Osteogenesis Imperfecta, Type I

A number sign (#) is used with this entry because OI type I (OI1) is caused by heterozygous mutation in the COL1A1 gene (120150) or the COL1A2 gene (120160).

Description

Osteogenesis imperfecta type I is a dominantly inherited, generalized connective tissue disorder characterized mainly by bone fragility and blue sclerae. In most cases, 'functional null' alleles of COL1A1 on chromosome 17 or COL1A2 on chromosome 7 lead to reduced amounts of normal collagen I.

Clinical Features

Osteogenesis imperfecta (see Byers, 1993) is characterized chiefly by multiple bone fractures, usually resulting from minimal trauma. Affected individuals have blue sclerae, normal teeth, and normal or near-normal stature (for growth curves, see Vetter et al., 1992). Fractures are rare in the neonatal period; fracture tendency is constant from childhood to puberty, decreases thereafter, and often increases following menopause in women and after the sixth decade in men. Fractures heal rapidly with evidence of a good callus formation, and, with good orthopedic care, without deformity. Hearing loss of conductive or mixed type occurs in about 50% of families, beginning in the late teens and leading, gradually, to profound deafness, tinnitus, and vertigo by the end of the fourth to fifth decade. Additional clinical findings may be thin, easily bruised skin, moderate joint hypermobility and kyphoscoliosis, hernias, and arcus senilis. Mitral valve prolapse, aortic valvular insufficiency, and a slightly larger than normal aortic root diameter have been identified in some individuals (Hortop et al., 1986), but it is not clear that these disorders are significantly more frequent than in the general population.

Radiologically, wormian bones are common but bone morphology is generally normal at birth, although mild osteopenia and femoral bowing may be present. Vertebral body morphology in the adult is normal initially, but often develops the classic 'cod-fish' appearance (Steinmann et al., 1991).

EYES

Individuals with OI type I have distinctly blue sclerae which remain intensely blue throughout life, in contrast to the sclerae in OI type III and OI type IV which may also be blue at birth and during infancy. The intensity of the blue fades with time such that these individuals may have sclerae of normal hue by adolescence and adult life (Sillence et al., 1993). In a likely heterogeneous group of 16 patients with OI syndromes, Kaiser-Kupfer et al. (1981) found low ocular rigidity and small corneal diameter and globe length; no correlation was found between rigidity of the eyeball and blueness of the sclera. The central corneal thickness was found to be significantly lower in 53 patients with OI than that in 35 patients with otosclerosis and in 35 control subjects (Pedersen and Bramsen, 1984).

Hartikka et al. (2004) found that patients with COL1A1 mutations more frequently had blue sclerae than those with COL1A2 mutations. In addition, patients with COL1A2 mutations tended to be shorter than those with COL1A1 mutations.

CARDIOVASCULAR SYSTEM

The prevalence and severity of cardiovascular involvement in OI type I was determined in a prospective study of patients of all ages (Pyeritz and Levin, 1981). Mitral valve prolapse occurred in 18% (3 times the prevalence in unaffected relatives) and rarely progressed to mitral regurgitation. Mean aortic root diameter was slightly but significantly increased and was associated with aortic regurgitation in 1 to 2%. No patient had suffered a dissection. Later, Hortop et al. (1986) studied 109 persons with nonlethal OI from 66 families. They could demonstrate no definite increase in the frequency of mitral valve prolapse over that to be expected in any group of persons. Aortic root dilatation was found by echocardiogram to be present in 8 of 66 persons with OI syndrome; dilatation was mild and unrelated to age of the patient but was strikingly aggregated in families. Of 109 persons surveyed, valvular disease was evident clinically in only 4 persons (aortic regurgitation in 2, aortic stenosis in one, and mitral valve prolapse in one). Hortop et al. (1986) stated that aortic root dilatation was seen in each of the different OI syndromes but strikingly segregated within certain families. They concluded that the mild and apparently nonprogressive nature of this lesion in OI argues against the use of beta-adrenergic blockade in affected individuals in the absence of systemic arterial hypertension.

Mayer et al. (1996) reported a 35-year-old woman with a mild form of OI1 who presented with spontaneous dissection of the right internal carotid artery and the right vertebral artery after scuba diving. She had no obvious head or neck trauma. Other than a history of easy bruising and bluish sclerae, she had no evidence of a connective tissue disorder. There had been no bone fractures or dental problems nor was there family history of vasculopathy. Genetic analysis identified a heterozygous mutation in the COL1A1 gene (G13A; 120150.0052).

EARS

In likely heterogeneous groups of patients with OI, about half of affected individuals have hearing loss that begins during the second decade as a conductive loss; older individuals have sensorineural losses (Riedner et al., 1980; Pedersen, 1984). In only 1 major study was a majority of patients with sensorineural pattern observed (Shapiro et al., 1982). A female-to-male preponderance of 2:1 has been reported (Shea and Postma, 1982). Hearing loss is different from otosclerosis.

