Loss of CBY1 results in a ciliopathy characterized by features of Joubert syndrome

Abstract Ciliopathies are clinically and genetically heterogeneous diseases. We studied three patients from two independent families presenting with features of Joubert syndrome: abnormal breathing pattern during infancy, developmental delay/intellectual disability, cerebellar ataxia, molar tooth sign on magnetic resonance imaging scans, and polydactyly. We identified biallelic loss‐of‐function (LOF) variants in CBY1, segregating with the clinical features of Joubert syndrome in the families. CBY1 localizes to the distal end of the mother centriole, contributing to the formation and function of cilia. In accordance with the clinical and mutational findings in the affected individuals, we demonstrated that depletion of Cby1 in zebrafish causes ciliopathy‐related phenotypes. Levels of CBY1 transcript were found reduced in the patients compared with controls, suggesting degradation of the mutated transcript through nonsense‐mediated messenger RNA decay. Accordingly, we could detect CBY1 protein in fibroblasts from controls, but not from patients by immunofluorescence. Furthermore, we observed reduced ability to ciliate, increased ciliary length, and reduced levels of the ciliary proteins AHI1 and ARL13B in patient fibroblasts. Our data show that CBY1 LOF‐variants cause a ciliopathy with features of Joubert syndrome.


| INTRODUCTION
Cilia are structurally and functionally complex organelles that reside on the surface of most cells in the vertebrate body plan.
Dysfunctional cilia cause a range of disorders in humans, called ciliopathies, which are clinically and genetically heterogeneous with a wide spectrum of phenotypes (Reiter & Leroux, 2017).
Cilia can be divided into two main types: primary cilia that are important signaling hubs and play a role in embryonic life during organogenesis and motile cilia that exert their action in epithelial cells, for example, in the lining of the respiratory tract. Joubert syndrome (JBTS; MIM #213300) is an early onset primary ciliopathy, accompanied by the characteristic mid-and hindbrain malformations, including the cerebellar anomaly designated as the "molar tooth sign", seen in the axial plane on cerebral magnetic resonance imaging (MRI; Romani et al., 2013).
Individuals with JBTS present with infantile breathing abnormalities (tachypnea and/or apnea), oculomotor apraxia, cerebellar ataxia, hypotonia, and psychomotor delay (Parisi, 2009). Multiple organ involvement and polydactyly is not unusual. Defective motile cilia, however, as in primary ciliary dyskinesia (PCD; MIM #244400), results in decreased clearance of respiratory tract secretions and is clinically an important differential diagnosis to cystic fibrosis. Individuals with PCD demonstrate chronic otosinopulmonary disease due to mucus retention, and also display other features, which may include heterotaxia and infertility (Zariwala et al., 1993). Despite major ultrastructural similarities between motile and primary cilia, concomitant dysfunction of both organellar subtypes is rarely documented (Horani & Ferkol, 2016;Mitchison & Valente, 2017). Here, we describe three individuals from two unrelated families with biallelic loss-of-function (LOF) variants in CBY1, presenting with features of JBTS.

| Ethical considerations
The ethics committees of all relevant participating institutions approved this study. Participant family members signed informed consent for genetic studies and for publication of clinical data.
The affected individuals reported in this study were connected through the data-sharing platform GeneMatcher (Sobreira et al., 2015).
The experiments performed on zebrafish were conducted in accordance with relevant institutional (Regierungspräsidium, Freiburg, Germany) and national guidelines and regulations for animal care and use.

