Nosology and classification of genetic skeletal disorders: 2019 revision

1 INTRODUCTION

Fifty years ago, in 1969, an international team of experts in radiology, orthopedic surgery, pediatrics, and genetics convened in Paris to develop an International Nomenclature of Constitutional Diseases of Bones (A Nomenclature for Constitutional (Intrinsic) Diseases of Bones, 1971; International Nomenclature of Constitutional Diseases of Bones, 1970; International Nomenclature of Constitutional Bone Diseases, 1971; McKusick & Scott, 1971). The goal was to reach an agreement on the nomenclature of several genetic skeletal disorders that were reported since the early 1960s. At that time, there was growing evidence that genetic skeletal disorders were more heterogeneous than previously thought. The medical community started to appreciate the clinical and radiographic diversity among individuals with a “constitutional” bone disorder. It had become clear that not all individuals with short limbs had achondroplasia and that not all individuals with a short trunk had Morquio syndrome. The rapid progress in the delineation of new entities prompted the group to update the Nomenclature on several occasions, with revisions in 1977, 1983, 1992, and 1997 (International Nomenclature of Constitutional Diseases of Bones, 1979, 1983; International Nomenclature and Classification of the Osteochondrodysplasias, 1998; Beighton et al., 1992; Lachman, 1998; Rimoin, 1997; Spranger, 1992). After the establishment of the International Skeletal Dysplasia Society (ISDS) in 1999, revisions of the Nomenclature (Nosology) were delegated to an expert committee nominated within the ISDS and representing a good mix of clinical, radiological, and genetics expertise. The first ISDS revision was published in 2002 and thereafter at regular intervals (Bonafe et al., 2015; Hall, 2002; Superti-Furga & Unger, 2007; Warman et al., 2011). Here, we provide the 2019 revision and 10th edition of the Nosology and Classification of Genetic Skeletal Disorders.

2 METHODOLOGY

The preparation of this paper started in September 2017, when members of the ISDS Nosology Committee met in Bruges, just before the 13th biannual ISDS meeting (September 20–23, 2017). Participants were: D. Cohn, V. Cormier-Daire, C. Hall, G. Mortier (chair), G. Nishimura, L. Sangiorgi, R. Savarirayan, D. Sillence, A. Superti-Furga, S. Unger, and M. Warman. The goal was to revise and update the last (2015) edition of the Nosology. Prior to this meeting, two to three curators were appointed for each group of disorders listed in the 2015 revision paper. Each member was assigned to one or more groups as the following: DC to groups 2, 3, 6, 7, 8, and 10; VCD to groups 1, 9, 15, 20, and 30; CH to groups 23, 24, 32, 33, and 36; DK to groups 7 and 9; GM to groups 2, 3, 10, 16, 28, 30, and 35; SM to groups 37–42; GN to groups 13, 18, 19, 21, 26, 28, 34, and 36; SR to group 7; LS to groups 25 and 29; RS to groups 1, 17, 19, and 33; DS to groups 22–27; ASF to groups 4, 11, 12, 20, 31, and 35; SU to groups 5, 8, 10, 14, and 22 and MW to groups 3, 16, and 29. They were responsible for reviewing the available literature and suggesting possible changes ahead of the meeting. During the meeting, the proposals by different curators were discussed and a consensus was reached on the general approach and methodology for the revision. After the meeting, drafts were circulated and continuously updated until August 4, 2019, which finally resulted in the current version.

The criteria used for inclusion of individual disorders were essentially unchanged from previous revisions and included the following: (a) the disorder should have significant skeletal involvement, corresponding to the definition of either skeletal dysplasias/dysostoses, metabolic bone disorders, or skeletal malformation/reduction syndromes; (b) the disorder should have achieved peer-reviewed publication status with listing in PubMed, OMIM, or another biomedical archive/database; (c) the disorder should have a genetic basis proven by pedigree or by occurrence of the same phenotype in unrelated families or by molecular analysis (mutation or linkage analysis); (d) the disorder should have nosologic autonomy; that is, it should represent an independent entity and not just a variation of an already existing entity. Each disorder that met these criteria received a separate listing as one entry regardless of the inheritance pattern or causal gene(s), unless there was evidence that the disorder encompassed phenotypically different conditions. For example, omodysplasia (group 17) is listed as two separate entities, because there is an important phenotypic difference between the autosomal dominant and autosomal recessive types, which is exemplified by their different genetic cause. On the other hand, the perinatally lethal form of osteogenesis imperfecta (OI type 2) (group 25) only receives one entry despite the different involved genes and inheritance patterns. In contrast to the previous revisions, it was decided not to list the protein anymore, since this information can be easily deduced from the gene. Gene symbols used were those approved by the HUGO Gene Nomenclature Committee. In addition to the OMIM number, the ORPHANET code (where available) was also listed for each disorder. Although some disorders could be classified into different groups, the committee chose to list each disorder only once to avoid redundancy in the Nosology.

