Identication of potential pathogenic genes causing occipitalization of the atlas

Occipitalization of the atlas (AO) is among the most common congenital or environmental malformations of the craniovertebral region, the rate of incidence of which ranges from 0.08%-2.76% in the general population. The fundamental pathogenic mechanisms and molecular etiology remain largely unknown.. Rare variants were detected in two genes (MEOX1 and MYO18B) with reported association with vertebral malformations, especially in AO. Through the use of a functional prediction tool, Meta-SVM, and genotype-phenotype analysis of 101 candidate genes related to vertebral segmentation abnormalities, TNS1 was identied as having the greatest probability of being associated with AO, followed by FGFR2, AGAP1, TBX2, LRP5, and TONSL. GO-BP and KEGG analyses were then performed on the candidate genes, the results indicating that Fanconi anemia-related genes may drive vertebral malformation in AO patients.. The of and 6 novel extending the spectrum of known mutants contributing to this Furthermore, functional gene enrichment analysis provided fresh insight into the related and etiology AO. In the current study, GO-BP analysis demonstrated that candidate genes of the AO patient cohort were signicantly enriched in the skeletal system and bone development. KEGG analysis found that these genes were enriched in the Fanconi anemia (FA) pathway, suggesting that FA pathogenic genes could be associated with AO malformation. FA is a relatively rare genetic disorder associated with a progressive decline in hematopoietic stem cells and bone marrow failure. FA is also characterized by a variety of developmental defects including skeletal malformations and short stature. More than 50% of children affected with FA have radial-ray abnormalities and many patients have early-onset osteopenia/osteoporosis. Although more than 20 FA genes have been identied, the underlying mechanism leading to bone malformations remains elusive.


Abstract Background
Occipitalization of the atlas (AO) is among the most common congenital or environmental malformations of the craniovertebral region, the rate of incidence of which ranges from 0.08%-2.76% in the general population. The fundamental pathogenic mechanisms and molecular etiology remain largely unknown..

Results
Rare variants were detected in two genes (MEOX1 and MYO18B) with reported association with vertebral malformations, especially in AO. Through the use of a functional prediction tool, Meta-SVM, and genotype-phenotype analysis of 101 candidate genes related to vertebral segmentation abnormalities, TNS1 was identi ed as having the greatest probability of being associated with AO, followed by FGFR2, AGAP1, TBX2, LRP5, and TONSL. GO-BP and KEGG analyses were then performed on the candidate genes, the results indicating that Fanconi anemia-related genes may drive vertebral malformation in AO patients..

Conclusions
The present study analyzed the WES pro le of a cohort of Chinese AO patients and identi ed 6 novel genes potentially associated with AO, extending the spectrum of known mutants contributing to this disorder. Furthermore, functional gene enrichment analysis provided fresh insight into the Fanconi anemia related pathway and etiology of AO.

Background
Occipitalization of the atlas (AO) is de ned as a form of congenital cervical spine dysplasia. AO was rst described by Rokitansky in 1844 and demonstrated roentgenographically by Schuller in 1911 (as described in Harcourt & Mitchell, 1990). The reported prevalence of AO in the general population has ranged from 0.08-2.76%, females affected at the same rate as males [1,2]. The phenotype affects all or part of the anterior arch, lateral mass, or posterior arch of the atlas, forming a bony connection with the occipital bone. AO is considered to be the main cause of bony deformities in the craniocervical junction, including "atlantoaxial dislocation" and "basilar invagination". Congenital atlanto-occipital fusion may affect the physiological development of the brain during growth and development in adolescence, simultaneously limiting the elongation of the slope [3,4]. Assimilated atlas leads to upward shifting of the axial vertebrae, and further resulting in atlantoaxial instability and compression to the brainstem and cranial nerves. The compression may cause progressive damage to spinal nerve function, blood circulation disorders and cerebrospinal uid circulation disorders. Symptoms include neck pain, breathing di culties, quadriplegia, and may even be life-threatening.
The principal etiology of AO is the impaired development of the bone of the cranio-cervical junction, leading to improper segmentation of the vertebrae. Recent studies have suggested that failed development of the craniocervical junction that underlies AO could be caused by defective somitogenesis in the cervical region resulting from gene mutations [5][6][7]. According to OMIM (Online Mendelian Inheritance in Man), Genecards, and MGI (mouse genome informatics)( MP:0010728), it has been reported that mutations in MEOX1(GCID:GC17M043640, MIM:600147), MYO18B(GCID:GC22P025742, MIM:607295), and the HOX family of genes are associated with AO [7][8][9]. A study by David (1999) of 30 cases of Klippel-Feil syndrome (KFS) found that a mutation in the Hox gene was the mechanism possibly responsible for atlas assimilation, while reduced or impaired Pax-1 gene expression may result in vertebral fusions [10]. Nevertheless, the etiology of AO remains unclear. To further elucidate the molecular basis for AO at the exome level, we herein present the molecular ndings of wholeexome sequencing among a Chinese cohort of 28 AO patients, further investigating rare variants using Meta-SVM, and GO-BP and KEGG analysis.

