Calcified Tissue International

, Volume 90, Issue 6, pp 439–449 | Cite as

Recent Advances in Osteogenesis Imperfecta



“Osteogenesis imperfecta” is a term used to describe a group of genetic disorders of variable phenotype usually defined by recurrent fractures, low bone mass, and skeletal fragility. Most cases are associated with mutations in one of the type I collagen genes, but in recent years several other forms have been identified with recessive inheritance. In most instances the latter result from mutations in genes encoding proteins involved in type I collagen’s complex posttranslational modification or in genes regulating bone matrix homeostasis. This article reviews the recent discoveries and an approach to classification and diagnosis. Bisphosphonates are widely used in patients with osteogenesis imperfecta, but some important questions about their optimal usage, their utility in children and adults with milder phenotypes, and their potential adverse effects are not yet resolved.


Bisphosphonate Matrix protein Osteogenesis imperfecta Pediatric bone disease Type I collagen 

Osteogenesis Imperfecta

The term “osteogenesis imperfecta” (OI) encompasses a group of genetic disorders usually defined by recurrent fractures, low bone mass, and skeletal fragility. Most cases are associated with mutations in the genes encoding either type I collagen (the most abundant and main structural protein of bone) or (in about 10% of cases) genes encoding proteins involved in type I collagen’s complex posttranslational modification and intracellular trafficking. However, they are not synonymous with OI as a number of genetic disorders affecting type I collagen or its posttranslational modification that have skeletal phenotypes are not usually included under the OI rubric. These include Caffey disease and the kyphoscoliosis, arthrochalasia, and dermatosparaxis types of the Ehlers-Danlos syndrome (types VI, VIIA/B, and VIIC, respectively). In addition, some genetic disorders of skeletal fragility included in classifications of OI are the consequence of mutations in key osteoblast genes (LRP5 and SP7) that code for proteins concerned with matrix homeostasis and are not directly related to collagen metabolism and matrix structure.

Type I Collagen

Type I collagen is initially synthesized in the rough endoplasmic reticulum as a precursor molecule (type I procollagen) that combines two proα1(I) and one proα2(I) peptide chains (coded by COL1A1 and COL1A2, respectively) in a triple helix. Proα1(I) and proα2(I) have similar structures, with a core triple-helical domain of 1,014 amino acids composed of uninterrupted Gly-Xaa-Yaa tripeptide repeats, flanked by propeptides at both the N- and C-terminal ends. During and after translation the three chains undergo extensive modification. Prolyl-4-hydroxylase converts virtually all Y-position proline residues to 4-hydroxyproline, an alteration that is essential for thermal stability of the assembled trimer. In the absence of this modification, the trimer melts (i.e., the individual chains unfold from the stable triple helix, at about 27°C, whereas with full hydroxylation the melting temperature is about 42°C). Some Y-position lysine residues within the triple-helical domain are hydroxylated by the enzyme lysyl hydroxylase-1 and glucose and galactose groups added by glycosyltransferases. Hydroxylation of these triple-helical residues is part of the pathway to form stable complex intermolecular cross-links that provide the tensile strength in tissues. Most of these modifications are completed during translation and occur on individual chains. If there is a delay in triple-helix folding, the process can continue but the physical properties of the chains and molecules are altered and contribute to an OI phenotype.

The three chains—proα1(I)2 and proα2(I)—that form a trimer interact through regions in the carboxyl-terminal propeptide of each chain. This creates the unusual situation in which the full-length chain must be maintained in an unfolded state while the carboxyl-terminal propeptides fold, associate, and then begin the process of triple-helix formation. Propogation of the collagen triple helix requires a number of enzymes and molecular chaperones to ensure correct folding and trimerization. These include peptidyl disulfide isomerase, which also forms part of the prolyl 4-hydroxylase complex and likely involves prolyl peptidyl cistrans isomerase B (also known as cyclophilin B). This protein can act on its own, to assist in the folding around prolyl residues, such as those in the carboxyl-terminal propeptide adjacent to cysteine residues, and as part of a complex that includes two additional proteins, cartilage-related protein and prolyl 3-hydroxylase, to modify certain triple-helical prolines. The function of this last process is not yet entirely clear, but when the complex is missing, the propagation of the triple helix is altered and modification of the chains increased.

Disulfide bonds between the carboxyl-terminal region of the chains require protein-disulfide isomerase and act to secure the three chains in a trimer. Lysine residues outside the major triple-helical domain of type I collagen and needed for the formation of mature intermolecular cross-links are hydroxylated by lysyl hydroxylase-2. These complex modifications, which are necessary for correct folding, assuring thermal stability of the triple helix, and cross-link formation between collagen molecules once they are secreted into the matrix, need to take place in an orderly and timely sequence; and various chaperone proteins, including HSP47 and FKBP65, help to regulate this process. Procollagen trimers are then transported via the Golgi network and packaged into membrane-bound organelles, where lateral aggregation, the intial phase of fibril formation, occurs. As secretion occurs, the procollagen molecules are further processed into mature type I collagen molecules by proteolytic cleavage of the N- and C-terminal propeptides (by the enzymes ADAMTS-2 and BMP1, respectively). Finally, the trimers are assembled into collagen fibrils and fibers [1, 2, 3] and anchored in those positions by intermolecular lysine-derived cross-links in a process that is begun by modification of specific residues by lysyl oxidase.


The Sillence classification, published more than 30 years ago, was the first systematic classification of OI phenotype [4]. Although the original numbering system is somewhat counterintuitive in that the ascending numbers do not correlate with severity, the distinctions between the nondeforming (type I), moderate (type IV), severe or progressively deforming (type III), and perinatal lethal forms (type II) remains clinically useful [4].

In recent years, the genetic complexity of the molecular basis of OI has become increasingly evident, and at the same time the extensive phenotypic variation arising from single loci has been documented clearly. The International Skeletal Dysplasia Society has suggested that it is untenable to try to maintain tight correlations between ‘‘Sillence types’’ and their molecular bases. They recommend retaining the essence of the Sillence classification to describe the phenotypic severity in OI, to free the clinical classification from direct molecular reference, and to limit the proliferation of new numbered ‘‘OI types’’ with each new genetic discovery [5]. When faced with a patient with possible OI, the clinician should firstly carefully define the phenotype and the inheritance; taken together these will usually point to the path by which one can identify the likely genetic defect.

