPF31 mutations. The fact that the non-penetrance in the family in this study could not be explained by any of the known mechanisms suggests the possible contribution of a novel modifier of PRPF31 penetrance.
Retinitis pigmentosa (RP), having a worldwide prevalence of approximately 1 in 4,000 individuals, is the most common form of inherited retinal dystrophy (IRD). In RP (also known as rod-cone degeneration), rod photoreceptors are initially more severely affected than cone photoreceptors; thus, the first clinical symptoms are usually night blindness and gradual restriction of the visual field. Ophthalmologic findings include characteristic pigmentation of the midperipheral retina, attenuation of retinal arterioles, and optic disc pallor . RP is one of the most genetically heterogeneous conditions in humans and can be inherited as autosomal dominant (AD), autosomal recessive, or X-linked. To date, more than 90 genes have been implicated in nonsyndromic RP, of which at least 31 are associated with an AD mode of inheritance (RetNet [Retinal Information Network]).
Pre-mRNA processing factor 31 (PRPF31) is a ubiquitous pre-mRNA splicing factor that is part of the U4/U6*U5 tri-small nuclear ribonucleoprotein complex of the spliceosome. The heterozygous mutations of PRPF31 are the major causes of adRP (RP11; OMIM No. 600138) , accounting for 6%–11% of cases in various populations . Most of more than 130 PRPF31 mutations reported to date are presumed loss-of-function variants, including frameshift, splice site, and nonsense or large-scale insertions or deletions . Genotype-phenotype correlation was observed between the mutation type and age of onset. The lowest age of onset is associated with nonsense and frameshift variants, followed by large deletions or splice variants. In-frame duplications, insertions, or missense variants show the highest median age of onset .
The abundance of loss-of-function mutations in patientswith RP11, including complete gene deletions, has led to a consensus view that haploinsufficiency is the disease mechanism in this form of RP . An interesting observation (although not statistically significant) is that patients with large-scale deletions (partial or complete gene loss) have a higher age of diagnosis than patients with other mutation types . On the basis of this observation, it could be postulated that an element of dominant negative effect is involved in cases of nonsense, frameshift, indel, in-frame, and missense variants compared with large deletions. This is further supported by recent works that have shown that at least some PRPF31 mutations are associated with a combined haploinsufficiency and dominant-negative disease mechanism [3,5].
A hallmark of PRPF31-associated RP is incomplete penetrance; that is, obligate carriers may be totally asymptomatic but can still pass pathogenic mutations to their offspring. This phenomenon has been reported worldwide in multiple families harboring various PRPF31 mutations . A major determinant of PRPF31 mutation penetrance is the expression level of the nonmutant PRPF31 allele. High expressors of this allele compensate for the deficiency of the mutant allele and have normal vision, whereas low expressors develop RP due to an insufficient amount of functional PRPF31 protein . Wild-type (wt) PRPF31 transcript abundance is a highly variable and heritable characteristic . Several factors contributing to the varied expression levels have been proposed, including expression quantitative trait loci (eQTLs) on chromosome 14, in trans with PRPF31 ; the variable expression level of CNOT3, encoding a subunit of the Ccr4-not transcription complex, which binds to and represses the transcription of the PRPF31 promoter [8,9]; and the number of cis-acting minisatellite repeat (MSR1) elements adjacent to the PRPF31 core promoter, which determines the level of PRPF31 transcriptional repression [10,11]. Based on these observations, the mechanism of incomplete penetrance in RP11 has been described as “variant haploinsufficiency,” in which the existence and/or the severity of disease depends on the type of the mutant allele inherited , and the levels at which this allele and the wt allele are expressed . Herein, we describe a family segregating adRP with incomplete penetrance due to an intronic mutation of the PRPF31 gene.
