Significant Evidence of Association Between Polymorphisms in ZNF533,Environmental Factors,and Nonsyndromic Orofacial Clefts in the Western Han Chinese Population

Abstract

The etiology of nonsyndromic orofacial clefts (NSOC) has been considered “complex” or “multifactorial.” Etiologic heterogeneity induces disparities in the results among different populations. The zinc finger protein 533 (ZNF533) and several environmental factors have been revealed to be associated with NSOC in several populations. We investigated three single-nucleotide polymorphisms (SNPs) and 10 environmental factors in 211 case–parent trios and 188 control individuals in the Western Han Chinese population to confirm the relationship between ZNF533, environmental factors, and the etiology of NSOC in the Western Han Chinese population. The transmission disequilibrium test, case–control analysis, multiple logistic regression, log-linear model, and conditional logistic regression were tested to confirm the contribution of the ZNF533 gene and environmental factors to the etiology of NSOC. Strong statistically significant evidence of association was found between the rs6757845 and rs1139 markers and NSOC. The haplotype G-G for rs6757845-rs1139 showed significant overtransmission among cleft lip with or without cleft palate (CL/P) trios and among cleft palate only trios. Additional 11 and 5 haplotypes were significantly overtransmitted and undertransmitted among CL/P and among cleft palate only trios, respectively. Maternal disease, use of medication, and passive smoking during the first trimester of pregnancy may increase the risk of NSOC. Maternal folic acid supplementation during the first trimester of pregnancy showed a protective effect on the etiology of NSOC. Genotype–environment interaction test showed a significant evidence of interaction effects between the genotypes at rs6757845 and maternal passive smoking during the first trimester among CL/P trios. These results confirm the effects of the ZNF533 gene and environmental factors on the etiology of NSOC.

Introduction

O rofacial clefts are among the most common birth defects in humans. These defects create serious physical, social, and emotional burdens for affected individuals and their families. The average occurrence is 1–2 per 1000 live births. However, there is a wide variability related to geographic origin, socioeconomic status, and subpopulation. Asian and Amerindian populations have the highest frequencies of this type of defect. Caucasian populations have an intermediate number of occurrences and African-derived populations have the lowest prevalence (Forrester and Merz, 2004). The ethnic Chinese population has a relatively high incidence of 1.62 per 1000 live births (Mossey and Little, 2002).

Although over 300 malformation syndromes include orofacial cleft, nonsyndromic forms represent majority of the cases. Orofacial clefts include cleft lip (CL), cleft lip and palate, and cleft palate only (CPO). CL and cleft lip and palate are always grouped into cleft lip with or without cleft palate (CL/P), according to their similarities in epidemiologic characteristics and embryologic timing.

Etiological studies have suggested that nonsyndromic orofacial clefts (NSOC) were considered as a “complex” or “multifactorial” disease. Both environmental and genetic factors contribute to the etiology of NSOC. Previous studies have revealed several environmental factors such as maternal smoking, vitamin intake, alcohol use, and medication to be involved in the pathogenesis of orofacial clefts (Shaw et al., 1995, Shaw and Lammer, 1999; Beaty et al., 2001; Little et al., 2004). The environmental agents are not the sole causes. Orofacial cleft occurrences are further increased by the joint effects of candidate genes and genotype–environment interaction. Several candidate genes have shown statistical evidence of association with NSOC across multiple studies (Jugessur and Murray, 2005). Disparities in the results always present in different studies. Linkage studies have revealed positive evidence for multiple chromosomal regions of linkage. A meta-analysis of 13 genome scans found that 16 different chromosomal regions showed statistically significant evidence of linkage, including suggestive evidence for linkage to 2p13 and stronger evidence for the 2q32–35 region (Marazita et al., 2004). Further, analysis of the chromosome 2 genes spanning three populations recently identified the zinc finger protein 533 (ZNF533) as a candidate gene (Beaty et al., 2006). Fifty-five single-nucleotide polymorphisms (SNPs) in ZNF533 gene regions were genotyped. Three SNPs at the 3′ end of the gene were significant at 1% level in the Taiwan CL/P subpopulation. In the middle region of the gene, only Taiwan and Maryland CL/P samples showed a significant evidence of SNPs and haplotypes. The SNPs nearby rs1517710 were significant at the 5% level in Singapore CL/P samples. The ethnic Chinese in both Singapore and Taiwan originated from southern China. The particular polymorphisms in ZNF533 associated with NSOC were different. The population stratification, sampling variation, the potential of heterogeneity, and the difference of genetic backgrounds may explain the disparities in the results.

Because of etiological heterogeneity, different samples of subpopulations may lead to different results. The identification of any pathogenic factor remains a challenge. We tested three SNPs in this study and analyzed environmental factors in 211 unrelated NSOC trios and 188 control individuals of the Western Chinese population. Transmission disequilibrium test (TDT), case–control analysis, multiple logistic regression, log-linear model, and conditional logistic regression were carried out to confirm the roles of ZNF533, enviromental factors, and genotype–environment interaction in the etiology of NSOC.

