After looking around further, I think it's most likely PAX1. There are PAX1 enhancer regions downstream of PAX1.
A PAX1 enhancer locus is associated with susceptibility to idiopathic scoliosis in females
A PAX1 enhancer locus is associated with susceptibility to idiopathic scoliosis in females
Idiopathic scoliosis (IS) is a common paediatric musculoskeletal disease that displays a strong female bias. By performing a genome-wide association study (GWAS) of 3,102 individuals, we identify significant associations with 20p11.22 SNPs for females (P=6.89 × 10−9) but not males (P=0.71). This association with IS is also found in independent female cohorts from the United States of America and Japan (overall P=2.15 × 10−10, OR=1.30 (rs6137473)). Unexpectedly, the 20p11.22 IS risk alleles were previously associated with protection from early-onset alopecia, another sexually dimorphic condition. The 174-kb associated locus is distal to PAX1, which encodes paired box 1, a transcription factor involved in spine development. We identify a sequence in the associated locus with enhancer activity in zebrafish somitic muscle and spinal cord, an activity that is abolished by IS-associated SNPs. We thus identify a sexually dimorphic IS susceptibility locus, and propose the first functionally defined candidate mutations in an enhancer that may regulate expression in specific spinal cells.
IS is a sexually dimorphic disease10. Girls and boys exhibit a striking difference in the prevalence of progressive IS, with girls having approximately tenfold greater risk of progressive curves that require operative treatment11. This dichotomy in female/male disease expression, and its correlation with the adolescent growth spurt have prompted investigations of hormonal influences in the development and progression of female IS6.
To discover new genetic risk factors for IS, we performed a two-stage GWAS in 3,102 individuals. Our results define a new susceptibility locus encoding associated SNPs that, surprisingly, are also associated with androgenic alopecia (AGA), or male pattern baldness. We find that the locus is specifically associated with female IS, suggesting that it contributes to the sexually dimorphic expression of the disease. By functional fine-mapping assays in zebrafish, we further define a sequence in the associated locus with enhancer activity that is abolished by IS-associated SNPs. Altogether, our results identify the first functionally characterized candidate mutations for IS susceptibility and expand our understanding of the role of non-coding regulatory elements in the disease. Our findings also suggest hypotheses to explain disease pathogenesis and provide the first insights into its puzzling sexual dimorphism.
Comparing our results to the National Human Genome Research Institute (NHGRI) GWAS catalogue22, we found that the chromosome 20 IS locus was previously associated with early-onset male pattern baldness (AGA). Similar to IS, AGA displays sexual dimorphism, that is, it is biologically unequal in males and females. However, unlike IS, disease progression in AGA (extent of hair loss) is generally more severe in males than in females36. We identified chromosome 20p11.22 SNPs that were previously associated with AGA and that were genotyped in our GWAS37, 38, 39. In this comparison, SNPs that were associated with IS and AGA displayed the opposite direction of effect for the two disorders (Supplementary Table 1). This observation suggested that sequences in the region conferring susceptibility to IS have a protective effect in AGA. To test whether the association we observed was sex-specific, we re-evaluated association with SNPs in the 20p11.22 locus after stratification by sex, that is, separating males and females. This analysis yielded evidence for association with IS in females but not males, with a combined Fisher’s P=6.88 × 10−9 in the former data set (Table 1 and Supplementary Tables 2 and 3).
Our investigation of the chromosome 20p11 locus provides the first genetic evidence to explain the puzzling sexual dimorphism that is a hallmark of IS. Besides susceptibility to progression, the pattern, onset and flexibility of deformity also differ between boys and girls10. Various hypotheses have been proposed to explain male/female differences in IS, including the existence of X-linked genetic risk factors and effects on circulating hormones. Neither mechanism has been clearly supported, although investigations have been limited6, 58. Our identification of a female-specific IS susceptibility locus suggests an underlying mechanism that is sensitive to the female milieu at the time of adolescence. Although we did not find evidence for oestrogen receptor-binding sites within the PEC7 enhancer locus itself, it is interesting to postulate that this locus increases risk of IS via downstream hormonal interactions. We note in this regard that the next-nearest gene, FOXA2, is implicated in sexually dimorphic gene expression via cooperation with androgen and oestrogen receptor59. It is possible that PEC7 regulates FOXA2. However, we did not detect Foxa2 expression in embryonic or postnatal mouse spine (data not shown) and consider it an unlikely candidate for IS susceptibility. PAX1 is also expressed in the adult scalp37. Whether variants in PEC7 affect this expression and drive association with early-onset male pattern baldness requires further study, but the overlapping genetic association suggests a possible correlation between the two sexually dimorphic conditions.
Comment