Reassessing the association: Evaluation of a polyalanine deletion variant of RUNX2 in non‐syndromic sagittal and metopic craniosynostosis

Abstract The RUNT‐related transcription factor RUNX2 plays a critical role in osteoblast differentiation, and alterations to gene dosage cause distinct craniofacial anomalies. Uniquely amongst the RUNT‐related family, vertebrate RUNX2 encodes a polyglutamine/polyalanine repeat (Gln 23 ‐Glu‐Ala 17 in humans), with the length of the polyalanine component completely conserved in great apes. Surprisingly, a frequent 6‐amino acid deletion polymorphism, p.(Ala84_Ala89)del, occurs in humans (termed 11A allele), and a previous association study (Cuellar et al. Bone 137 :115395;2020) reported that the 11A variant was significantly more frequent in non‐syndromic sagittal craniosynostosis (nsSag; allele frequency [AF] = 0.156; 95% confidence interval [CI] 0.126–0.189) compared to non‐syndromic metopic craniosynostosis (nsMet; AF = 0.068; 95% CI 0.045–0.098). However, the gnomAD v.2.1.1 control population used by Cuellar et al. did not display Hardy–Weinberg equilibrium, hampering interpretation. To re‐examine this association, we genotyped the RUNX2 11A polymorphism in 225 individuals with sporadic nsSag as parent–child trios and 164 singletons with sporadic nsMet, restricting our analysis to individuals of European ancestry. We compared observed allele frequencies to the non‐transmitted alleles in the parent–child trios, and to the genome sequencing data from gnomAD v.4, which display Hardy–Weinberg equilibrium. Observed AFs (and 95% CI) were 0.076 (0.053–0.104) in nsSag and 0.082 (0.055–0.118) in nsMet, compared with 0.062 (0.042–0.089) in non‐transmitted parental alleles and 0.065 (0.063–0.067) in gnomAD v.4.0.0 non‐Finnish European control genomes. In summary, we observed a non‐significant excess, compared to gnomAD data, of 11A alleles in both nsSag (relative risk 1.18, 95% CI 0.83–1.67) and nsMet (relative risk 1.29, 95% CI 0.87–1.92), but we did not replicate the much higher excess of RUNX2 11A alleles in nsSag previously reported ( p = 0.0001).