In silico analysis of single nucleotide polymorphisms (SNPs) in human FOXC2 gene

Introduction

Lymphedema is an abnormal accumulation of interstitial fluid, due to inefficient uptake and reduced flow, leading to swelling and disability, which mostly affects the extremities. It can be divided into primary and secondary lymphedema according to the underlying cause. Primary or hereditary lymphedema results from genetic damage, while secondary or acquired lymphedema is caused by lymphatic system malfunction, resulting from trauma, including surgery, radiotherapy, tumors, or infections (for example, parasitic infections)¹.

Hereditary lymphedema usually occurs as an autosomal dominant trait with allelic heterogeneity. The most common type of primary hereditary lymphedema, Milroy disease, can develop due to mutations in the vascular endothelial growth–factor receptor-3 gene (VEGFR-3; FLT4)².

Forkhead box (Fox) proteins are a family of transcription factors (TFs) that play a key role in cell development, cell cycle regulation, and other important biological processes³. FOXC2, was first identified as a transcription factor (TF) that plays a key role in the morphogenesis of the cardiovascular system⁴. Further studies revealed that FOXC2 was involved in lymphatic vascular development. In both humans and mice, FOXC2 is expressed in large amounts in the developing lymphatic vessels and in adult lymphatic valves⁵. FOXC2-deficient mice were demonstrated to have abnormal lymphatic patterning and failure to form lymphatic valves, which reveals the critical role of FOXC2 in lymphatic vascular development⁶. Truncating and missense mutations of FOXC2 have been discovered in patients with late-onset lymphedema (hereditary lymphedema II; OMIM 153200), often associated with distichiasis (double row of eyelashes), and sometimes ptosis (Lymphedema Distichiasis Syndrome [LDS]; OMIM 153400), and/or yellow nails (OMIM 153300)^7–9. LDS patients develop defects characterized by lymphatic and venous blood reflux, which means failure and/or absence of lymphatic and venous valves^10,11. On a molecular level, FOXC2 DNA binding sites are enriched in nuclear factor of activated T-cells 1 (NFATC1) consensus sequences, and the two TFs work in tandem during lymphatic vascular morphogenesis¹². FOXC2 controls the expression of proteins that are essential for lymphatic valve development, such as connexins. Thus, adequate control of their activity is extremely important for proper lymphatic vessel development and function¹³.

The aetiology governing the phenotypic variability remains unclear¹⁴. Lymphatic malformations could also result from slightly mutated germline alleles, which are difficult to access (and so to identify), mutations in regulatory regions of the DNA, epigenetic changes, or a combination of the three. Epigenome sequencing and whole exome sequencing may be needed for in-depth study of these mutations. The part genetics play in the development of lymphatic anomalies is highly complex, shown by the discovery of 23 mutated human genes¹³.

Few studies addressing FOXC2 mutations from a bioinformatics point of view have been published. Out of those, none have specifically focused on single nucleotide polymorphisms (SNPs). SNPs may affect codons of amino acids located at the forkhead active domain of the protein or other sites and may severely affect the function of the TFs. Furthermore, SNPs are feasible and cost effective to study using in silico analysis via available bioinformatics tools.

This study aimed to analyze all SNPs in the human FOXC2 gene and predict their effect on the structure, function, stability and regulation of its respective protein. The results of our study can be used in population studies to screen patients with hereditary lymphedema, and more importantly in phenotypic variations in lymphatic malformations of affected individuals.

Methods

Characterization of SNPs in 3’ and 5’ untranslated regions

PolymiRTS v3.0 (http://compbio.uthsc.edu/miRSNP/) is an integrated platform for analyzing the functional impact of genetic polymorphisms in miRNA seed regions and miRNA target sites²². We entered a second set of SNPs (containing only 5’ and 3’ SNPs) and acquired a list of the miRNAs affected by these mutations. The affected miRNAs might lead to a decrease/increase of the expression of FOXC2.

The analysis steps performed are summarized in Figure 1.

Figure 1. Flow chart of the analysis process.

Results

Project HOPE

Figure 2 shows a 3D model of rs121909107 and rs121909106 and the spatial effects of each on the respective domains they are a part of, in addition to their effects on neighbouring domains.

Figure 2. 3D representation of rs121909107 and rs121909106 and their effect of FOXC2.

