Multi‐gene panel sequencing in highly consanguineous families and patients with congenital forms of skeletal dysplasias

Skeletal dysplasias (SKDs) are a heterogeneous group of more than 750 genetic disorders characterized by abnormal development, growth, and maintenance of bones or cartilage in the human skeleton. SKDs are often caused by variants in early patterning genes and in many cases part of multiple malformation syndromes and occur in combination with non‐skeletal phenotypes. The aim of this study was to investigate the underlying genetic cause of congenital SKDs in highly consanguineous Pakistani families, as well as in sporadic and familial SKD cases from India using multigene panel sequencing analysis. Therefore, we performed panel sequencing of 386 bone‐related genes in 7 highly consanguineous families from Pakistan and 27 cases from India affected with SKDs. In the highly consanguineous families, we were able to identify the underlying genetic cause in five out of seven families, resulting in a diagnostic yield of 71%. Whereas, in the sporadic and familial SKD cases, we identified 12 causative variants, corresponding to a diagnostic yield of 44%. The genetic heterogeneity in our cohorts was very high and we were able to detect various types of variants, including missense, nonsense, and frameshift variants, across multiple genes known to cause different types of SKDs. In conclusion, panel sequencing proved to be a highly effective way to decipher the genetic basis of SKDs in highly consanguineous families as well as sporadic and or familial cases from South Asia. Furthermore, our findings expand the allelic spectrum of skeletal dysplasias.

through its ORIC Large Scale project from the Government of Balochistan including missense, nonsense, and frameshift variants, across multiple genes known to cause different types of SKDs.In conclusion, panel sequencing proved to be a highly effective way to decipher the genetic basis of SKDs in highly consanguineous families as well as sporadic and or familial cases from South Asia.Furthermore, our findings expand the allelic spectrum of skeletal dysplasias.

| INTRODUCTION
Inherited skeletal dysplasias (SKDs) are a heterogeneous group of developmental disorders of the skeleton, also known as osteochondrodysplasias, characterized by abnormal growth of bone and cartilage.The abnormal shape and size of the skeleton lead to disproportional long bones which can result in different types of SKDs. 1 Based on radiographic and molecular features SKDs are classified into distinct groups, including proportionate and disproportionate short stature, increased bone fragility, and skeletal deformities.SKDs show remarkable clinical and genetic heterogeneity that comprises more than 750 types. 2 Each type of SKD is rare, with the overall birth incidence rate worldwide estimated to be $1 of 3300 live births. 3Relatively common SKDs include osteogenesis imperfecta (OI) also known as brittle bone disease, osteopetrosis, achondroplasia, hypochondroplasia, campomelic dysplasia, and thanatophoric dysplasia. 4inically, SKDs can present with short stature, rhizomelic or mesomelic or acromelic limb shortening, bony deformities, or spine involvement.Genetic studies of inherited SKDs offer the opportunity to identify corresponding disease genes, thus providing significant clues in understanding the pathophysiology of these disorders. 5,6This can possibly contribute to the medical and surgical treatment options of the individuals affected with SKD in order to improve their quality of life and lifespan, as exemplified in the case of some genetic types of OI.
Over the past decade, next-generation sequencing (NGS) has enhanced the identification of variants in genes associated with rare diseases, including inherited skeletal dysplasia. 5,7,8Until recently the diagnosis of SKDs usually depended on the opinions of experienced clinicians or radiologists.However, multigene panel sequencing approaches now enable the identification of the underlying genetic cause of SKDs when pathogenic variants are identified, irrespective of the clinical diagnosis.We describe molecular characterization of SKD in individuals affected with SKDs from highly consanguineous Pakistani families as well as, sporadic and familial SKD cases from India by multi-gene panel sequencing.Thus, we show that using a multigene panel targeting a large number of genes associated with SKD is an effective approach in identifying causative variants.The Identification of novel variants in this study contributes to the understanding of phenotypic and genotypic aspects of SKDs.