Materials and Methods

Family assessment

Our samples comprised 211 case–parent trios and 188 control individuals (without major congenital malformation, had no family history of genetic disease, were born in the same region, and matched by sex ratio as possible). All participants were recruited between 2005 and 2007 from the Surgery Department of the West China Stomatology College of Sichuan University. Research protocols were reviewed and approved by the Hospital Ethics Committee of West China College of Stomatology, Sichuan University. Written informed consent and permission to draw a blood sample from the child during surgery was obtained from each mother. All probands were given a physicial examination to exclude either syndromic cases or those with associated congenital anomalies. Participants were asked about the history of orofacial clefts. All participants were self-identified as Western Han Chinese. The categories of cleft, laterality, severity, and gender among probands are shown in Table 1.

Table 1.

Categories of the Nonsyndromic Orofacial Clefts

	CL/P	CPO
Case	155	56
Sporadic case	147	54
Family history	8	2
Left unilateral cleft	43	—
Right unilateral cleft	70	—
Bilateral cleft	98	—
Complete cleft	92	4
Incomplete cleft	63	52
Male	106	25
Female	49	31

CL/P, cleft lip with or without cleft palate; CPO, cleft palate only.

Data collection

Questionnaire surveys were distributed to all participants' mothers. Environmental factors of orofacial clefts, especially in the period of 3 months after conception, were assessed, including maternal abortion history, smoking history (both active and passive smoking), alcohol consumption (maternal and paternal alcohol consumption), medication history (antibiotics, cold cure, and antiemetics), disease history, and vitamin (vitamin complex and folic acid) and calcium supplementation. According to the World Health Organization definition, passive smoking is defined as exposure for at least 15 min per day or >1 day per week (World Health Organization, 2001). All the participants were dichotomized by the exposures and classified as “exposed” or “unexposed.”

Genotyping

Venous blood samples (infant umbilical cord blood for control individual) were drawn from participants. Genomic DNA was extracted by phenol–chloroform extraction protocol. The three SNPs in ZNF533 were genotyped by restriction fragment length polymorphism–polymerase chain reaction. The primers and restriction enzymes are shown in Supplemental Table 1 (available online at www.liebertonline.com). All the reactions were done in duplicate. Agarose gels of restriction enzyme-digested fragments are shown in Supplemental Figure 1 (available online at www.liebertonline.com).

Statistical analysis

The Hardy–Weinberg equilibrium (HWE) was checked for two SNPs (rs3749123 and rs1139) in the parents of affected individuals and the control group. The power of test was calculated using the package for family-based association test (PBAT) software (www.biostat.harvard.edu/clange/default.htm). TDT was used to evaluate whether the proportion of transmitted target alleles from the heterozygous parents to the affected offspring deviated from those expected under Mendelian frequency in the case–parent trios. Pairwise linkage disequilibrium (LD) was computed as both D′ and r ² for all the SNPs using the Haploview program (www.broad.mit.edu/mpg/haploview/index.php/). Individual SNPs were tested using the family-based association test (FBAT) program (www.biostat.harvard.edu/fbat/default.html) to compare the genotype distribution observed in the “case” with its expected distribution under the null hypothesis: “no linkage and no association” or “no association in the presence of linkage.” Haplotype transmission disequilibrium (sliding windows consisting of two and three SNPs) was tested using the “hbat” command in the FBAT program. Case–control statistical analysis and χ ² analysis were performed for screening the genotypic and environmental factors in NSOC using the SPSS 11.5 statistical software package. Multiple logistic regression was used to analyze the association among the SNPs of ZNF533, environmental factors, and NSOC. The log-linear approach was used to compare the differential contributions of different factors to NSOC. The relative risks (RRs) with standard Wald confidence intervals (CI) were computed. Gene–environment interaction was tested using the case–parent trios design and conditional logistic regression analysis to avoid spurious results due to confounding or population stratification within the sample. Conditional logistic regression uses the observed genotype and “pseudo-sib” controls derived from the observed parental mating type as a matched set (Schaid, 1996; Beaty et al., 2002). Environmental exposures are introduced into the model as a covariate and coded “1” or “0” for “exposed” or “unexposed,” respectively.

Results

The rs6757845 genotype A/A frequency for our samples was found to be “zero,” which was in accordance with the NCBI dbSNP database (www.ncbi.nlm.nih.gov/SNP/) for Asian population. The rs6757845 did not meet HWE. HWE was checked for the rs3749123 and rs1139 among the parents and control group. The Hardy–Weinberg law strictly holds only under the conditions of random mating, no selection or migration, no mutation, no population stratification, and infinite population size. Deviations from HWE could be interpreted as genotyping errors or the occurrence of population stratification. Pearson's χ ² tests showed that two SNPs among the parents and control infants were compatible with HWE (Supplemental Table 2, available online at www.liebertonline.com). For excluding the genotyping errors, all three SNPs were genotyped twice.

