Abstract
In Drosophila, mutations in the Twist gene interact with mutations in the Snail gene. We show that the mouse Twist1 mutation interacts with Snai1 and Snai2 mutations to enhance aberrant cranial suture fusion, demonstrating that genetic interactions between genes of the Twist and Snail families have been conserved during evolution.
CRANIOSYNOSTOSIS, the premature fusion of the cranial sutures, is a significant medical problem, occurring in both syndromic and nonsyndromic forms (Wilkie 1997; Wilkie and Morriss-Kay 2001). Saethre-Chotzen syndrome (OMIM 101400) is one of the most common syndromic forms of craniosynostosis in humans. Clinical features of Saethre-Chotzen syndrome include craniosynostosis most commonly affecting the coronal suture (although other sutures can be affected), dysmorphic facial features (e.g., facial asymmetry, hypertelorism, and maxillary hypoplasia), brachydactyly, and cutaneous syndactyly (Jabs 2004). The locus for Saethre-Chotzen syndrome was mapped to chromosome 7p21-p22, and mutations in the TWIST1 (formerly TWIST) gene were identified in these individuals (el Ghouzzi et al. 1997; Howard et al. 1997; Rose et al. 1997). The TWIST1 gene encodes a basic helix-loop-helix transcription factor. The types of TWIST1 mutations found in these patients (insertions and deletions, nonsense mutations, and missense mutations in the DNA-binding domain) suggest that these mutations would lead to the production of truncated or nonfunctional TWIST1 protein. Microdeletions that encompass the entire TWIST1 gene have also been identified in Saethre-Chotzen syndrome patients (Johnson et al. 1998). These data indicate that the predominant cause of Saethre-Chotzen syndrome is haplo-insufficiency for the TWIST1 gene.
A null mutation in the mouse Twist1 gene results in embryonic lethality by 11.5 days of gestation in homozygotes (Chen and Behringer 1995) and in heterozygotes results in partially penetrant skeletal defects that replicate certain features of Saethre-Chotzen syndrome (el Ghouzzi et al. 1997; Bourgeois et al. 1998). We recently showed that, similarly to human patients with TWIST1 mutations, Twist1+/− heterozygous mice exhibit craniosynostosis primarily affecting the coronal sutures (Carver et al. 2002). The mammalian TWIST1 gene is a homolog of the Twist gene of Drosophila. Genetic evidence in Drosophila has demonstrated that mutations in the Twist gene interact with mutations in the Snail gene, which encodes a zinc-finger transcriptional repressor (Nieto 2002). The mesodermal phenotype of Snail/Twist double-homozygous mutant Drosophila embryos is much stronger than that of either single mutant. In the double mutants, no ventral furrow forms and the ventral epithelium of the embryos is indistinguishable from the neighboring ectoderm (Leptin 1991; Ip et al. 1992).
We previously described null mutations of two mouse Snail gene family members, Snai1 (formerly Snail) (Jiang et al. 1998) and Snai2 (formerly Slug) (Carver et al. 2001). Here we describe genetic interactions between the Twist1 mutation and the Snai1 and Snai2 mutations during cranial suture formation in mice. Twist1+/− Snai1+/− and Twist1+/− Snai2+/− double-heterozygous mice exhibit an enhanced incidence and severity of craniosynostosis compared to Twist1+/− single heterozygotes. We also describe a scoring matrix that provides a quantitative assessment of craniosynostosis in individual mice. Our results demonstrate that mutations of Snail family genes act as dominant enhancers of Twist1 haplo-insufficiency in mice and show that genetic interactions between genes of the Twist and Snail families have been conserved during evolution.
Enhanced craniosynostosis in Twist1+/−Snai1+/− and Twist1+/−Snai2+/− double-heterozygous mice:
Unlike humans, in rodents all but one of the cranial sutures remain patent (i.e., not fused) throughout the life span of the animal (Opperman 2000; Warren and Longaker 2001). Only the posterior portion of the interfrontal (metopic) suture normally fuses shortly after birth. We recently showed that Twist1+/− heterozygous mice exhibit partially penetrant craniosynostosis (Carver et al. 2002). Both complete and partial coronal suture fusion was observed in Twist1+/− mice, and fusion was either bilateral or unilateral. In addition, skulls exhibiting coronal craniosynostosis usually exhibited a second suture defect, a fusion of the distal portions of the occipitointerparietal suture separating the interparietal and supraoccipital bones (Carver et al. 2002).
