Determinants of Disease Penetrance in PRPF31-Associated Retinopathy

Retinitis pigmentosa 11 (RP11) is caused by dominant mutations in PRPF31, however a significant proportion of mutation carriers do not develop retinopathy. Here, we investigated the relationship between CNOT3 polymorphism, MSR1 repeat copy number and disease penetrance in RP11 patients and non-penetrant carriers (NPCs). We further characterized PRPF31 and CNOT3 expression in fibroblasts from eight RP11 patients and one NPC from a family carrying the c.1205C>T variant. Retinal organoids (ROs) and retinal pigment epithelium (RPE) were differentiated from induced pluripotent stem cells derived from RP11 patients, an NPC and a control subject. All RP11 patients were homozygous for the 3-copy MSR1 repeat in the PRPF31 promoter, while 3/5 NPCs carried a 4-copy MSR1 repeat. The CNOT3 rs4806718 genotype did not correlate with disease penetrance. PRFP31 expression declined with age in adult cadaveric retina. PRPF31 and CNOT3 expression was reduced in RP11 fibroblasts, RO and RPE compared with controls. Both RP11 and NPC RPE displayed shortened primary cilia compared with controls, however a subpopulation of cells with normal cilia lengths was present in NPC RPE monolayers. Our results indicate that RP11 non-penetrance is associated with the inheritance of a 4-copy MSR1 repeat, but not with CNOT3 polymorphisms.


Introduction
Inherited retinal diseases (IRDs) are the most common cause of blindness in the working-age population [1,2]. Retinitis pigmentosa (RP) is the most common form of IRDs accounting for almost half of the Australian national IRD registry [3]. Characterised by initial night vision impairment followed by progressive symmetrical constriction of the visual field, RP contributes to 64% of IRD-related blindness certification at a mean age of 48 years [1]. The Online Mendelian Inheritance in Man (OMIM) has listed RP genotype sequentially from RP1 to RP91. RP11, caused by mutations in the pre-mRNA processing factor 31 (PRPF31) gene (OMIM #600138) accounts for 5-10% of all the autosomal dominant forms of RP worldwide [4][5][6]. RP11 typically presents as a progressive rod-cone dystrophy, with a highly variable age of onset within and between affected families. There are two unique features of PRPF31-related disease. First, the carriers of pathogenic PRPF31 mutations can exhibit non-penetrance despite the manifestation of a severe RP phenotype in other family members carrying the same mutation [7,8]. Second, mutations in this gene lead to non-syndromic RP despite the ubiquitous expression of PRPF31 and its important role in splicing in all cell types. A deeper understanding of the molecular mechanisms underlying non-penetrance and non-syndromic manifestations of RP11 are essential for the development of safe and effective therapy for RP11.
The PRPF31 gene encodes the pre-mRNA processing factor-31 protein, which forms part of the U4/U6/U5 small nuclear ribonucleoprotein (snRNP) subunits of the spliceosome, a large RNP complex central to the process of pre-mRNA splicing [9][10][11][12][13]. Pre-mRNA splicing is an essential step in gene expression whereby non-coding intronic sequences are removed from the transcribed Pre-mRNA molecule to produce a mature mRNA transcript carrying the protein coding sequence. Several explanations have been put forward for the non-penetrance of PRPF31 mutations in asymptomatic carriers. Studies in immortalized lymphoblastoid cell lines (LCLs) harboring PRPF31 mutations demonstrated a link between PRPF31 expression and disease penetrance, with increased expression in non-penetrant carriers (NPCs) compared with affected RP11 patients [14][15][16]. Furthermore, genetic linkage studies showed disease non-penetrance was associated with inheritance of modifier sequences located within the 19q13.4 interval that contains the PRPF31 gene [14,17]. These observations led to the theory that disease non-penetrance was influenced by a penetrance-modifying sequence within 19q13.4 located in a gene proximal to PRPF31. In 2012, Venturini et al. proposed that this modifier gene was CNOT3 [18], which encodes a subunit of the Carbon Catabolite Repression-Negative On TATA-less (CCR4-NOT) transcriptional complex, a master regulator of global gene expression [19]. These authors identified a single nucleotide polymorphism (rs4806718, T/T), located within intron 17 of CNOT3 that was significantly associated with disease non-penetrance in two RP11 pedigrees [18]. Following on from their previous work, Rose et al. later published strong evidence indicating that a significant proportion of NPCs in RP11 families may also be attributed to the inheritance of a highly expressive, unaffected PRPF31 allele [20]. In that study, they identified the minisatellite repeat element 1 (MSR1) within the PRPF31 promoter, and further showed that a 4-copy repeat element was associated with higher expression of the PRPF31 allele in LCLs, while inheritance of a 3-copy repeat led to lower expression. In patients with PRPF31 mutations, haploinsufficiency of PRPF31 protein is thought to lead to a global increase in non-productive splicing, affecting the expression of a large number of retinal genes [21,22]. These global effects on splicing cause defects in cilia length in photoreceptor progenitor cells (PPCs) and retinal pigment epithelium (RPE) produced from induced pluripotent stem cells (iPSC) derived from patients with RP11 [21]. These cellular phenotypes may serve as a suitable disease model for studying the non-penetrant and non-syndromic features of RP11.
We recently described the clinical features of RP11 from eight unrelated families [23,24]. Here, we investigated the role of CNOT3 and the 4-copy MSR1 repeat in RP11 nonpenetrance in a cohort of 29 individuals carrying eight different pathogenic PRPF31 mutations from eight unrelated families. We examined the relative expression of PRPF31 and CNOT3 in dermal fibroblasts as well as induced pluripotent stem cell (iPSC)-derived retinal organoids (RO) and retinal pigment epithelium (RPE) and characterized the primary cilia in iPSC-derived RPE.

