Mutations in the COL1A2 gene of type I collagen that result in nonlethal forms of osteogenesis imperfecta

Although virtually all mutations that result in osteogenesis imperfecta (OI) affect the genes that encode the chains of type I procollagen, the effects of mutations in the COL1A2 gene have received less attention than those in the COL1A1 gene. We have characterized mutations in 4 families that give rise to different OI phenotypes. In three families substitutions of glycine residues by cysteine in the triple helical domain (a single example at position 259 and 2 families in which substitution of glycine at 646 by cysteine) have been identified, and in the fourth a G for A transition at position + 4 in intron 33 led to use of an alternative splice site and inclusion of 6 amino acids (val-gly-arg-ile-leu-phe) between residues 585 and 586 of the normal triple helix. The relation between position of substitution of glycine by cysteine in the COL1A2 gene does not follow the pattern developed in the COL1A1 gene. To determine how COL1A2 mutations produce OI phenotypes, we have produced a full-length mouse cDNA into which we plan to place mutations and examine their effects in stably transfected osteogenic cells and in transgenic animals.     

Mutation analysis of COL1A1 and COL1A2 in patients diagnosed with osteogenesis imperfecta type I-IV

Osteogenesis Imperfecta (OI) is a heterogeneous group of inherited disorders characterized by increased bone fragility, with clinical severity ranging from mild to lethal. To date, seven types of OI have been described, based on clinical phenotype and histological findings. Most patients with a clinical diagnosis of OI type I-IV have a mutation in the COL1A1 or COL1A2 genes which encode the two alpha chains of type I collagen, the major component of the bone matrix. Analysis of COL1A1 and COL1A2 in a cohort of 83 unrelated patients with OI type I-IV identified a total of 62 mutations. Thirty-eight appear novel, 26 in COL1A1, and 12 in COL1A2, and these are described here. The largest group consists of point mutations affecting glycine residues in the triple helical domain of the two alpha chains, predicted to disrupt protein folding and structure. This is in accordance with previously published data. A doublet GC deletion, an unusual 398 base deletion predicted to completely remove exon 20 of COL1A2, and a point mutation resulting in substitution of a conserved cysteine in the C-terminal propeptide are described. In addition rare mutations at the cleavage sites of the C-propeptide and the N-terminal signal peptide are described.     

Recurrence of lethal osteogenesis imperfecta due to parental mosaicism for a mutation in the COL1A2 gene of type I collagen. The mosaic parent exhibits phenotypic features of a mild form of the disease.

We have determined that a man, ascertained because he fathered a child with lethal osteogenesis imperfecta (OI) with each of two partners, is mosaic in both his germline and somatic tissues for a mutation in the COL1A2 gene which encodes the pro alpha 2(I) chain of type I procollagen. His dermal fibroblasts were previously shown to synthesize a population of cysteine-containing alpha 2(I) chains that were posttranslationally overmodified. DNA sequence analysis of COL1A2 cDNAs demonstrated that the cysteine-containing chain resulted from a point mutation (G to T) in the first position of the codon for the glycine at residue 472 of the triple helical domain. Genomic DNA from the one available affected infant contained the mutant and normal COL1A2 alleles in equal proportion. Examination of DNA from several tissues of the father showed that the mutant allele was present in approximately 40% of his sperm, 80% of his lymphocytes, and nearly 100% of his dermal fibroblasts. Despite the high level of mosaicism detected in somatic tissues, the only phenotypic manifestation of OI in the proband was that he was shorter than his unaffected male relatives and had mild dentinogenesis imperfecta. Thermal stability of type I collagen molecules containing the substitution was decreased, but to a lesser extent than for a nonlethal cysteine for glycine substitution at residue 259 of alpha 2(I), indicating that this measure of molecular stability may be of limited use in explaining the pathogenesis of osteogenesis imperfecta.     