| Clinical presentation of the patients
In Family A (FA), healthy, first-cousin parents from Pakistan had a total of seven pregnancies and four living children (Figure 1a). The two elder siblings FA.II-1 and FA.II-2 had a clinical diagnosis of JBTS (Table 1). In addition, they displayed organ manifestations likely not attributable to JBTS (Table S1). FA.II-1 was born at 36 weeks of gestation with birth weight of 2040 g (1 kg < 2.5 percentile). The occipitofrontal head circumference (OFC) at birth was 33 cm (2.5th percentile) and at 5 weeks 34 cm (1 cm < 2.5th percentile); and later OFC measurements stayed at 1 cm < 2.5th percentile. During the first year delayed eye contact and oculomotor apraxia became evident, together with recurrent episodes of tachypnea. Muscular hypotonia was noted from an early age. He walked at 2 years with a wide-based atactic gait. Cerebral MRI examinations at 1 and 8 years showed elongated and horizontally oriented superior cerebellar peduncles, vermis hypoplasia, and superior cerebellar foliar dysplasia 2180 | ( Figure 1b). He was operated for bilateral calcaneovalgus and pes planus. He was also noted to have agenesis of the lower lateral incisors, small teeth, and fifth finger clinodactyly. At 4 years and 9 months general development was delayed with language skills corresponding to 2.5 years. Speech was impaired due to dysarthria.
At 18 years, formal cognitive evaluation performed with the Wechsler Adult Intelligence Scale version IV Full Scale Intelligent Quotient (WAIS-IV FSIQ) concluded with IQ 64 (confidence interval, F I G U R E 1 Pedigree of Families A and B and cerebral MRI findings. (a) Pedigree of Family A showing the two affected siblings (black symbols) and carriers (symbols with the dot) of the CBY1 variant Chr22:g.39067079_39067080del within the family. The first two fetuses were not genotyped. (b−f) Cerebral magnetic resonance imaging examinations in Family A (FA.II-1, FA.II-2, FA.II-3), and Family B (FB.II-2). In the parasagittal views (b,c,f), the superior cerebellar peduncles (arrowheads) are more horizontally oriented, as opposed to the normal and more vertically oriented peduncle in (d). In the axial midbrain views (bʹ,cʹ,fʹ), cerebellar vermian foliar dysplasia is seen above the arrows. This anomaly is not present in (dʹ). The axial pons views (b″,c″,f″) show elongation of the superior cerebellar peduncles giving a mild "molar tooth" appearance, but not in (d″). Sagittal views (b-d) (Table S1), for example, he was operated for late infantile bilateral cataracts between 3 and 4 years, and he needed surgical correction for nasal septum deviation. At 3 years, he was diagnosed with celiac disease and anemia. From 13 years, he suffered subacute episodes of fatigue accompanied by poor appetite and pain in the thoracic cage, neck, and extremities. In parallel with this, serum creatine kinase (CK) concentrations were constantly elevated ranging from 350 to 1200 U/L (reference < 200 U/L) and plasma myoglobin was also elevated with a concentration of 500 ng/ml (reference, 0-85 ng/ml). Neurography and electromyography of upper and lower extremities at 16 and 19 years indicated combined myopathy and axonal (sensory and motor) neuropathy. A precise interpretation was difficult, but myopathy was considered the predominant pathological component. MRI examination of the lower extremities showed infiltration of fat in the muscles and biopsy from his right mid-gastrocnemius confirmed deposition of fat and connective tissue with extensive myopathic and neurogenic changes. At 21 years, he suffered episodes of nontraumatic nail bed bleedings of unknown etiology causing swelling and pain. At the age of 21 years, he was also diagnosed with hypogonadotropic hypogonadism with the following serum hormone concentrations: testosterone, 7.9 mmol/L (reference, 7.2-24 mmol/L); follicle-stimulating hormone, 18.4 IU/L (reference, 0.7-11.1 IU/L); luteinizing hormone, 12.5 IU/L (reference, 0.8-7.6 IU/L). On ultrasound examination, testicular volume was estimated to 3-4 ml for each, corresponding to pre-pubertal size.
FA.II-2 was born at term with left hand postaxial hexadactyly and fifth finger clinodactyly and normal birth measurements. Episodes of tachypnea were noted from infancy to early childhood and muscular hypotonia was present at an early age. She was referred for ophthalmological examination due to poor eye contact and peculiar thrusting of the head to one side with the eyes deviating to the opposite side, and was later diagnosed with saccadic initiation failure, also known as oculomotor apraxia. Cerebral MRI examination at 7 months, repeated at 6 and 18 years, revealed similar cerebellar abnormalities as in FA.II-1, compatible with JBTS ( Figure 1c). Delayed psychomotor development became evident from an early age, but