Bi-allelic loss-of-function variants in POC5 cause a syndromic retinal, endocrine, and neuromuscular ciliopathy

Abstract

Purpose: A homozygous loss-of-function (LoF) variant in POC5 was previously described in an individual with retinitis pigmentosa. We identified POC5 variants in 12 probands with a syndromic phenotype. We aim to define the phenotype spectrum and molecular mechanism associated with biallelic POC5 LoF variants.

Methods: We studied a cohort of 12 families with bi-allelic LoF POC5 variants and performed detailed phenotype analysis. POC5 localization studies were performed in 3 proband-derived fibroblast cell lines.

Results: Detailed phenotyping of probands with POC5 variants expands the phenotype spectrum beyond ocular manifestations. This syndrome causes not only rod-cone dystrophy but also diabetes mellitus with severe insulin resistance and partial lipodystrophy, kidney disease, and muscle cramps. The POC5 protein plays an essential role during cell cycle and cilium formation. Interestingly, POC5 localization studies in 3 proband-derived fibroblast cell lines show aberrant localization suggesting a ciliary defect. The phenotypes of the 12 families in this study fit well within the ciliopathy phenotype spectrum, except for lipodystrophy, which is not common in ciliopathies.

Conclusion: We describe a multiorgan syndrome caused by bi-allelic LoF variants in POC5. This underscores the pleiotropic effects of POC5 variants and highlights the significance of adipose tissue and metabolic dysfunction in ciliopathies.

Keywords: Ciliopathy; Insulin resistance; Muscle cramps; Renal disease; Syndrome.

Figure 1. Twelve families with POC5 LoF variants.

A. Pedigrees of the 12 families. Shaded symbols indicate the presence of various phenotype features in probands and their family members. Results of genetic analysis are indicated (− = POC5 variant). B. Schematic representation of the POC5 protein structure with its predicted functional domains. The white lines in the bar indicate the exon-exon junctions. The variants above the bar in black are LoF variants described in this study, and variants below the bar in gray have previously been associated with adolescent idiopathic scoliosis.

Figure 2. RNA and protein analysis of POC5-affected participants.

A. Targeted RNA analysis of POC5 in fibroblasts treated with (+) or without (−) CHX. B. Quantitative RNA analysis of POC5 in fibroblasts treated with (+) or without (−) CHX to inhibit NMD. Bars represent the mean with SD of POC5 expression relative to the controls (data from 2 or 3 replicates per fibroblast line). Statistical significance was calculated using analysis of variance with post-hoc multiple comparison using Dunnett’s test (* is P < .05). C and D. Reduced expression of POC5 transcript caused by homozygous variant c.686A>G in POC5 that leads to a 4-bp deletion (r.686_689del) in the POC5 transcript. C. Volcano plot of P3 showing expression outlier genes in orange with z-score >3 and P value < .01. POC5 has a P value of 5.9e⁻⁰⁷ and a z-score of −5.60. D. Gene ranking plot for POC5 showing P3 in red with a normalized count of 250 and unrelated control samples in gray with a mean of 544. E. Schematic visualization of the splice effect caused by variant c.686A>G seen in P3. CHX, cycloheximide; NMD, nonsense-mediated messenger RNA (mRNA) decay.

Figure 3. Representative ophthalmologic images of POC5 participants.

(1) Ophthalmologic imaging of RE of P10 (22-year-old female) with (A) wide-field fundus photograph showing peripheral bone-spicule hyperpigmentation, (B) wide-field green-light fundus autofluorescence image showing a clear hyperautofluorescent ring surrounding the central retina and, further peripherally, a ring of hypo-autofluorescent spots, (C) optical coherence tomography (OCT) image with preserved ellipsoid zone. (2) Ophthalmologic imaging of LE of P11 (54-year-old female) with (A) fundus photograph showing a waxy pallor visible on the optic nerve head, (B) blue-light fundus autofluorescence image is essentially normal centrally and shows a hypo-autofluorescent area inferonasally to the arcade, (C) subfoveal disruption of ellipsoid zone on OCT. (3) and (4) present P8 RE and P5 LE, respectively, with a similar ocular phenotype as P10. LE, left eye; RE, right eye.

Figure 4. Photographs of participants.

A. Acanthosis nigricans in neck of P3, acanthosis nigricans and striae in axilla of P3, and acanthosis nigricans in neck of P8. B. Full body photograph of P3 (26-year-old female) showing slim build with decreased subcutaneous fat and subtle central adiposity, full body photograph of P4 (46-year-old male) showing muscular appearance with decrease of subcutaneous fat, and arm and leg of P6 (48-year-old female) showing asymmetric fat distribution in her thighs and reduced subcutaneous fat in the upper arms. C. Frontal facial photographs of P4, P7, and P12 and lateral of P4, P3, and P12 showing variable dysmorphic features, including wide nasal base with thick nasal alae and a low-hanging columella, mandibular prognathia, and large ears. D. Photographs of hands of P3, P4, and P7 showing brachydactyly and loss of subcutaneous fat in P4.

Figure 5. Immunofluorescent imaging of POC5, centrin and cilium phenotype markers in fibroblasts from controls and P3, P8, P9, and P10.

A. Centriolar localization of POC5 in ciliated cells. The images are representative of the analysis based on 75 cells per cell line that was performed in triplicate. Centrioles and the proximal end of the cilium are visualized with GT335 (purple), the ciliary axoneme is stained with anti-acetylated-α-tubulin (red) and POC5 localization is visualized in green. The mother centriole or basal body is indicated with a white arrowhead and the scale bar represents 1 μm. B. Centriolar localization of centrin in ciliated cells. The images are representative of a n = 2 analysis based on 75 cells per cell line. Centrin is visualized in green, and pericentrin is visualized in red. C and D. No substantial differences were observed in cilium length (C) or ciliogenesis (D) between controls and affected participants. The cilium length was based on the combined signal of acetylated-α-tubulin and ARL13B and was measured in an automated manner as described by Doornbos et al. Cilia from affected P3 were only mildly shorter compared with controls (combined data from 2 control cell lines) (ie, 3.8 ± 0.2 and 4.19 ± 0.17 μm, respectively). Statistical significance was calculated using analysis of variance with post-hoc multiple comparison using Dunnett’s test (**** is P < .0001). E. Quantification of SHH assay showing the relative fluorescent intensity signal of SMO over GPR161 in the cilium. Fibroblasts were stimulated with SAG or DMSO as a negative control. F. Representative cilia images upon 48 hours of serum starvation with or without SAG stimulation. Panel 1 was used for ciliogenesis and cilium length measurements and panel 2 and 3 for the SHH assay. Staining for panel 1 includes anti-ARL13B (green), anti-acetylated-α-tubulin (red), and anti-PCNT (purple). Staining for panels 2 and 3 includes anti-SMO (green) and anti-GRP161 (red). Scale bar represents 1 μm. DMSO, dimethyl sulfoxide; PCNT, pericentrin; SAG, smoothened agonist; SHH, sonic hedgehog; SMO, smoothened.