Vertigo is frequently associated with otosclerosis in which the hearing loss clinically resembles that in OI. Vertigo is also common in basilar impression found in up to 25% of adult OI patients. To evaluate the cause, frequency, and characteristics of vertigo in OI, Kuurila et al. (2003) studied 42 patients by interview, clinical examination, and audiologic examination supplemented with electronystagmography (ENG) and lateral skull radiography. Audiometry showed hearing loss in 25 patients (59.5%). In 9 patients (21%), abnormal skull base anatomy was found in the forms of basilar impression, basilar invagination, or both. Vertigo, mostly of floating or rotational sensation of short duration, was reported by 22 patients (52.4%). Patients with hearing loss tended to have more vertigo than patients with normal hearing. Vertigo was not correlated with type of hearing loss or auditory brain stem response pathology. ENG was abnormal in 14 patients (33.3%). Kuurila et al. (2003) concluded that vertigo is common in patients with OI and that in most cases, it is secondary to inner ear pathology.

Hartikka et al. (2004) reported a correlative analysis between types of mutation in the COL1A1 and COL1A2 genes and OI-associated hearing loss. A total of 54 Finnish OI patients with previously diagnosed hearing loss or age 35 or more years were analyzed for mutations in COL1A1 or COL1A2. Altogether 49 mutations were identified, of which 41 were novel. No correlation was found between the mutated gene or mutation type and hearing pattern. The authors interpreted this to mean that the basis of hearing loss in OI is complex, and that it is a result of multifactorial, still unknown genetic effects.

SKIN

Using a suction-cup technique, Hansen and Jemec (2002) performed quantitative studies of skin mechanics (elasticity, distensibility, and hysteresis) in 10 patients with OI, 8 with type I and 2 with type III (259420), and 24 age-matched controls. Skin elasticity, distensibility, and hysteresis were significantly decreased in patients versus controls. OI type I patients had decreased distensibility and hysteresis but increased elasticity in comparison to the type III patients. The authors concluded that the skin of patients with OI is more stiff and less elastic than normal skin. These changes differ from age-related changes, which have been described as increased distensibility and viscosity (similar to hysteresis).

CRANIOFACIAL AND DENTAL FEATURES

To obtain baseline information on craniofacial development in OI patients who had not received bisphosphonate treatment, Waltimo-Siren et al. (2005) used lateral radiographs to analyze the size and form of the bony structures in heads of 59 consecutive patients with OI types I, III, or IV (Sillence classification). In OI type I they found linear measurements that were smaller than normal, indicating a general growth deficiency, but no remarkable craniofacial deformity. In OI types III and IV, the growth impairment was pronounced and the craniofacial form was altered as a result of differential growth deficiency and bending of the skeletal head structures. They found strong support both for an abnormally ventral position of the sella region due to bending of the cranial base and for a closing mandibular growth rotation. Vertical underdevelopment of the dentoalveolar structures and the condylar process were identified as the main reasons for the relative mandibular prognathism in OI. Waltimo-Siren et al. (2005) concluded that facial growth impairment would probably remain characteristic for many OI patients despite the widespread intervention with bisphosphonates and that orthodontic treatment should be further developed.

CLINICAL VARIABILITY

The disorder may exhibit considerable interfamilial and intrafamilial variability in the number of fractures and degree of disability. Rowe et al. (1985) reported a spectrum of disease severity within a 5-generation family. Those most severely affected exhibited more severe short stature and a mild degree of scoliosis relative to those who were less severely affected. Most striking were identical twins, the offspring of a mildly affected mother. Twin B was born small for gestational age, had had 12 fractures and was 150 cm tall (3rd centile) at 11 years of age. Her twin was born appropriate for gestational age and had had only 2 fractures at age 8 and 9 secondary to strenuous exercise; her current height was 162 cm (50th centile). This family study suggested that the severity of the disease is roughly correlated with the reduction in collagen I synthesis.

Willing et al. (1990) described 5 affected individuals of a 3-generation family with marked clinical variability. They wondered if there might be subtle biochemical differences between the family members with respect to the amount of the abnormal pro-alpha-1(I) chains produced or their intracellular fate, but no differences were observed. They noticed that the more severely affected family members had children with both mild and severe phenotypes, while the mildly affected individual had an offspring with a mild phenotype. This suggested to them that there might be some other, not identified, factor segregating independently in this family that acts to modulate the final phenotype.