with extra support she was able to attend regular school. Assessment Later, dyspnea and reduced working capacity became increasingly evident and lung function tests were pathological. She also suffered from a number of abnormalities and diseases without obvious connection to the JBTS, for example, nasal septum deviation, which was corrected by surgery (Table S1). Over the last 8 years, she needed more than 30 ophthalmological visits due to recurrent keratoconjunctivitis, a condition of uncertain etiology, which was treated with topical steroid ointment and artificial tears. Notably, keratoconjunctivitis (or blepharokeratoconjunctivitis) is reported to be more frequent in the Indian-Pakistani population (Viswalingam et al., 2005). She had urinary incontinence and recurrent urinary tract infections including pyelonephritis. No osmotic disturbance was FA.II-3, a female patient, was noted from an early age to have microcephaly and developmental delay. She also had bilateral cataracts (Table S1). Due to JBTS in the two elder siblings her development was followed closely. However, she was not noted to exhibit signs of oculomotor apraxia or ataxia and cerebral MRI was normal (Figure 1d). At age 12 years, she was assessed with WISC-V showing full scale IQ 69.
At 13.5 years, lack of secondary sexual characteristics lead to pelvic MRI and ultrasound examinations disclosing the absence of uterus and ovaries, and endocrinological assessment indicated primary ovarian insufficiency. She had normal feminine karyotype analysis and wholeexome sequencing (WES) did not reveal relevant abnormalities. In view of neuromuscular features in FA.II-1, CK was measured on several occasions, but was concluded to be normal (Table S1).
FA.II-4, a female, had microcephaly, bilateral cataract (operated) and nasal septum deviation (awaiting operation). At the age of 9 years, she started to complain of muscular pain and weakness. CK  Table S1.
The clinical picture of FA.II-1 and FA.II-4 were similar in regard to muscular pain and weakness, and CK elevation and the same neuromuscular disease. For comparison of features of hypogonadism, see Table S1.
In addition to her live children, mother FA.I-2 had two first trimester miscarriages; prenatal ultrasound scanning and autopsy were not performed. The last pregnancy of FA.I-2 concluded with an in utero death at 15 weeks (FA.II-5 in Figure 1a). On ultrasound scan, this fetus presented with dolichocephaly, scaphocephaly, and a thick nuchal fold. Fetal DNA of FA.II-5 tested negative for chromosomes 13, 15, 16, 18, 21, X, and Y aneuploidies by quantitative fluorescence polymerase chain reaction technique.
The parents in Family B (FB; Figure 1e) were first cousins of German origin. The mother reported a total of six first trimester miscarriages. Karyotyping was performed in one of these fetuses and yielded a normal male result (46,XY). Five years after the birth of a healthy daughter, the Family B proband was born; she was 10 years old at last clinical exam. She was born at 38 weeks of gestation by cesarean section and birth measurements were within the normal range. Postaxial hexadactyly was present on all four extremities ( Figure S1). Hypopigmented areas were irregularly scattered over her entire body ( Figure S1). Abnormal eye movements and psychomotor delay were noted from an early age. The muscle tone was reduced, and consequently, she wore splints during the day. Both eyes showed normal findings on funduscopy without evidence of retinal dystrophy and electroretinography was normal. Orthoptic examination revealed symptoms of Cogan II syndrome/oculomotor apraxia (MIM #257550).
Horizontal saccades were abnormal, while vertical saccades showed no deficits. She used head thrusts to compensate for the inability to accomplish voluntary horizontal saccades. Oculomotor apraxia became gradually less severe during childhood. Diadochokinesis was shown to be delayed, indicating cerebellar dysfunction. Axial cerebral MRI examination at 5 years showed horizontally oriented superior cerebellar peduncles, vermis hypoplasia, and superior cerebellar foliar dysplasia ( Figure 1f). Upper respiratory problems were present from an early age.
She had a constant running nose and recurrent otitis media, requiring bilateral tympanostomy tubes at 7 years. At 8 years, nNO concentrations were found in the normal range. At age 11.5 years, height was 154.5 cm (75-90th percentile) and weight 58.6 kg (97th percentile), and OFC was 55 cm (90th percentile). She had learning difficulties and attends a school for children with special needs. The clinical diagnosis of FB.II-2 was JBTS. The variant call files generated were analyzed using the Filtus program (Vigeland et al., 2016). We discarded variants with allelic frequency >0.01 in any of the databases used (gnomAD, ExAC,

