Congenital heart defects (CHDs) are a spectrum of diseases involving structural abnormalities of the heart and great vessels that affect about 1% of liveborn infants and occur in 10% of aborted fetuses (1). Left heart hypoplasia (LHH), a severe and complex phenotype of CHDs, refers to the underdeveloped left heart structures including left ventricle, aorta and mitral valve (2). Hypoplastic left heart syndrome (HLHS) is used to describe the most severe condition of LHH, and cause 23% of cardiac deaths in the first week of life without early surgery (3). Although a three-stage surgical palliation has significantly increased the 5-year survival rate (50–70%) (4), the extent of morbidity is considerable and life-expectance remains reduced, and some of them will still develop heart failure and need cardiac transplantation. To achieve substantive progress in the development of therapeutic intervention, the understanding of the formation of the 4-chambered vertebrate heart and the pathogenic mechanism causing LHH is needed.

Several lines of evidence support a genetic contribution to HLHS. A number of reports suggest HLHS is highly associated with chromosomal abnormalities (Turner, trisomy 13, 18, and 21) (5–7), as well as copy number variations (CNV) (DiGeorge, and Jacobsen syndrome) (8, 9). Genetic studies have also identified variants in several candidate genes such as NKX2.5, GJA1, NOTCH1, HAND1, FOXC2, FOXL1, IRX4, MYH6, and RBFOX2, in patients with HLHS (10–15), and these genes are very important for cardiogenesis.

The aim of this study is to identify the pathogenic variants in 3 fetuses with LHH by using next generation sequencing, and we found a de novo loss-of-function variant in KMT2D, and WDFY3 (WD40 repeat and FYVE domain containing 3) gene, respectively, and a de novo heterozygous microdeletion variant in chromosome 9q34.3, encompassing NOTCH1, which aid prenatal genetic counseling and understanding of the pathogenesis of LHH.

After performing genetic analysis, variants of two genes, KMT2D (NM_003482) and WDFY3(NM_014991), and a microdeletion involving NOTCH1 were detected in three fetuses, respectively. All the three variants were classified into pathogenic variants because they were LOF and de novo mutations, and absent in control population. The detailed clinical features and genetic information were shown in Table 1. No other pathogenic variants in known pathogenic genes, which were in line with family separation, were identified in the three cases.

Case 1 was identified to be a de novo heterozygous variant (c.10394delG, p.Gly3465Aspfs^*37) in the exon 36 of KMT2D gene (Figure 1A), which was reported previously and included in the HGMD (Accession number: CD114940) (20). At 23-week of gestation (wg), the ultrasound, shown in Figures 2A,B, revealed complex congenital heart defects (hypoplastic left heart, ventricular septal defect, double-outlet right ventricle, and perpetuate left superior vena cava), horseshoe kidney, and single umbilical artery, and the fetus presented arrhythmia. Mid-pregnancy ultrasound showed the fetus had increased nuchal translucency of 3.5 mm (normal value <2.5 mm) at 12 wg, and, at 18 wg, the tests of karyotyping and CMA analyses by amniocentesis were negative.

Case 2 was found to carry a de novo 150 kb microdeletion at 9q34.3 (chr9:139, 315, 643-139, 465, 759), resulting in deletion of full-length of NOTCH1, INPP5E, and SEC16A as well as part of PMPCA (Figure 1C). The ultrasound at 24 wg detected severe congenital heart defects, including small left heart (Figure 2C), coarctation of ascending aorta and aortic ache, and a small ventricular septal defect. Polydactyly and cystic adenomatoid of lung were also observed (Table 1). The nuchal translucency measurement and noninvasive prenatal testing were unremarkable.

Case 3 was identified to have a de novo heterozygous frameshift variant (c.349-c.350delAG, p.Ser117Xfs^*1) in the exon 6 of WDFY3 gene (Figure 1B), which was a novel pathogenic variant. Ultrasound of this fetus at 25 wg showed left heart dysplasia with mitral atresia and aortic valve stenosis (Figures 2D,E), and ventricular septum defect. No other abnormality was found during mid-pregnancy examination. Although, WDFY3 pathogenic variants were reported to cause microcephaly or macrocephaly (21, 22), the described fetus in this case had normal head circumference.

Although increasing evidence supports a genetic etiology (23–25), the precise pathogenesis of LHH or HLHS is unknown. In this study, we reported 2 de novo frameshift variations in the KMT2D and WDFY3 genes, respectively, and a 9q34.3 microdeletion of 150 kb encompassing NOTCH1, in three fetuses with LHH and other structural malformations.

