Association Analysis of Nonsyndromic Congenital Heart Disease and Tag Single Nucleotide Polymorphisms of TBX20 and Genes in the Ras-MAPK Pathway

Abstract

Aims:

The present study was performed to determine whether there are variants in TBX20 and genes of the Ras-MAPK pathway associated with nonsyndromic congenital heart disease (ns-CHD).

Materials and Methods:

A total of 223 ns-CHD patients and 273 healthy controls from China were selected as study subjects to perform an association analysis using 22 tag single-nucleotide polymorphisms (tag SNPs) located either in one of three genes in the Ras-MAPK pathway (MAP2K2, BRAF, and RAF1) or the TBX20 gene that have previously been associated with syndromic congenital heart disease.

Results:

The results showed that none of these tag SNPs are associated with ns-CHD.

Conclusions:

The results suggested that these disease-causing genes, which were previously discovered in familial cases, might not be the major genetic factors causing the development of ns-CHD in Chinese.

Introduction

Congenital heart disease (CHD) is a type of severe cardiac birth defect and is the main noninfectious cause of mortality in infants (Sadowski, 2009). Its incidence may be between 6 and 8% (Hoffman and Kaplan, 2002). CHDs with only congenital heart malformations are called nonsyndromic congenital heart diseases (ns-CHDs) and account for the majority of CHDs. Many CHD-causing genes were found, but most of them were syndromic CHD-causing genes (Ware and Jefferies, 2012). Some less harmful mutations in these syndromic CHD-causing genes may be just cause for heart defects. So, these genes also could be involved in the development of ns-CHDs.

The Ras-MAPK pathway is an important signaling pathway in cell growth and differentiation. Wright and Kerr reviewed gene mutations, which can cause familial syndromic CHDs in the Ras-MAPK pathway, and determined that the abnormalities of this pathway play an important role in the development of CHDs (Wright and Kerr, 2010). Therefore, the present study aimed to analyze whether genes in the Ras-MAPK pathway are associated with the development of ns-CHDs. The mutant MAP2K2 gene can cause syndromic CHD (cardio-facio-cutaneous syndrome [CFC]) (Rodriguez-Viciana et al., 2006; Schulz et al., 2008). MAP2K2 deletions in humans can also cause heart abnormalities (Nowaczyk et al., 2014). On the other hand, mice deficient in Map2k2 function are viable (Belanger et al., 2003). So, we thought that dysfunction of MAP2K2 may be the cause for heart abnormalities, but may not be lethal, and it makes the gene look like a candidate ns-CHD-causing gene in the population. About 50% CFC patients carried the mutant BRAF genes (Schulz et al., 2008), and the gene also can cause other syndromic CHDs, such as Noonan syndrome (NS) and LEOPARD syndrome (LS) (Sarkozy et al., 2009). So, the gene is chosen too. The RAF1 (CRAF) gene is located in similar position as BRAF in the ras protein mediated mitogen-activated protein kinase (Ras-MAPK) pathway. Mutations in RAF1 can cause NS (Pandit et al., 2007). So, it is also included. Except these three Ras-MAPK genes associated with syndromic CHDs, TBX20, which is not in the Ras-MAPK pathway, is included in our study. That is because mutations in TBX20 can cause familial ns-CHDs (Kirk et al., 2007). Association analysis was performed on tag single-nucleotide polymorphisms (tag SNPs) (Xu et al., 2007) of these genes in 223 ns-CHD patients and 273 healthy controls.

Materials and Methods

Sample collection

In the present study, peripheral blood samples were collected from ns-CHD patients and healthy controls. Patients and controls received cardiac ultrasound detection, and ns-CHD patients also received surgery. Patients who had extracardiac diseases, hereditary syndromes, left ventricular ejection fractions lower than 50%, coronary heart diseases, chronic liver/kidney dysfunction, or abnormal blood glucose and lipid levels were excluded. Healthy controls received cardiac ultrasound detection to confirm that they did not have abnormal heart development. The present study was approved by the Ethics Committee of Kunming Medical University and conformed to the Declaration of Helsinki and the ethical standards required by Kunming Medical University. Informed consent of the recruited volunteers was received. The peripheral blood samples (2 mL) were collected using EDTA as the anticoagulant and were stored at −80°C.

