Transforming growth factor alpha (TGFA) is a well-characterized mammalian growth factor. Since the first report of an association between DNA sequence variants at the TGFA genetic locus and nonsyndromic oral clefts, 47 studies have been carried out, producing conflicting results. In this review, the author synthesizes findings from published reports on the association between the TGFA gene and clefting in humans. Bias, lack of statistical power, and genuine population diversity can explain the diverse results. In the aggregate, TGFA is probably a genetic modifier of clefting in humans, which is consistent with the oligogenic model suggested for nonsyndromic oral clefts.

Keywords: cleft lip; cleft palate; epidemiology; genetics; TGFA; transforming growth factor alpha
Abbreviations: CI, confidence interval; FGFR1, fibroblast growth factor receptor 1; IRF6, interferon regulatory factor 6; LOD, logarithm of the odds; MSX1, muscle segment homeobox 1; MTHFR, 5,10-methylenetetrahydrofolate reductase; PAX9, paired box 9; PCR, polymerase chain reaction; TGFA, transforming growth factor alpha; TGFB3, transforming growth factor beta 3

Transforming growth factor alpha (TGFA) is a well-characterized mammalian growth factor. It has been mapped to chromosome 2p13 (1, 2), comprises 80 kilobases of genomic DNA, and consists of six exons (sizes: exon 1, 40 base pairs; exon 2, 57 base pairs; exon 3, 118 base pairs; exon 4, 150 base pairs; exon 5, 110 base pairs) (Figure 1).

click to view

Figure 1: Genomic structure of transforming growth factor alpha (TGFA). Based on the paper by Vieira et al. (114), with corresponding GeneBank entry AH013033. Black boxes are coding regions; white boxes are untranslated regions. Numbers indicate exons. Arrows and letters indicate the locations of the most-studied TGFA variants: (a) TaqI; (b) RsaI; (c) C3296T (C-to-T substitution at nucleotide 3296); (d, e, f) primer K; (f) C3827T (C-to-T substitution at nucleotide 3827); (g) primer P; and (h) BamHI.

Expression of the TGFA gene occurs in a wide spectrum of normal tissue from the preimplantation period in mouse embryos to adult life (3-8). During craniofacial development, TGFA is expressed at the medial edge epithelium of fusing palatal shelves (9, 10). In palatal cultures, TGFA promotes synthesis of extracellular matrix and mesenchymal cell migration, thereby ensuring the strength of the fused palate (11).

Although Tgfa is expressed in mice during palatogenesis, mice with a null mutation of the Tgfa gene have abnormal skin, hair, and eyes but do not have oral clefts (12, 13). Newborn epidermal growth factor receptor-negative/-negative mice have a high incidence of cleft palate, and this may explain the genetic correlation of human oral clefts with polymorphisms in TGFA (14), given that TGFA is a likely ligand for epidermal growth factor receptor. Cleft lip with or without cleft palate was first associated with polymorphisms in TGFA in 1989 (15), and the topic was reviewed in 1997 (16), 2001 (17, 18), and 2002 (19-21).

An extensive Human Genome Epidemiology (HuGE) review of the gene variants for TGFA was performed. Medline, PubMed, and EMBASE were searched using the keywords "transforming growth factor alpha" and "TGFA." Additional search words included "oral clefts," "cleft lip and palate," and "orofacial clefts." Reference lists from published articles were also reviewed, and journals related specifically to birth defects and clefting were searched by hand. The review included papers written in English, French, Spanish, and Portuguese, as well as Chinese or Japanese papers with English abstracts, that were published between 1986 and 2005.