Genetic evidence in Drosophila has demonstrated that mutations in the Twist gene interact with mutations in the Snail gene. To assess whether genetic interactions between the Twist1 gene and Snail family genes are evolutionarily conserved, we crossed Twist1+/− heterozygous mice to mice heterozygous for mutations of the Snail family genes Snai1 and Snai2. Alizarin-red/alcian-blue-stained skeletons were prepared from wild type, Twist1+/−, Snai1+/−, Snai2+/−, Twist1+/− Snai1+/−, and Twist1+/− Snai2+/− mice. Analysis of the skulls of these mice revealed a spectrum of cranial suture defects similar to what we observed previously in Twist1+/− heterozygotes (Figure 1). Most Twist1+/− Snai1+/− and Twist1+/− Snai2+/− double-heterozygous skulls appeared to be more severely affected than Twist1+/− skulls. However, to accurately assess to what degree craniosynostosis was enhanced in Twist1+/− Snai1+/− and Twist1+/− Snai2+/− double-heterozygous mice compared to Twist1+/− single heterozygotes, we required a quantitative measure of craniosynostosis in individual mice. Therefore, we devised a scoring matrix in which a suture was assigned a value from 0 (completely unfused) to 3 (completely fused; see Table 1 legend for details). Four sutures were scored for each skull—left coronal, right coronal, interfrontal, and occipitointerparietal—and each individual animal received a composite score that was the sum of the scores for the four sutures. The scoring matrix mean was determined for each genotype and was termed the craniosynostosis index (CI).
Mutations in the Snai1 and Snai2 genes enhance craniosynostosis of Twist1+/− mice. The unfused interfrontal, coronal, and sagittal sutures are indicated in the skull of a wild-type littermate (A). Aberrant suture fusion is typically more severe in Twist1+/− Snai1+/− (E) and Twist1+/− Snai2+/− (F) double-heterozygous mice than in any of the single heterozygotes (B–D). Skulls were prepared as described previously (Carver et al. 2002). Official nomenclature and references for the mutant mouse strains used in these studies are Twist1tm1Bhr (Chen and Behringer 1995), Snai1tm1Grid (Carver et al. 2001), and Snai2tm1Grid (Jiang et al. 1998). Mice were analyzed on a mixed C57BL/6J × 129S1/SvImJ background.
Craniosynostosis in progeny of the Twist1+/− × Snai1+/− cross
Use of this scoring matrix made it obvious that craniosynostosis was enhanced substantially in Twist1+/− Snai1+/− and Twist1+/− Snai2+/− double-heterozygous mice, compared to Twist1+/− mice. The CI of Twist1+/− heterozygotes in both crosses was 2.6. The CI of Twist1+/− Snai1+/− and Twist1+/− Snai2+/− double heterozygotes increased to 5.5 and 6.9, respectively (Tables 1 and 2), indicating that the double-heterozygous mice exhibited craniosynostosis of increased severity. The penetrance of the craniosynostosis phenotype was also increased in the double-heterozygous mice (Tables 1 and 2). One hundred percent of both Twist1+/− Snai1+/− and Twist1+/− Snai2+/− mice exhibited some form of cranial suture fusion. These data demonstrate that genetic interactions between genes of the Twist and Snail families have been conserved during evolution and that mutations in the mouse Snail family genes Snai1 and Snai2 act as dominant enhancers of the craniosynostosis phenotype of Twist1+/− haplo-insufficient mice. Our data revealing genetic interactions between mouse Snai1 and Twist1 mutations are supported by recent work demonstrating that the SNAI1 and TWIST1 proteins act cooperatively to inhibit expression of the p21WAF/Cip1 gene in a human osteoblast-like cell line (Takahashi et al. 2004).