A total of 223 ns-CHD patients were collected. There were 121 cases of atrial septal defect (ASD), 92 cases of ventricular septal defect (VSD), and 10 cases of combined ASD and VSD. A total of 273 healthy controls were also collected. All cases and controls were recruited from Yunnan and they were Han people of Southwest China. There was no significant difference (Student's t test, t = 0.092, p = 0.927) between the ages in the two groups (patient: 22.8 ± 17.3 years; controls: 22.7 ± 17.0 years). The composition of gender between the ns-CHD patients and healthy controls was also not significantly different (χ² = 0.26, p = 0.61).

Genes and tag SNP selection

Three genes (MAP2K2, BRAF, and RAF1) in the Ras-MAPK and TBX20 were selected for analysis. SNPs in these genes and within 1 kb of the five and three ends of these genes were used for constructing the tag SNPs. These SNPs were selected from the Beijing population of Han Chinese (CHB, Phase II+III) from the HapMap database (International HapMap, 2005) by the HaploView 4.2 software (Barrett et al., 2005). The tag SNP selection method was pairwise tagging only. The r² threshold value was required to be greater than or equal to 0.8. The minor allele frequency of tag SNPs in CHB was greater than 5%. A total of 25 tag SNPs were obtained using HaploView 4.2. Among them, 22 tag SNPs that were suitable for mass spectrometric analysis were selected.

DNA preparation and SNP genotyping

Peripheral blood DNA was extracted using the Axygen Mini DNA extraction reagent kit (Axygen, China). The concentrations of DNA were measured by using the value of 260/280 optical density. Sequenom Assay Designer 3.1 software (Sequenom) was used to design primers for these 22 tag SNPs. Polymerase chain reactions (PCRs) were performed by using these primers and amplification reagents (Transgen, China). PCR amplifications were performed according to the protocol of reagents. The productions of PCR were tested by mass spectrometry using the iPLEX Sequenom MassARRAY platform (Sequenom, San Diego, CA) for getting the genotypes. Genotyping was conducted without knowledge of the case-control status. In addition, 5% of the samples were randomly selected to repeat genotyping for confirmation, and the accordance ratio was 100%.

Data analysis

The chi-square test of the frequencies of alleles and genotypes was calculated using the SHEsis software (Shi and He, 2005) and the PLINK software under different genetic model (Purcell et al., 2007). The rare alleles were assumed to be harmful, and SNPs were excluded if they had a frequency less than 5% in at least one of the cells in χ² tests. The Hardy-Weinberg equilibrium (HWE) test was also performed using the SHEsis software. The significant cutoff values (α value) of multiple χ² tests (including the HWE) were all adjusted by Bonferroni correction. The independent sample t tests were calculated using the SPSS software.

Results

Association analysis

All 22 SNPs were successfully detected. The lowest HWE p value in the ns-CHD group was 0.083 (rs3773341), while the lowest p value in the control group was 0.018 (rs350916). After the Bonferroni correction, the cutoff value of statistical significance was α = 0.05/22 = 0.002. Therefore, all 22 SNPs met the HWE in ns-CHD patients and controls.

The results of allele association are shown in Table 1. Furthermore, association analyses were also performed using ASD or VSD as a case group (data not shown). All of these tag SNPs did not have a strong statistical association with disease. The lowest p values were present at the rs350916 locus of the MAP2K2 gene (p = 0.049) in the analysis for ns-CHD patients. However, after multiple detection correction, this positive association disappeared.

Table 1.

The Allele Association Analyses Between 22 Single-Nucleotide Polymorphisms and Congenital Heart Disease

