An Association Study of Candidate Gene Variants in Chinese Nonsyndromic Cleft Lip with or without Palate Subjects

Background: Nonsyndromic cleft lip with or without palate is a common birth defect of complex etiology involving the interaction of genetic factors and environmental exposures. Previous reports identied several candidate genes and risk variants of the disease, and their functions were veried by model animal studies. Methods: In order to depict the composition of the orofacial cleft susceptibility loci in Chinese population, we genotyped ten common SNPs of six genes (VAX1, MAFB, PAX7, ABCA4, NTN1 and NOG) in 249 nonsyndromic cleft lip with or without palate individuals, 62 nonsyndromic cleft palate only individuals and 480 controls. Results: Three loci: VAX1 rs7078160, MAFB rs11696257 and NTN1 rs4791774 showed signicant relevance with NSCL/P. Carrying both VAX1 rs7078160 and NTN1 rs4791774 further increased the risk, compared with carrying only one of them. Conclusions: This result supported that SNPs of genes VAX1, MAFB and NTN1 are associated with NSCL/P in Chinese subjects.


Background
Cleft lip with or without palate (CL/P) is among the most common congenital craniofacial defects, with an average incidence ranging from 1/500 to 1/1000 live births, based on different ethnic populations [1]. Approximately 70% of cases of CL/P occur with no other apparent structural abnormalities are de ned as nonsyndromic cleft lip with or without cleft palate (NSCL/P) [2]. The prevalence of NSCL/P in China was 1.22/1000 [3].It is considered to be a multifactorial disease with both genetic and environmental factors contributing to the etiology. Nonsyndromic cleft palate only (NSCPO) is to some extent different though not completely distinct from NSCL/P in genetic mechanism [4].
Nonsyndromic CL/P does not entirely follow Mendel's genetic characteristics, a large part of parental phenotype of cleft lip and palate patients is normal [5]. Genetic variations increase the risk of this phenotype, although they are common in population. The genome wide association studies found more than 40 candidate genes and their marker SNPs [6][7][8][9][10][11]. In addition to being a marker of susceptibility, several of these genes also have been linked to cleft lip and palate by animal model. Gene VAX1 located in chromosome 10q25 is expressed widely during development of craniofacial structures [6]. The Vax1 knockout mouse developed cleft palate [12]. A missense mutation H131Q in a strongly conserved sequence region of MAFB was found association with CL/P, and Mafb expression was upregulated during palatal fusion [11]. Mice lacking Pax7 showed malformations of the nasomaxillary complex [13]. Gene NOG is expressed in palate shelves and functions as a signaling molecule during embryonic development [14]. Inactivation of NOG showed cleft palate in a rat model [15]. Although deleterious changes in these genes can cause cleft in model animals, the contribution of polymorphisms remains to be investigated. And these variants may not play roles independently [16].
Dissimilar results of the gene contributions in the populations with different genetic backgrounds and living conditions were reported. For example, ABCA4 achieved more signi cant association among the Asian families compared to the European's while the 8q24 region showed an opposite result [11]. Therefore, we selected 10 SNPs of six candidate genes: VAX1, MAFB, PAX7, ABCA4, NTN1 and NOG to conduct this case-control study.
Allele analyses showed that VAX1 rs7078160 and MAFB rs11696257 in the NSCL/P cases were signi cantly different from those among the controls (Bonferroni method adjusted p-values were 0.020 and 0.00031, respectively). The risk allele frequencies in the case and the control groups were 48.6% and 40.3% for rs7078160 A and 60.1% and 48.5% for rs11696257 C, respectively. Genotype analyses under the additive model also identi ed the association of VAX1 rs7078160 and MAFB rs11696257 (p-values were 0.025 and 0.00044, respectively). The highest risk of this study was found in homozygote comparing CC vs. TT for MAFB rs11696257 (OR = 2.47, 95% CI: 1.58 to 3.87). Variants NTN1 rs4791774 passed the signi cant test under the dominate model (p-values were 0.030). For this SNP, the genotype GG + AG increased disease risk 1.63 times (95% CI: 1.18 to 2.25) comparing with the homozygote AA. There were no association found between the CPO group and all the SNPs. Then the CL/P (CL + CLP) and the CPO subjects were combined into one orofacial cleft case group to calculate. MAFB rs11696257 remained positive, but the signi cant level reduced, compared with the result when only NSCL/P subjects included (Table S1).
In order to test whether a second risk gene would further increase the disease risk, the number of subjects who carried any one or two of the three positive loci (VAX1 rs7078160 A, MAFB rs11696257 C and NTN1 rs4791774 G) were compared between NSCL/P patients and controls (Table 2). Individual, who carrying both risk genes VAX1 rs7078160 A and NTN1 rs4791774 G, has more than two times higher risk compared with these carrying only one of these two risk genes. These two genes are independent risk factors to each other.

