Genetic Variants Associated with Sweat Gland Phenotypes in 6210 Han Chinese Individuals

Sweat glands are widely distributed across human skin, playing an essential role in thermoregulation and body temperature maintenance, while sweat gland dysfunction is associated with various skin disorders, such as hyperhidrosis and anhidrosis (Asahina et al., 2015Asahina M. Poudel A. Hirano S. Sweating on the palm and sole: physiological and clinical relevance.Clin Auton Res. 2015; 25: 153-159Crossref PubMed Scopus (34) Google Scholar). Sweat gland density (SGD) varies significantly between and within ethnic groups (Lee et al., 2010Lee J.B. Kim T.W. Shin Y.O. Min Y.K. Yang H.M. Effect of the Heat-exposure on Peripheral Sudomotor Activity Including the Density of Active Sweat Glands and Single Sweat Gland Output.Korean J Physiol Pharmacol. 2010; 14: 273-278Crossref PubMed Scopus (20) Google Scholar). Despite the estimated heritability of SGD up to 0.66 (Scobbie and Sofaer, 1987Scobbie R.B. Sofaer J.A. Sweat pore count, hair density and tooth size: heritability and genetic correlation.Hum Hered. 1987; 37: 349-353Crossref PubMed Scopus (6) Google Scholar), the underlying genetic factors and mechanisms responsible for this variation remain largely unexplored. Here, we report the genome-wide association study (GWAS) on SGD, involving 6,210 Han Chinese individuals from two independent cohorts: the Taizhou Longitudinal Study (TZL, n=3,883) and the National Survey of Physical Traits (NSPT, n=2,327, Supplementary Figure S1). Genetic principal component analysis did not detect population sub-stratification (Supplementary Figure S2). Ethical approval was obtained from the Ethics Committees of Fudan University (14117) and the Shanghai Institutes for Biological Sciences (ER-SIBS-261410), and all participants provided written informed consent. The classical 'starch iodine solution' (Muller and Kierland, 1959Muller S.A. Kierland R.R. The use of a modified starch-iodine test for investigating local sweating responses to intradermal injection of methacholine.J Invest Dermatol. 1959; 32: 126-128Abstract Full Text PDF PubMed Scopus (9) Google Scholar) method was applied to detect active sweat glands. An advanced image analysis was then performed to quantify the number of glands per mm2 (Figure 1a, Supplementary Methods). SGD demonstrated a moderately right-skewed distribution in both cohorts (mean 5.02+1.47 to 5.37+1.70 glands/mm2, Supplementary Table S1). Notably, females displayed a trend toward higher Z-transformed SGD (Z-SGD), representing an increase of 0.4 standard deviations (P<0.006; Supplementary Table S2). Z-SGD decreased by 0.3-0.4 standard deviations for every 10 years increase in age (P<2.4×10-12). Single nucleotide polymorphism (SNP)-based heritability was estimated at 0.35, confirming the genetic component detectable by our microarray data. Our GWAS did not shown sign of genomic inflation (lambda<1.01, Figure 1b), and revealed 1p36.3 being significantly associated with Z-SGD in TZL (lead SNP rs190550090, P=1.01×10-8), which was successfully replicated in NSPT with similar allelic effects (P=1.03×10-3; Figure 1c-d). In the meta-analysis of two cohorts, the association at 1p36.3 was further strengthened (β=1.19, P=4.98×10-11; Figure 1c). SGD of rs190550090 heterozygous CT carriers was 1.52 glands/mm2 higher than that of TT carriers (Figure 1e). The frequency of derived C allele was low (∼1% in our sample), consistent with that of East Asians in the 1000 Genomes Project (1000GP). The C allele was virtually undetected in non-East Asian cohorts, suggesting population-specificity (Figure 1f; Supplementary Table S3). Rs190550090 is located in the intron of KCNAB2 (Potassium Voltage-Gated Channel Subfamily A Regulatory Beta Subunit 2), a gene involved in acetylcholine release, which is responsible for triggering sweat gland secretion (Shibasaki and Crandall, 2001Shibasaki M. Crandall C.G. Effect of local acetylcholinesterase inhibition on sweat rate in humans.J Appl Physiol (1985). 