Exploration of copy number variations and candidate genes in fetal congenital heart disease using chromosomal microarray analysis

Study participants

We selected 391 cases of CHD diagnosed using fetal echocardiography among fetuses who underwent fetal ultrasound systematic examination at Wuxi Maternal and Child Health Hospital from January 2019 to December 2023. 200 pregnant women with normal newborns were recruited as control group. Informed consent was obtained from the pregnant women and their families. This study was approved by the Ethics Committee of the Wuxi Maternal and Child Health Hospital (Approval Number: 2023-01-0628-14). The inclusion and exclusion criteria were referenced from our previous research [6].

Study methods

Statistical methods

Statistical software (Graph pad 10.4) was used for data analysis. Count data was expressed as the number of cases or percentages, mean standard deviation. Comparisons between groups were performed using the χ² test and one-way ANOVA. Statistical significance was set at p<0.05.

Basic information of the study participants

In total, 391 fetuses with CHD were included in this study. The prevalence of CHD of newborns in current district was about 1.5 %. Among them, 129 patients had isolated CHD (Group 1), accounting for 33 %; 144 patients had complex CHDs (Group 2), accounting for 36.8 %; and 118 patients had CHD combined with extracardiac structural abnormalities (Group 3), accounting for 30.2 %. There were 391 pregnant women of CHD fetuses in total, aged between 15 and 44 years, with a mean age of 29.0 ± 3.9 years. Compared to the control group’s average age of 29.6 ± 5.0 years among 200 pregnant women, no significant differences were observed. Additionally, the mean ages for Groups 1, 2, and 3 were 29.2 ± 3.7, 29.3 ± 3.8, and 29.1 ± 4.3 years, respectively, with no statistically significant differences detected among the groups. The gestational age for ultrasound diagnosis of fetal CHD ranged from 18 to 28 weeks, with an average of 23 ± 2.0 weeks.

Genetic analysis results of CHD

Isolated CHD group

The overall abnormality detection rate in the isolated CHD group was 11.6 % (15/129). Chromosomal karyotype analysis detected 3 abnormalities (2.3 %, 3/129), and chromosomal microarray analysis (CMA) detected 12 abnormalities (9.3 %, 12/129). The ratio of subtype composition to the isolated CHD group is summarized in Figure 1. The isolated CHD group was divided into four major categories: isolated valve abnormalities, vascular diseases, septal defects, and conotruncal malformations. Among them, the abnormal rate of conotruncal malformations was the highest, although there were no significant differences among the four categories (Table 1). The CNV variants detected in Group 1 can be seen in Table 2.

Figure 1:

Composition diagram of subtypes in the isolated CHD group.

Table 1:

The abnormality genetic results for different subtypes of the isolated CHD group.

Classifications of CHD (number)	Subtypes of CHD	Likely pathogenic/pathogenic CNVs	VUS variants	Karyotype abnormality	Overall abnormality rate (n)
A: Isolated Valve Abnormalities (19)	MVR, TVR, AVR, PVR, etc.	0	1	1	10.5 % (2/19)
B: Vascular Diseases (29)	RSAA, PLSVC, etc.	0	3	0	17.2 % (5/29)
C: Septal Defects (63)	ASD, VSD, AVSD, CECD, etc.	0	5	1	9.5 % (6/63)
D: Conotruncal Malformations (18)	TOF, DORV, PTA, PA, CoA, TGA, etc.	0	3	1	33.3 % (6/18)
Total (129)		0	12	3	11.6 % (15/129)

1. Full name of CHD subtypes can be seen in “Abbreviations of CHD subtypes”.

Table 2:

CNV results detected in the isolated CHD group.