Variants of known KFS-related genes in AO patients
Following the processing of the WES data and its interpretation relating to the variants, ltered rare variants were rst checked for their potential to be causative for AO in the reported KFS gene. As a result, 1 variant was found in MYO18B ( Table 2) and 1 in MEOX1 (Table 2). In MEOX1, a nonsynonymous SNV occipitalization of the atlas [13]. Consistent with the reported phenotype of craniocervical joint malformations in a MEOX1-de cient mouse model, the patient in the present study suffered from occipitalization of the atlas. However, there remain insu cient studies of pathogenic variants of MEOX1 associated with AO due to the limited numbers of samples.
A synonymous SNV (c.T1581G:p.A527A) in MYO18B was identi ed in patient AO763937, a 33 year-old male with occipitalization of the atlas and vertebral fusion at the C2-C3 level, tonsillar hernia, and syringomyelia. This variant of MYO18B was rst reported in a Saudi boy and girl that were unrelated and exhibited KFS with facial dysmorphism and nemaline myopathy [14]. There is still insu cient evidence in terms of its relevance to AO disease.

Variants identi ed in potential AO-associated genes
Based on rare variant ltering of the AO-associated list of candidate genes (Table S1), the functional prediction tools SIFT, Polyphen-2, Mutation taster, GERP, and CADD were combined to identify candidate genes which were relevant to the etiology of AO [15][16][17][18][19][20]. Furthermore, we compared the Meta-SVM score of rare coding variants for each candidate gene in the AO group with controls, nding 6 genes with a high Meta-SVM score that were related to vertebral malformation (Table 3) [21].
A reported nonsynonymous SNV (NM_001308022 c.2191T > A:p.S731T) in TNS1 was identi ed in patient AO757897, a 21 year-old female complaining of extremity adynamia. Fusion of the craniocervical joint at C0-C1 and C2-C3 was identi ed by radiological scanning, combined with cerebellar tonsillar hernia, basilar invagination, and syringomyelia. TNS1 codes for a protein localized to focal adhesions, regions of the plasma membrane where the cell attaches to the extracellular matrix. This protein crosslinks actin laments and contains an Src homology 2 (SH2) domain, often found in molecules involved in signal transduction [22]. Mutations of this gene have been reported in Cowden syndrome and Proteus syndrome [23]. Cowden syndrome is a genetic condition characterized by a large head, hamartomatous lesions affecting derivatives of ectodermal, mesodermal, and endodermal layers, macrocephaly, facial trichilemmomas, and scoliosis [24,25]. The association of TNS1 with AO or KFS has not been previously reported, thus the present ndings may represent a novel disease association [26].
A heterozygous variant of c.1052T > A (p.V351D) in the FGFR2 gene was detected at an ExAC allele frequency of 0.000025. This variant has been predicted as a pathogenic mutation using the functional prediction tools MutationTaster, SIFT, and Polyphen-2, with a CADD score of 33 and GERP score of 6.03. FGFR2 was reported to be associated with Pfeiffer and Crouzon syndromes, related mostly to the premature fusion of particular skull bones and cervical spine abnormalities [27][28][29].
The variant c.619C > T(p.H207Y) in gene TBX2 was detected in patient AO798026. TBX2 is a member of a phylogenetically conserved family of genes that share a common DNA-binding domain, the T-box. T-box genes encode transcription factors that regulate developmental processes [30,31]. This gene mutation can cause vertebral fusion and scoliosis [32,33]. Therefore, a signi cant association of TBX2 with AO may represent an expansion of diseases known to be associated with this gene mutation.
In AGAP1, LRP5, and TONSL, two other rare heterozygous variants were identi ed in different patients. It has been reported that these genes associate with vertebral or Chiari malformations [34][35][36][37][38][39]. The related information is displayed in Table 3.  Table 4 presents the Fanconi anemia-related genes that were identi ed in 17 patients. Based on the variants of ltered candidate genes, GO-BP and KEGG enrichment analysis was conducted for the AO cohort and controls, respectively. In GO-BP enrichment analysis, the candidate genes of the AO patient cohort were mostly concentrated in terms related to skeletal system development (p = 1.64 x 10-11), epithelial tube morphogenesis (p = 5.54 x 10-10), bone development (p = 3.42 x 10-11), the smoothened signaling pathway (p = 1.14 x 10-10), and epithelial tube formation (p = 9.03 x 10 − 9). In comparison, the controls were signi cantly enriched in terms related to skeletal system development (p = 1.