Currently, 356 single nucleotide polymorphisms and 20 insertion/deletion polymorphisms can be found in the May 2004 human genome assembly freeze in the University of California, Santa Cruz genome browser (http://www.genome.ucsc.edu/) . Table 1 and Figure 1 present the variants found to be most studied for oral clefts. There is a lack of information regarding the potential function of the variants presented in Table 1. The only consideration of this issue was in an Iowa study of cases of cleft palate only and the primer K variant (22). The authors located this variant in the 3'-untranslated region of the gene (Figure 1), within the same region as a transcribed 350-nucleotide polyadenylated, antisense mRNA species (23). If the antisense mRNA regulates the expression of TGFA by interacting with the K region of TGFA, the primer K variant may contribute to the cleft of the palate.

click to view

Table 1: Variants of the transforming growth factor alpha (TGFA) gene

The clinical studies reviewed were either case-control or family-based in design. The TaqI variant, first reported in 1987 (24) and characterized as a four-base-pair deletion in intron 5, is the one most studied in case-control studies of cleft lip and palate. In addition, the BamHI and RsaI variants (25) and the single nucleotide polymorphisms C3296T (rs2166975; a C-to-T substitution at nucleotide 3296) and C3827T (rs1058213; a C-to-T substitution at nucleotide 3827) (26) have also been studied. The published genotype frequencies for the BamHI variant are shown in Table 2, and those for the RsaI variant are shown in Table 3. Table 4 presents genotype frequencies for the TaqI variant from published studies and includes the frequencies of the populations in the Human Genome Diversity Cell Line Panel (27). The Diversity Cell Line Panel is a resource of 1,064 cultured lymphoblastic cell lines obtained from persons in different world populations. Lymphoblastic cell lines were collected from various laboratories by the Human Genome Diversity Project and the Fondation Jean Dausset-CEPH [Centre d'Etude du Polymorphisme Humain] to obtain unlimited supplies of DNA for studies of sequence diversity and history in modern human populations. Each cell line comes from a single individual. Samples were originally collected under nonrandom selection. The panel contains lymphoblastic cell lines from human populations on all continents. The maximum number of lymphoblastic cell lines from a population is 51. Twenty-five to 49 lymphoblastic cell lines are available from each of 21 population samples. Fourteen Chinese minority groups are represented by only 9-10 lymphoblastic cell lines each.

click to view

Table 2: Distribution of transforming growth factor alpha (TGFA) BamHI alleles*

click to view

Table 3: Distribution of transforming growth factor alpha (TGFA) RsaI alleles*

click to view

Table 4: Worldwide distribution of transforming growth factor alpha (TGFA) TaqI alleles*

Table 5 presents the genotype frequencies of the C3296T and C3827T variants.

click to view

Table 5: Worldwide distribution of the transforming growth factor alpha (TGFA) C3296T (C-to-T substitution at nucleotide 3296) and C3827T (C-to-T substitution at nucleotide 3827) alleles*

Most of the studies from which data were abstracted and are presented in Tables 2-5 were not population-based, with the exception of studies from Denmark, France, Norway, and the US states of California and Maryland. The Filipino data shown in Table 4 were from a hospital-based study. The frequencies presented in Tables 2-5 come from controls/unaffected persons.

Most of the studies described in Tables 2-5, as well as most of the populations from the Diversity Cell Line Panel, were small, as evidenced by wide 95 percent confidence intervals for the frequency of the least common genotype. In many instances in Tables 2-5, the data do not suggest Hardy-Weinberg equilibrium. The studies in Hardy-Weinberg disequilibrium either used a convenient sample as controls or obtained data from a very small sample. In Table 4 and Table 5, many populations from the Diversity Cell Line Panel were in Hardy-Weinberg disequilibrium, probably because sample sizes were very small. Genotyping error also cannot be discounted.

For Tables 2-5, in the case of overlap between studies, the report with the most thorough description was used to abstract genotype frequency data. The few inconsistencies between data presented in Tables 2-5 and data presented in later tables are due to small differences between the reported total numbers and the actual genotype information available.

