Discovering new hub genes of dilated cardiomyopathy

Abstract

Aims

Dilated cardiomyopathy (DCM) has a poor prognosis and exhibits a complex and diverse aetiology and genetic profile. The genes responsible for the pathogenesis of DCM have not been fully identified. The present study aimed to explore new hub genes of DCM by mining the human DCM databases and further by experimental validation.

Methods

Two gene expression profiles of human DCM (GSE9800 and GSE120895) in the Gene Expression Omnibus (GEO) database were analysed to identify the differentially expressed genes (DEGs) (DCM vs. normal) and to obtain the common DEGs (cDEGs, between GSE9800 and GSE120895) using bioinformatic methods. The cDEGs were subjected to Gene Ontology (GO) functional enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment, and the protein–protein interaction (PPI) networks and functional modules were constructed to screen the hub genes. The screened hub genes were identified using the Online Mendelian Inheritance in Man (OMIM) dataset, and their transcription and translation levels were further verified by real‐time quantitative PCR (RT‐qPCR) and western blotting using doxorubicin (DOX)‐treated H9C2 cardiomyocytes that simulate the cellular pathology of DCM, with phosphate‐buffered saline (PBS)‐treated H9C2 cells as a normal control.

Results

A total of 47 cDEGs were screened out, and 19 DCM‐associated hub genes were identified. Among the 19 hub genes, 6 genes (NFKBIB, PSMC4, PSMD3, RAD21, PRNP and STAT2) have not yet been reported as associated with DCM. Among the six genes, NFKBIB and PRNP showed up‐regulations, whereas PSMC4, PSMD3 and RAD21 exhibited down‐regulations in their mRNA and protein expression levels in DOX‐treated H9C2 cardiomyocytes compared with the control H9C2 cells (all P < 0.05). The remaining STAT2 showed a significant up‐regulation in its protein expression (P < 0.05), while its mRNA up‐regulation did not reach a statistical significance (P = 0.1082).

Conclusions

Six new hub genes of DCM (NFKBIB, PSMC4, PSMD3, RAD21, PRNP and STAT2) were identified by bioinformatic analysis and experimental validation in this study. These hub genes or their products may potentially be new diagnostic biomarkers or therapeutic targets for DCM.

Introduction

Dilated cardiomyopathy (DCM) is a heterogeneous cardiac disease characterized by left ventricular (LV) or biventricular chamber dilatation, wall thinning and systolic dysfunction. The clinical manifestations of DCM include adverse ventricular remodelling, arrhythmias, thrombo-embolism, progressive heart failure and sudden death, which severely compromise the quality and duration of patients' survival.¹ The aetiology of DCM is diverse, including infection, autoimmunity, inflammation, genetic predisposition, chronic alcohol abuse, and exposure to drugs or toxicants. Although the prognosis of DCM has been improved over the past few years due to clinical follow-up, early familial diagnosis and continued optimization of treatment, there is still a lack of targeted therapies for each aetiology due to its heterogeneity.^2,3

With the development of precision medicine, genetic testing may benefit the accuracies of diagnosis and prognosis of DCM.⁴ However, the genes involved in the development of DCM and their exact roles have not been fully identified. DCM has a heterogeneous genetic tendency, accounting for 35% of gene mutations and involving at least 250 genes with different functions, such as genes encoding cytoskeletons, myenteric segments, nuclear membranes and sheath bodies. These genes play important roles in both the phenotype and the progression of DCM, and 19 of them are considered to be genes of high level of evidence that are recommended to be routinely used in clinical genetic evaluation of DCM.^5,6 For example, mutations in titin and lamin A/C genes, which encode myocardial structural proteins, can lead to LV wall stiffness, thinning and fibrosis and finally the development of DCM.⁷ However, the current genetic knowledge for DCM is insufficient for clinical needs. Therefore, there is still an urgent clinical need to discover new key genes of DCM and to better understand their pathophysiological functions, hence to develop new diagnostic and therapeutic strategies for DCM.