Defining the Phenotype

The important points to be taken from the history and physical examination include a detailed family history of bone disease; the presence of fractures detected in utero or in the neonatal period; the nature, chronology, and outcome of subsequent fractures; growth velocity and current height and skeletal proportions; the presence and progression of long bone deformity, acetabular protrusion, and kyphoscoliosis. The distinction between mild, moderate, and severe phenotypes is broadly based on the number of fractures, the degree of deformity and growth impairment, and the age at which abnormalities are first recognized. Other phenotypic features of importance are head circumference, joint mobility (measured, e.g., on the Beighton scale) or joint contractures, scleral color, hearing, cardiac murmurs, and tooth abnormalities, particularly dentinogenesis imperfecta (Fig. 1a–c).
Fig. 1

Clinical features of OI in adults. a Blue sclerae in a woman and her daughter with OI of mild phenotype. The scleral color can vary substantially between patients and is commonly darker in infancy. b Dentinogenesis imperfecta in an adult patient with a mild phenotype. The lower teeth are commonly more severely affected than the upper. As is typically the case with dentinogenesis imperfecta, skin fibroblast studies indicated the production of both normal and abnormal collagen. c Hypermobility in an adult patient with a moderate-to-severe OI phenotype. d OI presenting as “postpartum osteoporosis.” This patient had a mild phenotype associated with a missense mutation in COL1A2. Four months after her first pregnancy she sustained a single vertebral fracture. Three weeks after her second pregnancy, age 38, she developed severe back pain with height loss. The radiograph shows multiple vertebral compression fractures

The standard biochemical tests of bone turnover are generally unhelpful in diagnosis. There are a few exceptions: plasma concentrations of procollagen-1 N-propeptide and procollagen-1 C-propeptide are low in patients with haploinsufficiency [6], and the ratios of pyridinoline to deoxypyrodinoline in urine are altered in the Bruck syndrome variants of OI (due to mutations in FKBP10 or PLOD2) where the formation of lysine cross-links is impaired. Plain radiography is very useful in defining the phenotype, but bone density testing is of little diagnostic value and can be difficult to perform and interpret when there is deformity or short stature. It is, however, commonly used in monitoring responses to bisphosphonate treatment.

Dominantly Inherited OI

Most patients with OI (~ 90%) have mutations in one of the type I collagen genes, COL1A1 or COL1A2. These are large genes (51 and 52 exons, respectively), and disease-causing mutations occur all along both. COL1A1 and COL1A2 mutations are dominantly inherited, and the phenotype can vary from the very mild to in utero lethal. It is generally not necessary to undertake in vitro studies of type I collagen or to sequence these genes if the diagnosis is clear from the family history and phenotype. However, further investigation is indicated if there are atypical features or if there is uncertainty about the diagnosis or a need for genetic counseling about recurrence risk, prenatal diagnosis, or preimplantation genetic diagnosis.

There are two general classes of mutations in the type I collagen genes that result in OI: those that cause a quantitative defect with synthesis of structurally normal type I procollagen at about half the normal amount (haploinsufficiency) and those that result in synthesis of a structurally abnormal collagen. The former are usually the result of premature termination codons in one COL1A1 molecule that initiate nonsense-mediated decay of the mRNA from the affected allele. These generally result in a mild, nondeforming phenotype with blue sclerae (type I in the original Sillence classification) [1]. The class of mutations that change the protein sequence in the triple-helical domain has a wide phenotypic range (from mild to lethal). The most prevalent mutations result in substitution for one of the invariant glycine residues that have a critical role in helix formation (glycine is the only amino acid residue small enough to be accommodated in the sterically restricted inner aspect of the helix). Other mutations cause alterations in splice sites that can lead to exon skipping, intronic inclusion, or activation of cryptic sites in introns or exons. Mutations affecting the C-propeptides of either of the chains are also common.

The test most frequently employed to distinguish these two classes uses cultured skin fibroblasts to examine the secretion of type I procollagen and the electrophoretic mobility of its constituent chains. With haploinsufficiency for COL1A1 only normal collagen is made but in reduced amount. Mutations that alter the sequence of the proα chains often slow folding of the triple helix and allow increased posttranslational modification so that the electrophoretic mobility of the constituent chains is altered (Fig. 2). This test may be normal in cells from individuals with mutations in the C-propeptide region that simply prohibit chain interaction and in some of the recently identified recessive types of OI (e.g., with FKBP10, PLOD2, SERPINH1, SERPINF1, SP7, and LRP5 mutations).
Fig. 2

Secretion and electrophoretic mobility of type I collagen from cultured skin fibroblasts. Cells from individuals with mild (type I, lanes 5 and 10), lethal (type II, lanes 4 and 9), and moderate (type IV, lanes 2 and 7) OI and controls (lanes 1, 3, 6, 8) were plated at confluent density, allowed to attach and spread, and then labeled with [3H]proline overnight in the presence of ascorbic acid. Proteins were harvested separately from the medium and cell layer, precipitated with ethanol, and then separated on 5% SDS polyacrylamide gels under reducing conditions to separate the constituent chains of type I and type III procollagen. In the medium from the mild OI (type I) cells, the ratio of proα1(I) to proα1(III) is markedly reduced and the amount of proα2(I) chains reduced, but there is no intracellular storage of the chains. This reflects haploinsufficiency for the COL1A1 mRNA and the requirement that there must be at least two proα1(I) chains in each type I procollagen molecule. The result is decreased production of type I procollagen. In cells from the individual with lethal OI (type II) there is a marked shift in the electrophoretic mobilities of some of the proα1(I) and proα2(I) chains (lane 4). In addition, there is significant intracellular retention of the abnormal molecules inside the cells (lane 9). These findings reflect the substitution of a glycine near the carboxyl-terminal end of the triple-helical domain by a larger amino acid and the slow folding and increased posttranslational modification of the unwound chains. The sample from the patient with moderate OI (type IV) has a much more subtle defect in chain mobility, represented by the blurring of the space between the proα1(I) and proα1(III) chains, and illustrates one of the aspects of the “art” of interpreting protein studies