No.	Ultrasonic manifestation	Results of CMA	Segment of CNV length	Variant type	Number of protein-coding genes	OMIM genes	Genes associated with CHD	Dosage sensitivity genes
A-1	SUA	arr[hg19] 2q13 (110,876,775-110,980,295) × 1	103.5 Kb	VUS	2	NPHP1, MTLN	N/A	N/A
A-2	PLSVC	arr[hg19] Xp22.31 (6,455,150-8,135,568) × 3	1.6 MB	VUS	4	STS, PUDP, VCX, PNPLA4	N/A	STS (HI score: 3)
A-3	TGA	arr[hg19] 15q15.3 (44,096,004–44,696,149) × 1	600.1 Kb	VUS	5	FRMD5, HYPK MFAP1, WDR76	N/A	N/A
A-4	PTA	arr[hg19] 6q23.3q25.1 (138,351,191-149,830,858) × 2, hmz	11.4 Mb	VUS	43	PERP, ARFGEF3, PBOV1, HEBP2, NHSL1, CCDC28A, REPS1, HECA, TXLNB, CITED, NMBR, VTA1, ADGRG6, HIVEP2, AIG1, ADAT2, PEX3, FUCA2, PHACTR2, etc.	PERP, ECT2L, CITED2, PLAGL1, TAB2	HIVEP2 (HI score: 3)
A-5	DORV	arr[GRCh37] 3p24.3 (19358886–19655517) × 1	297 kb	VUS	1	KCNH8	N/A	N/A
A-6	VSD	arr[GRCh37] 8p22 (14836264_15984355) × 3	1.15 Mb	VUS	3	MSR1, SGCZ, TUSC3	N/A	N/A
A-7	VSD	arr[GRCh37] 9q34.3 (139295495_139654647) × 1	359 kb	VUS	12	PMPCA, AGPAT2, INPP5E, NOTCH1, ENTR1, SEC16A, EGFL7, DIPK1B, LCN10, LCN6, LCN8	NOTCH1	NOTCH1 (HI score: 3)
A-8	VSD	arr[GRCh37] 7q35 (145,573,563_146,998,389) × 3	1.42 Mb	VUS	1	CNTNAP2	N/A	N/A
A-9	TVR	arr[GRCh37] 9p21.3p21.2 (25,096,870_26,425,831) × 3	1.33 Mb	VUS	1	TUSC1	N/A	N/A
A-10	PLSVC	arr[GRCh37] 5q14.2q14.3 (82,281,502_83,636,624) × 1	1.36 Mb	VUS	5	VCAN, XRCC4, TMEM167A, HAPLN1, EDIL3	VCAN	N/A
A-11	ASD	arr[GRCh37] 2q13q14.1 (113,972,494_114,804,521) × 1	832 kb	VUS	7	PAX8, ZNG1B, FOXD4L1, RABL2A, SLC35F5, ACTR3, IGKV1OR2-108	PAX8	PAX8: (HI score: 1)
A-12	VSD	arr[GRCh37] 3q26.1 (164,956,837_167,403,859) × 3	2.45 Mb	VUS	2	BCHE, PDCD10	N/A	N/A

1. Full name of CHD subtypes can be seen in “Abbreviations of CHD subtypes”.

Comparison of maternal age between the chromosomal abnormalities CHD group and normal chromosome group

Figure 2 illustrates that maternal ages among three groups – those with isolated CHD, complex CHD, and CHD with extracardiac structural abnormalities – were stratified according to fetal chromosomal status: chromosomal abnormalities and normal chromosomes. No statistically significant differences were found between the CHD groups and the control group (p>0.05). Furthermore, no significant differences were detected between any of the CHD groups themselves.

Figure 2:

Comparison of maternal age between the chromosomal abnormality CHD group and normal chromosome group.

GO enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis results for isolated and complex CHD groups

OmicShare tools were used to perform GO analysis on isolated and complex CHD at three levels: biological process (BP), cellular component (CC), and molecular function (MF). Figures 3–5 show an overview of the enrichment results of the GO analysis for both CHD groups.

Figure 3:

Top 15 significantly enriched GO terms (cellular component) for isolated CHD (A) and complex CHD (B).

Figure 4:

Top 15 significantly enriched terms (molecular function) for isolated CHD (A) and complex CHD (B).

Figure 5:

Top 15 significantly enriched GO terms (biological process) for isolated CHD (A) and complex CHD (B).

KEGG pathway analysis for isolated and complex CHD groups

We conducted KEGG pathway analysis for both CHD types. The top 15 enriched terms and the number of genes involved in these terms for isolated and complex CHD in the KEGG pathways are shown in Figure 6. The top three significantly enriched KEGG pathways for isolated CHD were as follows: bacterial invasion of epithelial cells, SNARE interactions in vesicular transport, and cell adhesion molecules. The top three significantly enriched KEGG pathways for complex CHD were as follows nucleocytoplasmic transport, cell adhesion molecules, and mRNA surveillance pathway.