With the advent of the 21st century, bioinformatic technology has been widely used to uncover potential molecular targets for diseases and to discover differentially expressed genes (DEGs) and their possible signalling pathways associated with diseases.^8,9 For example, Gholipour et al. identified ACSBG1 and DEFA4 genes as diagnostic biomarkers in coronary heart disease by bioinformatic analysis.¹⁰ Fan and Hu identified 10 key genes that were expected to be potential biomarkers or therapeutic targets in heart failure using network analysis.¹¹ The aim of this study was to search for DCM hub genes using bioinformatics and to validate these genes by both online database analyses and experimental verification including real-time quantitative PCR (RT-qPCR) and western blotting and to obtain potential new causative genes for DCM. Our findings of new DCM hub genes may have the potential to be new diagnostic biomarkers or therapeutic targets for DCM.

Materials and methods

Microarray data analysis

The gene expression profiles of GSE9800 and GSE120895 were downloaded from the Gene Expression Omnibus (GEO) database (). The platforms for GSE9800 and GSE120895 were GPL887 and GPL570, respectively. The sequencing samples of both datasets were human myocardial tissues. The GSE9800 and GSE120895 datasets included a total of 59 DCM patient samples (12 and 47, respectively) and 17 normal cardiac samples (9 and 8, respectively).

Data processing and analysis of DEGs

The raw data were downloaded in txt format and normalized and annotated with probe information using the ‘limma’ software package of Bioconductor (). DEGs were identified with fold change (FC) >1.5 and false discovery rate (FDR) <0.05. The VENNY tool () was used to perform the intersection of the DEGs of the two datasets to obtain the overlapping DEGs [i.e., common DEGs (cDEGs)]. Here, DEGs referred to the DEGs between DCM patients and normal control subjects.

Gene Ontology (GO) and KEGG pathway analyses

In this study, cDEGs referred to the DEGs shared by the two examined datasets (GSE9800 and GSE120895). To comprehensively understand the functions of cDEGs and related signalling pathways, we used the ‘ClusterProfiler’ R package for GO functional enrichment analysis and ‘Metascape’¹² for Kyoto Encyclopedia of Genes and Genomes (KEGG) signalling pathway enrichment analysis. P < 0.05 and FDR < 0.05 were considered statistically significant.

Construction of protein–protein interaction (PPI) networks and modules and screening of hub genes

The Reactome database (Version 2014, ) provided a dataset comprising 217 249 functional interaction pairs. These interactions were aggregated from established PPI resources—BioGRID, the Database of Interacting Proteins (DIP), the Human Protein Reference Database (HPRD), I2D, IntAct and MINT—alongside gene co-expression evidence derived from high-throughput experimental platforms. These included yeast two-hybrid screening, affinity purification coupled with mass spectrometry (AP-MS) and genome-wide DNA microarray profiling.¹³ The interaction information was imported into Cytoscape software (Version 3.2.1, ) for PPI network construction. Important functional modules and hub genes in the PPI network were identified using Molecular Complex Detection (MCODE) analysis.

Identification and validation of hub genes in Online Mendelian Inheritance in Man (OMIM)

In OMIM (), we entered the keyword ‘dilated cardiomyopathy’ and looked for genes that have been reported in the literature associated with DCM. These genes were then intersected with the 19 hub genes selected by MCODE using the Venn diagram, thus to obtain the intersected genes related to DCM reported in the literature.

Cell culture and treatment

The H9C2 myocardial cell line was purchased from Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences (Shanghai, China). Doxorubicin (DOX) was purchased from Aladdin Company (Shanghai, China). H9C2 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% foetal bovine serum in a humidified atmosphere with 5% CO₂ at 37°C until cell confluency reached 80%–90%. Then the H9C2 cells were treated with 5 μM DOX for 24 h to construct a cell model of DCM^14,15 for further in vitro experiments. Negative control H9C2 cells were treated with phosphate-buffered saline (PBS) buffers. Cell viability was determined using a CCK-8 assay kit (AR1199-10, BOSTER). Cell morphology was observed under a light microscope (AE2000, Motic).