Lack of correlation between the type of COL1A1 or COL1A2 mutation and hearing loss in osteogenesis imperfecta patients

Osteogenesis imperfecta (OI) is caused by mutations in COL1A1 and COL1A2 that code for the alpha1 and alpha2 chains of type I collagen. Phenotypes correlate with the mutation types in that COL1A1 null mutations lead to OI type I, and structural mutations in alpha1(I) or alpha2(I) lead to more severe OI types (II-IV). However, correlative analysis between mutation types and OI associated hearing loss has not been previously performed. A total of 54 Finnish OI patients with previously diagnosed hearing loss or age 35 or more years were analyzed here for mutations in COL1A1 or COL1A2. Altogether 49 mutations were identified, of which 41 were novel. The 49 mutations represented the molecular genetic background of 41.1% of the Finnish OI population. A total of 38 mutations were in COL1A1 and 11 were in COL1A2. Of these, 16 were glycine substitutions and 16 were splicing mutations in alpha1(I) or alpha2(I). In addition, 17 null allele mutations were detected in COL1A1. A total of 32 patients (65.3%) with a mutation had hearing loss. That is slightly more than in our previous population study on Finnish adults with OI (57.9%). The association between the mutation types and OI type was statistically evident. Patients with COL1A1 mutations more frequently had blue scleras than those with COL1A2 mutations. In addition, patients with COL1A2 mutations tended to be shorter than those with COL1A1 mutations. However, no correlation was found between the mutated gene or mutation type and hearing pattern. These results suggest that the basis of hearing loss in OI is complex, and it is a result of multifactorial, still unknown genetic effects.     

CRTAP and LEPRE1 mutations in recessive osteogenesis imperfecta

Autosomal dominant osteogenesis imperfecta (OI) is caused by mutations in the genes (COL1A1 or COL1A2) encoding the chains of type I collagen. Recently, dysregulation of hydroxylation of a single proline residue at position 986 of both the triple-helical domains of type I collagen alpha1(I) and type II collagen alpha1(II) chains has been implicated in the pathogenesis of recessive forms of OI. Two proteins, cartilage-associated protein (CRTAP) and prolyl-3-hydroxylase-1 (P3H1, encoded by the LEPRE1 gene) form a complex that performs the hydroxylation and brings the prolyl cis-trans isomerase cyclophilin-B (CYPB) to the unfolded collagen. In our screen of 78 subjects diagnosed with OI type II or III, we identified three probands with mutations in CRTAP and 16 with mutations in LEPRE1. The latter group includes a mutation in patients from the Irish Traveller population, a genetically isolated community with increased incidence of OI. The clinical features resulting from CRTAP or LEPRE1 loss of function mutations were difficult to distinguish at birth. Infants in both groups had multiple fractures, decreased bone modeling (affecting especially the femurs), and extremely low bone mineral density. Interestingly, "popcorn" epiphyses may reflect underlying cartilaginous and bone dysplasia in this form of OI. These results expand the range of CRTAP/LEPRE1 mutations that result in recessive OI and emphasize the importance of distinguishing recurrence of severe OI of recessive inheritance from those that result from parental germline mosaicism for COL1A1 or COL1A2 mutations.     

Mutational spectrum of type I collagen genes in Korean patients with osteogenesis imperfecta

Mutations in the type I collagen genes COL1A1 and COL1A2 are responsible for the dominantly inherited connective tissue disorder osteogenesis imperfecta (OI). The severity of OI is diverse, ranging from perinatal lethality to a very mild phenotype that is characterized by normal stature and the absence of deformities. Although there have been several studies on the mutational spectra of COL1A1 and/or COL1A2 in Western populations, very few cases have been reported from Asia. In this study, we investigated 67 unrelated Korean probands with OI and used nucleotide sequence analysis to detect COL1A1 and COL1A2 mutations. Thirty-five different mutations were identified in the two genes, including 24 novel mutations. Among the 35 kinds of detected mutations, 15 were glycine substitutions (seven in COL1A1 and eight in COL1A2), one was a nonsense mutation, four were frameshift mutations in COL1A1, three were in-frame duplications in COL1A2, and 12 were splice site mutations (seven in COL1A1 and five in COL1A2). Until now, mutations in the COL1A1 and COL1A2 genes known to cause OI were unique and rarely repeated in other families. Interestingly, the c.982G>A (p.Gly328Ser) mutation in COL1A2 was found recurrently and was the causative mutation in five independent OI probands. Haplotype analysis of the COL1A2 gene revealed that four probands from five independent OI probands with c.982G>A (p.Gly328Ser) had a common haplotype. Our clinical data showed the heterogeneity even within a specific genotype, which suggested the complex expression of this disease.     

