INTRODUCTION

Clefts of the lip and/or palate are among the most common congenital defects worldwide, characterized by their high genetic component and affect 1 in every 700 live births,[1] of which 70% of the cases are non-syndromic and 30% are syndromic.[2-4] Syndromic forms are usually caused by chromosomal aberrations or monogenic diseases, while non-syndromic forms derive from the interaction between genetic and environmental factors.[2,3]

The upper lip and palate are formed by the fusion of components of the frontonasal process and the maxillary processes; and failures in these events during development cause the appearance of the cleft lip with or without a cleft palate.[5,6] Thus, murine and knock-out mouse models have allowed the identification of several genes and even different chromosomal loci as involved in cleft formation.[7,8]

For example, the Transforming Growth Factor-B3 Gene (transforming growth factor B3 [TGFB3], 14q24.3) is expressed in the epithelial cells of the medial border of the palatal processes and participates in events including: Membrane degradation, metalloproteinase-mediated remodeling of the extracellular matrix, and epithelialmesenchymal transformation, which together generate the confluence of processes during palate formation.[9] Interestingly, studies in ethnically contrasting populations have verified positive findings between TGFB3 gene alterations and non-syndromic cleft lip with or without cleft palate (NSCL/P).[10,11] Most of these studies have been performed in white and Asian populations, thus reflecting the heterogeneity of the cleft phenotype as well as variability in environmental risk factors.[10-12]

Similarly, it has been verified that the fibroblast growth factors (FGFs) and its receptors play a crucial role through the regulation of cell proliferation, differentiation, and motility which are necessary for the development of the palate and upper lip.[13] Mutations in the genes coding these molecules have been shown to contribute significantly to the development of syndromic orofacial clefts, such as Apert syndrome.[14] In addition, case–control studies have shown evidence of the association of single-nucleotide polymorphisms (SNPs) in FGFs genes with susceptibility or decreased risk of NSCL/P, depending on the polymorphism and the gene analyzed.[15]

However, no single candidate gene has been consistently identified in all studies.[10-15] Therefore, the present study aims to perform a systematic review of the possible association between polymorphisms in TGFB3 and FGFs genes and NSCL/P.

MATERIAL AND METHODS

Based on the guidelines for preferred reporting elements for systematic reviews in the RevMan manual,[16] the specific PICO question[17] “What is the association of polymorphisms in TGFB3 and FGFs genes with NSCL/P?,” was developed being:

(P) Participants: Patients with cleft lip and palate.

(I) Interventions of interest: TGFB3 and FGFs genes.

(C) Controls: Groups of patients without cleft lip and palate.

(O) Outcome measures: Would be the association of polymorphism with cleft lip and palate.

The relevant studies for the present systematic review were identified in the digital databases of PubMed, Scielo, BVS, AJODO, and Google Scholar, and the last search in all databases was performed on January 31, 2022. The search terms used were the following: “Gene polymorphism and cleft lip,” “gene polymorphism and cleft palate,” “gene polymorphism and cleft lip and cleft palate,” “genetic and cleft lip,” and “genetic and cleft palate.”

Two reviewers (Andrea Soledasd Quizhpi-Quito, Diego Mauricio Bravo-Calderon) independently selected studies by examining all titles and abstracts. Any association study in humans, and meeting the following criteria were included in the study: the outcome of interest was NSCL/P, the polymorphisms studied were TGFB3 and FGFs, presenting sufficient data, that is, allele/genotype frequency between cases and controls; or their odds ratio (OR) with 95% confidence interval (CI). Reference lists of selected articles were also reviewed to identify additional relevant publications. On the other hand, animal studies, family-based studies, case reports, and publications with insufficient information such as letters from the author were excluded and, specifically for the TGFB3 gene, studies before December 2013 were not considered because the most recent systematic review on the subject was published on the aforementioned date.[18] In sequence, the lists of articles were compared, validated and disagreements were resolved by consensus among the reviewers. [Figure 1] Summarizes the literature search strategy according to PRISMA guidelines. The pattern of the present systematic review was customized to summarize mainly relevant data.

Figure 1:: Flow chart for item selection.

Export to PPT

Data extraction

A data extraction sheet was developed in Microsoft Excel, from the selected studies, the information was extracted from each study including: title, author, year of publication, database, type of gene, SNPs, type of study, and being genetic in nature, it was checked if they performed Hardy–Weinberg equilibrium (HWE).

Risk of bias assessment

Study quality was independently assessed using a risk of bias assessment for genetic association studies.[19,20] Briefly, the following domains were analyzed: Selection bias, information bias, and HWE assessment, whereby, each item was scored as 2; 1 or 0; corresponding to low risk of bias, high risk of bias, or unclear/insufficient information, respectively. Total scores ranged from 0 (worst) to 13 (best).

