A novel AXIN2 mutation and perinatal risk factors contribute to sagittal craniosynostosis: evidence from a monochorionic diamniotic twin family

Background: Craniosynostosis, the premature fusion of one or more cranial sutures, affects approximately 1 in every 2000-2500 live births. The etiology of sagittal craniosynostosis, the most prevalent form of isolated craniosynostosis, involves interplay between genetic and environmental insults, such as perinatal risk factors. However, the detailed information still remains largely unknown. Methods: The proband, a female monochorionic twin diagnosed with sagittal craniosynostosis, as well as her healthy twin sister and parents, were enrolled in our study. The obstetric medical records were retrospectively reviewed. Genetic cause was investigated by whole exome sequencing (WES) and further conrmed by Sanger sequencing. Results: We identied a novel heterozygous mutation of AXIN2 (c.1181G>A) in monochorionic twins and their father. However, only the proband presented sagittal craniosynostosis. This mutation results in the replacement of Arg at residue 394 by His (p.R394H). Arg 394 is located at the GSK3β binding domain of the AXIN2 protein, which is highly conserved across species. The obstetric medical records revealed that the proband had additional persistent breech presentation and intrauterine growth restriction, except for the perinatal risk factors shared by the twins. Conclusions: Based on the role of AXIN2 in craniosynostosis development and deleterious prediction in silico of AXIN2 (c.1181G>A: p.R394H), we speculate that this particular mutation confers susceptibility to sagittal craniosynostosis, while extra environmental insults are also involved in the pathogenesis of this case. Our ndings provide a new evidence for the gene-environment interplay in understanding pathogenesis of craniosynostosis.


Introduction
Craniosynostosis (CS), de ned as premature fusion of one or more cranial sutures, affects approximately 1 in every 2000-2500 live births [1]. It usually involves as an isolated condition (non-syndromic craniosynostosis, NCS), but may also be associated with other malformations as part of complex syndromes (syndromic craniosynostosis) [2]. Sagittal craniosynostosis is the most prevalent form of NCS, which accounts for 40-58% of all NCS cases [3]. Sagittal suture premature closure restricts the widen of the skull, thus causing the scaphocephaly deformity and other adverse neurologic outcomes [4].
Monochorionic (MC) twins, sharing almost the same genome, offer a unique opportunity to study the gene/environment interactions, for the healthy twin as an ideal control. Discordant phenotypes and lack of 100% concordance between MC twins emphasize the interplay between genetic and environmental in uences in the etiology of the disease[16].
In our study, we reported a female monochorionic twin, who was diagnosed with sagittal craniosynostosis, carried a novel heterozygous mutation of AXIN2 (c.1181G>A: p.R394H). This mutation also exists in the patient's father and her monochorionic sister; however, neither of whom present craniosynostosis. Intriguingly, the obstetric history documented that the proband had a persistent breech presentation and intrauterine growth restriction. We speculate that this particular AXIN2 mutation confers susceptibility, and additional environmental insults eventually trigger the sagittal craniosynostosis to occur.

Clinic examination
All participants signed the informed consent and received physical examination in the craniofacial clinic.
Whole exome sequencing and data analysis Genomic DNA, extracted from peripheral blood samples (I-1, I-2, II-1 and II-2) and skull periosteum tissue (II-1), 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 ltered out. All variants were further annotated using the ANNOVAR tool with the information of the Exome Aggregation Consortium (ExAC) Browser, Sorting Intolerant From Tolerant (SIFT), Polymorphism Phenotyping v2 (PolyPhen-2), MutationTaster, Online Mendelian Inheritance in Man (OMIM), Gene Ontology and KEGG Pathway databases [19][20][21][22][23][24][25]. The work ow of our genetic analysis is shown in Fig. 1.

Structural analysis
Three-dimensional models of the wild-type and mutant site (p.R394H) of AXIN2 protein were designed using the I-TASSER [27]. The models were manipulated and visualized using the PyMOL software (PyMOL Molecular Graphics System, DeLano Scienti c, San Carlos, CA).

