Journal of Oral Hygiene & Health
Open Access

NSOFC: Nonsyndromic Orofacial Cleft; CL(P): Cleft Lip with or without Palate; CP: Isolated Cleft Palate; IRF6: Interferon Regulatory Factor 6; GEIs: Gene-Environment Interactions; PCR: Polymerase Chain Reaction; TDT: Transmission Disequilibrium Test; FBAT: Family-Based Association Test

Mutations in interferon regulatory factor 6 (IRF6), located on 1q32.2, are responsible for the two autosomal dominant orofacial cleft syndromes, Van Der Woude and popliteal pterygium syndrome [1-3]. It was also the first identified nonsyndromic orofacial cleft (NSOFC) susceptibility locus [4] and has been the only candidate gene consistently found to have a significant association with NSOFC across multiple studies in many regions of the world, for example, China and Europe [5-7].

Blanton et al. [8] confirmed that the association between SNPs at IRF6 and NSOFC varied between ethnic groups and that there was a need for further evaluation of IRF6 variations across populations to better determine its role in NSOFC [8].

Gene-environment interactions (GEIs) have been suggested to play an important role in the etiology of NSOFC. Gene-environment interactions are defined as the co-participation of genetic and environmental risk factors in the same causal mechanism to promote disease development [9]. One of the important applications of GEI studies is to help public health researchers develop strategies for targeted intervention to allow risk-factor modification based on individual genetic profiles [10].

This study investigated whether (a) IRF6 was also associated with cleft lip with or without palate (CL(P)) and isolated cleft palate (CP) in Saudi Arabia and (b) there was any association between IRF6 polymorphisms (rs2013162, rs2235375, and rs2235371) and maternal environmental exposures, especially folic acid supplementation in the etiology of NSOFC.

Patients

This paper is a part of a series of studies aimed at determining the prevalence and investigating the etiology of NSOFC in Saudi Arabia. Three papers on the prevalence and environmental risk factors related to NSOFC have been published [11-13].

We obtained samples from 11 referral hospitals in three main cities in Saudi Arabia: Riyadh: King Saud Medical City Hospital, King Fahad Medical Cities, Riyadh Armed Hospital, and Riyadh National Guard Hospital; Jeddah: King Fahad Hospital, Abdulaziz University Hospital, Al-Messadia Maternity Hospital, Al-Azizia Maternity Hospital, King Abdulaziz Medical City, and King Fahad Armed Forces Hospital; and Medina: Medina Children and Maternity Hospital. The inclusion criteria for this study population included non-syndromic cleft lip with or without cleft palate or cleft palate children who were 18 months or younger and had been admitted to neonatal, plastic surgery, or orthodontic units, while the controls were healthy, unaffected, age-and gender-matched children recruited from the vaccination clinic. A total of 171 cases and 189 controls (infant–parent triads) were recruited for the study during January 2010–December 31, 2011. Ethical approvals were obtained from King Abdulaziz University Hospital (359-10, 2010), the Ministry of Health (10-079, 2010), and the Institutional Research Review Boards (IRBs) of the military hospitals (429/2011). Parents consented to participate in the study before sample collection.

Methods

Saliva was sampled from both parents and infants using Oragene (500) (for adults) and (575) (for infants). DNA extraction was carried out using the QIAamp DNA Mini Kit (catalog# 51306; Qiagen, Canada) according to the manufacturer’s protocol, with modification of the starting quantity of saliva (500 instead of 200 μl). The purity and quantity of extracted DNA were measured using a Nanodrop 2000 spectrophotometer (Thermo Scientific, Delaware, USA). The DNA quality was also evaluated using agarose gel electrophoresis (1% in 1× TBE buffer).

We used this DNA for subsequent amplification with restriction digestion polymerase chain reaction (PCR), TaqMan® Genotyper Software, and TaqMan® SNP Genotyping Assays. Three polymorphisms at IRF6 were analyzed: rs2013162, rs2235375, and rs2235371.

To assess GEIs in the etiology of NSOFC, we hypothesized that environmental factors and maternal exposures strengthen or diminish the effect of maternal rare alleles, which in turn affect the risk of having an infant with NSOFC. These factors included maternal medication and supplementation, maternal diseases, maternal stress and parental smoking; we obtained this information through parental interviews. The type of oral cleft was not sub-grouped to CL/P and CP, as the number of cases was not large enough to subdivide or stratify them. Two study designs were used to measure GEI: case-only and case-control study designs.

