Integrative Bioinformatics Analysis Revealed Mitochondrial Defects Underlying Hypoplastic Left Heart Syndrome

Introduction

Hypoplastic left heart syndrome (HLHS) remains one of the most complex congenital cardiac diseases (CHD) and is characterized by the hypoplasia of the left ventricle, the stenosis or atresia of the mitral and aortic valve, and the underdevelopment of the aorta. Because the left ventricle and its components are incapable of supporting the systemic circulation, the disease is life-threatening without surgical intervention.¹ A three-stage surgical palliation that includes the final Fontan surgery has revolutionized the surgical treatment with HLHS and made the survival with HLHS now possible.²

HLHS has been uniformly considered as a polygenic disease with a multifactorial inheritance pattern.^1,3 Recent studies have identified a number of transcription factors (TFs) that play a fundamental role in the cardiac development, such as NK2 homeobox 5 (NKX2.5) and notch receptor 1 (Notch1).³ And somatic mutations in these TFs are also recognized as causal factors for HLHS.^3,4 Dysregulation of microRNAs (miRNAs) has been proved as related to the occurrence of CHD in a variety of studies.⁵ The down-regulation of miR-592 has been demonstrated to inhibit the Notch signaling by up-regulating the potassium channel tetramerization domain containing 10 (KCTD10) expression, thereby protecting the mice from the hypoplastic heart and CHD.⁶ Liu et al found that the deletion of miR-133a resulted in the ectopic gene expression of the cardiac smooth muscle and abnormal cardiomyocyte proliferation.⁷ Through single-cell RNA profiling of hiPSC-derived endocardium and human fetal heart tissue with an underdeveloped left ventricle, Miao et al had identified a developmentally impaired endocardial population in HLHS.⁸ In addition, HLHS is also correlated with the Turner syndrome, trisomy 18 syndrome, DiGeorge syndrome, Jacobsen syndrome, Noonan syndrome, etc., suggesting the complexity of its underlying mechanisms.³

While HLHS patients have improved prognosis, 10-year transplant-free survival stands at only 39–50% for HLHS patients.^9,10 The high morbidity/mortality is largely associated with ventricular dysfunction and acute heart failure (AHF).¹⁰ The Phase I TICAP (Transcoronary Infusion of Cardiac Progenitor Cells in Patients With Single-Ventricle Physiology) trial (NCT01273857)^11,12 and Phase II PERSEUS randomized, controlled trial (NCT01829750) enrolled patients with various single-ventricle conditions-including HLHS and they were treated by intracoronary infusion of cardiosphere-derived cells (CDCs) after palliative operations. An improvement in ejection fraction of the RV, somatic growth, reduced heart failure status, and quality of life were observed in treated patients compared with those in controls.¹³ However, therapies that have been developed for HF in adults are still inefficient for treating HF in HLHS.¹⁴ The detailed molecular mechanisms underlying the HF in HLHS is to a large extent elusive despite recent advancements.¹⁵ A prior work focused on the peripheral blood mononuclear cells (PBMC) oxygen consumption rate (OCR) measured from single-ventricle congenital heart disease developing HF demonstrated the mitochondrial respiration defects including higher maximal respiratory capacity and respiratory reserve.¹⁶ Another study by Xu et al suggested that intrinsic mitochondrial dysfunction is linked with cardiac dysfunction and heart failure risk in HLHS. In this study, the induced pluripotent stem cells (iPSC) were established from HLHS patients and differentiated into cardiomyocytes (iPSC-CM). These HLHS patient iPSC-CM demonstrated the reduced mitochondrial membrane potential as well as diminished mitochondrial oxygen consumption rate.¹⁷ Hence, comprehensively exploring the novel therapeutic strategies is urgently warranted in order for effective treatments of HF in HLHS.³

Bioinformatics analysis provides important clues in understanding the molecular mechanisms of diseases, exploring the novel biomarkers related to the disease diagnosis and prognosis, and investigating the possible therapeutic targets in the early stages of diseases. A previous report by Liu and his colleagues on the HLHS mouse heart tissue suggested a mitochondrial maturation defect, and another study by Ricci identified the alternative mRNA splicing patterns in the pathogenesis of HLHS based on GSE23959. In HLHS, over 1800 mRNAs were differentially spliced and the most significant alterations in KEGG pathways involved the oxidative phosphorylation in mitochondria.^16,18 To further unravel the molecular basis underlying the HF in HLHS, integrative bioinformatics analysis was performed to unravel the underlying genes and mechanisms involved in HLHS. A previous report by Liu and his colleagues on the HLHS mouse heart tissue suggested a mitochondrial maturation defect and another study by Xu et al demonstrated intrinsic mitochondrial dysfunction linked with cardiac dysfunction and heart failure risk in HLHS.^16,17 Of note, our results revealed the mitochondrial-related key genes, regulatory pathways, TFs and miRNAs underlying HLHS, which added to deepen the understanding of mitochondrial maturation and function defects in HF in HLHS.