Glycine to tryptophan substitution in type I collagen in a patient with OI type III: a unique collagen mutation

We report a unique glycine substitution in type I collagen and highlight the clinical and biochemical consequences. The proband is a 9 year old Turkish boy with severely deforming osteogenesis imperfecta (OI). Biochemical analysis of (pro) collagen type I from a skin fibroblast culture showed both normal and overmodified α chains. Molecular analysis showed a G>T transversion in the COL1A2 gene, resulting in the substitution of glycine by tryptophan at position 277 of the α2(I) collagen chain. Glycine substitutions in type I collagen are the most frequent cause of the severe and lethal forms of OI. The phenotypic severity varies according to the nature and localisation of the mutation. Substitutions of glycine by tryptophan, which is the most voluminous amino acid, have not yet been identified in type I collagen or any other fibrillar collagen. The severe, though non-lethal OI phenotype associated with this mutation may appear surprising in view of the huge size of the tryptophan residue. The fact that the mutation resides within a so called "non-lethal" region of the α2(I) collagen chain supports a regional model in phenotypic severity for α2(I) collagen mutations, in which the phenotype is determined primarily by the nature of the collagen domain rather than the type of glycine substitution involved.


Keywords: osteogenesis imperfecta; COL1A2; tryptophan; collagen     

A novel mutation combining with rs66612022 in a Chinese pedigree suggests a new pathogenesis to osteogenesis imperfecta via whole genome sequencing

Osteogenesis imperfecta (OI) is a rare heritable disease with systemic connective tissue disorder. Most of the patients represent autosomal dominant form of OI, and are usually resulting from the mutations in type I collagen genes. However, the gene mutations reported previously only account for ∼70% of the OI cases. Here, in a Chinese OI family, we examined seven patients and nine normal individuals using the whole genome sequencing and molecular genetic analysis. The mutation of rs66612022 (COL1A2:p.Gly328Ser) related to glycine substitution was found in the seven patients. Moreover, we identified a novel missense mutation (HMMR:p.Glu2Gln). Interestingly, the individuals of this family with both the mutations were suffering from OI, while the others carried one or none of them are normal. The mutations of COL1A2 and HMMR and their combined effect on OI would further expand the genetic spectrum of OI.     

Osteogenesis imperfecta: prospects for molecular therapeutics

Osteogenesis Imperfecta (OI) is a dominant negative disorder of connective tissue. OI patients present with bone fragility and skeletal deformity within a broad phenotypic range. Defects in the COL1A1 or COL1A2 genes, coding, respectively, for the alpha1 and alpha2 chains of type I collagen, are the causative mutations. Over 150 mutations have been characterized. Both quantitative defects, such as null COL1A1 alleles, and qualitative defects, such as glycine substitutions, exon skipping, deletions, and insertions, have been described in type I collagen. Quantitative and structural mutations are associated with the milder and more severe forms of OI, respectively. A more detailed relationship between genotype and phenotype is still incompletely understood; several models have been proposed and are being tested. Transgenic and knock-out murine models for OI have previously been created. We have recently generated a knock-in murine model (the Brittle mouse) carrying a typical glycine substitution in type I collagen. This mouse will permit a better understanding of OI pathophysiology and phenotypic variability. It will also be used for gene therapeutic approaches to OI, especially mutation suppression by hammerhead ribozymes. The present review will provide an update of OI clinical and molecular data and outline gene therapeutic approaches being tested on OI murine models for this disorder.     

Genetic and biochemical analyses of Israeli osteogenesis imperfecta patients

Osteogenesis imperfecta (OI) is clinically characterized by abnormal bone fragility, with most patients harboring heterozygote germline mutations in the COL1A1 or COL1A2 genes that encode the chains of type I procollagen, the major protein in bone. More than 250 mutations in both genes in OI patients have been reported, mostly missense mutations affecting glycine residues in the triple helical domains of the two chains. These mutations disrupt protein folding and structure, and their effects often can be detected by the analysis of proteins synthesized but cultured fibroblasts or, less often, osteoblasts. In this study, mutational analysis of all the COL1A1 and part of the COL1A2 was performed using exon-specific PCR amplification followed by denaturing gradient gel electrophoresis (DGGE) analysis and complemented by DNA sequencing in 57 Israeli OI patients from 55 unrelated families. Protein analysis was also performed using cultured fibroblasts obtained from a subset of these OI patients. Of 57 OI patients analyzed, 35 had OI type 1, 12 has OI type III, 8 had OI type IV, and 2 had OI type II. Fourteen different pathogenic mutations (10 novel) were identified in the COL1A1 gene: 3 missense, 5 nonsense, 3 insertion/deletion frameshift, 2 splice junction mutations, and 1 in frame deletion. We conclude that COL1A1 mutations underlie a subset of Israeli OI patients, that most commonly in OI type I, the mutations are contained within the COL1A1 gene, and that there are no predominant mutations in Jewish OI patients. Lastly, the use of protein analyses complements genetic analyses.     