There is a wide range in the TGFA TaqI, C3296T, and C3827T (Table 4 and Table 5) allele frequencies across different studies. Some populations show a remarkably high frequency of the TGFA TaqI C2 (rare) allele, including Biaka Pygmies, Chinese Han, Danish, Japanese, and Filipinos. For the C3296T and C3827T alleles, the Melanesians and Papuans have the T allele for both loci as the most common one. The C allele is the most common for both variants in all other populations studied.

As ancestral haplotypes propagate through a population, their physical length is reduced by recombination events. Thus, genotypes at nearby markers are not independent, and their association may reflect ancestral founding haplotypes. Most of the TGFA-cleft association studies relate to TaqI, BamHI, and RsaI polymorphisms, but there is no information on linkage disequilibrium for these three markers. Therefore, linkage disequilibrium between TaqI and BamHI, TaqI and RsaI, and BamHI and RsaI marker alleles was calculated from published haplotype data (15, 28, 29) (Table 6). TaqI and BamHI marker alleles are not in linkage disequilibrium. However, TaqI and RsaI marker alleles present borderline linkage disequilibrium, while BamHI and RsaI are strongly linked. Future studies should avoid generating data for both of the strongly linked variants. Combined genotypes of the TaqI and BamHI variants would provide the most informative data.

click to view

Table 6: Results of linkage disequilibrium analysis for the transforming growth factor alpha (TGFA) variant alleles TaqI, BamHI, and RsaI

Linkage disequilibrium analysis of the three variants is also reported for the Human Genome Diversity Cell Line Panel (Table 7). For this analysis, the populations were pooled by geographic origin, and the D' statistic was calculated using the software GOLD (30). This statistic measures the difference between the observed and expected (under independence) numbers of haplotypes bearing one marker allele and the other marker allele. D' depends strongly on marker allele and disease allele frequencies. Values higher than 0.9 are considered to be in strong linkage disequilibrium, and a value equal to 1.0 indicates complete linkage disequilibrium (31).

click to view

Table 7: Results of linkage disequilibrium analysis for transforming growth factor alpha (TGFA) variant alleles in the Human Genome Diversity Cell Line Panel*

Linkage disequilibrium calculations can provide evidence for how close in time mutation events resulting in single nucleotide polymorphisms occurred in a given population. The TGFA TaqI allele is in weak linkage disequilibrium with the C3296T allele in Adygei and Russians (D' = 0.071) and in Pakistanis (D' = 0.077), suggesting that these two mutation events are ancient (probably more than 50-100 generations old) or arose independently two or more times. The TGFA TaqI site shows weak linkage disequilibrium with the C3827T allele in French and French Basques (0.008). Several population groups have the C3296T and C3827T alleles in weak linkage disequilibrium.

Nonsyndromic oral clefts
Isolated or nonsyndromic oral clefts (those occurring in people with no other structural or developmental abnormalities) are common congenital anomalies in humans. Typically, oral clefts are anatomically divided into two groups: cleft lip with or without cleft palate (hereafter called cleft lip/palate) and cleft palate only. The prevalence at birth of cleft lip/palate among persons of European ancestry is generally near 1 in 1,000 livebirths; for cleft palate only, the prevalence at birth is lower (1 in 2,500 livebirths), but there is substantial variability and higher prevalence at birth in Northern Europeans. The only demographic variable that has been consistently associated with the prevalence of nonsyndromic oral clefts is ethnicity. Compared with European descendants, prevalence is higher in Asians and American Indians and lower in persons of African descent (32, 33).

Since the first report of an association between TGFA and oral clefts (15), some studies, but not all, have replicated this finding. Tables 8-10 summarize results from studies that investigated the possible association/linkage between oral clefts and the TGFA locus. The tables present data from all reports available, including multiple reports on basically the same data set, to allow appreciation of the different findings obtained within the same population.

click to view

Table 8: Results from case-control studies of the association between the transforming growth factor alpha (TGFA) gene and oral clefts*

click to view

Table 9: Results from meta-analyses of the association between the transforming growth factor alpha (TGFA) TaqI marker and oral clefts*

click to view

Table 10: Results from family-based studies of association/linkage between the transforming growth factor alpha (TGFA) gene and oral clefts*

The first studies suggested a stronger genetic effect than was found by subsequent studies. Both bias and genuine population diversity might explain why early studies tended to overestimate the disease predisposition conferred by TGFA polymorphisms (17). Tables 2-5 show clearly that most of the studies contained very small series. Any future publication of the results of an association study (whether negative or positive) should be accompanied by a meta-analysis of all similar studies. Accordingly, individual researchers should also publish or make easily available information that will facilitate future meta-analysis, including relevant genotype and phenotype data
(20).