In this study, the tenets of the Declaration of Helsinki were followed. This study was approved by the institutional review boards of the participating medical centers, and written informed consent was obtained from all participants. Ophthalmic examination included measurement of the best-corrected visual acuity, visual field testing, slit-lamp biomicroscopy of the anterior segment and ophthalmoscopic examination after pharmacological pupillary dilatation, spectral-domain optical coherence tomography (SD-OCT), and full field electroretinography (ff-ERG).
Genomic DNA was extracted from venous blood samples using a high-salt solution in accordance with the standard protocol . Whole exome sequencing (WES) of Subject II:1 was performed at 3billion (Seoul, South Korea) using xGen Exome Research Panel v2 (Integrated DNA Technologies, Coralville, IA) and Novaseq 6000 (Illumina, San Diego, CA). Sequence reads were aligned to the reference human genome (GRCh37/hg19). Variants were called via the Franklin web-based pipeline by Genoox. The genotyping of family members was performed using PCR amplification of the relevant DNA segments with specific primers (Appendix 1), followed by Sanger sequencing.
An in vitro evaluation was performed using a minigene-based assay. To create wt and mutant minigene constructs, a 1,028-bp DNA fragment harboring PRPF31 exons 10, 11, and 12, and the introns between them was PCR amplified from the genomic DNA of the patient (Appendix 1). Fragments harboring either the wt or mutant allele were inserted in the pCMV-Script mammalian expression vector (Stratagene, La Jolla, CA). Constructs were transfected into COS-7 cells using the jetPEI transfection reagent (Polyplus-transfection, Illkrich, France). Cells were cultured in a DMEM culture medium supplemented with 10% fetal bovine serum (Biologic Industries, Beit Ha'emek, Israel) and maintained at 37 °C and 5% CO2. Twenty-four hours after transfection, total RNA was extracted from the cells using TRI Reagent (Sigma-Aldrich, St Louis, MO) and treated with RQ1 RNase-free DNase (Promega, Madison, WI). Reverse transcription was performed with 1 μg of DNase-treated total RNA in a 20-μl reaction volume using 200 U of M-MLV reverse transcriptase and 100 ng of random primers (Stratagene). Two microliters of cDNA were subjected to PCR amplification with a forward primer in exon 10 and a reverse primer in the vector (T7; Appendix 1). The PCR products were subcloned into the pGEM-Teasy vector (Promega), and several independent clones of each product were sequenced. The COS-7 cells used for this experiment were authenticated by genotyping 4 STR markers (D17S1304, D5S1467, D4S2408, and D19S245). Genotyping was performed with PCR amplification of each STR and direct sequencing, as described by Almeida et al. . The results (in terms of the repeat number for each STR) were compatible with those described for COS-7 cells (data not shown) .
For the in vivo evaluation, total RNA was isolated from fresh blood lymphocytes using TRI Reagent, treated with RQ1 RNase-free DNase, and reverse transcribed as described earlier. PCR amplification was performed with forward and reverse primers in exons 10 and 12, respectively (Appendix 1). For sequencing, the PCR products were cloned into the pGEM-Teasy vector (Promega).
Amplicon-based next-generation sequencing (NGS) was performed at the Technion Genomics Center. Lymphocyte-derived RT-PCR products (described in the previous section, “In vivo splicing analysis”) were prepared for NGS according to the “16S Library Preparation protocol” by Illumina. This included PCR product purification, barcode attachment, and second purification of the final libraries. The concentration of each library was measured using the Qubit dsDNA HS Assay Kit (Life Technologies, Thermo Fisher Scientific, Waltham, MA), and the size was determined using TapeStation 4200 with the High Sensitivity D1000 Kit (Agilent, Santa Clara, CA). All libraries were mixed into a single tube with equal molarity. DNaseq data were generated on Illumina MiSeq as 150 paired-end reads. Demultiplexing was executed using bcl2fastq version v22.214.171.1242 software (Illumina), with one barcode mismatch allowed and a minimum trimmed read length of 35. Quality control was assessed using Fastqc (v0.11.8), 150-bp paired-end reads were trimmed for adapters, low-quality 3′, and a minimum length of 30 using CUTADAPT (v1.10). Next, the reads were simultaneously mapped to two references: long.fasta, including exons 10, 11, and 12, and short.fasta, including exons 10 and 12. Mapping was performed using the BBSplit tool from BBTools (version 38.95) with default parameters, where ambiguous reads were binned to the first best reference.
