Novel TONSL variants cause SPONASTRIME dysplasia and associate with spontaneous chromosome breaks, defective cell proliferation and apoptosis

Abstract

SPONASTRIME dysplasia is an ultrarare spondyloepimetaphyseal dysplasia featuring short stature and short limbs, platyspondyly, depressed nasal bridge with midface hypoplasia and striated metaphyses. In 2019, an autosomal recessive inheritance was demonstrated by the identification of bi-allelic hypomorphic alleles in TONSL. The encoded protein has a critical role in maintaining genome integrity by promoting the homologous recombination required for repairing spontaneous replication-associated DNA lesions at collapsed replication forks. We report a 9-year-old girl with typical SPONASTRIME dysplasia and resulted in carrier of the novel missense p.(Gln430Arg) and p.(Leu1090Arg) variants in TONSL at whole-exome sequencing. In silico analysis predicted that these variants induced thermodynamic changes with a pathogenic impact on protein function. To support the pathogenicity of the identified variants, cytogenetic analysis and microscopy assays showed that patient-derived fibroblasts exhibited spontaneous chromosomal breaks and flow cytometry demonstrated defects in cell proliferation and enhanced apoptosis. These findings contribute to our understanding of the molecular pathogenesis of SPONASTRIME dysplasia and might open the way to novel therapeutic approaches.

Graphical Abstract

Introduction

SPONASTRIME dysplasia (MIM: 271510) is an ultrarare autosomal recessive spondyloepimetaphyseal dysplasia that mainly features short stature with short limbs, platyspondyly particularly affecting the lumbar metameres, midface hypoplasia with depressed nasal bridge and striated metaphyses (1). Additional features include hypoplastic/short dental roots, coxa vara, limited elbow extension, congenital cataracts, hypogammaglobulinemia and neutropenia (2). Since its first description in 1983, no more than 15 patients have been described and its molecular basis remained undiscovered until 2019, when two research groups reported bi-allelic hypomorphic alleles in TONSL as the unique molecular cause of SPONASTRIME dysplasia in 20 families (2,3).

TONSL maps to 8q24.3 and encodes the Tonsoku-like DNA repair protein that interacts with DNA replication and repair factors, including anti-silencing function 1 (ASF1), minichromosome maintenance complex component helicases (MCM helicases), H3-H4 histones and methanesulfonate-sensitivity protein 22-Like protein (MMS22L) (4,5). The encoded protein TONSL is a genome caretaker which plays a crucial role as a regulator of DNA integrity in human cells. Proteomic analysis revealed that TONSL forms a complex with MMS22L. MMS22L-TONSL complex is recruited to regions of single strand DNA (ssDNA) and promotes the homologous recombination (HR) required for repairing spontaneous replication-associated DNA lesions at collapsed replication forks by stimulating displacement of ssDNA-binding replication protein RPA and assembly of RAD51, the central strand exchange factor in all HR repair reactions (6).

RNAi-mediated depletion of TONSL from human cells causes a high level of double-strand breaks during DNA replication (5). Functional studies in SPONASTRIME dysplasia subject-derived cell lines revealed increased amounts of spontaneous replication fork stalling and chromosomal aberrations (2). Early embryonic lethality in a knock-in Tonsl−/− mouse model and the discovery of reduced length, spinal anomalies, reduced numbers of neutrophils, and early lethality in a Tonsl−/− zebrafish model confirm that altered TONSL function is associated with defects in DNA replication and HR-dependent repair processes, and chromosomal instability (2,3).

In this study, we report a 9-year-old girl with typical SPONASTRIME dysplasia carrying two novel hypomorphic variants in TONSL. To better test the pathogenicity of these novel variants and dissect the biological mechanism(s) that leads to disease, we explored molecular aspects of primary fibroblast patient lines by employing multiple investigative strategies. We demonstrated that patient-derived fibroblasts exhibit increased levels of spontaneous chromosomal breaks, reduced cell proliferation and enhanced apoptosis.