There has been considerable variation in study designs, markers used, and percentages of patients with a positive family history, such that direct comparisons are difficult. The wide range in TGFA TaqI allele frequencies across different studies (3-20 percent) suggests that heterogeneity between populations may exist (34, 35). A meta-analysis (16) showed evidence of statistically significant heterogeneity between European-descendant cleft lip/palate patients from different studies before 1997, which could reflect differences in allele frequency, percentage of positive family history, cleft severity, and ethnicity. Interestingly, this same study reported similar allele frequencies for controls comprising Australians of predominantly Anglo-Celtic descent, French of Alsatian ancestry, Britons, and US European descendants from California, Iowa, Maryland, or Philadelphia, Pennsylvania (15, 28, 36-40). However, the present review does not support this statement (Table 4). Much of the variation we can see in the TaqI, C3296T, and C3827T marker allele frequencies could be due to chance, since many of the series were small. In addition, there is evidence of selection bias for these non-population-based series.

The author of the meta-analysis (16) concluded that the lack of significant heterogeneity between such diverse groups of European descendants suggests that TGFA allele frequencies are unlikely to be dramatically influenced by ethnicity. However, this may be not true for all cases. Danes have a frequency of the "rare" TGFA TaqI allele that is at least 10 percent higher than that of other European populations tested. In addition, the frequency of cleft lip/palate in Scandinavia is among the highest in the world. When case-control studies are conducted in regions admixed by Danish migrants, investigators may inadvertently select a population that has a higher frequency of the "rare" TGFA marker allele in the case group (41). This could explain the results of studies conducted in the US state of Iowa (15, 22, 26, 34, 42>-44), where there is substantial Northern European mixing (45). b

However, it is unlikely that the association between TGFA and oral clefts that has been reported in studies using family-based controls or the transmission disequilibrium test is due to the confounding influence of ethnicity (i.e., population stratification). These findings provide evidence against the ethnicity bias (34, 46-48), because the affected-family-based controls and transmission disequilibrium tests are not subject to the potentially confounding influence of population stratification (49).

There is a consistent pattern of positive findings in Australia, Chile, France, and Great Britain and negative findings in some Asian populations, such as Asian Indians, Chinese, Filipinos, and Turks (Table 8 and Table 10). Studies in North American populations present somewhat contradictory findings (see Table 8), which may result from a lack of statistical power. If TGFA has a small effect on clefting, this may be missed in assessments of both type I error and type II error.

The evidence regarding an association between genetic variation at the TGFA locus and cleft lip/palate was considered inconclusive in the first meta-analysis (16). A second meta-analysis showed a small effect of the TGFA TaqI marker (17). The current review revisited additional studies that included not only case-control and family-based approaches but also linkage. There is evidence that TGFA plays a small but significant role in cleft lip/palate and that lack of power to detect very mild effects is the main reason for the conflicting results. Investigators should move forward in the direction of functional studies to define the role of the TGFA variants described in Table 1.

The first genome-wide scan published for cleft lip/palate in European descendants studied 92 British affected sibling pairs and found maximum logarithm of the odds (LOD) scores equal to 0.66 at chromosome 2p13 (the TGFA locus) (50). This study also was unable to demonstrate the involvement of a single locus of major effect in cleft lip/palate. Genome-wide scans done in Chinese, Syrian, Turkish, and West Bengali families also found positive LOD scores for the 2p region (48, 51-53). One study of 10 families from Argentina, Mexico, and the United States did not show any linkage with the TGFA locus (54).

