The incidence of congenital heart disease (CHD) varies between 4 and 8 in every 1,000 live births globally (1). However, it is considerably higher in the prenatal population; the percentage of miscarriages and elective abortions in pregnant women with structural CHD is reportedly 15 and 5%, respectively (2,3). The development and formation of the human heart is an intricate process. Unfavorable environmental and embryotoxic factors, genetic variations, numerical and structural chromosomal aberrations (e.g. trisomy of chromosome 21), as well as chromosomal microdeletions (e.g. DiGeorge syndrome), may all interfere with this process, thus leading to CHD or even miscarriage in some cases (4–6). The genetic etiology of CHD has been studied extensively over the last decade; a number of germ line mutations in cardiac transcription factors (7–13), including NK2 homeobox 5 (NKX2-5) (11,14), GATA binding protein 4 (GATA4) (15–17), T-box 20 (18), and Notch1 (19) have been validated. However, further research is required to better understand the underlying mechanisms of CHD. Next generation sequencing (NGS), in addition to its advantageous cost, accuracy and efficiency, has proven to be successful in identifying concordant variants in patients with the same disease (20). Genome-wide coverage may allow for a nonbiased approach, as it is not restricted to certain pre-selected regions. The conventional Sanger sequencing approach has been used to validate the candidate discordant variants obtained from NGS (21,22).

In order to obtain a comprehensive understanding of the effect of genetic variants on CHD, the present study used NGS to sequence the whole exome of a stillborn fetus diagnosed with CHD. In addition to a number of known CHD-inducing genes, genes with a poor association were additionally identified. The results provide a more complete understanding of the effect of specific genetic variants on CHD.

To comprehensively investigate the association between germline mutations and CHD susceptibility, the present study was completed in two consecutive steps. Whole exome sequencing of the stillborn fetus discovered 17,601 SNP sites, spanning 98% of the target region (Table II). These variants were subsequently annotated according to the dbSNP and HGP databases. 17,302 missense sites and 309 indel sites were then selected as the candidate genetic variants (Table II). In the subsequent analysis, genes associated with heart formation, development and cardiovascular disease were selected for further consideration. A number of known causative genes for congenital heart malformations, including cysteine rich with EGF like domains 1 (CRELD1) (30), tolloid like 1 (TLL1) (31), cbp/P300 interacting transactivator with Glu/Asp rich carboxy-terminal domain 2 (CITED2) (32,33) and myocardin (MYOCD) (33), were detected in the present study. However, no mutations in the exons of additional pivotal genes, including GATA4, GATA6, NKX2-5, T-box transcription factor and heart and neural crest derivatives expressed 2, were identified (Table III).

Out of the candidate genes identified, a variant of IRX1 (c.718C>G, p.Gln240Glu), which is an important gene involved in early heart development and limb formation (34), was selected for further investigation. The c.718C>G variant and whole exon 2 of the IRX1 gene were screened in 215 non-syndromic patients with CHD and 249 healthy control subjects by PCR and Sanger sequencing. The c.718C>G variant was detected in a male (age, 5 years) with TGA and an atrial septal defect (Fig. 1, Table IV). In addition, a novel variant, c893G>A, p.S298N, was detected in a 1-year-old male with total anomalous pulmonary venous drainage (Fig. 1; Table IV). An additional previously identified variant, c.1142C>A (p.A381E; dpSNP cluster ID, rs530506520) was identified in a 3-year-old male with a ventricular septal defect phenotype (Fig. 1; Table IV). However, no non-synonymous variant was detected in the 249 healthy control subjects. Conservation analysis demonstrated that glutamine 240 and serine 298 residues are highly conserved among different mammalian species, while alanine 381was only moderately conserved (Fig. 2).

NGS has become a powerful tool for identifying concordant variants in patients with the same disease. In previous studies it has successfully identified the causative gene of monogenic diseases (35), as well as a number of cancers, autoimmune diseases (36) and neurodegenerative diseases (37). In the present study, NGS was used to sequence the exome of a stillborn fetus with tricuspid atresia and complete TGA. In total, 17,302 missense sites and 309 indel sites were selected as candidate genetic variants. Out of these, a number of known cardiovascular disease-associated variants were identified, including CRELD1 (30), CITED2 (32), MYOCD (33), transmembrane protein 43 (38), TLL1 (31), IRX-1, IRX-3 and IRX-5. Therefore, the authors hypothesize that CRELD1, CITED2 and TLL1 genetic variants may have been responsible for the development of CHD in the fetus. It is possible that the IRX-1, IRX-3 and IRX-5 variants may have additionally contributed to CHD development.

The IRX gene is highly conserved among vertebrates. A total of 6 IRX genes (IRX1-IRX6) are organized in two cognate clusters of three genes, IRX1, IRX2, IRX4 and IRX3, IRX5, IRX6, respectively (39,40). The IRX gene exhibits restricted temporal and spatial expression patterns during murine neural and cardiac development (41). IRX4 was the first cardiac transcription factor identified to be expressed in the ventricles alone at all stages of heart development (40). In chicken embryos, aberrant expression of Irx4 affects heart chamber development (42). In mice, targeted inactivation of Irx4 led to aberrant ventricular gene expression, including reduced expression of the basic helix-loop-helix transcription factor (40). In this previous study, adult Irx4Dex2/Dex2 mice developed cardiomyopathy characterized by cardiac hypertrophy and impaired contractile function (40). Cardiac expression of Irx1, Irx2 and Irx5 may partially compensate for loss of Irx4 function (41).

In the present study, the coding sequence of the IRX1 gene was screened in sporadic non-syndromic patients with CHD and healthy volunteers. The number of missense mutations identified was higher in CHD patients when compared with healthy volunteers (3 of the 215 CHD cases vs. 0 of the 249 controls). These results further support the notion that disrupted IRX1 may be insufficient to induce a CHD phenotype, and that variants of the IRX1 gene only contribute to an increased susceptibility of CHD.

In the present study, the whole exome of a stillborn fetus with tricuspid atresia and complete TGA was sequenced. A number of missense mutations in known CHD-associated genes, including CRELD1, CITED2 and TLL1 were detected. In addition, the missense mutation rate of IRX1 was observed to be higher in patients with sporadic CHD when compared with normal healthy volunteers. This suggests that genetic variants of IRX1 may contribute to the development of CHD.