Results

Clinical report

The proband was a 9-year-old girl, daughter of unaffected and unrelated parents. The girl was born at 37 gestation weeks after a pregnancy complicated by ultrasound detection of short limbs and intrauterine growth retardation in the third trimester. At birth, weight was 1550 g (−3.27 Z-score), length 37 cm (−5.21 Z-score), head circumference 31 cm (−2.07 Z-score) and Apgar score 6¹/8⁵. Neonatal period, breastfeeding and nutrition were normal. Primary and secondary dentition was delayed. Growth parameters were always significantly below the third centile, but full endocrinological panel screening was repeatedly normal. IgM, IgG and IgA serum levels, and hematocrit were repeatedly normal. In particular, any sign of immunoglobulin deficiency and neutropenia was annotated. Urinary glycosaminoglycans levels, serum anti-gliadin antibodies, standard karyotyping, SNParray and FGFR3 mutational hot spots screening were normal or negative. She held the head up at 4 months, crawled at 12 months, walked alone at 24 months, said first words at 36 months and first sentences at 4.5 years. At the time of examination, the patient had a previous diagnosis of global developmental delay and needed special education programs, speech therapy twice/week and physical therapy twice/week. At 9 years of age, height was 81.1 cm (−8.78 Z-score), weight 10.5 kg (−9.22 Z-score) and head circumference 46 cm (−5.90 Z-score). Physical examination disclosed frontal bossing, depressed nasal bridge, broad nose, midface hypoplasia, low-set ears, long face, obtuse angles of the mandible, small teeth, short limbs with rhizomelia and relatively normal extremities, genua valga, flatfeet and generalized joint hypermobility (Fig. 1A). Total body radiographic study demonstrated hypo/aplasia of the nasal bones, increased antero-posterior diameter of the cranial vault, severe hypoplasia of the dental roots, shortened long bones with predominant involvement of the humeri, osteopathia striata involving the distal femoral metaphyses and the proximal tibial metaphyses, shortened femoral necks, flattened proximal epyphises of the femurs and humeri, generalized platyspondyly with biconcave aspect of the lumbar vertebrae (Fig. 1B–F). Full ophthalmological and audiological examinations resulted negative.

Figure 1

Clinical and radiological features of the proband. (A) Full body picture showing short-limb dwarfism, rhizomelia, genua valga and facial features. (B) Close-up of the teeth with severe hypoplasia of the roots. (C) Midface hypoplasia, hypo/aplasia of the nasal bones and increased anterio-posterior diameter of the neurocranium. (D) Rhizomelia; (E) platyspondyly and biconcave aspect of the lumbar vertebrae; (F) osteopathia striata.

Molecular finding and in silico analysis

In search of the causative gene, we performed a trio whole-exome sequencing (WES) on DNA from the patient and her unaffected parents. Variants were prioritized according to the criteria reported in the American College of Medical Genetics and Genomics guidelines, listed in the Methods section. Specifically, variants with a minor allele frequency (MAF) > 0.01 were excluded from the analysis. Next, we applied filtering criteria for all nonsynonymous single nucleotides variants and likely gene-disrupting events (frameshift indels, stop gains, stop losses or splice site alterations). To further reduce the number of candidate genes, we filtered for de novo heterozygous, rare compound heterozygous and homozygous variants. We removed genes for which the variants also occurred in other unrelated individuals (Supplementary Material, Table S2). This filtering reduced the number of potential candidate genes to fifth for de novo heterozygous, six for homozygous (E4F1, PCBD2, PHF1, TNXB, LAIR2 and ZBTB9) and five for compound heterozygous variants (AGBL1, CCDC71, HYOU1, MYH14 and TONSL).

Interestingly, while our study was underway, two research groups reported bi-allelic hypomorphic alleles in TONSL (NM_013432) as the unique molecular cause of SPONASTRIME dysplasia (2,3). In agreement with these reports, among the five candidate genes carrying compound heterozygous variants, we detected in the proband’s DNA two previously unpublished bi-allelic missense variants in TONSL: the c.1289A>G variant in the exon 10 predicted to cause the p.(Gln430Arg) amino acid substitution and the c.3269T>G variant in the exon 21 which causes the incorporation of an Arginine at the Leucine 1090 residue (Fig. 2A). The c.1289A>G and c.3269 T>G variants were found in the proband’s mother and father, respectively.

