Identification of two novel COL10A1 heterozygous mutations in two Chinese pedigrees with Schmid-type metaphyseal chondrodysplasia

Schmid-type metaphyseal chondrodysplasia (MCDS; MIM 156500), the most common type of metaphyseal chondrodysplasia, is an autosomal dominant congenital disorder that is characterized by short stature, waddling gait, coxa vara and bowing of the long bones (primarily involving the femur) [1, 2]. The distinctive clinical features of MCDS also include leg pain, enlarged capital femoral epiphyses or partial metaphyseal abnormalities of the upper limbs [3]. The typical radiographic findings of MCDS are widening and irregularity of the growth plates, especially in the distal femur and proximal tibia [4‐6], but only a small percentage of patients have involvement of the upper limbs [7]. In addition, it should be noted that all patients affected with MCDS exhibit normal extraskeletal manifestations [8].

COL10A1 (MIM 120110), located on chromosome 6q21-q22.3 [9], is closely associated with MCDS when pathogenic variants occur in this gene [10‐12]. The product of the COL10A1 gene is the α1 chain of type X collagen, which composes the type X collagen by forming a homotrimer [5]. Type X collagen is a member of the collagen superfamily of structural macromolecules, which has a unique expression pattern localized to the hypertrophic chondrocytes of growth plate cartilage [13]. The function of type X collagen, as a short-chain minor collagen of cartilage, is to play an important role in fetal chondrogenesis and endochondral ossification [5, 14]. Therefore, specific mutations in COL10A1 are likely to result in occurrence of MCDS.

To date, over fifty mutations in COL10A1 have been reported in the Human Gene Mutation Database (http://​www.​hgmd.​cf.​ac.​uk/​ac/​) or relevant literature, and 51 variants of those to our knowledge are directly associated with MCDS (Additional file 1) [15, 16]. Based on different mutation types, the mutations in COL10A1 can be divided into two categories: single amino acid residue lack or substitution resulting from missense mutations or small deletions (Class I) and truncated peptide attributed to nonsense mutations or frameshift mutations (Class II) [17, 18]. Of note, the frequencies of clinical findings and underlying pathogenic mechanisms of the two types are completely different.

The present study describes the phenomenon of incomplete dominance and summarizes several potential genetic patterns in two unrelated Chinese pedigrees with MCDS. Moreover, mutation screening used to support diagnosis was performed in the two unrelated Chinese MCDS families.

There is no doubt that MCDS is an autosomal dominant inherited disorder, resulting from heterozygous mutation of the COL10A1 gene [4, 10, 17]. In the present study, probands with MCDS and affected family members were identified in two large independent Chinese pedigrees by means of typical clinical findings and genetic analysis. Interestingly, although these individuals suffered from identical disease, there were still differences in the severity of clinical manifestations among all the patients in each family. We attributed the phenomenon of differential performance to incomplete dominance based on the presence of an identical pathogenic mutation in each family, possibly caused by the existence of a modifier gene and environmental differences [19]. Recently, Forouhan et al. proposed that ATF6α and ATF6β play important roles in modulating disease severity in MCDS mice by positively or negatively regulating the endoplasmic reticulum stress response [20], which we considered to be the associated mechanism of the incomplete dominance phenomenon. Moreover, carbamazepine, a drug which stimulates intracellular proteolysis and alleviates endoplasmic reticulum stress, effectively reduced the disease severity in the model of MCDS [21]. However, further molecular experiments are needed.

Depending on the differential expressivity of all affected members in two Chinese families (Table 1), we have summarized the following possible rules on the pedigrees affected with MCDS [6, 16]. First, based on the onset age, these patients were characterized by delayed dominance, which only occurred months or even a year after birth, at 6 to 18 months old in this study, rather than at birth. Furthermore, the severity of disease was closely associated with onset age presenting a negative correlation; that is, the earlier the onset age, the more severe the condition. For example, decreased quality of life, including unstable standing or waddling gait was observed in patients whose onset ages were only around 6 months old. Conversely, only short stature without other deformities was exhibited in those whose onset ages were 10 months old or later. In addition, we found that there were possible, but not significant, potential differences in gender susceptibility in MCDS. Moreover, despite the trend that male patients were more severely affected than females, as observed in family 2, we still could not draw a firm conclusion due to the rather small sample size.

