A missense variant in SLC39A8 is associated with severe idiopathic scoliosis

Abstract

Genetic factors predictive of severe adolescent idiopathic scoliosis (AIS) are largely unknown. To identify genetic variation associated with severe AIS, we performed an exome-wide association study of 457 severe AIS cases and 987 controls. We find a missense SNP in SLC39A8 (p.Ala391Thr, rs13107325) associated with severe AIS (P = 1.60 × 10^-7, OR = 2.01, CI = 1.54-2.62). This pleiotropic SNP was previously associated with BMI, blood pressure, cholesterol, and blood manganese level. We replicate the association in a second cohort (841 cases and 1095 controls) resulting in a combined P = 7.02 × 10^-14, OR = 1.94, CI = 1.63-2.34. Clinically, the minor allele of rs13107325 is associated with greater spinal curvature, decreased height, increased BMI and lower plasma manganese in our AIS cohort. Functional studies demonstrate reduced manganese influx mediated by the SLC39A8 p.Ala391Thr variant and vertebral abnormalities, impaired growth, and decreased motor activity in slc39a8 mutant zebrafish. Our results suggest the possibility that scoliosis may be amenable to dietary intervention.

Fig. 1

Manhattan plot showing AIS exome-wide association results. Linear regression p-values (−log10) for 52,480 single-nucleotide polymorphisms with minor allele frequency > 0.01. One SNP (rs13107325) was significantly associated with AIS after Bonferroni correction for multiple tests

Fig. 2

Clinical characteristics of AIS cases with different rs13107325 genotypes. Comparison of a scoliosis curve severity (Cobb angle), b height, c body-mass index, and d plasma manganese and iron levels in AIS patients who are carriers of the SLC39A8 391T allele compared to AIS cases who are non-carriers. For Cobb angle, Ala/Ala (N = 940) were compared to the combined Ala/Thr (N = 242) and Thr/Thr (N = 26). For height, Ala/Ala (N = 361) were compared to the combined Ala/Thr and Thr/Thr carriers (N = 123). For plasma manganese and iron, only Ala/Ala (N = 14) and Ala/Thr (N = 14) were measured due to lack of plasma samples from Thr/Thr carriers. Lines are median, 1st and 3rd quartiles. Error bars are 95% confidence intervals. All comparisons were performed using t-tests

Fig. 3

Cellular uptake of ions with SLC39A8 overexpression in HEK293 cells. Comparison of uptake of either radio-labeled a Mn²⁺, b Fe²⁺, or c Zn²⁺ with overexpression of either vector control, SLC39A8 wild-type or SLC39A8 A391T in HEK293 cells. All measurements were the average of three replicate experiments. Lines are median, 1st and 3rd quartiles. Error bars are 95% confidence intervals. All comparisons were performed using t-tests. ***P = 5 × 10⁻⁴

Fig. 4

Skeletal abnormalities and growth deficit in slc39a8−/− zebrafish. a Thoracic and caudal vertebral malformations causing mild spinal curvature in slc39a8−/− zebrafish (bottom) compared to WT (top) stained with alizarin red. Magnified images show vertebral fusions in KO fish. Scale bar 5 mm. b slc39a8−/− zebrafish have more spinal abnormalities than WT fish. c Reduced length of slc39a8−/− zebrafish (N = 20) compared to wild-type (N = 20) at 9 mpf ***p < 10⁻⁶, t-test. Lines are mean, first and third quartiles

Fig. 5

Movement deficits in zebrafish slc39a8−/− larvae. Zebrafish movements were quantified for 25 min without disturbing (total movement) followed by a tap (stimulus response) at 4 dpf. a, b Stimulus response visualized as a heat map following tap showing reduced movement in slc39a8−/− larvae compared to WT. c Total movement is less in slc39a8 hets and slc39a8−/− larvae compared to WT but normalizes with addition of 200 mM Mn for 24 h with little effect on WT. ***p < 0.001. **p < 0.01, t-test. Lines are median, 1st and 3rd quartiles. Error bars are 95% confidence intervals. All comparisons were performed using t-tests