Materials and Methods

Microarray Data

The gene expression profile data GSE23959¹⁸ based on the platform of GPL5188 (Affymetrix Human Exon 1.0 ST Array) was obtained from the Gene Expression Omnibus (GEO) database, National Center for Biotechnology Information (NCBI) (https://www.ncbi.nlm.nih.gov/geo/). The dataset available in this analysis was uploaded by Ricci et al, which includes 16 samples containing 10 healthy controls and 6 patients with HLHS. And the HLHS neonates were diagnosed based upon clinical features including hypoplasia/atresia of the ascending aorta, various degrees of underdevelopment of the aortic valve, mitral valve, and left ventricle (LV) cavity, and retrograde flow in the aortic arch as determined by conventional 2-D echocardiography. The myocardial samples were isolated from the right ventricles (RVs) of 6 HLHS neonates and all subjects underwent the stage 1 Norwood reconstruction. As for the control samples, the myocardial tissue of RVs and left ventricles (LVs) were obtained from 5 newborns each newborn with normal cardiac anatomy, who died due to non-cardiac causes. An overview of the detailed information on the samples analyzed in our study is shown in Table 1.

Table 1 Details of Clinical Samples from GSE23959

Identification of Differentially Expressed Genes (DEGs)

The Perl script was applied to convert the gene IDs to official gene symbols after downloading the probe expression matrix file GSE23959 series matrix.txt. If multiple probes match the same gene, the mean value of probes is calculated as the final value of this gene. The dataset was then normalized using the Normalized Between Arrays function of the Limma package in RStudio software (Version 1.2.5001).¹⁹ After processing, the Limma package was applied to screen the DEGs in this dataset.¹⁹ The DEGs were screened out following the standard of a corrected p value <0.05 and |log₂ fold change (FC)| ≥1.00. Volcano and heatmap were constructed using the RStudio software with the ggplot2 package and pheatmap package, respectively.^20,21 A list of DEGs including the up-regulated and down-regulated genes was saved for the following integration analysis.

Functional and Pathway Enrichment Analysis of DEGs

Metascape (http://metascape.org) is an integrated online tool providing comprehensive gene annotations and analyses for input list of genes.²² To explore the biological functions of DEGs in HLHS, Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were performed using Metascape with the criteria of minimum overlap >3, p value cut off <0.01, and minimum enrichment score >1.5. And the results were visualized by Hiplot, a comprehensive web platform for scientific data visualization (https://hiplot.com.cn).²³

PPI Network Construction and Analysis of Modules

The Search Tool for the Retrieval of Interacting Genes (STRING) online database (https://string-db.org/) (version: 11.0), which provides the direct and indirect associations of proteins or genes, was utilized to detect further details of molecular interactions of DEGs.²⁴ Medium confidence (minimum required interaction score >0.4) was chosen as a threshold and all positive interactions were included. Subsequently, the PPI network was visualized by Cytoscape software (vision 3.8.2) (https://cytoscape.org/).²⁵ Then, a Cytoscape plugin known as the Molecular Complex Detection (MCODE) was applied to find out significant modules with degree cutoff = 2, node score cutoff = 0.2, k-Core = 2, max. Depth = 100 as a filter criterion.²⁶ The modules with a MCODE score greater than 4 and containing more than five nodes were regarded as the key modules. Moreover, the overlapping genes were identified as hub genes according to the score calculated by three topological algorithms through the cytoHubba plugin of Cytoscape including Maximum Correlation Criterion (MCC), Density of Maximum Neighborhood Component (DMNC) and Maximum Neighborhood Component (MNC).²⁷ The enrichment analysis was further performed for the key modules and hub genes through Metascape, respectively.²²

