Two Compound Heterozygous Variants in SNX14 Cause Stereotypies and Dystonia in Autosomal Recessive Spinocerebellar Ataxia 20

Autosomal Recessive Spinocerebellar Ataxia 20, SCAR20, is a rare condition characterized by intellectual disability, lack of speech, ataxia, coarse facies and macrocephaly, caused by SNX14 variants. While all cases described are due to homozygous variants that generally result in loss of protein, so far there are no other cases of reported compound heterozygous variants. Here we describe the first non-consanguineous SCAR20 family, the second Portuguese, with two siblings presenting similar clinical features caused by compound heterozygous SNX14 variants: NM_001350532.1:c.1195C>T, p.(Arg399*) combined with a novel complex genomic rearrangement. Quantitative PCR (Q-PCR), long-range PCR and sequencing was used to elucidate the region and mechanisms involved in the latter: two deletions, an inversion and an AG insertion: NM_001350532.1:c.[612+3028_698-2759del;698-2758_698-516inv;698-515_1171+1366delinsAG]. In silico analyses of these variants are in agreement with causality, enabling a genotype-phenotype correlation in both patients. Clinical phenotype includes dystonia and stereotypies never associated with SCAR20. Overall, this study allowed to extend the knowledge of the phenotypic and mutational spectrum of SCAR20, and to validate the role of Sorting nexin-14 in a well-defined neurodevelopmental syndrome, which can lead to cognitive impairment. We also highlight the value of an accurate clinical evaluation and deep phenotyping to disclose the molecular defect underlying highly heterogeneous condition such as intellectual disability.

Autosomal Recessive Spinocerebellar Ataxia 20, SCAR20, is a rare condition characterized by intellectual disability, lack of speech, ataxia, coarse facies and macrocephaly, caused by SNX14 variants. While all cases described are due to homozygous variants that generally result in loss of protein, so far there are no other cases of reported compound heterozygous variants. Here we describe the first non-consanguineous SCAR20 family, the second Portuguese, with two siblings presenting similar clinical features caused by compound heterozygous SNX14 variants: NM_001350532.1:c.1195C>T, p.(Arg399 * ) combined with a novel complex genomic rearrangement. Quantitative PCR (Q-PCR), long-range PCR and sequencing was used to elucidate the region and mechanisms involved in the latter: two deletions, an inversion and an AG insertion: NM_001350532.1:c. [612+3028_698-2759del;698-2758_698-516inv;698-515_1171+1366delinsAG]. In silico analyses of these variants are in agreement with causality, enabling a genotype-phenotype correlation in both patients. Clinical phenotype includes dystonia and stereotypies never associated with SCAR20. Overall, this study allowed to extend the knowledge of the phenotypic and mutational spectrum of SCAR20, and to validate the role of Sorting nexin-14 in a well-defined neurodevelopmental syndrome, which can lead to cognitive impairment. We also highlight the value of an accurate clinical evaluation and deep phenotyping to disclose the molecular defect underlying highly heterogeneous condition such as intellectual disability.
Herein, we describe the first non-consanguineous family with two siblings affected by SCAR20, carrying two SNX14 compound heterozygous variants, of which only one, p.(Arg399 * ), was identified by exome sequencing. Phenotypic similarity to SCAR20, prompted further characterization of the other allele that subsequently was found to contain a complex genomic rearrangement in SNX14: NM_001350532.1:c. [612+3028_698-2759del;698-2758_698-516inv;698-515_1171+1366delinsAG]. Interestingly, siblings also show dystonia and stereotypies, besides the pathognomonic SCAR20 clinical features, representing the first implication of SNX14 in movement impairment. Further functional studies are warranted to understand the underlying molecular mechanisms.

Clinical Report
Patient II:1 is the first female child of a healthy nonconsanguineous Portuguese couple ( Figure 1A). She is currently 29-year-old. She was referred to a pediatric neurology consultation at 2 months for febrile seizures. Family history is negative for neurodevelopmental disorders except for a paternal second-cousin with Prader-Willi syndrome. Gestation and delivery were uneventful except for gestational diabetes at 37 weeks, born at 39 weeks of gestation with weight, 4000 g, length, 51 cm and occipitofrontal circumference (OFC), 36.5 cm. At 9 months she was very hypotonic with no head control; Growth parameters were normal except for OFC. Additional features include diminished reflexes in the lower limbs, bilateral talipes equinovarus with a dystonic hallux, kyphosis, neurosensorial hearing loss, macrocephaly, and facial dysmorphisms: high forehead, hypertelorism, depressed nasal bridge, large base to nose, anteverted nares and thick lips ( Figure 1B). She never acquired independent walking or speech and severe intellectual disability was evident. Although speechless she communicates with the eyes and hands. She also had frequent dystonic manual stereotypies with hands apart.
The male patient II:2, currently 19-year-old, was referred for examination at 3 months when hypotonia and lack of head control were noted. Gestation and delivery were uneventful, born at 40 weeks of gestation with weight, 3250 g; length, 50 cm and OFC, 36.5 cm. Psychomotor development was also severely delayed, with no independent walking or speech acquisition. Macrocephaly with normal additional growth parameters, talipes, diminished reflexes in the lower limbs, bilateral talipes equinovarus with a dystonic hallux, dystonic posture hands, kyphosis, hand stereotypies and facial features similar to his sister were also present ( Figure 1B). He was also very sociable and communicated with facial expressions.
Brain MRI ( Figure 1C) showed diffuse cerebellar atrophy, involving both the vermis and the hemispheres, as well as pontine atrophy, normal spectroscopy and absence of other malformations. Previous normal investigations included metabolic studies such as lysosomal storage parameters, karyotype, array comparative genomic hybridization (aCGH). Following medical genetics referral, the family was selected for research based on the observed putative autosomal recessive inheritance pattern.

