Association between IRF6 SNPs and Oral Clefts in West China

Abstract

Analyses of previous data have confirmed the contribution of the IRF6 gene to susceptibility to non- syndromic oral clefts (NSOC) in some populations. We tested for associations between the rs2013162, rs2235375, and rs2235371 polymorphisms in IRF6 and the risk of NSOC, using both case-parent trio and case-control designs on samples from western China. Our study group consisted of 332 persons with NSOC, their parents (289 mothers and 243 fathers for 206 complete trios for these three SNPs), and 174 control individuals. We found strong evidence of over- and under-transmission of the C allele (the Val allele) at rs2235371, and the C allele at rs2235375 in cleft case-parent trios (P = 0.013 and P = 0.000, respectively). There were significant differences in the frequency distributions of both genotypes and alleles when cases were compared with control infants at rs2235371 and rs2235375. Five specific haplotypes showed significant over- and under-transmission. These results further support a role for IRF6 variants in western Chinese populations.

INTRODUCTION

Non-syndromic oral clefting (NSOC) is a complex or multifactorial disease which has a high prevalence, 1.62/1000 live births, in China (Mossey and Little, 2002). The etiology of NSOC involves both environmental and genetic factors (Lidral and Murray, 2004). Based on both linkage and association studies, several genes appear to influence risk for NSOC, but it has been very difficult to replicate findings across different populations (Marazita et al., 2002a,b).

One approach to studying the genetics of NSOC is to evaluate genes known to control malformation syndromes which include oral clefts, especially to those phenotypically closest to non-syndromic NSOC. Such genes may serve as a starting point for unraveling genetic variants controlling risk for NSOC. Recently, the interferon regulatory factor-6 (IRF6) gene on 1q32.2 has been shown to harbor mutations responsible for the majority of persons with van der Woude and popliteal pterygium syndromes (Kondo et al., 2002). Moreover, studies investigating the role of IRF6 in the occurrence of isolated clefts (apparently non-syndromic cases) showed strong evidence of over-transmission of the C allele at an intragenic polymorphism rs2235371 (which results from a C→T substitution of nucleotide 820 in the coding region, creating a valine→isoleucine substitution at amino acid 274 in the protein-binding domain of IRF6, also denoted V274I) in Asian and South American populations. Up to 12% of non-syndromic cleft lips with/without cleft palate can be attributed to variants in IRF6, and the recurrence risk in families with one affected child tripled if a variant of IRF6 was segregated in the family (Zucchero et al., 2004). Analysis of 7 other single-nucleotide polymorphisms (SNPs) in and around IRF6 has shown several distinct haplotypes demonstrating altered transmission in Iowan and Danish trios (Zucchero et al., 2004). Confirmatory studies in Italian, European-American, Belgian, Taiwanese, Singaporean, and Korean families with cleft lip with or without cleft palate (CL/P), respectively, have strengthened the evidence that IRF6 is important in the etiology of non-syndromic oral clefts (Zucchero et al., 2004; Blanton et al., 2005; Ghassibé et al., 2005; Scapoli et al., 2005; Srichomthong et al., 2005; Park et al., 2007; Jugessur et al., 2008).

Risk of CL/P associated with particular variants in IRF6 may differ among ethnic groups, however. Up to now, no large studies of this association have been conducted in a Chinese population, especially in western China. Here we tested three SNPs (rs2235371, rs2013162, and rs2235375) in the IRF6 gene using 206 unrelated NSOC case-parent trios from west China to confirm the role of IRF6 in the etiology of non-syndromic oral clefts.

MATERIALS & METHODS

Family Assessment

Our study sample consisted of 332 persons with NSOC, their parents (289 mothers and 243 fathers, including 206 complete trios), and 174 control individuals (having no congenital malformation of the body, no family history of genetic disease, being born in the same region, and, as much as possible, with the same gender distribution as in the affected persons). All participants were recruited between 2004 and 2006 from the Surgery Department of the West China Stomatology College of Sichuan University. Informed consent was obtained from each participant prior to enrolling in the study, as approved by the local ethics committee. Among the affected persons, 320 were sporadic (i.e., clefts with no family history of oral clefts), whereas 12 had one or more relatives with a cleft and were therefore classified as ‘familial’. To assess the non-syndromic status of affected persons, a physician screened all participants for the presence of associated anomalies or syndromes, and only those determined to have an isolated cleft were included in this current study. Participants were asked about a history of oral clefts among first and second relatives. All participants were self-identified as western Chinese. Gender, type of cleft, laterality, and severity among affected persons are shown in Appendix Table 1.