The power was computed using Monte-Carlo simulation, underlying an additive genetic model at the 5% test level. Statistical parameters were defined as follows: population prevalence of the disease was 0.00162, sample size for the FBAT was 172, and rs1139 was identified as disease locus. The power of the FBAT was 0.948.

LD was computed as both D′ and r ² using the Haploview program, and the results showed a weak LD among markers (D′ < 0.3, r ² < 0.3; Table 2).

Table 2.

Pairwise Linkage Disequilibrium Measured as D′ (Above the Diagonal) and r ² (Below the Diagonal)

Marker	rs3749123	rs6757845	rs1139
rs3749123		0.062^a	0.292^a
rs6757845	0.0010^b		0.179^a
rs1139	0.016^b	0.028^b

Lewontin D′ values.

r ² values.

TDT and FBAT analysis of alleles and genotypes

TDT was undertaken on case–parent trios with heterozygous informative parents. At rs3749123, allelic TDT showed no evidence for association of CL/P or CPO. G allele showed a highly statistically significant evidence of overtransmission at rs6757845 for CL/P and CPO (p < 0.01). The odds ratios (ORs) for allele G at rs6757845 were 6.454 (95% CI: 3.743–11.131) for CL/P and 7.109 (95% CI: 2.459–20.553) for CPO. At rs1139, allele G showed excess transmission among CL/P (p = 0.000) and CPO (p = 0.006). The ORs for transmission of allele G at rs1139 were 2.201 (95% CI: 1.552–3.120) for CL/P and 2.382 (95% CI: 1.283–4.423) for CPO (Table 3).

Table 3.

Allelic Transmission Disequilibrium Test Results for 211 Case–Parent Trios

		CL/P					CPO
Marker	Allele	T/NT	χ²	p-Value	OR	95% CI	T/NT	χ²	p-Value	OR	95% CI
rs3749123	T	128/117	1.090	0.296	1.219	0.840–1.770	44/53	1.766	0.184	0.675	0.377–1.207
	C	96/107					48/39
rs6757845	G	170/108	51.514	0.000	6.454	3.742–11.131	51/33	15.429	0.000	7.109	2.459–20.553
	A	20/82					5/23
rs1139	G	157/106	19.855	0.000	2.201	1.552–3.120	58/40	7.685	0.006	2.382	1.283–4.423
	A	105/156					28/46

T/NT indicates transmitted or nontransmitted counts from heterozygous parents to offsprings. Bold values indicate p < 0.01.

OR, odds ratio; CI, confidence interval.

FBAT analysis confirmed the results of TDT. Allele G and genotype G/G at rs6757845 showed a high overtransmission among CL/P trios (p < 10⁻⁶, collectively) and CPO trios (p = 0.001069 and p = 0.001069, respectively). Allele A and genotype G/A at rs6757845 showed a significant undertransmission among CL/P and CPO trios. At rs1139, allele G and genotype G/G showed a significant overtransmission among CL/P trios (Z = 3.833, p = 0.000126; Z = 2.953, p = 0.003143, respectively) and CPO trios (Z = 2.546, p = 0.010909; Z = 2.145, p = 0.031972, respectively). Correspondingly, allele A at rs1139 showed a significant undertransmission among CL/P and CPO trios. Genotype A/A at rs1139 showed a statistically significant evidence of undertransmission among CL/P trios (Z = −2.858, p = 0.004259; Table 4).

Table 4.

Family-Based Association Test Results for Single-Nucleotide Polymorphisms in ZNF533, Using a Biallelic Genetic Model (Additive and Genotype)

		CL/P				CPO
Marker	Allele/genotype	A. freq.	Family no.^a	Z	p-Value	A. freq.	Family no. ^a	Z	p-Value
rs3749123	T	0.555	112	0.889	0.373843	0.515	44	−1.172	0.241317
	C	0.445	112	−0.889	0.373843	0.485	44	1.172	0.241317
rs6757845	G	0.838	94	6.070	0.000000*	0.879	27	3.272	0.001069*
	A	0.162	94	−6.070	0.000000*	0.121	27	−3.272	0.001069*
rs1139	G	0.497	131	3.833	0.000126*	0.532	41	2.546	0.010909**
	A	0.503	131	−3.833	0.000126*	0.468	41	−2.546	0.010909**
rs3749123	TT	0.309	87	1.084	0.278193	0.236	32	−0.847	0.397191
	TC	0.492	112	−0.756	0.449692	0.558	44	0.000	1.000000
	CC	0.199	66	−0.201	0.840781	0.205	27	0.933	0.350688
rs6757845	GG	0.676	94	5.987	0.000000*	0.758	27	3.272	0.001069*
	GA	0.324	94	−5.570	0.000000*	0.242	27	−3.272	0.001069*
rs1139	GG	0.213	89	2.953	0.003143*	0.297	31	2.145	0.031972**
	GA	0.569	131	−0.087	0.930377	0.468	41	−0.781	0.434880
	AA	0.218	88	−2.858	0.004259*	0.234	19	−1.588	0.112240

Bold values indicate *p < 0.01 and **p < 0.05.