TF-DEGs and miRNA-DEGs Regulatory Network Construction

The TFs fine-tune the downstream target genes at the pre-transcriptional stage and play a key role in HLHS. The TRRUST (version 2) is a manually curated database of human and mouse transcriptional regulatory networks, of which the TF-targeted regulatory interactions have been verified experimentally.²⁸ The TRRUST was used to discover the TFs for DEGs and the interaction pairs whose FDR < 0.05 were visualized in Cytoscape. The target miRNAs of DEGs were predicted using a miRNA-centric network visual analytics platform named miRNet (version 2.0), which integrates data from 14 different miRNA databases.²⁹ After obtaining the interactions between all DEGs and related microRNAs, the interaction pairs with the number of connected target genes in the top 10 were visualized in Cytoscape.

Results

The Identification of DEGs in HLHS

The gene expression profile of the GSE23959 was normalized and shown in Figure 1A. A total of 17,513 genes were detected in the myocardial samples, of which 135 up-regulated genes (adjusted p-value <0.05, log2 (Fold Change) >1.00) and 75 down-regulated genes (adjusted p-value <0.05, log2 (Fold Change) < −1.00) were identified as significant DEGs. Compared to the normal controls, the distribution of DEGs in HLHS patients was manifested in a volcano plot (Figure 1B). The hierarchical clustering analysis revealed that the DEGs were well clustered between HLHS tissues and normal tissues, as shown in the heatmap (Figure 1C). The details of 210 DEGs are presented in Supplementary Table S1.

Figure 1 The identification of DEGs between the HLHS patients and normal controls. The box plot of gene expression data after normalization (A). The volcano plot of genes detected in HLHS in which the red dots represented the up-regulated genes and the blue dots represented the down-regulated genes (B). The heatmap of DEGs (adjusted p-value<0.05 and |log2(FC)|≥1.00) in which the up-regulated genes were in red color and the down-regulated genes were in blue color (C). Abbreviations: DEG, differentially expressed gene; HLHS, hypoplastic left heart syndrome; FC, fold change.

The Functional and Pathway Enrichment Analysis

The up-regulated and down-regulated DEGs were uploaded to Metascape to further analyze their crucial biological functions, and the top ten enrichment results of each term were shown in Supplementary Table S2. The DEGs were mainly enriched in the generation of precursor metabolites and energy, cofactor metabolic process, monocarboxylic acid metabolic process, cellular amino acid metabolic process and response to decreased oxygen levels, etc., by the biological process (BP) analysis (Figure 2A). The results of cellular component (CC) in GO showed that the DEGs were mainly involved in mitochondrial envelope, mitochondrial matrix, ficolin-1-rich granule, sarcolemma and mitochondrial proton-transporting ATP synthase complex and etc., of which most were mitochondrial composition correlated (Figure 2B). For the molecular function (MF) group, oxidoreductase activity, cofactor binding, protein homodimerization activity, catalytic activity, acting on a tRNA, proton-transporting ATP synthase activity and rotational mechanism and etc., were the enriched terms (Figure 2C). Additionally, KEGG pathway analysis indicated that the DEGs were abundant in oxidative phosphorylation, carbon metabolism, valine, leucine and isoleucine degradation, cardiac muscle contraction, renal cell carcinoma, etc. (Figure 2D, Supplementary Table S3). The results above indicated that the DEGs were mainly enriched in the mitochondria-related biological processes, subcellular components, molecular functions and signaling pathways.

Figure 2 The GO and KEGG pathway enrichment analysis of DEGs. The top ten biological process (A), cellular component (B), molecular function (C) and KEGG pathway (D) of DEGs. Abbreviations: GO, gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; DEG, differentially expressed gene.

The PPI Network Construction and Analysis of Modules

To assess the relationship between the DEGs, the STRING online database was utilized for further analysis. A total of 155 DEGs, including 104 up-regulated genes and 51 down-regulated genes, were finally filtered into the PPI network. The PPI network involved 155 nodes and 363 edges with a local clustering coefficient of 0.412 and the PPI enrichment p-value < 1.0 × 10⁻¹⁶ (Figure 3). And 15 hub genes, such as NDUFB6, UQCRQ, SDHD, ATP5H, were identified based on three topological analysis methods.