RESULTS

The search in the databases PubMed, Scielo, BVS, AJODO, and Google Scholar allowed us to locate a total of 373 studies, being that, after the exclusion of 33 duplicate records, 340 remained for further review. Of these 340 articles, 334 studies were discarded because after reading the abstracts, these articles clearly did not meet the inclusion criteria [Figure 1].

The reasons for ineligibility were mainly due to non-relevant genes or populations and studies published in years before the range established in the present review. Thus, there were finally six publications that were included in the systematic review, four for the TGFB3 gene, and two for the FGFs genes. The PRISMA flow chart provides an overview of the selection process of the polymorphisms studied and the number of studies selected [Figure 1].

[Table 1] describes the general characteristics of the selected studies. Among the selected articles, five had a case–control design, and one was of cross-sectional type. Regarding ethnicity, two studies were conducted in Asian, two in European, and two in Latino populations. Particularly for the TGFB3 gene, genetic variations located in exons 1–7[10-21]and the SNPs: rs2300607, rs2268625,[11] rs369477964, rs375973742, and rs117462711,[22] were studied.

Table 1:: General characteristics of the included studies about TGFB genes.

Title Author Country Ethnia Study design Number of cases Number of controls Snp/Exon Locus Variant/ Allele OR 95% IC P-value Screening of Ghazali Malasia European Cross – 96 96 EXON 1–7 TGFB3 g. 15812T.G 0.313 (0.082– P=0.089 transforming growth et al. 2014 sectional 1.191) factor beta 3 and g. 15966A.G 0.356 (0.92– P=0.315 jagged2 genes in the 1.382) Malay population with non-syndromic cleft lip with or without cleft palate Association between Oner and Turqui European Case 80 125 EXON 1 TGFB3 Pro10Leu CT 2.18 CT CT the transforming Tastan control TT 1.46 (0.94–5.06) P=0.06 growth factor beta 1 2016 TT TT gene polymorphisms (6.67–3.19) P=0.66 and Turkish patients Arg25Pro GC 1.31 GC GC with non-syndromic CC 0.29 (0.48–3.64) P=0.79 cleft lip with/without C C P=0.80 cleft palate (0.48–3.52) Association of Aljabeiti Guatemala American Case 237 266 rs2300607 TGFB3 rs2300607 P=0.020 TGFβ3 variants with et al. 2017 control rs2268625 A/T non-syndromic cleft rs2268625 P=0.041 lip and palate in C/T guatemalan population TGFB3, MSX1, and Kumari India Asia Case 162 108 rs3917219 TGFB3 rs3917219 P=0.008 MMP3 as candidates for et at. 2018 control rs369477964 rs369477964 P<0.001 NSCL+P in an Indian rs375973742 rs375973742 population rs117462711 rs117462711

In the study conducted by Ghazali et al. 2014, exons 1–7 were analyzed, finding significant results for exon 4, indicating that individuals harbored the g.15812T.G variant with an OR of 0.313 and 95% CI: 0.082–1.191, P = 0.089; while, for the g.15966A.G variant with an OR of 0.356 and 95% CI: 0.092– 1.382. P = 0.315 [Table 1].

On the other hand, the study by Oner and Tastan, in 2017, revealed two polymorphisms in exon 1: Pro10Leu and Arg25Pro, analyzing a total of 80 cases and 125 control patients. The differences in frequencies of CT and TT genotypes in the case group were not statistically significant compared to controls, OR = 2.18; 95% CI = 0.94–5.6; P = 0.06 and OR = 1.46; 95% CI = 0.67–3.19; P = 0.315, respectively. For the Arg25Pro polymorphism, there were 74 subjects out of the group of 80 cases. The frequency of the GC genotype was higher compared to the control group OR = 1.31 95% CI: 0.48–3.64, P = 0.79, but was not statistically significant. The frequency of C alleles was higher in cases than in controls OR = 0.29, 95% CI: 0.48–3.52, P = 0.8, but the differences were not significant [Table 1].

SNP rs2300607 A/T, from the study by Aljabeiti et al. 2017, had a total of 237 cases and 168 controls, the χ2 test revealed a statistically significant difference in the distribution of genotypes between cases and controls with the following P = 0.0206; χ2 = 0.7626 [Table 1]. Likewise for SNP rs2268625 C/T, 237 cases and 98 controls were analyzed, indicating that the distribution of rs2268625 C/T SNP genotypes differed significantly between cases and controls with P = 0.041 [Table 1].

In the study by Kumari et al. 2018, the SNP rs3917219 was genotyped in 162 cases and 108 controls but revealed no association with FLAPNS. The SNP rs369477964 and rs375973742 were present in two cases and two controls, of the 162 cases and 108 controls analyzed, and finally, the SNP rs117462711 registered only two cases, obtaining P < 0.001 [Table 1].