Figure 2

Schematic representation of TONSL gene structure and protein domains. (A) Coding regions are highlighted with black boxes and introns with horizontal lines. TONSL variants identified here are showed. Electropherograms showing DNA sequencing analysis of PCR product amplified with primers targeting TONSL exons 10 and 21 of the DNA proband and her unaffected parents. Nucleotide sequences are provided. Asterisk indicates variants position. (B) Schematic representation of the protein domains of TONSL. TONSL consists of 1.378 amino acids that contain eight TPRs, three Ankyrin repeats (ANK), an UBL domain and seven LRRs. Evolutionary conservation of the Gln430 and Leu1090 residues found in our patient. (C) Modeling of two regions of the TONSL protein. The first region included between residues 1 and 674 contains p.(Gln430Arg) variant. The second region included between residues 1021 and 1378 contains p.(Leu1090Arg) variant. Both variants are highlighted by a vertical rectangle.

No further clinically relevant variant was identified in the remaining genes. Direct Sanger sequencing and carrier analysis confirmed that the patient presented both the variants and that her parents were heterozygous (Fig. 2B). We interrogated the Genome Aggregation Database (GnomAD) (http://gnomad.broadinstitute.org/) to evaluate the frequency of the alternate alleles in different sub-populations: the c.1289A>G, p.(Gln430Arg) and c.3269T>G, p.(Leu1090Arg) variants have a total MAF of 0.00000845 and 0.0000326, respectively. Protein sequences alignment of the heterologous regions including the affected Gln430 and Leu1090 residues of human TONSL was generated by using the Clustal Omega tool and showed that the affected residues were evolutionarily conserved (Fig. 2B). The Gln430 residue is located between the tetratricopeptide repeat (TPR) and three Ankyrin repeats of TONSL, while Leu1090 residue is located in the third leucine-rich repeat (LRR) motif at the C-terminal region of the protein (Fig. 2B). The impact of novel variants on protein function was confirmed by bioinformatics analysis using different online analysis tools including MutationTaster, SIFT, VarSome and PolyPhen-2 and REVEL (Table 1).

Table 1

Characteristics of TONSL variants identified in our subject with SPONASTRIME dysplasia

Variant	c.1289A>G	c.3269T>G
Predicted AA change	p.(Gln430Arg)	p.(Leu1090Arg)
rs-ID	dbSNP: rs776042221	dbSNP: rs137933176
Parental origin	Mother	Father
LOVD-ID	#0000665985	#0000665986
Frequency
gnomAD - European (non-Finnish)
Allele count	2	7
Allele number	106.852	109.926
Number of homozygotes	0	0
Allele frequency	0.00001872	0.0000637
gnomAD–Total
Allele count	2	8
Allele number	236.682	245.380
Number of homozygotes	0	0
Allele frequency	0.000008450	0.0000326
Varsome
Verdict	VUS	VUS
Rules	PM2 BP4	PM2 PP3
Pathogenicity	Pathogenic Benign	Pathogenic
Strength	Moderate Supporting	Moderate Supporting
Predictors
DANN-score	0.9546	0.9976
PolyPhen2 (score)	Damaging (0.996)	Damaging (0.992)
SIFT (score)	Damaging (0.004)	Damaging (0.001)
Eigen (raw coding)	Benign (0.3919)	Pathogenic (0.5696)
Mutation Assessor	Medium (2.085)	High (4.195)
MutationTaster (accuracy)	Disease causing (0.9999)	Disease causing (0.9999)
LRT (score)	Deleterious (0.000006)	Deleterious (0.000006)
FATHMM-MKL (coding score)	Damaging (0.9662)	Damaging (0.8333)
MetaSVM (score)	Tolerated (−0.1857)	Damaging (0.5027)
MetaLR (score)	Tolerated (0.4408)	Damaging (0.6167)
REVEL (score)	Damaging (0.461)	Damaging (0.729)

PM2: absent from controls (or at extremely low frequency if recessive) in Exome Sequencing Project, 1000 Genomes Project, or Exome Aggregation Consortium; BP4: multiple lines of computational evidence suggest no impact on gene or gene product (conservation, evolutionary, splicing impact, etc.); PP3 definition: multiple lines of computational evidence support a deleterious effect on the gene or gene product (conservation, evolutionary, splicing impact, etc.); PP3: multiple lines of computational evidence support a deleterious effect on the gene or gene product (conservation, evolutionary, splicing impact, etc.); DANN score: the value range is 0–1, with 1 given to the variants predicted to be the most damaging; Polyphen-2 (http://genetics.bwh.harvard.edu/pph2/); SIFT (http://sift.jcvi.org/); Eigene (http://www.columbia.edu/∼ii2135/eigen.html); Mutation Taster (http://www.mutationtaster.org/); Mutation assessor (http://mutationassessor.org/r3/). LRT (http://www.genetics.wustl.edu/jflab/lrt_query.html); FATHMM-MKL (http://fathmm.biocompute.org.uk/fathmmMKL.htm).