Clinical information
The pedigree came from Wuhan, Hubei Province, China. The female proband (II-1), the elder monochorionic diamniotic (MCDA) twin ( Fig. 2a-c and Addition le: Fig. S1-3), was diagnosed with sagittal craniosynostosis 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) (Addition le: Fig. S1-3). Her father was 36-year-old at that time. The 26 days after FET, four-dimensional ultrasound scan con rmed two embryos were inside a gestational sac (approximately 22 mm × 13 mm); one embryo was 2.8 mm in length and the fetal heart rate was 107 per minute, the other embryo was 3.3 mm in length and the fetal heart rate was 118 per minute (Addition le: Fig. S1).
At 13 weeks of gestation, IgG of cytomegalovirus and herpes simplex virus were detected in the mother's serum by the toxoplasmosis, rubella, cytomegalovirus and herpes simplex (TORCH) test (Addition le: Table S1). At 14 weeks of gestations, maternal dietary assessment indicated that the mother had inadequate intakes of energy, protein, fat, vitamins and several other minerals (Addition le: Table S2); and non-invasive prenatal testing demonstrated low risks for Trisomy 13, 18 and 21 (Addition le: Table  S3). At 17 weeks of gestations, elevated levels of thyroid-stimulating hormone (TSH) and urinary iodine were detected (Addition le: Table S4). At 27 weeks of gestations, the mother was diagnosed with gestational diabetes mellitus (fasting blood glucose level, 7.38 mmol/L; postprandial one-hour blood glucose level, 15.30 mmol/L; postprandial two-hour blood glucose level, 11.03 mmol/L).
At 28 weeks of gestations, the proband (II-1) and her twin younger sister (II-2) were born via spontaneous vaginal delivery (Fig. 2a, b). The proband (II-1), who had been in persistent breech position on the left side of mother's uterus (Fig. 2b), was rst recorded by ultrasound scan at 23.5 weeks of gestation; her birth weight was 880 g (50-90 centile[28]). The younger sister (II-2), who had been in cephalic position on the right side of mother's uterus (Fig. 2b), weighted 990 g (90-97 centile[28]) at birth. Only the proband presented the typical appearance of sagittal craniosynostosis (Fig. 2c). While no signs of craniofacial deformity were found in her parents (I-1, I-2) and sister (II-2) (Fig. 2c). Cranial index, which represents the ratio of maximum cranial width to maximum cranial length [29], is decreased in the head shape of sagittal craniosynostosis. In our case, the cranial index of the proband and her sister are 70.9%, 83.2%, respectively. Taken together, the clinical ndings indicate that the proband, who developed sagittal craniosynostosis, had additional persistent breech presentation and intrauterine growth restriction, except for the perinatal risk factors shared by the twins.

Mutation analysis
The whole-exome sequencing (WES) was applied to uncover 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-speci c. As geneenvironment interactions have been revealed in the pathogenesis of craniosynostosis [5,15], we wonder whether the adverse intrauterine exposures (environment) triggered the susceptible individual (II-1) to develop sagittal craniosynostosis. Based on this hypothesis, we re-analyzed our sequencing data and identi ed a heterozygous missense mutation of AXIN2 (c.1181G>A, p.R394H) in the leukocytes of subjects I-1, II-1 and II-2, and skull periosteum tissue of subject II-1. The ndings were further validated by Sanger sequencing (Fig. 3a-b). The nucleotide sequence showed a heterozygous G to A transition at nucleotide 1181 (c.1181 G>A) of the coding sequence, which resulted in the replacement of Arg at residue 394 by His of AXIN2 protein.

Conservation analysis
Arg394, located at the GSK3β binding domain (amino acid 327 to 413 according to the UniProt Consortium) of the AXIN2 protein (Fig. 3c) [30], is conserved across species (Fig. 3d, e) and was predicted to be potentially deleterious by in silico analysis (Table 1).
Structural analysis I-TASSER structural analysis predicted that the p. R394H substitution may affect the spatial structure of the GSK3β binding domain in AXIN2 protein (Fig. 4 a, b). Taken together, our results indicate that AXIN2 (c.1181G>A, p.R394H) mutation may confer deleterious of sagittal craniosynostosis.

