Exome sequencing of 1190 non-syndromic clubfoot cases reveals HOXD12 as a novel disease gene

Abstract

Background: Clubfoot, presenting as a rigid inward and downward turning of the foot, is one of the most common congenital musculoskeletal anomalies. The aetiology of clubfoot is poorly understood and variants in known clubfoot disease genes account for only a small portion of the heritability.

Methods: Exome sequence data were generated from 1190 non-syndromic clubfoot cases and their family members from multiple ethnicities. Ultra-rare variant burden analysis was performed comparing 857 unrelated clubfoot cases with European ancestry with two independent ethnicity-matched control groups (1043 in-house and 56 885 gnomAD controls). Additional variants in prioritised genes were identified in a larger cohort, including probands with non-European ancestry. Segregation analysis was performed in multiplex families when available.

Results: Rare variants in 29 genes were enriched in clubfoot cases, including PITX1 (a known clubfoot disease gene), HOXD12, COL12A1, COL9A3 and LMX1B. In addition, rare variants in posterior HOX genes (HOX9-13) were enriched overall in clubfoot cases. In total, variants in these genes were present in 8.4% (100/1190) of clubfoot cases with both European and non-European ancestry. Among these, 3 are de novo and 22 show variable penetrance, including 4 HOXD12 variants that segregate with clubfoot.

Conclusion: We report HOXD12 as a novel clubfoot disease gene and demonstrate a phenotypic expansion of known disease genes (myopathy gene COL12A1, Ehlers-Danlos syndrome gene COL9A3 and nail-patella syndrome gene LMX1B) to include isolated clubfoot.

Keywords: exome sequencing; genetic diseases, inborn; genetic research; orthopedics; sequence analysis, DNA.

Figure 1

Flow chart of gene burden analysis. During data cleaning (see ‘Methods’ section), exomes of clubfoot cases and in-house controls went through principal component analysis (PCA) to retain data anchored to non-Finish European (NFE) population in 1000 genome phase III data. Testing Rare vAriants using Public Data (TRAPD) method was used in gene burden analysis between 857 unrelated isolated clubfoot probands and 2 independent controls (1043 in-house controls and 56 885 gnomAD NFE controls) in the autosomal dominant mode. Multiple filters were applied in order to select the deleterious/likely deleterious variants in each gene (see ‘Methods’ section). AC, allele count.

Figure 2

Localisation and segregation of HOXD12 variants. (A) Localisation of HOXD12 variants identified in the study along the protein. HOXD12 variants co-segregated with clubfoot in multiplex families are marked in red. HOXD12 variants in singletons are marked in blue. Yellow motif represents homeobox domain. (B) HOXD12 variants co-segregate with clubfoot in four multiplex families with complete penetrance. a.a., amino acid.

Figure 3

Pedigrees of segregated COL12A1 families. Variants in COL12A1 co-segregate with clubfoot in five multiplex families, potentially in two families and one proband carries a de novo COL12A1 variant.