Discussion
NSCL/P is considered as a multifactorial disease resulting from the interaction between genetic and environmental factors. We conduct an association study for 10 SNPs and support that MAFB rs11696257, VAX1 rs7078160 and NTN1 rs4791774 increased NSCL/P risk in Chinese subjects. The ABCA4 polymorphisms also showed weak association but they had not passed the multiple test adjustment. Similar results were found in MAFB rs13041247 and PAX7 rs4920520. But for VAX1 rs4752028 and PAX7 rs766325, the frequencies were nearly equal in the cases and the controls (32.21% vs. 32.29% and 16.67% vs. 16.99%). The minor allele frequency of NOG rs17760296 is only about 1.5%, so the contribution of this variant must be limited in Chinese population. This study indicates that CL/P and CPO should be considered as two phenotypes caused by different genetic reasons, at least for these positive genes.
The highest risk was found in MAFB, which is a transcription factor located in chromosome 20q12. This gene encodes a basic leucine zipper transcription factor and also associated with another disease with maxillary hypoplasia phenotype named Multicentric Carpotarsal Osteolysis Syndrome. Expression analysis in mouse embryos revealed its function in lip and palate morphogenesis especially palatal fusion [17], which reminded the role of gene IRF6. Beaty's GWAS identi ed the association of two MAFB SNPs rs13041247 and rs11696257 with NSCL/P [11], but no positive result was obtained for the rst one in this study.
In the population that acquired both VAX1 rs7078160 A and NTN1 rs4791774 G, the risk of disease increased, higher than the risk variant obtained from either of these two genes. SNP rs7078160 A signi cantly associated with NSCL/P in rs4791774 G positive subgroup (AG or GG, p = 0.0014). Similar result discovered for rs4791774 G in rs7078160 A positive subgroup (AG or AA, p = 0.018). These two genes play independent roles in the disease onset.
VAX1 encodes a conserved homeobox transcription factors and involves in the regulation of development and morphogenesis. Mice heterozygous for the Vax1 mutation were fertile and appeared normal, although homozygous exhibited craniofacial malformations including cleft palate [12]. NTN1 encodes a laminin-related secreted protein and plays a critical role in axon guidance, cell migration and adhesion during development. Mice that lack Ntn1 die during the perinatal period with a cleft palate phenotype [18]. High-level NTN1 protein was observed in the mesenchyme, especially along the basement membrane of the palatal shelves [19]. Taken together, VAX1 is involved in the development of the tissue structure of the palate, while NTN1 may ensure the cell adhesion of the palate aps. They may play different roles in different aspects of CL/P onset. Although both SNPs are not codon variants, the double-site detection can be developed as a better early warning marker for cleft lip and palate risk in Chinese.

Conclusions
The SNPs of genes VAX1, MAFB and NTN1 are associated with NSCL/P in Chinese subjects.

Sample Collection
This study involved 311 nonsyndromic orofacial clefts cases and 480 controls. Subjects were collected from Shanghai Ninth People's Hospital a liated to Shanghai JiaoTong University School of Medicine, which were physically screened and were carefully diagnosed by at least two physicians. Healthy controls were also recruited from Shanghai. The case group consisted of 57 cleft lip (CL), 192 cleft lip and palate (CLP), and 62 cleft palate only (CPO) patients. Gender ratio is 1.75:1.
DNA samples were extracted from peripheral blood using Flexi Gene DNA Kit (Qiagen, Germany). The DNA was measured for concentration and purity and then stored at -20 ℃.
Primer Design and PCR PCR primers were designed for these ten SNPs based on hg19 of the human genome. All amplicons were in the range of 200 to 300bp. For each primer pair designed, the forward and reverse primers were tagged with a common sequence 1 (CS1: 5′-CCTACACGACGCTCTTCCGATCT-3′) and common sequence 2 (CS2: 5′-AGTTCCTTGGCACCCGAGAATTCCA-3′), respectively. The primer pairs were synthesized from Shanghai Morgen Biotechnology Co., Ltd.
All the primer pairs were divided into 3 combinations, each combination of 3 or 4 primer pairs. Every DNA sample was ampli ed in separate multiplex PCR reactions (HotStarTaq, Qiagen, Germany) with these primer combinations and mixed after electrophoresis. PCR products were cleaned up by AMPure XP Beads (Beckman Coulter, CA).

Barcoding and Illumina Sequencing
Barcoding was performed in a 20 μL reaction mixture that contained 8 μL of the cleaned up PCR products, 10 µl of KAPA2G Robust hotstart ready mix (Kapa Biosystems, USA), 1µmol/L barcode F primers and 1 µmol/L barcode R primer. The barcoded PCR products from various samples were cleaned up by the AMPure XP Beads (Beckman Coulter, CA).
Puri ed PCR product library was quanti ed using a Qubit Fluorometer. According to library quantitation, the PCR product was pooled together with equal mole. Puri ed libraries were sequenced on a MiSeq Benchtop Sequencer or a NextSeq 500 sequencer (Illumina Inc., San Diego, CA) using protocol. The quality of sequence reads were checked by FastQC algorithm. SNPs were identi ed using Genome Analysis Toolkit (GATK) and annotated by Annovar software.

Sequence Data Analysis
Demultiplexed compressed FASTQ les were generated from BCL by bcl2fastq Conversion Software v1.8.4 (Illumina, San Diego, CA). For all successful sequencing runs, read depth was 1600x at any given position, with 2700x mean coverage across the entire targeted sequence, and 100% of bases above Q30 at 2 * 150 bp. The variant calling and the coverage of each captured region were analyzed by an in-house developed bioinformatics pipeline, based on the general analysis algorithm pipeline. Brie y, the reads were mapped to the hg19 version of the human reference genome, ltered to remove off-target and poorquality reads. Variants were identi ed and annotated. The variants and annotation results were transformed into Excel sheets.

Statistical Analysis
Statistical analysis was performed with the SPSS 11.0 statistical software package (SPSS Inc., Chicago, IL, USA) or R (Version 3.0.2: www.r-project.org/). Differences in genotype and allele frequencies were analyzed using Pearson χ 2 test or Fisher's exact test and with a p<0.05 taken as being signi cant after multiple testing adjustment. Odds ratio and 95% CI of alleles had also been calculated. Power analysis was estimated assuming a prevalence of NSCL/P in China of 0.0012

Declarations
Ethics approval and consent to participate The study was approved by the Ethical Committee of Chinese National Human Genome Center (2014-09). Written informed consent was obtained from all participants or their legal guardians.

Consent for publication
Not applicable Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.