Statistical analysis

A transmission disequilibrium test (TDT) was measured using the family-based association test (FBAT) and PLINK. PLINK was also used to measure the effect of parents of origin. We compared genotype and allele frequencies among CL(P) and CP cases with those for controls using the chi square test and adjustment with Bonferroni correlation (P ≤ 0.05) using SPSS. The degree of association between genotype and allele frequencies with NSOFC was estimated by measuring the odds ratio (OR) and respective 95% confidence intervals (95% CI) using an online program found at http://www.quantpsy.org/chisq/chisq.htm.

To assess GEIs, two study designs were used: case-only and casecontrol. In the case-only study design, interactions between maternal SNPs (genotypes and allele) and environmental factors were analyzed by measuring the distribution of maternal genotypes and alleles according to exposure/no-exposure to environmental factors among oral cleft cases. The common allele and null maternal exposure were set as references in our calculation. For the type of drinking water source, tap water with the common allele was set as the reference. Analysis was carried out using the chi square test with SPSS version 16 (SPSS Inc., Chicago, IL, USA), and a P-value of less than 0.05 was considered to indicate significance. Their degrees of association with NSOFC were estimated by measuring the OR and respective 95% CIs. If one cell contained a 0 value, only the P-value was calculated. No analyses were carried out if more than one cell contained a 0 value.

Multi-nominal logistic regression analysis was carried out using SPSS to overcome confounding factors. Factors that showed significant GEIs when analyzed alone in the previous analysis were entered in the logistic regression.

For the case-control design, we compared differences between the frequency of maternal genotypes (homozygous common allele, homozygous rare allele, and heterozygous allele) for different environmental factors using a chi-square test; we adjusted the P-value by Bonferroni correction methods between cases and controls using SPSS. We determined ORs and 95% CIs for factors with P-values ≤ 0.05 by using MedCalc (user-friendly statistical software). For the type of drinking water source, tap water with the common allele was set as a reference. If one cell contained a 0 value, only the P-value was calculated. No analyses were carried out if more than one cell contained a 0 value.

From the 171 cases and 189 controls, 10 NSOFC cases were not examined for sub-phenotype classification, and 16 control fathers refused to provide saliva samples. In addition, genotyping values were not obtained for some samples. This resulted in 120 CL(P) and 33 CP paternal genotypes compared to 158 controls, 127 CL(P) and 34 CP maternal genotypes compared to 185 controls, and 123 CL(P) and 32 CP infant genotypes compared to 178 controls, according to IRF6 rs2013162 genotype variance. For rs2235375, 122 CL(P) and 33 CP paternal genotypes compared to 169 controls, 127 CL(P) and 34 CP maternal genotypes compared to 189 controls, and 126 CL(P) and 35 CP infant genotypes compared to 188 controls were analyzed. For IRF6 rs2235371 genotype variance, 120 CL/P and 33 CP paternal genotype compared to 170 controls, 126 CL/P and 33 CP maternal genotypes compared to 189 controls, and 124 CL/P and 35 CP infant genotypes compared to 187 controls were analysed.

Almost all cases (99.4%) and controls (98.8%) were homozygous at IRF6 rs2235371 for the common CC allele. As the number of rare alleles was negligible TDT using FBAT and PLINK analyses were not carried out for this polymorphism.

Statistically significant over-transmission of the common C allele in CL(P) cases (P= 0.018) was found using FBAT and PLINK analyses. PLINK testing also showed significant reduction of CL(P) risk with the IRF6 rs2013162 rare A allele (P= 0.018, OR: 0.644) (Table 1). The IRF6 variant was over-transmitted from the paternal side but this finding was not statistically significant (P= 0.06), (Table 2).

Table 1: Transmission Disequilibrium test (TDT) results for IRF6 rs2013162 and rs2235375 variants among NSOFC infant-parental triads and its sub-phenotypes (CL(P) and CP) using FBAT analysis.

Table 2: Testing IRF6 rs2013162 andrs2235375 alleles for transmission disequilibrium and parent of origin using PLINK analysis for NSOFC infant-parental triads and its sub-phenotypes (CL(P) and CP).

There were no significant differences between cases and controls in the paternal IRF6 rs2013162 genotype (P= 0.08 for CL(P), P= 0.92 for CP), infant IRF6 rs2013162 genotype (P= 0.18 for CL(P) and P= 0.32 for CP), and maternal CP IRF6 rs2013162 genotype (P= 0.187) (Table 3). However, there were significant differences in maternal IRF6 rs2013162 genotype for CL(P) (P= 0.024) in cases compared to controls.