A family with homozygous and heterozygous p.Gly337Ser mutations in COL1A2

Osteogenesis imperfecta (OI) is commonly caused by monoallelic mutations in COL1A1 or COL1A2. Biallelic mutations are extremely rare. Only five previous reports have identified seven OI patients with homozygous mutations in COL1A2. OI is a genetically and phenotypically heterogeneous disorder which challenges an establishment of genotype-phenotype correlation. Notably, more than thirty patients with OI possess the heterozygous mutation, p.Gly337Ser, in COL1A2. Their clinical severity ranges from mild OI type I to severe types III and IV. Here, we report a 17-year-old Thai female with recurrent bone fractures, short stature, blue sclerae, triangular face, missing teeth, dentinogenesis imperfecta (DI), skeletal deformities, and scoliosis. She was diagnosed with OI type III. Her parents were second-cousin-once-removed. The father was a professional Thai boxer. Both had normal bone mineral density, no history of bone fractures, and only teeth problems. They were diagnosed with DI without OI. Whole exome sequencing identified that the proband harbored the homozygous mutation, c.1009G > A (p.Gly337Ser), in exon 19 of COL1A2 while her parents were heterozygous for this mutation. This study reports the eighth child with OI and the homozygous mutation in COL1A2; and the first two individuals with the heterozygous p.Gly337Ser mutation in COL1A2 causing an isolated DI without OI.     

Hearing loss in individuals with osteogenesis imperfecta in North America: Results from a multicenter study

Hearing loss (HL) is an extra‐skeletal manifestation of the connective tissue disorder osteogenesis imperfecta (OI). Systematic evaluation of the prevalence and characteristics of HL in COL1A1/COL1A2‐related OI will contribute to a better clinical management of individuals with OI. We collected and analyzed pure‐tone audiometry data from 312 individuals with OI who were enrolled in the Linked Clinical Research Centers and the Brittle Bone Disorders Consortium. The prevalence, type, and severity of HL in COL1A1/COL1A2‐related OI are reported. We show that the prevalence of HL in OI is 28% and increased with age in Type I OI but not in Types III and IV. Individuals with OI Types III and IV are at a higher risk to develop HL in the first decade of life when compared to OI Type I. We also show that the prevalence of SNHL is higher in females with OI compared to males. This study reveals new insights regarding prevalence of HL in OI including a lower general prevalence of HL in COL1A1/COL1A2‐related OI than previously reported (28.3 vs. 65%) and high prevalence of SNHL in females. Our data support the need in early routine hearing evaluation in all types of OI that can be adjusted to the severity of the skeletal disease.     

Linkage analysis in dominantly inherited osteogenesis imperfecta

The only serious attempts at linkage in osteogenesis imperfecta (OI) have shown that the disease is linked to type 1 collagen genes in all families studied in which it segregrates as a clear mendelian dominant trait. For prenatal diagnosis the probability that a new family is linked can be taken as greater than 0.95 and this figure is augmented as more meioses are studied. Some phenotype correlations, notably between the OI type IV phenotype and linkage to COL1A2 and between presenile hearing loss in OI type I and linkage to COL1A1, can be used to improve risk estimates substantially in families where there are no segregation data to distinguish whether COL1A1 or COL1A2 is the mutant locus.     