Informative family.

A. freq., allele frequency; Z, vector of the large sample Z statistic.

TDT and FBAT analyses showed stronger association between the rs6757845 and rs1139 markers and NSOC. No significant association was obtained between rs3749123 marker and NSOC.

Haplotypes in TDT analysis

Preceding studies have shown that haplotype analyses have a greater power than those based on single markers. TDT analysis based on haplotypes was carried out using genotype data to confirm the association between ZNF533 and NSOC. Sliding window haplotypes consisting of two and three SNPs showed a significant evidence of overtransmission and undertransmission among CL/P trios and CPO trios. The haplotype G-G (for rs6757845-rs1139) showed a significant evidence of overtransmission among CL/P trios and CPO trios. T-G-G (for rs3749123-rs6757845-rs1139), T-G (for rs3749123-rs6757845), and T-G (for rs3749123-rs1139) were significantly overtransmitted among CL/P trios. C-G (for rs3749123-rs6757845) was significantly overtransmitted among CPO trios. Conversely, the haplotypes C-A (for rs3749123-rs6757845), T-A (for rs3749123-rs1139), and A-A (for rs6757845-rs1139) showed a significant evidence of undertransmission among CL/P trios and CPO trios. C-A-A, T-A-A, and T-A-G (for rs3749123-rs6757845-rs1139), T-A (for rs3749123-rs6757845), and A-G (for rs6757845-rs1139) were significantly undertransmitted among CL/P trios. T-G-A (for rs3749123-rs6757845-rs1139) was significantly undertransmitted among CPO trios (Table 5). Sliding window haplotypes of size 2 were more informative than the three-SNP haplotypes, whereas haplotypes of size 3 were more significant. The haplotypes composed of rs3749123 showed a significant evidence of association with NSOC, whereas the single marker showed no statistically significant association with NSOC.

Table 5.

Transmission Disequilibrium Test Based on Haplotype Results

		Marker haplotype
	No.	rs3749123	rs6757845	rs1139	H. freq.	Z	p-Value
CL/P
	h1	T	G	G	0.275	5.360	0.000000
	h2	T	G	—	0.471	3.366	0.000762
	h3	T	—	G	0.310	4.019	0.000058
	h4	—	G	G	0.436	5.365	0.000000
	h5	C	A	A	0.055	−4.453	0.000008
	h6	T	A	A	0.047	−3.308	0.000940
	h7	T	A	G	0.036	−2.719	0.006541
	h8	T	A	—	0.082	−4.040	0.000053
	h9	C	A	—	0.081	−4.389	0.000011
	h10	T	—	A	0.244	−2.945	0.003230
	h11	—	A	A	0.101	−5.288	0.000000
	h12	—	A	G	0.062	−3.051	0.002284
CPO
	h1	C	G	—	0.432	2.433	0.014987
	h2	—	G	G	0.492	2.696	0.007025
	h3	T	G	A	0.165	−2.367	0.017944
	h4	C	A	—	0.055	−2.887	0.003892
	h5	T	—	A	0.209	−3.332	0.000864
	h6	—	A	A	0.083	−3.856	0.000115

H. freq., haplotype frequency.

Case–control statistical analysis, multiple logistic regression, and log-linear analysis

The χ ² analysis was performed for screening the etiopathogenic factors. Eight genotypic and environmental factors—G/G genotype at rs1139, maternal use of medication during the first trimester, maternal disease history during the first trimester, maternal passive smoking during the first trimester, maternal vitamins supplementation during the first trimester, maternal folic acid supplementation during the first trimester, maternal calcium supplementation during the first trimester, and paternal alcohol consumption—showed a significant evidence of a difference for NSOC cases compared with controls (Supplemental Table 3, available online at www.liebertonline.com).

Multiple logistic regression analysis was carried out using the data of these significant factors. The regression coefficient, OR, and 95% CI were calculated. Five factors entered into the regression equation. G/G genotype at rs1139 (OR = 8.091, 95% CI: 3.811–17.177), maternal disease history during the first trimester (OR = 3.701, 95% CI: 1.430–9.577), maternal use of medication during the first trimester (OR = 4.544, 95% CI: 1.908–10.821), and maternal passive smoking during the first trimester (OR = 5.073, 95% CI: 2.836–9.076) were the risk factors for NSOC. Maternal folic acid supplementation during the first trimester was demonstrated as a protective factor in the etiology of NSOC (OR = 0.316, 95% CI: 0.124–0.807; Table 6).

Table 6.

Multiple Logistic Regression Analysis Results for the Association Among the Single-Nucleotide Polymorphisms of ZNF533, Environmental Factors, and Nonsyndromic Orofacial Clefts

Factors	Regression coefficient	χ²	p-Value	OR	95% CI
G/G at rs1139	2.091	29.629	0.000	8.091	3.811–17.177
Maternal medication history during first trimester	1.514	11.694	0.001	4.544	1.908–10.821
Maternal disease history during first trimester	1.309	7.278	0.007	3.701	1.430–9.577
Maternal passive smoking during first trimester	1.624	29.950	0.000	5.073	2.836–9.076
Maternal folic acid supplementation during first trimester	−1.151	5.810	0.016	0.316	0.124–0.807

OR, odds ratio; CI, confidence interval.