In the future it is probable that the analysis of cultured cells will be relegated to a second-tier investigation for the diagnosis of OI. There are problems with the test: first, it is invasive; second, it requires facilities in which to grow cells and experienced personnel to understand and interpret the results; and third, the time to results is often 8–10 weeks. The ease of genomic sequence analysis has led to the development of gene-based tests for identification of mutations in all the genes now associated with OI. This strategy uses a single approach, which can be adapted to new technologies. In addition, the data are digital and qualitative rather than analogue and quantitative. The only “art” required is knowledge of the effects of different classes of mutations and the ability to transmit the interpretations in a facile manner. Further, the results of genetic studies are readily translated into family studies, prenatal diagnosis, and preimplantation diagnosis and allow a base for genotype/phenotype analysis. Some mutations are missed by this approach, but the analysis of proteins and mRNA from cultured cells can often point to the effects of mutations and then to the target genes and locations in the genes. In addition, the effects of splice-site mutations cannot be predicted from sequence alone, although, to some extent, can be inferred from correlation with phenotype. All in all, DNA sequence analysis seems likely to provide more “bang for the buck” or a richer set of data for prediction than studies of collagens. A major limitation is access to the full library of mutations identified by all the diagnostic laboratories.

With COL1A1 and COL1A2 mutations the relationship between genotype and phenotype is complex. A number of recent publications have addressed this topic, and the data are summarized in Table 1 [1, 7, 8, 9, 10]. In brief, in patients with haploinsufficiency the phenotype is milder in almost all respects than those producing an abnormal collagen (the exception is the high prevalence of hearing impairment in the former, but this may in part reflect the longer life expectancy of these patients). The presence of clinically overt dentinogenesis imperfecta almost always indicates abnormal collagen. Most lethal mutations arise from substitutions for glycine in the triple-helical domain, with an almost 2:1 preference for mutations in COL1A1 [1].
Table 1

Genotype–phenotype relationships in OI resulting from COL1A1 or COL1A2 mutations

Haploinsufficiency: compared to helical-domain mutations

 Nonlethal, nondeforming, taller

 Blue sclerae usual

 Normal life expectancy

 Fewer fractures—fracture rate falls after adolescence

 No dentinogenesis imperfecta

 Fewer skull-base abnormalities

 Sensorineural deafness common

 Greater size-adjusted lumbar spine areal BMD

 Greater cortical bone width

 Wormian bones less common (28% vs. 81%)

Helical-domain mutations—with production of an abnormal collagen

 82% are glycine substitutions: → serine, arginine, or cysteine are the most common (~78% of proα1(I) and 63% of proα2(I) substitutions)

  33% of glycine substitutions are lethal in proα1(I), 20% lethal in proα2(1)

  Substitutions of large amino acid residues (arginine, valine, glutamic acid, aspartic acid) for glycine in proα1(I) beyond position ~200 are usually lethal

  Glycine substitutions at N-terminal end of either proα1(I) or proα2(I) are nonlethal and not associated with dentinogenesis imperfecta

 18% are splice-site, exon-skipping, intronic inclusion, or activation of cryptic sites—rarely lethal

C-propeptide mutations

 Mutations in proα1(I) severe to lethal, in proα2 (I) mild to moderate phenotype

Data compiled from references 1, 7–10, 30

There are some rare examples of apparently specific phenotypes associated with genotypes—for example, mutations at the C-peptide cleavage site in COL1A1 may result in a relatively high bone mass with skeletal fragility [11], and substitutions for glycine near the very end of the triple-helical domain of proα2(I) encoded in exon 49 of the COL1A2 gene have been associated with brachydactyly and intracerebral hemorrhage, in addition to a severe OI phenotype [12].

To date, only one dominantly inherited OI variant not related to COL1A1 and COL1A2 mutations has been described. This variant (type V) has a mild-to-moderate phenotype and is distinguished clinically by hyperplastic callus formation (particularly in the lower limbs), calcification of the forearm interosseous membrane (causing limitation of pronation/supination), anterior dislocation of the radial head and a radiodense metaphyseal band immediately adjacent to the growth plate in growing patients, white sclerae, and no dentinogenesis imperfecta. Bone biopsies reveal irregular, mesh-like lamellation [13, 14]. The underlying genetic locus has not yet been identified.

Recessively Inherited OI

This is a fast-moving field; in the past decade the genetic bases of 10 new OI variants have been discovered, seven (or possibly eight) of which result from mutations in genes encoding proteins involved in the posttranslational modification of type I procollagen [15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28]. The rate of discovery shows no sign of slowing, with the technique of exome sequencing now permitting the identification of genetic causes in families of small size. One of the main interests in these variants is that they help to define the minimal degree of change to type I collagen that can result in an OI phenotype, and that could eventually lead to effective therapies. Phenotypic data on the recessively inherited forms of OI is summarized in Table 2, although it should be noted that for some of these disorders very few individuals have been described, so our appreciation of the clinical spectrum is likely to change as further cases are found.
Table 2

Autosomal recessive forms of OI





Skeletal phenotype

Wormian bones

Scleral color

Dentinogenesis imperfecta


Other features


Genes regulating matrix structure



Cartilage-associated protein




Light blue



Rhizomelia, popcorn metaphyses; head circumference small




Prolyl 3-hydroxylase 1




Light blue



Impaired calvarial calcification, popcorn metaphyses




Peptidyl-prolyl isomerase B










FK506-binding protein 10 (FKBP65)