Table 1.

SNP Information, Primers, and Incision Enzymes

SNP i.d.	Position	Allelesa	Primer	Restriction Enzyme	Fragments (bp)
a Common allele listed first, in bold.
rs2013162	in 5 extron	C/A	F 5′CCCTGGGATGAGAAGGATAA 3′	DdeI (TC^TC)	CC 264 AA 202,62
			R 5′ACCTCTGACTCCCACTTGCT 3′		CA 264,202,62
rs2235375	6,7 splice	C/G	F 5′ ACAGGAAAGAGTCTATAA	TaqI(T^CGA)	CC 151 GG 118,33
			TAGAAGCAGAAGATC 3′		CG 151,118,33
			R5′ CCCAAAACTGAACCCCTGGAGAT 3′
rs2235371	in 7 extron	C/T	F 5′ GAGTCACAGGGATGAACAGG 3′	DpnII (^GATC)	CC 263 TT 217,46
			R 5′ GCTTCTGCTTCTCATTGGTA 3′		CT 263,217,46

Genotyping

The three single-nucleotide polymorphisms (SNPs) evaluated in this study were rs2013162, rs2235375, and rs2235371 in the IRF6 gene (Table 1). The V274I SNP was run because it is very polymorphic in the Asian population. The other two SNPs, rs2013162 and rs2235375, in the IRF6 gene were used in two previous studies, which allowed for an easy comparison of the results. Venous blood samples (infant umbilical cord blood for control individuals) were drawn from all participants, and DNA was extracted by a phenol chloroform extraction protocol and amplified by polymerase chain-reaction (PCR) with primers as shown in Table 1. Genotyping for the 3 SNPs was carried out by restriction digestion of the PCR products with DdeI, ApolI, and DpnII incision enzymes, respectively (Figs. A, B, C). We set 3 to 5 samples as positive controls in each of the reaction groups. The positive controls could be digested by restriction endonuclease and were identified by direct sequencing. If some samples were found to be different, we repeated the experiment. We also selected 10% samples at random to repeat the experiment. The results were coincident with the prior genotyping.

Figure.

Restriction fragment length polymorphism patterns for IRF6 rs2013162 (A), rs2235375 (B), and rs2235371 (C). (A) Lane M is the 50-bp DNA marker (pUC 19 DNA/MspI), with the arrowhead denoting the 250-bp band. Lanes 1 and 6 represent individuals who are homozygous AA, 3 and 4 represent homozygous CC, and 2 and 5 represent heterozygous CA, respectively. (B) Lane M is the 50-bp DNA marker, with the arrowhead denoting the 150-bp band. Lane 1 represents individuals who are homozygous C, 3 represents homozygous GG, and 2 and 4 represent heterozygous CG, respectively. (C) Lane M is the 50-bp DNA marker, with the arrowhead denoting the 250-bp band. Lane 2 represents individuals who are homozygous TT, 1 and 4 represent homozygous CC, and 3 represents heterozygous CT, respectively.

Statistical Analysis

Hardy-Weinberg equilibrium (HWE) was assessed for all polymorphisms in parents of affected persons and the control group. Case-control statistical analysis was performed with the SPSS 11.0 statistical software package (SPSS Inc., Chicago, IL, USA). Comparison of genotype and allele frequencies among case groups (CL, CP, and CL/P cases) and the control group was analyzed by the χ² test. A transmission disequilibrium test (TDT) was used to test for excess transmission of the target alleles in these case-parent trios. Pairwise linkage disequilibrium (LD) was computed as both D′ and r² for all SNPs, by use of the Haploview program (http://www.broad.mit.edu/mpg/haploview/index.php/). The Family Based Association Test (FBAT) package (http://www.biostat.harvard.edu/fbat/default.html) was used to test the null hypothesis that transmission of alleles from parents to child is independent of phenotype, which is equivalent to testing the composite H_o: no linkage or no linkage disequilibrium (LD) (Horvath et al., 2001). Additionally, haplotype transmission disequilibrium was calculated with the same FBAT software package.