Maternal disease and use of medication showed high association. It was shown that both of them increased the risk of NSOC. The log-linear approach was used to compare the differential contributions. The result showed that the factor of maternal medication use during the first trimester had greater contribution to NSOC than the factor of maternal disease history (RRm = 7.06, 95% CI: 1.13–44.26; RRd = 4.43, 95% CI: 1.85–10.58, respectively). The risk of NSOC was 1.59-fold higher among the offsprings of mothers who used medications, compared with those who had disease history during the first trimester (RR = 1.59).

ZNF533 genotype–environment interaction effects

The interaction effects of the risk genotypes at three SNPs and the four factors entered into the regression equation were examined using the case–parent trios design and the conditional logistic regression analysis. The trios were stratified into “exposed” and “unexposed” groups. The likelihood ratio test was used to test for interaction effects by comparing the model with environmental exposures to the model with environmental factors “unexposed.” The results showed a significant evidence of genotype–environment interaction between the genotypes at rs6757845 and maternal passive smoking during the first trimester among CL/P trios (likelihood ratio test = 6.051, p = 0.049). Offsprings carrying homozygous G allele were 3.786 times more likely to be affected if mothers were exposed to environmental tobacco smoke during the first trimester (OR = 3.786, 95% CI: 2.101–6.822). G/A genotype at rs6757845 showed a protective effect on the etiology of CL/P (OR = 0.264, 95% CI: 0.147–0.476; Table 7).

Table 7.

Conditional Logistic Regression Analysis Results of Polymorphisms in ZNF533 Genotype–Environment Interaction Effects

		CL/P			CPO
SNPs	Genotype	Counts of case	OR (95% CI)	LRT ^a (df; significance)	Counts of case	OR (95% CI)	LRT ^a (df; significance)
rs3749123
No passive smoking	TT	9	0.951 (0.310–2.919)	2.3431 (2; 0.310)	4	0.356 (0.071–1.792)	0.123 (2; 0.940)
No passive smoking	TC	22	1.260 (0.527–3.008)		8	0.493 (0.135–1.795)
Passive smoking	TT	29	1.412 (0.644–3.096)		7	0.736 (0.211–2.565)
Passive smoking	TC	30	0.869 (0.436–1.732)		14	0.799 (0.308–2.076)
rs6757845
No passive smoking	GG	15	1.154 (0.549–2.425)	6.051 (2; 0.049)	6	6.000 (0.722–49.837)	0.117 (2; 0.943)
No passive smoking	GA	13	0.867 (0.412–1.821)		1	0.167 (0.020–1.384)
Passive smoking	GG	53	3.786 (2.101–6.822)		16	4.000 (1.337–11.965)
Passive smoking	GA	14	0.264 (0.147–0.476)		4	0.250 (0.084–0.748)
rs1139
No passive smoking	GG	14	2.455 (0.888–6.786)	0.221 (2; 0.895)	10	1.536 (0.220–10.748)	3.019 (2; 0.221)
No passive smoking	GA	25	1.700 (0.769–3.762)		4	0.421 (0.070–2.541)
Passive smoking	GG	32	3.843 (1.702–8.678)		9	4.948 (0.888–27.575)
Passive smoking	GA	39	2.018 (1.007–4.046)		15	3.928 (0.858–17.979)

Bold value indicates p < 0.05.

Compared with the model without maternal environmental exposures.

SNP, single-nucleotide polymorphism; LRT, likelihood ratio test.

Discussion

In a previous study (Beaty et al., 2006), ZNF533 showed a consistent evidence of linkage and disequilibrium in Maryland, Singapore, and Taiwan populations for NSOC. But for individual SNPs, high levels of heterogeneity appeared among three populations. The SNP rs3749123 showed a significant evidence of linkage and disequilibrium in the Singapore CL/P trios. The SNP rs1139 was individually significant in the Taiwan CL/P trios, whereas rs6757845 showed statistically significant evidence of linkage and disequilibrium in the Maryland and Taiwan CL/P trios. In our study, rs3749123 allelic and genotypic TDT showed no significant evidence of association with CL/P or CPO trios. At rs6757845, highly significant appearances of distortion in allelic and genotypic transmission were found in both CL/P and CPO trios. Allele G and G/G homozygotes showed significant overtransmission and association with the etiology of NSOC. This model is consistent with a recessive genetic effect of the G allele. In our samples, no transmitted A/A genotype at rs6757845 was found, which may explain the weak LD between rs3749123 and rs6757845 (D′ = 0.062, r ² = 0.0010) despite their close proximity. Allelic TDT also showed the G allele at rs1139 to be significantly overtransmitted in both CL/P and CPO trios. Allele G was identified to be a risk factor for NSOC. G/G homozygotes at rs1139 were significantly overtransmitted among CL/P and CPO trios, and correspondingly, A/A homozygotes were significantly undertransmitted among CL/P trios. This result confirmed the recessive genetic effect of the ZNF533 gene.