Figure 3 The PPI network of DEGs constructed using the Cytoscape, including 155 nodes and 363 edges. The up-regulated genes were depicted as the circles and the down-regulated genes as the squares. The red and large node showed a high degree, and the yellow and small node showed a low degree. Abbreviations: PPI, protein–protein interaction network; DEG, differentially expressed gene.

Subsequently, the MCODE plugin of Cytoscape software, which detects densely connected regions in large PPI networks, was applied to explore significant modules. A total of ten modules were generated, and only two modules with a MCODE score greater than four and containing more than five nodes were screened out. The module 1 showed 13 nodes linked via 76 edges with an MCODE score of 12.667 and UQCR10 (ubiquinol-cytochrome c reductase, complex III subunit X) was regarded as seed gene. All DEGs of module 1 were up-regulated and GO and KEGG analysis showed that these genes projected the oxidative phosphorylation, mitochondrial inner membrane and respiratory chain complex, etc., as top terms, which were all mitochondrial components and functions related (Figure 4A and C and Supplementary Table S4). The module 2 was consisted of 5 nodes and 9 edges with an MCODE score of 4.500 and MRPS28 (mitochondrial ribosomal protein S28) was considered as seed gene. The enrichment analysis showed that these up-regulated genes were abundant in the ribosomal subunit, mitochondrial translational elongation and large ribosomal subunit, and mitochondrial components and functions were still the most relevant (Figure 4B and D and Supplementary Table S4).

Figure 4 The network analysis of DEGs. The clustered module 1 identified from the whole PPI network (A). The enrichment analysis of the module 1 by the Metascape (C). The clustered module 2 identified from the whole PPI network (B). The enrichment analysis of the module 2 by the Metascape (D). The Venn diagram showing the number of overlapping genes among the three topological analysis methods (E). The interactions among the fifteen hub genes (F). The enrichment analysis of the hub genes by the Metascape (G). Abbreviations: DEG, differentially expressed gene; PPI, protein–protein interaction network.

Furthermore, 3 topological analysis methods were adopted to identify the hub genes in PPI network. Fifteen hub genes were obtained after overlapping the genes according to these ranked means in cytoHubba plugin (Figure 4E and F). And the details of these hub genes are shown in Table 2. According to the enrichment results, the hub genes were mainly correlated with the oxidative phosphorylation, oxidoreductase complex, proton-transporting ATP synthase activity and mitochondrial respiratory chain complex III. Similar to the results of enrichment of modules, the hub genes were abundant in the biological processes and molecular functions of mitochondria (Figure 4G and Supplementary Table S5).

Table 2 Hub Genes Identified Through Three Topological Analysis Methods

The TF-DEGs and miRNA-DEGs Regulating Network Analysis

TF-DEGs regulating network was constructed with the help of TRRUST exploring the relationships between TFs and DEGs (Figure 5). As a result, a total of 85 interactions among 23 TFs, 10 up-regulated DEGs and 19 down-regulated DEGs were identified. In this network, both NFKB1 and RELA possessed the similar maximum quantity of regulatory genes. NFKB1 regulated 8 DEGs such as VEGFA, NAMPT and IGF2BP2, while RELA regulated 8 DEGs such as ABCA1, CD74 and NPPB. Additionally, the top targeted DEGs for TFs was VEGFA that was regulated by 10 TFs and HLA-DRA that was regulated by 5 TFs. Interestingly, both NFKB1 and RELA belongs to the NF-κB family and this indicated the possible involvement of this canonical signaling pathways in mitochondrial dysfunction in HLHS.^30–32

Figure 5 TF-target DEGs regulatory network. The green and red nodes for the DEGs, and the blue diamonds for the TFs. Abbreviations: TF, transcription factor; DEG, differentially expressed gene.

Then, a total of 9062 interaction pairs were predicted, including 1949 miRNAs and 204 DEGs. The miRNAs were ranked according to the number of target genes and the top ten were screened out, and the relation pairs were finally visualized in Cytoscape (Figure 6). hsa-mir-1-3p had the highest connectivity with 119 target DEGs. Besides, hsa-mir-124-3p and hsa-mir-16-5p had 99 pairs and 87 pairs with DEGs, respectively. Of note, both hsa-mir-124 and hsa-mir-16 have been reported to mediate the NF-κB signaling and mitochondrial functions.^33–35 Thus these above-mentioned results should promote further in-depth investigation of HLHS mechanism by which specific miRNAs dysregulates the mitochondrial dysfunction via NF-κB signaling.