Overview shows the protein in grey and the affected amino acid in purple, the close-up shows the side chains of both the wild-type (Arginine) and the mutant residue (Histidine) at position 121 and colored green and red respectively. For rs121909106; overview shows the protein in grey and the affected amino acid in purple, the close-up shows the side chains of both the wild-type (Serine) and the mutant residue (Leucine) at position 125 are shown and colored green and red, respectively.

GeneMANIA

Figure 3 shows the gene-gene interactions of FOXC2. The most important interactions are with FOXB1, FOXD1, FOXD2 and FOXD3 (Figure 3) (additional data in Supplementary material).

Figure 3. Gene-gene interaction network of FOXC2 with paths colored according to their functions.

Discussion

This study aimed to analyze the SNPs identified in the FOXC2 gene. We found a total of 473 SNPs, 2 of which were predicted to adversely affect the function of the resulting protein. One mutation leads to translation of a histidine instead of an arginine at position 121, and the mutant residue is located near a highly conserved region²⁰. The mutation was reported previously in a patient suffering from lymphedema-distichiasis syndrome by Berry et al. (2005), and the effects that the R121H mutation would have on the DNA-binding part of the forkhead domain (FHD) of FOXC2 was predicted. It was predicted that R121H would impair FOXC2 protein’s ability to bind DNA and act as a transcription activator, and this was confirmed by biochemical studies. Also, the mutation resulted in the mislocalization of FOXC2. Berry et al. (2005) determined that the mutation results in a non-functional protein and that this leads to hereditary LDS²³. The residue affected by rs121909107 is part of an inter pro-domain named Fork Head Domain Conserved Site 2 (Interpro, IPR030456), and it is annotated with the following Gene-Ontology (GO) terms to indicate its function: sequence-specific DNA binding (GO:0043565) and sequence-specific DNA binding transcription factor activity (GO:0003700).

The change of a serine into a leucine at position 125 means that a nonpolar amino acid will be replaced with a polar one. The original wild-type residue and newly introduced mutant residue differ in their electrochemical properties. The mutation results in incorporating an amino acid with a different level of hydrophobicity, this will affect hydrogen bond formation, and the mutant residue is located near a highly-conserved region²⁰. This mutation matches a previously described variant in affected members of families with lymphedema-distichiasis syndrome, previously reported by Mangion et al.²⁴ and Bell et al.⁷. The mutation was not identified in 100 normal chromosomes. This mutation results in a wide range of phenotypes from minimal distichiasis to severe lymphoedema and congenital heart disease⁷. The residue affected by rs121909106 is part of an inter prodmain named Winged Helix-Turn-Helix DNA-Binding Domain (IPR011991) and it is annotated with the following GO terms to indicate its function: nucleic acid binding (GO:0003676) and nucleic acid binding transcription factor activity (GO:0001071). No coordination was found between the two SNPs, and they are not known to belong to any haplotype.

Our results from GeneMANIA showed that FOXC2 interacts with a lot of genes (Figure 3), mainly functioning to control connective tissue, tube and epithelial tissue development. These results were proved important by Mangion et al.²⁴ and Bell et al.⁷, which showed that mutations in FOXC2 lead to developing LDS.

PolymiRTS results showed SNPs and INDELs in miRNA target sites: target sites disrupted by SNPs and INDELs in miRNA seeds and target sites created by SNPs and INDELs in miRNA seeds (Table 2–Table 4). Three miRNAs that are worth noting are hsa-miR-6886-5p (with a CS of 7), hsa-miR-6886-5p (with a CS of 7) and hsa-miR-6720-3p (with a CS of 6), which are affected by the SNPs rs201118690, rs6413505, rs201914560, respectively. This points to the possibility that the areas affected by those SNPs have an evolutionary important function.

Conclusions

In conclusion, there are many SNPs that affect FOXC2 gene; some are predicted to be harmful, such as rs121909106 and rs121909107, but most are not. miRNAs were affected by SNPs in the 3’ and 5’ untranslated regions of FOXC2 gene and three are noteworthy, hsa-miR-6886-5p, hsa-miR-6886-5p and hsa-miR-6720-3p, due their high conservation score.

Computational biology tools are very powerful, especially when provided with good data and used by experts. However, bioinformatics tools have their limitations; most importantly the fact that their results are but predictions, meaning that the information they provide us with requires confirmation using various methods such as functional studies.

Data availability

Dataset 1: SNPs used for PolymiRTS analysis. doi, 10.5256/f1000research.10937.d153064²⁵

Dataset 2: SNPs used for other analyses. doi, 10.5256/f1000research.10937.d153065²⁶