Disrupted alternative splicing for genes implicated in splicing and ciliogenesis causes PRPF31 retinitis pigmentosa

Abstract

Mutations in pre-mRNA processing factors (PRPFs) cause autosomal-dominant retinitis pigmentosa (RP), but it is unclear why mutations in ubiquitously expressed genes cause non-syndromic retinal disease. Here, we generate transcriptome profiles from RP11 (PRPF31-mutated) patient-derived retinal organoids and retinal pigment epithelium (RPE), as well as Prpf31+/- mouse tissues, which revealed that disrupted alternative splicing occurred for specific splicing programmes. Mis-splicing of genes encoding pre-mRNA splicing proteins was limited to patient-specific retinal cells and Prpf31+/- mouse retinae and RPE. Mis-splicing of genes implicated in ciliogenesis and cellular adhesion was associated with severe RPE defects that include disrupted apical - basal polarity, reduced trans-epithelial resistance and phagocytic capacity, and decreased cilia length and incidence. Disrupted cilia morphology also occurred in patient-derived photoreceptors, associated with progressive degeneration and cellular stress. In situ gene editing of a pathogenic mutation rescued protein expression and key cellular phenotypes in RPE and photoreceptors, providing proof of concept for future therapeutic strategies.

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1

Characterisation of RP11 - RPE cells revealed polarity and functional defects. a Schematic of RPE differentiation timeline; b Bright-field images of iPSC-derived RPE: representative examples from at least ten independent experiments, scale bar 100 μm; c Immunostaining for basolateral markers BEST1 and Na+/K+-ATPase: representative images from three independent experiments, scale bar 50 μm; d Correct basolateral distribution of collagen IV (C-IV) and apical MERTK in unaffected control (WT3) but not RP11 RPE cells: representative images from three independent experiments, scale bar 50 μm; e, f ELISA assays for apical and basal secretion of PEDF and VEGF, respectively, in control and RP11 - RPE cells; g Trans-epithelial resistance measurements revealed a significant difference between patient and RP11 - RPE cells; h Reduced phagocytic capacity in RP11 - RPE cells. Statistical significance is calculated for MFI (mean fluorescence intensity) values. e–h Data shown as mean ± SEM, n = 3. Statistical significance of pairwise comparisons is indicated by n.s.: not significant; ***p < 0.001; ****p < 0.0001 (Student’s paired t test). b–h Data obtained from RPE at week 21 of differentiation

Fig. 2

Generation of retinal organoids following long-term suspension culture. a Schematic representation of iPSC differentiation to retinal organoids; b Bright-field images showing development of retinal neuroepithelium over time, scale bar 50 μm; c Immunostaining of retinal organoids showing the expression of cell-specific markers; b–c representative examples from iPSC-derived retinal organoids from RP11S2 patient are shown, scale bar 25 μm apart from ARL13B, where scale bar is 10 μm; d TEM revealed the presence of outer limiting-like membrane (white arrows), inner segments (IS), connecting cilia (CC) and developing outer segments (OS) in retinal organoids after 21 weeks in culture, top panel scale bars: 10 μm, 500 nm, 500 nm, 10 μm, bottom panel scale bars: 5 μm, 2 μm, 500 nm, 500 nm; e At 43 weeks in culture, TEM showed that patient photoreceptors contained apoptotic nuclei with electron dense structures of condensed chromatin (white arrow) and stress vacuoles (black stars). d, e Representative examples of three independent experiments

Fig. 3

PRPF31 expression in patient-specific cells and effects on pre-mRNA splicing. a Gel electrophoresis showing the presence of a long mutant transcript (LM) isoform for the exon 11 deletion in patient-specific cells. The short mutant (SM) isoform is present only upon inhibition of NMD with puromycin (indicated by + ); b The bar graph shows wild-type PRPF31 mRNA in patient cells relative to controls from a, b. Data are representative of at least three independent repeats, RO retinal organoids; c Wild-type PRPF31 is significantly reduced in patient RPE cells and less notably in retinal organoids. The LM form and reduced SART1 is observed only in the patient RPE cells; d The bar graph shows wild-type PRPF31 levels in patient cells relative to normal cells quantified from c, n = 3; e, f Patient RPE cells and retinal organoids exhibit a notable defect in the alternative splicing of E1A minigene reporter. Schematic representation of alternative splice variants of the E1A reporter (e) and denaturing PAGE and autoradiography using a phosphoimager (f), n = 3; g Northern blot analysis showing the level of snRNAs in various normal and patient cells. Total RNA was isolated from each sample and snRNA levels were analysed by denaturing PAGE followed by Northern blotting using probes against U1, U2, U4, U5, U6 and 5S rRNA (top). The levels of snRNAs were quantified and normalised to the amount of 5S rRNA (bottom), n = 2. All error bars represent SEM

