This study enrolled a pedigree with complex CHD and 1,792 patients with sporadic CHD at the Children’s Hospital of Fudan University. All procedures were conducted in alignment with the principles of the Declaration of Helsinki (2013 revision). Written informed consent was obtained from the parents or legal guardians of all participating patients. The study protocol received approval from the Ethics Committee of the Children’s Hospital of Fudan University [No. (2021)429].

Genomic DNA was extracted from peripheral blood samples of the patients and their parents using the QIAamp DNA Blood Mini Kit (QIAGEN, Germany). WES was performed on the proband and her parents by Gemple Biotech Co., Ltd. (Shanghai, China). Targeted sequencing of PDIA4 was conducted on cases of sporadic CHD. The criteria for screening rare and deleterious variants were as follows: (1) loss-of-function variants; (2) missense mutations with absence or minor allele frequency ≤ 0.01% in gnomADv2_exome_EAS and a combined annotation dependent depletion (CADD) score >20. Allele frequencies from healthy controls were obtained from East Asians in the Genome Aggregation Database Version 2 (gnomADv2), accessible at https://gnomad.broadinstitute.org/. The same filtering criteria were applied to both patient and control datasets to ensure consistency.

Reference control data were sourced from GnomADv2, which includes 125,748 exome sequences. Considering ethnic specificity, we restricted our analysis to East Asian individuals (n = 9,197) from gnomADv2 as the control group. The filtering criteria for rare deleterious variants in the control population were consistent with those applied to our CHD patients.

HEK293T cells were cultured in Dulbecco’s modified Eagle’s medium (1×) (Gibco, United States) containing 10% fetal bovine serum (Excell, Uruguay). HUVECs were maintained in endothelial cell medium (ScienCell, United States). All cells were incubated at 37 °C and 5% CO₂. For passaging, cells were washed with phosphate-buffered saline (PBS) (Biosharp, China) and then detached using 0.25% trypsin-EDTA (Gibco, United States)

Short hairpin RNA (shRNA) sequences targeting PDIA4 were designed and synthesized by Genomeditech Company and subsequently inserted into PGMLV lentiviral vectors (Supplementary Table S1). The constructed lentiviral vector contained a puromycin resistance gene for the selection of stable PDIA4-knockdown cell lines. HEK293T cells were seeded into 10-cm dishes and grown to reach 80% confluence. They were then transfected with a mixture of the constructed PGMLV-PDIA4-shRNA plasmid (6.8 μg) and the lentiviral packaging plasmids PMDL2 (3.4 μg) and PSPAX (5.2 μg), using Lipofectamine 3000 transfection reagent following the manufacturer’s guidelines. Viral supernatants were collected 48 h after transfection and concentrated according to standard protocols. HUVECs were seeded in 6-well plates and infected with the concentrated lentiviral supernatant when they reached approximately 90% confluence. After 72 h, infected cells were selected with puromycin (Beyotime, China) for 2 weeks. Knockdown efficiency was ultimately confirmed by Western blot analysis.

Total RNA was extracted using TRIzol reagent (Thermo Fisher Scientific, United States). RNA was reverse-transcribed into complementary DNA (cDNA) using the PrimerScript™ RT Reagent Kit (Takara, Japan). Quantitative polymerase chain reaction (PCR) (qPCR) was performed using TB Green® Premix Ex Taq™ (TaKaRa, Japan) on a QuantStudio Real-Time PCR System (Thermo Fisher Scientific, United States), according to the manufacturer’s guidelines. Primer sequences are listed in Supplementary Table S1.

HUVECs were lysed in protein lysate (Beyotime, China) containing 50× protease inhibitor (Beyotime, China), 50× salubrinal (Beyotime, China), and 50× EDTA (Beyotime, China). Lysates were centrifuged, and protein concentrations were measured using a BCA Protein Assay Kit (TaKaRa, Japan). Protein samples, boiled after mixing with 5 x loading buffer, were separated by 7.5% SDS-polyacrylamide gel (Epizyme, China) and subsequently transferred to a PVDF membrane (Millipore, United States). The membrane was blocked with 5% non-fat milk at ambient temperature for 2 h and then incubated at 4 °C overnight with primary antibodies against PDIA4 (Proteintech, 1D5F3) at 1:5000, β-catenin (CST, 8480) at 1:1000, lamina/c (CST, 2032) at 1:1000, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (CST, 2118) at 1:1000. After six washes with TBST (Beyotime, ST673), it was incubated with secondary antibodies (CST, 7074/7076) diluted 1:2000 for 2 h at room temperature. Finally, protein bands were visualized with the ChemiDoc Imaging System (Bio-Rad XRS+).