Live/dead cell staining

Cell death was evaluated using calcein AM/propidium iodide (PI) staining. H9C2 cells (4 × 10⁴/mL) were inoculated in 24-well plates and incubated for 24 h. When the cell density was confluent to 80%–90%, DOX (5 μM) was added to the 24-well plates to intervene for 24 h. Cells treated with DOX-free medium were used as a control. Then the solution was discarded, and each well was washed with PBS twice. Under light-protected conditions, cells were stained with 300 μL of a dual fluorescent dye solution containing 2 μM calcein AM and 4 μM PI. After thorough mixing, the cells were incubated with the staining solution for 30 min at room temperature. Subsequently, the dye solution was carefully aspirated and discarded, followed by two washes with PBS. Cellular viability was then assessed using an IX51 inverted fluorescence microscope (Olympus, Japan) with appropriate filter sets for live/dead cell differentiation.

Detection of intracellular reactive oxygen species (ROS)

Intracellular ROS levels were detected using the DCFH-DA fluorescent probe. Cells were washed twice with PBS. Under light-avoidance conditions, 300 μL of prepared DCFH-DA (10 μM) dye was added to each well and incubated at 37°C for 20 min, and each well was washed twice with PBS, and the intracellular production of ROS was observed using a fluorescent microscope.

RT-qPCR

Total RNA was extracted from DOX-treated H9C2 cells and negative control cells using TRIzol reagent (MF034-01, Mei5bio), followed by examination for the concentration and quality of RNA by NanoDrop One^C (Thermo Fisher Scientific), and the RNA was reverse-transcribed into cDNA using M5 Super qPCR RT kit with gDNA remover (MF166-01, Mei5bio). Then, qPCR was performed to detect the transcription levels of the seven predicted hub genes (PRKCA, NFKBIB, PSMC4, PSMD3, RAD21, PRNP and STAT2) and ANP and BNP genes in DOX-treated H9C2 cells (5 μM DOX for 24 h) and DOX-untreated H9C2 cells (negative control) using 2× M5 HiPer SYBR Premix EsTaq (MF787-01, Mei5bio) on the PCR instrument (Mx3005P, STRATAGENE). Primer sequences are shown in Table 1. Relative abundance of the transcripts was calculated using the 2^−ΔΔCt method.

Table 1 Primer sequences for real-time quantitative PCR.

Gene	Primer sequence (5′–3′)
GAPDH	F: CTCACAGCAGCATCTCGACAAGAG	R: CTCACAGCAGCATCTCGACAAGAG
ANP	F: TAGGAGACAGTGACGGACAA	R: GAAGAAGCCCAGGGTGAT
BNP	F: CAGCTCTCAAAGGACCAAGG	R: CGATCCGGTCTATCTTCTGC
PRNP	F: AGCGTGTGGTGGAGCAGATG	R: GGAAGATGAGGAAGGAGATGAGGAG
RAD21	F: ACCAGTGCTTCCAACCTCCTC	R: ATATACCACCATCATTGCCTTCTCC
PSMC4	F: GGATGAGCAGAAGAACCTGAAGAAG	R: AACTGACCAATGACCAACGGAATG
PSMD3	F: TTGCCAAGTTCAACCAAGTCCTG	R: TCATACGCACACCTGTCTTAATCAC
NFKBIB	F: GCCTCAGATACCTACCTCACTCAG	R: CTTCATCACGCAGCTCCTCTTC
STAT2	F: GCAGACTTCAGCAAGCGAGAG	R: CATCATTCCAGAGATCCTTCAGGTG
PRKCA	F: CGGACGACACGGAATGACTTC	R: GTTGCCTTCTTCATCTCCTTCTGG

