Osteogenesis imperfecta (OI) is a heritable disorder characterised by bone fragility and increased fracture risk, ranging in severity from mild to perinatal lethal. Mutations in 17 genes have been implicated in this disorder; more than 90% of cases of OI are caused by heterozygous mutations in COL1A1 or COL1A2. Most of the remainder are recessively inherited and result from mutations in genes encoding proteins involved in the synthesis of type I collagen. SERPINH1 encodes the collagen chaperone HSP47 that binds to arginine rich sequences in the type I procollagen trimers and provides the final steps in the folding and stabilisation of the triple helical domain. Loss of both alleles in mice results in early embryonic lethality. Homozygous missense mutations in dachshunds and in one child resulted in a moderately severe form of OI with continuing fractures despite bisphosphonate treatment. In both instances, the identified missense mutations resulted in substitution of different interacting leucine residues by proline. We describe a family with non-consanguineous unaffected parents who have two children with moderate short stature, low bone density, fractures, and no evidence of dentinogenesis imperfecta. Both children are heterozygous for two mutations (allele 1; allele 2); one of each was derived from each parent. Initial study identified only a frameshift mutation in the last exon that did not lead to nonsense-mediated mRNA decay (c.1233dupT, p.Asp412*). Analysis of the expression of the two alleles in cultured cells indicated that the other allele was expressed, but was only about half as abundant as the frameshift-containing allele. High density CGH identified a 5.3kb deletion upstream from the translation start site that removed a region of DNase sensitivity identified in the ENCODE data set. This allele was inherited from the father, and the mRNA in his cells was also expressed at a low level, confirming that this domain has a regulatory function for SERPINH1. This class of mutation is rarely found in individuals with genetically determined disorders and would not be detected by whole exome sequence analysis. While whole genome sequence analysis could identify it, targeted gene studies may be the more efficient way of identification, especially in families such as this one, in which only one mutant allele is found, and appropriate cells are available to measure expression levels.
Disclosure: The authors declared no competing interests.