2001; 90: 757-762Crossref PubMed Scopus (55) Google Scholar). Interestingly, 139 participants had absence of sweat glands (ASG; < -2 SD), with a significantly higher risk of dry skin based on questionnaires (OR = 1.03, 95% CI 1.01-1.05, P=0.003), and susceptibility to crack or wrinkle skin (OR = 1.02, 95% CI 1.00-1.03, P=0.004). Additionally, they displayed a higher number of facial pigment spots (β = 0.03, 95% CI 0.01-0.05, P=0.003). Further meta-analysis of the absence of ASG identified two significantly related loci. The first locus is at 1q41 downstream of KCNK2, another member of the potassium channel family playing a crucial role in heat-sensitive sweating (Kang et al., 2005Kang D. Choe C. Kim D. Thermosensitivity of the two-pore domain K+ channels TREK-2 and TRAAK.J Physiol. 2005; 564: 103-116Crossref PubMed Scopus (203) Google Scholar). The derived G allele of rs77384957 was associated with an increased risk of the ASG (OR=3.05, 95% CI: 2.12-4.39, P=1.98×10-9), with consistent effects in both cohorts (PNSPT=2.32×10-7; PTZL=5.28×10-3; Figure 1d-e). The frequency of G allele was low in our sample and in the 1000GP East Asians, while nearly absent outside of East Asia (Figure 1f; Supplementary Table S3). The second locus is at 18q21.3 downstream of TNFRSF11A (TNF receptor superfamily member 11a), inducing the activation of NF-kappa B, known to regulate sweat gland development and sweat production (Lu et al., 2016Lu C.P. Polak L. Keyes B.E. Fuchs E. Spatiotemporal antagonism in mesenchymal-epithelial signaling in sweat versus hair fate decision.Science. 2016; 354Crossref Scopus (96) Google Scholar). The T allele of rs3018355 was associated with an increased risk of the ASG (OR = 2.50; 95% CI: 1.87-3.35; P=8.25×10-10). The homozygote TT was not observed due to the low allelic frequency (1%), while it was common in populations outside of East Asia (fEUR = 0.14; fAFR = 0.25; Figure 1f; Supplementary Table S3). No genome-wide significant loci associated with excessive sweat gland (n=151; > 2SD) were found. For the three above nominated SNPs, a gender-stratified analysis revealed consistent SNP effects across both sexes. Fine-mapping analysis showed all of the three SNPs had high posterior probabilities (>0.95) with regulatory roles in gene expression (Supplementary Table S3). Specifically, rs190550090 region showed distinct active enhancer signatures and ChIP-Seq data (Landt et al., 2012Landt S.G. Marinov G.K. Kundaje A. Kheradpour P. Pauli F. Batzoglou S. et al.ChIP-seq guidelines and practices of the ENCODE and modENCODE consortia.Genome Res. 2012; 22: 1813-1831Crossref PubMed Scopus (1328) Google Scholar) further revealed its role as a binding site for several transcription factors, including ZNF76, POU5F1, MYC, KLF4, and FOXA2 (Figure 2a). Hi-C study demonstrated that rs190550090 interacts with the promoter of KCNAB2, and it located in the canonical E-box motif that is known to be recognized and bound by transcription factors, thereby activating gene transcription (Massari and Murre, 2000Massari M.E. Murre C. Helix-loop-helix proteins: regulators of transcription in eucaryotic organisms.Mol Cell Biol. 2000; 20: 429-440Crossref PubMed Scopus (1138) Google Scholar) (Figure 2b). We performed luciferase reporter assay in HEK293 and A375 cell lines and the results showed that the transcriptional activity of the sequence containing C allele exhibited significantly higher than that containing T allele in both cell lines, which verified the potential modulating activity of rs190550090 (Figure 2c-d). Detailed results for rs77384957 and rs3018355 were provided in Supplementary Text and Supplementary Figure S3. No significant positive selection signals were detected at these three loci, as expected due to the low allelic frequencies (Supplementary Figure S4; Supplementary Table S4). In summary, we identified three loci associated with sweat gland phenotypes. Two are specific to East Asians and situated near potassium channel genes, highlighting the channels' functional role in sweat gland development and the risk of chronic idiopathic anhidrosis. These loci may exhibit different mutations in non-East Asian populations, warranting further investigation through population-specific candidate gene analysis. The third locus, prevalent in non-East Asian populations and showing uniform linkage