The position of the three SNPs located on both ends of ZNF533 gene. The TDT based on haplotypes showed a high distortion of transmission. Five haplotypes overtransmitted from the heterozygous parents to the affected offspring. The result further confirms the association between ZNF533 and NSOC.

The overtransmitted genotypes were identified as risk factors for NSOC. The FBAT and TDT analyses suggested T/T genotype at rs3749123; although no significant evidence of association was found, G/G genotypes at rs6757845 and rs1139 were found to be overtransmitted. These genotypes were introduced into the case–control analysis. G/G genotype at rs6757845 showed no significant difference between the case and control children. The approach with FBAT in the trios compared the alleles transmitted to affected offspring with the expected distribution of alleles among offspring, or with the nontransmitted alleles. Only the families with heterozygous parents were informative for FBAT and TDT. Ninety trios among NSOC families in our samples were noninformative for rs6757845 family-based analysis. This population structure may explain the discrepancy between the family-based studies and the case–control analysis.

The association tests of individual SNPs showed heterogeneity among different subpopulations, even with the same origin (Singapore and Taiwan both originated from China). Haplotypes of ZNF533 were consistently significant for both CL/P and CPO in the Western Chinese population. The consistency of statistical evidence across different subpopulations suggested that this gene is a candidate for the etiology of NSOC.

The function of ZNF533 has relatively little recognition. This gene belongs to the family of zinc-finger genes, which code for the proteins of transcription factors, playing an essential role in altering gene expression. ZNF533 is thought to be a repressor of transcription, but the specific targets are yet not known (O'Geen et al., 2007). It is widely expressed (Strausberg et al., 2002) in the tissues of the developing palate and lip, according to the data of the Craniofacial and Oral Gene Expression Network (COGENE-http://hg.wustl.edu/COGENE). Recent studies have revealed the correlation of the ZNF533 deletion and the mental retardation. The patients exhibit mental retardation and multiple congenital anomalies, including high-arched palate, cleft lip, and cleft palate, which further supports the role of this gene in orofacial clefts (Monfort et al., 2008). The causative mechanism of ZNF533 for NSOC, the gene–gene interaction, and the gene–environment interaction are totally unclear and worth further investigation.

Epidemiological studies have revealed several environmental factors associated with the risk of orofacial clefts. Maternal use of medications, such as retinoids, anticonvulsants, folate antagonists, and corticosteroids, increases the risk of having a child with orofacial clefts (Chan et al., 1996; Carmichael and Shaw, 1999; Hernández-Diaz et al., 2000; Artama et al., 2005). In our samples, the overwhelming majority of mothers who used medications during the first trimester had a history of influenza, which was the main categories of the maternal disease history. Antibiotics and the cold medications were the main maternal use categories (data not shown). The multiple logistic regression analysis showed a 4.544-fold increase in the risk of NSOC among the offspring of mothers who used medications compared with those who had no medication use history (OR = 4.544, 95% CI: 1.908–10.821). Women with disease history were 3.701 times more likely to have NSOC offspring (OR = 3.701, 95% CI: 1.430–9.577). The log-linear model was used to compare the differential contributions with the etiology of NSOC. Mothers who used medication during the first trimester were 1.59 times more likely to have NSOC offspring, compared with those who had disease history (RR = 1.59).

Maternal smoking is one of the most widely investigated risk factors for NSOC. A meta-analysis based on 24 published studies has proved that women who smoke during the first trimester of pregnancy have an increased risk of having NSOC offspring. Moreover, a dose–response relationship between the risk of NSOC and the number of cigarettes smoked was supported by the published studies (Little et al., 2004). Smoking included both active and passive smoking. Relatively little is known about the impact of passive smoking on NSOC. In the Asian region, more than 60% of men are smokers, and only a few percentage of women smoke. The rates of exposure to environmental tobacco smoke in women are higher in China (World Health Organization, 2001). In our study, no mother was a smoker. Higher rates of passive smoking were found in our samples. Multiple logistic regression analysis showed 5.073-fold increase in the risk of having NSOC offspring among the mothers who were exposed to environmental smoke during the first trimester (OR = 5.073, 95% CI: 2.836–9.076).

Folic acid has been proved to be a protective factor for the neural tube defects and to be associated with the risk of NSOC (Ray et al., 2002; Johnson and Little, 2008). Although several studies have investigated the relationship between folic acid and NSOC, the results are inconsistent. Maternal folic acid supplementation showed a protective effect for the reduced risk of NSOC (OR = 0.316, 95% CI: 0.124–0.807) in our study.

