Osteogenesis imperfecta (OI) is a clinically and genetically heterogeneous disorder associated with bone fragility and susceptibility to fractures after minimal trauma. OI type V has an autosomal-dominant pattern of inheritance and is not caused by mutations in the type I collagen genes COL1A1 and COL1A2. The most remarkable and pathognomonic feature, observed in ∼65% of affected individuals, is a predisposition to develop hyperplastic callus after fractures or surgical interventions. To identify the molecular cause of OI type V, we performed whole-exome sequencing in a female with OI type V and her unaffected parents and searched for de novo mutations. We found a heterozygous de novo mutation in the 5′-untranslated region of IFITM5 (the gene encoding Interferon induced transmembrane protein 5), 14 bp upstream of the annotated translation initiation codon (c.−14C>T). Subsequently, we identified an identical heterozygous de novo mutation in a second individual with OI type V by Sanger sequencing, thereby confirming that this is the causal mutation for the phenotype. IFITM5 is a protein that is highly enriched in osteoblasts and has a putative function in bone formation and osteoblast maturation. The mutation c.−14C>T introduces an upstream start codon that is in frame with the reference open-reading frame of IFITM5 and is embedded into a stronger Kozak consensus sequence for translation initiation than the annotated start codon. In vitro, eukaryotic cells were able to recognize this start codon, and they used it instead of the reference translation initiation signal. This suggests that five amino acids (Met-Ala-Leu-Glu-Pro) are added to the N terminus and alter IFITM5 function in individuals with the mutation. Osteogenesis imperfecta (OI) is a clinically and genetically heterogeneous disorder associated with bone fragility and susceptibility to fractures after minimal trauma. OI type V has an autosomal-dominant pattern of inheritance and is not caused by mutations in the type I collagen genes COL1A1 and COL1A2. The most remarkable and pathognomonic feature, observed in ∼65% of affected individuals, is a predisposition to develop hyperplastic callus after fractures or surgical interventions. To identify the molecular cause of OI type V, we performed whole-exome sequencing in a female with OI type V and her unaffected parents and searched for de novo mutations. We found a heterozygous de novo mutation in the 5′-untranslated region of IFITM5 (the gene encoding Interferon induced transmembrane protein 5), 14 bp upstream of the annotated translation initiation codon (c.−14C>T). Subsequently, we identified an identical heterozygous de novo mutation in a second individual with OI type V by Sanger sequencing, thereby confirming that this is the causal mutation for the phenotype. IFITM5 is a protein that is highly enriched in osteoblasts and has a putative function in bone formation and osteoblast maturation. The mutation c.−14C>T introduces an upstream start codon that is in frame with the reference open-reading frame of IFITM5 and is embedded into a stronger Kozak consensus sequence for translation initiation than the annotated start codon. In vitro, eukaryotic cells were able to recognize this start codon, and they used it instead of the reference translation initiation signal. This suggests that five amino acids (Met-Ala-Leu-Glu-Pro) are added to the N terminus and alter IFITM5 function in individuals with the mutation. Osteogenesis imperfecta (OI [MIM 166200, 166210, 259420, 166220, 610967, 613982, 610682, 610915, 259440, 613848, 610968, and 613849 for type I to XII of the disease]) is a clinically and genetically heterogeneous disorder associated with bone fragility and susceptibility to fractures after minimal trauma. Some individuals with OI also have hearing loss, dentinogenesis imperfecta, hypermobility of joints, and/or blue sclera. The original Sillence classification, introduced in 1979, uses clinical and radiological features to differentiate between four types: OI type I (mild nondeforming, with blue sclera), type II (perinatal lethal), type III (progressive deforming), and type IV (moderately deforming, with normal sclera).1Sillence D.O. Senn A. Danks D.M. Genetic heterogeneity in osteogenesis imperfecta.J. Med. Genet. 