Zebrafish microinjections
For synthesis of sense RNA, we cloned full-length zebrafish cby1 into pCS2+. Sense RNA was prepared from ApaI-linearized cby1-pCS2+ using the SP6 mMessage mMachine Transcription Kit (Thermo Fisher Scientific). Morpholinos (MOs) and sense RNA were diluted in 0.1-M KCl to concentrations of 4 and 0.02 μg/μl, respectively. One nanoliter of this dilution was injected through the chorion of 1-cell stage embryos. To attenuate possible off target effects, a p53 MO (Robu et al., 2007) was coinjected 1.5-fold to the other MOs used. The following translation/splicing-blocking (TB/SB) antisense MOs (Gene Tools; www.gene-tools.com/morpholino; antisense oligos) were used: (ACAGAAACGTGTACTTACTGTATGT), and a standard control We used the online tool (http://crispr.mit.edu/) to design efficient guide RNAs (gRNA) targeting genomic cby1 in zebrafish. We used a gRNA targeting the sequence 5ʹ-AGCGATGCAGACTTTCGAGGTGG-3ʹ (protospacer adjacent motif sequence underlined) in exon2 of cby1. We Whole mount in situ hybridization (ISH) analysis and immunostaining ISH was performed as previously described (Lu et al., 2017).

Immunoblotting
For each condition, 30 dechorionated zebrafish embryos at 1 dpf were deyolked by incubation in calcium-free Ringer's solution

Microscopy and image acquisition
Brightfield images of whole mount in situ embryo stains were taken using an Axioplan2 microscope with Axiocam camera and using Ax-

Statistical analysis and quantification
All data represent results from one of at least three independent experiments, which showed consistent results. Numbers of embryos used for analysis are indicated in the respective bar chart. Data were analyzed by Student's t test (two-sided, unpaired); error bars represent the standard error of the mean (SEM). Immunoblot signals were quantified using Gel-Pro Analyzer 6.0, INTAS and normalized to respective loading controls. Statistical analysis for qPCR results was performed using the GraphPad Prism software and significance was calculated with the one-sample t test.

| Studies on fibroblast cultures
Cell culture Skin fibroblasts obtained from individual FA.II-1, FA.II-2, and FB.II-2 and from three controls were cultivated in DMEM (Life Technologies) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin.