Haplotype analysis of collagen type I genes in the general population and in osteogenesis imperfecta families

The allele frequencies of 2 new polymorphic markers of collagen type I proalpha 1 (COL1A1) and proalpha 2 (COL1A2) genes were determined in a random sample of chromosomes by polymerase chain reaction. The minor allele frequencies were 0.27 for COL1A1/+88Mn1I, and 0.39 for COL1A2/1446 PvuII RFLPs, respectively. These 2 polymorphisms increased the combined (PIC) values we previously determined in the Italian population with Southern blotting procedures, from 0.71 at the COL1A1 locus to 0.81, and from 0.71 at the COL1A2 locus to 0.88, respectively. With a combination of these markers, we have carried out the segregation analysis of 4 new families in which osteogenesis imperfecta (OI) segregated as a dominant trait. The disease segregated with COL1A1 in 2 OI type I families, and with COL1A2 in one OI type IV family. In one OI type I family the concordant locus was uncertain. This analysis was extended to the 7 dominant OI families we previously reported: in 3 out of 11 pedigrees either locus still could not be excluded, indicating the need for more genetic markers. COL1A1 and COL1A2 haplotype frequencies were compared in normal and OI chromosomes: no preferential association of the disease with a given haplotype was detected. The correlation between affected locus and clinical aspects is discussed.     

Thirty-three novel COL1A1 and COL1A2 mutations in patients with osteogenesis imperfecta types I-IV

Osteogenesis imperfecta (OI) is a heritable disease of bone characterized by low bone mass and bone fragility. Six different types of OI have been described to date, based on clinical phenotype and histological findings. The genetic defect in many patients with OI types I-IV is due to mutations in the genes encoding type I collagen, while patients with OI types V and VI show no evidence of mutations in the COL1A1/COL1A2 genes. Here we report thirty-three novel mutations in patients with types I-IV OI. Sixteen mutations were in COL1A1 and seventeen were in COL1A2. Most mutations resulted in substitutions for glycine: one of these, a doublet GG>CC transversion, created a unique Gly-->Pro missense mutation in the triple helical domain of COL1A2. Two rare triple helical Gly-->Glu substitutions in COL1A2 are also described. In addition, there were six single-base deletion mutations resulting in frameshifts, seven splice junction mutations, and a 9-bp triple helix insertion associated with a severe (OI II) phenotype. The variety of mutations described in the COL1A1/COL1A2 genes giving rise to an OI phenotype is in accordance with the clinical heterogeneity of the disease. Hum Mutat 17:434, 2001.     

Osteogenesis imperfecta type III: mutations in the type I collagen structural genes, COL1A1 and COL1A2, are not necessarily responsible.

Most forms of osteogenesis imperfecta are caused by dominant mutations in either of the two genes, COL1A1 and COL1A2, that encode the pro alpha 1(I) and pro alpha 2(I) chains of type I collagen, respectively. However, a severe, autosomal recessive form of OI type III with a comparatively high frequency has been recognised in the black populations of southern Africa. We preformed linkage analyses in eight OI type III families using RFLPs associated with the COL1A1 and COL1A2 loci to determine whether mutations in the genes for type I collagen were responsible for this form of OI. Recombination between the OI phenotype and polymorphic markers at both loci was shown in three of the eight families investigated. The combined lod scores for the eight families were -10.6 for COL1A1 and -11.2 for COL1A2. Further, we examined the type I procollagen produced by skin fibroblast cultures derived from 15 affected and 12 unaffected subjects from the above eight families plus one further family. We found no evidence for defects in the synthesis, structure, secretion, or post-translational modification of the chains of type I procollagen produced by any of the family members. These results suggest that mutations within or near the type I collagen structural genes are not responsible for this form of OI.     

COL1-related overlap disorder: A novel connective tissue disorder incorporating the osteogenesis imperfecta/Ehlers-Danlos syndrome overlap

The 2017 classification of Ehlers-Danlos syndromes (EDS) identifies three types associated with causative variants in COL1A1/COL1A2 and distinct from osteogenesis imperfecta (OI). Previously, patients have been described with variable features of both disorders, and causative variants in COL1A1/COL1A2; but this phenotype has not been included in the current classification. Here, we expand and re-define this OI/EDS overlap as a missing EDS type. Twenty-one individuals from 13 families were reported, in whom COL1A1/COL1A2 variants were found after a suspicion of EDS. None of them could be classified as affected by OI or by any of the three recognized EDS variants associated with COL1A1/COL1A2. This phenotype is dominated by EDS-related features. OI-related features were limited to mildly reduced bone mass, occasional fractures and short stature. Eight COL1A1/COL1A2 variants were novel and five recurrent with a predominance of glycine substitutions affecting residues within the procollagen N-proteinase cleavage site of α1(I) and α2(I) procollagens. Selected variants were investigated by biochemical, ultrastructural and immunofluorescence studies. The pattern of observed changes in the dermis and in vitro for selected variants was more typical of EDS rather than OI. Our findings indicate the existence of a wider recognizable spectrum associated with COL1A1/COL1A2.     