| Genetic analyses
Briefly, to identify the disease-causing gene in the seven enrolled families segregating autosomal recessive SKDs, we first excluded relatively common loci/genes for SKD, for example GDF5, FGFR3, TRPV4, and BMPR1B either directly by Sanger sequencing or by genotyping STR (short tandem repeat) markers flanking or within the candidate gene in the DNA of the affected as well as the unaffected family members, when available (data not shown).All seven highly consanguineous families were excluded for the common genes/loci and the 27 sporadic and familial patients with SKDs were subjected to multigene panel sequencing in order to identify the causative variants.
For NGS-based genotyping, we used a custom-designed SureSelect XT gene panel (Agilent, Santa Clara, CA, USA) containing all coding exons of 386 genes associated with skeletal disorders at the time of conducting this study (skeletal disease-associated genome [sDAG] panel 9,10 ; Table S1).After library preparation and enrichment, sequencing was performed using MiSeq (Illumina, San Diego, CA, USA).Sequencing reads were mapped to the haploid human reference genome release (GRCh37) using BWA MEM.Variants were called with the GATK toolkit and annotated by ANNOVAR.For variant filtering and prioritization, the tools GeneTalk and PhenIX were used. 11,12 find candidate variants we used a filtering strategy, which was found to be effective in our previous studies. 7,134][15][16] The clinical significance and novelty of the identified candidate variants were assessed in ClinVar and HGMD ( ® Professional 2023.3)databases.Analysis of splice site effects was performed by NNSPLICE (http://www.fruitfly.org/seq_tools/splice.html).Moreover, for all the missense variants, conservation of the changing residue was determined by Phy-loP100way scores (PhyloP100way scores: À20.0 to 10.003 max, and are most conserved; UCSC multiz alignments of 100 vertebrates, hg38).
After filtering the variants, the candidate variants were further prioritized using Exomiser, 17 which predominantly prioritizes variants based on minor allele frequency (MAF), predicted pathogenicity, and the presence of a variant in the gene known (or functional relevance) in a disease similar to the patient's phenotype as well as the inheritance model.

Validation of NGS-detected variants and cosegregation analyses
were performed in all familial cases by Sanger sequencing after PCR amplification using reverse and forward primers flanking the respective variant.However, segregation analyses of variants by Sanger sequencing were not performed in sporadic and familial SKD cases.
Sequencing was performed on an ABI 3100 genetic analyzer (ThermoFisher Scientific).The nomenclature of the variants reported here is according to the standard guidelines as defined by Human Genome Variation Society (HGVS).

| RESULTS
In this study, we investigated seven large, highly consanguineous families from Pakistan with at least three affected individuals per family and a cohort of 27 sporadic and familial patients from India with different types of SKDs.
In the consanguineous families, we performed multi-gene panel sequencing and were able to identify the underlying genetic cause in five families corresponding to a diagnostic yield of 71%.Segregation analysis in 14 affected patients further confirmed the pathogenicity of the variants.
In the sporadic and familial SKD cases, we identified the pathogenic variants in 12 out of a total of 27 individuals corresponding to a diagnostic yield of 44%.The zygosity across these 12 patients included: heterozygous variants in 8 patients and biallelic variants in 4 patients (3 homozygous and 1 compound heterozygous variants; Table 1).Segregation analysis was not possible for most of the sporadic and familial SKD cases because DNA samples of their family members were not available.
Looking at the allelic heterogeneity across all cases from both cohorts, we found 8 missense, 4 nonsense, and one frameshift variant out of the total 12 positive cases from sporadic and familial SKD patients (Table 1).Whereas, in the large consanguineous families, we identified five distinct variants in five different genes associated with congenital forms of skeletal dysplasia (Table 2).Of the seven families, five families were found to carry disease-causing variants in NPR2, MATN3, PAPSS2, CCN6, and DYM genes.The mutational spectrum across these five genes included two missense variants, one nonsense variant, and one small deletion leading to a frameshift (Table 2).Consistent with the autosomal recessive pattern of inheritance observed in the consanguineous families, all five identified variants were homozygous and segregated with the disease in the extended families.
The minor allele frequency (MAF) of most of the the variants identified has not been reported in large publicly available databases such as dbSNP and gnomAD, except for the missense variants c.2101G > C; p.Ala701Pro (rs137853116) in IFT80, c.361C > T; p.Arg121Trp (rs104893637) in MATN3, c.835C > T; p.Arg279Trp (rs104893915) in SLC26A2 and c.679G > T; p.Asp227Tyr (rs781145233) in PEX7 in the sporadic familial cases from India (Table 1).and a nonsense variant c.156C > A; p.Cys52Ter (rs121908901) in CCN6 as well as a missense variant c.1022G > C; p.Arg346Pro (rs752982930) in the PAPSS2 with very rare MAF in dbSNP and gnomAD in the consanguineous families (Table 2).Furthermore, the residues changed by the identified missense variants were found to be highly conserved with high positive Phy-loP100way scores (Tables 1 and 2).
We found some of the identified variants particularly interesting.
Most notable is a rare homozygous missense variant in MATN3 associated with recessive spondyloepimetaphyseal dysplasia (SEMD), which provided significant insights into its unusual recessive effect.Additionally, cases 15-1684 and 16-2417 are highlighted for their manifestation of rare SKDs with skeletal and growth implications, underscoring the significance of genetic diagnosis and counseling.We discuss these two instances in detail below.