Discussion
AO is a relatively rare disorder characterized by congenital synostosis of the occipital bone and cervical vertebrae due to a formation or segmentation defect.
However, articles related to the etiology of the mechanism of AO are limited and focused mostly on the type of surgery performed. Thus, this 28 patient cohort is relatively large compared with other reported studies of AO in which to identify the molecular differences by WES. The pathogenesis of the developmental defect in AO was thought to be attributable to the somites at the embryonic stage resulting from a disruption or mutations in genes regulating segmentation or resegmentation, but the underlying etiology of AO remains unclear [40,41]. Previous studies have suggested that mutations in members of the HOX gene family, MYO18B or MEOX1, are associated with the occipitalization of the atlas in a mouse model [42]. In the present study, we detected rare variants in known AO-related genes. However, these novel nonsynonymous mutations in MEOX1 or MYO18B were still of uncertain signi cance.
Comprehensive analysis of gene-related phenotypes and functional prediction tools can identify pathogenic genes associated with AO. As a result, we identi ed 6 genes (TNS1, FGFR2, AGAP1, TBX2, LRP5, and TONSL) with a high Meta-SVM score among candidate genes. TNS1 is involved in cell migration, brilar adhesion formation, cartilage development, and linking signal transduction to the cytoskeleton system. Pathogenic mutations in TNS1 are frequently associated with Cowden and Proteus syndromes. Cowden syndrome is a genetic condition characterized by a large head size, facial trichilemmomas, and scoliosis. Wang et al. reported that the TNS1 mutation may result in adolescent scoliosis in 2020 [26]. Mutations in FRFR2 have been previously associated with Pfeiffer and Crouzon syndromes. These are rare developmental defect disorders with premature fusion of particular skull bones and cervical spine abnormalities. Meina et al. found three generations of a family with Crouzon syndrome carrying a mutation of FGFR2 in 2019, in which extremely severe scoliosis, heterotopic ossi cation, and osteoarthritis had manifested. It has been reported that mutations in TBX2 can cause fusion in vertebral anomalies and variable endocrine and T-cell dysfunction [27]. Yinlin et al. demonstrated that TBX2 was important for skeletal development, especially during vertebral segmentation [32]. These reports reinforce the possibility of a key role for TNS1 and FGFR2 in the pathogenesis of congenital occipitalization of the atlas.
In the current study, GO-BP analysis demonstrated that candidate genes of the AO patient cohort were signi cantly enriched in the skeletal system and bone development. KEGG analysis found that these genes were enriched in the Fanconi anemia (FA) pathway, suggesting that FA pathogenic genes could be associated with AO malformation. FA is a relatively rare genetic disorder associated with a progressive decline in hematopoietic stem cells and bone marrow failure. FA is also characterized by a variety of developmental defects including skeletal malformations and short stature. More than 50% of children affected with FA have radial-ray abnormalities and many patients have early-onset osteopenia/osteoporosis. Although more than 20 FA genes have been identi ed, the underlying mechanism leading to bone malformations remains elusive.
FA and AO are poorly understood because few reports of these two disorders have been published. Houten et al. described a case of a female with FA and AO in 2013 [43]. McGaughran et al. suggested a relationship between FA and vertebral abnormalities [44]. Mazon et al. reported a mouse model with an FA mutant gene that displayed skeletal anomalies such as vertebral fusion and defective bone mineralization [45]. In the present study, 53.6% of patients (15 /28) carried FA-related mutant variants and KEGG analysis indicated that this cohort was affected by FA genetic factors. Furthermore, it has been demonstrated that mice with a mutant FA gene are susceptible to the teratogenic effects of ethanol. Such teratogenicity is largely due to acetaldehyde, a by-product of ethanol catabolism [45].
It is currently unclear how aldehyde and ethanol damage DNA in vivo, but in vitro they can directly modify bases, allowing DNA-DNA and DNA-protein crosslinks. It is likely that acetaldehyde-mediated DNA damage drives FA gene mutant related disorders. In summary, our ndings here provide fresh insight into FA in AO.
The study had a number of limitations. The limited sample size could be a cause of false-positive results. Sequencing studies with larger sample sizes are required to reveal the underlying association between phenotype and genotype.