While, for FGFs, the following SNPs were studied: rs6790664, rs11717284, rs1464942, rs12106855, rs1875735,[23] rs34010, and rs13317.[15] Regarding the FGF SNPs, of the genotypic frequencies observed in the study conducted by De Aquino, with 300 cases and 365 controls, none showed an association with NSCL/P, likewise the controls did not show significant differences. In the SNPs studied by Rafiqdoost et al., a statistically significant difference was observed between cases and controls in the rs34010 CA genotype with the following OR values = 0.29, 95% CI = 0.16–0.55, P = 0.001. With respect to the FGF rs13317 AG variant, no statistically significant values were found between the genotypic frequencies of patients and controls in the models analyzed [Table 2].

Table 2:: General characteristics of the included studies about FGFs genes.

Title Author Country Ethnia Study design Number of cases Number of controls Snp/Exon Locus Variant/Allele OR 95% IC P-value Polymorphisms in FGF12, VCL, CX43 and VAX1 in Brazilian patients with non-syndromic cleft lip with or without cleft palate Aquino
et al. 2013 Brasil American Case control 300 385 rs6790664
rs11717284
rs1464942 rs12106855 rs1875735 FGF rs6790664 1.11 (0.87–1.41) P=0.56 rs11717284 0.74 (0.58–0.95) P=0.91 rs1464942 0.94 (0.72–1.24) P=0.5 rs12106855 1.12 (0.87–1.43) P=0.53 rs1875735 0.93 (0.72–1.18) P=0.76 Investigation of FGF1 and FGFR gene polymorphisms in a group of Iranian patients with non-syndromic cleft lip with or without cleft palate Rafiqdoost et al. 2014 Irani Asia Case control 100 100 rs34010 FGF1 rs34010 C/A 0.29 (0.16–0.55) P=0.001 rs13317 rs13317 A/G 0.84 (0.46–1.56) P=0.588

As shown in [Table 3], the quality analysis verified a score that varied between 6 and 10 points for the studies on the TGFB3 gene, and as for HWE, only 50% performed it.

Table 3:: Bias risk analysis of included studies about TGFB genes.

Parameters Articles on TGFB gene polymorphisms Ghazali
et al. 2014 Oner and Tastan 2016 Aljabeiti
et al. 2017 Kumari
et al. 2018 2 1 0 2 1 0 2 1 0 2 1 0 Representativeness of the cases Consecutives/randomly selected from the population of cases with a
clearly defined sampling frame Consecutive/randomly selected from the case population without a
clearly defined sampling frame or broad inclusion/exclusion criteria x x x x No selection method is described Representativeness of the controls Controls were drawn consecutively/randomly from the same sampling
frame (neighborhood/community) as cases Controls were drawn consecutively/randomly from a different sampling
frame as cases x x x x Not described Determination of flaps Clearly described objective criteria for the diagnosis of NSCL/P x x x x Diagnosis of NSCL/P by patient self-report or patient history Not described Verification of controls Controls were tested to rule out NSCL/P x x x Controls were subjects who did not report NSCL/P; no objective tests x Not described Evaluation of the partnership Evaluate the association between genotypes and NSCL/P with
appropriate statistics and adjustment for confounding factors x x To evaluate the association between genotypes and NSCL/P with
appropriate statistics without adjustment for confounding factors x x Inappropriate statistics used Genotyping test* Genotyping performed under the “blind” condition without knowledge
of the sample group Not blinded or not mentioned x x x x Hardy–Weinberg equilibrium Hardy–Weinberg equilibrium in the control group x x Hardy–Weinberg disequilibrium in the control group Without testing Hardy–Weinberg equilibrium x x Total, risk of bias 8 9 6 10

On the other hand, the value ranged between 9 and 10 points for the investigations on FGFs genes, and all FGFs polymorphisms measured HWE [Table 4].

Table 4:: Bias risk analysis of included studies about FGFs genes.

Parameters Articles on FGF gene polymorphisms Aquino et al. 2013 Rafiqdoost et al. 2014 2 1 0 2 1 0 Representativeness of the cases Consecutives/randomly selected from the population of cases with a clearly
defined sampling frame Consecutive/randomly selected from the case population without a clearly
defined sampling frame or broad inclusion/exclusion criteria x x No selection method is described Representativeness of the controls Controls were drawn consecutively/randomly from the same sampling frame
(neighborhood/community) as cases Controls were drawn consecutively/randomly from a different sampling
frame as cases x x Not described Determination of flaps Clearly described objective criteria for the diagnosis of NSCL/P x x Diagnosis of NSCL/P by patient self-report or patient history Not described Verification of controls Controls were tested to rule out NSCL/P x x Controls were subjects who did not report NSCL/P; No objective tests Not described Evaluation of the partnership Evaluate the association between genotypes and NSCL/P with appropriate
statistics and adjustment for confounding factors To evaluate the association between genotypes and NSCL/P with appropriate
statistics without adjustment for confounding factors x x Inappropriate statistics used Genotyping test* Genotyping performed under the “blind” condition without knowledge of
the sample group x Not blinded or not mentioned x Hardy–Weinberg equilibrium Hardy–Weinberg equilibrium in the control group x x Hardy–Weinberg disequilibrium in the control group Without testing Hardy–Weinberg equilibrium Total, risk of bias 10 9 DISCUSSION