As a result of the modeling phase, we reconstructed two regions of TONSL by Phyre2, one ranging from the first amino acid to 674 and the second from 1021 to 1378. As an attempt to identify meaningful genotype–phenotype correlations, we assessed and compared the impact of p.(Gln430Arg) and p.(Leu1090Arg) on the stability of these TONSL regions, by measuring their induced thermodynamic change. In detail for p.(Gln430Arg), the difference in free energy (ΔΔG = ΔGmut− ΔGwt) resulted negative and in the range to classify this variant as highly stabilizing [ΔΔG = −2.74 kcal/mol (±0.6925 kcal/mol)]. The difference in free energy caused by p.(Leu1090Arg), resulted positive and in the range to classify this variant as slightly destabilizing [ΔΔG = 0.98 kcal/mol (±0.03 kcal/mol)] (Fig. 2C).

Altogether, our in silico analysis predicted a pathogenic impact of both variants on protein function.

Bi-allelic TONSL variants results in genome instability and cell cycle arrest

In an attempt to explore the pathogenicity of TONSL variants found in our patient, we employed functional studies on established primary dermal fibroblasts lines from the affected individual and her parents.

Chang et al. (3) reported that the functional impairment of TONSL leads to genome instability, which results in cell-cycle arrest and inhibition of cell division. Based on these evidences, we evaluated the effect of TONSL bi-allelic variants on genome stability by assessing metaphase spreads from subject-derived fibroblast cell lines for increased spontaneous chromosome breakage. We analyzed 100 metaphases per sample (proband and her parents). Chromosomal analysis of the proband’s parents was normal (Fig. 3A). Conversely, patient-derived fibroblast cell line showed levels of spontaneous chromosomal aberrations which included gaps, chromatid breaks, chromosome breaks and multiradial figures. These results prompted us to carry out other molecular analyses to corroborate the effect of TONSL variants on the level of spontaneous DNA damage. To this end, patient’s, parents’ and unaffected control’s fibroblast cells were subjected to immunofluorescence with an antibody against p-H2AX, a protein that marks regions of DNA double-strand breaks and contributes to the recruitment of DNA repair factors to the site of DNA damage (18). We performed a quantification analysis of confocal images of p-H2AX staining and we found a significant increase in the levels of p-H2AX foci formation in patient’s cell line compared to that of her parents and control (Fig. 3B). Taken together, these data confirm the role of TONSL in DNA damage response and genome instability.

Figure 3

Patient-derived fibroblasts exhibit increased levels of spontaneous chromosomal aberrations, reduced cell proliferation and enhanced apoptosis. (A) Representative images of metaphases spreads prepared from human fibroblast cell lines established from patient and her unaffected parent’s skin-biopsies. A minimum of 100 metaphases was counted for each experiment. Metaphases spreads of proband’s mother and father showing normal chromosomes, respectively. Metaphase spreads of proband show chromosomes breaks. (B) Confocal study showed the nuclear localization of p-H2AX foci in patient, her parents and healthy subject fibroblast lines, staining with anti-p-H2AX antibody. After acquisition, the intensity of Alexafluor 568 signal was analyzed by measuring the relative intensity of pixels representative for each ROI corresponding of a single cell by LAS-X software. The graph reports means ± standard deviation (SD) of H2AX intensity values from 100 cells for each experiment (^*P < 0.001, n = 2). (C) Cell counts after 24 h and 96 h of incubation. Quantifications were expressed as cells/ml; error bars represent SD of three replicates. (D) Apoptosis of primary fibroblasts derived from healthy subject, patient and her parents were analyzed by flow cytometry by PI staining. Results are expressed as mean ± SEM of triplicate experiments. (E) Cell cycle distribution of primary fibroblasts derived from the patient and her unaffected parents. (F) All pathogenic variants in TONSL identified in SPONASTRIME-affected individuals to date are displayed on the TONSL protein. At the top and bottom of the image are represented the TONSL variants reported in Chang et al. (3) and Burrage et al. (2), respectively. The variants identified here are shown in red.