The etiology of NSOC is thought to be determined by genetic or environmental factors. Recent studies have focused on the joint effects of candidate genes and environmental risk factor interactions. The genetic susceptibility could be modulated by several environmental agents. The interaction between the genetic and environmental factors has been proven to increase the risk of NSOC (e.g., maternal smoking and offspring genotypes for transforming growth factor α, acetyl-N-transferase 1, and glutathione S-transferase theta-1 (Shaw et al., 1996; Lammer et al., 2004). In our study, the case–parent trio design and the conditional logistic regression analysis were used to test the genotype environment interaction. The interaction effects of risk genotypes at three SNPs and four selected environmental factors, which entered into the regression equation, were tested. The genotypes at rs6757845 showed heterogeneity of genetic effects between CL/P-affected offsprings of mothers exposed and unexposed to environmental tobacco smoke. Offsprings who carried G/G genotype and whose mothers were exposed to environmental tobacco smoke have 3.786-fold increased risk of CL/P (OR = 3.786). The genetic effects of genotypes at rs6757845 may be susceptible to the effects of maternal passive smoking and hence increase the risk of CL/P. The modest sample size of informative families was the limitation of the study. Studies with larger sample size would be needed to investigate the interaction effects of ZNF533 polymorphism and the maternal passive smoking during the first trimester of pregnancy.

In summary, our study confirmed that the ZNF533 gene is associated with the etiology of NSOC. Different results suggest the etiologic heterogeneity in the Western Chinese population. The causative locus in ZNF533 and the mechanism of ZNF533 for the etiology of NSOC are indefinite and need further investigation.

Footnotes

Acknowledgments

This research was supported by the National Science Funds of China (30660198). The authors thank all the participants who donated samples for the international study of oral clefts in this research.

Disclosure Statement

No competing financial interests exist.

References

Artama

, Auvinen

, Raudaskoski

, Isojärvi

2005. Antiepileptic drug use of women with epilepsy and congenital malformations in offspring. Neurology, 64:1874–1878.

Beaty

T.H.

, Hetmanski

J.B.

, Fallin

M.D.

, Park

J.W.

, Sull

J.W.

, McIntosh

, Liang

K.Y.

, Vanderkolk

C.A.

, Redett

R.J.

, Boyadjiev

S.A.

, Jabs

E.W.

, Chong

S.S.

, Cheah

F.S.

, Wu-Chou

Y.H.

, Chen

P.K.

, Chiu

Y.F.

, Yeow

, Ng

I.S.

, Cheng

, Huang

, Ye

, Wang

, Ingersoll

, Scott

A.F.

2006. Analysis of candidate genes on chromosome 2 in oral cleft case-parent trios from three populations. Hum Genet, 120:501–518.

Beaty

T.H.

, Hetmanski

J.B.

, Zeiger

J.S.

, Fan

Y.T.

, Liang

K.Y.

, VanderKolk

C.A.

, McIntosh

2002. Testing candidate genes for non-syndromic oral clefts using a case-parent trio design. Genet Epidemiol, 22:1–11.

Beaty

T.H.

, Wang

, Hetmanski

J.B.

, Fan

Y.T.

, Zeiger

J.S.

, Liang

K.Y.

, Chiu

Y.F.

, Vanderkolk

C.A.

, Seifert

K.C.

, Wulfsberg

E.A.

, Raymond

, Panny

S.R.

, McIntosh

2001. A case-control study of nonsyndromic oral clefts in Maryland. Ann Epidemiol, 11:434–442.

Chan

, Hanna

, Abbott

, Keane

R.J.

1996. Oral retinoids and pregnancy. Med J Aust, 165:164–167.

Carmichael

S.L.

, Shaw

G.M.

1999. Maternal corticosteroid use and risk of selected congenital anomalies. Am J Med Genet, 86:242–244.

Forrester

M.B.

, Merz

R.D.

2004. Descriptive epidemiology of oral clefts in a multiethnic population, Hawaii, 1986-2000. Cleft Palate Craniofac J, 41:622–628.

Hernández-Díaz

, Werler

M.M.

, Walker

A.M.

, Mitchell

A.A.

2000. Folic acid antagonists during pregnancy and the risk of birth defects. N Engl J Med, 343:1608–1614.

Johnson

C.Y.

, Little

2008. Folate intake, markers of folate status and oral clefts: is the evidence converging? Int J Epidemiol, 37:1041–1058.

10.

Jugessur

, Murray

J.C.

2005. Orofacial clefting: recent insights into a complex trait. Curr Opin Genet Dev, 15:270–278.

11.

Lammer

E.J.

, Shaw

G.M.

, Iovannisci

D.M.

, Van Waes

, Finnell

R.H.

2004. Maternal smoking and the risk of orofacial clefts: susceptibility with NAT1 and NAT2 polymorphisms. Epidemiology, 15:150–156.

12.

Little

, Cardy

, Munger

R.G.

2004. Tobacco smoking and oral clefts: a meta-analysis. Bull World Health Organ, 82:213–218.

13.

Marazita

M.L.

, Murray

J.C.

, Lidral

A.C.

, Arcos-Burgos

, Cooper

M.E.

, Goldstein

, Maher

B.S.

, Daack-Hirsch

, Schultz

, Mansilla

M.A.

, Field

L.L.

, Liu

Y.E.