1979; 16: 101-116Crossref PubMed Scopus (1563) Google Scholar The majority of individuals with the clinical diagnosis OI types I–IV have heterozygous mutations in one of the two genes encoding the α chains of collagen type 1, COL1A1 (MIM 120150) and COL1A2 (MIM 120160). OI types I–IV are inherited in an autosomal-dominant manner, and the mutations result in quantitative and/or qualitative defects in type 1 collagen production by osteoblasts.2Barsh G.S. Byers P.H. Reduced secretion of structurally abnormal type I procollagen in a form of osteogenesis imperfecta.Proc. Natl. Acad. Sci. USA. 1981; 78: 5142-5146Crossref PubMed Scopus (93) Google Scholar, 3Byers P.H. Tsipouras P. Bonadio J.F. Starman B.J. Schwartz R.C. Perinatal lethal osteogenesis imperfecta (OI type II): A biochemically heterogeneous disorder usually due to new mutations in the genes for type I collagen.Am. J. Hum. Genet. 1988; 42: 237-248PubMed Google Scholar, 4Chu M.L. Williams C.J. Pepe G. Hirsch J.L. Prockop D.J. Ramirez F. Internal deletion in a collagen gene in a perinatal lethal form of osteogenesis imperfecta.Nature. 1983; 304: 78-80Crossref PubMed Scopus (104) Google Scholar, 5Williams C.J. Prockop D.J. Synthesis and processing of a type I procollagen containing shortened pro-alpha 1(I) chains by fibroblasts from a patient with osteogenesis imperfecta.J. Biol. Chem. 1983; 258: 5915-5921Abstract Full Text PDF PubMed Google Scholar A substantial number of persons with OI do not have a mutation in one of the collagen genes. Many of these are individuals with recessive mutations in a growing list of genes that encode proteins involved in, among other activities, the posttranslational processing or modification of type 1 collagen (CRTAP [MIM 605497],6Morello R. Bertin T.K. Chen Y. Hicks J. Tonachini L. Monticone M. Castagnola P. Rauch F. Glorieux F.H. Vranka J. et al.CRTAP is required for prolyl 3- hydroxylation and mutations cause recessive osteogenesis imperfecta.Cell. 2006; 127: 291-304Abstract Full Text Full Text PDF PubMed Scopus (413) Google Scholar LEPRE1 [MIM 610339],7Cabral W.A. Chang W. Barnes A.M. Weis M. Scott M.A. Leikin S. Makareeva E. Kuznetsova N.V. Rosenbaum K.N. Tifft C.J. et al.Prolyl 3-hydroxylase 1 deficiency causes a recessive metabolic bone disorder resembling lethal/severe osteogenesis imperfecta.Nat. Genet. 2007; 39: 359-365Crossref PubMed Scopus (351) Google Scholar PPIB [MIM 123841],8Barnes A.M. Carter E.M. Cabral W.A. Weis M. Chang W. Makareeva E. Leikin S. Rotimi C.N. Eyre D.R. Raggio C.L. Marini J.C. Lack of cyclophilin B in osteogenesis imperfecta with normal collagen folding.N. Engl. J. Med. 2010; 362: 521-528Crossref PubMed Scopus (137) Google Scholar and BMP1 [MIM 112264]9Asharani P.V. Keupp K. Semler O. Wang W. Li Y. Thiele H. Yigit G. Pohl E. Becker J. Frommolt P. et al.Attenuated BMP1 function compromises osteogenesis, leading to bone fragility in humans and zebrafish.Am. J. Hum. Genet. 2012; 90: 661-674Abstract Full Text Full Text PDF PubMed Scopus (163) Google Scholar, 10Martínez-Glez V. Valencia M. Caparrós-Martín J.A. Aglan M. Temtamy S. Tenorio J. Pulido V. Lindert U. Rohrbach M. Eyre D. et al.Identification of a mutation causing deficient BMP1/mTLD proteolytic activity in autosomal recessive osteogenesis imperfecta.Hum. Mutat. 2012; 33: 343-350Crossref PubMed Scopus (162) Google Scholar), the final quality control of procollagen formation (SERPINH1 [MIM 600943]11Christiansen H.E. Schwarze U. Pyott S.M. AlSwaid A. Al Balwi M. Alrasheed S. Pepin M.G. Weis M.A. Eyre D.R. Byers P.H. Homozygosity for a missense mutation in SERPINH1, which encodes the collagen chaperone protein HSP47, results in severe recessive osteogenesis imperfecta.Am. J. Hum. Genet. 2010; 86: 389-398Abstract Full Text Full Text PDF PubMed Scopus (256) Google Scholar and FKBP10 [MIM 607063)]12Alanay Y. Avaygan H. Camacho N. Utine G.E. Boduroglu K. Aktas D. Alikasifoglu M. Tuncbilek E. Orhan D. Bakar F.T. et al.Mutations in the gene encoding the RER protein FKBP65 cause autosomal-recessive osteogenesis imperfecta.Am. J. Hum. Genet. 