To better explore whether the chromosomal defects gave rise to altered proliferation rate, cell cycle distribution and apoptosis, we first characterized the growth properties of primary fibroblasts by performing a cell count analysis. We observed a marked lower proliferation rate of patient’s cell cultures (Fig. 3C). Specifically, after 96 h of growth, these cells increased only 1.25 fold, a significant delayed proliferation compared to that of her unaffected father and mother, and control (3.85 fold, 7.18 fold and 5.35 fold, respectively). Together with hypoproliferation, an increase of apoptosis percentage occurred in patient’s fibroblasts compared to the cells of parents and control (37.76% vs ≈ 10%, respectively), suggesting that TONSL aberrant function may be critical and bring to apoptotic response as a consequence of an increased amount of collected unrepaired lesions (Fig. 3D).

In order to strengthen these results, we analyzed the cell cycle progression by flow cytometry. We discriminated fibroblasts in G1, S and G2 phases and found that while cell cycle distribution of parents was comparable to control’s cells, patient’s cells resulted in a significant reduction of G1 fraction (49% vs ≈ 68% and 74%, respectively) and in an accumulation of G2-phase (46% vs ≈ 28% and 23%, respectively) (Fig. 3E). Our results confirm that bi-allelic TONSL variants of TONSL induce a massive cell death in the patient’s cell line.

Discussion

We reported an additional Italian girl with typical SPONASTRIME dysplasia and the previously unpublished p.(Gln430Arg) and p.(Leu1090Arg) variants in TONSL. Both variants were annotated in GnomAD with a very low frequency (MAF 0.00000845 and 0.0000326, respectively) and in silico analysis was in accordance with a presumed disrupting effect on protein function. Metaphase spreads of subject-derived cultured dermal fibroblasts showed an increased rate of spontaneous chromosome breakage in according to Burrage et al. (2), who, for the first time, demonstrated increased genomic instability in patients’ fibroblasts and rescue of this phenotype upon re-expression of the wild-type TONSL by a flag-tagged lentiviral vector. We also provided experimental evidence that the resulting chromosomal instability might impact the growth properties of primary fibroblasts and induce enhanced apoptosis.

To date, TONSL is the unique gene causing SPONASTRIME dysplasia as recently pointed out in two back-to-back papers (2,3). A total of 23 individuals from 21 families, including the present subject, with SPONASTRIME dysplasia and variants in TONSL have been reported (Fig. 3F); bi-allelic variants were identified in 18 families, while only a single mutated allele was found in the remaining three (2,3). Recessive variants in TONSL have been also identified in four subjects from three families with spondylometaphyseal dysplasia but without a clinical diagnosis of SPONASTRIME dysplasia (2). Only individual P04 in Chang et al. (3) showed a homozygous genotype, while all others resulted (compound) heterozygous. Overall, a total of 35 different deleterious variants were identified in TONSL (Fig. 3F). Thirty-three variants were private, while p.(Arg934Trp) and p.(Glu487Lys) recurred in five unrelated individuals with SPONASTRIME dysplasia each. As individuals’ ethnicity was reported only in one of the two back-to-back papers, any speculation on founder effect or ethnic-specificity cannot be proposed for either of the two recurrent variants.

Of the 35 identified deleterious variants, 17 were null alleles (six nonsense, eight frameshift and three variants affecting canonical splice sites), 15 missense, one was a non-canonical splice site variant (c.122-5C>G), one a single gene deletion (exon 23) and one a four base intronic deletion generating two different mutant mRNAs (c.1291-11_1291-14delCCTC). Seven variants fall in the N-terminal TPRs domain, three in the ANK domain, three in the ubiquitin-like (UBL) domain, four in the C-terminal LRR domain and all the remaining affect inter-domain protein regions (Fig. 3F). Deleterious variants are, therefore, distributed along the entire protein region without clear-cut mutational ‘hot spots.’ Available data do not support any significant genotype–phenotype correlations. In particular, key phenotypic features, such as the presence/absence of metaphyseal striae, peculiar facial features and immunological involvement seem to occur independently from the underlying genotype. More observations are needed to better explore this point. Finally, although TONSL is currently the unique gene associated with SPONASTRIME dysplasia (and related phenotypes), the description of typical families without identified variants in TONSL (2) and three subjects with a single mutated allele (3) prompt to consider the existence of further mutational mechanisms affecting TONSL and/or locus heterogeneity, perhaps involving proteins which interact with TONSL.