Cell proliferation was assessed using the Cell Counting Kit-8 (CCK8, Dojindo, Japan). HUVECs were seeded at a density of 8000 cells per well in 96-well plates. CCK8 solution (10 µl) was added to each well at 0 h, 24 h, 48 h, and 72 h. Absorbance at 450 nm was measured after incubation for 2 h.

Cell proliferation was assessed using the BeyoClick 5-ethynyl-2′-deoxyuridine (EdU)-555 Cell Proliferation Detection Kit (Beyotime, China). HUVECs were seeded in 6-well culture plates and cultured to approximately 80% confluence. The EdU working buffer was added to each well, followed by a 2-h incubation. Fluorescently labeled cells were imaged using a fluorescence microscope. ImageJ software was utilized to analyze the final data.

HUVECs were seeded into 6-well culture dishes and cultured to full confluence. A 10 µl pipette tip was used to form a wound. Floating cells were removed by washing twice with PBS, and cells were then incubated in serum-free medium. Cell migration dynamics were documented and quantified by measuring the residual wound area from microscopic images captured at 0 h and 8 h with ImageJ software.

Total RNA was extracted from three independent biological replicates with TRIzol reagent (Invitrogen, United States), following the manufacturer’s instructions. The subsequent reverse transcription and high-throughput sequencing procedures were performed by Novogene Company (Shanghai, China), following standard service protocols.

Cells were harvested using the same procedure used for passaging and collected as a cell pellet. Cytoplasmic protein extraction reagent A (200 μL; Beyotime, P0028) was added to each 20 μL cell pellet, followed by vortexing for 5 s and incubation on ice for 10 min. Cytoplasmic protein extraction reagent B (10 μL; Beyotime, P0028) was then added, and the samples were vortexed for an additional 5 s and incubated on ice for 1 min. The lysates were centrifuged at 12,000 × g at 4 °C for 5 min, and the supernatants containing cytoplasmic proteins were collected.

The remaining pellet was resuspended in 50 μL of nuclear protein extraction reagent (Beyotime, P0028), vortexed vigorously for 15–30 s, and incubated on ice for 30 min, with intermittent vortexing every 1–2 min. Following centrifugation at 12,000 × g at 4 °C for 10 min, the supernatants containing nuclear proteins were collected for subsequent analysis.

All data are presented as the mean ± standard deviation (SD) from three independent experiments. Statistical analyses were performed using GraphPad Prism 9.5. Student’s t-test and Pearson’s chi-squared test were applied as appropriate. p < 0.05 was considered statistically significant. Figures were generated using GraphPad Prism 9.5.

To investigate potential genetic variants in CHD, we performed WES on an infant with complex CHD, including tetralogy of Fallot (TOF), mesocardia, ventricular septal defect (VSD), atrial septal defect (ASD), partial anomalous pulmonary venous return (PAPVR) of the right lung, and congenital pulmonary dysplasia, along with her unaffected parents (Figures 1C–H). We found that the proband carried a de novo mutation (NM_004911: c.1249G>A; p.V417I) in PDIA4 (OMIM: 620018), which was absent in the gnomADv2 database, indicating it is rare (Figure 1A). In silico pathogenicity prediction using multiple bioinformatics tools suggested that this mutation is potentially deleterious. MutationTaster classified it as “disease causing,” while PolyPhen-2 predicted it to be “probably damaging” with a confidence score of 0.994, and the combined annotation-dependent depletion (CADD) score was 20.7 (Table 1). Then, we collected peripheral venous blood from the family members and confirmed the de novo mutation by Sanger sequencing (Figure 1B).

The identified de novo mutation is located in exon 8 of the PDIA4 gene, resulting in the substitution of valine with isoleucine at amino acid position 417, which is primarily located within the b' domain of PDIA4 (Figures 2A,B), a non-catalytic domain potentially involved in substrate binding (Freedman et al., 2002). Evolutionary conservation analysis via the UCSC Genome Browser revealed that the amino acid at position 417 is highly conserved across diverse species, indicating its functional importance (Figure 2C). Structural modeling of PDIA4 was generated by SWISS-MODEL. Subsequent analysis in PyMOL revealed an inter-residue distance of 2.9 Å (<5 Å) between valine 417 and its contacting residues, indicating potential interactions between these residues. (Figure 2D).