Figure 6 MicroRNA-target DEGs regulatory network. The green and red nodes for the DEGs, and the pink V-shapes for the miRNAs. Abbreviations: miRNA, microRNA; DEG, differentially expressed gene.

Discussion

In this study, a total of 210 DEGs including 135 up-regulated genes and 75 down-regulated genes were identified. From the results of enrichment analysis of DEGs, multiple terms in BP, CC, MF and KEGG pathway were found to be associated with mitochondrial components and functions. Generation of precursor metabolites and energy, cofactor metabolic process, cristae formation, aerobic respiration, glucose import across plasma membrane, response to decreased oxygen levels, mitochondrial gene expression, mitochondrial electron transport and ubiquinol to cytochrome c were the top enriched terms in BP. The results of CC showed that the DEGs were mainly related to the mitochondrial envelope, mitochondrial matrix, mitochondrial proton-transporting ATP synthase complex, mitochondrial respiratory chain complex III, mitochondrial intermembrane space and intrinsic component of mitochondrial inner membrane. And the oxidoreductase activity, proton-transporting ATP synthase activity, rotational mechanism, cofactor binding and NADH dehydrogenase activity were found to be the most important MFs. The KEGG pathway analysis indicated that oxidative phosphorylation and carbon metabolism, ubiquinone biosynthesis, eukaryotes, 4-hydroxybenzoate ≤ ubiquinone were predominantly enriched.

Similar results were obtained in the subsequent analysis of modules and hub genes. The enrichment results of modules included the oxidative phosphorylation, mitochondrial inner membrane, respiratory chain complex, mitochondrion organization, mitochondrial proton-transporting ATP synthase complex, coupling factor F(o), mitochondrial respiratory chain complex III, and mitochondrial translational elongation. The oxidative phosphorylation, oxidoreductase complex, proton-transporting ATP synthase activity, rotational mechanism, mitochondrial respiratory chain complex III and mitochondrial matrix were found to be the main results of enrichment of hub genes.

Accumulating evidence has pointed towards the mitochondrial maturation and function defects in the pathogenesis of HF in HLHS. Ricci et al have previously identified the DEGs and alternative mRNA splicing patterns in the pathogenesis of HLHS. In HLHS, over 1800 mRNAs were differentially spliced and the most significant alterations in KEGG pathways involved the oxidative phosphorylation in mitochondria.¹⁸ A previous report by Liu and his colleagues on the HLHS mouse heart tissue suggested a mitochondrial maturation defect of fewer cristae, changed shape and decreased size. And the transcriptome profiling showed the metabolic and mitochondria-related pathways among the top impacted pathway.¹⁶ Another study by Xu et al demonstrated the intrinsic mitochondrial dysfunction is linked with cardiac dysfunction and heart failure risk in HLHS. In this study, the induced pluripotent stem cells (iPSC) were established from HLHS patients and differentiated into cardiomyocytes (iPSC-CM). These HLHS patient iPSC-CM demonstrated the reduced mitochondrial membrane potential as well as diminished mitochondrial oxygen consumption rate. Importantly, the patients with CM mitochondrial and differentiation defects had poorer clinical prognosis with severe ventricular dysfunction.¹⁷ A prior work focused on the peripheral blood mononuclear cell (PBMC) oxygen consumption rate (OCR) measured from single-ventricle congenital heart disease developing HF demonstrated the mitochondrial respiration defects including higher maximal respiratory capacity and respiratory reserve.¹⁶ A recent report using iPSC-CM generated from HLHS patients showed uncompensated mitochondrial-mediated oxidative stress underlying early HF in HLHS. Early-HF patient iPSC-CM showed the mitochondrial permeability transition pore (mPTP) opening, mitochondrial hyperfusion and respiration defects, which were associated with early-HF and poorer outcomes.³⁶

Our results found that seed and hub genes were mostly involved in the mitochondrial functional and metabolic pathways. UQCR10 has been involved in the mitochondrial oxidative phosphorylation and myocardial contraction in cardiomyocytes.³⁷ MrpS28 encodes a mitochondrial ribosomal protein, which is essential for mitochondrial ribosomal assembly, mitochondrial translation and oxidative phosphorylation.³⁸ NDUFB6, a mitochondrial gene, plays a major role in reactive oxygen species (ROS) production.³⁹ UQCRQ has been associated with the mitochondrial dysfunction and has been found to be up-regulated in mice with ischemic and dilated cardiomyopathy.^40,41 As one of the four subunits of mitochondrial respiratory complex II, SDHD regulates reserve respiratory capacity and cell survival in cardiac myocytes.^42,43 And the expression of ATP5H was apparently increased in the myocytes with mitochondrial dysfunction induced by fluoride.⁴⁴