disequilibrium patterns, suggests a more detectable association in these groups. Despite the low allele frequencies of these loci, which may inherently be more susceptible to chance associations compared to common alleles, our findings enhance the understanding of genetic influences on sweat gland diversity and their potential impact on skin-related conditions. The GWAS summary statistics can be accessed from the National Omics Data Encyclopedia (http://www.biosino.org/node/) using the project identification document OEP004627. Please note that the usage of data must fully comply with the Regulations on Management of Human Genetic Resources in China. Due to privacy concerns and restrictions imposed by the Institutional Review Board, individual genotype and phenotype data cannot be shared. However, other relevant data supporting the key findings of this study can be found in the letter, Supplementary Materials, or obtained from the corresponding author upon reasonable request. The authors state no conflict of interest. We thank members of the Wang laboratory for helpful discussions. This work was supported by the Strategic Priority Research Program of the Chinese Academy of Sciences (grant number XDB38020400 to SW), CAS Project for Young Scientists in Basic Research (grant number YSBR-077 to SW), National Science and Technology Basic Research Project (grant number 2015FY111700 to SW), CAMS Innovation Fund for Medical Sciences (2019-I2M-5-066 to SW), National Natural Science Foundation of China (32070579 and 32370664 to ZY), the Natural Science Foundation of Henan (222300420067 to ZY), and National Natural Science Foundation of China (32200482 to JL). Correspondence regarding the luciferase assays and population comparative analyses should be addressed to ZY ([email protected]) and FL ([email protected]), respectively. Author Contributions Statement Conceptualization: SW; Data Curation: WC; Methodology: WC, SW, FL; Data collection: WC, KX, JL, MW, QP, YL, FW, LW, SD, WX, JG; Formal analysis: WC; Visualization: SW, WC, JL; Funding Acquisition: SW; Supervision: SW, FL, ZY;Writing-Original Draft Preparation: WC, SW, FL; Writing- Review and Editing: All authors. Download .xlsx (.08 MB) Help with xlsx files Download .xlsx (.08 MB) Help with xlsx files Supplementary Figure S2. The principal components analysis of 6,210 samples and reference samples from the HapMap.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Supplementary Figure S3. Functional annotation of the absence of active sweat glands associated SNPs rs77384957 and rs3018355. (a) The absence of active sweat glands associated SNP rs77384957 is bound by multiple histone modifications. (b) ChIP-Seq analysis reveals that the lead SNP rs77384957 serves as a binding site for transcription factors BRD4, MED1, RELA, PPARG, CTCF, AHR, and SMARCA4. (c) ChIP-Seq analysis demonstrates that the absence of active sweat glands phenotype associated SNP rs3018355 serves as a binding site for the transcription factor ZNF384.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Supplementary Figure S4. Composite of multiple signals (CMS) for the region extending 500 kb upstream and downstream of the Z-SGD-associated SNP rs190550090 (top) and the absence of active sweat glands associated SNP rs77384957 (bottom). The yellow highlight indicates the loci associated with phenotypes, while the green color represents the KCNAB2 genes.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Supplementary Figure S5. The heatmap of pairwise LD (r2) in the three identified marker SNPs (rs190550090 at 1p36.3, rs77384957 at 1q41, and rs3018355 at 18q21.3).View Large Image Figure ViewerDownload Hi-res image Download (PPT)Supplementary Figure S6. Image classification according to the qualityView Large Image Figure ViewerDownload Hi-res image Download (PPT)Supplementary Figure S7. A strong correlation and consistent agreement between the number of sweat glands recognized by our developed image recognition method and those manually labeled.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Supplementary Figure S8: Eigenvalues of genomic PCs.View Large Image Figure ViewerDownload Hi-res image Download (PPT)