EMQN best practice guidelines for the laboratory diagnosis of osteogenesis imperfecta

Osteogenesis imperfecta (OI) comprises a group of inherited disorders characterized by bone fragility and increased susceptibility to fractures. Historically, the laboratory confirmation of the diagnosis OI rested on cultured dermal fibroblasts to identify decreased or abnormal production of abnormal type I (pro)collagen molecules, measured by gel electrophoresis. With the discovery of COL1A1 and COL1A2 gene variants as a cause of OI, sequence analysis of these genes was added to the diagnostic process. Nowadays, OI is known to be genetically heterogeneous. About 90% of individuals with OI are heterozygous for causative variants in the COL1A1 and COL1A2 genes. The majority of remaining affected individuals have recessively inherited forms of OI with the causative variants in the more recently discovered genes CRTAP, FKBP10, LEPRE1,PLOD2, PPIB, SERPINF1, SERPINH1 and SP7, or in other yet undiscovered genes. These advances in the molecular genetic diagnosis of OI prompted us to develop new guidelines for molecular testing and reporting of results in which we take into account that testing is also used to 'exclude' OI when there is suspicion of non-accidental injury. Diagnostic flow, methods and reporting scenarios were discussed during an international workshop with 17 clinicians and scientists from 11 countries and converged in these best practice guidelines for the laboratory diagnosis of OI.     

Consortium for osteogenesis imperfecta mutations in the helical domain of type I collagen: regions rich in lethal mutations align with collagen binding sites for integrins and proteoglycans

Osteogenesis imperfecta (OI) is a generalized disorder of connective tissue characterized by fragile bones and easy susceptibility to fracture. Most cases of OI are caused by mutations in type I collagen. We have identified and assembled structural mutations in type I collagen genes (COL1A1 and COL1A2, encoding the proalpha1(I) and proalpha2(I) chains, respectively) that result in OI. Quantitative defects causing type I OI were not included. Of these 832 independent mutations, 682 result in substitution for glycine residues in the triple helical domain of the encoded protein and 150 alter splice sites. Distinct genotype-phenotype relationships emerge for each chain. One-third of the mutations that result in glycine substitutions in alpha1(I) are lethal, especially when the substituting residues are charged or have a branched side chain. Substitutions in the first 200 residues are nonlethal and have variable outcome thereafter, unrelated to folding or helix stability domains. Two exclusively lethal regions (helix positions 691-823 and 910-964) align with major ligand binding regions (MLBRs), suggesting crucial interactions of collagen monomers or fibrils with integrins, matrix metalloproteinases (MMPs), fibronectin, and cartilage oligomeric matrix protein (COMP). Mutations in COL1A2 are predominantly nonlethal (80%). Lethal substitutions are located in eight regularly spaced clusters along the chain, supporting a regional model. The lethal regions align with proteoglycan binding sites along the fibril, suggesting a role in fibril-matrix interactions. Recurrences at the same site in alpha2(I) are generally concordant for outcome, unlike alpha1(I). Splice site mutations comprise 20% of helical mutations identified in OI patients, and may lead to exon skipping, intron inclusion, or the activation of cryptic splice sites. Splice site mutations in COL1A1 are rarely lethal; they often lead to frameshifts and the mild type I phenotype. In alpha2(I), lethal exon skipping events are located in the carboxyl half of the chain. Our data on genotype-phenotype relationships indicate that the two collagen chains play very different roles in matrix integrity and that phenotype depends on intracellular and extracellular events.     

Segregation analysis of dominant osteogenesis imperfecta in Italy.

We have performed linkage analysis in seven Italian families, in which mild osteogenesis imperfecta (OI) segregated as a dominant trait, by means of six DNA restriction fragment length polymorphisms (RFLPs) of type I collagen genes. OI type I was linked to the alpha 1(I) gene (COL1A1) in two families, and to the alpha 2(I) gene (COL1A2) in one family. OI type IV segregated with COL1A2 in two families. In two OI type I families, the molecular genetic data were insufficient for exclusion of one gene. Four DNA polymorphisms were particularly informative for cosegregation analysis of OI in Italian kindreds.