RESULTS

Hardy-Weinberg equilibrium was assessed for the 3 SNPs among parents of all cases combined and among control infants as samples of the Western Chinese population. Significant deviation from Hardy-Weinberg expectations could reflect genotyping errors or true heterogeneity in the general population, and could bias our statistical tests. All χ² tests using genotype frequencies of parents of affected children and control individuals showed that these unrelated individuals conformed to Hardy-Weinberg Equilibrium (Appendix Tables 2, 3, 4).

Observed allele frequencies and their resulting genotypes among affected children, their mothers, their fathers, and control individuals are presented in Appendix Tables 2, 3, and 4. The distributions of genotypes and alleles among affected persons and their parents were compared with those among control individuals: At rs2013162, no difference was found (P = 0.897, P = 0.782). At rs2235375 and rs2235371, significant differences for both genotypic (P = 0.001, P = 0.000) and allelic (P = 0.000, P = 0.000) distributions between affected persons and control individuals were found, but not for the ‘cleft palate alone’ group. There were no differences in the distributions between the affected persons’ parents and control individuals at all 3 SNPs. We computed odds ratios of an individual being affected, given the common allele genotype, by constructing 2x2 tables comparing affected persons with control individuals, while combining the other two genotypes.

TDT analysis was carried out on participants with heterozygous-informative parents (Table 2). At rs2013162, this allelic TDT showed no evidence for association of CL/P. At rs2235375, highly significant evidence of linkage in the presence of disequilibrium (P < 0.01) was seen for CL/P. At rs235371, highly significant evidence of linkage in the presence of disequilibrium (P < 0.05) was also seen for CL, CLP, and CL/P, but not among trios with a cleft palate case only (CP) (Appendix Tables 5, 6, 7).

Table 2.

Transmission Disequilibrium Test (TDT) Summary

	SNP	Allelea	A. Frequencyb	T/NTc	Pd
a Common alleles are in bold.
b Common allele frequency was calculated based on all normal controls’ chromosomes.
c Transmission/non-transmission of the common allele counts from heterozygous parents.
d P-value for TDT.
1	rs2013162	C/A	54.3%	177/184	0.705
2	rs2235375	C/G	57.6%	146/213	0.000
3	rs2235371	C/T	55.5%	202/114	0.013

FBAT analysis also showed stronger associations between the rs2235375 and rs2235371 markers and NSOC and confirmed the allelic TDT results. These statistics were also significant in the combined group of all cleft trios. However, no significant association was obtained when we genotyped the marker rs2013162.

To increase information, we also conducted haplotype-based transmission disequilibrium analyses using genotype data. The haplotype was made of 2 and 3 alleles at each SNP. Significant results of the FBAT examination of the haplotypes are presented in Table 3.

Table 3.

Haplotypes Showing Transmission Distortion

	Haplotype	H. Frequency	Z	P
P-value for transmission distortion.
Order of SNPs: rs2013162-rs2235375-rs2235371 (V274I).
X indicates that the SNP was not included in creating the haplotype.
H. frequency: Haplotype frequency.
Z: a vector of the Family Based Association Test (FBAT) statistic.
1	C G X	0.318	3.204	0.001356
2	C C X	0.242	−3.326	0.000880
3	C X T	0.113	−2.600	0.009333
4	C G C	0.278	3.468	0.000525
5	C C T	0.077	−3.007	0.002636

Linkage disequilibrium between markers was calculated by D′ and r² (Appendix Table 8). In our study population, linkage disequilibrium between the two markers was less significant than in the Italian and Filipino studies.

DISCUSSION

In previous studies (Zucchero et al., 2004; Srichomthong et al., 2005), a significant association was reported between NSOC and the rs235371 polymorphism in IRF6. In particular, the C allele was over-transmitted to persons with CL/P from Asian and South American populations. In the current study, we used the case-control, TDT, and carried out family-based association studies to replicate these results in our NSOC cohort collected in Western China. Our results were significant in rejecting the null hypothesis of no linkage or no association between this marker and CL, CLP, and CL/P, but not for CP alone.