To further evaluate the association between PDIA4 and CHD, we performed targeted sequencing of PDIA4 in a cohort of 1792 sporadic CHD patients. The clinical profile summary of the cohort is summarized in Supplementary Table S2, with patients classified into seven CHD subtypes (Geng et al., 2014). The results identified 19 rare deleterious PDIA4 variants that fulfilled the filtering criteria (a. loss of function; b. missense mutations: absence or minor allele frequency ≤0.01% in gnomADv2_exome_EAS, and CADD score >20) (Table 2). These variants were distributed across distinct functional domains of the PDIA4 protein. (Figure 2E). Applying identical filtering criteria, we identified 24 rare deleterious variants from the control group in the gnomADv2_exome_EAS database (Supplementary Table S3). Comparison of variant frequencies revealed a significantly higher burden of rare deleterious PDIA4 variants in CHD patients (19/3,584, 0.53%) than in controls (24/18,394, 0.13%). We found that these variants were associated with the occurrence of CHD (Person’s chi-squared test: OR: 4.08, 95% CI: 2.23–4.76, p = 7.46e−7) in Table 3. As shown in Supplementary Table S4, these variants were predominantly enriched in conotruncal defect (CTD) patients. We stratified the study cohort into a CTD group and a non-CTD group. The results demonstrated that the frequency of rare deleterious PDIA4 variants in the CTD group (14/1,386, 1.01%) was higher than that in the non-CTD group (5/1,672, 0.2) in Supplementary Table S5 (Pearson’s chi-squared test: OR: 3.40; 95% CI: 1.22–9.46; p = 0.013).

To explore the functional role of PDIA4 in HUVECs, we generated PDIA4-knockdown (shPDIA4) HUVECs using shRNA-mediated lentiviral transduction. Western blot analysis confirmed a marked reduction in PDIA4 protein expression in the shPDIA4 group compared with the scrambled shRNA-transduced negative control (shNC) (Figures 3A,B). Cell proliferation was evaluated using EdU and CCK-8 assays. The EdU assay revealed a significant decrease in the proportion of proliferating cells following PDIA4 knockdown (Figures 3C,D). Consistently, the CCK-8 assay indicated significantly reduced cell proliferation in the shPDIA4 group compared with the shNC group at 72 h (Figure 3E). A wound-healing assay was performed to assess the effect of PDIA4 deficiency on cell migration. Quantitative analysis indicated that PDIA4 deficiency significantly attenuated the migratory capacity of HUVECs at 8 h (Figures 3F,G).

Transcription profiling and functional enrichment analyses were performed in the shPDIA4 and shNC groups to explore the molecular mechanism by which PDIA4 modulates HUVEC function. A total of 3,875 genes were notably upregulated and 3,955 were markedly downregulated in the shPDIA4 group compared with the shNC group, as visualized in the volcano plot (Figure 4A). Kyoto Encyclopedia of Genes Genome (KEGG) pathway analysis of the differentially expressed genes revealed significant enrichment of pathways related to cardiac development, including the WNT signaling pathway and cytoskeleton in muscle cells (Figure 4B). Gene Ontology (GO) enrichment analysis identified biological processes related to cell proliferation and heart development (Figure 4C).

As shown in Figure 4D, a heatmap highlighted the upregulation of multiple inhibitors of the canonical WNT/β-catenin signaling pathway, including DKK2, CXXC4, and SFRP1 (Shih et al., 2007; Filleur et al., 2009; Hirata et al., 2009). RT-qPCR validation confirmed a significantly increased expression of these genes in the shPDIA4 group compared with the shNC group (Figure 4E). Given the central role of β-catenin in canonical WNT signaling, the protein level of β-catenin and its downstream effector MYC, which is associated with cell proliferation, were further examined (Wang et al., 2023a; Qu et al., 2018). Western blotting demonstrated that total cellular levels of β-catenin and MYC were downregulated in the shPDIA4 group (Figures 4F–H). In addition, the main activated sign of the canonical WNT/β-catenin signaling pathway is the accumulation of β-catenin in the nucleus (Logan and Nusse, 2004). Nuclear and cytoplasmic fractionation assays revealed impaired β-catenin nuclear translocation in the shPDIA4 group (Figure 4I), indicating suppression of WNT/β-catenin signaling activity.