Discussion
Craniosynostosis, a highly heterogeneous disease, is caused by genetic mutations, adverse exposures as well as their interactions. The study of gene/environment interactions provides the foundation for clarifying the pathogenesis of craniosynostosis, identifying susceptibility individuals, discerning modi able environmental risk factors and establishing effective strategies for prevention and early diagnosis. In our study, we demonstrated that a novel heterozygous AXIN2 (c.1181G>A: p.R394H) mutation was presented in three members of the family, including the monochorionic twins and their father. However, only the female proband with an additional environmental insult, that is, persistent breech presentation and intrauterine growth restriction in our case, developed overt craniosynostosis. We assume that this AXIN2 mutation predisposes to sagittal craniosynostosis but needs extra environmental insults to initiate the disease.
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 [31][32][33]. Deletion or mutation of AXIN2 attribute to premature suture closure and craniosynostosis in humans and mice [6,32]. Moreover, phenotype data available in International Mouse Phenotyping Consortium (IMPC, https://www.mousephenotype.org/) show that all the female and male homozygous AXIN2 knockout mice present calvarial malformation. For heterozygous AXIN2 deletion mice, only 2/7 females develop an abnormal head shape; however, the males are not observed any noticeable distinctions, which indicates the incomplete penetrance of AXIN2 mutations. In this study, an AXIN2 heterozygous missense mutation (c.1181G>A) was identi ed in peripheral blood samples of subjects I-1, II-1 and II-2 (Fig. 1a), which indicates that the proband inherits the mutation from her father. The function of AXIN2 (c.1181G>A) has not been reported according to the lasted updates in ClinVar (Record last updated Dec 17, 2019, https://www.ncbi.nlm.nih.gov/clinvar/variation/234485/). Our ndings indicate that this mutation is conserved across species and was likely to be deleterious by in silico prediction tools; however, only the proband suffered sagittal craniosynostosis. We assumed that the reason of phenotypic segregation in our case is probably because AXIN2 (c.1181G>A) mutation possesses incomplete penetrance, thus making it insu ciency to trigger the disease alone. However, further clinical observations and animal studies are still needed to validate our hypothesis.
Prenatal risk factors, such as maternal thyroid disorders, gestational diabetes, drug use, malnutrition, virus infectious, intrauterine constraint, twin gestation, premature delivery, are associated with craniosynostosis and likely increase the susceptibility in an already genetically predisposed infant [7,[11][12][13][14]. Monochorionic (MC) twins provide exceptional chances to decipher the interplay among genetic and environment in uences in the pathogenesis of premature suture fusion [34]. In our study, the mother of monochorionic twins suffered the majority of risk factors prenatally, however, only the infant with breech presentation and intrauterine growth restriction experienced sagittal craniosynostosis. Our ndings corroborate another well-established gene-environment interaction model of NCS, which substantiates almost the same environmental insults ultimately determining phenotype [15]. In all, we highlight that the intrauterine growth restriction and breech position, deserves particularly attention in predicting sagittal craniosynostosis occurrence.
However, our gene-environment interaction fashion was observed in the context of AXIN2 (c.1181G>A) mutation with small sample size and unforeseen bias, which is unlikely to provide accurate assessments for all types NCS. And we also cannot evaluate the epigenetic role in pathogenesis of this case. Further epidemiological and experimental studies are encouraged to give a more unambiguous pattern.

Conclusion
In summary, we described a monochorionic twin with AXIN2 (c.1181G>A) mutation suffered sagittal craniosynostosis in the insult of prenatal risk factors (intrauterine growth restriction and breech position). Our ndings provide a new evidence for the gene-environment interplay in etiology of NCS, which will be informative in the molecular diagnosis and genetic counselling in clinic.    Figure 1 Work ow for the genetic analyses using WES and Sanger sequencing in pedigree diagnosed with sagittal craniosynostosis.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.