Figure 6:

KEGG pathway analysis for isolated CHD (A) and complex CHD (B).

Comparison of chromosome aberration rates among different CHD groups

Reports indicate that approximately 8–10 % of CHD cases are associated with numerical chromosomal abnormalities, with the most common types being trisomy 21, trisomy 18, trisomy 13 syndrome, and Turner syndrome [7]. In this study, among 391 CHD cases, 28 cases of numerical chromosomal abnormalities were detected, including 24 cases of trisomy 21, 18, 13, and trisomy 16, as well as 4 cases of Turner syndrome and Klinefelter syndrome. The overall abnormal detection rate by karyotype analysis was 9.7 % (38/391), which was similar to that reported in the literature [8], 9]. Among the three CHD classifications, no numerical chromosomal abnormalities were found in the isolated and complex CHD groups; and the overall detection rate of chromosomal abnormality in all CHD cases was 23.5 % (92/391). It can be seen that the more complex the disease, the higher the rate of overall chromosomal abnormalities, which was consistent with what Xu reported [10].

Comparison of abnormal CNV rates among different CHD groups

In the current study, CMA detected an additional 20 pathogenic CNVs in all CHD cases, increasing the positive detection rate by 5.1 % compared with karyotype analysis. In the isolated CHD, complex CHD, and CHD with other extracardiac abnormalities groups, the additional detection rates by CMA were 9.3 % (12/129), 9.7 % (14/144), and 23.7 % (28/118), respectively. The pathogenic CNV rate in the isolated CHD group was significantly lower than that in the CHD group with extracardiac abnormalities, and there was no statistical difference compared with the complex CHD group. Additionally, in the isolated CHD, complex CHD, and CHD groups with other extracardiac abnormalities, the CNV detection rates were 9.3 , 10.4, and 26.3 % respectively. The detection rate of overall CNVs in CHD patients with other extracardiac abnormalities was significantly higher than that in the other two groups, suggesting that the more complex the disease, the higher the fetal abnormality rate. For fetuses with CHD with no abnormalities detected by karyotype and CMA, there may still be gene-related variations or associations with maternal factors, environmental factors [11], etc., which require further exploration.

Involved genes within CNV detected in isolated CHD and complex CHD group

The development process of the heart is very complex, it is more related to genetic mutations in addition to chromosomal abnormalities and CNVs. We further investigated the genes potentially associated with CHD that may be involved in the CNV regions. The study has compiled a list of CNVs and the involved genes detected in cases of single CHD and complex CHD group. Based on Table 3, it can be observed that in the complex CHD group (Group 2), three cases of pathogenic CNVs are all heterozygous deletion variants. As for case B-1 in Group 2, the genes associated with CHD were ETS1, JAM3, ROBO4, etc. Animal experiments have confirmed that gene loss of function caused by missense or nonsense mutations in the EST1 gene leads to ventricular developmental abnormalities [12]. As for gene JAM3, a member of the junction adhesion molecule family, the deletion of JAM3 could lead to hypoplastic left heart (HLH) or VSD [13]. As for case B-9, the correlated genes were CEP152, FBN1. A cohort study of the whole-exome sequencing indicate that CEP152 insertions or deletions can lead to CHD [14]. Although it had reported that missense mutations of FBN1 can be correlated to cardiac defect [15], genes within the CNV region that are dose-sensitive, perhaps would play a crucial role in the development of CHD. In case B-13, the associated and dose-sensitive gene was NIPBL (HI score:3), it had proved that insertion deletion of NIPBL could contribute to cardiac defects [16]. Among the remaining 12 cases of VUS with CNVs in complex CHD group, only 2 of these CNVs contain candidate genes related to CHD. As for case B-6 and B-16, candidate gene was BPGM and SOX4, respectively. A large cohort study has proved deletion of BPGM could lead to cardiac defect [17]. Although it was suggestted that the loss of SOX4 gene could contribute to a change in the corresponding protein expression levels, the clinical phenotype is not only limited to AVSD but also includes conditions such as learning disabilities and ocular abnormalities [18]. The association between candidate gene (SOX4) and the occurrence of CHD is not yet clear and requires further investigation.

