Validation of a genome-wide association study implied that SHTIN1 may involve in the pathogenesis of NSCL/P in Chinese population

Orofacial clefts are among the most common birth defects in humans worldwide. A large-scale, genome-wide association study (GWAS) in the Chinese population recently identified several genetic risk variants for nonsyndromic cleft lip with or without cleft palate (NSCL/P). We selected 16 significant SNPs from the GWAS I stage (P < 1.00E-5) that had not been replicated to validate their association with NSCL/P in 1931 NSCL/P cases and 2258 controls. Ultimately, we identified a NSCL/P susceptibility loci (rs17095681 at 10q25.3, intron of SHTN1 and 27.2 kb downstream of VAX1, Pmeta = 3.80E-9, OR = 0.64) in Chinese Han and Hui populations. This locus was not high LD with the reported loci in 10q25.3. It was a newly identified independent locus in 10q25.3 associated with NSCL/P. These results imply that SHTIN1 may involve in the pathogenesis of NSCL/P advance our understanding of the genetic susceptibility to NSCL/P.

Scientific RepoRts | 6:38872 | DOI: 10.1038/srep38872 switch and the formation of oral periderm 3,4 . A functional variant of IRF6 (rs642961), located within the promoter sequence and disrupting the binding site for the transcription factor AP2a, also significantly increases the risk for NSCL/P 5 .
With recent advances in high-density single nucleotide polymorphism (SNP) genotyping arrays and statistical methodology, genome-wide association studies (GWAS) have heralded a new era of gene discovery for complex diseases. To date, five GWASs and one genome-wide meta-analysis on NSCL/P have been performed, which identified 13 loci or genes (8q24, IRF6, MAFB, ABCA4, NOG, VAX1, PAX7, EPHA3, THADA, SPRY2, TPM1, 8q21. 3 and CREBBP) [6][7][8][9][10][11] that exceeded genome-wide significant levels. It has been shown that the 8q24 region harbors remote cis-acting enhancers that control Myc expression in the developing face. Deletion of this regulating interval in mice results in mild alteration of facial morphology, including CL/P 12  The GWAS performed by Sun et al. in a large scale of Chinese population identified several risk genetic variants for NSCL/P 11 . In this project, Two GWASs were performed, validating 30 loci that were significant in both GWAS studies. However, some loci that exceeded the significance threshold (P < 1E-5) in the first GWA study but were not significant or imputed successfully in the second GWA study have yet to be validated. In this paper, we selected 16 such SNPs and validated them in Chinese population to further identify susceptibility loci/genes for NSCL/P. We identified one locus showing significant association with NSCL/P risk.

Results
Validation results. Sixteen SNPs were selected for validation based on a stepwise series of criteria (see Methods), none of which were in high LD with each other. We then performed the validation of these 16 SNPs in 1668 Chinese Han cases and 1924 Chinese Han controls from multiple hospitals in China. Three SNPs showed evidence of association with NSCL/P (P meta < 0.05, Supplementary Table S1), but the association direction of two of these SNPs was not in concordance with the GWAS stage. One SNPs, rs17095681 at 10q25.3 (P meta = 8.50E-05, OR = 0.70) displayed consistent association with NSCL/P, and the significant direction was in concordance with the GWAS stage (Table 1; Supplementary Table S1). We then validated rs17095681 in Chinese Hui population, which is a minority of China, including 263 case and 334 control from General Hospital of Ningxia Medical University. The validation result was close to significance (P = 6.42E-2) and the association direction was in concordance with the GWAS stage (OR = 0.64). The Meta analysis P-values of Han and Hui population validations was 1.50E-5 (Table 1). In the combined analysis of GWAS and validation stages, rs17095681 showed strong evidence of association (P meta = 3.80E-9, OR = 0.64), which reached the genome-wide significance level among the Chinese Han and Hui samples (P meta < 5.00E-8, Table 1). Another reported significant locus in 10q25.3, rs7078160, was also validated in these validation samples 11 . To test the independence of these two SNPs, We fix one SNP as conditional factor and analysis the association between the other SNP and NSCL/P. The conditional analysis results indicated that the effect of rs17095681 was not correlated with rs7078160 (P value of rs17095681 was P meta = 7.82E-4 conditioned on rs7078160, Supplementary Table S2). The LD analysis indicated that rs17095681 was not linkage with rs7078160 in four validation cohort (r 2 < 0.1, Supplementary Table S3). These results indicated that rs17095681 may be an independent locus associated with NSCL/P at 10q25.3 in Chinese Han and Hui populations.

Analysis of different genetic models.
For the significant SNP rs17095681, we used other genetic models, the additive model, allelic model and genotypic model for further analysis. We observed that this locus significant under the dominant model achieved similar results under the additive model (P meta = 8.96E-5), allelic model (P meta = 2.56E-5) and genotypic model (P meta = 9.85E-6) in validation stage (Table 2). In the combined analysis of the two stages, this locus achieved similar results with dominant model under the genotypic (het) model (P meta = 3.40E-9). It also achieved near genome-wide significance in additive (P meta = 6.50E-8) and allelic model (P meta = 5.65E-8). In Summary, the SNP rs17095681 was associated with NSCL/P significantly in additive model, allelic model and genotypic model.