2010; 86: 551-559Abstract Full Text Full Text PDF PubMed Scopus (232) Google Scholar), or osteoblast differentiation (SP7/OSX [MIM 606633]13Lapunzina P. Aglan M. Temtamy S. Caparrós-Martín J.A. Valencia M. Letón R. Martínez-Glez V. Elhossini R. Amr K. Vilaboa N. Ruiz-Perez V.L. Identification of a frameshift mutation in Osterix in a patient with recessive osteogenesis imperfecta.Am. J. Hum. Genet. 2010; 87: 110-114Abstract Full Text Full Text PDF PubMed Scopus (215) Google Scholar).14van Dijk F.S. Byers P.H. Dalgleish R. Malfait F. Maugeri A. Rohrbach M. Symoens S. Sistermans E.A. Pals G. EMQN best practice guidelines for the laboratory diagnosis of osteogenesis imperfecta.Eur. J. Hum. Genet. 2012; 20: 11-19Crossref PubMed Scopus (99) Google Scholar The molecular pathomechanism leading to OI in individuals with recessive mutations in SERPINF1 (MIM 172860) is currently unclear.15Becker J. Semler O. Gilissen C. Li Y. Bolz H.J. Giunta C. Bergmann C. Rohrbach M. Koerber F. Zimmermann K. et al.Exome sequencing identifies truncating mutations in human SERPINF1 in autosomal-recessive osteogenesis imperfecta.Am. J. Hum. Genet. 2011; 88: 362-371Abstract Full Text Full Text PDF PubMed Scopus (271) Google Scholar, 16Homan E.P. Rauch F. Grafe I. Lietman C. Doll J.A. Dawson B. Bertin T. Napierala D. Morello R. Gibbs R. et al.Mutations in SERPINF1 cause osteogenesis imperfecta type VI.J. Bone Miner. Res. 2011; 26: 2798-2803Crossref PubMed Scopus (143) Google Scholar In 2000, a COL1A1/2-mutation-negative group of individuals with autosomal-dominant OI was delineated. These individuals had been classified as having OI type IV, but had some clinical features clearly distinguishing them from other persons with OI.17Glorieux F.H. Rauch F. Plotkin H. Ward L. Travers R. Roughley P. Lalic L. Glorieux D.F. Fassier F. Bishop N.J. Type V osteogenesis imperfecta: A new form of brittle bone disease.J. Bone Miner. Res. 2000; 15: 1650-1658Crossref PubMed Scopus (376) Google Scholar They were then categorized as having OI type V. The most remarkable and pathognomonic feature of OI type V, observed in approximately 65% of affected individuals, is a predisposition to develop a hyperplastic callus after fractures or surgical interventions, predominantly in long bones.18Cheung M.S. Glorieux F.H. Rauch F. Natural history of hyperplastic callus formation in osteogenesis imperfecta type V.J. Bone Miner. Res. 2007; 22: 1181-1186Crossref PubMed Scopus (62) Google Scholar Other features are a radiodense band at the growth plate and interosseus membrane calcification of the forearms.17Glorieux F.H. Rauch F. Plotkin H. Ward L. Travers R. Roughley P. Lalic L. Glorieux D.F. Fassier F. Bishop N.J. Type V osteogenesis imperfecta: A new form of brittle bone disease.J. Bone Miner. Res. 2000; 15: 1650-1658Crossref PubMed Scopus (376) Google Scholar These characteristics are not consistently found in all individuals with OI type V. On polarized light microscopy in individuals with OI type V, an irregular pattern of bone lamellation is seen. Dentinogenesis imperfecta and blue sclera have not been observed, and there is no evidence for altered collagen migration in gel electrophoresis. The pathogenesis of OI type V is unknown. In our specialized unit for children with OI at the Children's Hospital of the University of Cologne, we have treated one girl (proband 1) and one boy (proband 2) with the classical clinical diagnosis OI type V. Their diagnoses were based on the fact that both had displayed radiologically confirmed episodes of hyperplastic callus formation after low-trauma fractures (Figure 1). In proband 1, there was also a metaphyseal band visible on the distal femur; such a band is another typical radiological finding in individuals with OI V (Figure 1C). In proband 2, callus formation in the tibia developed without a severe fracture, and a biopsy excluded an osteosarcoma. The histological report confirmed reactive callus tissue without signs of malignancy. Neither of the children had a calcified interosseus membrane of the forearm. The two children exhibited similar clinical symptoms and disease severity, initially classified as OI type IV. There were no intrauterine fractures reported, and birth length and weight were normal. Dentinogenesis imperfecta was not present, and there were no clinical signs of hearing impairment. The sclerae were grayish, and no ophthalmological problems were