Immunofluorescence (IF) analysis
Fibroblasts between passages 4 and 9 were seeded in six-well plates (BD Bioscience), 60,000 cells in each well, allowing growth on sterilized glass cover slips (  docking of the basal bodies to the plasma membrane (Burke et al., 2014;Voronina et al., 2009). CBY1 also antagonizes the transcriptional coactivator beta-catenin, a key mediator of the canonical Wnt signaling pathway (Takemaru et al., 2003). CBY1 was, therefore, considered a putative disease-causing gene in the two families, who shared clinical features of a ciliopathy.
As the hypogonadism and the neuromuscular disease observed in Family A did not segregate with the CBY1 variant and cannot be easily ascribed to a ciliopathy, we reanalyzed the WES data available in search for additional disease-causing genes. However, we did not identify additional relevant pathogenic variants for the neuromuscular disease (Table S3, panel C) and for the hypogonadism (Table S3, panel D).
Diagnostic microarray analyses in FA.II-1 (180k aCGH + SNP; Agilent Technologies) did not reveal any potentially pathogenic copy number variation. We also reanalyzed WES or targeted gene sequencing data (as described in (Bachmann-Gagescu et al., 2015)) of 638 families with various ciliopathy diagnoses of unknown genetic cause. We did not identify any additional families with biallelic variants in CBY1 (Table S4), suggesting that CBY1 dysfunction is an ultrarare cause of a hitherto uncharacterized ciliopathy.
Interestingly, CBY1 p.Asn23Profs*24 was described in a study aiming at assessing the contribution of CBY1 to obesity (F. Q. Li et al., 2007). Obesity is documented in ciliopathies including JBTS (Thomas et al., 2015). CBY1 has been identified as a proadipogenic factor required for adipocyte differentiation and the p.Asn23Profs*24 was found in one control subject, which lead the authors to suggest that decreased CBY1 activity may have a protective effect for obesity (Van Camp et al., 2013). Patient FB.II-2, homozygous for p.Asn23-Profs*24, at age 11.5 years had weight of 58.6 kg, >97th percentile, and height of 154.5 cm, 75th percentile. Based on these growth measurements, FB.II-2 was obese (https://www.cdc.gov/ healthyweight/bmi/calculator.html). However, more individuals should be studied to draw a conclusion on the possible effect CBY1 p.Asn23Profs*24 on obesity.
3.2 | Depletion of cby1 in the zebrafish resulted in ciliopathy-related phenotypes We investigated the involvement of CBY1 in ciliary function using zebrafish, whose genome has only one CBY1 orthologue. Zebrafish Cby1 protein displays a high degree of evolutionary conservation, with 71%-75% identity to its vertebrate counterparts ( Figure S2). To determine the temporal and spatial expression pattern of cby1 in zebrafish, we performed semiquantitative RT-PCR and whole mount ISH analysis. The analysis revealed that cby1 was expressed throughout zebrafish embryogenesis and in several organs of the adult zebrafish (Figures 2a-c and S2). We detected maternal cby1 transcripts at early cleavage stages and revealed that cby1 was ex-  Love et al., 2010;Steere et al., 2012;Takemaru et al., 2003;Voronina et al., 2009). In zebrafish MZoval (ovl; ift88) mutants that lack all cilia, defects in Hedgehog, but not in Wnt signaling were observed (Huang & Schier, 2009). Through qPCR analyses, we, therefore, studied potential effects of zebrafish Cby1 loss of function on the expression of Hedgehog and Wnt signaling-associated genes.
In MZcby1 mutant embryos, we observed a significant reduction in gli1 expression compared with respective control, ptc1 was, however, not affected ( Figure S3). Expression of the Wnt signaling genes axin2 and wnt8a was not affected in MZcby1 mutant embryos ( Figure S3). These results indicate that the ciliary defects in MZcby1 mutants partially impair Hedgehog signaling. The results in zebrafish support that reduced Cby1 expression causes a kidney defect, leading to a ciliopathy phenotype observed also in Cby1 −/− mice (Lee et al., 2014).
None of the three affected individuals described displayed gross renal anomalies. However, to further explore if and how loss of CBY1 affects the kidneys in humans, additional affected individuals are required for long-term follow-up.

| Primary cilia of the affected individuals presented structural anomalies
The CBY1  Reduced numbers of primary cilia were documented in embryonic fibroblasts from a Cby1 −/− mouse model, accompanied by reduced levels of ARL13B and AHI1 (Lee et al., 2014). Therefore, we assessed by IF the fraction of ciliated cells and the ciliary structure in fibroblasts from FA.II-1, FA.II-2, and FB.II-2 and from three controls using acetylated tubulin and CEP164 antibodies as ciliary markers.
We detected a reduction in the number of ciliated fibroblasts (p < .0001) and significantly increased ciliary length (p < .0001) after serum starved for 72 h (Figure 4b,c) and for 24 h ( Figure S5). IF also showed significantly reduced AHI1 (p < .0001) and ARL13B  (Aguilar et al., 2012;Dafinger et al., 2011;Hynes et al., 2014;Shi et al., 2017). We, therefore, examined the effect of loss of CBY1 on SMO recruitment to the ciliary membrane in 72h serum-starved fibroblasts before and after stimulation with the SMO agonist for 24 h.
The fraction of SMO-positive ciliated cells was 2%-4% without stimulation and 45%-51% upon SAG stimulation. No difference in the fraction of SMO-positive cells was measured between cells from affected individuals compared with controls ( Figure S8).