CLASSIFICATION

Using clinical, radiographic, and genetic criteria, Sillence et al. (1979) developed the classification currently in use into types I to IV: a dominant form with blue sclerae, type I; a dominant form with normal sclerae, type IV (166220); a perinatally lethal OI syndrome, type II (166210); and a progressively deforming form with normal sclerae, type III. The biochemical and linkage studies support the broad validity of the classification but confirm that it is incomplete. Although biochemical and genetic studies will provide the basis of the most rational classification, even such a detailed scheme probably will never predict correctly the evolution of OI in every affected individual, because of the still unexplained variability of expression seen in many families (Byers, 1993).

Bauze et al. (1975) divided their 42 patients with OI into mild, moderate, and severe groups according to deformity of long bones. None of the 17 patients in the mild group had scoliosis or white sclerae. The terms 'congenita' and 'tarda' now have limited usefulness, since they do not specify the mode of inheritance or basic biochemical defects.

Levin et al. (1980) concluded that dominant type I OI separates clearly into families in which affected persons have opalescent teeth and those in which dentinogenesis imperfecta (DGI) is absent. In 5 families, all members whose teeth were studied radiographically and by scanning electron microscopy had opalescent teeth. In 2 families the teeth of all affected persons were normal. Some members of both classes of families had blue sclerae and others did not. These 2 forms of OI were designated type IA and IB, depending on the presence or absence, respectively, of DGI. Paterson et al. (1983) found that patients with associated DGI (type IA) have more severe disease, with a greater fracture rate and greater likelihood of growth impairment, than do type IB patients.

Superti-Furga et al. (2007) discussed the 2006 revisions to the Nosology of Constitutional Disorders of Bone by the Nosology Group of the International Skeletal Dysplasia Society and provided a comprehensive table of the new classification scheme.

Biochemical Features

Byers (1993) summarized that 'functional null' alleles, i.e., silent alleles or mutations leading to excluded proteins, are the most common biochemical and genetic features of OI type I, although structural mutations in COL1A1 and COL1A2 leading to the synthesis of abnormal procollagen I can occasionally produce the OI type I phenotype.

Assessing reports of biochemical findings in the OI syndromes is difficult because the phenotype and genetics generally are not specified. Most studies deal, no doubt, with heterogeneous groups of patients. Several forms of OI were among the earliest of the inherited disorders of collagen biosynthesis and structure to be studied using cultured dermal fibroblasts from affected individuals (Martin et al., 1971; Penttinen et al., 1975). Cells cultured from patients who, in retrospect, would be considered to have OI type I, synthesized less procollagen I than did controls, but the mechanism by which production was decreased was not determined. These studies were extended from culture to tissue.

Francis et al. (1974) concluded that patients with OI and blue sclerae tend to have a reduced amount of collagen that has normal stability, as measured by resistance to depolymerization by pronase, heat, or cold alkali, whereas those with white sclerae have a normal amount of collagen with reduced stability; they suggested that a defect in cross-linking of collagen is present in the severe form of the disease.

Sykes et al. (1977) and, in a slightly extended study, Francis et al. (1981), found an increased ratio of collagen III to I in dermis and interpreted this as indicating a deficiency of collagen I. In studies of 44 patients with OI, Cetta et al. (1983) found in the largest category, the mild form, also an increased ratio of collagens III to I in skin and, in addition, an increased ratio of hydroxylysine diglycoside to monoglycoside in skin collagen.

Rowe et al. (1981) proposed that an additional criterion for OI type I is the production of a reduced quantity of collagen I. Among the cases of osteogenesis imperfecta with reduced synthesis of pro-alpha-1 chains, considerable heterogeneity is likely to emerge at the level of gene structure, as in the case of the globin genes in the thalassemias. Barsh et al. (1982) found that cultured skin fibroblasts from 3 patients produced half-normal levels of procollagen type I. Furthermore, the OI cells contained equimolar amounts of pro-alpha-1(I) and pro-alpha-2(I) chains, which suggested that trimer assembly and secretion were limited by the level of pro-alpha-1(I) chain synthesis. The 'extra' pro-alpha-2(I) chain in the OI cells was in a non-disulfide bonded configuration and apparently contributed to an increased level of intracellular degradation. The results of Barsh et al. (1982) suggested that the stoichiometry of the pro-alpha chains in procollagen I is determined by the conformation of the chains rather than by the ratio in which they are synthesized, that molecules containing more than a single pro-alpha-2(I) chain are not assembled, and that the production of collagen I can be regulated by controlling synthesis of only one of its subunits.

Rowe et al. (1985) demonstrated that reductions in collagen I production and in the ratio of alpha-1(I) to alpha-2(I) mRNA are clearly segregated with affected individuals within the 5 generation family. Rowe et al. (1985) further suggested that the severity of the disorder is roughly correlated with the reduction in collagen I synthesis.

