Mutations in SEC24D cause autosomal recessive osteogenesis imperfecta

To the Editor, Osteogenesis imperfecta (OI) is a genetically heterogeneous group of disorders characterized by decreased bone mass and a propensity to fractures. To date, mutations in 17 genes have been associated with OI, with the large majority of affected individuals carrying heterozygous mutations in either COL1A1 or COL1A2. The autosomal recessively inherited forms of OI are rarer and caused by biallelic mutations in BMP1, CREB3L1, CRTAP, FKBP10, LEPRE1, PLOD2, PPIB, SERPINF1, SERPINH1, SP7, SPARC, TMEM38B, or WNT1 (1, 2). Recently, compound heterozygous mutations in SEC24D were shown to cause a novel severe form of syndromic OI additionally characterized by large skull ossification defects and craniofacial malformations including macrocephaly, frontal bossing, down-slanting palpebral fissures, midface hypoplasia, micrognathia and thereby resembling – at least in part – patients with Cole Carpenter syndrome (CLCRP2; OMIM 616294) (3). Here, we present the first patient with a more classical OI phenotype, carrying two novel compound heterozygous mutations in SEC24D. Although her anterior fontanelle remains enlarged, she has no other features in keeping with CLCRP2. The index patient is the only child to non-consanguineous Chinese parents. Short long bones were noted on prenatal ultrasound since the 18th week of gestation. She presented during the neonatal period with blue-gray sclerae, osteopenia, Wormian bones, and deformities of the long bones (Fig. 1a,b). No dysmorphic craniofacial features were present. Her coronal sutures were patent and her anterior fontanelle was enlarged (aged 13 months) (Fig. 1c2), measuring 2× 2cm at the last visit (aged 16 months). She sustained a fracture to her right proximal femur at the age of 12 months, followed by a transverse fracture of her left tibia at age 15 months (Fig. 1c,d). These fractures were surgically treated and her recovery was satisfactory. The radiological evolution of her phenotype showed a mild spontaneous improvement with time (Fig. 1a–d). After exclusion of mutations in all known OI genes, we performed trio whole-exome sequencing (WES) on the index patient and her parents using the NimbleGen SeqCap EZ Human Exome Library v2.0 enrichment kit on an Illumina HiSeq2000 sequencer (illumina, San Diego, CA, USA). Informed consent was obtained from the parents, and the study was approved by the ethics committee of the Medical Faculty of the University of Cologne. WES data analysis and filtering of variants were carried out using ‘Varbank’, the exome analysis pipeline of the Cologne Center for Genomics (CCG, University of Cologne, Germany). We applied the following WES variant filtering criteria: coverage >6 reads, a minimum quality score of 10, an allele frequency ≥25%, a minor allele frequency <0.1% in the 1000 Genomes database and not annotated in the ExAC database, the Exome Variant Server (NHLBI Exome Sequencing Project), or the in-house WES datasets of the CCG. Using these filter criteria, we identified compound heterozygous, novel mutations in SEC24D in the patient (Fig. 1e). The parents were confirmed to be heterozygous carriers and mutations were confirmed by Sanger sequencing (Fig. 1e). The c.113dupC mutation in exon 2 is predicted to cause a frameshift and premature stop codon (p.T39Nfs*16). The second mutation, c.2496G>T (p.Q832H), affects the last base of exon 19. In vitro analysis of constructs harboring the c.2496G>T mutation revealed abnormal splicing (Fig. 2). SEC24D (OMIM 607186; RefSeq NM_014822.2) is a component of the COPII complex, which is responsible for the export of several proteins from the endoplasmic reticulum (ER), including procollagen. The SEC24D mutations underlying CLCRP2 described by Garbes et al. (3) were a combination of a truncating and a missense mutation, functionally resulting in insufficient procollagen export from the ER and dilated ER tubules. Our patient has none of the CLCRP2-associated facial features and her anterior fontanelle, although enlarged, is not comparable with the dramatic skull ossification defects described in the original SEC24D patients and in corresponding animal models (3). Rather, enlarged fontanelles and delayed closure have been described in patients with several other typical OI subtypes (1, 4). Therefore, our patient’s phenotype is likely due to a milder functional effect of the identified mutations. Our report thus serves to expand both the phenotypic and the mutational spectrum of the SEC24D-associated diseases and establishes SEC24D as a further gene underlying non-CLCRP2-related autosomal recessively inherited OI.