We evaluated 186 individuals with single-suture craniosynostosis for CNVs in order to identify novel candidate genes for craniosynostosis. Within our cohort, 7.5% of individuals carried a deletion or duplication involving one or more genes that has not been previously reported in unaffected individuals. We identified three individuals with CNVs >2 Mb. Although two of these are inherited, when compared to a large control cohort, we found a slight excess of events >2 Mb [3/186 (1.6%) vs. 8/2,493 (0.3%); P
= 0.036, Fisher's exact test]. We also note that all of the events >1 Mb disrupted at least one gene. By comparison, only 44% of events >1 Mb in our control set of 2,493 individuals disrupted one or more genes. Although it is possible that gene disruption, in addition to or distinct from gene dosage, may influence phenotypic outcome in craniosynostosis our data did not reach statistical significance for this hypothesis (P
= 0.06). To date, there is only one published study investigating the role of submicroscopic chromosomal rearrangements in a small series of patients with syndromic craniosynostosis using a variety of methods [Jehee et al., 2008
]. That study identified chromosome abnormalities in a large fraction of affected individuals (42%), suggesting that gene dosage may be an important mechanism in craniosynostosis. We note that the fraction of individuals with rare CNVs is substantially smaller than that found by Jehee et al. [2008
]. However, the individuals in our cohort—with isolated single-suture craniosynostosis—are probably less severely affected than those in the previous study.
The most intriguing CNV is present in two individuals in our cohort who are first cousins: a 1.1 Mb duplication encompassing the entire RUNX2
gene. Loss-of-function mutations in RUNX2
cause the autosomal dominant disorder, cleidocranial dysplasia (OMIM 119600), characterized by dysgenesis (or agenesis) of the clavicles and delayed closure of the anterior fontanelle [Mundlos et al., 1997
; Cunningham et al., 2006
]. In addition, the majority of affected individuals also have dental abnormalities including supernumerary (extra) permanent teeth in ~70% of patients. There are many lines of evidence that suggest the duplication of RUNX2
may be causative in these two individuals including the converse dental phenotype of hypodontia to that seen in individuals with heterozygous RUNX2
is a pro-osteogenic protein and is considered the principal osteogenic master switch [Lian et al., 2004
] as its activity is necessary and sufficient for osteoblast differentiation. RUNX2
is expressed in fusing cranial sutures and is upregulated in mice with heterozygous knock-in mutations of Fgfr1
and Pfeiffer-type craniosynostosis [Zhou et al., 2000
] and in patients with non-syndromic craniosynostosis [Nacamuli et al., 2003
]. RUNX2 protein function is repressed by TWIST1, mutations of which cause Saethre–Chotzen syndrome (OMIM 101400), a craniosynostosis syndrome characterized by coronal synostosis, facial asymmetry, ptosis small ears, and occasional syndactyly [el Ghouzzi et al., 1997
; Howard et al., 1997
; Krebs et al., 1997
mutations have been shown to disrupt the association of TWIST1 and RUNX2, preventing RUNX2 repression [Bialek et al., 2004
Taken together, these data suggest that increased expression of RUNX2
is associated with craniosynostosis. To our knowledge, there is only one other report of a patient with a duplication encompassing the entire RUNX2
gene [Wilkie et al., 2006
]. The patient is described with unicoronal synostosis and learning difficulties and has duplication that is significantly larger than the one that we describe (3.4 Mb, [Colella et al., 2007
]). However, this additional case report further supports the likely pathogenicity of the duplications in our two affected patients. The mother of 1007 was confirmed to be a carrier of the duplication, and the father of 1019 is presumed to carry the same duplication; neither is known to have had synostosis suggesting incomplete penetrance for that phenotype, but both are reported to have hypodontia. We propose that increased dosage of RUNX2
leads to susceptibility to premature suture fusion and tooth abnormalities, and additional genetic, epigenetic, or environmental factors influence the final phenotypic outcome.
We evaluated expression of RUNX2 in osteoblast cells from the two patients with a duplication. Although this study was underpowered due to the availability of only two affected cell lines, the average expression of RUNX2 in these samples was higher than all affected and unaffected cell lines. These data should be interpreted with caution but are supportive of a role of RUNX2 overexpression as a cause of single-suture craniosynostosis.
One child with sagittal synostosis and mild developmental delay has a 3.34 Mb duplication of chromosome 3p25. Although inherited from his unaffected father, similar duplications have not been reported in large control populations. In addition, Jehee et al. [2008
] report a smaller, inherited duplication of the same region on 3p25 in 1/45 individuals with synostosis and additional anomalies. Because the duplication was inherited, the authors concluded that it was not likely to be causative. An alternative explanation is that the region contains one or more dosage-sensitive genes important for craniosynostosis with decreased penetrance. Another patient in our series carries a ~2.5 Mb duplication of 1q43 that is maternally inherited. We are unaware of any similar duplications in published studies of control individuals. Although phenotypes associated with trisomy 1q43 have been described [Morava et al., 2004
], the region that is duplicated in those cases is much larger than the duplication in our patient.
We also identified a 4 Mb deletion of 9q22 in a young girl with metopic synostosis and developmental delay. There are several cases previously reported with overlapping deletions of this region [Boonen et al., 2005
; Redon et al., 2006
; Nowakowska et al., 2007
], and at least two of these cases also had metopic synostosis. Our case, one of the smallest of the deletions reported to date, provides additional support for the presence of a craniosynostosis gene in the region and narrows the critical region. The deletion in our case disrupts the ROR2
gene and deletes 32 additional genes. There are several genes within the deletion region that could contribute to the synostosis phenotype. Mutations in ROR2
result in autosomal dominant brachydactyly type B and in the autosomal recessive skeletal dysplasia, Robinow syndrome. Two other genes within the region are involved in bone formation: OMD
(osteomodulin) and OGN
(osteoglycin). Importantly, not all individuals with deletions encompassing the region deleted in our patient have craniosynostosis, again suggesting decreased penetrance for the phenotype.
Finally, we also detected a deletion or duplication between 35 and 730 kb in eight individuals. Each of these involves one or several genes, each of which is a potential candidate gene for craniosynostosis (). It will be necessary to evaluate additional affected individuals for deletion, duplication, and point mutations of candidate genes to determine whether they truly play a significant role in craniosynostosis. Likewise, each large event (with the exception of the RUNX2 duplication) was found in a single individual; identification of recurrent or overlapping events in additional patients will help narrow critical regions and candidate genes. We propose that changes in gene dosage of critical genes, including RUNX2, increases susceptibility to premature suture fusion.