Bruck 1






Contractures, scoliosis, acetabular protrusion




Telopeptide lysyl hydroxylase

Bruck 2










Heat shock protein 47







Renal stones




Bone morphogenetic protein 1










Pigment-derived epithelium factor







Hyperosteoidosis with abnormal lamellation


Genes regulating matrix homeostasis



LDL receptor-related protein 5

Osteoporosis pseudoglioma






Blindness; heterozygotes have low bone mass











Delayed tooth eruption


Collagen-Related Genes

In patients at the severe-to-lethal end of the OI spectrum the question of dominant inheritance is, for obvious reasons, rarely put to the test; and in Sillence’s original classification the perinatal lethal and progressive deforming types (types II and III) were thought to be recessively inherited. However, it was not until 2007 that mutations in CRTAP were identified in patients without mutations in COL1A1 or COL1A2 but with excess posttranslational modification of type I collagen, indicative of delayed folding of the triple helix [15, 16]. CRTAP encodes cartilage-related protein, part of the three-protein complex responsible for the 3-hydroxylation of proline at position 986 of the triple helix in the proα1(I) chain. Very soon after, patients with similar phenotypes with mutations in the genes for the other components of this complex, LEPRE1 (prolyl 3-hydroxylase, PH3) and PPIB (cyclophilin B), were identified [17, 18]. It is not clear that the importance of this enzyme complex is due solely to its effect on prolyl 3-hydroxylation. The genetic test of substituting the proline 986 residue has not been done either by nature or in the laboratory. The 3-hydroxylation of additional X-position proline residues may be important to make normal molecules, and cyclophilin B, a prolyl cistrans isomerase, may have other targets in type I procollagen chains; thus, the effects of mutations in any of these genes are likely to be biochemically complex.

Bruck syndrome describes the occurrence of neonatal fractures with contractures of the legs (and, in some cases, the arms), which is occasionally misclassified as arthrogryposis. Two causative genes, FKBP10 (Bruck 1) and PLOD2 (Bruck 2), have been identified [19, 20, 21, 22]. FKBP10 mutations seem to be the more common, and as more cases have come to light, a milder phenotype presenting in late childhood or adolescence with long bone fractures, acetabular protrusion, and scoliosis has been recognized [21]. FKBP65, the protein product of FKBP10, is a prolyl cistrans isomerase which seems to have multiple substrates, among them being lysyl hydroxylase 2 (encoded by PLOD2) and perhaps LH1 and HSP47 (see the following). The effects of mutations in each gene probably share phenotypic features because of this interaction. One child with a severe phenotype has been described with homozygosity for a missense mutation in the gene SERPINH1, which encodes another collagen chaperone, HSP47 [23].

Among the most recent discoveries are the association of mutations in the gene SERPINF1 with a variant known as type VI OI [24, 25]. SERPINF1 encodes pigment epithelium-derived factor (PEDF), a secreted glycoprotein of uncertain function in bone; and it is not clear at present whether it acts on collagen metabolism or in some other way. PEDF can be measured in normal serum and is undetectable in the serum of patients with this variant of OI. Bone from these individuals has the distinctive feature of a marked increase in osteoid, with an unusual “fish-scale” pattern when viewed under polarized light. In addition, the response to bisphosphonate therapy in affected children may be disappointing in terms of reducing fracture rates and improving mobility [25].

Recessive disorders occur most commonly in areas of the world where consanguineous marriage takes place or where there are mutations in relatively small founder populations. Many instances of the latter are seen in recessively inherited OI. Examples include the splice-site mutation in LEPRE1 (IVS5 + 1G-T) seen both in West Africa and in people of African descent dispersed to the Americas (through the Atlantic slave trade) [15, 16] and the insertion mutation in FKBP10 (c.948dupT) seen in Samoa [21]. Knowledge of such locally occurring mutations can significantly aid diagnosis.

Non-Collagen-Related Genes

The osteoporosis pseudoglioma syndrome (originally designated the ocular form of OI) is an example of how genetic discoveries can advance understanding of bone metabolism and open up therapeutic possibilities. This disorder is due to mutations in the gene LRP5, which encodes lipoprotein receptor protein 5 (LRP5), a key regulator of osteoblast function. LRP5 affects bone accrual during growth and is important for the establishment of peak bone mass but is not directly related to type I collagen metabolism. The skeletal phenotype of the osteoporosis pseudoglioma syndrome is relatively mild, but the distinguishing feature is the ocular involvement, which is due to failure of the normal involution of the hyaloid vasculature [27]. The remnant forms a fibrous mass (“pseudoglioma”) in the globe, causing retinal detachment and blindness; nearly all subjects have near complete loss of vision by the age of 25 years [28]. Heterozygosity for LRP5 loss-of-function mutations differs from other recessive forms of OI in that carriers clearly have an osteopenic phenotype [28]. Some patients with familial, idiopathic, or juvenile osteoporosis are heterozygous for LRP5 mutations. Of note is that dominant gain-of-function mutations in LRP5 give rise to a dense bone phenotype because of activation of the same pathway in bone that is diminished by the recessive loss-of-function mutations.

More recently, a child with a moderate OI phenotype has been identified with homozygous mutations in SP7. This gene encodes osterix, a transcription factor specifically expressed in osteoblasts in the developing skeleton [29].

OI in Adults

Children with severe phenotypes have significantly reduced life expectancy, predominantly because of problems associated with respiratory insufficiency and pulmonary hypertension, secondary to kyphoscoliosis and small lung volumes. Although survival of patients with severe phenotypes into adulthood is not uncommon only ~ 20% survive beyond the age of 40 [30]. Thus, most patients who survive into adulthood have a mild or moderate phenotype. In these patients the fracture rate typically falls after puberty (as it does for most forms of childhood osteoporosis; the strength of long bones is proportional to the fourth power of their external radius, so bones get stronger as they grow wider). Vertebral fractures can occur in the postpartum period (Fig. 1d), and the fracture rate for both vertebral and long bones increases after the menopause in women [31] and in later life in men.

Deafness is a common problem for adults, particularly those with a mild phenotype due to haploinsufficiency. About one-third of such patients are affected by the age of 30 and one-half by the age of 50, although the proportion affected does not change much after that age. Hearing impairment shows distinct familial trends, being common in some families and infrequent in others [32]. Cardiac valve dysfunction (aortic or mitral regurgitation) is a recognized but uncommon feature seen in people with OI, but the mechanistic relationship between the two remains unclear. Valve replacement may be necessary, but the surgery is often complicated and should be undertaken in centers with experience [33].

Bisphosphonate therapy in adults with OI has been explored in three relatively small randomized controlled trials: using intravenous neridronate, given 3-monthly for 2 years [34]; oral alendronate, given daily for 3 years [35]; or oral risedronate, given weekly for 2 years [36]. All of these studies showed that bisphosphonate treatment had statistically significant effects, increasing bone density at the spine and hip, and decreasing bone turnover markers; but there were was no difference in the fracture rate, which was generally low. Given the relative rarity of OI, it will be difficult to demonstrate efficacy at fracture reduction for any intervention.