Gene	CHR	SNP	BP	A1	F_A	F_U	A2	χ ²	p	p_bonf	FDR_BH
RAF1	3	rs3773341	12565033	G	0.419	0.423	A	0.015	0.904	1	0.904
	3	rs2648309	12565833	A	0.022	0.033	G	0.995	0.319	1	0.773
	3	rs15997	12572570	T	0.099	0.104	C	0.088	0.766	1	0.847
	3	rs1532534	12584755	G	0.040	0.053	T	0.885	0.347	1	0.773
TBX20	7	rs10247437	35172773	T	0.126	0.097	C	2.037	0.154	1	0.773
	7	rs2191883	35223678	C	0.280	0.306	T	0.774	0.379	1	0.773
	7	rs17171924	35231619	C	0.040	0.048	A	0.305	0.581	1	0.798
	7	rs17675148	35238663	C	0.222	0.245	T	0.751	0.386	1	0.773
	7	rs336284	35244360	G	0.401	0.440	A	1.469	0.225	1	0.773
	7	rs1003549	35245879	C	0.054	0.044	T	0.518	0.472	1	0.798
	7	rs336286	35250200	A	0.184	0.194	T	0.169	0.681	1	0.847
BRAF	7	rs1267647	78279708	T	0.164	0.172	C	0.126	0.722	1	0.847
	7	rs3748093	78293871	A	0.170	0.150	T	0.750	0.387	1	0.773
	7	rs1476868	78298609	T	0.081	0.093	C	0.494	0.482	1	0.798
	7	rs1267606	78341206	C	0.083	0.079	A	0.059	0.809	1	0.847
	7	rs17161747	78352160	C	0.202	0.218	G	0.385	0.535	1	0.798
	7	rs11762469	78407632	T	0.132	0.145	A	0.315	0.575	1	0.798
MAP2K2	19	rs350916	4034776	G	0.305	0.365	A	3.890	0.049	1	0.761
	19	rs1979013	4038948	A	0.249	0.275	G	0.846	0.358	1	0.773
	19	rs350910	4040661	G	0.043	0.070	A	3.303	0.069	1	0.761
	19	rs1823059	4055994	A	0.267	0.275	G	0.078	0.780	1	0.847
	19	rs12459940	4058872	C	0.294	0.324	T	1.063	0.302	1	0.773

Bold indicates p < 0.05.

A1, minor allele name (based on whole sample); A2, major allele name; BP, physical position; CHR, chromosome; F_A, frequency of A1 allele in cases and controls; FDR_BH, Benjamini and Hochberg (1995) step-up FDR; F_U, frequency of A1 allele in controls; P_bonf, Bonferroni single-step adjusted p-values; χ², basic allelic test chi-square (1df).

The results of genotype association under genotypic model, dominant model, and the recessive model using ns-CHD as a case group are shown in Table 2. Finally, 10 tag SNPs met the conditions (frequencies less than 5%). The lowest p values (p = 0.012) were found at the rs350916 locus of the MAP2K2 gene in recessive inheritance model. However, after multiple detection correction (α = 0.05/10 = 0.005), this statistically positive association disappeared. We also analyzed the genotypes of ASD and VSD patients and did not find a statistically significant association (data not shown).

Table 2.

The Association Analyses Between the Genotypes of 10 Tag Single-Nucleotide Polymorphisms and Congenital Heart Disease

Test	Gene	CHR	SNP	A1	A2	AFF	UNAFF	χ ²	p	pbonf
Genotypic model df = 2	RAF1	3	rs3773341	G	A	41/105/77	56/119/98	0.681	0.711	1
	TBX20	7	rs2191883	C	T	11/103/109	25/117/131	3.346	0.188	1
	TBX20	7	rs17675148	C	T	6/87/130	15/104/154	2.382	0.304	1
	TBX20	7	rs336284	G	A	33/113/77	49/142/82	1.553	0.460	1
	BRAF	7	rs1267647	T	C	6/61/156	6/82/185	0.515	0.773	1
	BRAF	7	rs17161747	C	G	11/68/144	13/93/167	0.717	0.699	1
	MAP2K2	19	rs350916	G	A	13/110/100	34/131/108	6.547	0.038	0.3788
	MAP2K2	19	rs1979013	A	G	14/83/126	20/110/143	0.879	0.644	1
	MAP2K2	19	rs1823059	A	G	18/83/122	18/114/141	1.223	0.543	1
	MAP2K2	19	rs12459940	C	T	20/91/112	27/123/123	1.316	0.518	1
Dominant model df = 1	RAF1	3	rs3773341	G	A	146/77	175/98	0.101	0.751	1
	TBX20	7	rs2191883	C	T	114/109	142/131	0.039	0.843	1
	TBX20	7	rs17675148	C	T	93/130	119/154	0.178	0.673	1
	TBX20	7	rs336284	G	A	146/77	191/82	1.137	0.286	1
	BRAF	7	rs1267647	T	C	67/156	88/185	0.274	0.601	1
	BRAF	7	rs17161747	C	G	79/144	106/167	0.607	0.436	1
	MAP2K2	19	rs350916	G	A	123/100	165/108	1.407	0.236	1
	MAP2K2	19	rs1979013	A	G	97/126	130/143	0.840	0.359	1
	MAP2K2	19	rs1823059	A	G	101/122	132/141	0.462	0.497	1
	MAP2K2	19	rs12459940	C	T	111/112	150/123	1.316	0.251	1
Recessive model df = 1	RAF1	3	rs3773341	G	A	41/182	56/217	0.353	0.552	1
	TBX20	7	rs2191883	C	T	11/212	25/248	3.255	0.071	0.7122
	TBX20	7	rs17675148	C	T	6/217	15/258	2.380	0.123	1
	TBX20	7	rs336284	G	A	33/190	49/224	0.883	0.347	1
	BRAF	7	rs1267647	T	C	6/217	6/267	0.126	0.722	1
	BRAF	7	rs17161747	C	G	11/212	13/260	0.008	0.930	1
	MAP2K2	19	rs350916	G	A	13/210	34/239	6.280	0.012	0.1221
	MAP2K2	19	rs1979013	A	G	14/209	20/253	0.211	0.646	1
	MAP2K2	19	rs1823059	A	G	18/205	18/255	0.399	0.528	1
	MAP2K2	19	rs12459940	C	T	20/203	27/246	0.122	0.727	1