| A consanguineous family with a homozygous missense variant in MATN3
This highly consanguineous family DW-20 included three affected individuals VI:5, VI:6, and VI:7 aged 15, 18, and 17 years, respectively (Figure 1A).All affected individuals, in general, had bowing of both knees and elbows since childhood, waddling gait as well as proportionate short stature with no facial dysmorphism (Figure 1B,C).X-ray radiographs of the affected individuals (VI:6, VI:7) revealed metaphyseal expansion.Dysplastic epiphyses were also observed in both the wrists and knees of both patients VI:6 and VI:7.A lateral x-ray of the lumbar spine revealed oval vertebrae with a bullet-like appearance, flattened surfaces (platyspondyly), and insignificant posterior curvature.The unaffected father V:7 is atypical of SEMD (Figure 1C).This study Note: All the patients with heterozygous variants were observed with unaffected parents.Genomic positions are according to GRCh37/hg19 human genome assembly.Abbreviations: gnomAD, genome aggregation database; het, heterozygous; hom, homozygous; SIFT, sorting intolerant from tolerant; n.a., not applicable.
to affect the vWFA (von Willebrand factor A domain) domain of the MATN3 protein, with a predicted deleterious effect by Polyphen-2, SIFT, and has a CADD score of 25.9 (Figure 1E).The amino acid residue Threonine "Thr" at position 195 of the MATN3 protein (NM_002381.5)affected by this variant was found conserved across species (Figure 1F).

| Case 15-1684
A 7-year-old female child presented with deformity in her legs and short stature.On examination, she had a prominent forehead, flat facial profile (Figure 2A), blue sclerae, narrow thorax, pectus carinatum, brachydactyly, tapering fingers, genu varum, and flat feet.A radiographic skeletal survey revealed generalized osteoporosis, platyspondyly, thoracic kyphoscoliosis, small and square iliac wings, coxa valga, short femoral necks, dysplastic epiphyses, and flared metaphyses esp.evident at elbows and knees, and bowing of both radii.X-rays also showed ectopic calcifications in the chest (Figure 2A).She had one prolonged admission to the hospital fol-

| Case 16-2417
A three-year-old female child was brought with a complaint of not gaining adequate height.She had short stature, mild facial dysmorphism with retrognathia, and short hands, with ulnar deviation of hands.
The child was, however, good in studies and in extra-curricular activities.The anthropometric evaluation revealed a weight of 10.4 kg (zscore: À2.23 SD) and a height of 74 cm (z-score: À5.61 SD).There were no associated malformations in the heart and kidneys.
There was a positive family history.Deformity and shortening of bilateral upper limbs were also observed in the mother and maternal aunt.
Both these sisters also had short stature.X-rays showed a bowed T A B L E 2 List of variants identified in five consanguineous families segregating SKDs.