detected. Both children had recurrent fractures of long bones and vertebral compression fractures with resulting wedge-shaped vertebrae, indicating generally reduced bone stability. Bone mineral density was below the normal range, and both children developed slight bowing of the extremities and short stature during childhood. They were able to walk short distances. For longer distances, they were dependent on a wheelchair or walking aids. The clinical findings with regard to the two children are summarized in Table 1. No organ system other than the skeletal system was affected. We did not perform biochemical collagen analyses or molecular tests to exclude mutations in genes known to be involved in the pathogenesis of other OI types. The parents of the affected children were healthy, nonconsanguineous, and of German descent, and they had no overt signs of reduced bone mineral density and no history of previous fractures. DXA measurements have not been performed in the parents. There were no other family members affected by OI. Thus, the family histories were compatible with underlying heterozygous de novo mutations in an as-yet-unknown autosomal gene of the two affected children.Table 1Clinical Features of Two Children with Osteogenesis Imperfecta Type VClinical FindingsProband 1Proband 2Disease severitymoderatemoderateOI typeaAccording to the expanded Sillence classification.17VVAge at first presentation1 year 7 months8 years 7 monthsTime of follow up3 years 2 months6 years 6 monthsHyperplastic-callus formations after fractures before bisphosphonate treatmentyesyesAge at start of bisphosphonate treatment1 year 7 months (iv. neridronate)9 years 9 months (oral residronate)Birth length and birth weightnormalnormalConfirmed prenatal fracturesnonoAge at first nontraumatic fracture5 months18 monthsColor of scleranormalnormalDentinogenesis imperfectanonoHypermobility of jointsnonoHearing impairmentnonoOld fractures of extremitiescAt first presentation.yesyesSpine abnormalitiescAt first presentation.wedge-shaped vertebraebiconcave vertebraeScoliosisnonoCalcified membrane interosseanonoMetaphyseal bandsyesnoAnterior dislocation of the radial headnonoSevere bowing of extremities (not due to fractures)nonoAge and region of first fracture with hyperplastic callus formation8 months, right femur1 year 10 months, forearmReduced joint mobility due to callus formationsyesyesWeight at first visit kg/BMI (SD) [age]7.3/15.3 (−0.8) [1.6 years]25.8/16.5 (0.1) [8.6 years]Weight at last visit kg/BMI (SD) [age]12.9/14.0(−1.0) [4.8 years]45.0/18.9 (−0.4)[15.1 years]Length at first visit cm (SD) [age]69 (−2.1) [1.6 years]123 (−1.8) [1.6 years]Length at last visit cm (SD) [age]96 (−2.8) [4.8 years]154 (−2.4) [15.1 years]Retarded gross motor functionsdelayed standing and walkingnoMobility at last visit if not limited by fractures (BAMFbBrief assessment of motor function.19)79IntelligencenormalnormalSerum calcium levelcAt first presentation.normalnormalSerum alkaline phosphatasecAt first presentation.normalelevatedSerum procollagen type 1 C-peptidecAt first presentation. (marker for osteoblastic activity)not measurednormalUrinary deoxypyridinolinecAt first presentation. (marker for osteoclastic activity)elevatedelevatedFirst available bone density, DXA ap spine (z score) before bisphosphonate treatment−3.3−2.8Last available bone density, DXA ap spine (z score) during bisphosphonate treatment−1.9−3.0Period between DXA measurements2 years 2 months6 years 6 monthsAbbreviations are as follows: DXA, dual-energy X-ray absorptiometry; ap, anterior-posterior; BMI, body mass index; SD, standard deviation.a According to the expanded Sillence classification.17Glorieux F.H. Rauch F. Plotkin H. Ward L. Travers R. Roughley P. Lalic L. Glorieux D.F. Fassier F. Bishop N.J. Type V osteogenesis imperfecta: A new form of brittle bone disease.J. Bone Miner. Res. 2000; 15: 1650-1658Crossref PubMed Scopus (376) Google Scholarb Brief assessment of motor function.19Cintas H.L. Siegel K.L. Furst G.P. Gerber L.H. Brief assessment of motor function: reliability and concurrent validity of the Gross Motor Scale.Am. J. Phys. Med. Rehabil. 