cDNA was synthesized using the qPCRBIO High-Quality cDNA Synthesis Kit (PCR Biosystems, London, UK). qPCR reactions were prepared using 2X qPCRBIO Fast qPCR SyGreen Blue Mix (PCR Biosystems, London, UK) and performed using the QuantStudio3 Real-Time PCR System (Applied Biosystems, Thermo Fisher Scientific). Data were analyzed using the ∆∆CT method. Gene expression values were normalized against the GAPDH expression level. The primer sequences are presented in Appendix 1.
TB209 is a Muslim Arab family from Northern Israel, segregating adRP. The affected individuals (subjects II:1 and III:1; Figure 1A) were a mother and her son, aged 54 and 32 years old at the time of the last follow-up, respectively. Both reported childhood-onset night blindness as the initial symptom, which was associated with subjectively worsening dark and light adaptation, followed by a progressive, concentric restriction of the visual field. Subject II:1 also complained of decreasing visual acuity in both eyes several years prior. A posterior subcapsular cataract was present in both eyes of both subjects. The ophthalmoscopic findings included pallor of the optic nerve head, cellophane macular reflex, and narrowed retinal arteries with perivascular and midperipheral bone spicule pigmentation of the retina. Macular imaging using SD-OCT revealed loss of the outer nuclear layer (ONL) and inner segment ellipsoid zone (EZ) band, with preservation of this landmark only in the foveal center in Subject II:1. This patient also had focal vitreomacular traction with cystoid macular changes mainly in the right eye. In Subject III:1, the SD-OCT imaging revealed relative preservation of the EZ and ONL in the fovea, whereas ff-ERG at the age of 29 years was attenuated to a non-measurable level (Table 1 and Figure 2).
To identify the genetic cause of the disease in family TB209, WES analysis (including copy number variation detection) was performed on a DNA sample from Subject II:1. A heterozygous variant at position +5 of intron 11 in the PRPF31 gene (NM_ 015629) was identified (c.1146+5G>T; IVS11+5G>T; Figure 1B, C). Through in silico analysis, this variant was predicted to alter the splicing of intron 11 (Table 2). The same variant appeared to be heterozygous in Subject III:1 (affected) and in his unaffected brother (Subject III:3; Figure 1A). This novel variant has never been reported in patients with IRD (neither in the literature nor in ClinVar); is not registered in public databases, including the Genome Aggregation Database, Trans-Omics for Precision Medicine (TOPMed) Bravo, The Greater Middle East (GME) Variome, and the Exome Sequencing Project (ESP) 6500; and is absent in more than 1000 exomes of Israeli patients with IRD.
We initially used an in vitro splicing assay approach that enabled us to examine each allele (wt vs. mutant) separately and evaluate the effect of c.1146+5G>T on splicing. For this purpose, we created two minigene constructs (wt and mutant) harboring PRPF31 exons 10 to 12 and the introns between them, downstream of a CMV promoter. Constructs were transfected into COS-7 cells, followed by RNA extraction and RT-PCR analysis. To specifically detect PRPF31 transcripts obtained from the minigene constructs (and not from the endogenously expressed gene), we performed PCR amplification with a forward primer derived from exon 10 and a reverse primer derived from the expression vector (Figure 3A). The wt construct yielded both the expected product harboring exons 10, 11, and 12 (exon 11+) and a shorter product in which exon 11 was skipped (exon 11−). The mutant construct mainly yielded the shorter product (exon 11−; Figure 3A,B). The skipping of exon 11 was expected to cause a frameshift that yields an aberrant truncated protein (p.Tyr359Serfs*29).