Wenstrup et al. (1990) correlated clinical severity in nonlethal variants of OI with the nature of the alteration in the alpha chains of procollagen I secreted by cultured fibroblasts. Cells from 40 probands secreted about half the normal amount of normal procollagen I and no identifiable abnormal molecules; these patients were generally of normal stature, rarely had bone deformity or dentinogenesis imperfecta, and had blue sclerae. Cells from 74 other probands produced and secreted normal and abnormal procollagen I molecules; these patients were usually short and had bone deformity and dentinogenesis imperfecta, and many had gray or blue-gray sclerae. In cells from yet another 18 probands, Wenstrup et al. (1990) were unable to identify altered procollagen I synthesis or structure.

Gauba and Hartgerink (2008) reported the design of a novel model system based upon collagen-like heterotrimers that can mimic the glycine mutations present in either the alpha-1 or alpha-2 chains of type I collagen. The design utilized an electrostatic recognition motif in 3 chains that can force the interaction of any 3 peptides, including AAA (all same), AAB (2 same and 1 different), or ABC (all different) triple helices. Therefore, the component peptides could be designed in such a way that glycine mutations were present in zero, 1, 2, or all 3 chains of the triple helix. They reported collagen mutants containing 1 or 2 glycine substitutions with structures relevant to native forms of OI. Gauba and Hartgerink (2008) demonstrated the difference in thermal stability and refolding half-life times between triple helices that vary only in the frequency of glycine mutations at a particular position.

By differential scanning calorimetry and circular dichroism, Makareeva et al. (2008) measured and mapped changes in the collagen melting temperature (delta-T(m)) for 41 different glycine substitutions from 47 OI patients. In contrast to peptides, they found no correlation of delta-T(m) with the identity of the substituting residue but instead observed regular variations in delta-T(m) with the substitution location on different triple helix regions. To relate the delta-T(m) map to peptide-based stability predictions, the authors extracted the activation energy of local helix unfolding from the reported peptide data and constructed the local helix unfolding map and tested it by measuring the hydrogen-deuterium exchange rate for glycine NH residues involved in interchain hydrogen bonds. Makareeva et al. (2008) delineated regional variations in the collagen triple helix stability. Two large, flexible regions deduced from the delta-T(m) map aligned with the regions important for collagen fibril assembly and ligand binding. One of these regions also aligned with a lethal region for Gly substitutions in the alpha-1(I) chain.

Other Features

Dickson et al. (1975) reported a quantitative and qualitative abnormality of noncollagenous proteins of bone.

Lancaster et al. (1975) found a consistent morphologic abnormality of cultured skin fibroblasts: irregular packing of aggregated cells and an irregular tessellated appearance of individual fibroblasts. Boright et al. (1984) showed that dermal fibroblasts derived from individuals with OI type I take longer than control cells to reach confluency, have a lower cell density at stationary phase and have an abnormal cell shape as judged by the increased ratio of width to length. An increase in population doubling time of fibroblasts derived from individuals with the milder form of OI was also observed by Rowe and Shapiro (1982).

Inheritance

The mode of inheritance is autosomal dominant. Penetrance of blue sclerae is 100 percent, while penetrance of hearing loss is clearly age-dependent (Garretsen and Cremers, 1991). Paternal age effect for increased risk of new mutations has been documented although it appears to be considerably lower than, for example, in achondroplasia (100800). In 10 cases with OI type I presumed to have arisen by new mutation, the mean paternal age was increased by 2.1 years (Sillence et al., 1979), whereas in 38 other cases it was significantly increased by 2.9 years (Carothers et al., 1986).

Blumsohn et al. (2001) confirmed the presence of a small paternal age effect in apparently sporadic OI. The study evaluated patients born in England, Wales, and Scotland between 1961 and 1998. For 357 apparently sporadic cases among 730 eligible cases, the mean age of fathers at the birth of their children was 0.87 years greater than expected (p = 0.01). The relative risk was 1.62 for fathers in the highest quintile of paternal age compared with fathers in the lowest quintile.

Mapping

In all but 1 of 11 families with OI tarda, Sykes et al. (1986) found that the disorder segregated with either the COL1A1 locus or the COL1A2 locus. In 1 small family, segregation occurred with both genes, but this disorder clearly cannot be linked to both; had further meioses been available, the OI gene would probably have segregated independently of at least 1 of the 2 loci. Tsipouras (1987), also, concluded that mild OI is genetically heterogeneous and that 1 or more loci other than COL1A1 and COL1A2 may be involved in the causation of phenotypically indistinguishable autosomal dominant OI.