Human Tissues
The collection and use of human tissues in this study was approved by the University

Clinical Assessment
Patients with genetically confirmed PRPF31 mutation underwent comprehensive routine ophthalmic assessment including detailed medical and ocular history, best-corrected distance visual acuity (BCVA), complete slit lamp eye examination, functional assessment and structural multimodal retinal imaging (see below). A clinical diagnosis of RP was made based on history of impaired night vision, funduscopic findings of attenuated retinal vessels, bone spicules and retinal atrophy, and the observation of a profound reduction in dark-adapted full-field ERG response, where available. The NPCs were determined based on lack of symptoms of nyctalopia or constricted visual field, absence of typical fundus features on retinal examination or multimodal imaging and a normal visual field if available, despite the presence of a known disease associated PRPF31 variant. Detailed clinical descriptions and pedigrees for the eight families in our studies have previously been published [20,21].

Genetic Diagnosis and Pathogenicity Assessment
DNA samples were sourced, processed and stored as detailed previously [25]. Genetic screening of RP11 patients was previously reported [23,24]. Variants are described in accordance with Human Genome Variation Society recommendations [26]. Nucleotide one corresponds to the A of the start codon, ATG. The OMIM accession number and gene reference sequence utilised are #600138 and NM_015629.3, respectively. Pathogenicity was assessed as previously reported [27] and interpreted in accordance with current guidelines proposed by the American College of Medical Genetics and Genomics and the Association for Molecular Pathology [28], and Jarvik and Browning [29].

CNOT3 Polymorphism and Copy Number of MSR1
For genotyping the CNOT3 rs4806718 single nucleotide polymorphism, DNA was extracted from blood samples or cultured fibroblasts using the FlexiGene DNA kit (Qiagen, Hilden, Germany). PCR amplification was followed by Sanger sequencing. Primers used in this study are listed in Supplementary Table S1. The number of copies of the MSR1 repeat element adjacent to the PRPF31 promoter was assessed using PCR analysis. Primers were designed to amplify gDNA across the reference MSR1 region (hg 19 coordinates chr19:54617996-54619393), as previously reported by Rose et al. [20].

Dermal Fibroblast Culture
Human dermal fibroblasts were obtained from the Western Australian Retinal Disease Study Biobank and cultured in DMEM medium supplemented with 10% fetal calf serum and antibiotic-antimycotic (ThermoFisher, Waltham, MA, USA). Media was changed twice per week. Fibroblasts were harvested for RNA analysis across three separate passages for cells from each patient. For reprogramming and gene expression measurements, early passage fibroblasts were used (<p8).

Cadaveric Adult Human Retina
Adult human eyes were obtained from 7 male and 4 female donors aged 22-64 years from the Lions Eye Bank, Western Australia. The mean time (±SD) from death to retinal tissue harvesting was 18 (±5) h. The retina was dissected from the posterior eye cup and stored at −80 • C in RNAlater. Total RNA was isolated from whole retinae using TRIzol Reagent (ThermoFisher, Waltham, MA, USA).

Induced Pluripotent Stem Cells
Fibroblasts from a 28 year old male RP11 patient and a 54yo female NPC from the same RP11 family ( c.1205C > A variant) were previously reprogrammed into iPSC [30]. For this study, one additional iPSC line from this RP11 patient and NPC were utilized. Additional iPSC lines were generated from two unrelated RP11 patients, one carrying a deletion of exons 2-8 and the other carrying the c.267del variant in PRPF31. For controls, we utilized iPSC lines previously generated from a 24 year old healthy control subject. Human iPSC were cultured in feeder-free conditions, on geltrex (ThermoFisher, Waltham, MA, USA) coated culture plates in mTeSR1 medium (StemCell Technologies, Vancouver, BC, Canada), as previously described [30].

Retinal Organoid Differentiation
Human iPSC colonies were dissociated into small pieces by mechanical dissociation after 2-4 min of incubation with EDTA buffer (0.5 mM EDTA and 30 mM NaCl in DPBS without Ca 2+ or −Mg 2+ ). Cell clusters were seeded into suspension culture plates and cultured in DMEM/F-12 medium (ThermoFisher, Waltham, MA, USA) containing MEM Non-Essential Amino Acids Solution (NEAA, ThermoFisher, Waltham, MA, USA), B27 (ThermoFisher, Waltham, MA, USA), human recombinant IGF-1 (StemCell) and KnockOut Serum Replacement (KOSR, ThermoFisher, Waltham, MA, USA). The KOSR concentrations were gradually decreased from 20% for the first 5 days, to 15% until day 12 and to 10% until day 35). From day 35, KOSR was removed from the media. Half media changes were performed every day, with full media changes every second-third day, up to 200 days of culture.

RPE Differentiation
In addition to RO, the abovementioned protocol for RO differentiation resulted in the production of highly pigmented 'RPE organoids' that are largely comprised of RPE cells. These RPE organoids were removed from RO cultures after 6-8 weeks of differentiation and transferred into geltrex-coated 24-well plates with RPE media containing DMEM/F-12 medium with 1x NEAA, 1x B27, human recombinant 10 ng/mL IGF-1 (Stem Cell Technologies) and 4% KOSR (ThermoFisher, Waltham, MA, USA). Media was changed every third day. RPE organoids formed pigmented cell clusters within 4-6 weeks (RPEp0). RPE clusters were dissociated by incubation with TrypLE Express Enzyme for 20 min at 37 • C, and passaged into new geltrex coated wells using a split ratio of 1:5. Pure RPE monolayers were usually established after two passages. RPE monolayers were used for experiments at passage 3-4, 4 weeks after plating.