Bisphosphonate Therapy in Children and Adolescents

Physiotherapy, rehabilitation, and orthopedic surgery are the mainstay of treatment for children and adolescents with OI; and the best results are obtained when undertaken by skilled multidisciplinary teams. Intravenous bisphosphonates were first suggested as treatment to improve bone fragility in children with severe OI 25 years ago [37], and although not subjected to randomized, placebo-controlled trials, bisphosphonate treatment has rapidly become established as a standard of care. Compared to historical controls, intravenous bisphosphonate therapy is associated with improvements in bone pain, well-being, longitudinal growth and muscle strength, and vertebral and long bone mass as well as with a reduced fracture rate [38]. In an important series of studies the Montreal group has shown clearly how bisphosphonate therapy has effects on both trabecular and cortical bone mass. In OI, trabeculae are reduced in number and abnormally thin. With bisphosphonate treatment, the number (but not the thickness) of trabeculae is increased. During endochondral growth most primary trabeculae are lost in the conversion of primary into secondary spongiosa, but bisphosphonate treatment inhibits the resorption of primary trabeculae, permitting more to survive as secondary trabeculae [38]. Because of slow periosteal bone formation, the long bones in OI are typically narrow (although this is often partially compensated by a relative narrowing of the marrow cavity that conserves cortical bone width). During normal growth, cortical width is determined by bone modeling in which bone resorption at the endosteal surface takes place in parallel with periosteal new bone formation. Bisphosphonates inhibit the former process (but not the latter), permitting an increase in cortical thickness and, thus, improvement in bone strength [38]. Bisphosphonates do not increase bone width, and of course, if an abnormal collagen is produced, then bone quality will also remain unaffected. In adults, bisphosphonates increase bone mineralization and decrease bone turnover [39]; but in OI the bones are already hypermineralized [40], so this mechanism is unlikely to contribute to any improvement in bone strength.

The mode of action of bisphosphonates is the same in all forms of OI and, indeed, probably across all varieties of childhood osteoporosis. Intravenous pamidronate is the most widely used bisphosphonate; but the effects are generic, and regimens employing longer-acting agents such as zoledronate are being used increasingly.

Bisphosphonate therapy is not without side effects. In addition to the well-known first dose reaction, uveitis has been reported; and in children prolonged bisphosphonate use can impair metaphyseal modeling [41]. Intermittent intravenous therapy produces the characteristic metaphyseal lines, where remnants of calcified cartilage from the growth plate have accumulated (Fig. 3a–c). Low bone turnover induced by bisphosphonates can impair bone healing, particularly after corrective osteotomy [42]. Several important questions concerning bisphosphonate treatment of moderate to severe OI in children remain unresolved. These include how treatment should be monitored, whether the regimen should be modified for long-term use, and if and when treatment should be discontinued.
Fig. 3

Effects of prolonged bisphosphonate treatment on bone. a Femoral radiograph from a 2.5-year-old boy with severe OI who was treated with intravenous bisphosphonates from the age of 2 months. The femoral diaphysis is narrow (typical in OI), but the metaphysis is wide because of bisphosphonate-associated impairment of bone modeling. Horizontal white lines coincide with his pamidronate infusions. b Upper tibia and femur from the same child age 10. He had been treated with pamidronate infusions from age 2 months to 6 years and with zoledronate infusions from the age of 8 onward. The bone laid down in the 2 years off bisphosphonate treatment is notably less dense. The junction between treated and nontreated bone may be vulnerable to fracture [47]. c A transiliac bone biopsy from the same child aged 10 (hematoxylin and eosin stain) showing extensive retention of mineralized cartilage (blue stain) with relatively little bone tissue (pink stain). Calcified cartilage contributes to greater “bone density” but may not be resistant to fracture. Note also the giant osteoclasts detached from the surface of bone (arrows) that are a common finding in bisphosphonate-treated bone [48]

Debate continues on the use of bisphosphonates in children with mild OI and whether oral administration is just as effective as intravenous. Children with mild OI have less to gain from treatment and potentially more to lose from adverse events. Reports on two randomized controlled trials of oral bisphosphonates in children with predominantly (but not exclusively) mild forms of OI have been recently published [43, 44]. These showed that, at the doses given, both oral risedronate and alendronate increased spinal bone density and reduced bone turnover markers, but there was no improvement in fracture rate over a 2-year period. A randomized controlled trial of oral risedronate in children with moderate or severe OI had similar results, but there was a suggestion that the active treatment slowed the progression of bone deformity [45]. The authors pointed out that in both their control and bisphosphonate-treated groups the fracture rate fell with increasing age, emphasizing the challenge that future studies will need to be powered adequately to demonstrate a fracture outcome.

The use of anabolic agents to increase bone mass and size in children with severe forms of OI is an attractive theoretical option, but there are formidable practical problems. Parathyroid hormone is contraindicated in childhood because of concerns about the possibility of inducing osteosarcoma, so to date most attention has been focused on growth hormone. Anabolic agents generally increase bone turnover, which makes the development of deformity more likely, so the coadministration of an inhibitor of bone resorption would most likely be necessary. One small trial has looked at the effects of 1-year growth hormone treatment in conjunction with neridronate vs. neridronate alone [46]. The growth hormone–treated subjects had greater increases in bone mass at various sites and improved growth velocity, but the study was underpowered to detect any difference in fracture rates. Experimental approaches such as bone marrow transplantation, stem cell transplantation, and correction of the mutated gene may eventually come to fruition but are not currently ready for clinical trial.


There have been substantial advances in the understanding of OI in recent years. The main progress has been in understanding the genetic bases of this heterogenous group of disorders that has provided better information for genetic counseling and opportunities for prenatal diagnosis. The mode of action and risks and benefits of bisphosphonate therapy are being clarified. Particular challenges for future OI research will be to design therapeutic trials that can convincingly demonstrate effects on deformity and fracture and to determine the role of mutations in determining responses to treatment.



My sincere thanks go to Dr. Peter Byers, University of Washington, for his great help in preparing this article and for providing Fig. 2. The radiographs and the histology in Fig. 3 are reproduced with kind permission of Dr. Paul Hofman and Dr. Michael Dray, respectively.