Western blotting

Total protein was extracted from DOX-treated H9C2 cells and negative control cells using radioimmunoprecipitation assay (RIPA) lysis buffer (AR0102, BOSTER) containing 1 mM phenylmethylsulfonyl fluoride (PMSF) (AR1178, BOSTER) according to the instructions. The protein concentration was measured using the BCA protein quantification kit (AR1189, BOSTER). The proteins were isolated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to polyvinylidene fluoride (PVDF) membrane. Blots were probed by antibodies including anti-IKB beta (gene symbol: NFKBIB) (dilution, 1:500; Cat No. CPA1808), anti-PSMC4 (dilution, 1:500; Cat No. CMA4116), anti-PSMD3 (dilution, 1:500; Cat No. CPA1962), anti-RAD21 (dilution, 1:500; Cat No. CPA4988), anti-CD230 (gene symbol: PRNP) (dilution, 1:500; Cat No. CMA4411), anti-STAT2 (dilution, 1:500; Cat No. CPA9449) and anti-β-actin (dilution, 1:500; Cat No. bsm-33036M). The signals of protein–antibody complexes were exposed and scanned using the Bio-Rad ChemiDoc system. ImageJ was used for greyscale analysis.

Statistical analysis

Data were statistically analysed using GraphPad Prism 9.0. Data were expressed as mean ± standard deviation (SD). Comparisons between two groups were performed using Student's t test. P < 0.05 and FDR < 0.05 were considered statistically significant. Each experiment was repeated at least three times.

Results

DEGs and cDEGs in GSE9800 and GSE120895 datasets

A total of 428 DEGs were screened out from the GSE9800 dataset, including 226 up-regulated genes and 202 down-regulated genes (Figure 1A,C), and a total of 1806 DEGs were screened from the GSE120895 dataset, including 1773 up-regulated genes and 33 down-regulated genes (Figure 1B,D). The DEGs from the two datasets were then intersected by Venn diagram, and 47 cDEGs were obtained (Figure 2).

[IMAGE OMITTED. SEE PDF]

[IMAGE OMITTED. SEE PDF]

GO and KEGG enrichment analyses of cDEGs

Results of GO enrichment analysis displayed that the cDEGs were mainly enriched in thrombin-activated receptor signalling pathway, regulation of organic acid transport, response to growth hormone, regulation of amino acid transport, response to cold and so forth [biological process (BP)] (Table 2). Results of KEGG pathway enrichment analysis showed that the enriched signal pathways were mainly involved in hepatitis B, influenza A, Epstein–Barr virus infection, lipid and atherosclerosis, cAMP signalling pathway and so forth (Table 3).

Table 2 GO enrichment results of cDEGs.

ID	Description	Count	P value	FDR
GO:0070493	Thrombin-activated receptor signalling pathway	3	0.0000	0.0058
GO:0032890	Regulation of organic acid transport	4	0.0000	0.0279
GO:0060416	Response to growth hormone	3	0.0001	0.0405
GO:0051955	Regulation of amino acid transport	3	0.0001	0.0405
GO:0009409	Response to cold	3	0.0001	0.0405
GO:0034764	Positive regulation of transmembrane transport	5	0.0001	0.0405
GO:0045933	Positive regulation of muscle contraction	3	0.0002	0.0458
GO:0015911	Long-chain fatty acid import across plasma membrane	2	0.0002	0.0458
GO:0031652	Positive regulation of heat generation	2	0.0002	0.0458
GO:0034599	Cellular response to oxidative stress	5	0.0003	0.0488

Table 3 KEGG pathway enrichment results of cDEGs.