3 RESULTS

The updated Nosology comprises 461 disorders classified within 42 different groups (Table 1). The overall number of groups remains unchanged in comparison to the previous (2015) revision but two groups have changed names. Group 18 is now the “Bent bone dysplasia group” instead of “Campomelic dysplasia and related disorders,” hereby referring to the common radiographic sign of bent bones in these disorders. Group 19 changed from “Slender bone dysplasia group” to “Primordial dwarfism and slender bones group.” Genomic alterations affecting 437 different genes have been found in 425 of the listed disorders. Pathogenic variants in one gene can cause several phenotypes (e.g., groups 1, 2, 5, 6, and 8) and one phenotype can be caused by variants in several genes (e.g., groups 9 and 25). Pathogenic variants in FGFR3, COL2A1, COMP, NPR2, and ACAN can cause mild phenotypes such as isolated short stature or premature degenerative joint disease. However, these conditions were considered not to meet the inclusion criteria (mainly lack of significant skeletal involvement) and were therefore not incorporated into the Nosology. Phenotypes showing locus heterogeneity but clinically and/or radiographically almost indistinguishable from each other were included as one entry in the Nosology. Examples include the autosomal dominant form of multiple epiphyseal dysplasia (group 10), microcephalic osteodysplastic primordial dwarfism (group 19), rhizomelic chondrodysplasia punctata (group 21), and the severe, infantile form of osteopetrosis (group 23). For osteogenesis imperfecta, the more phenotypically based Sillence classification was kept as in the previous revisions of the Nosology (Van Dijk & Sillence, 2014). Several new entities have been added to the SEMD group (group 13). These disorders were previously ill-defined and classified as “aspecific” or “unknown” types of SEMD. The use of exome or whole genome sequencing has solved their molecular mystery and has now rendered them the status of separate and well-defined entities within the Nosology.

4 DISCUSSION

As with the previous revisions, the major challenge was keeping up with the rapid pace at which new entities are described and new genes are discovered. Next generation sequencing has revolutionized genetic medicine and this advance is also reflected in the field of genetic bone disorders. For many of these disorders, the molecular defect has now been identified. In this new edition of the Nosology, the causal gene or genomic alteration is listed for 92% (425/461) of the disorders. Previously, the percentages of disorders that had been “solved” genetically were 58% (215/372) for the 2006 revision, 69% (316/456) for the 2010 revision, and 88% (385/436) for the 2015 revision (Bonafe et al., 2015; Superti-Furga & Unger, 2007; Warman et al., 2011). The 2010 revision was published when the application of massively parallel sequencing to Mendelian genetic diseases was just beginning (Ng et al., 2010). Not only well-known entities, previously carefully delineated based on their clinical and/or radiographic features, are being “solved” at the genetic level, but also new disorders and their causal genes are being discovered and reported at a rather rapid pace. The latter is often the result of a “genotype first—phenotype later” approach in individuals with an “unknown” skeletal dysplasia and facilitated through web-based tools such as GeneMatcher, which enables the comparison of phenotypes among patients with pathogenic variants in a newly identified gene (Sobreira et al., 2015).

The 437 disease-causing genes listed in the 2019 Nosology are functionally diverse, involved in a broad range of cell biologic processes, and cause disease by a variety of mutational mechanisms. They not only code for tissue-specific proteins that are essential for the formation and maintenance of bone and cartilage but also encode proteins that have a more ubiquitous role such as regulating gene transcription, cell division, or intracellular transport. While many disease-causing genes have clear roles in skeletal development (e.g., those involved in NOTCH, WNT, TGFβ, or BMP signaling), the skeletal roles for other genes are not yet clear. For example, pathogenic variants in mitochondrial proteins can cause a skeletal dysplasia, which is surprising, since skeletal manifestations are uncommon for most mitochondrial disorders (Dikoglu et al., 2015; Girisha et al., 2019; Mehawej et al., 2014; Peter et al., 2019; Royer-Bertrand et al., 2015). In addition, genes that do not encode proteins are also responsible for skeletal disorders. A well-known and longstanding example is cartilage hair hypoplasia that is caused by pathogenic variants in RMRP, which encodes an RNA component of the mitochondrial RNA processing endoribonuclease. Interestingly, the current Nosology now includes the first example of pathogenic variants in a miRNA causing a skeletal dysplasia (SED, MIR140 type; group 15) (Grigelioniene et al., 2019). Alterations in regulatory sequences, residing outside the genes, are another well-established cause of skeletal disorders. As a general rule, these disorders are characterized by defects in early skeletal development and patterning and tend to affect a particular set of bones in the skeleton (dysostoses). They usually do not present as true chondrodysplasias having widespread epiphyseal or metaphyseal changes. For example, pathogenic variants in an upstream cis-regulatory enhancer of the SHH gene (known as the ZPA regulatory sequence) can cause a spectrum of limb malformations ranging from preaxial polydactyly and triphalangeal thumb to the more severe Werner mesomelic dysplasia (Girisha et al., 2014; Wieczorek et al., 2010). Structural variations and translocations within the vicinity of the HOXD cluster locus on chromosome 2 have been reported in several limb malformations, including mesomelic dysplasia Kantaputra type (Kantaputra et al., 2010). Similarly, a particular deletion encompassing four protein coding genes on 6p22.3 has been identified in three unrelated patients with mesomelic dysplasia Savarirayan type (Flöttmann et al., 2015). In this paper, the authors provide evidence that haploinsufficiency for the deleted genes is not the mutational mechanism but rather the disruption of topologically associated domains in this region. By the deletion, two regulatory boundaries are removed and new limb enhancers are brought into close proximity of the ID4 gene, a phenomenon known as enhancer adoption.

The classification and organization of disorders have not been changed significantly compared with the previous editions. The Nosology still remains “hybrid” in nature in the sense that the classification is not always based on the same criteria. Some diseases are grouped based on the causal gene, others are listed together, because they share common radiographic features, and still others are brought together because of a similar clinical course (lethality) or involvement of similar parts of the skeleton. A Web-based Nosology with a clinical, radiographic, and molecular annotation for each disorder and with links to different databases would not only solve this classification issue but would also enable more specific searches per gene, pathway, or clinical/radiographic feature.

Regular revisions of the Nosology on Skeletal Disorders are important. The Nosology can serve as a diagnostic aid for clinicians who care for individuals with a skeletal disorder. In addition, it can facilitate the recognition of new entities and be a guide in the interpretation of new genetic variants. Finally, the Nosology can foster and enhance research by providing a catalogue of genes with important roles in skeletal biology.