| DISCUSSION
In this study, we performed panel sequencing of 386 genes (Table S1) known or thought to associate with SKDs in seven large highly consanguineous families from Pakistan and a cohort of 27 sporadic and or familial patients with congenital SKDs from India.We detect causative variants in a majority of the highly familial cases and less than half of the sporadic and familial SKD cases.Our study highlights the remarkable genetic heterogeneity of SKDs, which is evident by the identification of causative variants in 20 different genes (Tables 1 and 2) involved in bone morphogenesis.We found several novel variants in genes associated with SKDs including genes in which very few pathogenic variants have been reported so far.The identification of causative variants in genes with few reported variants in patients with SKD further confirms their potential role in SKD.
Overall, in the 27 cases with sporadic and familial SKD, we identified 12 disease-causing variants including five heterozygous missense, four nonsense, one frameshift variant, and three homozygous missense variants in 12 different genes associated with different types of SKDs (Table 1).All of the identified variants in highly consanguineous families and in cases with sporadic and familial SKD were either very rare (MAF <0.01%) or absent in gnomAD (Tables 1 and 2) and predicted to be deleterious by several Bioinformatics tools such as Polyphen-2, SIFT, and CADD score.Most of the variants were either pathogenic, likely pathogenic or variants of uncertain significance (VUS) according to ACMG criteria (Tables 1 and 2).
In the familial cases, we identified five disease-causing variants in 5 different genes (NPR2, MATN3, PAPSS2, CCN6, and DYM).Among these, three were unknown p.Tyr225Cys in NPR2; p.Thr19Arg in trafficking in the endoplasmic reticulum. 28The child interestingly had micromelia and chest wall calcifications.Hyoid calcifications and tracheobronchial calcifications are previously reported.Furthermore, the heterozygous nonsense variant c.412G > T (p.Glu138Ter) in SHOX, detected in an affected girl with familial short stature, facial dysmorphism, and abnormalities of hands, likely leads to SHOX truncation.However, she and her mother did not have typical Madelung deformity described earlier.Predominantly, small deletions or deletions of the entire SHOX gene have been reported, accounting for around 80% of the cases, whereas point variants are likely less common. 29Therefore, the DDR2 and SHOX-associated variants in this study not only expand the phenotypic spectrum but also highlight the broader clinical relevance of these variants.
Furthermore, there was a significant difference in the diagnostic yield between the patients from large consanguineous families and sporadic and or familial SKD cases.The detection rate was likewise higher in consanguineous families (71%) than in the cases with sporadic and or familial SKD (44%) cases. 30Our study, therefore, suggests that the NGS multigene panel sequencing approach targeting a large number of genes (386 genes; Table S1) associated with SKDs is a powerful tool, especially in a cohort of patients from large consanguineous families.SKDs (genes not on the panel), which cannot be detected by NGS panel sequencing. 19 summary, we have demonstrated the molecular characterization of SKD in two different cohorts.Our findings not only expand the phenotypic and genotypic data of SKDs, but will also offer an immediate benefit to the families for carrier testing and molecular prenatal diagnosis in selected families.A confirmatory diagnosis enables better genetic counseling and prognostication in affected families.
Further studies are warranted to identify hotspot variants or population-specific variants in different geographic areas.

2 | MATERIALS AND METHODS 2 . 1 |
Recruitment of patients and DNA extraction A total of seven large highly consanguineous families, comprising a total of 22 patients, segregating autosomal recessive SKDs from Pakistan and 27 sporadic and familial patients suggestive of SKDs from India were enrolled in this study.The majority of the cases were isolated, however, consanguinity in some cases could not be ruled out.Genomic DNA extraction from peripheral blood samples was performed according to standard procedures.The protocol in this research was approved by the "Institutional Review Board (IRB# 00010538) of BUITEMS, Quetta, Pakistan," the "Institutional Ethics Committee of Kasturba Hospital, Manipal (IEC 164/2014), India," and the "Institute Ethics Committee (Ref: IEC/2021/SPL122) PGIMER, Chandigarh, India".Written informed consent was obtained from all the participants.Investigations were conducted according to the Declaration of Helsinki.Initial diagnosis of the disease of the affected individuals in each family, as well as sporadic and familial SKD patients, was made based on physical examination, photographs, and radiographs available.
We identified the homozygous missense variant c.584C > G (p.-Thr195Arg) in MATN3 causing spondyloepimetaphyseal dysplasia in this family (Figure1D).So far very few homozygous missense variants have been reported in MATN3 (HGMD ® Professional 2023.3).The identified variant was found in exon 2 of MATN3, which is predicted T A B L E 1 List of variants identified in 12 of 27 sporadic and familial patients with SKDs.
lowing a chest infection and breathing difficulty, where she was tracheotomized and received oxygen therapy.She was subsequently discharged on intermittent home oxygen therapy, which was needed for over 5 years.She had very limited height gain over the period.HRCT chest was done which revealed calcification involving tracheal and bronchial cartilage as well as the lobar and segmental division of bronchi.The possibility of spondyloepimetaphyseal dysplasia short limb hand (SEMD-SL) type with abnormal calcification was kept.Testing for the most common two achondroplasia variants was negative.TRPV4 variant analysis was also negative.Targeted next-generation sequencing (NGS) revealed a homozygous pathogenic variant c.2107 T > C (p.Ser703Pro) in DDR2.At the age of 14 years, her height was less than 100 cm (micromelia).She had a large head (preserved head growth) with hypoxemia requiring oxygen therapy at night but was mentally alert and responsive.The homozygous missense variant c.2107 T > C (p.Ser703Pro) in DDR2 in addition to SMED is associated with growth implication and, thus expands the phenotypic spectrum of DDR2 mutations.
Genomic positions are according to GRCh37/hg19 human genome assembly.Abbreviations: Chr, chromosome; gnomAD, genome aggregation database; hom, Homozygous; SIFT, sorting intolerant from tolerant; n.a., not applicable.radius and hypoplastic ulna in the child, and lower limb x-rays were normal.The variant analysis by targeted next-generation sequencing (NGS) panel for skeletal dysplasia revealed a heterozygous variant in the SHOX gene in the child c.412G > T (p.Glu138Ter).The variant was confirmed also in the mother by Sanger sequencing.The inheritance pattern suggested is pseudoautosomal dominant and disease due to F I G U R E 1 Family (DW-20) with homozygous missense variant in MATN3 gene associated with arSEMD matrlin 3 type.Pedigree of family (DW-20) segregating c.584C > G (p.Thr195Arg) variant.Co-segregation of variants/genotypes is represented with: C, reference (wild type) allele; G, mutant allele (A).Photographs of affected individuals (VI:5, VI:6, and VI:7) and of her hands and feet showing features of arSEMD, matrilin 3 type (B).X-ray radiographs of thorax, pelvis, upper legs with knee joints, hands, and feet of affected individuals (VI:6 and VI:7) from family DW-20 showing signs SEMD, matrilin 3 type and of the unaffected father atypical of spondyloepimetaphyseal dysplasia (V:7) showing absence of signs typically found in individuals heterozygous for MATN3 variant (C).A part of MATN3 Sequence chromatogram of the unaffected mother (V:6), an affected (VI:5), and an unaffected (VI:2) showing the c.584C > G variant in family DW-20 (D).Schematic representation of MATN3 protein with the (p.Thr195Arg) missense variant in vWFA domain (modified from Pei, M et al. 18 ; (E).Multiple sequence alignments of MATN3 orthologs showing the changing residue Threonine "T" marked in the red stripe which is conserved across species (F).[Colour figure can be viewed at wileyonlinelibrary.com]SHOX variant in this family is Leri Weill dyschondrosteosis (LWD) (Figure 2B).