ID	Description	Count	P value	FDR
hsa05161	Hepatitis B	4	0.0001	0.0263
hsa05164	Influenza A	4	0.0001	0.0263
hsa05169	Epstein–Barr virus infection	4	0.0003	0.0298
hsa05417	Lipid and atherosclerosis	4	0.0004	0.0298
hsa04024	cAMP signalling pathway	4	0.0004	0.0298
hsa04931	Insulin resistance	3	0.0007	0.0389

PPI network of cDEGs and modules of PPI network

The PPI network of cDEGs was constructed, which had 76 nodes, 189 edges and 42 cDEGs (Figure 3A). Using degree ≥ 4 as the selection criterion, 29 significant genes were obtained. These genes were analysed using MCODE (degree cutoff = 2, k core = 2, node score cutoff = 0.2 and max depth = 100) to generate two functional modules of the PPI network, which were Module A (13 nodes, score = 4.5) (Figure 3B) and Module B (6 nodes, score = 3.2) (Figure 3C) and included a total of 19 hub genes. Module A had 13 genes, and Module B had 6 genes. Through an extensive review of the literature, we found that 13 of the 19 hub genes have been reported to be associated with DCM, including PRKCA, UBB, PPARA, NFKB1, SRC, STAT6, MYC, AKT1, STAT3, HTT, NFKBIA, JUN and EP300, while the remaining 6 hub genes (NFKBIB, PSMC4, PSMD3, RAD21, PRNP and STAT2) have not been reported associated with DCM.

[IMAGE OMITTED. SEE PDF]

Identification and validation of DCM hub genes in OMIM

The 526 DCM-related genes reported in literatures were searched through OMIM. These genes were intersected with the above 19 hub genes obtained from the MCODE analysis. One intersecting gene, PRKCA, was found (Figure 4). PRKCA has been reported to be associated with DCM, suggesting the credibility and accuracy of this study.

[IMAGE OMITTED. SEE PDF]

Cytological features of DCM in DOX-treated H9C2 cardiomyocytes

The CCK-8 assay showed that the survival rate of DOX-treated H9C2 cells was significantly decreased compared with the negative control H9C2 cells (P < 0.0001) (Figure 5A). Light microscopy showed that the morphology of control cells was normal, the nuclei were oval, and the cells were long pike-shaped and tightly arranged. DOX-treated H9C2 cells showed a DCM-like damaging morphology, such as cell crumpling, flat-round cell shape, decrease of membrane adherence ability and massive cell death (cell floating) in the culture medium (Figure 5B). Live/dead cell staining showed that a large number of DOX-treated cells were dead compared with the control cells (Figure 5C). The ROS detection assay showed that the ROS level of DOX-treated H9C2 cells was significantly higher than that of the control cells (Figure 5D). These results indicate that DOX damaged cardiomyocytes and enhanced oxidative stress; these features are consistent with the pathological changes of cardiomyocytes in DCM.

[IMAGE OMITTED. SEE PDF]

Verification of the hub genes by RT-qPCR

To determine whether the DCM cell model was successfully established, we examined the transcription levels of two representative myocardial markers of heart failure, atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP), in DOX-treated H9C2 cells by RT-qPCR. Results showed that the mRNA levels of ANP and BNP were significantly increased in DOX-treated H9C2 cells compared with the negative control H9C2 cells (Figure 6A). These results further suggest that H9C2 cardiomyocytes exhibited DCM and heart failure-like phenotypes in response to DOX, indicating that our DCM cell model was successfully established.

[IMAGE OMITTED. SEE PDF]

The transcription levels of the seven predicted hub genes (PRKCA, NFKBIB, PSMC4, PSMD3, RAD21, PRNP and STAT2) in H9C2 cells were also examined by RT-qPCR. Results showed that the mRNA levels of six predicted hub genes (PRKCA, NFKBIB, PSMC4, PSMD3, RAD21 and PRNP) were significantly altered in DOX-treated H9C2 cells compared with negative control cells (P < 0.05). Among the six genes, three (PRKCA, NFKBIB and PRNP) were up-regulated, while the other three (PSMC4, PSMD3 and RAD21) were down-regulated, in response to the DOX challenge (Figure 6B). Gene STAT2 showed a trend of up-regulation in its mRNA level but did not reach a statistical significance (Figure 6B). We adjusted the P value of RT-qPCR verification results with FDR, and all FDRs were <0.05 (Table S1). Note that the GEO database mining results (shown in Figure 3B), the OMIM database validation results (shown in Figure 4) and the RT-qPCR verification results (shown in Figure 6B) all indicated that PRKCA was a hub gene of DCM, further confirming the high credibility of our study.