Our in silico analysis predicted a pathogenic impact of both variants on TONSL protein function. This computational evidence was corroborated by our functional studies carried out on subject-derived fibroblasts. In agreement with Burrage et al. (2), metaphase spreads analysis and confocal studies showed that the patient’s cell line showed spontaneous chromosomal aberrations and an increase of the level of p-H2AX foci formation. These data strongly confirm that the TONSL activity impairment results in replication defects and genome instability. Intriguingly, as a consequence of the chromosomal defects and DNA damage, the patient-derived fibroblasts exhibited reduced cell proliferation capacity and enhanced apoptosis. Our results indicate that cells with bi-allelic variants of TONSL are unable to assemble the needed protein complex for effective DNA repair. Thus, DNA damage reaches a pro-apoptotic threshold inducing massive cell death. From these studies, we conclude that the identified TONSL variants alter TONSL functional activity, and that this compromise potentially leads to SPONASTRIME dysplasia.

The MMS2L-TONSL complex mediates recovery from replication stress and HR via physical interactions with a set of cooperating proteins, including the heterotrimeric replication protein A (RPA1–3), MCMs, ASF1 and the FACT complex (composed by SSRP1 and SUPT16H) (4). In this way, the complex facilitates RAD51-mediated recombination at regions of perturbed replication (19). This mechanism is ubiquitous and crucial for basic cell survival and functions. Reasonably, germline variants in genes encoding these proteins should have so devastating consequences on cell homeostasis and early embryogenesis to be hardly compatible with life. Accordingly, in SPONASTRIME dysplasia, the hypomorphic effect of all identified variants, perhaps, explains the need for a residual function of TONSL for survival. Interestingly, recessive variants in RAD51 cause Fanconi anemia of complementation group R, which share prenatal/postnatal short stature and increased chromosome breakage with SPONASTRIME dysplasia. A single patient with Meier-Gorlin syndrome, another syndromic condition with severe growth restriction, has been described with bi-allelic deleterious variants in MCM5 (20). Finally, 10 individuals with an immunodeficiency syndrome featuring natural killer cell deficiency, increased chromosome breakage and glucocorticoid deficiency have been described with a homozygous variant in MCM4 (21). Therefore, it is tempting to speculate that, with an increasing number of reports, a community of syndromes associated with dysregulation of the MMS2L-TONSL complex and the related DNA repair machinery could be delineated.

In conclusion, this study expands the TONSL mutational spectrum associated with SPONASTRIME dysplasia, contributes to the understanding of the molecular pathogenesis of this disorder and offers the rationale for the design of novel therapeutic targets.

Materials and Methods

Family enrollment, genomic DNA extraction and quantification

The family provided signed informed consent regarding publishing their data and photographs and to underwent blood sampling for the following investigations. This study is conducted in accordance with the 1984 Helsinki declaration and its following modifications, and conservation and use of biological samples for scientific purposes were approved by the local ethics committee (protocol no. GTB12001). Molecular testing carried out in this report is based on the routine clinical care of our institution. Genomic DNA was extracted from patients’ and her parents’ peripheral blood by using Bio Robot EZ1 (Qiagen, Solna, Sweden). The DNA was quantified with Nanodrop 2000 C spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and QUBIT assay (Thermo Fisher Scientific, Waltham, MA, USA).

WES analysis

Proband’s and parents’ DNA was analyzed by WES by using SureSelect Human Clinical Research Exome (Agilent Technologies, Santa Clara, CA), according to manufacturer instructions. This is a combined shearing-free transposase-based library prep and target-enrichment solution, which enables comprehensive coverage of the entire exome. This system enables a specific mapping of reads to target deep coverage of protein-coding regions from RefSeq, GENCODE, CCDS and UCSC Known Genes, with excellent overall exonic coverage and increased coverage of HGMD, OMIM, ClinVar and ACMG targets. Sequencing was performed on a NextSeq 500 platform (Illumina, San Diego, CA) by using the MidOutput flow cells (300 cycles), with a minimum expected coverage depth of 70x. The average coverage obtained was 98X, 46X and 67X for proband, her mother and father, respectively.