Molecular Investigation
Genomic DNA (gDNA) was obtained from peripheral blood of index, brother and parents, using salting out methods (Miller et al., 1988).
Copy number variants calling based on ES data was performed using CoNIFER 2 (Krumm et al., 2012). CNVs with an absolute Z-score greater than 1.7 were considered for analysis. All deletions were considered disruptive, as were duplications in known fully penetrant microdeletion/duplication regions and intragenic CNV duplications. CNVs that overlapped known regions of partial penetrance were considered separately (Firth et al., 2009).

Variant Filtering, in silico Analyses and Segregation Studies
Variants passing following filters were selected for clinical correlation: (a) frequency < 1% (dbSNP, GnomAD Browser, and local databases); (b) gene component, that is, exon and 18-year-old, showing macrocephaly, high forehead, hypertelorism, depressed nasal bridge, large base to nose, anteverted nares and thick lips and younger brother (II:2, right), at 8-year-old, presenting with similar phenotypic features. (C) Brain MRI scan from patient II:2: Sagittal T1 SPGR (i) and coronal T2-weighted (ii) images showing diffuse cerebellar atrophy, involving both the vermis and the hemispheres, as well as pontine atrophy. There is no evidence of dentate calcification. Brain MRI form patient II:1 showed overlapping abnormalities.
Segregation studies of the putative pathogenic variants identified by ES were performed by Sanger sequencing analysis using the BigDye R Terminator v3.1 cycle sequencing kit (Applied Biosystems TM , Foster City, CA, United States), after purification of the PCR products with Illustra TM ExoStar TM 1-Step, (GE Healthcare Life Sciences R , Little Chalfont, United Kingdom).

Screening for Low Covered Regions, Genomic Quantitative PCR (Q-PCR) and Long-Range PCR
Integrative Genomics Viewer (IGV) software, version 2.2.11 (Robinson et al., 2011) was used to manually analyze exome reads coverage-depth. Low covered exons were then sequenced for detection of SNVs or indels, and a low cycle number PCR was performed to screen for short CNVs. Symmetric and asymmetric amplification studies were preformed essentially as described above.
Comparative C T method ( C T ) was used to detect constitutional copy number variants in SNX14 gene. The TTR exon 2 was used as endogenous control. Q-PCR assay was carried out using the PowerUp TM SYBR R Green Master Mix (Thermo Fisher Scientific, Waltham, MA, United States) and 7500 fast Real-Time PCR System (Applied Biosystems TM ), following manufacturer's instructions.
To ensure equivalent efficiency of amplification between each pair of primers, the average threshold cycle value (C T ) was calculated based on the C T obtained for each of the two replicates from a log 10 dilution series of gDNA (20, 10, 5, 2, and 1.25 ng/µl). Data obtained were subjected to linear regression analysis through the semi-logarithmic regression and calculation of the linear correlation value (R 2 ). Only primers that lead to R 2 values close to 1 were considered.
In the final computational analysis of the relative copy number variants of each exon, the average C T and associated standard deviations (SD) was calculated for each sample. These values were normalized by the endogenous control, finding the C T and respective SD values. The values of C T (and respective SD) were determined by the difference between the C T of all samples and the normal control. The relative quantification (RQ) was finally determined by the formula 2 − C T (or 2 − C T ± SD), considering RQ = 1 for the normal control (Weksberg et al., 2005).
To amplify the genomic region encompassing SNX14 introns 7 to 13, by long-range PCR, the RANGER DNA Polymerase (Bioline reagents, London, United Kingdom) was used, following the manufacturer's instructions.