As for the isolated CHD group (Table 2), there are 12 VUS variants, with 5 microduplications, 6 microdeletions and one ROH. Only 3 cases of microdeletions contain candidate genes associated with CHD. NOTCH1 (in case A-7 of VSD), which is correlated to cardiac lesions, definitely is an important candidate gene for CHD. Besides missense mutation of NOTCH1 can lead to right-sided cardiac lesion [19], NOTCH1 gene has a dosage sensitivity of haploinsufficient which related to Adams-Oliver syndrome 5. Roifman et al. [20] had described deletion at 7q34.3 including NOTCH1 can result in cardiac defects with BAV, CoA, HLH, etc. In case A-10 (PLSVC), the candidate gene was VCAN, a mouse model study has proved the homozygous for the exon 7 deletion in VCAN can result in VSD [21]. In case A-11(ASD), the correlated gene is PAX8. A study of PAX8 null mice murine hearts has found that PAX8 may mediate the development of VSD. The potential mechanism may correlate to the abnormal expression of PAX8, which leads to apoptosis and imbalance in proliferation, consequently resulting in septal dysplasia [22].

The correlation between maternal age and chromosomal abnormality of CHD groups

A recent cohort study has reported that maternal age ≥40 years were associated with smaller systolic and diastolic LV diameters [23]. Nonetheless, in the current study, no significant differences in maternal age were observed among any CHD subgroups (Figure 2). While it is well-known that advanced maternal age is a risk factor for offspring chromosomal abnormalities, we discovered no correlation between maternal age and the incidence of CHD in offspring.

GO and KEGG pathway analysis results for isolated and complex CHD groups

At the cellular component level for isolated CHD (Figure 3), the glycoprotein complex, which ranks second in significant enrichment, most commonly includes the dystrophin-glycoprotein complex (DGC), which is widely present in myocardial and skeletal muscles. A multicenter study has shown that this complex is associated with muscular dystrophies or heart diseases [24]. The intracellular membrane system, had the largest number of enriched genes, including VTA1, VCAN, UFD1, and other genes involved in the occurrence of CHD. A mouse model showed that deletion of exon 7 of the VCAN gene prevents the expression of some mRNA splice forms and alters the heart morphology [21]. A study of family members suggested that phosphorylation of UFD1 may be involved in the pathogenesis of VSD [25]. The most significantly enriched terms at the cellular component level were more complex in the complex CHD group than in the isolated one. Integrator complex is a regulatory factor that plays a key role in RNA polymerase II-mediated transcription. The desmosome is an adhesive intercellular connection crucial for tissues that bear mechanical stress, such as the myocardium and gastrointestinal mucosa [26]. Telomere shortening has also been linked to certain heart diseases. An animal experiment demonstrated that low-power infrared laser irradiation of the relative mRNA levels of genes in heart tissue can lead to telomere shortening and decreased relative mRNA levels in LPS-induced AL1 animals [27].

At the molecular function level (Figure 4), for isolated CHD, the top three significantly enriched terms were benzodiazepine receptor binding, magnesium ion transmembrane transporter activity, and soluble N-ethylmaleimide-sensitive factor attachment protein receptor binding (SNARE binding). The benzodiazepine receptor binding site, which is ranked first, exists in many tissues and cell types. Peripheral benzodiazepine receptors (PBR), one of the benzodiazepine receptor binding site, is present in almost all peripheral tissues of mammals. A previous study showed that PBR receptors can be located on the plasma membrane of the heart and are coupled to calcium ion channels [28]. SNARE proteins facilitate the delivery of various substances within eukaryotic cells and defects in these processes are associated with various human diseases such as diabetes, ciliopathies, etc. [29]. For complex CHD, the top three significantly enriched terms were sulfuric ester hydrolase activity, methylated histone binding, and metallodipeptidase activity. Barker et al. [30] proposed that potential epigenetic mechanisms include modified gene expression through altered DNA methylation, histone acetylation, and stem cell allocation. Therefore, problems with the binding process of methylated histones can affect fetal heart development. Peptidase inhibitor activity showed the largest number of enriched genes in the complex CHD group. A study proposed that SerpinE2 is translocated to cardiac fibroblasts by low-density lipoprotein receptor-related protein 1 and urokinase plasminogen activator receptoron, which activates specific signaling pathways [31]. Therefore, peptidase inhibitor activity is likely to affect heart function.