The quality of the generated sequences (a.k.a. reads) was checked using FastQC. Reads were eventually trimmed with Trimmomatic (7) before being aligned to the hg19 version of the human reference genome with Bowtie2 (8). Mapped reads were recalibrated based on quality constraints and then processed by GATK ver. 3.7 (McKenna, 2010 #232for the variant calling analytical step). Variants [single nucleotide variants (SNVs) and short insertion/deletions] were annotated by ANNOVAR (9). dbSNP, v150, ClinVar and ExAC v0.3 (10,11) were queried to retrieve allele frequency information in the general population and clinical associated phenotypes for known variants. The deleteriousness of variants was checked by querying the dbNSFP v3.5 for pre-computed functional predictions (12).

Variants underwent a series of consecutive filtering steps, according to the criteria reported in the American College of Medical Genetics and Genomics guidelines (13). Firstly, variants described as benign and likely benign were excluded. Then, remaining variants were classified based on their clinical relevance as pathogenic, likely pathogenic, or as a variant of uncertain significance (VUS) according to the following criteria: (i) nonsense/frameshift variant in genes previously described as disease-causing by haploinsufficiency or loss-of-function; (ii) missense variant located in a critical or functional domain; (iii) variant affecting canonical splicing sites (i.e. ±1 or ± 2 positions); (iv) variant absent in allele frequency population databases; (v) variant reported in allele frequency population databases, but with a MAF significantly lower than expected for the disease; (vi) variant predicted and/or annotated as pathogenic/deleterious in ClinVar and/or LOVD. Common (MAF > 0.01) and synonymous variants were discarded. Finally, only segregating variants among relatives were selected for further in-vitro validation.

Variant validation

The candidate variants identified by WES were confirmed by Sanger sequencing using the proband’s and her parents’ DNA. Primer sequences are reported in Supplementary Material, Table S1. The amplified products were subsequently purified by using ExoSAP-IT PCR Product Cleanup Reagent (Thermo Fisher Scientific, USA) and sequenced by using BigDye Terminator v1.1 sequencing kit (Thermo Fisher Scientific, USA). The fragments obtained were purified using DyeEx plates (Qiagen, Germany) and resolved on ABI Prism 3130 Genetic Analyzer (Thermo Fisher Scientific, USA). Sequences were analyzed using the Sequencher software (Gene Codes, Ann Arbor, MI). The identified variants were resequenced in two independent experiments. Both TONSL variants have been submitted to the LOVD (Leiden Open Variation Database, individual ID # 00301520, https://databases.lovd.nl/shared/variants/0000665985#00021634).

Variant designation

Nucleotide variant nomenclature follows the format indicated in the Human Genome Variation Society (HGVS) (http://www.hgvs.org) recommendations. DNA variant numbering system refers to cDNA. Nucleotide numbering uses +1 as the A of the ATG translation initiation codon in the reference sequence, with the initiation codon as codon 1.

Conservation of TONSL amino acid

Evolutionary conservation of TONSL p.Gln430 and p.Leu1090 residues was investigated with protein sequence alignment generated by Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/) and compared with data provided by UCSC Database (https://genome.ucsc.edu).

In silico variant analysis

Two protein regions of TONSL were reconstructed using Protein Homology/analogY Recognition Engine ver. 2.0 (Phyre2) (14) and enabling the intensive modeling mode, which performs the modeling of a protein using multiple templates and ab-initio techniques. In these regions, the stability of the p.(Gln430Arg) and p.(Leu1090Arg) variants was investigated thermodynamically through the FoldX algorithm. A standalone version of FoldX is downloadable from http://foldx.crg.es (15). It was run with standard parameters. For each region, FoldX computed the total energy of the wild-type and mutated proteins, as a proxy of their overall stability, and the Van der Waals inter-residue clashes, as energy penalization factors. Thus, models were minimized, before assessing their stability, namely all the side chains were slightly moved to reduce the Van der Waals’ clashes.