Among the predicted TFs in this study, some have been correlated with cardiac development and function. Both NFKB1 and RELA are members of the Rel/NFKB family of TFs and are involved in the regulation of immunity, inflammation, cell proliferation and apoptosis.⁴⁵ Zhang et al found that the functional-94 insertion/deletion ATTG polymorphism in the promoter of NFKB1 is associated with the susceptibility to CHD.⁴⁶ The RELA subunit of NF-κB is essential for the transition from the committed precursors into mature cardiomyocytes, and activation of TLR3/NF-κB pathway enhances the myocardial maturation.⁴⁷ This indicated the possible involvement of these canonical signaling pathways in this complex CHD.^48,49

MiRNAs are a class of small single-stranded RNA containing 19–24 nucleotides and fine-tunes the post-transcriptional gene expression. The dysregulation of miRNA expression has been found in a variety of cardiovascular diseases.⁵⁰ The top 10 miRNAs were screened using the miRNet online tools, including hsa-let-7b-5p, hsa-mir-16-5p, hsa-mir-26a-5p, hsa-mir-27a-3p, hsa-mir-34a-5p, hsa-mir-1-3p, hsa-mir-124-3p, hsa-mir-191-5p, hsa-mir-155-5p, hsa-mir-129-2-3p. As a member of the miR-155 family, hsa-miR-155-5p have been used as a biomarker to distinguish the patients with hypertrophic cardiomyopathy from the normal controls.^51,52 Moreover, hsa-miR-155-5p could regulate the mitochondrial biogenesis by regulating the expression level of TFAM.^53,54 Hsa-miR-191-5p regulates the expression of PGC-1α, which is a key molecule in the process of mitochondrial production.⁵¹ Interestingly, both hsa-mir-124 and hsa-mir-16 have been demonstrated to regulate the NF-κB signaling and mitochondrial functions. In 2020, Chen et al revealed that long non-coding RNA NEAT1 could inhibit miR-124/NF-κB signaling pathway and regulate the expression of inflammatory factors to protect cardiomyocytes from As₂O₃ damage. It has also been proved that miR-124 was involved in the regulation of mitochondrial apoptosis in a variety of diseases.^55–57 And Zhao et al found that miR-124 could inhibit the expression of CD151, leading to the aggravation of heart failure.⁵⁸ In the apoptosis of neutrophils exposed to Pb, miR-16-5p activated the mitochondrial apoptotic pathway and death receptor pathway by down-regulating the expression of PiK3R1 and IGFR1.⁵⁹ Besides, miR-16 could mediate the expression of CD40 and inhibit the NF-κB signaling pathway to alleviate LPS-induced myocardial cell injury.⁶⁰ Thus, these above-mentioned results should promote further in-depth investigation of HLHS mechanism by which specific miRNAs dysregulate mitochondrial dysfunction via NF-κB signaling.

Limitations

There were several limitations that should be highlighted to interpret the results. First, confounding factors such as age, gender, BMI, and differences between the LV and RV of the samples were not carefully considered in our study. The results of this study must be interpreted with caution. Second, the HLHS group and the healthy group were not well matched to age, and some results may be due to differences in age. Third, the enrichment results for up- and down-regulated genes were not performed separately, and we might miss some of the valuable information. Fourth, due to the small number of clinical samples included in GSE23959, there may be selection bias in the study design and the sample size needs to be further expanded. Lastly, in further studies, HLHS patients with baseline HF should be considered to explore the mechanisms of HF.

Conclusion

In summary, a comprehensive analysis of DEGs was performed for HLHS, and the results demonstrated that the mitochondrial structure and function may be particularly relevant in the pathophysiological mechanisms of HF in HLHS. In addition, the network analysis detected 15 hub genes, 23 TFs and 10 miRNAs closely related to HLHS. These results have provided the evidence of novel mechanisms as well as promising therapeutic targets for HLHS and await in-depth investigation in the future.