An Axin2 mutation and perinatal risk factors contribute to sagittal craniosynostosis: evidence from a Chinese female monochorionic diamniotic twin family

Craniosynostosis, defined as premature fusion of one or more cranial sutures, affects approximately 1 in every 2000–2500 live births. Sagittal craniosynostosis (CS), the most prevalent form of isolated craniosynostosis, is caused by interplay between genetic and perinatal environmental insults. However, the underlying details remain largely unknown. The proband (a female monochorionic twin diagnosed with CS), her healthy co-twin sister and parents were enrolled. Obstetric history was extracted from medical records. Genetic screening was performed by whole exome sequencing (WES) and confirmed by Sanger sequencing. Functional annotation, conservation and structural analysis were predicted in public database. Phenotype data of Axin2 knockout mice was downloaded from The International Mouse Phenotyping Consortium (IMPC, http://www.mousephenotype.org). Obstetric medical records showed that, except for the shared perinatal risk factors by the twins, the proband suffered additional persistent breech presentation and intrauterine growth restriction. We identified a heterozygous mutation of Axin2 (c.1181G > A, p.R394H, rs200899695) in monochorionic twins and their father, but not in the mother. This mutation is not reported in Asian population and results in replacement of Arg at residue 394 by His (p.R394H). Arg 394 is located at the GSK3β binding domain of Axin2 protein, which is highly conserved across species. The mutation was predicted to be potentially deleterious by in silico analysis. Incomplete penetrance of Axin2 haploinsufficiency was found in female mice. Axin2 (c.1181G > A, p.R394H, rs200899695) mutation confers susceptibility and perinatal risk factors trigger the occurrence of sagittal craniosynostosis. Our findings provide a new evidence for the gene-environment interplay in understanding pathogenesis of craniosynostosis in Chinese population.


Introduction
Craniosynostosis (CS), defined as premature fusion of one or more cranial sutures, affects approximately 1 in every 2000-2500 live births [1]. CS contains an isolated condition (non-syndromic craniosynostosis, NCS) and complex syndromes (with other malformations, syndromic craniosynostosis) [2]. Sagittal craniosynostosis is the most prevalent form of NCS, accounting for 40-58% of all NCS cases [3]. Sagittal suture premature closure restricts the widen of the skull and then causes the scaphocephaly deformity and other adverse neurologic outcomes [4].
Monochorionic (MC) twins, sharing almost the same genome, offer a unique opportunity to study the geneenvironment interactions, for the healthy twin as an ideal control. Discordant phenotypes between MC twins emphasize the interplay between genetic and environmental influences in etiologies of the disease [16].
In our study, we found that a heterozygous mutation of Axin2 (c.1181G > A, p.R394H, rs200899695) exist in Chinese female monochorionic sisters and their father. However, only the proband, suffering a persistent breech presentation and intrauterine growth restriction, was diagnosed with sagittal craniosynostosis. Public database indicated that Axin2 (c.1181G > A, p.R394H, rs200899695) mutation was not detected in Asian population. According to data from The International Mouse Phenotyping Consortium (IMPC, http:// www. mouse pheno type. org), about 22% female Axin2 heterozygous knockout mice developed abnormal head shape before delivery. Thus, we speculate that this particular Axin2 mutation leads to haploinsufficiency in female with incomplete penetrance, and additional environmental insults eventually trigger the occurrence of sagittal craniosynostosis.

Clinic Examination and Information
All participants signed the informed consent and received physical examination by two experienced surgeons in Children's Hospital of Nanjing Medical University. All samples used in our study were in compliance with the informed consent and agreement of patients. This study was approved by the ethics committee of Nanjing Medical University.
Clinical information of pregnancy history and infant clinical data were obtained from hospital medical records. Head CT scan of the healthy child was performed upon the request of the parents.

Whole Exome Sequencing and Data Analysis
Genomic DNA, extracted from peripheral blood samples (proband, proband's co-twin sister, the parents) and proband's skull periosteum tissue, was subjected to whole-exome sequencing (WES) on the platform of Genergy Biotechnology, Shanghai, China. Raw reads were aligned to the human genome reference assembly (GRCh37/hg19) using the Burrows-Wheeler Aligner [17]. The Picard software was employed to remove PCR duplicates and evaluate the quality of variants. DNA variants was called and analyzed using the Genome Analysis Toolkit [18]. The variants with read depths less than 4 × were filtered out. All variants were further annotated [19][20][21][22][23][24][25]. The workflow of genetic analysis was shown in Fig. 1.

Structural Analysis
Three-dimensional models of the wild-type and mutant Axin2 protein were constructed by I-TASSER [27] and visualized using the PyMOL software (PyMOL Molecular Graphics System, DeLano Scientific, San Carlos, CA).