Cell cultures

Primary dermal fibroblasts were established from skin biopsies of the patient, her unaffected father and mother, and healthy control and cultured in Dulbecco’s-Modified Eagle Medium/Nutrient Mixture F-12 (D-MEMF12) (Thermo Fisher Scientific, USA), plus 10% fetal bovine serum (FBS) (Thermo Fisher Scientific, USA) and 1% streptomycin and penicillin (Thermo Fisher Scientific, USA), as previously described (16) and grown in a 5% CO₂ incubator at 37°C. Fibroblast cultures were deposited into Genomic and Genetic Disorders Biobank (GGDB) at the Fondazione IRCCS-Casa Sollievo della Sofferenza, San Giovanni Rotondo, Foggia, Italy.

Cytogenetic analysis

Patient- and parents-derived fibroblast cell lines were trypsinized and incubated overnight with Colcemid (Thermo Fisher Scientific, Scotland UK) at a final concentration of 0.4 μg/ml to arrest cells in mitosis. Cells were trypsinized again for 10 min, harvested and centrifugated to remove the medium. The pellet was re-suspended in 10 ml of hypotonic solution for 20 min at 37°C and fixed in Carnoy solution (methanol: acetic acid, 3:1, v/v). The cells were spread on ice glass slides and stained in Giemsa solution (5% in sodium chloride physiological solution). Metaphases were examined under a Leica DM2500 (Leica, Wetzlar, Germany) microscope using 20X and 100X objective lenses.

Confocal study

For immunocytochemical analysis, fibroblasts from the patient, her parents and a healthy control were plated in 12-wells plates and subjected to immunofluorescence analysis, as previously reported (17). Cells were counterstained with anti-phospho-H2AX (p-H2AX) antibody (1:100 diluition, #2577S, Cell Signaling, Dellaertweg, Netherlands) for 2 h at room temperature followed by incubation with Alexa Fluor 568 goat anti-rabbit IgG (1:500 diluition, #A11011 Thermo Fisher Scientific, Waltham, Massachusetts, USA), finally with 4′,6-diamidino-2-phenylindole (DAPI, Molecular Probes, #D1306). Cells were examined on a Leica TCS SP8 confocal microscopy (Leica, Wetzlar, Germany). All confocal images were obtained using the necessary filter sets for Alexa Fluor 568 using a X63 (1.2 numerical aperture) water immersion objective. Acquisition of data was performed with the same intensity settings and quantified measuring the relative intensity of pixels representative for each region of interest (ROI) corresponding to a single cell by eLAS X software (Leica, Wetzlar, Germany). Statistical analysis was performed using unpaired, two-tailed Student’s t-test (Excel software) (^*P < 0.001).

Cell cycle, apoptosis analysis and cell proliferation assays

Primary fibroblast cultures established from individuals were grown in DMEM/F12 supplemented with 20% fetal calf serum, 100 units/ml penicillin and 100 ìg/ml streptomycin. The cells (1 × 10⁶) were harvested, washed in 1% BSA, fixed in chilled 70% ethanol, treated with RNase A (100 μg/ml RNase, Sigma Aldrich, St Louis, USA) for 15 min at 37°C and stained with propidium iodide (PI) (10 μg/ml Sigma Aldrich, St Louis, USA) in the dark for 30 min at room temperature.

Data acquisition of cell cycle distribution and apoptosis was carried out using a flow cytometry (FACS Calibur BD Biosciences, Milan, Italy). Data analysis was performed with CellQuestPro software (BD Biosciences, Milan, Italy). For cell proliferation assay, primary fibroblast cells of each individual were seeded in triplicate at a density of 4 × 10⁴ per well on 12 well plates. After 24 h and 96 h of incubation, cell counts were performed using hemocytometer and results were expressed in the number of cells/ml.

Acknowledgments

The authors thank the family for their kind availability in sharing the findings within the scientific community. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Conflict of Interest statement. All authors declare no conflict of interest concerning this work.

Funding

This work was supported by the Ricerca Corrente 2018-2020 Program from the Italian Ministry of Health and Scientific Collaboration Project ‘Undiagnosed Rare Diseases’ Istituto Superiore di Sanità, Fasc. X7E.