Bold indicates p < 0.05.

AFF, Genotypes in cases (A1A1/A1A2/A2A2 in genotypic model, (A1A1+A1A2)/A2A2 in dominant model, A1A1/(A1A2+A2A2) in recessive model); p, unadjusted p.

Linkage disequilibrium analysis

We also performed linkage analysis on the 22 selected tag SNPs using our data (Fig. 1). The results showed that the linkage disequilibrium of these SNPs in Han people living in southern China (Yunnan) is similar to that of the CHB population (all r² ≤ 0.8 except rs1823059-rs12459940). So, these tag SNPs can work well in Yunnan people.

FIG. 1.

The linkage disequilibrium plot. LD plots using the SNP data of present study and r² x100 values for the 22 tag SNPs in the four genes. SNPs, single nucleotide polymorphisms.

Discussion

Wright and Kerr found that many gene mutations in the Ras-MAPK pathway could induce syndromic CHDs (Wright and Kerr, 2010). For example, MAP2K2 gene mutations could induce CFC (Rodriguez-Viciana et al., 2006; Schulz et al., 2008; Nowaczyk et al., 2014); BRAF gene mutations could induce CFC syndrome (Niihori et al., 2006; Schulz et al., 2008), NS, and LS (Sarkozy et al., 2009); and RAF1 gene mutations could also induce NS (Razzaque et al., 2007) and LS (Pandit et al., 2007). This implied that these genes were important in the development of heart and variants in them maybe also correlate with the ns-CHDs. However, based on results in the present study, all of the tag SNPs in these three genes did not correlate with ns-CHDs. So, variants in MAP2K2, BRAF, and RAF1 may not be the main factors influencing ns-CHD development. However, we found that the rs350916 locus in the MAP2K2 gene was associated with ns-CHD both in gene frequency and in genotype composition before applying statistical multiple detection correction. If the pathogenic variants of MAP2K2 are present, the association analysis may also result in weak significance when the patients carrying these variants are few. That is because the few numbers of these patients are not enough to affect the distribution of individual numbers in the case and control groups. Therefore, if this association is not a statistical error, then some variants with low frequencies in MAP2K2 could affect ns-CHD development.

The present study also analyzed tag SNPs in the TBX20 gene. The results showed that the TBX20 gene is also not the major gene affecting ns-CHD development. Development of ns-CHD caused by the TBX20 mutation was first discovered in a family that presented with only the cardiac malformation phenotype (Kirk et al., 2007). Mutations in the TBX20 gene were found to directly cause ns-CHDs. However, previous reports showed that the TBX20 mutation was only a minor influencing factor in ns-CHD development in various populations. For example, in 203 cases of ns-CHD patients in northern China, only three patients carried nonsynonymous mutations in the TBX20 gene that might cause the disease (Liu et al., 2008). Additionally, in 170 cases of German ns-CHD patients, only one case carried the TBX20 gene mutation. The present study also used the tag SNP method to confirm that TBX20 gene variants are not the major mutations influencing the development of ns-CHDs.