To evaluate the effect of c.1146+5G>T on splicing in vivo, we performed an RT-PCR analysis of leukocyte-derived RNA samples for all available family members, with primers located in exons 10 and 12. As expected, exon 11+ transcripts were detected in all tested individuals. Exon 11− transcripts were mainly detected in the individuals heterozygous for the c.1146+5G>T allele, although faint products could also be observed in the wt individuals (Figure 4A). To quantify the amount of exon 11+ relative to that of exon 11− transcripts in the wt and heterozygous individuals, we applied amplicon-based NGS to the RT-PCR products of family members. The quantitative analysis of the NGS results confirmed that skipping of exon 11 occurred in both the wt and mutant individuals. Exon 11− reads composed 2%–5% of all unique reads in the wt individuals and 8%–10% in the individuals heterozygous for the c.1146+5G>T allele (Figure 4B, C).
As mentioned earlier, Subject III:3 was heterozygous for the c.1146+5G>T mutation. The clinical data of this individual were not available, but at the age of 32 years, he subjectively reported good vision and indicated that he was examined by an ophthalmologist, with no significant findings. Considering that both his mother and brother experienced onset of night blindness in childhood and that his brother's ff-ERG was non-recordable at the age of 29 years, it is unlikely that he was affected, although late onset of disease symptoms cannot be excluded.
One factor that affects the penetrance of PRPF31 mutations is the expression level of the nonmutant PRPF31 allele . CNOT3 negatively regulates the expression of PRPF31, and in some studies, the lymphoblasts of asymptomatic PRPF31 mutation carriers were found to have higher PRPF31 expression levels and lower CNOT3 expression levels than those of symptomatic carriers [8,9]. Hence, we evaluated several factors that have been suggested to correlate with the non-penetrance of PRPF31 mutations. One was the CNOT3rs4806718 single-nucleotide polymorphism (SNP; C/T). A correlation between the T allele of this polymorphism and non-penetrance was previously suggested . In family TB209, the CNOT3rs4806718 genotype did not correlate with disease penetrance, as all three carriers of the PRPF31 mutation, both symptomatic (subjects II:1 and III:1) and asymptomatic individuals (Subject III:3), were heterozygotes for this SNP. We also determined the number of cis-acting MSR1 elements adjacent to the PRPF31 core promoter; four-copy repeats have been associated with non-penetrance [10,11]. In family TB209, no such correlation was found, as all family members, including the non-penetrant individual, were homozygous for the three-copy MSR1 repeat. We then evaluated the relative expression levels of PRPF31 and CNOT3 in lymphocyte samples from TB209 family members using qPCR. The PCR primers used for PRPF31 amplification were located in exons 3 and 5, and therefore amplified all PRPF31 transcripts, including both the wt and mutant. However, given the low abundance of exon 11− transcripts (Figure 4), most amplification products were expected to represent wt transcripts. The analysis revealed that subjects III:1 (symptomatic) and III:3 (asymptomatic) had similar and reduced expression levels of PRPF31 (approximately 50% of the average expression level in the wt controls; Figure 5). The CNOT3 expression level showed no statistically significant difference between the two individuals and was similar to the average expression level in the wt controls (Figure 5). Finally, we genotyped the family members for three SNPs in the PRPF31 gene to reconstruct the PRPF31-linked haplotype in chromosomes harboring the wt rather than the mutant allele. The analysis revealed that subjects III:1 and III:3 inherited different copies of chromosome 19 from their father (Appendix 2). These results support the presence of a genetic modifier leading to the non-penetrance in Subject III:3 in cis with PRPF31, although the effect of such a putative modifier on the PRPF31 expression level was not apparent in the lymphocytes.