  1. 1.
    Marini JC, Forlino A, Cabral WA, Barnes AM, San Antonio JD, Milgrom S, Hyland JC, Körkkö J, Prockop DJ, De Paepe A, Coucke P, Symoens S, Glorieux FH, Roughley PJ, Lund AM, Kuurila-Svahn K, Hartikka H, Cohn DH, Krakow D, Mottes M, Schwarze U, Chen D, Yang K, Kuslich C, Troendle J, Dalgleish R, Byers PH (2007) 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. Hum Mutat 28:209–221PubMedCrossRefGoogle Scholar
  2. 2.
    Canty EG, Kadler KE (2005) Procollagen trafficking, processing and fibrillogenesis. J Cell Sci 118:1341–1353PubMedCrossRefGoogle Scholar
  3. 3.
    Krane SM (2008) The importance of proline residues in the structure, stability and susceptibility to proteolytic degradations of collagens. Amino Acids 35:703–710PubMedCrossRefGoogle Scholar
  4. 4.
    Sillence DO, Senn A, Danks DM (1979) Genetic heterogeneity in osteogenesis imperfecta. J Med Genet 16:101–116PubMedCrossRefGoogle Scholar
  5. 5.
    Warman ML, Cormier-Daire V, Hall C, Krakow D, Lachman R, LeMerrer M, Mortier G, Mundlos S, Nishimura G, Rimoin DL, Robertson S, Savarirayan R, Sillence D, Spranger J, Unger S, Zabel B, Superti-Furga A (2011) Nosology and classification of genetic skeletal disorders: 2010 revision. Am J Med Genet 155A:943–968PubMedCrossRefGoogle Scholar
  6. 6.
    Cundy T, Horne A, Bolland M, Gamble G, Davidson J (2007) Bone formation markers in adults with mild osteogenesis imperfecta. Clin Chem 53:1109–1114PubMedCrossRefGoogle Scholar
  7. 7.
    Rauch F, Lalic L, Roughley P, Glorieux FH (2010) Genotype–phenotype correlations in nonlethal osteogenesis imperfecta caused by mutations in the helical domain of collagen type I. Eur J Hum Genet 18:642–647PubMedCrossRefGoogle Scholar
  8. 8.
    Rauch F, Lalic L, Roughley P, Glorieux FH (2010) Relationship between genotype and skeletal phenotype in children and adolescents with osteogenesis imperfecta. J Bone Miner Res 25:1367–1374PubMedGoogle Scholar
  9. 9.
    Semler O, Cheung MS, Glorieux FH, Rauch F (2010) Wormian bones in osteogenesis imperfecta: correlation to clinical findings and genotype. Am J Med Genet A 152:1681–1687Google Scholar
  10. 10.
    Cheung MS, Arponen H, Roughley P, Azouz ME, Glorieux FH, Waltimo-Siren J, Rauch F (2011) Cranial base abnormalities in osteogenesis imperfecta: phenotypic and genotypic determinants. J Bone Miner Res 26:405–413PubMedCrossRefGoogle Scholar
  11. 11.
    Faqeih E, Roughley P, Glorieux FH, Rauch F (2009) Osteogenesis imperfecta type III with intracranial hemorrhage and brachydactyly associated with mutations in exon 49 of COL1A2. Am J Med Genet A 149:461–465Google Scholar
  12. 12.
    Lindahl K, Barnes AM, Fratzl-Zelman N, Whyte MP, Hefferan TE, Makareeva E, Brusel M, Yaszemski MJ, Rubin CJ, Kindmark A, Roschger P, Klaushofer K, McAlister WH, Mumm S, Leikin S, Kessler E, Boskey AL, Ljunggren O, Marini JC (2011) COL1A1 C-propeptide cleavage site mutations cause high bone mass osteogenesis imperfecta. Hum Mutat 32:598–609PubMedCrossRefGoogle Scholar
  13. 13.
    Glorieux FH, Rauch F, Plotkin H, Ward L, Travers R, Roughley P, Lalic L, Glorieux DF, Fassier F, Bishop NJ (2000) Type V osteogenesis imperfecta: a new form of brittle bone disease. J Bone Miner Res 15:1650–1658PubMedCrossRefGoogle Scholar
  14. 14.
    Cheung MS, Glorieux FH, Rauch F (2007) Natural history of hyperplastic callus formation in osteogenesis imperfecta type V. J Bone Miner Res 22:1181–1186PubMedCrossRefGoogle Scholar
  15. 15.
    Morello R, Bertin TK, Chen Y, Hicks J, Tonachini L, Monticone M, Castagnola P, Rauch F, Glorieux FH, Vranka J, Bachinger HP, Pace JM, Schwarze U, Byers PH, Weis M, Fernandes RJ, Eyre DR, Yao Z, Boyce BF, Lee B (2006) CRTAP is required for prolyl 3- hydroxylation and mutations cause recessive osteogenesis imperfecta. Cell 127:291–304PubMedCrossRefGoogle Scholar
  16. 16.
    Cabral WA, Chang W, Barnes AM, Weis M, Scott MA, Leikin S, Makareeva E, Kuznetsova NV, Rosenbaum KN, Tifft CJ, Bulas DI, Kozma C, Smith PA, Eyre DR, Marini JC (2007) Prolyl 3-hydroxylase 1 deficiency causes a recessive metabolic bone disorder resembling lethal/severe osteogenesis imperfecta. Nat Genet 39:359–365PubMedCrossRefGoogle Scholar
  17. 17.
    Baldridge D, Schwarze U, Morello R, Lennington J, Bertin TK, Pace JM, Pepin MG, Weis M, Eyre DR, Walsh J, Lambert D, Green A, Robinson H, Michelson M, Houge G, Lindman C, Martin J, Ward J, Lemyre E, Mitchell JJ, Krakow D, Rimoin DL, Cohn DH, Byers PH, Lee B (2008) CRTAP and LEPRE1 mutations in recessive osteogenesis imperfecta. Hum Mutation 29:1435–1442CrossRefGoogle Scholar
  18. 18.
    Barnes AM, Carter EM, Cabral WA, Weis M, Chang W, Makareeva E, Leikin S, Rotimi CN, Eyre DR, Raggio CL, Marini JC (2010) Lack of cyclophilin B in osteogenesis imperfecta with normal collagen folding. N Engl J Med 362:521–528PubMedCrossRefGoogle Scholar