Verification of the hub genes by western blotting

We used western blotting to further validate the translation levels of the six predicted new hub genes (NFKBIB, PSMC4, PSMD3, RAD21, PRNP and STAT2). Compared with the negative control cells, the protein expression levels of genes NFKBIB, PRNP and STAT2 were significantly up-regulated (P < 0.05) in DOX-treated H9C2 cells, while the protein expressions of genes PSMC4, PSMD3 and RAD21 were significantly down-regulated (P < 0.05) (Figure 7). FDR adjustments were performed on the P values; all the FDRs were <0.05 (Table S2). These FDR results were consistent with those of RT-qPCR.

[IMAGE OMITTED. SEE PDF]

Discussion

This study focused on finding some new hub genes of DCM. To our knowledge, we for the first time identified six new hub genes (NFKBIB, PSMC4, PSMD3, RAD21, PRNP and STAT2), which have not been reported to be associated with DCM. Here, we used a classic DCM cell model (H9C2 myocardial cell line treated with DOX)^14,16 to validate the hub genes screened from human DCM datasets, as DOX can induce DCM-like pathological changes, such as oxidative stress, gradual loss of myofibril and cellular vacuolation, and finally leads to heart failure.^17,18 The discovery of these new hub genes may help to find new pathological mechanisms of DCM and to develop new biomarkers or therapeutic targets for DCM.

We identified that protein kinase C alpha (PRKCA) is a hub gene of DCM; this is not a new finding but suggests that our study has a high credibility and persuasiveness because PRKCA has been proven a key gene of DCM. PRKCA is a member of the serine–threonine-specific protein kinase C family, which is involved in the regulation of cell proliferation, apoptosis, differentiation, migration, cardiac hypertrophy and inflammation.^19,20 Previous studies found PRKCA to be highly expressed in DCM,^21,22 and the present study is consistent with these previous studies.

The first identified new hub gene of DCM in this study is NFKBIB, which encodes NF-κB inhibitory protein β. We suggest that NFKBIB may be a causative gene in the development of DCM. NF-κB is a redox-sensitive transcription factor and a major transcriptional regulator of immune response, apoptosis and cell growth control genes.²³ During heart failure, NF-κB is thought to worsen cardiac remodelling or dysfunction primarily through activation and thus enhancement of pro-inflammatory pathways.^24,25 Both NFKBIA and NFKBIB are inhibitory genes of the NF-κB pathway, encoding NF-κB inhibitory proteins α and β, respectively.²⁶ Studies have shown increased activation and expression of NFKBIA in both humans and mice with DCM,^27,28 while there was yet no study to show the association of NFKBIB with DCM. Single nucleotide polymorphisms of NFKBIB have been reported to be associated with a variety of viral infections.²⁹ The expression of NFKBIB is significantly up-regulated in inflammatory breast cancer³⁰ and is also altered in some diseases such as rheumatoid arthritis.³¹ Our finding on NFKBIB as a DCM hub gene may add to the knowledge of DCM pathogenesis.