Quantitative PCR
Total mRNA was isolated using TRIzol Reagent (ThermoFisher, Waltham, MA, USA) and cDNA was synthesized using the RT 2 First Strand Kit (Qiagen, Hilden, Germany). qPCR reactions were prepared using RT 2 SYBR Green qPCR Mastermixes (Qiagen, Hilden, Germany) and performed using the CFX Connect Real-Time System (BioRad Laboratories, Hercules, CA, USA). Data was analyzed using the ∆∆CT method. Gene expression values were normalized against GAPDH expression. For fibroblast analysis, gene expression values were calculated as fold-change compared with expression level in fibroblasts derived from a 24-year old male control subject. For retinal organoid analysis, gene expression values were calculated as fold-change compared with the mean expression value derived from three cadaveric human retinal samples from working aged adults (28-60 years old).
For RPE analysis gene expression values were calculated as fold-change compared with control iPSC-derived RPE. Primers used in this study are listed in Supplementary Table S1.

Immunostaining
Cells were fixed with 4% paraformaldehyde for 15 min at 37 • C, washed, then permeabilized using 0.3% Triton X-100 in phosphate buffered saline (PBS) for 15 min. The cells were then incubated in blocking buffer (5% normal goat serum and 0.3% Triton X-100 (Sigma-Aldrich, St. Louis, MO, USA) in PBS) for 1 h at room temperature. Primary antibodies were diluted in blocking buffer and applied at 4 • C, overnight. Secondary antibodies were diluted in blocking buffer and applied for 2 h at room temperature. Nuclei were stained with DAPI (1 µg/mL). Antibodies and dilution factors used in this study are listed in Supplementary Table S2. Cells were examined using the Olympus BX60 fluorescence microscope and images were captured using Olympus DP-Controller 3.1.1.267 acquisition software (Olympus Corporation, Tokyo, Japan).

Scanning Electron Microscopy
For scanning electron microscopy (SEM), RPE cells were passaged onto Geltrex coated Millicell hanging cell culture inserts with pore size of 0.4 µm (Millipore Sigma, Burlington, MA, USA). After 4 weeks of culture, inserts were removed following Lynn's method and fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) for 20 min at room temperature. Critical dehydration and platinum coating were performed according to a previously published SEM sample preparation protocol [31]. Finally, SEM samples were visualized using a field emission scanning electron microscope (Carl Zeiss Supra 55 FESEM, Zeiss, Oberkochen, Germany).

Cilia Length Analysis
Primary cilia in RPE monolayers were labeled by immunostaining for ARL13B followed by counterstaining with DAPI. Cells were imaged using the Nikon Instruments A1 Confocal Laser Microscope. For primary cilium length measurements, maximum projection intensity images were generated from confocal stacks using NIS Elements Viewer (Version 4.11.0, Laboratory Imaging) and cilia lengths measured using Image J64 software (National Institute of Health, Bethesda, MD, USA). Mean cilia lengths were compared using the student t-test, with p < 0.05 considered significant. Comparison of DAPI stained nuclei counts and cilia counts in each field indicated >98% of cells displayed a primary cilium. Unciliated cells were not included in our analysis.

Demographics of RP11 Patients
A total of 29 individuals from eight unrelated RP11 families contributed DNA for CNOT3 and MSR1 genotyping. Five individuals from four families had no symptoms or signs of RP. Affected individuals had a mean ± SD age of 14 ± 11 years at the onset of symptoms. Clinical phenotypes have been previously published [23,24].

The CNOT3 rs4806718 Polymorphism Does Not Correlate with RP11 Disease Penetrance
Amongst the five NPCs, three (60%) carried the C/T genotype whilst two (40%) carried the T/T genotype at the rs4806718 locus. Amongst the 24 affected RP11 patients, 16 (66%) carried the C/T genotype, four (17%) carried the T/T genotype and four (17%) carried the C/C genotype at the rs4806718 SNP ( Table 1). The higher prevalence of the T/T rs4806718 genotypes in NPCs compared to affected RP11 patients (40% vs. 17%) was not statistically significant (χ 2 = 1.37, p = 0.24) suggesting this locus does not strongly influence disease penetrance in our RP11 cohort.

The 4-Copy MSR1 Repeat Is Associated with RP11 Disease Penetrance
Genotyping of the MSR1 repeat element demonstrated the inheritance of a homozygous 3-copy MSR1 repeat element (3/3 genotype) in all affected patients, while three of the five NPCs carried a 4-copy repeat element (3/4 genotype) (Table 1). Together, these results demonstrate disease non-penetrance is associated with the inheritance of PRPF31 alleles containing 4-copy MSR1 repeats in their promoter region.