At the biological process level for isolated CHD (Figure 5), the top three most significantly enriched terms were outflow tract (OFT) morphogenesis, parathyroid gland development, and positive regulation of bicellular tight junction assembly. The mature OFT is between the contractile myocardial chamber and vast embryonic vascular network, and its abnormalities can lead to the formation of various CHD [32]. The parathyroid gland can affects heart and vascular development by downregulating G protein-coupled receptors in myocardial endothelial cells [33]. There are two enriched genes for the positive regulation of bicellular tight junction assembly, and genome-wide CNV analysis supports that NPHP1 affects cardiovascular development and functional pathways [34]. Zhang et al. [35] recently showed that claudin-5 (CLDN5) downregulation leads to myocyte atrophy and myocardial dysfunction. For complex CHD, the top three most significantly enriched terms were ncRNA 3′-end processing, snRNA 3′-end processing, and snRNA processing. SnRNAs are an abundant class of basic non-coding small RNAs in the eukaryotic nucleus and cytoplasm and plays a crucial role in pre-mRNA splicing. Therefore, abnormal processing of snRNAs in cardiomyocytes may affect heart ontogenesis and development [36].

As shown in Figure 6, among the significantly enriched KEGG pathways, the cell adhesion molecule pathway was involved in the development of both isolated and complex CHD. The significantly enriched pathways in isolated CHD were related to bacterial invasion of epithelial cells and SNARE interactions in vesicular transport. Early studies have shown that vesicular transport in the myocardium can occur from the capillary lumen through endothelial cells into the intercellular space and then into cardiomyocytes through endocytosis [37]. Peters et al. [38] demonstrated the presence of multiple SNARE proteins in cardiomyocytes and revealed that VAMP-1, VAMP-2, and synaptotagmin-1 are localized to subpopulations of secretory granules containing atrial natriuretic peptide in atrial myocytes. In contrast, the significantly enriched pathways in complex CHD were mainly associated with nucleocytoplasmic transport, mRNA surveillance pathways, and leukocyte transendothelial migration. Zhang’s research [39] confirmed two mutation sites located in RNA-binding motif 20, a splicing factor mainly expressed in the heart that regulates the splicing process, promotes nucleocytoplasmic transport, and the formation of protein-RNA condensates. The abnormal process subsequently leads to cardiac dysfunction and abnormal protein localization. A large-sample study reported a correlation between the high expression of circulating exogenous RNAs and left ventricular remodeling, they demonstrated that three microRNAs (miR-20a, miR-106b, and miR-17) target a common set of messenger RNAs that participate in cardiomyocyte biological processes through specific pathways, such as growth/cell cycle and apoptosis [40]. A Turkish study showed that wild-type genotypes and alleles of JAM-A rs790056 (TT genotype and T allele) and LFA-1 rs8058823 (AA genotype and A allele) are risk factors for CHD [41]. A possible mechanism is that the 3′ untranslated region (3′-UTR) gene region (rs8058823) regulates LFA-1 expression and mediates mRNA stability, degradation, and subcellular localization [42].

Strengths and limitations

To our knowledge, genetic etiology represents a critical factor in the development of the CHD, molecular mechanisms underlying CHD are complex and essential for understanding cardiac development. In current study, GO annotation and KEGG pathway analysis indicated intricate molecular pathways in complex CHD, which contributes to provide a theoretical foundation for understanding the genetic mechanisms underlying CHD.

However, there are some limitations in the study. Firstly, the sample size was not sufficiently large, and it is necessary to recruit multicenter CHD samples to expand our association analysis between genetic factors and CHD. Secondly, our findings did not examine gene variations related to the CHD cases, which would be further investigated by our team in future. Lastly, the analysis of CHD KEGG pathways requires further investigation of the potential molecular mechanisms of CHD using animal or functional experiments.