Sykes et al. (1990) studied segregation of the COL1A1 and COL1A2 genes in 38 dominant osteogenesis imperfecta pedigrees. None of the 38 pedigrees showed recombination between the OI gene and both collagen loci. All 8 pedigrees with OI type IV (166220) segregated with COL1A2. On the other hand, 17 type I pedigrees segregated with COL1A1 and 7 with COL1A2. The concordant locus was uncertain in the remaining 6 OI type I pedigrees. The presence or absence of presenile hearing loss was the best predictor of the mutant locus in OI type I families, with 13 of the 17 COL1A1 segregants and none of the 7 COL1A2 segregants showing this feature. By linkage analysis in 7 autosomal dominant osteogenesis imperfecta families in Italy, Mottes et al. (1990) showed that the COL1A1 gene was implicated in 2 families and the COL1A2 gene in 1 family with OI type I. The COL1A2 gene was implicated in 2 families with OI type IV. In 2 OI type I families, the molecular genetic data were insufficient for exclusion of one gene.

Molecular Genetics

Byers (1993) summarized that 'functional null' alleles are the most common genetic features of OI type I. The mechanism by which the synthesis of pro-alpha-1(I) chains is decreased remains a difficult problem to solve. A variety of mutations, such as deletion of an allele, promoter and enhancer mutations, splicing mutations, premature termination, as well as other mutations that result in the inability of pro-alpha-1(I) chains to assemble into molecules, would presumably result in the same biochemical picture and the same phenotype.

In some individuals, the decreased production of pro-alpha-1(I) chains by fibroblasts results from about half-normal steady-state levels of the mRNA (Rowe et al., 1985). Later studies on these cells indicated that there is a defect in splicing of the pre-mRNA of COL1A1 that prohibits transport of the product of the mutant allele to the cytoplasm; the ratio of pro-alpha-1(I) to pro-alpha-2(I) mRNA was 1:1 in the cytoplasm instead of the normal 2:1, whereas the ratio was 4:1 in the nucleus instead of the normal 2:1 (Genovese and Rowe, 1987). Furthermore, a novel species of alpha-1(I) mRNA present in the nuclear compartment was not collinear with a cDNA probe (Genovese et al., 1989). In another individual with OI type I, Stover et al. (1993) demonstrated a G-A transition in the first position of the splice donor site of intron 26 which resulted in inclusion of the entire succeeding intron in the mature mRNA that accumulated in the nuclear compartment; apparently because no abnormal pro-alpha-1(I) chains were synthesized from the mutant allele, the clinical phenotype of this individual was mild. In a large study, Willing et al. (1992) showed that among 70 individuals with OI type I 23 from 21 families were heterozygous at the COL1A1 polymorphic MnlI site. As shown by primer extension with nucleotide-specific chain termination, there was in each case marked diminution in steady-state mRNA levels from one COL1A1 allele. Loss of an allele through deletion or rearrangement was not the cause of the diminished COL1A1 mRNA levels. Only in one family has the causative mutation been identified; an A-G transition in the obligatory acceptor splice site of intron 16 resulted in skipping of exon 17 in the mRNA which represented only 10% of the total COL1A1 mRNA. Further, linkage studies in 38 additional families have demonstrated no evidence of deletion of those regions of the COL1A1 gene used for linkage analysis (Sykes et al. (1986, 1990)) and confirmed that most individuals with the OI type I phenotype have mutations linked to the COL1A1 gene. In some families, a similar phenotype is thought to result from mutations in the COL1A2 gene (Sykes et al. (1986, 1990); Wallis et al., 1986), but the clinical criteria by which the diagnosis of OI type I is made are not always clear. Willing et al. (1990) described a 5-bp deletion near the 3-prime end of one COL1A1 allele that resulted in a reading frame shift 12 amino acid residues from the normal terminus of the chain and predicted an extension of 84 amino acid residues beyond the normal termination site. Although the abnormal mRNA could be translated in vitro, it proved extremely difficult to identify the abnormal chains in cells; it appeared that although the mRNA was present in normal amount, the protein product was unstable. This mutation provides a model of how many different mutations in the COL1A1 gene could produce the OI type I phenotype by resulting in the synthesis of half the normal amount of a functional pro-alpha-1(I) chain.

In an effort to further understand the reasons for diminished COL1A1 transcript levels in OI type I, Willing et al. (1995) investigated whether mutations involving key regulatory sequences in the COL1A1 promoter, such as the TATAAA and CCAAAT boxes, are responsible for the reduced levels of mRNA. They used PCR-amplified genomic DNA in conjunction with denaturing gradient gel electrophoresis and SSCP to screen the 5-prime untranslated domain, exon 1, and a small portion of intron 1 of the COL1A1 gene. In addition, direct sequence analysis was performed on an amplified genomic DNA fragment that included the TATAAA and CCAAAT boxes. In a survey of 40 unrelated probands with OI type I in whom no causative mutation was known, Willing et al. (1995) identified no mutations in the promoter region and there was 'little evidence of sequence diversity among any of the 40 subjects.'