PRPF31 and CNOT3 Expression Is Reduced in RP11 Patient Fibroblasts
Dermal fibroblast cultures were derived from eight patients with RP11 and one asymptomatic non-penetrant carrier (NPC) from one RP11 family carrying the c.1205C > A variant [23], as well as three unrelated control subjects, aged 24, 46 and 77 years old (yo). PRPF31 and CNOT3 mRNA expression was measured in fibroblast cDNA samples by quantitative PCR (qPCR) analysis. The mean level of PRPF31 expression across the RP11 patient group (n = 8) was not significantly different from the control group (n = 3, p = 0.055). Interestingly, mean PRPF31 expression levels in fibroblasts from RP11 patients in their teens (< 20yo, n = 3) were similar to controls, while expression in fibroblasts derived from adult patients (> 28yo) was significantly reduced compared with control fibroblasts (n = 5, p = 0.0173, Figure 1A).
CNOT3 expression was significantly reduced in the RP11 patient group (p = 0.0170) compared with the control group ( Figure 1B). Within the RP11 patient group, teenaged patients showed significantly lower levels of CNOT3 expression than controls (n = 3, p = 0.0095). Mean CNOT3 expression levels were similarly reduced in adult RP11 patients, however this reduction was not significant compared with controls (p = 0.0717). Together, these results suggest that PRPF31 and CNOT3 expression is reduced in fibroblasts carrying the c.1205C > A variant.

The 4-Copy MSR1 Repeat Is Associated with RP11 Disease Penetrance
Genotyping of the MSR1 repeat element demonstrated the inheritance of a homozygous 3-copy MSR1 repeat element (3/3 genotype) in all affected patients, while three of the five NPCs carried a 4-copy repeat element (3/4 genotype) (Table 1). Together, these results demonstrate disease non-penetrance is associated with the inheritance of PRPF31 alleles containing 4-copy MSR1 repeats in their promoter region.

PRPF31 and CNOT3 Expression Is Reduced in RP11 Patient Fibroblasts
Dermal fibroblast cultures were derived from eight patients with RP11 and one asymptomatic non-penetrant carrier (NPC) from one RP11 family carrying the c.1205C>A variant [23], as well as three unrelated control subjects, aged 24, 46 and 77 years old (yo). PRPF31 and CNOT3 mRNA expression was measured in fibroblast cDNA samples by quantitative PCR (qPCR) analysis. The mean level of PRPF31 expression across the RP11 patient group (n = 8) was not significantly different from the control group (n = 3, p = 0.055). Interestingly, mean PRPF31 expression levels in fibroblasts from RP11 patients in their teens (<20yo, n = 3) were similar to controls, while expression in fibroblasts derived from adult patients (>28yo) was significantly reduced compared with control fibroblasts (n = 5, p = 0.0173, Figure 1A).
CNOT3 expression was significantly reduced in the RP11 patient group (p = 0.0170) compared with the control group ( Figure 1B). Within the RP11 patient group, teenaged patients showed significantly lower levels of CNOT3 expression than controls (n = 3, p = 0.0095). Mean CNOT3 expression levels were similarly reduced in adult RP11 patients, however this reduction was not significant compared with controls (p = 0.0717). Together, these results suggest that PRPF31 and CNOT3 expression is reduced in fibroblasts carrying the c.1205C>A variant. (B) were calculated for the healthy control subject group (24-77 years old, n = 3) and the affected RP11 patient group (16-51 years old, n = 8). The RP11 group was further subdivided into teenage (<20 years old, n = 3) and adult (>28 years old, n = 5) patients. Statistical significance was calculated using a two-tailed t-test (* p < 0.05, ** p < 0.01, compared with control group). The RP11 group was further subdivided into teenage (<20 years old, n = 3) and adult (>28 years old, n = 5) patients. Statistical significance was calculated using a two-tailed t-test (* p < 0.05, ** p < 0.01, compared with control group).

PRPF31 mRNA Expression Is Reduced with Age in Cadaveric Adult Retina
PRPF31 mRNA levels were higher in adult human retina, iPSC-derived retinal organoids (ROs) and RPE cells compared with fibroblasts, with the highest levels detected in RPE monolayers ( Figure 2A). In retinal tissues obtained from donated human eyes, PRPF31 expression was found to decrease with advancing donor age ( Figure 2B Linear regression analysis demonstrated a moderate negative correlation between PRPF31 expression and age (R 2 = 0.495, Figure 2C). In contrast, CNOT3 expression was similar in retinal tissues from old and young subjects and did not correlate with age (R 2 = 0.043, Figure 2B-C).
PRPF31 mRNA levels were higher in adult human retina, iPSC-derived retinal organoids (ROs) and RPE cells compared with fibroblasts, with the highest levels detected in RPE monolayers (Figure 2A). In retinal tissues obtained from donated human eyes, PRPF31 expression was found to decrease with advancing donor age ( Figure 2B), with older donors (57-64 years old) showing significantly reduced PRPF31 expression compared with younger donors (22-41 years old, p = 0.0189). Linear regression analysis demonstrated a moderate negative correlation between PRPF31 expression and age (R 2 = 0.495, Figure 2C). In contrast, CNOT3 expression was similar in retinal tissues from old and young subjects and did not correlate with age (R 2 = 0.043, Figure 2B-C).

PRPF31 mRNA Expression Is Reduced in RP11 Retinal Organoids
To examine PRPF31 expression in retinal cells, we differentiated iPSC derived from the RP11 patient and NPC carrying the c.1205C>A mutation or a control subject into RO ( Figure 3A). After six months of differentiation, expression of the retinal markers PAX6, OTX2, CRX and RCVRN was upregulated in RO compared with undifferentiated iPSC (Supplementary Figure S1A). Additionally, RT-PCR analysis demonstrated the retinal specific transcript variant of BBS8 was expressed in RO cultures (Supplementary Figure  S1B).