A targeted scan of candidate loci chosen on the basis of previous suggestive linkage and/or association with human families or suggestive animal-model data was carried out in three independent studies (47, 55, 56). Suggestive linkage and/or association between cleft lip/palate and the TGFA locus was found for Colombians (p = 0.08), Filipinos (p = 0.01), and North Americans from Ohio (p = 0.005) and Boston, Massachusetts/Texas (p = 0.014). This same approach was used to study candidate loci for cleft palate only in 24 Finnish families, but no linkage with the TGFA locus was found (57).

A meta-analysis of 13 genome scans (574 multiplex families, 3,584 genotyped individuals) of published and unpublished studies from Argentina, Australia, China, Colombia, England, India, Mexico, the Philippines, Syria, Turkey, and the United States (Iowa, Ohio, and Pennsylvania) showed suggestive linkage results (heterogeneity LOD score = 2.67; p = 0.001) for the TGFA locus on chromosome 2 (58).

Tooth agenesis
One study investigated the role of TGFA in tooth agenesis. In a Brazilian population, the affected-family-based controls and transmission disequilibrium tests showed an association between the TGFA C3827T marker and nonsyndromic tooth agenesis (p = 0.01 and p = 0.02, respectively). These results were confirmed by testing of the haplotype of TGFA TaqI-C3296T-C3827T by transmission disequilibrium test (p = 0.02). Interestingly, cases with at least one incisor missing showed a borderline association with the TGFA markers C3296T (p = 0.06) and C3827T (p = 0.05), which supports the hypothesis that distinct types of teeth have independent genetic influences. No interactions with markers in the muscle segment homeobox 1 (MSX1) gene or the paired box 9 (PAX9) gene could be seen (59).

There is strong evidence supporting the possibility that cleft lip/palate and tooth agenesis could be related. In a Dutch family, an MSX1 stop-mutation was associated with a concomitant cleft lip/palate and tooth agenesis phenotype (60). Syndromic forms of clefting, such as Van der Woude syndrome (caused by mutations in the interferon regulatory factor 6 (IRF6) gene) and autosomal-dominant Kallmann syndrome (caused by mutations in the fibroblast growth factor receptor 1 (FGFR1) gene), can present with oral clefts and tooth agenesis (61, 62). Patients with cleft lip/palate can have a frequency of tooth agenesis as much as six times higher than that of the general population (63, 64), and mice that are null for Msx1 and Pax9 have craniofacial anomalies that include cleft palate and tooth agenesis (65, 66). The association of TGFA with both cleft lip/palate and tooth agenesis is more evidence that these two defects share genetic predisposing factors.

Cancer
The role of the TGFA TaqI variant in cutaneous malignant melanoma (67, 68), breast cancer (69), and oral cancer (70) has been studied. The role of TGFA in human cancer is still unknown.

For nonsyndromic oral clefts, gene-gene and gene-environment interactions have been suggested for TGFA.

An Australian study presented no evidence for an interaction between TGFA and the retinoic acid receptor variants (28). Retinoic acid, a naturally occurring form of vitamin A, is a recognized teratogen for cleft palate.

A genetic marker (D2S378) close to the TGFA gene showed LOD scores higher than 3.0 when Italian families linked to the 6p23 markers were analyzed (71). This result suggests not only a role for the TGFA locus in human clefting but also an interaction with a gene mapped at chromosome 6p23 in the development of the cleft.

A Norwegian study presented evidence of a strong effect of the TGFA TaqI rare allele among children homozygous for one common variant of the MSX1 gene (72). That study did not suggest any possible interaction between TGFA and the transforming growth factor beta 3 (TGFB3) gene. However, a South American study did not provide evidence of an interaction between TGFA and MSX1 (73).