MATN3;
and p.Leu53GlyfsTer13 in DYM and two were known disease-causing variants p.Cys52Ter in CCN6 and p.Arg346Pro in PAPSS2 (Table2).19,20Of the 17 variants identified in this study, the homozygous missense variant c.584C > G (p.Thr195Arg) in MATN3 gene causing spondyloepimetaphyseal dysplasia (SEMD) was particularly interesting.Variants in MATN3 have been found to cause two allelic disorders, that is, SEMD MATN3 type associated with homozygous variants of MATN3 and multiple epiphyseal dysplasia (MED) associated with heterozygous variants of MATN3.21,22It is worth noting that, to date, about 27 heterozygous dominant variants have been reported in the MATN3 gene (HGMD, 2023.3).21However, only four homozygous recessive variants in MATN3 have been reported to cause autosomal recessive SEMD.[22][23][24]Consistent with earlier reports, variants causing both disorders SEMD and MED, are frequently located in exon 2 of the MATN3 gene which encodes the vWFA domain.21,22The recessive effect of this variant (c.584C > G; p.Thr195Arg) is very interesting since the heterozygous variant p.-Thr195Lys that also leads to exchange to a positively charged amino acid (Lysine) was shown to cause autosomal dominant multiple epiphyseal dysplasia (MED).25 Homozygous Matn3 loss of function in mice does not cause any skeletal consequences.26Therefore, MATN3 missense mutation are assumed to lead to a pathological gain of function through abnormal folding and ER stress.The homozygous missense variant c.2107 T > C (p.Ser703Pro) in DDR2, identified in a patient with SEMD-SL, 27 is also associated with growth implications.This variant is predicted to affect the tyrosine kinase domain of the DDR2 protein.Missense variants in the kinase domain of DDR2 have been shown previously in proteins' abnormal Consistent with the literature, the failure to diagnose SKDs patients by NGS panel sequencing in this study could be attributed to large structural variants (SVs), copy number variants (CNVs), deep intronic variants, or variants in novel genes not yet associated with F I G U R E 2 A Child (15-1684) with DDR2 variant causing SEMD short limb hand type presenting with severe micromelia; x-ray showing meso-acromelic shortening with bell-shaped thorax (tracheostomy in situ) and ectopic calcifications and bowing of radii (A).Baby (16-2417) shows deformed ulna and bowed radius, Ulna is shorter than radius leading to ulnar deviation at wrist (B).[Colour figure can be viewed at wileyonlinelibrary.com]