PRPF31 mRNA Expression Is Reduced in RP11 Retinal Organoids
To examine PRPF31 expression in retinal cells, we differentiated iPSC derived from the RP11 patient and NPC carrying the c.1205C > A mutation or a control subject into RO ( Figure 3A). After six months of differentiation, expression of the retinal markers PAX6, OTX2, CRX and RCVRN was upregulated in RO compared with undifferentiated iPSC (Supplementary Figure S1A). Additionally, RT-PCR analysis demonstrated the retinal specific transcript variant of BBS8 was expressed in RO cultures (Supplementary Figure S1B).
PRPF31 expression was significantly reduced in RP11 patient RO (p = 0.0231) compared with control RO. In contrast, NPC RO showed an intermediate PRPF31 expression level that was not significantly different from control RO (p = 0.2821) or RP11 patient RO (p = 0.0875) ( Figure 3B). No significant differences in CNOT3 expression were detected between control, NPC and RP11 patient RO ( Figure 3C).

of 17
PRPF31 expression was significantly reduced in RP11 patient RO (p = 0.0231) compared with control RO. In contrast, NPC RO showed an intermediate PRPF31 expression level that was not significantly different from control RO (p = 0.2821) or RP11 patient RO (p = 0.0875) ( Figure 3B). No significant differences in CNOT3 expression were detected between control, NPC and RP11 patient RO ( Figure 3C). Each bar represents the mean from 3-4 retinal organoids from each subject. Error bars indicate standard deviation. Statistical significance was calculated using a two-tailed t-test (* p < 0.05). Expression of the spliceosome genes PRPF31, PRPF4, PRPF6, PRPF8 and SNRNP200 was measured in day 35 retinal organoid cultures (D) derived from control subjects (black bars), the NPC (dark grey bars) and the RP11 patient (light grey bars) by qPCR. Error bars indicate standard deviation.

Expression of Spliceosome Genes Is Reduced in RP11 RetinalOorganoids
In addition to reduced PRPF31 expression, early ROs (day 35) derived from the RP11 patient and the NPC displayed reduced expression of the spliceosome genes PRPF6, PRPF8 and SNRNP200 compared with control ROs. In contrast, PRPF4 showed similar levels of expression in RP11 patient, NPC and control ROs ( Figure 3D).

PRPF31 and CNOT3 Expression Is Reduced in RP11 iPSC-Derived RPE
RPE cells were differentiated from iPSC derived from a control subject, an RP11 patient and an NPC carrying the c.1205C>A variant, an RP11 patient carrying a deletion of exons 2-8 and an RP11 patient carrying the c.267del variant ( Figure 4A). RPE cells formed

Expression of Spliceosome Genes Is Reduced in RP11 RetinalOorganoids
In addition to reduced PRPF31 expression, early ROs (day 35) derived from the RP11 patient and the NPC displayed reduced expression of the spliceosome genes PRPF6, PRPF8 and SNRNP200 compared with control ROs. In contrast, PRPF4 showed similar levels of expression in RP11 patient, NPC and control ROs ( Figure 3D).

PRPF31 and CNOT3 Expression Is Reduced in RP11 iPSC-Derived RPE
RPE cells were differentiated from iPSC derived from a control subject, an RP11 patient and an NPC carrying the c.1205C > A variant, an RP11 patient carrying a deletion of exons 2-8 and an RP11 patient carrying the c.267del variant ( Figure 4A). RPE cells formed pigmented monolayers comprised of RPE65, bestrophin, MITF, CRALBP and ZO1 immunopositive cells (Supplementary Figure S2A-C). Upregulation of the RPE markers RPE65 and BEST1 was demonstrated in cultured iPSC-derived RPE by qPCR (Supplementary Figure S2D).
PRPF31 expression was significantly reduced (p < 0.001) in RPE derived from the three RP11 patients and the NPC, compared with control RPE, with the patient carrying the exon 2-8 deletion showing the lowest levels ( Figure 4B). CNOT3 expression was also significantly reduced in RP11 (p < 0.05-0.001) and NPC RPE (p < 0.01) compared with control RPE ( Figure 4C). These results demonstrate that expression of PRPF31 was reduced both in RP11 patient and NPC iPSC-derived RPE cells, suggesting the 4-copy MSR1 repeat associated with disease non-penetrance leads to increased expressivity of the PRPF31 allele in neural retinal cells, but not in RPE cells.

of 17
RPE65 and BEST1 was demonstrated in cultured iPSC-derived RPE by qPCR (Supplementary Figure S2D).
PRPF31 expression was significantly reduced (p < 0.001) in RPE derived from the three RP11 patients and the NPC, compared with control RPE, with the patient carrying the exon 2-8 deletion showing the lowest levels ( Figure 4B). CNOT3 expression was also significantly reduced in RP11 (p < 0.05-0.001) and NPC RPE (p < 0.01) compared with control RPE ( Figure 4C). These results demonstrate that expression of PRPF31 was reduced both in RP11 patient and NPC iPSC-derived RPE cells, suggesting the 4-copy MSR1 repeat associated with disease non-penetrance leads to increased expressivity of the PRPF31 allele in neural retinal cells, but not in RPE cells.  PRPF31 (B) and CNOT3 (C) expression were normalized to GAPDH and expressed as fold-change compared with values obtained from control RPE. Each bar represents the mean from 3 independent RPE culture wells. Error bars indicate standard deviation. Statistical significance was calculated using a twotailed t-test (* p < 0.05, ** p < 0.01, *** p < 0.001).

