Cleft lip with or without cleft palate (CL/P) is a congenital malformation with complex aetiology. In addition to genetic predisposition, environmental risk factors may play an important role in the pathogenesis of this birth defect. Among genetic causes, a number of CL/P susceptibility genes can be found among components that control the folate plasma concentration, since folic acid supplementation in early pregnancy has been seen to have a protective effect against CL/P development (van Rooij et al. 2004). Recent case-control studies carried out by our group on a variant form of methylenetetrahydrofolate reductase (MTHFR) (C677T), indicated it as a risk factor for CL/P (Martinelli et al. 2001; Pezzetti et al. 2004). Such findings have prompted the search for additional cleft-associated variants in folate-metabolism enzymes. Thus, our group investigated, by linkage disequilibrium, folate receptor genes that mediate the delivery of 5-methyltetrahydrofolate to the interior, exterior, within or between cells (Scapoli et al. 2005). Highly negative multipoint LOD scores were obtained in linkage analysis using two microsatellites mapping on the cluster of FOLR1 and FOLR2, indicating that a large role in clefting for variation in these folate receptors is unlikely.
More recently, again our group tested the hypothesis of an involvement in CL/P aetiology by an additional four genes belonging to the folate pathway: transcobalamins (TCN1 and TCN2), methionine synthase (MTR), and MTR reductase (MTRR) (Martinelli et al. 2006). We demonstrated that polymorphisms in TCN2 can influence the risk of developing CL/P. However, given the high clinical and genetic heterogeneity of the disease, we believe that other genes acting in the folate pathway could be involved in CL/P onset and deserve further investigation. Thus, on the basis of this data, we propose, in the present paper, the study of methylenetetrahydrofolate dehydrogenase 1 (MTHFD1), a gene that could represent an attractive candidate in CL/P aetiology, given its key role in folate metabolism.
MTHFD1 is a folate-dependent cytoplasmic, nicotinamide adenine dinucleotide phosphate-dependent, trifunctional enzyme comprising three activities: N5,N10-methylenetetra-hydrofolate dehydrogenase, N5,N10-methenyltetrahydro-folate cyclohydrolase, and N10-formyltetrahydrofolate synthetase (OMIM 172460). The enzymes catalyze sequential interconversion of tetrahydrofolate derivatives required for purine, methionine, and thymidylate synthesis.
To date, 1958A>G transition of MTHFD1 has been associated with maternal genetic risk of having a child with neural tube defects (NTDs) (Brody et al. 2002, Parle-McDermott et al. 2006), and with an increased risk for the children (De Marco et al. 2006), but in just one paper the authors investigated genotypes with the scope of revealing any effect on the risk of having a child with CL/P (Mostowska et al. 2006). However, Mostowska and colleagues could not find any association between maternal MTHFD1 A1958G genotypes and the risk of CL/P-affected, while they got positive evidence in analysing MTR, another gene involved in the folate metabolism that we had also previously investigated, obtaining no evidence of its involvement in CL/P (Martinelli et al. 2006).
Due to the key role of the MTHFD1 gene in such an important metabolic pathway, our aim in this study was to verify its possible contribution in determining cleft among a large sample study of CL/P familial and sporadic patients and their parents.
For this investigation we used a case-parent triad design. The sample study consists of 216 CL/P patients and their parents (212 mothers and 206 fathers), with Italian ancestry. Among them, 120 cases have a positive familial history, while 96 are considered sporadic or non-familial because no other relative shares the same malformation. CL/P was the unique disorder affecting the probands. To assess the nonsyndromic CL/P status, the patients and relatives were asked specific questions about the presence in their family of other somatic and/or neurological disorders and the use in pregnancy of clefting drugs such as phenytoin, warfarin, and ethanol. After informed consent, a peripheral blood sample was drawn from each individual and DNA extracted from peripheral blood cells.
Two non-synonymous SNPs at the MTHFD1 gene were chosen for this investigation, rs1950902 (G401A, R134K), and rs2236225 (A1958G, R653Q), located on exon 6 and 20 respectively (ref. seq. NM_005956.2). R134K polymorphism was investigated by Alw26I (NEB, Hertfordshire, U.K.) restriction endonuclease digestion of PCR products previously obtained by using For 5′-AATGAAACAGTC-ATTGAGGTCAC-3′ Rev 5′-TGCATGCTTTCATTTA-TAATATGTTT-3′ primers, whereas R653Q was tested by the MspI (NEB) restriction enzyme and PCR reactions were carried out employing the primers For 5′- TACAAAC-CCTTCTGGGCCAAC-3′ and Rev 5′- CCAAATCAA-TTCCATACCGTTGA-3′.
Restriction fragments were separated by electrophoresis on 10% polyacrylamide gels and visualized by SybrGreen (SIGMA, Milan, Italy) staining. In order to increase throughput, up to three successive loadings of the same gel were performed.
Hardy-Weinberg equilibrium and allelic frequency comparisons among groups, i.e. fathers, mothers and patients, were performed using the on line tool for case-control studies at http://ihg.gsf.de/cgi-bin/hw/hwa1.pl. Single marker analyses were performed with the transmission disequilibrium test (TDT), which examines the transmission of alleles from heterozygous parents (Spielman et al. 1993). TDT and haplotype analysis were performed with the program TDTPHASE, part of the UNPHASED package version 2.403 (Dudbridge, 2003). Haplotype analysis was restricted to phase-certain haplotypes only, in a conditional logistic regression model. Linkage disequilibrium between markers were calculated with the D' and r2 statistics from parental haplotypes by use of the ldmax program from the GOLD package (Abecasis & Cookson, 2000).
Genotypes were obtained for 98.1% and 99.7% of samples, for G401A and A1958G respectively. A summary of observed genotypes and allele frequencies of analyzed genetic variants are summarized in Table 1. The Hardy–Weinberg equilibrium was observed in both case and parent groups. Linkage disequilibrium analyses revealed no association between CL/P and investigated polymorphisms alleles or haplotypes (Table 2). The Global P value for haplotype analysis was also not significant. On the other hand, a borderline undertransmission of the most frequent haplotype 401G/1958G was observed, that may indicate a possible protective role for this haplotype, (P value = .07). The linkage disequilibrium between markers was relatively low, (D' = .30; r2= .01). Such results agree with data available from the HAPMAP project (http://www.hapmap.org/).
In summary, our results among the Italian population did not support any association between the MTHFD1 polymorphisms and clefting, partially confirming the data reported by Mostowska and colleagues regarding the Polish population. Altogether, data from both studies do not support MTHFD1 involvement in CL/P onset among Europeans. Nevertheless, a borderline undertransmission of the most frequent haplotype was observed. This may represent an insignificant random deviation from the equilibrium or, alternatively, it may reflect a minor role for MTHFD1 in clefting. If this is the case, a much bigger sample size may be required to demonstrate such a small effect.