Although less common than 'functional null' allele mutations, there are several examples in which the synthesis of abnormal procollagen I molecules can produce the OI type I phenotype. In one family (Nicholls et al., 1984), cells cultured from the affected mother and son, but not those from the normal daughter, synthesized alpha-1(I)-chains bearing a cysteine residue within the protease-resistant domain of the collagen molecule, a region from which that residue is normally absent. Although it was initially thought that the cysteine substitution was at the X or Y position of the Gly-X-Y repeating unit of the alpha-1(I) chain in the carboxyl-terminal peptide CB6 (Steinmann et al., 1986), peptide sequence analysis and sequencing of the cDNA demonstrated that the mutation resulted in the substitution of a glycine by cysteine in position 1017 in the telopeptide, 3 amino acid residues from the carboxy-terminal to the end of the triple helix (Cohn et al., 1988; Labhard et al., 1988).

Other substitutions of cysteine for glycine within the triple helical domain of the alpha-1(I) chain at residue 94 (Starman et al., 1989; Nicholls et al., 1990; Shapiro et al., 1992; Byers, 1993) also produce mild forms of OI, perhaps compatible with OI type I (see, e.g., 120150.0002 and 120500.0038). Byers et al. (1983) described an isolated patient with a mild to moderate form of OI: blue sclerae, a height of 147 cm, deformity as a consequence of poor orthopedic treatment, and hearing loss. Her cells synthesized a pro-alpha-2(I) chain in which approximately 30 amino acid residues were deleted from the triple-helical domain, in the CB4 peptide, a domain in which phosphoproteins important to bone calcification may bind and in which crosslinks may form. Subsequent studies indicated that a point mutation at the consensus splice-donor site resulted in the skipping of exon 12 (amino acids 91-108) from about half the COL1A2 transcripts (Rowe et al., 1990). Zhuang et al. (1993) showed that deletion of 19 bp from +4 to +22 of intron 13 of COL1A2 caused skipping of exon 13 in about 88% of the transcripts, whereas 12% of the transcripts were normally spliced; procollagen I containing the mutated pro-alpha-2(I) chain had reduced thermal stability and was only poorly secreted from the cells.

A woman with 'postmenopausal osteoporosis' was reported by Spotila et al. (1991) to be heterozygous for a serine-to glycine substitution at position 661 of the alpha-2(I) triple-helical domain. Since her 3 sons, who inherited the mutation, had experienced fractures as adolescents, the diagnosis of 'mild OI cannot be fully excluded' according to the authors' view; one of the sons was homozygous for the mutation due to partial isodisomy for maternal chromosome 7 (Spotila et al., 1992). All these findings suggest that other point mutations in the COL1A1 gene, and perhaps in the COL1A2 gene (as suggested also by linkage studies), could lead to a phenotype similar to that produced by 'functional null' allele mutations.

Diagnosis

The diagnosis is based on clinical and genetic criteria. In sporadic cases, the diagnosis may be difficult, and secondary osteoporosis and nonaccidental injury has to be ruled out. In women with severe 'postmenopausal osteoporosis' careful clinical investigation and a thorough personal and family history quite often reveals OI type I. While the direct molecular characterization is not feasible in the majority of cases at present, demonstration of reduced synthesis of procollagen I by dermal fibroblasts is indicative for the disorder. Lynch et al. (1991) discussed the problem of making the prenatal diagnosis of OI type I on the basis of linkage.

De Vos et al. (2000) reported the achievement of a normal twin pregnancy after preimplantation genetic diagnosis for osteogenesis imperfecta type I. Because 2 blighted ova were seen on ultrasound at 7 weeks' gestation, the pregnancy was terminated. The female partner with OI type I carried a 1-bp deletion in exon 43 of the COL1A1 gene, resulting in a premature stop codon in exon 46. The nonfunctional allele was predicted to result in the synthesis of too little type I procollagen.

Byers et al. (2006) published practice guidelines for the genetic evaluation of suspected OI.

Clinical Management

Fractures in OI are treated with standard orthopedic procedures appropriate for the type of fracture and the age, and heal rapidly with evidence of good callus formation (sometimes with hypertrophic callus formation) and without deformity. Regular hearing evaluations after adolescence and early stapedectomy or stapedotomy are recommended. In postmenopausal women with OI, a long-term physical therapy program to strengthen the paraspinal muscles, together with estrogen and progesterone replacement, adequate calcium intake, and perhaps calcitonin or fluoride administration, may be specifically indicated (for review, see Steinmann et al., 1990).