Fig. 4

RNA-seq analysis of alternative splicing in fibroblasts, iPSC, RPE, retinal organoids and Prpf31+/− retina. a rMATS analysis showing that RP11 - RPE have the highest percentage of transcripts containing retained introns (RI) and alternative 3′ splice sites (A3SS); b Gene Ontology enrichment analysis showing biological and cellular processes affected by alternative splicing, respectively, in human cells; c Gel electrophoresis of RT-PCR for the indicated genes in RPE and retinal organoids derived from patient RP11VS and unaffected control WT3. Sizes (in bp) for major and minor isoforms (arrowheads), and percentage-spliced-in (PSI) values, are indicated; d Sashimi plots for the indicated genes for validation of alternative splicing events in RPE and retinal organoids derived from RP11 patients (blue) and unaffected controls (red). Data are representative of at least three independent experiments. Green highlights in Sashimi plots indicate alternative splicing events with the number of junction reads indicated for each event; e, f Gene Ontology enrichment analysis showing biological and cellular processes affected by alternative splicing, respectively, in mouse Prpf31+/− retinae and RPE. Data are representative of at least three independent experiments

Fig. 5

RP11 - RPE cells and photoreceptors have defective ciliogenesis and cilia morphology. a Immunostaining of RPE with cilia markers ARL13B (green) and pericentrin (red), with representative images shown from n = 3 independent experiments, scale bar 10 μm; b Quantification of cilia length and incidence showing significant reduction across both parameters in RP11 patients compared to the controls; c, d 2D TEM and 3D SBFSEM images showing shorter cilia in RP11 - RPE cells, with abnormal bulbous morphology, with representative images shown from n = 3 independent experiments, scale bar 500 nm (c), 1 μm (d); e Immunostaining of photoreceptors with cilia marker ARL13B (red), with representative images shown from n = 3 independent experiments; f Quantification of cilia length and frequency in photoreceptors showing significant reduction in RP11 patients compared to the controls; g, h 2D TEM and 3D SFBSEM images showing shorter cilia in patient-derived photoreceptors, with abnormal bulbous morphology, with representative images shown from n = 3 independent experiments, scale bar 500 nm (g), 1 μm (h). b, d, f, h Data shown as mean ± SEM, n = 3. Statistical significance of the indicated comparisons is indicated by n.s. not significant; ***p < 0.001; ****p < 0.0001 (one-way ANOVA test with Dunnett’s post hoc test correction for multiple testing)

Fig. 6

PRPF31 loss causes defects in cilia incidence and structural organisation. aPRPF31 siRNA knockdown in human hTERT-RPE1 cells causes a significant decrease in cilia incidence (lower left) and PRPF31 protein levels (lower right) compared to scrambled negative control (siScr) siRNA, scale bar: 10 μm; b Gli1 reporter assays of Shh activity measured in NIH3T3-GL cells following knockdown for Ptch1 (positive control), scrambled negative control siRNA (siScr) and Prpf31. Cells were treated with either 100 nM SAG or vehicle control for 48 h, as indicated. Assays results are expressed in arbitrary units of the ratio of firefly: Renilla luciferase activities; c Ciliary localisation of IFT88 (green) in primary cilia of hTERT-RPE1 cells (visualised by staining for γ-tubulin and poly-glutamylated tubulin; red) showing mislocalisation of IFT88 (arrowheads) at ciliary tips following PRPF31 knockdown. Bar graph quantitates the percentage of cilia with IFT88 at their tip. Scale bar: 1 μm; d Visualisation and quantitative analysis of the transition zone protein CC2D2A (green) and ARL13B (red); e Visualisation and quantitative analysis of the transition zone (TZ) protein RPGRIP1L (green) and cilia (γ-tubulin and poly-glutamylated tubulin; red) showing mislocalisation of RPGRIP1L from the TZ into the ciliary axoneme (arrowheads) following PRPF31 knockdown. f, g Ciliary localisation of IFT88 and RPGRIP1L (green) in RP11 - RPE cells showing mislocalisation of IFT88 (arrowheads) at ciliary tips and RPGRIP1L from the TZ into the ciliary axoneme (arrowheads). a–g Data shown as mean ± SEM, n = 3. Statistical significance of pairwise comparisons is indicated by n.s. not significant; *p < 0.05; **p < 0.01; ***p < 0.001 (Student’s unpaired t test). c–f Scale bar: 1 μm

Fig. 7

Gene correction of PRPF31 mutation results in reversal of cellular and functional phenotypes in RPE and photoreceptors. a–c CRISPR/Cas9 correction of the PRPF31 deletion in exon 11; d, e Quantification of cilia length and incidence in PRPF31- and WT-RPE. f TEM analysis of PRPF31-edited RPE cilia showing morphologically normal cilia, scale bar: 500 nm; g Increased phagocytosis in PRPF31-edited RPE. h–j Restoration of apical–basal polarity in PRPF31-edited RPE, scale bar: 50 μm. k, l Quantification of cilia length and frequency in PRPF31- and WT-photoreceptors; m TEM analysis of PRPF31-edited photoreceptor cilia showing morphologically normal cilia, scale bar: 500 nm. c–e, g–i, k, l Data shown as mean ± SEM, n = 3. Statistical significance of pairwise comparisons is indicated by n.s.: not significant; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 (Student’s paired t test)

Fig. 8

Schematic diagram that summarises pathogenic pathways in the RP11 disease state. Altered pre-mRNA splicing in iPSC-derived RPE and retinal organoids is associated with severe RPE defects and disrupted cilia in photoreceptors