The second and third identified new hub genes of DCM in this study are PSMC4 and PSMD3. These two genes may also be causative genes for DCM. PSMC4 and PSMD3 encode ATPase proteasome 26S subunit 4 and non-ATPase proteasome 26S subunit 3, respectively. These two genes belong to the mitochondrial genes and regulate many cellular processes including protein degradation, cell cycle, apoptosis and DNA damage repair. In addition, these two genes play a key role in the maintenance of protein homeostasis, and their aberrant expressions can lead to an imbalance of protein homeostasis and mitochondrial dysfunction.^32–34 PSMC4 and PSMD3 are involved in a variety of diseases such as pulmonary hypertension and chronic granulocytic leukaemia.^32,35 In overweight/obese children, aberrant expressions of PSMC4 and PSMD3 can cause impaired proteasome efficiency, perilipin 2 (PLIN2) overexpression and atheropathy.³⁶ However, the relationship between PSMC4/PSMD3 and DCM has not been reported. We found that the mRNA levels of PSMC4 and PSMD3 were both down-regulated in the DCM cell model. This change may affect mitochondrial function and thus lead to the development of DCM. Here, we did not go deep into how PSMC4 and PSMD3 affect the development of DCM; this issue warrants further study.

The fourth identified DCM hub gene in this study is RAD21. RAD21 protein is a key component of the cohesin complex. The cohesin complex plays a key role in sister chromatid cohesion, cell cycle, DNA damage repair, transcriptional regulation and protein modification degradation.^37,38 RAD21 is a frequently amplified oncogene and encodes a subunit of cohesin complex. Aberrant RAD21 expression is closely associated with ovarian cancer and small cell lung cancer.^39,40 Some studies have shown that mutations in RAD21 can cause abnormal development of cardiac neural crest, leading to congenital heart disease.⁴¹ Till now, no study has been reported to show the relationship between RAD21 and DCM. Here, we found altered expression of RAD21 in a DCM cell model, but the exact role of RAD21 in DCM still needs further study.

The fifth identified new hub gene of DCM in this study is PRNP, which encodes prion protein (PrPC). Mutations of PRNP have been associated with a variety of disorders, including autonomic failure, Creutzfeldt–Jakob disease and Alzheimer's disease.^42,43 PrPC is involved in cell survival, oxidative stress, immune modulation, metal ion transport, cellular adhesion and transmembrane signalling.^44,45 PrPC prevents oxidative stress-induced cardiotoxicity. Mutations of PRNP can lead to increased mitochondrial hyperoxidation and altered membrane potential and dysfunction.⁴⁶ Currently, there is no report to show the relationship between PRNP and DCM. We identified that PRNP is a hub gene of DCM and showed transcription up-regulation of this gene in a DCM cell model, but the exact role of PRNP in DCM also needs further investigation.

The sixth identified new hub gene of DCM in this study is STAT2. STAT2 is a key member of the JAK/STAT signalling pathway. This pathway is important in cytokine signalling and is considered one of the central communication nodes in cellular function, regulating cell proliferation, differentiation, apoptosis, inflammation, injury, oxidative stress and so forth.^47,48 STATs are both signalling factors and transcription factors and have seven members in their family: STAT1, STAT2, STAT3, STAT4, STAT5a, STAT5b and STAT6. Enhanced phosphorylation and activation of STAT1, STAT3 and STAT6 have been observed in DCM, which in turn cause alterations in the structure and function of the left ventricle,⁴⁹ whereas STAT2 has not yet been associated with DCM. STAT2 differs from other STATs, as it is the only STAT that does not bind to the primordial γ-activation site, and the biological functions of STAT2 are mainly antivirus and immunomodulation.⁵⁰ It has been observed that STAT2 anticipated in the migration of rat cardiac fibroblasts and reduction of α-smooth muscle actin, causing the fibroblasts to display an antifibrotic effect.⁵¹ Here, we showed that the transcription level of STAT2 tended to be up-regulated in H9C2 cells in response to DOX stress but did not reach the statistical threshold (P > 0.05) (shown in Figure 6B). What is more, the translation level of STAT2 was up-regulated (P < 0.05) (shown in Figure 7). This phenomenon may be related to experimental replication numbers, individual differences, disease model or disease progression stage. Nonetheless, the role of STAT2 in DCM deserves in-depth study.