The molecular structure of type X collagen is a homotrimer of three X (α1) chains, each consisting of a 463 amino acid Gly-X-Y collagenous domain (COL1) flanked by a 38-residue N-terminal noncollagenous domain (NC2) and a 161-residue C-terminal noncollagenous domain (NC1) (Fig. 4c) [9]. In addition, there is an 18-residue signal peptide ahead of the NC2 domain. To date, a total of 51 mutations of the COL10A1 gene resulting in MCDS have been reported (Additional file 1). All of the identified mutation sites of COL10A1 associated with MCDS, including mutations in the present study, are located in the NC1 domain [4, 8, 12, 22‐28], except for two missense mutations in the signal peptide and one in the triple helical domain.

As for genotype–phenotype correlations, the varying clinical findings were closely associated with the different molecular mechanisms of pathogenesis between the cases caused by protein-altering variants and those resulting from protein-truncating variants [29]. First, all of the protein-altering variants in COL10A1 were missense variants, most of which were located in the NC1 domain. The function of the NC1 domain is mainly to assist the folding of the peptide chain so that it can combine to form a homotrimer [5]. Once the NC1 domain becomes impaired, the collagen X (α1) chains are prevented from aggregating and instead form non-functional polymers, which tend to promote harmful accumulation of invalid products and even initiate the endoplasmic reticulum stress response [5, 30]. Meanwhile, the quantity of correctly-folded collagen X is reduced, and therefore functional haploinsufficiency is the most likely cause of the MCDS [14, 31, 32]. These cases caused by missense variants exhibited relatively late-onset ages (generally after 8 months old) and mild or moderate manifestations [4]. Among these variants, the three variants that are not located in the NC1 domain are associated with much later-onset ages and milder manifestations of MCDS, but most of those located in the NC1 domain lead to the more severe forms [25, 27]. On the other hand, COL10A1 nonsense mutations in cartilage tissue lead to removal of the mutant mRNA by nonsense-mediated mRNA decay (NMD), which is the pathogenic molecular mechanism of nonsense mutations in MCDS [18, 29]. In addition to the functional haploinsufficiency, more complicated molecular mechanisms, such as dominant negative effects were involved in the pathogenesis [8, 29, 33]. The cases caused by COL10A1 protein-truncating variants showed earlier onset ages (6 months old or earlier) and more severe clinical or radiographic manifestations, such as restricted motion of the joints, compared with those carrying protein-altering variants [6, 8, 28].

In the present study, two novel variants resulted in moderate but differential clinical features in affected individuals. Overall, the genotype–phenotype correlation of cases in this study was consistent with previous reports [4], but the incomplete dominance of phenotype is the first report in MCDS. In terms of pathogenesis, one of the substitutions (p.Phe589Ile) affects a hydrophobic area and the other (p.Lys616Glu) is predicted to affect the surface of the assembled trimer (Fig. 4a, b). The substitution p.Phe589Ile weakens the hydrophobicity of the wild type residue, which is likely to seriously impact the assembly and stability of the hydrophobic channel and thus hinder collagen X trimerization. The other substitution (p.Lys616Glu) changes the residue site from alkaline to acidic, potentially destroying the combination of trimeric collagen X into supramolecular structures within the cartilage pericellular matrix. Together, these biochemical and pathophysiological processes may explain the underlying mechanisms of MCDS in the two present pedigrees. In future, in vitro or in vivo functional studies will be performed to gather more evidence.

In summary, we identified two novel COL10A1 heterozygous missense mutations [c.1765 T > A (p.Phe589Ile)] and [c.1846A > G (p.Lys616Glu)] in two relatively large unrelated Chinese pedigrees with MCDS. The genetic analysis facilitated diagnosis of the disease and further expanded the spectrum of COL10A1 mutations associated with MCDS patients. In addition, our research revealed the phenomenon of incomplete dominance and summarized several potential genetic patterns in the two Chinese pedigrees with MCDS.