  19. 19.
    Alanay Y, Avaygan H, Camacho N, Utine GE, Boduroglu K, Aktas D, Alikasifoglu M, Tuncbilek E, Orhan D, Bakar FT, Zabel B, Superti-Furga A, Bruckner-Tuderman L, Curry CJ, Pyott S, Byers PH, Eyre DR, Baldridge D, Lee B, Merrill AE, Davis EC, Cohn DH, Akarsu N, Krakow D (2010) Mutations in the gene encoding the RER protein FKBP65 cause autosomal-recessive osteogenesis imperfecta. Am J Hum Genet 87:551–559CrossRefGoogle Scholar
  20. 20.
    Kelley BP, Malfait F, Bonafe L, Baldridge D, Homan E, Symoens S, Willaert A, Elcioglu N, Van Maldergem L, Verellen-Dumoulin C, Gillerot Y, Napierala D, Krakow D, Beighton P, Superti-Furga A, De Paepe A, Lee B (2011) Mutations in FKBP10 cause recessive osteogenesis imperfecta and Bruck syndrome. J Bone Miner Res 26:666–672PubMedCrossRefGoogle Scholar
  21. 21.
    Schwarze U, Cundy T, Pyott S, Christiansen H, Hegde MR, Bank R, Pals G, Ankala A, Connelly K, Seaver L, Yandow S, Raney H, Babovich-Vicsanovic D, Stoler J, Ben-Neriah Z, Segal R, Al-Aqeel A, Siderius L, Hannibal M, Hudgins L, McPherson E, Clemens M, Sussman MD, Steiner R, Mahan J, Smith R, Anyane-Yeboa K, Chong K, Uster T, Aftimos S, Sutton VR, Davis EC, Weis MA, Eyre D, Byers PH (2012) Mutation in FKBP10, which encodes a 65kD FK506 binding prolyl cis-trans isomerase, results in recessive forms of osteogenesis imperfecta and is the first Bruck syndrome (contractures and fractures) locus. Submitted for publicationGoogle Scholar
  22. 22.
    Van der Slot AJ, Zuurmond A-M, Bardoe AFJ, Wijmenga C, Pruijs HEH, Sillence DO, Brinckmann J, Abraham DJ, Black CM, Verzijl N, DeGroot J, Hanemaaijer R, TeKoppele JM, Huizinga TWJ, Bank RA (2003) Identification of PLOD2 as telopeptide lysyl hydroxylase, an important enzyme in fibrosis. J Biol Chem 278:40967–40972PubMedCrossRefGoogle Scholar
  23. 23.
    Christiansen HE, Schwarze U, Pyott SM, AlSwaid A, Al Balwi M, Alrasheed S, Pepin MG, Weis MA, Eyre DR, Byers PH (2010) Homozygosity for a missense mutation in SERPINH1, which encodes the collagen chaperone protein HSP47, results in severe recessive osteogenesis imperfecta. Am J Hum Genet 86:389–398PubMedCrossRefGoogle Scholar
  24. 24.
    Martinez-Glez V, Valencia M, Caparros-Martin JA, Aglan M, Temtamy S, Tenorio J, Pulido V, Lindert U, Rohrbach M, Eyre D, Giunta C, Lapunzina P, Ruiz-Perez VL (2012) Identification of a mutation causing deficient BMP1/mTLD proteolytic activity in autosomal recessive osteogenesis imperfecta. Hum Mutat 33:343–350PubMedCrossRefGoogle Scholar
  25. 25.
    Becker J, Semler O, Gilissen C, Li Y, Bolz HJ, Giunta C, Bergmann C, Rohrbach M, Koerber F, Zimmermann K, de Vries P, Wirth B, Schoenau E, Wollnik B, Veltman JA, Hoischen A, Netzer C (2011) Exome sequencing identifies truncating mutations in human SERPINF1 in autosomal-recessive osteogenesis imperfecta. Am J Hum Genet 88:362–371PubMedCrossRefGoogle Scholar
  26. 26.
    Homan EP, Rauch F, Grafe I, Lietman C, Doll JA, Dawson B, Bertin T, Napierala D, Morello R, Gibbs R, White L, Miki R, Cohn DH, Crawford S, Travers R, Glorieux FH, Lee B (2011) Mutations in SERPINF1 cause osteogenesis imperfecta type VI. J Bone Miner Res 26:2798–2803PubMedCrossRefGoogle Scholar
  27. 27.
    Gong Y, Slee RB, Fukai N, Rawadi G, Roman–Roman S, Reginato AM, Wang H, Cundy T, Glorieux FH, Lev D, Zacharin M, Oexle K, Marcelino J, Suwairi W, Heeger S, Sabatakos G, Apte S, Adkins WN, Allgrove J, Arslan-Kirchner M, Batch JA, Beighton P, Black GC, Boles RG, Boon LM, Borrone C, Brunner HG, Carle GF, Dallapiccola B, De Paepe A, Floege B, Halfhide ML, Hall B, Hennekam RC, Hirose T, Jans A, Juppner H, Kim CA, Keppler-Noreuil K, Kohlschuetter A, Lacombe D, Lambert M, Lemyre E, Letteboer T, Peltonen L, Ramesar RS, Romanengo M, Somer H, Steichen-Gersdorf E, Steinmann B, Sullivan B, Superti-Furga A, Swoboda W, van den Boogaard MJ, Van Hul W, Vikkula M, Votruba M, Zabel B, Garcia T, Baron R, Olsen BR, Warman ML (2001) LDL receptor–related protein 5 (LRP5) affects bone accrual and eye development. Cell 107:513–523PubMedCrossRefGoogle Scholar
  28. 28.
    Ai M, Heeger S, Bartels CF, Schelling DK, Osteoporosis-Pseudoglioma Collaborative Group (2005) Clinical and molecular findings in osteoporosis-pseudoglioma syndrome. Am J Hum Genet 77:741–753PubMedCrossRefGoogle Scholar
  29. 29.
    Lapunzina P, Aglan M, Temtamy S, Caparros-Martin JA, Valencia M, Leton R, Martinez-Glez V, Elhossini R, Amr K, Vilaboa N, Ruiz-Perez VL (2010) Identification of a frameshift mutation in osterix in a patient with recessive osteogenesis imperfecta. Am J Hum Genet 87:110–114PubMedCrossRefGoogle Scholar
  30. 30.
    Paterson CR, Ogston SA, Henry RM (1996) Life expectancy in osteogenesis imperfecta. BMJ 1996:312–351Google Scholar