2003; 82: 33-41Crossref PubMed Scopus (29) Google Scholarc At first presentation. Open table in a new tab Abbreviations are as follows: DXA, dual-energy X-ray absorptiometry; ap, anterior-posterior; BMI, body mass index; SD, standard deviation. To identify the molecular cause of OI type V, we decided to perform whole-exome sequencing in proband 1 and her unaffected parents and to search for de novo mutations. The study was approved by the ethics committee of the Ludwig-Maximilians University of Munich, and written informed consent was obtained from the probands' parents. We used Agilent SureSelectXT All Exon V4 target enrichment and 100 bp paired-end runs on an Illumina HiSeq 2000 according to the manufacturers' protocols. Across the three samples the mean target coverage was 59×, and on average, 84% of bases were covered to a minimum depth of 20. Data were analyzed in Oxford Gene Technology’s exome analysis pipeline. Briefly, reads were aligned to the reference GRCh37 using bwa 0.61.20Li H. Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform.Bioinformatics. 2009; 25: 1754-1760Crossref PubMed Scopus (26648) Google Scholar Local realignment was performed around indels with the Genome Analysis Toolkit (GATK v1.4).21DePristo M.A. Banks E. Poplin R. Garimella K.V. Maguire J.R. Hartl C. Philippakis A.A. del Angel G. Rivas M.A. Hanna M. et al.A framework for variation discovery and genotyping using next-generation DNA sequencing data.Nat. Genet. 2011; 43: 491-498Crossref PubMed Scopus (7101) Google Scholar Optical and PCR duplicates were marked in BAM files with Picard 1.62. Original HiSeq base quality scores were recalibrated with GATK TableRecalibration, and variants were called with GATK UnifiedGenotyper. We hard filtered indels and SNPs according to the Broad Institute's best-practice guidelines (available online) to eliminate false-positive calls. Variant annotation was done with a modified version of ENSEMBL’s Variant Effect Predictor.22McLaren W. Pritchard B. Rios D. Chen Y. Flicek P. Cunningham F. Deriving the consequences of genomic variants with the Ensembl API and SNP Effect Predictor.Bioinformatics. 2010; 26: 2069-2070Crossref PubMed Scopus (1204) Google Scholar To determine de novo mutations in the proband, we assessed violations of the Mendelian inheritance rules in the trio by using GATK SelectVariants.21DePristo M.A. Banks E. Poplin R. Garimella K.V. Maguire J.R. Hartl C. Philippakis A.A. del Angel G. Rivas M.A. Hanna M. et al.A framework for variation discovery and genotyping using next-generation DNA sequencing data.Nat. Genet. 2011; 43: 491-498Crossref PubMed Scopus (7101) Google Scholar A Mendelian violation was scored if the child's genotype differed unexpectedly from that of the known parental genotypes at the same position. This captured positions in which the child (1) harbors an allele that is not present in either parent or (2) is homozygous for an allele present in one but not both parents. To be called as a Mendelian violation, the genotype in the proband and both parents needed to have a minimum genotype QUAL score of 50. QUAL scores are calculated in the genotyping phase and are the Phred-scaled probability that the polymorphism exists at this site given the underlying sequencing data. The value will increase with greater depth of coverage at the genotype position. All but one of the called Mendelian inconsistancies had read depths of >20 in the proband at the given positions. Sixteen genes with autosomal variants were thought to contain Mendelian violations. Eleven of the 16 variants were novel, i.e., they were not annotated in dbSNP132, and these were analyzed in more detail. Manual inspection of the genotypes at the read-level revealed that three of the genes with Mendelian violations that were associated with point mutations did not result from the detection of a de novo variant that was detected in the proband but absent in both parents. Rather, they resulted from a putative reversion of this sequence to the reference sequence in the proband for whom one parent was homozygous nonreference. These variants were discarded. Three other Mendelian violations associated with indel positions were rated as artifacts as a result of inconsistencies in the genotype calls in one or both of the parents, a common problem in the course of indel calling in lower-coverage regions. None of the six