A Brazilian study did not find evidence of an interaction between the rare TGFA TaqI allele and the 677T allele of the 5,10-methylenetetrahydrofolate reductase (MTHFR) gene (74). However, a Norwegian study found a stronger effect of the homozygous form of the rare TGFA TaqI allele in children with one or two copies of the T allele at MTHFR C677T (relative risk = 4.0, 95 percent confidence interval (CI): 1.1, 13.9) than in children who were homozygous for the C allele (relative risk = 1.7, 95 percent CI: 0.2, 15.7) (75).

Environmental factors have been more extensively studied with regard to the association between TGFA genetic variants and oral clefts, and this research was recently reviewed (76). Among these environmental factors, maternal cigarette smoking during pregnancy presents the most compelling case for an interaction, because it has long been associated with a moderate increase in the risk of oral clefts (40, 77-84), though some studies have not confirmed such an association (39, 85-87). In a meta-analysis of the published literature (88), summary odds ratios associated with maternal smoking during pregnancy were 1.34 (95 percent CI: 1.25, 1.44) for cleft lip/palate and 1.22 (95 percent CI: 1.10, 1.35) for cleft palate only.

Evidence of interaction between TGFA marker alleles and maternal cigarette smoking during pregnancy in the risk of oral clefts can be seen in some studies, but not all (Table 11). The biologic rationale for studying the interaction between TGFA and cigarette smoking is that bronchial epithelial cells, which respond to oxidants present in cigarette smoke by producing interleukin-8, make several ligands for the epidermal growth factor receptor, including TGFA (89).

click to view

Table 11: Results from studies of the interaction between transforming growth factor alpha (TGFA) genetic variants and cigarette smoking

Besides smoking, the use of vitamin supplements, ethanol, and recreational drugs and urinary tract infection have been evaluated (44, 75, 90, 91). Only periconceptional multivitamin use showed evidence for a TGFA-nutrient interaction in risk of clefting (75, 90). Compared with infants who were homozygous for the common TGFA TaqI genotype and whose mothers used multivitamins, increased clefting risks were observed for infants with the C2 genotype (homozygous and/or heterozygous) whose mothers did not use multivitamins. Risk estimates were 3.0 (95 percent CI: 1.4, 6.6) for infants with isolated cleft lip/palate in California and 4.5 (95 percent CI: 1.3, 15.7) for infants with isolated cleft palate only in Norway.

There are no laboratory tests available as of yet, and laboratory testing is not indicated.

Molecular methods for determining the presence of the TGFA variants listed in Table 1 have been published (22, 24, 26, 92, 93). All of the studies reviewed extracted genomic DNA from blood samples or blood-spot filter cards or used exfoliated oral cells. Genotyping methods used in the studies were consistent with the standard techniques of polymerase chain reaction (PCR), PCR-restriction fragment length polymorphism, kinetic PCR, and direct sequencing.

The TGFA TaqI, RsaI, BamHI, and HinfI allelic variants can be determined by Southern blot or PCR-restriction fragment length polymorphism assay in agarose gels. The Southern blot allelic fragments will be detected using probes for the region of the gene indicated in primary references; they are: TaqI, 3.0 kilobases (common allele) and 2.7 kilobases (rare allele); RsaI, 1.5 kilobases (common allele) and 1.2 kilobases (rare allele); BamHI, 7.0 kilobases (common allele) and 4.0 kilobases (rare allele); and HinfI, 2.9 kilobases (common allele) and 2.5 kilobases (rare allele). For the TGFA TaqI variant, a PCR assay with allelic fragments of 117 base pairs (common allele C1) and 113 base pairs (rare allele C2) is available (92).

The P primer variant alleles have been detected by single-strand conformation polymorphism. The products are fragments 369 (common allele) and 365 (rare allele) base pairs long (22).