, Prescott

, Malcolm

, Winter

, Ray

, Moreno

, Valencia

, Neiswanger

, Wyszynski

D.F.

, Bailey-Wilson

J.E.

, Albacha-Hejazi

, Beaty

T.H.

, McIntosh

, Hetmanski

J.B.

, Tunçbilek

, Edwards

, Harkin

, Scott

, Roddick

L.G.

2004. Meta-analysis of 13 genome scans reveals multiple cleft lip/palate genes with novel loci on 9q21 and 2q32-35. Am J Hum Genet, 75:161–173.

14.

Monfort

, Roselló

, Orellana

, Oltra

, Blesa

, Kok

, Ferrer

, Cigudosa

J.C.

, Martínez

2008. Detection of known and novel genomic rearrangements by array based comparative genomic hybridisation: deletion of ZNF533 and duplication of CHARGE syndrome genes. J Med Genet, 45:432–437.

15.

Mossey

P.A.

, Little

2002. Epidemiology of oral clefts: a international perspective. In: Cleft Lip and Palate: From Origin to Treatment. Wyszynski

D.F.

Oxford University Press: Oxford, England, 127–158.

16.

O'Geen

, Squazzo

S.L.

, Iyengar

, Blahnik

, Rinn

J.L.

, Chang

H.Y.

, Green

, Farnham

P.J.

2007. Genome-wide analysis of KAP1 binding suggests autoregulation of KRAB-ZNFs. PLoS Genet, 3e891–11.

17.

Ray

J.G.

, Meier

, Vermeulen

M.J.

, Boss

, Wyatt

P.R.

, Cole

D.E.

2002. Association of neural tube defects and folic acid food fortification in Canada. Lancet, 360:2047–2048.

18.

Schaid

D.J.

1996. General score tests for associations of genetic markers with disease using cases and their parents. Genet Epidemiol, 13:423–449.

19.

Shaw

G.M.

, Lammer

E.J.

, Wasserman

C.R.

, O'Malley

C.D.

, Tolarova

M.M.

1995. Risks of orofacial clefts in children born to women using multivitamins containing folic acid periconceptionally. Lancet, 346:393–396.

20.

Shaw

G.M.

, Lammer

E.J.

1999. Maternal periconceptional alcohol consumption and risk for orofacial clefts. J Pediatr, 134:298–303.

21.

Shaw

G.M.

, Wasserman

C.R.

, Lammer

E.J.

, O'Malley

C.D.

, Murray

J.C.

, Basart

A.M.

, Tolarova

M.M.

1996. Orofacial clefts, parental cigarette smoking, and transforming growth factor-alpha gene variants. Am J Hum Genet, 58:551–561.

22.

Strausberg

R.L.

, Feingold

E.A.

, Grouse

L.H.

, Derge

J.G.

, Klausner

R.D.

, Collins

F.S.

, Wagner

, Shenmen

C.M.

, Schuler

G.D.

, Altschul

S.F.

, Zeeberg

, Buetow

K.H.

, Schaefer

C.F.

, Bhat

N.K.

, Hopkins

R.F.

, Jordan

, Moore

, Max

S.I.

, Wang

, Hsieh

, Diatchenko

, Marusina

, Farmer

A.A.

, Rubin

G.M.

, Hong

, Stapleton

, Soares

M.B.

, Bonaldo

M.F.

, Casavant

T.L.

, Scheetz

T.E.

, Brownstein

M.J.

, Usdin

T.B.

, Toshiyuki

, Carninci

, Prange

, Raha

S.S.

, Loquellano

N.A.

, Peters

G.J.

, Abramson

R.D.

, Mullahy

S.J.

, Bosak

S.A.

, McEwan

P.J.

, McKernan

K.J.

, Malek

J.A.

, Gunaratne

P.H.

, Richards

, Worley

K.C.

, Hale

, Garcia

A.M.

, Gay

L.J.

, Hulyk

S.W.

, Villalon

D.K.

, Muzny

D.M.

, Sodergren

E.J.

, Lu

, Gibbs

R.A.

, Fahey

, Helton

, Ketteman

, Madan

, Rodrigues

, Sanchez

, Whiting

, Madan

, Young

A.C.

, Shevchenko

, Bouffard

G.G.

, Blakesley

R.W.

, Touchman

J.W.

, Green

E.D.

, Dickson

M.C.

, Rodriguez

A.C.

, Grimwood

, Schmutz

, Myers

R.M.

, Butterfield

Y.S.

, Krzywinski

M.I.

, Skalska

, Smailus

D.E.

, Schnerch

, Schein

J.E.

, Jones

S.J.

, Marra

M.A.

Mammalian Gene Collection Program Team. 2002. Generation and initial analysis of more than 15,000 full-length human and mouse cDNA sequences. Proc Natl Acad Sci USA, 99:16899–16903.

23.

World Health Organization. 2001. Women and the Tobacco Epidemic: Challenges for the 21st Century. Jonathan

M.S.

World Health Organization: Geneva, Switzerland, 17–19.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.08 MB