Microvilli Densities Are Reduced in RP11 iPSC-Derived RPE
Ultrastructural analysis of iPSC-derived RPE by scanning electron microscopy revealed differences in cell morphology between controls and RP11 patients. Control RPE demonstrated regularly spaced, polygonal morphology with high surface densities of microvilli. In contrast, RPE derived from NPC and RP11 patient iPSC displayed a flattened morphology, with patches of elongated or hypertrophic cells with reduced microvilli densities on their apical surfaces ( Figure 5). . RPE monolayers were differentiated from iPSC derived from a control subject, a non-penetrant carrier (NPC) and RP11 patient carrying the c.1205C > A mutation, an RP11 patient carrying a deletion in exons 2-8 and an RP11 patient carrying the c.267del mutation (A). PRPF31 (B) and CNOT3 (C) expression were normalized to GAPDH and expressed as fold-change compared with values obtained from control RPE. Each bar represents the mean from 3 independent RPE culture wells. Error bars indicate standard deviation. Statistical significance was calculated using a two-tailed t-test (* p < 0.05, ** p < 0.01, *** p < 0.001).

Microvilli Densities Are Reduced in RP11 iPSC-Derived RPE
Ultrastructural analysis of iPSC-derived RPE by scanning electron microscopy revealed differences in cell morphology between controls and RP11 patients. Control RPE demonstrated regularly spaced, polygonal morphology with high surface densities of microvilli. In contrast, RPE derived from NPC and RP11 patient iPSC displayed a flattened morphology, with patches of elongated or hypertrophic cells with reduced microvilli densities on their apical surfaces ( Figure 5).

Primary Cilia Length Is Reduced in RP11 iPSC-Derived RPE
To investigate the effects of reduced PRPF31 expression on ciliogenesis in iPSC-derived RPE cells, RPE monolayers derived from the healthy control subject, the RP11 patient and NPC carrying the c.1205C>A variant, the RP11 patient carrying the c.267del variant and the RP11 patient carrying the exon 2-8 deletion were labeled with an antibody targeting the primary cilium marker ARL13B. More than 98% of the cells from both control and RP11 iPSC-derived RPE monolayers displayed a primary cilium, however RPE derived from both RP11 patients and the NPC showed increased frequencies of short, 'stunted' cilia ( Figure 6A). Quantification of cilia length showed significantly shortened cilia in the RP11 patients compared with RPE from the healthy control (p < 0.001). The mean length of NPC RPE primary cilia was longer than that measured in RP11 patient RPE (i < 0.001), but shorter than primary cilia of healthy control RPE (p < 0.001, Figure 6B).
In healthy control RPE primary cilia lengths were normally distributed around a median value of 2 μm. In contrast, the distributions of cilia lengths were skewed towards shorter lengths in RP11 patient RPE, with two patients showing peak frequencies at 0.5 μm and one patient showing peak frequency at 1 μm ( Figure 6C). These results indicate the majority of RPE cells derived from RP11 patients displayed stunted cilia. Interestingly,

Primary Cilia Length Is Reduced in RP11 iPSC-Derived RPE
To investigate the effects of reduced PRPF31 expression on ciliogenesis in iPSCderived RPE cells, RPE monolayers derived from the healthy control subject, the RP11 patient and NPC carrying the c.1205C > A variant, the RP11 patient carrying the c.267del variant and the RP11 patient carrying the exon 2-8 deletion were labeled with an antibody targeting the primary cilium marker ARL13B. More than 98% of the cells from both control and RP11 iPSC-derived RPE monolayers displayed a primary cilium, however RPE derived from both RP11 patients and the NPC showed increased frequencies of short, 'stunted' cilia ( Figure 6A). Quantification of cilia length showed significantly shortened cilia in the RP11 patients compared with RPE from the healthy control (p < 0.001). The mean length of NPC RPE primary cilia was longer than that measured in RP11 patient RPE (i < 0.001), but shorter than primary cilia of healthy control RPE (p < 0.001, Figure 6B).
Genes 2021, 12, 1542 12 of 17 two peaks were observed in the primary cilia length distribution of the NPC RPE, suggesting two populations of cells with median cilia lengths of 1 μm (stunted) and 2 μm (normal) ( Figure 6C).

Discussion
Since the identification of PRPF31 as the causal gene underlying RP11, multiple hypotheses have been proposed to explain the features associated with this autosomal dominant inherited retinal disease. For example, the retinal specificity of RP11 has been linked to high levels of PRPF31 expression and increased dependence on alternative splicing in retinal cells [32][33][34]. Another feature of RP11 is the variable penetrance of disease among patients carrying heterozygous PRPF31 null alleles. Although inheritance of a single PRPF31 null allele is associated with the development of RP11 in the majority of patients (75-95%, depending on the population studied), most reported RP11 cohorts included individuals who carried the pathogenic mutation but did not develop retinal disease [14,15,17,18,35]. The factors underlying disease non-penetrance in these individuals have been the focus of a number of studies, since their identification could provide novel insights to enhance clinical disease prognostication as well as suggesting new targets for treatment. Given the retinal specificity of PRPF31 deficiency, it is important that potential mechanisms of non-penetrance uncovered in unaffected cell types are validated in the affected retinal cell types. Currently, only one other study examining the effect of PRPF31 mutations in iPSC-derived retinal tissues from RP11 patients has been published [21]. Here, we provide the first report characterizing retinal organoids and RPE cells generated from a non-penetrant carrier. In this study, we examined two proposed determinants of non-penetrance, CNOT3 and MSR1, using fibroblasts and iPSC-derived retinal cells from patients with RP11.
The CNOT3 gene encodes a subunit of the Ccr4-NOT transcriptional complex, a master regulator of global gene expression [19]. Ccr4-NOT complexes regulate gene expression through modulation of transcription by binding promoter elements, as well as through modulation of RNA stability by binding directly to mRNA [36,37]. CNOT3 has been proposed to function as a tumor suppressor gene and was found to be mutated at In healthy control RPE primary cilia lengths were normally distributed around a median value of 2 µm. In contrast, the distributions of cilia lengths were skewed towards shorter lengths in RP11 patient RPE, with two patients showing peak frequencies at 0.5 µm and one patient showing peak frequency at 1 µm ( Figure 6C). These results indicate the majority of RPE cells derived from RP11 patients displayed stunted cilia. Interestingly, two peaks were observed in the primary cilia length distribution of the NPC RPE, suggesting two populations of cells with median cilia lengths of 1 µm (stunted) and 2 µm (normal) ( Figure 6C).