WES identified a de novo heterozygous frameshift variant (c.10394delG) in the KMT2D gene, resulting in a truncated protein, in Case 1. The KMT2D encodes a histone H3K4 methytransferases that is associated with transcriptionally active genes. Mutations in KMT2D are most frequently, up to 70%, found in patients with Kabuki syndrome (KS, MIM: #147920), which is a rare, autosomal dominant multiple congenital anomaly syndrome characterized by recognizable facial features, global developmental delay, intellectual disability, short stature, and musculoskeletal abnormalities (26). Approximately 55% of patients with KMT2D mutations have been reported to present CHDs, including both septal defects and aortic coarctation (27). Studies in animal models have indicated KMT2D is essential for regulating cardiac gene expression during heart development and is required for the establishment of the primary and secondary heart fields (28, 29). Recently, a study including 80 fetuses with congenital cardiac left-sided lesions showed that KMT2D was the most frequently mutated gene (10.6%) followed by NOTCH1 (6.1%) (30). NOTCH1, encoding one of core receptors in Notch signaling pathway, is a well-known CHD-related gene, and loss-of-function variants in this gene confer a higher risk for and segregates with left-sided-CHD (31, 32). Deletion of the whole NOTCH1 gene was previously reported to cause non-syndromic Tetralogy of Fallot and HLHS (32, 33). Our Case 2 carried a 9q34.3 microdeletion surrounding full-length NOTCH1, and presented underdeveloped left heart and ventricular septal defect. In addition to NOTCH1, the microdeletion contained three other protein-coding genes, PMPCA, INPP5E, and SEC16A according to human hg19/GRCh37. However, our q-PCR assay confirmed that the microdeletion might not destroy the PMPCA gene which contained 13 exons, since the relative copies of PMPCA exon2 and exon 13 in the fetus was comparable to that in its parents (Figure 3). INPP5E were associated with autosomal recessive diseases, while SEC16A had not been reported to be related to a known disease. Deletion of the whole NOTCH1 gene was the most likely cause for LHH of the fetus in case 2.

Although a clear prenatal genetic diagnosis can be made only when definitely pathogenic genes and variants are identified, prenatal WES is invaluable for the diagnosis of causal genes but also for the discovery of candidate genes. In case 3, we did not detect any pathogenic variants in well-established CHD pathogenic genes, but identified a frameshift (c.349-c.350delAG, p.Ser117Xfs^*1) variant in WDFY3, thus, we propose novel candidate gene for early cardiac development supported by functional and phenotyping evidence in animal model (34). WDFY3 is located on chromosome 4q21.23, encompassing 74 exons and encoding a protein with 3,526 amino acids. The encoded protein is an adaptor protein involved in selective degradation of protein aggregates by autophagy, and is critical for brain development (35, 36). It was included in OMIM and associated with microcephaly 18, an autosomal dominant disorder. It is extremely intolerant to loss-of-function (LOF) mutation with a pLI score of 1. Pathogenic WDFY3 variants caused mild to moderate neurodevelopmental delay and intellectual disability, most of which (8/13) were loss-of-function variant and putatively lead to haploinsufficiency (21).

Homozygous knockout of this gene in mice model would lead to variable cardiac anomalies, such as ventricular septum defect, double outlet right ventricle, aortic overriding, thinning of ventricular wall, ventricular dilation and disorganized trabeculation, and abnormal NOTCH1 signaling in the heart of mice (34), suggesting that WDFY3 is critical for cardiac development. However, this report did not mention whether heterozygous WDFY3 knockout mice presented visible heart abnormalities (34). Unfortunately, we did not obtain the heart tissue of the aborted fetus to examine the impacts of WDFY3 variant on signaling pathways, such as NOTCH1 and Wnt pathways related to early cardiac development. This is the first report of a human phenotype caused by a frameshift variant in WDFY3, more cases and further functional studies are needed before we make a conclusion that WDFY3 is pathogenic candidate gene for LHH. In this study, the pregnancy was terminated due to the poor prognosis of LHH and the risk of neurodevelopmental delay. The pregnant woman and her family were informed that the recurrence risk is very low except that the variant is moasic in germline cells.

In conclusion, next generation sequencing is a powerful tool for prenatal genetic diagnosis, in particular, for fetuses with LHH. Apart from two known genes, KMT2D and NOTCH1, we reported a potential pathogenic gene WDFY3, and this is the first case of a fetal CHD with a frameshift variant in WDFY3. Before this genetic variant is used to genetic counseling, more sequencing data and functional studies are needed to determine the contribution of this genetic variant to CHDs.

Genetic variants in KMT2D and NOTCH1 previously have been reported in patients with CHDs, however, WDFY3 is reported for the first time to be possibly related to CHD in human. Genetic factor is an important factor for the pathogenesis of LHH.

The data of KMT2D p.Gly3465Aspfs^* and WDFY3 p.Ser117Xfs^*1 presented in the study are deposited in the ClinVar database, and the accession numbers were SCV001449147 and SCV001449148, respectively.

The studies involving human participants were reviewed and approved by Wuhan Children's Hospital (Wuhan Maternal and Child Healthcare Hospital) ethics committee. The patients/participants provided their written informed consent to participate in this study.

HX and ZA conceived the study. CL, WW, ZM, DZ, CJ, and ZY clinically analyzed the patients. LS and TL conducted genetic data interpretation. LS and HX performed literature review and prepared the manuscript. ZA, ZY, and HX revised the manuscript. All authors reviewed and approved the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.