The K primer allelic variants are determined using a combination of single-strand conformation polymorphism and denaturing gradient gel electrophoresis. The primers for this four-allele polymorphism amplify a 345-base-pair fragment. In a single-strand conformation polymorphism gel, allele 3 is the fastest-migrating band, and alleles 2 and 4 comigrate. In a denaturing gradient gel electrophoresis analysis, alleles 1 and 4 comigrate. By performing both experiments, it is possible to distinguish the four alleles, especially if positive controls with known genotypes are included (22).

For the variants C3296T and C3827T, kinetic PCR- or direct-sequencing-based assays have been described. All other variants described in Table 1 were originally detected by direct sequencing. Older techniques (single-strand conformation polymorphism, denaturing gradient gel electrophoresis, or Southern blot) could probably be replaced by newer genotyping methods using available sequence data.

Other potential public health applications are dependent on confirmation that particular mutations or variants increase the risk of oral clefts or cancer.

Genetic epidemiologic data support the hypothesis of a small effect of TGFA on clefting in humans. The attributable risk of TGFA for clefts was calculated to be between 1.21 and 1.23, or a 20 percent increase in risk to offspring and siblings attributed to TGFA (94).

The magnitude of the association between TGFA and oral clefts in persons of European descent is 0.62; that is, the frequency with which the "rare" TGFA marker allele is transmitted from heterozygous parents to affected offspring is 62 percent instead of the expected 50 percent (95). This statistic further demonstrates the effect of TGFA on oral clefts in humans.

While no missense, stop-codon, or splice variants were detected that could provide direct evidence of TGFA protein dysfunction in clefting, five mutations in 3'-untranslated conserved regions, which could play a role in message stability or tissue-specific targeting, were described (26). These mutations were not found in controls. In aggregate, these mutations showed a marginal association, suggesting that, as a group, such mutations may be responsible for clefting. Although this is only weakly supportive evidence—since the statistical evidence as a whole continues to support a role of TGFA in clefting and since the exact consequences of mutations in the 3'-untranslated region are not yet fully understood—TGFA remains on any list of candidate genes for clefting.

The Tgfa knockout mice demonstrated no cleft phenotype (13, 14), suggesting that Tgfa may act as a modifier gene rather than being a necessary and sufficient determinant (96, 97). There is evidence supporting this in the studies presented in Tables 8-10 (see "Highlights" column in Table 8 and "Reported results" column for Asian Indians and Italians in Table 10). One hypothesis is that the TGFA locus modifies the expression (severity) of the cleft lip/palate trait. However, it is not clear what aspect of expression (presence or absence of palate fusion) is influenced by the TGFA locus.

In summary, the role of TGFA in clefting appears small but significant, and mutations in this gene may represent a rare cause of clefting in humans. The conflicting results seen in the literature are partially caused by differences in both study design and populations. TGFA is probably a genetic modifier of clefting in humans, which is concordant with the oligogenic model suggested for nonsyndromic oral clefts.

Investigators in future studies should focus on understanding the possible role of common polymorphic variants in the development of oral clefts. The possible interaction between TGFA and other clefting-related genes, such as MSX1, must also be explored. A more complete clinical description of affected persons, including the severity and laterality of clefts and the presence of hypodontia and other dental anomalies, might be useful in future studies. To address these issues, investigators will need to use study designs that remove bias due to differences in family history, clinical description (cleft type, severity, laterality, association with other oral and craniofacial anomalies), genetic markers, and ethnic background (to clarify possible differences in association patterns for distinct population groups).

• Table 1
• Table 2
• Table 3
• Table 4
• Table 5
• Table 6
• Table 7
• Table 8
• Table 9
• Table 10
• Table 11
• Figure 1

List of References

• Atlas of Genetics and Cytogenetics in Oncology and Haematology (last accessed 2/2008)
• Cancer Genetics Web (last accessed 2/2008)
• Gene Cards
• Human Protein Reference Database
• Information Hyperlinked Over Proteins
• Murray Laboratory genetics information server, University of Iowa
• Mutation Database
• University of California, Santa Cruz Genome Bioinformatics browser, version 24
• Utah Genome Depot