Discussion
Since the identification of PRPF31 as the causal gene underlying RP11, multiple hypotheses have been proposed to explain the features associated with this autosomal dominant inherited retinal disease. For example, the retinal specificity of RP11 has been linked to high levels of PRPF31 expression and increased dependence on alternative splicing in retinal cells [32][33][34]. Another feature of RP11 is the variable penetrance of disease among patients carrying heterozygous PRPF31 null alleles. Although inheritance of a single PRPF31 null allele is associated with the development of RP11 in the majority of patients (75-95%, depending on the population studied), most reported RP11 cohorts included individuals who carried the pathogenic mutation but did not develop retinal disease [14,15,17,18,35]. The factors underlying disease non-penetrance in these individuals have been the focus of a number of studies, since their identification could provide novel insights to enhance clinical disease prognostication as well as suggesting new targets for treatment. Given the retinal specificity of PRPF31 deficiency, it is important that potential mechanisms of non-penetrance uncovered in unaffected cell types are validated in the affected retinal cell types. Currently, only one other study examining the effect of PRPF31 mutations in iPSC-derived retinal tissues from RP11 patients has been published [21]. Here, we provide the first report characterizing retinal organoids and RPE cells generated from a non-penetrant carrier. In this study, we examined two proposed determinants of non-penetrance, CNOT3 and MSR1, using fibroblasts and iPSC-derived retinal cells from patients with RP11.
The CNOT3 gene encodes a subunit of the Ccr4-NOT transcriptional complex, a master regulator of global gene expression [19]. Ccr4-NOT complexes regulate gene expression through modulation of transcription by binding promoter elements, as well as through modulation of RNA stability by binding directly to mRNA [36,37]. CNOT3 has been proposed to function as a tumor suppressor gene and was found to be mutated at high frequencies in T-cell lymphoblastoid leukemia cells [38], while in colorectal and non-small cell lung cancers, CNOT3 overexpression was shown to promote proliferation [39,40]. Knockdown of CNOT3 led to tumor formation in a sensitized Drosophila melanogaster model [38] and inflammatory phenotypes in mice [41,42]. In 2012, Venturini et al. demonstrated an inverse relationship between PRPF31 and CNOT3 expression in LCLs from six RP11 patients and four NPCs, suggesting the higher PRPF31 expression observed in NPC LCLs may result from reduced transcriptional repression by CNOT3-containing Ccr-NOT complexes. Supportively, the authors identified a single nucleotide polymorphism (rs4806718), located within intron 17 of CNOT3, which was significantly associated with disease non-penetrance in 38 subjects from two RP11 pedigrees. Furthermore, CNOT3 was shown to directly bind to the PRPF31 promoter, and inhibition of CNOT3 with siRNA increased PRPF31 expression in cultured ARPE19 cells [18]. While these results suggested CNOT3 might regulate PRPF31 expression in LCLs, the linkage disequilibrium between these two neighboring genes has made it difficult to untangle the effects of the rs4806718 SNP from the possibility that it could be marking a contiguous PRPF31 allele. Indeed, the authors noted the correlation between inheritance of rs4806718 and disease non-penetrance was 'not perfect' and our results in a cohort of 29 subjects from eight families demonstrated the rs4806718 genotypes present in the NPCs were also present in affected RP11 patients, suggesting this SNP does not strongly influence disease penetrance in these families. In this study, we sought to examine whether CNOT3 expression was correlated with disease penetrance and PRPF31 expression in an RP11 family by measuring expression levels in fibroblasts as well as iPSC-derived retinal cells. Our results showed both PRPF31 and CNOT3 expression were reduced in fibroblasts, RO and RPE derived from affected RP11 patients and an unaffected NPC. In support of our observations, Buskin et al. previously demonstrated mis-splicing of CNOT3 mRNA in RP11 patient-iPSC derived RPE [21]. Although expression levels were not quantitated, inefficient splicing of CNOT3 pre-mRNA would reduce levels of mature mRNA due to frameshifting of coding sequences and nonsense mediated decay. Together with the results of Buskin et al. [21], our results suggest PRPF31 haploinsufficiency leads to mis-splicing of CNOT3, resulting in reduced CNOT3 expression in RP11 patient cells.
In addition to CNOT3 mis-splicing, Buskin et al. identified genes associated with the spliceosome as one of the most common subsets affected by PRPF31 haploinsufficiency in RP11 patient derived RO and RPE [21]. Supportively, knockdown of PRPF31 expression in organotypic adult human retinal cultures was recently shown to cause mis-splicing of transcripts encoding the spliceosome subunits PRPF3, PRPF8, PRPF4 and PRPF19 [22]. Here, we observed decreased expression of PRPF6, PRPF8 and SNRNP200 in RP11 patientderived RO compared with RO derived from a healthy control, suggesting splicing defects initially caused by PRPF31 deficiency may be further amplified through reductions in expression of other components of the splicing machinery.