Discussion
In this validation study of NSCL/P, we identified one SNP at 10q25.3, rs17095681, which were significantly associated with NSCL/P risk in Chinese Han and Hui populations. It reached the genome-wide significance threshold (P meta = 3.40E-9) in the combined analysis. The SNP rs17095681 is located in the intron of SHTN1 (also known as KIAA15598) and 27.2 kb downstream of VAX1 (ventral anterior homeobox 1, Fig. 1). It has been reported that four SNPs with P GWAS < E-4 are located in a 30-kb region that is 50 kb downstream of VAX1 8 . The SNP rs17095681 is not in high LD (r 2 > 0.8) with these SNPs. Rs7078160 has been successfully validated in a Chinese population 11 . The SNPs rs17095681 and rs7078160 are 33 kb apart from each other, but the effect of rs17095681 is not correlated with rs7078160 (Supplementary Table S2), indicating that rs17095681 is in an independent block associated with NSCL/P. SHTN1 code a linker molecule shootin 1 that couples F-actin retrograde flow and the cell adhesion molecule (CAM) L1-CAM 13 at neuronal growth cones to promote neuronal polarization and axon outgrowth. The attractive axon guidance molecule netrin-1 14,15 induces Pak1-mediated shootin1 phosphorylation in axonal growth cones 16 which in turn enhances the coupling between F-actins and shootin1, thereby promoting the traction    forces for axon outgrowth. It has been reported that netrin-1 gene NTN1 is associated with NSCL/P 9,11,17 . NTN1 encodes the protein NETRIN 1, which plays a role in the developing the nervous system by promoting both axonal outgrowth and axonal guidance in pathfinding [18][19][20][21] . In addition, NTN1 was up regulated in dental pulp stem cell cultures from NSCL/P patients 11 . This information suggested that SHTIN1 and NTN1 play important roles in the development of NSCL/P. The other gene near SNP rs17095681, VAX1 was also been reported associated with NSCL/P. Mice with homozygous Vax1 mutations display craniofacial malformations including cleft palate 22 . Two individuals with a 10q terminal deletion syndrome with breakpoints in 10q25 have been reported, one with a submucous cleft palate 23 and the other with a cleft lip 24 .
In summary, in this validation of a NSCL/P GWAS in Chinese populations, we identified a susceptibility locus at 10q25.3 that reached genome-wide significance. The rs1709568 SNP at 10q25.3 is located in an independent block and therefore is not related to any previously reported associated loci. Genes near this locus participate in the processes of neuronal axon outgrowth and cell migration. It is known that cell migration are crucial for oralfacial development. Further studies with larger sample sizes are warranted to replicate our findings. Fine mapping around this locus and related functional studies should also be performed to elucidate the molecular mechanisms underlying the observed associations.

Methods
Study populations. We performed the validation study using samples from four regions in China, including 1931 NSCL/P cases and 2258 controls. A summary of all cases and controls in the study is provided in Table 3. Chinese Hui cases and 334 Chinese Hui control). All cases were recruited in local hospitals and independently confirmed as NSCL/P by two gynecologic pathologists during routine diagnosis. Syndromic cleft lip or palate patients and cleft palate-only patients were excluded. Controls were recruited in local hospitals for individuals receiving routine physical examinations or healthy newborns whose parents volunteered to donate their umbilical cord blood. All controls were clinically assessed to be without cleft lip or palate or family history of cleft lip or cleft palate (including first, second, and third degree relatives). The cases and controls were frequency-matched for age and gender. At recruitment, informed consent was obtained from each subject. This study was approved by the ethics committees of Guangdong Maternal and Child Health Care Hospital, Western China Hospital of Stomatology Sichuan University, the Institute of Stomatology, Nanjing Medical University and General Hospital of Ningxia Medical University and the methods were carried out in accordance with the approved guidelines. SNP selection and genotyping. SNPs for the replication stage were selected using the following criteria: (i) SNPs with P < 1.00E-5 in the first GWA study but were not significant or imputed successfully in the second GWA study; (ii) only the SNP with the lowest P-value was selected when multiple SNPs were observed but in strong linkage disequilibrium (LD) (r 2 > = 0.8); (iii) primers could be successfully designed using Sequenom primer design software; and (iv) SNPs had not previously been validated. A total of 16 SNPs that matched these criteria were included in the replication stage. Genotyping of replicates was conducted by the Sequenom MassARRAY system at Beijing CapitalBio Technology Company, Beijing, China.
Quality control at the replication stage. We excluded SNPs with a call rate < 90% or a deviation from Hardy-Weinberg equilibrium (P < 0.05) in the controls. All 16 SNPs exceed quality control and were used for further analysis.
Association analysis in the replication and combined stage. For the replication studies, associations between SNP genotypes and disease status were assessed in a dominant model in PLINK v1.07 (http://pngu.mgh. harvard.edu/Bpurcell/plink/) using logistic regression modeling with gender as a covariate. Joint analyses of all combined samples at the validation stage and GWAS stage were conducted by using either the random effects model (I 2 > 25%) or by using the fixed-effect model (I 2 < 25%). Another genetic model (the additive, allelic and genotypic model) was also calculated for the associated SNPs. The chromosome regions of significant loci were plotted using an online tool, LocusZoom 1.1 (http://csg.sph.umich.edu/ locuszoom/).  Male  308  236  330  287  310  280  273  333  162  252  170  164   Female  196  219  167  210  170  202  173  189  83  171  93  170   Table 3. Sample characteristics of cases with NSCL/P and controls.