Bembi et al. (1997) described the results of treatment of 3 children with OI type I with cyclic intravenous infusions of aminohydroxypropylidene bisphosphonate (pamidronate). Each of the children had repeated bone fractures and low bone density. The rationale for pamidronate therapy in OI is based on the fact that bisphosphonates inhibit osteoclastic bone resorption; this leads to increased bone density and possibly to reduced risk of fracture. Bembi et al. (1997) reported a clear clinical response over the 22- to 29-month treatments, with a striking reduction in the frequency of new fractures. They also observed an effect on bone density. There were no notable adverse effects during therapy.

In an uncontrolled observational study involving 30 children aged 3 to 16 years with severe osteogenesis imperfecta, Glorieux et al. (1998) administered pamidronate intravenously at 4- to 6-month intervals for 1.3 to 5.0 years. They observed a sustained reduction in serum alkaline phosphatase concentrations and in the urinary excretion of calcium and type I collagen N-telopeptide. Increases in the size of the vertebral bodies suggested that new bone had formed. The mean incidence of radiologically confirmed fractures decreased by 1.7 per year (P less than 0.001). Treatment with pamidronate did not alter the rate of fracture healing, the growth rate, or the appearance of growth plates. Mobility and ambulation improved in 16 children and remained unchanged in the other 14. The children with severe osteogenesis imperfecta treated by Glorieux et al. (1998) fell into the type III (259420) and type IV (166220) categories of osteogenesis imperfecta.

Marini (1998) commented that fluoride and calcitonin treatment in OI had proved unsuccessful. The bisphosphonates are synthetic analogs of pyrophosphate, a natural inhibitor of osteoclastic bone resorption. They have been useful in the treatment of osteoporosis, Paget disease of bone, and fibrous dysplasia.

Lee et al. (2001) performed a prospective open label study to determine the efficacy and safety of pamidronate in 6 children with OI (3 had OI type I, 2 had type III, and 1 had type IV). The dose was 1.5 mg/kg bimonthly over 12 to 23 months. The number of fractures decreased from median of 3 (range 1-12) to 0 fractures per year (range 0-4) and all patients experienced improved bone mineral density and decreased serum alkaline phosphatase.

Astrom and Soderhall (2002) performed a prospective observational study using disodium pamidronate (APD) in 28 children and adolescents (aged 0.6 to 18 years) with severe OI or a milder form of the disease, but with spinal compression fractures. All bone metabolism variables in serum (alkaline phosphatase, osteocalcin, procollagen-1 C-terminal peptide, collagen-1 teleopeptide) and urine (deoxypyridinoline) indicated that there was a decrease in bone turnover. All patients experienced beneficial effects, and the younger patients showed improvement in well-being, pain, and mobility without significant side effects. Vertebral remodeling was also seen. They concluded that APD seemed to be an efficient symptomatic treatment for children and adolescents with OI.

Lindsay (2002) reviewed the mechanism, effects, risks, and benefits of bisphosphonate therapy in children with OI. He stated that the clinical course and attendant morbidity for many children with severe OI is clearly improved with its judicious use. Nevertheless, since bisphosphonates accumulate in the bone and residual levels are measurable after many years, the long-term safety of this approach was unknown. He recommended that until long-term safety data were available, pamidronate intervention be reserved for those for whom the benefits clearly outweighed the risks.

Rauch et al. (2002) compared parameters of iliac bone histomorphometry in 45 patients (23 girls, 22 boys) with OI types I, III, or IV before and after 2.4 +/- 0.6 years of treatment with cyclical intravenous pamidronate (age at the time of the first biopsy, 1.4 to 17.5 years). There was an increase in bone mass due to increases in cortical width and trabecular number. The bone surface-based indicators of cancellous bone remodeling, however, were decreased. There was no evidence of a mineralization defect in any of the patients.

Rauch et al. (2003) evaluated the effect of intravenous therapy with pamidronate on bone and mineral metabolism in 165 patients with OI types I, III, and IV. All patients received intravenous pamidronate infusions on 3 successive days, administered at age-dependent intervals of 2 to 4 months. During the 3 days of the first infusion cycle, serum concentrations of ionized calcium dropped and serum PTH levels transiently almost doubled. Two to 4 months later, ionized calcium had returned to pretreatment levels. During 4 years of pamidronate therapy, ionized calcium levels remained stable, but PTH levels increased by about 30%. Rauch et al. (2003) concluded that serum calcium levels can decrease considerably during and after pamidronate infusions, requiring close monitoring especially at the first infusion cycle. In long-term therapy, bone turnover is suppressed to levels lower than those in healthy children. The authors stated that the consequences of chronically low bone turnover in children with OI were unknown.