The primary aim of the present study was to identify the molecular mechanism of adRP in an Israeli Muslim Arab family. Genetic analysis revealed a novel heterozygous intronic mutation of PRPF31, c.1146+5G>T. The affected individuals carrying this mutation had classic symptoms of RP with childhood onset. The clinical findings from these patients were similar to those reported previously in patients with other PRPF31 mutations . To evaluate the effect of c.1146+5G>T on splicing, we performed in vitro and in vivo splicing analyses. The results of the analyses indicated that some degree of alternative splicing of exon 11 is normal and that the c.1146+5G>T allele further weakens the intron 11 donor splice site and enhances the skipping of intron 11. At the protein level, skipping of exon 11 was expected to cause a frameshift that yields an aberrant truncated protein (p.Tyr359Serfs*29). The presence of such a truncated protein was not confirmed experimentally. If it was indeed generated, it could be associated with either loss of function and/or a dominant negative effect . Nevertheless, although exon 11− transcripts were more abundant in the individuals harboring the c.1146+5G>T allele than in the wt individuals, the relative frequency of these transcripts was low in both groups. Moreover, the total PRPF31 expression level in the carriers of the c.1146+5G>T allele was approximately half of the expression level in the wt individuals. These results indicate that the mutant transcript was probably unstable and may be subjected to the NMD mechanism and support haploinsufficiency as the disease mechanism associated with the c.1146+5G>T mutation. In this context, it is of special interest to mention another PRPF31 mutation, c.1115_1125del. This mutation is located in exon 11 and can inactivate an exonic splicing enhancer and thus promote the skipping of exon 11 during mRNA splicing and result in an out-of-frame premature termination codon in exon 12, similar to the effect of c.1146+5G>T . Exon 11− transcripts generated as a result of this mutation were shown to undergo degradation through the NMD mechanism .
One family member carried the c.1146+5G>T mutation with no phenotypic expression. This phenomenon has been previously reported in families segregating PRPF31 mutations . However, we could not identify the factors contributing to the reduced penetrance in this family. The CNOT3rs4806718 genotype did not correlate with disease penetrance in this family, which is in agreement with a previous study  that did not replicate the original observation made by Venturini et al. . RP11 disease non-penetrance was also associated with the presence of a four-copy MSR1 repeat in the PRPF31 promoter in many RP11 asymptomatic carriers [10,11]. In family TB209, all family members, including the asymptomatic carrier, were homozygous for the three-copy MSR1 repeat. PRPF31 penetrance correlates, at least in part, with the expression level of the nonmutant PRPF31 allele , which reversely correlated with the CNOT3 expression level [8,9]. In family TB209, the asymptomatic carrier had a reduced PRPF31 expression level, which was similar to the expression level in his affected brother. The CNOT3 expression levels were similar in both individuals and the wt controls. However, these expression levels were measured in blood lymphocytes and may not reflect the expression levels in retinal photoreceptors. For example, in a study performed by McLenachan et al. , the PRPF31 expression level of a non-penetrant carrier was not higher than that of symptomatic individuals when measured in skin fibroblasts but was an intermediate between wt and symptomatic individuals when measured in retinal organoid cultures (differentiated from induced pluripotent stem cells). It is possible that the asymptomatic carrier in family TB209 had an increased expression level of the wt PRPF31 allele in his retina, which compensated for the mutant allele. Alternatively, the non-penetrance in this individual might have resulted from other genetic and/or environmental factors that are still to be determined.
In conclusion, this report expands the spectrum of pathogenic mutations in PRPF31 and further demonstrates the importance of intronic mutations. Moreover, it demonstrates the phenomenon of incomplete penetrance, which was previously associated with PRPF31 mutations. The fact that non-penetrance in the family in this study could not be explained by any of the known mechanisms suggests the possible contribution of a novel modifier of PRPF31 penetrance.
Appendix 1. Primers used in this study.
Appendix 2. PRPF31-linked haplotypes in family TB209.
We are grateful to all family members for their participation in this study. We thank Elizabeta Ginzburg, Margalit Peleg and Liat Linde for amplicon-based NGS. This work was supported by research grant BR-GE-0518–0734 from the Foundation Fighting Blindness to TB, DS and EB.