Although the present study did not find ns-CHD-associated variants, other genes in the Ras-MAPK pathway are still worth further investigation. Additionally, genes that had weak associations, such as the MAP2K2 gene, and genes that previous studies found to be associated with ns-CHD, such as the TBX20 gene, are still worth whole-gene sequencing to determine whether they carried low-frequency variants that caused the ns-CHD.

Footnotes

Acknowledgments

This study was supported by some research grants, including Chinese National Natural Science Foundation (NO. 31571304, NO.81660052) and Application of Basic Research Project funding by Yunnan Provincial Science and Technology Department and Kunming Medical University.

Author Disclosure Statement

No competing financial interests exist.

References

Barrett

, Fry

, Maller

, et al. (2005) Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics, 21:263-265.

Belanger

, Roy

, Tremblay

, et al. (2003) Mek2 is dispensable for mouse growth and development. Mol Cell Biol, 23:4778-4787.

Hoffman

, Kaplan

(2002) The incidence of congenital heart disease. J Am Coll Cardiol, 39:1890-1900.

International HapMap

(2005) A haplotype map of the human genome. Nature, 437:1299-1320.

Kirk

, Sunde

, Costa

, et al. (2007) Mutations in cardiac T-box factor gene TBX20 are associated with diverse cardiac pathologies, including defects of septation and valvulogenesis and cardiomyopathy. Am J Hum Genet, 81:280-291.

Liu

, Shen

, Li

, et al. (2008) T-box transcription factor TBX20 mutations in Chinese patients with congenital heart disease. Eur J Med Genet, 51:580-587.

Niihori

, Aoki

, Narumi

, et al. (2006) Germline KRAS and BRAF mutations in cardio-facio-cutaneous syndrome. Nat Genet, 38:294-296.

Nowaczyk

, Thompson

, Zeesman

, et al. (2014) Deletion of MAP2K2/MEK2: a novel mechanism for a RASopathy?. Clin Genet, 85:138-146.

Pandit

, Sarkozy

, Pennacchio

, et al. (2007) Gain-of-function RAF1 mutations cause Noonan and LEOPARD syndromes with hypertrophic cardiomyopathy. Nat Genet, 39:1007-1012.

10.

Purcell

, Neale

, Todd-Brown

, et al. (2007) PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet, 81:559-575.

11.

Razzaque

, Nishizawa

, Komoike

, et al. (2007) Germline gain-of-function mutations in RAF1 cause Noonan syndrome. Nat Genet, 39:1013-1017.

12.

Rodriguez-Viciana

, Tetsu

, Tidyman

, et al. (2006) Germline mutations in genes within the MAPK pathway cause cardio-facio-cutaneous syndrome. Science, 311:1287-1290.

13.

Sadowski

(2009) Congenital cardiac disease in the newborn infant: past, present, and future. Crit Care Nurs Clin North Am, 21:37-48, vi.

14.

Sarkozy

, Carta

, Moretti

, et al. (2009) Germline BRAF mutations in Noonan, LEOPARD, and cardiofaciocutaneous syndromes: molecular diversity and associated phenotypic spectrum. Hum Mutat, 30:695-702.

15.

Schulz

, Albrecht

, Arici

, et al. (2008) Mutation and phenotypic spectrum in patients with cardio-facio-cutaneous and Costello syndrome. Clin Genet, 73:62-70.

16.

Shi

, He

(2005) SHEsis, a powerful software platform for analyses of linkage disequilibrium, haplotype construction, and genetic association at polymorphism loci. Cell Res, 15:97-98.

17.

Ware

, Jefferies

(2012) New Genetic Insights into Congenital Heart Disease. J Clin Exp Cardiolog, S8. pii 003.

18.

Wright

, Kerr

(2010) RAS-MAPK pathway disorders: important causes of congenital heart disease, feeding difficulties, developmental delay and short stature. Arch Dis Child, 95:724-730.

19.

, Kaplan

, Taylor

(2007) Tag SNP selection for candidate gene association studies using HapMap and gene resequencing data. Eur J Hum Genet, 15:1063-1070.