Transcript Analysis
RNA was obtained from peripheral blood samples from both patients, using 5 PRIME PerfectPure TM RNA Blood Kit (Fisher Scientific International, Inc., Hampton, NH, United States). A region harboring exons 5 to 15 was amplified using SuperScript R One-Step -RT-PCR with Platinum R Taq kit (Invitrogen TM , Carlsbad, CA, United States) following manufacturer's instructions. PCR products were separated on a 2% agarose gel, purified using Corning R Costar R Spin-X R plastic centrifuge tube filters (Merck KGaA, Darmstadt, Germany), and further sequenced as described above.
Complex genomic rearrangements, such as the described herein, are believed to be caused by errors in the DNA doublestand breaks (DSBs) repair system (Carvalho and Lupski, 2016). NHEJ, HR and the recently discovered TMEJ are some of the DSBs repair mechanism already known Lieber, 2008;Hustedt and Durocher, 2016;Schimmel et al., 2019). Alu-Alu recombinations are especially common (Sen et al., 2006;Han et al., 2008), which lead us to speculate   -2759del;698-2758_698-516inv;698-515_1171+1366delinsAG]. The 2240 bp deletion occurred between introns 7 (yellow line) and 8 (green line), with breakpoints at L1PA16 and LTR39 elements; the partial intron 8 inversion, is followed by a 4648 bp deletion, that occurred between introns 8 and 13 (blue line), within a AluSx and a AG insertion (vertical black arrow). Black triangles indicate the orientation of each fragment at the mutated allele. (C) A potential structure of the mutant allele is shown in the lower panel in relation to the canonical genomic structure at the top. Lower arrows indicate each SNX14 introns and respective orientation.
that an Alu pairing can be the inciting event. This however, cannot explain the full SNX14 rearrangement, as the deletion breakpoints are within divergent types of retroelements, LTRs (one of the oldest primate-specific L1 sequences) and Alu sequences precluding juxtaposition of homologous domains. Since a LINE-1 element appears partially deleted, it is plausible to hypothesize that a SINE-VNTR-Alu (SVA) might be in the origin of this mutated allele, as SVAs are presumably mobilized by trans LINE-1 reverse transcriptase (Hancks and Kazazian, 2010). However, without a stretch of a new retrotransposon or a polyA tail to embody a newly retrotransposed sequence, this also seems unlikely. Interestingly, pathogenic CNVs in other neurodevelopmental genes such as PLP1, MECP2, APP, and SNCA, are believed to be caused by defects in microhomologymediated DNA repair mechanism (Lee et al., 2007). In fact, Lee et al. (2007) characterized a genomic rearrangement in PLP1 gene, originated by FoSTeS, a mechanism previously known to cause inversions accompanied by duplications and deletions in other organisms such as Escherichia coli (Lee et al., 2007). Although we cannot explain the exact origin of the rearrangement, probably caused by a cascade of errors, it is tempting to speculate that SNX14 gene is prone to the emergence of complex rearrangements as it harbors a large number and variety of repetitive retroelements.
The loss of non-coding RNAs (ncRNAs) at the SNX14 locus could contribute toward the severity of the SCAR20 phenotype, as ncRNAs are implicated in the regulation of protein coding host gene expression (Boivin et al., 2018). However, according to the NONCODE database 5 (assessed at June 22nd 2020) (Liu et al., 2005) there is no ncRNA at these positions in SNX14 gene, so their involvement is unlikely.
Notwithstanding recent developments in sequencing technologies, this kind of genomic rearrangements are overlooked and encumber the identification of the genetic defect particularly in highly heterogeneous diseases such as autosomal recessive intellectual disability (ARID) (Hennekam and Biesecker, 2012). This report highlights the detailed phenotypic evaluation following a multidisciplinary approach as the key for the identification of SNX14 gene involvement.

CONCLUSION
This study discloses the first non-consanguineous SCAR20 family caused by compound heterozygous pathogenic variants in SNX14 gene. Additionally, this study also emphasizes laboratory-clinician crosstalk as key element to guide successful genetic investigation and, therefore, to increase ARID diagnostic yield. At the end, we broaden the SNX14 phenotypic and mutational spectrum, describing two patients with dystonia and stereotypies beyond the typical clinical features of SCAR20, carrying p.(Arg399 * ) and a novel rearrangement encompassing introns 7 to 13, undetected by ES. 5 http://www.noncode.org/index.php

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The study involving human participants were reviewed and approved by the medical

AUTHOR CONTRIBUTIONS
NM: conception and study design, ES data analysis and interpretation, involved in drafting and writing. GS: collected the clinical data, co-writing medical results, interpretation and discussion. CS: shared molecular data analysis and interpretation. IM: data analysis and interpretation, involved in critical reviewing of the manuscript. BR: shared molecular data analysis and interpretation. RS and MP: critical revision, final review, and editing. AB: ES data interpretation, manuscript final critical revision and editing. TT: collected the clinical data, neurological history, co-writing medical results, interpretation and discussion, final review and editing. PJ: conception and study design, involved in drafting and manuscript final revision and editing. All authors gave approval to the final version of the manuscript.