Clinical information
The pedigree came from Wuhan, Hubei Province,  ) and the results of Sanger sequencing were shown. The gDNA samples were extracted from peripheral leucocytes of the parents and twin sisters (a-d) and from skull periosteum tissue of the proband (e). Please note that the DNA sequences are shown in + strand, but Axin2 is located on-strand. The father and two twins have this missense mutation (a, c, d, e), but the mother does not carry this mutation (b) at the age of 9 months in the Department of Neurosurgery, Children's Hospital of Nanjing Medical University. She was born to non-consanguineous parents without family history of craniosynostosis. Her mother, a 30-year-old Chinese female, conceived MCDA twins by frozen embryo transfer (FET) (Supplementary Fig.  S1-3). Her father was 36-year-old at that time. Twentysix days after FET, two embryos inside a gestational sac (approximately 22 mm × 13 mm) was confirmed by four-dimensional ultrasound scan ( Supplementary Fig.  S1). Embryo length were 2.8 mm and 3.3 mm, respectively. Fetal heart rate were 107 per minute and 118 per minute, respectively (Supplementary Fig. S1). At 13 weeks of gestation, IgG of cytomegalovirus and herpes simplex virus were detected in the mother's serum (Supplementary Table S1). At 14 weeks of gestations, dietary assessment indicated that the mother had inadequate intakes of energy, protein, fat, several vitamins and minerals, while excessive consumption of carbohydrate (Supplementary Table S2). At 17 weeks of gestations, decreased thyroid-stimulating hormone (TSH) level and increased level of urinary iodine were detected (Supplementary Table S3). At 27 weeks of gestations, the mother was diagnosed with gestational diabetes mellitus (Supplementary Table S3).
The ultrasound scan detection indicated that the proband (II-1) had been in persistent breech position on the left side of uterus, while the younger sister (II-2) had been in cephalic position (Fig. 2a, b). At 28 weeks of gestations, the proband (II-1) and twin younger sister (II-2) were born via spontaneous vaginal delivery (Fig. 2a, b). Birth weight of proband (II-1) and co-twin (II-2) was 880 g (ranking 50 th -90 th centiles for postnatal weight of infants [31]) and 990 g (ranking 90 th -97 th centiles for postnatal weight of infants [31]), respectively. The proband (II-1) were diagnosed with sagittal craniosynostosis and intrauterine growth restriction (Fig. 2c), while no signs of craniofacial deformity were detected in parents (I-1, I-2) and co-twin sister (II-2) (Fig. 2c). Cranial index, which represents the ratio of maximum cranial width to maximum cranial length, is decreased in patients with sagittal craniosynostosis [32]. In our case, the cranial index of the proband and co-twin sister was 70.9%, 83.2%, respectively. In addition, tooth agenesis, oral clefts or colorectal cancer were not detected in this family through physical examination by clinical doctor as well as medical history inquiry. And the other family members have no history of tooth agenesis, oral clefts or colorectal cancer. Taken together, clinical records indicate that the proband suffered sagittal craniosynostosis, persistent breech presentation and intrauterine growth restriction, except for other shared perinatal risk factors of the twins. Whole-exome sequencing was applied to identify the potential genetic etiology leading to sagittal craniosynostosis in our case. Due to the low incidence rate of sagittal craniosynostosis [3], we focused on private and/or rare (minor allele frequency, MAF < 0.01) variants on exons or splicing sites (Fig. 1). However, none of candidate germline or somatic mutations were proband-specific. As gene-environment interactions have been demonstrated in the pathogenesis of craniosynostosis [5,15], we wonder whether the intrauterine risk exposures (environment factors) triggered the susceptible individual to develop sagittal craniosynostosis. Based on this hypothesis, we re-analyzed our sequencing data and identified a heterozygous missense mutation of Axin2 (c.1181G > A, p.R394H, rs200899695) in the leukocytes of subjects I-1, II-1 and II-2, and skull periosteum tissue of subject II-1. This finding was further validated by Sanger sequencing ( Fig. 3a-e). The frequency of Axin2 (c.1181G > A, p.R394H, rs200899695) mutation of global population is 0.000849, 0.000637, 0.00008 in gnomAD, ExAC, ALFA database, respectively, while is not detected in Asian population (Table 1).

Functional annotation and structural analysis and of Axin2 (c.1181G > A: p.R394H, rs200899695)
Axin2 (c.1181G > A: p.R394H, rs200899695) mutation is predicted to be potentially deleterious by in silico analysis (Table 3) and the wild type Axin2 (c.1181G) loci is located within a region modified by H3K4Me3 and H3K27Ac (Fig. 5 a). In addition, p. R394H substitution is predicted to affect spatial structure of AXIN2 GSK3β binding domain (Fig. 5 b, c).