Zeitlin et al. (2003) analyzed longitudinal growth during cyclical intravenous pamidronate treatment in children and adolescents (ages 0.04 to 15.6 years at baseline) with moderate to severe forms of OI types I, III, and IV and found that 4 years of treatment led to a significant height gain.

Rauch et al. (2006) assessed the effect of long-term pamidronate treatment on the bone tissue of children and adolescents with OI. Average areal bone mineral density (aBMD) increased by 72% in the first half of the observation period, but by only 24% in the second half. Mean cortical width and cancellous bone volume increased by 87% and 38%, respectively, between baseline and the first time point during treatment (P less than 0.001 for all changes). Rauch et al. (2006) concluded that the gains that can be achieved with pamidronate treatment appear to be realized largely in the first 2 to 4 years.

Rauch et al. (2006) studied the effect of pamidronate discontinuation in pediatric patients with moderate to severe OI types I, III, and IV. In the controlled study, 12 pairs of patients were matched for age, OI severity, and duration of pamidronate treatment. Pamidronate was stopped in one patient of each pair; the other continued to receive treatment. In the observational study, 38 OI patients were examined (mean age, 13.8 years). The intervention was discontinuation of pamidronate treatment for 2 years. The results indicated that bone mass gains continue after treatment is stopped, but that lumbar spine aBMD increases less than in healthy subjects. The size of these effects is growth dependent.

In lethal forms of osteogenesis imperfecta caused by mutation in either the COL1A1 gene or the COL1A2 gene, the mutations result in the synthesis of abnormal chains of procollagen that bind to normal chains synthesized by the same cells and destroy their biologic activity in a classic dominant-negative manner. Chamberlain et al. (2004) developed a strategy to inactivate the mutated alleles in cells of the bone marrow called mesenchymal stem cells (MSCs), or marrow stromal cells. They chose MSCs because these cells are easily obtained from a patient, they engraft and differentiate into many tissues after infusion in vivo, and allogeneic MSCs had produced promising results in a previous trial involving patients with osteogenesis imperfecta (Prockop et al., 2003; Horwitz et al., 2002). Chamberlain et al. (2004) designed a gene construct that targeted exon 1 of the COL1A1 gene. They predicted that, on insertion, the construct would both inactivate COL1A1 and confer resistance to the antibiotic neomycin. To insert the gene construct efficiently into MSCs, they used an adeno-associated virus as a vector. The results obtained with MSCs from 2 patients with osteogenesis imperfecta was highly encouraging. In 31 to 90% of the cells that became resistant to neomycin, the gene construct had inserted itself into either the wildtype or the mutated COL1A1 allele. In all cultures of the neomycin-resistant cells, most signs of the dominant-negative protein defect were corrected--apparently because the cells in which the mutated allele was inactivated began to produce an adequate amount of wildtype collagen. Most importantly, the quality of bones synthesized by the altered MSCs was improved. Prockop (2004) commented on the promising nature of the approach as well as some of the problems.

In a cohort of 540 individuals with OI studied longitudinally, Bellur et al. (2016) conducted a study to address whether cesarean delivery has an effect on at-birth fracture rates and whether an antenatal diagnosis of OI influences the choice of delivery method. They compared self-reported at-birth fracture rates among individuals with OI types I, III (259420), and IV (166220). When accounting for other covariates, at-birth fracture rates did not differ based on whether delivery was vaginal or by cesarean section. Increased birth weight conferred conferred higher risk for fractures irrespective of the delivery method. In utero fracture, maternal history of OI, and breech presentation were strong predictors for choosing cesarean delivery. The authors recommended that cesarean delivery should not be performed for the sole purpose of fracture prevention in OI, but only for other maternal or fetal indications.

Population Genetics

In the county of Fyn, where approximately 9% of the Danish population lives, Andersen and Hauge (1989) identified 48 patients with osteogenesis imperfecta, of whom 17 were born between January 1, 1970 and December 31, 1983. Of the 17, 12 had type I, 2 had type II, 2 had type III, and 1 had type IV. The point prevalence at birth was 21.8/100,000 and the population prevalence was 10.6/100,000 inhabitants. All ethnic and racial groups seem to be similarly affected (Byers, 1993).

History

Kozma (2008) provided a detailed historical review of skeletal dysplasias in ancient Egypt, with an example of presumed osteogenesis imperfecta.

Animal Model

Bonadio et al. (1990) reported that the heterozygous Mov-13 mouse, which has a murine retrovirus integrated within the first intron of The Col1a1 gene, is a good model for the mild autosomal dominant form of OI. The animals showed morphologic and functional defects in mineralized and nonmineralized connective tissue and progressive hearing loss.

Aihara et al. (2003) demonstrated that mice with a targeted mutation of the Col1a1 gene had ocular hypertension. They suggested an association between intraocular pressure regulation and fibrillar collagen turnover.