Craniosynostosis in progeny of the Twist1+/− × Snai2+/− cross
Polydactyly in Twist1+/−, Twist1+/−Snai1+/−, and Twist1+/−Snai2+/− mice:
Another phenotype exhibited by Twist1+/− haplo-insufficient mice is preaxial polydactyly of one or both hind feet (el Ghouzzi et al. 1997; Bourgeois et al. 1998). We did not observe polydactyly or any other limb abnormalities in wild type, Snai1+/−, or Snai2+/− mice. Polydactyly was observed in mice heterozygous for the Twist1 mutation (Figure 2). Combining the data for the Twist1 × Snai1 and Twist1 × Snai2 crosses, the penetrance of polydactyly was 50% in Twist1+/− mice (n = 52), 48% in Twist1+/− Snai1+/− (n = 23) mice, and 22% in Twist1+/− Snai2+/− mice (n = 18). However, there was an increased incidence of animals containing two polydactylous hind feet in Twist1+/− single heterozygotes, compared to Twist1+/− Snai1+/− or Twist1+/− Snai2+/− double-heterozygous mice (Figure 2).
Hind-feet polydactyly in Twist1+/−, Twist1+/− Snai1+/−, and Twist1+/− Snai2+/− mice. (A) Twist1+/−. (B) Twist1+/− Snai1+/−. (C) Twist1+/− Snai2+/−. The severity and penetrance of polydactyly was similar in these three genotypes. However, there was an increased incidence of animals containing two polydactylous hind feet in Twist1+/− mice.
Implications for Saethre-Chotzen syndrome:
Two groups have analyzed cohorts of patients exhibiting both syndromic and nonsyndromic craniosynostosis for mutations in the SNAI1 gene (Paznekas et al. 1999; Twigg and Wilkie 1999). No mutations were detected in 127 patients examined in these studies. No studies reporting the analysis of SNAI2 mutations in patients with craniosynostosis have been published. However, homozygous deletions of the SNAI2 gene have been observed in two patients with Waardenburg syndrome type 2 (Sanchez Martin et al. 2002), and heterozygous deletions have been observed in three patients with piebaldism (Sanchez Martin et al. 2003). A SNAI2 missense mutation also was detected in 1 of 150 unrelated spina bifida patients (Stegmann et al. 1999). This mutation caused the exchange of a conserved amino acid in the region upstream of the first zinc finger of the SNAI2 protein. However, the same mutation was found in the proband's unaffected father, leading to uncertainty as to whether the observed SNAI2 mutation was causally related to the neural tube defect.
While there currently is no evidence that mutations in the SNAI1 and SNAI2 genes cause craniosynostosis in humans, no studies in human populations have examined whether mutations in the SNAI1 or SNAI2 genes can modify the penetrance or expressivity of Saethre-Chotzen syndrome or other forms of syndromic or nonsyndromic craniosynostosis. It is possible that genetic interactions similar to those observed in Twist1+/− Snai1+/− and Twist1+/− Snai2+/− double-heterozygous mice may occur in human Saethre-Chotzen syndrome patients and that particular SNAI1 or SNAI2 alleles may influence the penetrance or severity of Saethre-Chotzen syndrome phenotypes. This hypothesis could be tested by determining whether segregation of different SNAI1 or SNAI2 alleles in families with Saethre-Chotzen syndrome correlates with phenotypic severity. Future study of Twist1+/− Snai1+/− and Twist1+/− Snai2+/− double-heterozygous mice may lead to additional insights into the pathogenesis of craniosynostosis in humans.
Acknowledgments
We thank Weidong Zhang of the Computational Biology Resource for help with the statistical analysis. This work was supported by a grant from the National Institutes of Health (NIH; HD34883) to T.G. and a subcontract to T.G. under NIH Project Center grant DE13078 to Johns Hopkins University. This work was also supported by a grant from the National Cancer Institute (CA34196) to the Jackson Laboratory.
Footnotes
Communicating editor: N. A. Jenkins
- Received January 28, 2005.
- Accepted February 17, 2005.
- Genetics Society of America