Discussion
Craniosynostosis, a highly heterogeneous disease, is caused by genetic mutations, adverse environmental exposures and their interactions. Identifying the pathoetiology of craniosynostosis gives light to uncover susceptibility individuals, discern environmental risk factors and establish effective strategies for prevention and early diagnosis. In our study, we demonstrated that a heterozygous Axin2 (c.1181G > A: p.R394H, rs200899695) mutation was presented in the monochorionic twins and their father, but not in the mother. However, only the female proband, who was received additional environmental insults (persistent breech presentation and intrauterine growth restriction), developed sagittal craniosynostosis. We assume that this Axin2 mutation predisposes to sagittal craniosynostosis but extra environmental insults are needed to initiate the disease. Prenatal risk factors, including intrauterine constraint, twin gestation, premature delivery, maternal thyroid disorders, gestational diabetes, malnutrition, virus infectious, increase the susceptibility of craniosynostosis in genetically predisposed infants [7,[11][12][13][14]. Research of monochorionic (MC) twins provide exceptional opportunity to decipher the interplay between genetic and environment risks on the occurrence of premature suture fusion [34]. In our study, monochorionic twins suffered the majority of risk factors prenatally, however, only the infant with breech presentation and intrauterine growth restriction presented sagittal craniosynostosis. Thus, intrauterine growth restriction and breech position deserves particularly attention in causing sagittal craniosynostosis.
It has been well accepted that AXIN2 is essential for normal calvarial morphogenesis by directly targeting β-catenin, orchestrating the crosstalk of Wnt, BMP, FGF signaling pathways and maintaining suture cell stemness [35][36][37]. Deletion or mutation of Axin2 attribute to craniosynostosis in humans and mice [6,36]. Moreover, phenotype data available in The International Mouse Phenotyping Consortium (IMPC, https:// www. mouse pheno type. org/) show that all female and male homozygous Axin2 knockout (Axin2 −/− ) mice present abnormal head shape malformation. For heterozygous Axin2 deletion (Axin2 +/− ) mice, a total of 2/9 females develop abnormal head shape at E18.5;    however, the male Axin2 +/− mice are not. These results indicate haploinsufficiency of Axin2 in female mice with incomplete penetrance. In this study, an Axin2 heterozygous missense mutation (c.1181G > A: p.R394H, rs200899695) was identified in peripheral blood samples of subjects I-1, II-1 and II-2 (Fig. 1a), suggesting that the proband inherits the mutation from her father. The wild type Axin2 (c.1181G) loci is conserved across species and in the region may be modified by H3K4Me3 or H3K27Ac in seven cell lines by ChIP-Seq assay. H3K4Me3 modification is known to mark genes that are essential for the identity and function. H3K4Me3 breadth contains information that ensures transcriptional precision at the identified genes [38]. H3K27Ac, a robust mark of active enhancers and promoters, has been demonstrated to be strongly correlated with gene expression and transcription-factor binding [39]. However, much more functional annotation and experiments are needed to clarify whether Axin2 (c.1181G > A: p.R394H, rs200899695) could affect the H3K4Me3 or H3K27Ac modification. In addition, Axin2 missense mutation (c.1181G > A: p.R394H, rs200899695) is likely to be deleterious by in silico predication; however, only the proband received additional risk factors (persistent breech presentation and intrauterine growth restriction) developed sagittal craniosynostosis. We assumed that phenotypic segregation in our case was probably due to Axin2 (c.1181G > A) mutation possesses incomplete penetrance, thus making it insufficiency to trigger the disease alone. Our findings corroborate another well-established gene-environment interaction model of NCS, which substantiates the same environmental insults ultimately determining phenotype [15]. However, our gene-environment interaction fashion was observed in the context of Axin2 (c.1181G > A, p.R394H, rs200899695) mutation and female individual, further clinical observations, animal and mechanistic studies are needed to validate the hypothesis.
However, there are some limitations of our study. Experiments on whether Axin2 (c.1181G > A, p.R394H, rs200899695) mutation could affect the H3K4Me3 or H3K27Ac modification of AXIN2 are needed. We only offered a plausible explanation of Axin2 (c.1181G > A, p.R394H, rs200899695) mutation and perinatal risk factors contribute to sagittal craniosynostosis in a Chinese female monochorionic diamniotic twin family. Given the small sample size and inherent ascertainment bias, the identification of further families in the setting of this mutation would help to clarify the clinical implications.

Conclusion
Based on the results of monochorionic twins, we demonstrated Axin2 (c.1181G > A, p.R394H, rs200899695) mutation led to haploinsufficiency with incomplete penetrance in female, and additional prenatal risk factors (intrauterine growth restriction and breech position) were indispensable to trigger the occurrence of sagittal craniosynostosis in this Chinese female monochorionic diamniotic twin family. These findings provide new evidence for the gene-environment interplay in understanding etiologies of NCS.