Following on from their previous work examining CNOT3 as a potential modifier of RP11 disease penetrance, Rose et al. published strong evidence indicating that a significant proportion of non-penetrance in RP11 families could be attributed to the inheritance of a highly expressive, unaffected PRPF31 allele [20]. In that study, the authors identified an MSR1 repeat element within the PRPF31 promoter, and further showed that a 4-copy repeat element was associated with higher expression of PRPF31 in RP11 patient LCLs, while inheritance of a 3-copy repeat led to lower expression. In a cohort of 42 affected RP11 patients and 29 unaffected NPCs, 4-copy repeats were not found in any affected RP11 patients, but were present in almost one third of NPCs. The frequencies of the 4-copy MSR1 variant of the PRPF31 allele within different ethnic populations was strongly correlated with the frequencies of non-penetrance measured in RP11 cohorts derived from these populations, suggesting the MSR1 repeat element is a major determinant of nonpenetrance in RP11 [20,43]. In support of these previous observations, we found that a 4-copy MSR1 repeat element was present in 3/5 unaffected NPCs from different RP11 pedigrees, but was not found in any affected RP11 patients. Consistent with previous observations in LCLs, RO derived from an NPC carrying the 4-copy MSR1 repeat showed an intermediate level of PRPF31 expression between RO derived from one of her affected sons and control RO levels. Interestingly, inheritance of the 4-copy MSR1 repeat element did not correlate with significant differences in PRPF31 expression in RPE cells derived from the NPC, suggesting the effects of this promoter element could be specific to some cell types, such as neural retinal cells and LCLs. However, further studies with larger numbers of NPC patients will be required to determine the relationship between MSR1 genotype and expression of the unaffected PRPF31 allele in NPC retinal cells.
In dermal fibroblasts, reductions in PRPF31 expression were more pronounced in cells from adult RP11 patients, while cells from teenaged patients had higher levels of PRPF31 mRNA. Additionally, we demonstrated PRPF31 expression in the adult human retina significantly decreases with age. Age-related declines in PRPF31 expression highlight the need for age-matched control subjects in studies performed in patient-derived primary cell cultures. Notably, the NPC studied here was the eldest subject (54yo) in the family and expressed 65% of the PRPF31 mRNA levels measured in fibroblasts derived from the closest age-matched control (46yo), and almost double the levels of the closest age-matched RP11 patient (51yo) from the same family. These results are consistent with the modest increases in PRPF31 expression previously reported in NPC LCLs compared with LCLs from affected RP11 patients [18] and suggest that the inheritance of the 4-copy MSR1 repeat element is associated with increased PRPF31 expression in NPC fibroblasts, compared with an age-matched RP11 patient carrying homozygous 3-copy repeats.
Although dermal fibroblasts and LCLs represent accessible cell models for examining the expression of ubiquitously expressed genes such as PRPF31, their use in the identification of disease mechanisms is limited by a lack of a retinal cell phenotype. Inz'this study, we confirmed the ciliopathy phenotype previously reported by Buskin et al. in RP11 patient derived RPE cells [21] and further showed the variable penetrance of this phenotype in RPE cells derived from the NPC. Compared with RPE derived from a control subject, RPE cells derived from the unaffected NPC showed similar reductions in PRPF31 expression levels as RPE cells derived from an affected RP11 patient, however, cilia length distributions in NPC RPE revealed the presence of two populations of RPE cells. Although the majority of cells were affected by a 'stunted' cilia phenotype (62% < 1.5 µm), a sizable subpopulation of cells displayed normal length cilia (38% > 1.5 µm). It is possible that these two populations represent RPE cells below and above a critical threshold of PRPF31 expression. Ultrastructural characterization by SEM supported this conclusion, with patches of flattened, elongated RPE cells interspersed with unaffected cells present in NPC RPE monolayers. It is difficult to predict how defective ciliogenesis in RPE cells in vitro might reflect disease processes in vivo, since RP11 patients are born with functional RPE. It is likely that in the absence of paracrine support from the vasculature, resident immune cells and neural retina, cultured RPE cells are more prone to disturbances in cellular homeostasis than their in vivo counterparts. However, despite these potential limitations, the stunted cilia phenotype in RP11 patient iPSC-derived RPE could provide a useful measure for screening of pharmacological agents with potential for treating RP11 through increasing PRPF31 expression in RPE cells.

Conclusions
Together, our results demonstrate reduced PRPF31 expression in retinal cells derived from RP11 patients was associated with reduced CNOT3 expression and that CNOT3 polymorphism did not correlate with non-penetrance. We showed that inheritance of the 4-copy variant of the MSR1 repeat element was associated with disease non-penetrance. We further demonstrate a 'stunted' cilia phenotype in RPE monolayers derived from RP11 patients, with variable penetrance in the NPC. This variable penetrance at the cellular level suggests NPC-derived RPE cells are balanced on the edge of PRPF31 insufficiency, further suggesting that small increases in PRPF31 expression may be sufficient to restore splicing defects in haploinsufficient retinal cells.  Informed Consent Statement: Informed consent was obtained from all subjects involved in this study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.