| Studies of motile cilia did not reveal anomalies in the patients
Because all three affected individuals presented with airway symptoms, and Cby1 −/− mice exhibited chronic upper respiratory infection F I G U R E 4 Reduced fraction of ciliated cells and increased primary cilia length in fibroblasts from patients. (a) Fibroblasts from two controls (images from one control shown) and from the individuals FA.II-1, FA.II-2, FB.II-2 were fixed after 72 h of serum starvation and stained for ARL13B and CEP164 (Alexa488, green), CBY1 (Cy3, red), and nuclei (Hoechst, blue). The CBY1 signal was detected at the transition zone in the control, but not in cells from FA.II-1, FA.II-2, FB.II-2. Scale bars are 10 µm in the left panel and 2 µm in the magnified pictures. (b) Fibroblasts from two controls and from the affected individuals FA.II-1, FA.II-2, FB.II-2 were fixed after 72h serum starvation and stained for acetylated tubulin, CEP164, and nuclei (Hoechst). The numbers of ciliated fibroblasts were significantly reduced in FA.II-1, FA.II-2, and FB.II-2 (p < .0001) compared with the two pooled controls (C in the figure). The numbers of cells analyzed are indicated on top of the bars. (c) Staining for ARL13B and polyglutamylated tubulin and nuclei (Hoechst) detected a significant increase in ciliary length in serum-starved fibroblasts from the affected individuals FA.II-1, FA.II-2, FB.II-2 compared with the two pooled controls (C). Median length difference between cilia in cells from affected individuals and control cells were 0.44, 0.43, and 1.65 µm, respectively (p < .0001). The numbers of cells analyzed are indicated on top. (d) Intensity measurement for AHI1 costained with polyglutamylated tubulin and Hoechst detected significantly reduced AHI1 signal intensities in serum-starved fibroblasts from FA.II-1, FA.II-2, and FB.II-2 compared with the two pooled controls (C) (p < .0001). The numbers of cilia analyzed are indicated on top of the bars. (e) Intensity measurements for ARL13B costained with polyglutamylated tubulin and Hoechst detected significantly reduced ARL13B signal intensities in serum-starved fibroblasts from FA.II-1, FA.II-2, and FB.II-2 compared with the two pooled controls (C) (p < .0001). The numbers of cilia analyzed are indicated on top of the bars and otitis media (D. Love et al., 2010;Voronina et al., 2009), we also analyzed motile cilia of the respiratory epithelium obtained by nasal brush biopsies in FA.II-2 and FB.II-2 by IF, transmission electron microscopy, and high-speed video microscopy analysis. These experiments did not reveal any anomalies ( Figure S9).

| CONCLUSION
We describe biallelic CBY1 LOF-variants in three individuals presenting with clinical features of JBTS. In line with findings described for JBTS, we detected structural and functional cilia defects in fibroblasts from the affected individuals. These anomalies included increased length of primary cilia and a reduced fraction of ciliated fibroblasts. We also demonstrated that depletion of Cby1 in zebrafish causes ciliopathy-related phenotypes. Comprehensive reanalysis of WES data obtained from a large cohort of individuals from different parts of the world with a clinical diagnosis of ciliopathy did not result in the identification of any additional putative biallelic pathogenic variants in CBY1, indicating that loss of CBY1 is most probably an ultrarare cause of disease.

ACKNOWLEDGMENTS
We thank the families for participating in the study. The exome sequencing for Family A and the RNA sequencing services were provided by the Norwegian High-Throughput Sequencing Centre, a national technology platform supported by the "Functional Geno-