Table 3: Distribution of IRF6 rs2013162 and rs2235375 infant-parental triad genotypes according to NSOFC phenotypes (CL(P) and CP) and compared to controls. There were missing samples; 23 (8 cases and 15 controls) paternal samples, four maternal control samples and 17 infant samples (6 cases and 11 controls). The phenotype diagnosis of ten NSOFC cases are missing.

After chi-square adjustment and Bonferroni correlation were carried out, it was found that the maternal homozygous AA rare allele genotype was significantly more prevalent in controls than in CL(P) (P<0.05). Furthermore, the heterozygous CA genotype was significantly more prevalent in cases than in controls. IRF6 rs2235375 variants did not show any association with both CL(P) and CP infant– parent triads following either FBAT, PLINK or chi-square testing in the Saudi population.

Gene-environment interactions

The maternal IRF6 rs2013162 rare homozygous allele genotype (AA) and heterozygous genotype (CA) were compared to the common homozygous allele genotype (CC) in NSOFC cases for environmental factors (Table 4). Mothers who had a homozygous rare allele genotype (AA) and, during their first trimester, had used antipyretics (P= 0.012, OR: 9, 95% CI: 1.62–49.91), experienced pain (P=0.013, OR: 6.57; 95% CI: 1.49–29.01), suffered from depression (P=0.048, OR: 8.11; 95% CI: 1.01–64.84) or who, during the pregestation period, had taken folic acid supplements (P=0.047, OR: 8.22, 95% CI: 1.03– 65.72), were more likely to have an offspring with NSOFC.

Table 4: Relationship between maternal IRF6 (rs2013162 and rs2235375) genotypes and different environmental factors, including: maternal illness, medication, stress, supplements, paternal smoking and maternal domestic exposure using case-only study design.

The maternal IRF6 rs2013162 rare A allele was compared to the common C allele in NSOFC cases in relation to different environmental factors (Table 5). Mothers who had a rare A allele and were using antipyretics in the first trimester (OR: 2.34; 95% CI: 1.12–4.9) or folic acid supplements during the pregestation period (OR: 2.57 and 95% CI: 1.23–5.37) were significantly more likely to have infants with NSOFC. Mothers who had a rare A allele and experienced fever during the 3-month pregestation period (OR: 0.28, 95% CI: 0.09–0.81), underwent X-ray (OR: 0.19, 95% CI: 0.07–0.5) or ingested iron (OR: 0.57, 95% CI: 0.33–0.97) during the first trimester, or who drank tap water as their main drinking water source rather than bottled water (P= 0.043, OR: 0.51, 95% CI: 0.26–0.98), were significantly less likely to have infants with NSOFC.

Table 5: Relationship between maternal IRF6 (rs2013162 and rs2235375) allele frequency and different environmental factors, including: maternal illness, medication, stress, supplements, paternal smoking and maternal domestic exposure using case-only study design.

Mothers who had a rare homozygous allele genotype (AA) for rs2235375, were using folic acid supplements during the pregestation period (P=0.045, OR: 0.23, 95% CI: 0.05–0.97), or complained of stress (P=0.016, OR: 0.3, 95% CI: 0.1–0.8) were significantly less likely to have infants with NSOFC.

For the case-control study approach for the three maternal IRF6 rs2013162 genotypes (Supplementary table S1), statistically significant case and control difference in the maternal homozygous rare allele genotype (AA) frequency for mothers using folic acid supplementation was found. Significantly more control mothers used folic acid supplements in the first trimester compared to mothers with NSOFC.

We found significantly higher maternal homozygous common allele genotype (CC) rs2235375 frequencies in cases compared to controls in maternal illness in the pre-gestation and first trimester periods, flu/ common cold infection in the pre-gestation period, fever in the pregestation period, and for maternal folic acid supplementation.

Multinominal logistic regression analysis was carried out to identify significant GEI among NSOFC cases. The homozygous common allele genotype for each SNP included was set as a reference for analysis (Table 6). For maternal rs2013162, the homozygous rare allele genotype (AA) was significantly related to antipyretics in the first trimester (P=0.027, OR: 10.18, 95% CI: 1.31–79.1) and abdominal pain in the first trimester (P=0.031, OR: 7.4, 95% CI: 1.2–45.51) among NSOFC cases. The heterozygous allele genotype (AC) was significantly related to pre-gestation folic acid supplementation (P=0.0167, OR: 6.78, 95% CI: 1.41–33.49) and occurrence of fever pre-gestation (P=0.025, OR: 0.23 and 95% CI: 0.06–0.83) among NSOFC cases.