  31. 31.
    Paterson CR, McAllion S, Stellman JL (1984) Osteogenesis imperfecta after the menopause. N Engl J Med 310:1694–1696PubMedCrossRefGoogle Scholar
  32. 32.
    Paterson CR, Monk EA, McAllion SJ (2001) How common is hearing impairment in osteogenesis imperfecta? J Laryngol Otol 115:280–282PubMedCrossRefGoogle Scholar
  33. 33.
    Bonita RE, Cohen IS, Berko BA (2010) Valvular heart disease in osteogenesis imperfecta: presentation of a case and review of the literature. Echocardiography 27:69–73PubMedCrossRefGoogle Scholar
  34. 34.
    Adami S, Gatti D, Colapietro F, Fracassi E, Braga V, Rossini M, Tato L (2003) Intravenous neridronate in adults with osteogenesis imperfecta. J Bone Miner Res 18:126–130PubMedCrossRefGoogle Scholar
  35. 35.
    Chevrel G, Schott A-M, Fontanges E, Charrin JE, Lina-Granade G, Duboeuf F, Garnero P, Arlot M, Raynal C, Meunier PJ (2006) Effects of oral alendronate on BMD in adult patients with osteogenesis imperfecta: a 3-year randomized placebo-controlled trial. J Bone Miner Res 21:300–306PubMedCrossRefGoogle Scholar
  36. 36.
    Bradbury LA, Barlow S, Geoghegan F, Hannon RA, Stuckey SL, Wass JAH, Russell RGG, Brown MA, Duncan EL (2012) Risedronate in adults with osteogenesis imperfecta type I: increased bone mineral density and decreased bone turnover, but high fracture rate persists. Osteoporos Int 23:285–294PubMedCrossRefGoogle Scholar
  37. 37.
    Devogelaer JP, Malghem J, Maldague B, Nagant de Deuxchaines C (1987) Radiological manifestations of bisphosphonate treatment with APD in a child suffering from osteogenesis imperfecta. Skel Radiol 16:360–363CrossRefGoogle Scholar
  38. 38.
    Rauch F, Glorieux FH (2004) Osteogenesis imperfecta. Lancet 363:1377–1385PubMedCrossRefGoogle Scholar
  39. 39.
    Gamsjaeger S, Buchinger B, Zwettler E, Recker R, Black D, Gasser JA, Eriksen EF, Klaushofer K, Paschalis EP (2011) Bone material properties in actively bone-forming trabeculae in postmenopausal women with osteoporosis after three years of treatment with once-yearly zoledronic acid. J Bone Miner Res 26:12–18PubMedCrossRefGoogle Scholar
  40. 40.
    Roschger P, Fratzl-Zelman N, Misof BM, Glorieux FH, Klaushofer K, Rauch F (2008) Evidence that abnormal high bone mineralization in growing children with osteogenesis imperfecta is not associated with specific collagen mutations. Calcif Tissue Int 82:263–270PubMedCrossRefGoogle Scholar
  41. 41.
    Land C, Rauch F, Glorieux FH (2006) Cyclical intravenous pamidronate treatment affects metaphyseal modeling in growing patients with osteogenesis imperfecta. J Bone Miner Res 21:374–379PubMedCrossRefGoogle Scholar
  42. 42.
    Munns CF, Rauch F, Zeitlin L, Fassier F, Glorieux FH (2004) Delayed osteotomy but not fracture healing in pediatric osteogenesis imperfecta patients receiving pamidronate. J Bone Miner Res 19:1779–1786PubMedCrossRefGoogle Scholar
  43. 43.
    Rauch F, Munns CF, Land C, Cheung M, Glorieux FH (2009) Risedronate in the treatment of mild pediatric osteogenesis imperfecta: a randomized placebo-controlled study. J Bone Miner Res 24:1282–1289PubMedCrossRefGoogle Scholar
  44. 44.
    Ward LM, Rauch F, Whyte MP, D’Astous J, Gates PE, Grogan D, Lester EL, McCall RE, Pressly TA, Sanders JO, Smith PA, Steiner RD, Sullivan E, Tyerman G, Smith-Wright DL, Verbruggen N, Heyden N, Lombardi A, Glorieux FH (2011) Alendronate for the treatment of pediatric osteogenesis imperfecta: a randomized placebo-controlled study. J Clin Endocrinol Metab 96:355–364PubMedCrossRefGoogle Scholar
  45. 45.
    Bishop N, Harrison R, Ahmed F, Shaw N, Eastell R, Campbell M, Knowles E, Hill C, Hall C, Chapman S, Sprigg A, Rigby A (2010) A randomized, controlled dose-ranging study of risedronate in children with moderate and severe osteogenesis imperfecta. J Bone Miner Res 25:32–40PubMedCrossRefGoogle Scholar
  46. 46.
    Antoniazzi F, Monti E, Venturi G, Franceschi R, Doro F, Gatti D, Zamboni G, Tato L (2010) Growth hormone in combination with bisphosphonate treatment in osteogenesis imperfecta. Eur J Endocrinol 163:479–487PubMedCrossRefGoogle Scholar
  47. 47.
    Rauch F, Cornibert S, Cheung M, Glorieux FH (2007) Long-bone changes after pamidronate discontinuation in children and adolescents with osteogenesis imperfecta. Bone 40:821–827PubMedCrossRefGoogle Scholar
  48. 48.
    Cheung MS, Glorieux FH, Rauch F (2009) Large osteoclasts in pediatric osteogenesis imperfecta patients receiving intravenous pamidronate. J Bone Miner Res 24:669–674PubMedCrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC 2012

Authors and Affiliations

  1. 1.Department of Medicine, Faculty of Medical & Health SciencesUniversity of AucklandAucklandNew Zealand

Personalised recommendations