variants that were discarded affected the coding region of a transcript or a gene with a known function in bone metabolism. The remaining five variants were followed up by Sanger sequencing of genomic DNA of the proband and her parents. Only one of these variants, a heterozygous C>T transition in the 5′-UTR of IFITM5 (the gene encoding Interferon induced transmembrane protein 5; NM_00102595.1), affected the transcribed region of a gene, and only two of the candidates were located within genes that have a known function in bone homeostasis: an intronic heterozygous A>C variant in RUNX1 (MIM 151385, NM_001754.4), located 1307 bp downstream of exon 2, and the IFITM5 variant. Sanger sequencing revealed that three variants, including the intronic RUNX1 variant, were false positives. That the RUNX1 variant was a false positive is well compatible with the whole-exome sequencing data; the RUNX1 variant call was error prone because of low coverage (10×) in the proband and the variant's genomic location within a repeat region adjacent to a deletion. Two variants were confirmed by Sanger sequencing as true heterozygous de novo mutations in the proband: a variant c.6620-32_-34del in intron 18 of KIAA0947 (NM_015325.1), encoding a protein of unknown function, and the IFITM5 5′-UTR variant, located 14 bp upstream of the annotated translation initiation codon of the gene. The algorithms Splice View23Rogozin I.B. Milanesi L. Analysis of donor splice sites in different eukaryotic organisms.J. Mol. Evol. 1997; 45: 50-59Crossref PubMed Scopus (138) Google Scholar and Spliceport24Dogan R.I. Getoor L. Wilbur W.J. Mount S.M. SplicePort—An interactive splice-site analysis tool.Nucleic Acids Res. 2007; 35: W285-W291Crossref PubMed Scopus (180) Google Scholar did not predict that the intronic variant c.6620-32_-34del in KIAA0947 would alter the adjacent splice acceptor site. Therefore, a single de novo mutation in proband 1 remained as a strong candidate for causing the phenotype: the heterozygous variant c.−14C>T in IFITM5 (Figure 2A and Figure S1). This variant is located in the transcribed region of a gene that encodes a protein with a function in bone. Sanger sequencing of both coding IFITM5 exons identified no other mutation in the proband. Primer details are available upon request. As a next step, we performed Sanger sequencing of the IFITM5 5′-UTR and both coding exons on genomic DNA samples of proband 2 and his unaffected parents. Intriguingly, this individual was heterozygous for the same 5′-UTR mutation c.−14C>T, that we had identified in proband 1, whereas both parents were homozygous for the reference sequence at this site (Figure 2B). Again, no other mutation was detected in the coding region of IFITM5 in the proband. We confirmed paternity in both proband-parent trios by genotyping 11 microsatellite markers (data not shown; correct paternity in the first proband-parent trio can also be derived from the low number of Mendelian violations in the inheritance of variants). To exclude that this 5′-UTR sequence alteration represents a polymorphism, we referred to the Exome Variant Server (NHLBI Exome Sequencing Project [ESP], Seattle, WA). The genomic position mutated in the two children with OI type V (chr11:299,504 in hg19) was covered in more than 5,150 individuals of European American and African American descent with an average read depth of 14-fold. None of these individuals displayed the IFITM5 mutation c.−14C>T. The de novo mutation rate in humans is ∼1.1 × 10−8 per position per haploid genome, and there are on average 70 new mutations arising in the diploid genome per generation and 0–3 de novo mutations detectable in the approximately 50 MB of genomic DNA covered by a typical whole-exome sequencing project.25Keightley P.D. Rates and fitness consequences of new mutations in humans.Genetics. 2012; 190: 295-304Crossref PubMed Scopus (106) Google Scholar, 26Roach J.C. Glusman G. Smit A.F. Huff C.D. Hubley R. Shannon P.T. Rowen L. Pant K.P. Goodman N. Bamshad M. et al.Analysis of genetic inheritance in a family quartet by whole-genome sequencing.Science. 2010; 328: 636-639Crossref PubMed Scopus (793) Google Scholar, 27Vissers L.E. de Ligt J. Gilissen C. Janssen I. Steehouwer M. de Vries P. van Lier B. Arts P. Wieskamp N. del Rosario M. et al.A de novo paradigm for mental retardation.Nat. Genet. 