The common “C” variant has been identified as a risk factor in Asian or South American populations, but the evidence was less strong in European populations (Zucchero et al., 2004). However, previous articles discussed not only the rs235371 variant, but also the more general genetic background of IRF6, including an extensive haplotype analysis of an additional 2–35 SNPs in the IRF6 gene. Most importantly, this haplotype-based analysis has recently been confirmed in Italian, European-American, and Belgian CL/P families (Blanton et al., 2005; Ghassibé et al., 2005; Scapoli et al., 2005) and in Asian populations (Park et al., 2007). Isolated forms of CL/P have a complex etiology, with multiple genetic and environmental factors contributing (Beaty et al., 1997). Whether these SNPs are causative, or whether they are only in linkage disequilibrium with the disease-causing variants is yet to be elucidated. Significant results observed from SNPs other than p.V274I (rs2235371) suggested that V274I itself is not causal, but rather in LD with some causal mutation in IRF6. Thus, the rs235371 variant itself is not directly useful for genetic counseling in individual families (Chakravarti, 2004). This will await more specific identification of the risk allele within those haplotypes conferring greatest risk, which may differ among different populations.

In our study, both the C/C and T/T homozygotes at rs2235371 showed significant differences from the control group, with more C/C genotypes and fewer T/T genotypes among affected persons (P < 0.005). (Note that, in our study, all alleles are reported for the forward strand of the chromosome.) This pattern is consistent with a recessive effect of the C allele. Such a recessive effect was further confirmed by the observation of a significant difference in the frequency distribution of genotypes among probands compared with unaffected persons (P < 0.001); the frequency of the C/C genotype was increased, and the frequencies of T/T genotypes were reduced among probands as compared with unaffected control individuals.

In this study, we analyzed the other two SNPs, rs2235375 and rs2013162, in the IRF6 locus that had earlier shown, in TDT analysis, the strongest association with NSOC in European-derived populations (in Italian, American [Texas], and Belgian populations), and in Asian Chinese populations (Taiwan and Singapore; Park et al., 2007), respectively. We also detected significant under-transmission of the C common allele and significant over-transmission of the G allele for marker rs2235375 (P = 0.002 and P = 0.002, respectively). But, interestingly, for the rs2013162 marker, no significant association was obtained in our samples, which differed greatly from previous findings (Blanton et al., 2005; Ghassibé et al., 2005; Scapoli et al., 2005; Park et al., 2007). This is not surprising, given that two different ancestral populations were studied.

However, in haplotype analyses of NSOC triads, we observed, among the most significant results of the haplotypes, that the frequent C alleles for marker rs2013162 were always found to be carried. The most significant results were obtained for the C-G-C (order: rs2013162-rs2235375-rs2235371) haplotype (P = 0.000525). All haplotypes carrying the C alleles at marker rs2235375 and T alleles at rs2235371 (C-C, C-T, C –C-T haplotypes) showed significant under-transmission.

Another implication of the IRF6 polymorphism is that persons with isolated cleft lip or palate should no longer be considered a monolithic group, but, in future studies, should be stratified according to their individual genotypes. It is also imperative to examine environmental exposures in women who give birth to a child with CL/P while stratifying by their IRF6 genotypes to increase the statistical power of the study. It is possible that environmental exposure will have a greater effect in some pregnancies (e.g., those in which the fetus has the C/C genotype at rs2235371) than in others, resulting in an increased chance of expression in fetuses with specific high-risk genotypes. If environmental factors do interact with the IRF6 pathway, it may become possible to identify the mechanism of action of specific environmental triggers.

In summary, our findings confirm that the IRF6 gene is associated with increased risk of NSOC. This study further confirms the impact of IRF6 on the risk for cleft lip/palate in a Western China population, and supports findings from South American, European, and Asian populations. Replication of positive findings in complex diseases has been difficult, but since seven different studies have yielded significant results of an association between IRF6 and NSOC, further mechanistic studies are warranted.

This is the first study to confirm the impact of IRF6 on the risk for cleft lip/palate in a Western China population.