Table 6: Multi-nominal logistic regression analysis for case-only study design gene-environmental interaction including significant factors.

We found a significant over transmission of the common C allele at IRF6 rs2013162 among CL(P) cases and suggest that GEIs may be associated with IRF6 polymorphisms in the Saudi Arabian population.

The first SNP selected was IRF6 rs2013162. FBAT showed significant over-transmission of the common C allele in CL(P) cases (P=0.018). In addition, PLINK analysis showed a protective effect and an association between the rare A allele and CL(P) (P=0.018 and OR: 0.644 for CL(P)). A similar finding was reported by Scapoli et al. [14], who detected an over-transmission of the common allele for rs2013162 (P=0.004) and all haplotypes carrying these common alleles among 219 Italian CL(P) trios. This is also similar to the finding reported by Park et al. [15] and Blanton et al. [16]. In addition, the rs2013162 rare allele exhibited a genome-wide significant relationship with CL(P) in a study by Beaty et al. [17]. Stratification of the sample population using Asian ancestry as a factor yielded stronger evidence of association with IRF6 rs62013162 [17].

In our study, PLINK analysis showed transmission of the IRF6 rs2013162 variants from the paternal side (P=0.06). Although the relationship was not significant, it agrees with the review by Anderson et al. [18] on male-mediated developmental toxicity; it was suggested mutated DNA were more frequently inherited from fathers than mothers. Brinkworth [19] suggested that the paternal genetic inherence effect occurs through genomic instability or apoptosis suppression of the germ cell. In addition, our study found that maternal gene variants showed significant differences between cases and controls, with the more heterozygous allele genotype in CL(P) cases and the more homozygous rare allele genotype in controls (P=0.024). This suggests a maternal rare allele effect that was not previously described. Ludwig et al. [20], in their Weinberg’s log-linear model analysis on IRF6 gene variants in Central European patients, found no difference in the risk for CL(P) between maternal- and paternal-derived alleles.

The other two IRF6 SNPs that were analyzed in this study were rs2235375 and rs2235371. Huang et al. [21] in Western China reported an association between the transmission of rs2235375 and rs2235371 (C/T) markers and CL(P). However, our research did not indicate any association between the transmission of rs2235375 and NSOFC. This agrees with a case-control study in a Mexican population where there was no significant difference between CL(P) cases and controls in the frequency of rs2235375 (P=0.08) [22]. Also a case-control triad study by Zhou et al. [23], in a Chinese population, found no significant association between rs2235371 and CL(P) using FBAT analysis.

As the frequency of the rs2235371 rare allele was very low in our study, it was not possible to carry out any analysis. The Ensemble Genome project report stated that the prevalence of the rare T allele in rs2235371 was calculated to be only 2% in a European population compared to 41% in an Asian population [24]. Similarly, in their fourpopulation case-parent trios, Park et al. [15] found that the rare allele frequency of rs2235371 in European Americans was too low to be reported.

Our study did not find any relationship between IRF6 and CP. This finding is supported by other studies that investigated the relationship between NSOFC and IRF6 and included CP cases in their sample [6,14,21].

Gene environment interaction

Only a few studies have considered the role of GEIs in the etiology of NSOFC. The factors that were studied were mainly folic acid and vitamin supplements, smoking, and maternal passive smoking. Furthermore, only a small number of gene variants were analyzed [25-28]. In our study, we focused on the interaction between environmental factors and maternal genes to assess any direct maternal effect occurring during pregnancy.

One of the concerns in GEI studies is the study power. False-positive and false-negative outcomes were reported in studies with small sizes [29]. To improve the reliability of our results, two study designs and three statistical analyses were carried out for GEI. Consistency of association between the findings of different study designs in GEI analyses reinforces the potential significance of these factors.

The environmental risk factor that showed a significant interaction with maternal IRF6 rs2013162 in both study designs and statistical approaches was maternal folic acid supplementation during the pregestation period. This significant association concurs with a population based case-control study in Northern Netherlands that reported “duration of exposure-response effect” increased the risk of cleft lip [30].

In contrast, in the case-control study design, maternal folic acid in the first trimester were significantly more ingested by the controls with homozygous rare AA allele genotype mothers than cases (P= 0.003), indicating a joint protective effect.