2010; 42: 1109-1112Crossref PubMed Scopus (624) Google Scholar Therefore, a conservative estimate that takes into account the fact that the IFITM5 mutation affects a hypermutable cytosine-phosphate-guanine (CpG) dinucleotide would be that the likelihood to detect identical de novo mutations in two individuals with the same phenotype by chance is below 5 × 10−6. We concluded that the IFITM5 mutation c.−14C>T is the molecular cause of OI type V in both affected children. IFITM5 (alternatively called Bone-restricted interferon-induced transmembrane protein-like protein [BRIL]) is highly enriched in osteoblasts.28Moffatt P. Gaumond M.H. Salois P. Sellin K. Bessette M.C. Godin E. de Oliveira P.T. Atkins G.J. Nanci A. Thomas G. Bril: A novel bone-specific modulator of mineralization.J. Bone Miner. Res. 2008; 23: 1497-1508Crossref PubMed Scopus (116) Google Scholar, 29Hanagata N. Li X. Morita H. Takemura T. Li J. Minowa T. Characterization of the osteoblast-specific transmembrane protein IFITM5 and analysis of IFITM5-deficient mice.J. Bone Miner. Metab. 2011; 29: 279-290Crossref PubMed Scopus (83) Google Scholar, 30Hanagata N. Takemura T. Monkawa A. Ikoma T. Tanaka J. Phenotype and gene expression pattern of osteoblast-like cells cultured on polystyrene and hydroxyapatite with pre-adsorbed type-I collagen.J. Biomed. Mater. Res. A. 2007; 83: 362-371Crossref PubMed Scopus (31) Google Scholar Its function has mainly been studied in mice and rats. In these animals, Ifitm5 expression peaks during osteoblast maturation around the early mineralization stage,28Moffatt P. Gaumond M.H. Salois P. Sellin K. Bessette M.C. Godin E. de Oliveira P.T. Atkins G.J. Nanci A. Thomas G. Bril: A novel bone-specific modulator of mineralization.J. Bone Miner. Res. 2008; 23: 1497-1508Crossref PubMed Scopus (116) Google Scholar, 29Hanagata N. Li X. Morita H. Takemura T. Li J. Minowa T. Characterization of the osteoblast-specific transmembrane protein IFITM5 and analysis of IFITM5-deficient mice.J. Bone Miner. Metab. 2011; 29: 279-290Crossref PubMed Scopus (83) Google Scholar suggesting a role in bone formation. Mouse IFITM5 has a similarity of 88% to human IFITM5.28Moffatt P. Gaumond M.H. Salois P. Sellin K. Bessette M.C. Godin E. de Oliveira P.T. Atkins G.J. Nanci A. Thomas G. Bril: A novel bone-specific modulator of mineralization.J. Bone Miner. Res. 2008; 23: 1497-1508Crossref PubMed Scopus (116) Google Scholar The expression pattern of Ifitm5 during embryonal development is similar to that observed for Osterix (Sp7),29Hanagata N. Li X. Morita H. Takemura T. Li J. Minowa T. Characterization of the osteoblast-specific transmembrane protein IFITM5 and analysis of IFITM5-deficient mice.J. Bone Miner. Metab. 2011; 29: 279-290Crossref PubMed Scopus (83) Google Scholar the human ortholog of which is mutated in a rare autosomal-recessive form of OI.13Lapunzina P. Aglan M. Temtamy S. Caparrós-Martín J.A. Valencia M. Letón R. Martínez-Glez V. Elhossini R. Amr K. Vilaboa N. Ruiz-Perez V.L. Identification of a frameshift mutation in Osterix in a patient with recessive osteogenesis imperfecta.Am. J. Hum. Genet. 2010; 87: 110-114Abstract Full Text Full Text PDF PubMed Scopus (215) Google Scholar The available expression data from adult rodents argue for a role of Ifitm5 not only during bone formation in embryogenesis but also during postnatal development. In Ifitm5−/− mice, the long bones are 15%–25% shorter at birth than in Ifitm5+/− mice, and they are sometimes severely bent, a symptom that partially resolves by adulthood.31Hanagata N. Li X. Osteoblast-enriched membrane protein IFITM5 regulates the association of CD9 with an FKBP11-CD81-FPRP complex and stimulates expression of interferon-induced genes.Biochem. Biophys. Res. Commun. 2011; 409: 378-384Crossref PubMed Scopus (38) Google Scholar However, bone morphometric parameters for the tibiae did not differ significantly between Ifitm5−/− and Ifitm5+/− mice, suggesting that loss of Ifitm5 has no effect on bone mass. Increased bone fragility or hyperplastic-callus formation has also not been reported for Ifitm5-deficient mice. Other bone deficiencies observed in newborn mice, such as a less-c