The findings found in each study analyzed are detailed below. Once the articles have been selected, it should be noted that, for the TGFB gene, two papers, despite not performing the HWE balance, we consider that they have a high score in the quality analysis; this is because they meet other methodological parameters but emphasize the need to perform the HWE analysis in genetic studies.[24]

Although NSCL/P is among the most common congenital defects, the exact genetic and environmental events associated with its pathogenesis are still unknown.[11-23] Identification of the genetic alterations causing NSCL/P could lead to a better understanding of the molecular basis of the disease, and of craniofacial development.[22-23]

The present systematic review focuses on two genes, TGFB3 and FGF, which are involved in facial development, and some of their variants have been shown to be risk factors for NSCL/P, although to different degrees.[10-22]

Particularly for the TGFB3 gene, Ghazali et al., in 2014, considered the g.15812T>G variant as a polymorphism because it was present in more than 1% of the study population, and also indicates that the side most affected by clefting was the left side, which was affected 2.8 times more frequently than the right, possibly because blood is more abundant on the right side than on the left.[10]

On the other hand, Oner and Tastan, in 2016, indicates that no statistically significant differences were found in the Pro10Leu and Arg25Pro variants in exon 1.[21] However, it is demonstrated that genetic evidence of TGFB is functionally required for secondary palate formation.[10-21]

Likewise, in the study conducted by Aljabeiti et al. 2017, in the population of Guatemala, the results showed, for the first time, the association between TGFB3 rs2268625 C/T polymorphism and NSCL/P, which coincide with the results found in the study Ichikawa et al., in a Japanese population in the rs2300607 A/T polymorphism.[25] In this study, comparisons of allele frequencies for both mutations were made, finding a higher proportion of mutated alleles in cases compared to controls.[11] Therefore, it is suggested that both mutations of the TGFB3 gene are involved in the etiology of NSCL/P.[11]

In relation to FGFs genes, in the study conducted by De Aquino et al. 2013, in a Brazilian population, no significant association of FGF12 with NSCL/P was found, the author indicates that this is due to the sample size used in the study; therefore, it is possible that associations of polymorphisms and cleft risk were overlooked, indicating that their study was based on samples from European studies,[23] and also performed additional analyses on the dominant and recessive genetic models which also revealed no differences in the distribution between groups.[23]

In the SNPs studied by Rafiqdoost et al., in 2014, a statistically significant difference was observed between NSCL/P patients and the control group in SNP rs34010.[15] The FGF1 rs34010 CA genotype had a markedly higher frequency in controls compared to NSCL/P; and in the CA + AA genotype, it was more frequent in the control group.[15] With respect to the FGF1 rs13317 AG variant, no statistically significant values were found between genotypic frequencies.[15] Highlighting the protective role of the FGF1 rs34010 C/A polymorphism, therefore, the FGFs signaling pathway has constant functions in lip and palate morphogenesis, and the perturbation of its expression patterns sometimes leads to cleft pathogenesis.[15]

On the other hand, in the study conducted by De Aquino et al. 2013, in a Brazilian population, no significant association of FGF12 with NSCL/P was found, the author indicates that this is due to the sample size used in the study; therefore, it is possible that associations of polymorphisms and cleft risk were overlooked, indicating that his study was based on samples from European studies.[23]

In summary, understanding the etiology of NSCL/P patients may help to predict and prevent its occurrence in the future. Specifically, the present review did not verify significant associations between gene polymorphisms TGFB3 and FGF with NSCL/P, and these results may be based on the limited number of studies exploring these relationships and the ethnic variability of the populations among the investigations analyzed.

Moreover, facial development and its alterations, such as NSCL/P, are not only determined by genes, but also by their interaction with environmental factors such as habits, nutrition, and trauma.[26]

On the other hand, it should be noted that in this review age or gender was not considered as a determining factor because they are birth defects and measurements of genetic expressions are taken among affected individuals.

CONCLUSION

None of the genetic variations that show the association is verified in different populations; therefore, there is not enough scientific validation regarding the association between TGFB and FGF polymorphism and NSCL/P. Therefore, the findings of the different studies suggest the need for new research with samples composed of a larger number of individuals in different populations, which should be performed with all the standards for genetic studies, thus allowing to have certified information that would allow a better understanding of the molecular basis of the disease.