Conclusions
In conclusion, the results represent an AO study in the Chinese population in which we identi ed a number of candidate genes able to correlate with AO and several FA related mutant genes that correlate with AO etiology. The present research contributes to a more comprehensive understanding of the phenotypegenotype association in AO...

Recruitment of a cohort of AO patients
From Feb 2019 to Feb 2020, 33 Han Chinese ethnicity patients diagnosed with AO at Xuanwu Hospital were consecutively recruited to this study.
We obtained the demographic information, clinical symptoms, physical examination results, and a detailed medical history.
Radiological evaluation, including anteroposterior, lateral neutral, and exion-extension plain radiographs, computer tomography (CT) scan, and magnetic resonance imaging (MRI) were performed on all patients in order to determine a robust clinical diagnosis. We also record segmentation of congenitally fused vertebrae and cervical scoliosis.Each radiographic evaluation was performed by a trained neurosurgery surgeon and reviewed by an another observer blinded to the radiographic evaluation.
The study was approved by Department of Scienti c Research and the Ethics Committee of Xuanwu Hospital in China.
Genomic DNA preparation and whole-exome sequencing Genomic DNA was extracted from 33 patients peripheral blood lymphocytes. We use Illumina libraries following whole-exome capture to perform WES.
Sequencing was carried out on an Illumina HiSeq 4000 platform by using 150-bp paired-end reads mode (Illumina, San Diego, CA, USA).

Variant annotation and interpretation
Variations in each individual were called using a GATK best practice pipeline. In detail, prior to beginning the aligning process, the quality of raw sequencing reads was evaluated using FastQC v0.11.9 [46]. FastQC: A Quality Control Tool for High Throughput Sequence Data [Online]. Available online at: http://www.bioinformatics.babraham.ac.uk/projects/fastqc/]. Low-quality reads and contaminated adaptor sequences were removed from further analysis using the paired trimming mode of the cutadapt v3.2 tool, using the parameters: "-q 20 -m 100A [forward_adaptor] a [reverse_adaptor] [47]. The adaptor sequences were obtained from the sequencing service provider. Remaining sequencing reads were aligned against the human reference genome build 37 (b37) using the mem algorithm of the bwa v0.7.17 tool Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. arXiv:1303.3997v2 [q-bio.GN]]. Variants were called using the HaplotypeCaller algorithm of GATK v4.18. Called variants were further ltered using the CNNScoreVariants and FilterVariantTranches modules of GATK v4.18. vcftools v0.1.12b was used to remove variants that did not pass the lter described above [48].Functional annotation of the remaining variants from the step above was conducted using ANNOVAR and the InterVar pipeline [49]. Disease-causing states were further adjusted according to the clinical information of each patient in accordance with the mutation classi cation protocol of the American College of Medical Genetics and Genomics (ACMG) [15]. The relationship between mutated genes and human disease and phenotype were obtained from DisGeNET [50]. GO-BP and KEGG enrichment analysis for the AO cohort and controls, the AO patient cohort were mostly concentrated in terms related to skeletal system development (p=1.64 x 10-11), epithelial tube morphogenesis (p=5.54 x 10-10), bone development (p=3.42 x 10-11);for KEGG enrichment analysis, the AO patient cohort was signi cantly enriched in pathways related to Fanconi anemia (p=3.53 x 10-12) Figure 2 GO-BP and KEGG enrichment analysis for the controls. The controls were signi cantly enriched in terms related to skeletal system development (p=1.18 x 10-5), embryonic organ development (p=0.0002), ear development (p=4.12 x 10-5); the controls were substantially enriched in pathways related to hepatocellular carcinoma (p=6.88 x