Other environmental risk factors were either significant in the caseonly study design or in the case-control study designs. In the case-only study design, antipyretic medication and fever showed GEI in both analyses (genotypes and allele analysis). This highlights the importance of future studies to verify the influence of maternal disease symptoms on the function of maternal genes and their effect on the embryonic development.

Moreover, in the case-control study, mothers who were homozygous for the common allele (CC), and who consumed antibiotics during the pregestation period or were exposed to illness and common cold/ flu were significantly more likely to have an infant with NSOFC. In addition, paternal waterpipe smoking associated with maternal common allele (CC) is more likely prone to have an infant with NSOFC. This indicates a synergic effect between environmental risk factors and the gene variant that when isolated were found to be associated with an increased risk of oral clefts.

Maternal stress was an environmental risk factor that showed a significant interaction with maternal IRF6 rs2235375 in both study designs and all GEI statistical approaches. However, the pregestational use of folic acid supplementation and presence of maternal IRF6 rs2235375 were found to be associated with NSOFC in the case-only study design. This finding was not supported by the Velázquez-Aragón et al. [22] Mexican study. However, these differences may be related to differences in the study setting.

In addition, although association with the multivitamin supplementation interaction with IRF6 rs2235375 was significant in the case-control analysis (P= 0.026) it was not supported by the Wu et al. [28] population-based study that was carried out in China. In addition, the number of subjects exposed to multivitamins in our study was small and a significant relationship was only found in one analysis and in the pregestation period.

To draw conclusions regarding definitive GEIs, we need to overcome the limitation of the study sample size. In addition, a GWAS study using a log-linear modeling approach is necessary. However, achieving this number is difficult, particularly because NSOFC is a rare disease and sub-phenotyping is necessary. In addition, there are multiple risk factors that contribute to these diseases [9,30].

There is a great value in preliminary studies that may be underpowered for definitive findings, but that act as instruments for preliminary description of GEI and for generating hypotheses that can then be further tested in studies with adequate power [29]. Accordingly, we expect that our findings will play a role in directing future research to identify possible gene-environment risk factors that may allow prevention of NSOFC through public health strategies.

The differences between our findings and those of previous studies could indicate that the Saudi population has a different genetic etiology of NSOFC compared to other populations worldwide. TDT and PLINK analyses showed that the IRF6 rs2013162 rare A allele was not a marker for CL(P) risk as the common C allele showed significant transmission in CL(P) cases. Maternal exposure to antipyretics, folic acid, fever, antibiotics, illnesses, common cold/flu paternal waterpipe smoking, stress, and x-rays and/or chemicals could significantly interact with the maternal IRF6 (rs2013162 and rs2235375) gene variants, affecting the risk of having a child with oral cleft. Because of the genetic heterogeneity of OFC and the evidence for population specific genetic etiology, adequately powered GWAS studies combined with further investigation of environmental factors and GEI are recommended.

This project was funded by the Deanship of Scientific Research, King Abdulaziz University, Jeddah (Grant No. 4/165/1431); and was conducted with the support of the University of Dundee Dental School World Health Organization Collaborating Centre for Oral Health and Craniofacial Anomalies. The authors also acknowledge the support of Princess Al-Jawhara Albrahim of Excellence for Hereditary Disorders (PACER-HD) where the laboratory experiments of this project were carried out.

The authors thank the research committees of the Ministry of Health in Riyadh, Jeddah, and Medina, the research committees and of King Saud Medical City, Riyadh National Guard Hospital, King Fahad Medical City, King Fahad Armed Hospital, and King Abdulaziz Medical City, and Dr Hassan Al-Naeem at King Fahad Hospital; Zamzam Ebrahim Al-Hakami and Nouf Al-Beshri at Al-Messadia Maternity Hospital; Dr Safinaz Salamah and Ebtisam Hussain at Al-Azizia Maternity Hospital; Mervat Ali Sayed and all nurses at King Abdulaziz University Hospital; Dr Mosleh Saad Alharbi, Dania Baeasa, and Dr Mamoon Daghestani at King Abdulaziz Medical City; and Dr Manal Al-Malik, Dr Fawzia Sabbagh, Dr Mawahib Abuauf, and Mariam Malope at King Fahad Armed Forces Hospital. Dr. Wamda Helal, Dr. Ahmed Mustafa Hamdan, and Dr. Bassem Mohamad Gesrha at King Saud Medical City.