Limitations

This study mainly focused on finding new hub genes of DCM; further extensive experimental validations on these new hub genes were not performed, such as validations on the functions of these genes at animal and clinical levels. However, the study may open a window for researchers to further investigate the roles of the six identified hub genes in DCM development.

Conclusions

We for the first time identified six new hub genes of DCM, including NFKBIB, PSMC4, PSMD3, RAD21, PRNP and STAT2. These findings provide new basic information for future mechanistic exploration of DCM. These six genes or their products may also have the potential to be developed as biomarkers as well as therapeutic targets for DCM.

Conflict of interest statement

None declared.

Funding

This study was financed by the Shanxi ‘1331’ Project Quality and Efficiency Improvement Plan (1331KFC) and partially by grants from the National Natural Science Foundation of China (82170523, 82170294 and U22A6008).

References

Mao Y, Liu Y, Zhang J, Zhu Y, Chen X, Sun L, et al. A nomogram predicting atrial fibrillation in patients with dilated cardiomyopathy. Chin Med J (Engl) 2022;135:991‐993. doi:10.1097/CM9.0000000000001926

Seferovic PM, Polovina M, Bauersachs J, Arad M, Gal TB, Lund LH, et al. Heart failure in cardiomyopathies: A position paper from the Heart Failure Association of the European Society of Cardiology. Eur J Heart Fail 2019;21:553‐576. doi:10.1002/ejhf.1461

Alonso‐Villa E, Bonet F, Hernandez‐Torres F, Campuzano O, Sarquella‐Brugada G, Quezada‐Feijoo M, et al. The role of microRNAs in dilated cardiomyopathy: New insights for an old entity. Int J Mol Sci 2022;23:13573. doi:10.3390/ijms232113573

Sammani A, Baas AF, Asselbergs FW, Te Riele A. Diagnosis and risk prediction of dilated cardiomyopathy in the era of big data and genomics. J Clin Med 2021;10:921. doi:10.3390/jcm10050921

Weintraub RG, Semsarian C, Macdonald P. Dilated cardiomyopathy. Lancet 2017;390:400‐414. doi:10.1016/S0140‐6736(16)31713–5

Jordan E, Peterson L, Ai T, Asatryan B, Bronicki L, Brown E, et al. Evidence‐based assessment of genes in dilated cardiomyopathy. Circulation 2021;144:7‐19. doi:10.1161/CIRCULATIONAHA.120.053033

Giri P, Mukhopadhyay A, Gupta M, Mohapatra B. Dilated cardiomyopathy: A new insight into the rare but common cause of heart failure. Heart Fail Rev 2022;27:431‐454. doi:10.1007/s10741‐021‐10125‐6

Cheng Q, Chen X, Wu H, Du Y. Three hematologic/immune system‐specific expressed genes are considered as the potential biomarkers for the diagnosis of early rheumatoid arthritis through bioinformatics analysis. J Transl Med 2021;19:18. doi:10.1186/s12967‐020‐02689‐y

Xue B, Li S, Jin X, Liu L. Bioinformatics analysis of mRNA and miRNA microarray to identify the key miRNA‐mRNA pairs in cisplatin‐resistant ovarian cancer. BMC Cancer 2021;21:452. doi:10.1186/s12885‐021‐08166‐z

Gholipour A, Shakerian F, Zahedmehr A, Irani S, Malakootian M, Mowla SJ. Bioinformatics analysis to find novel biomarkers for coronary heart disease. Iran J Public Health 2022;51:1152‐1160. doi:10.18502/ijph.v51i5.9430

Fan S, Hu Y. Integrative analyses of biomarkers and pathways for heart failure. BMC Med Genomics 2022;15:72. doi:10.1186/s12920‐022‐01221‐z

Zhou Y, Zhou B, Pache L, Chang M, Khodabakhshi AH, Tanaseichuk O, et al. Metascape provides a biologist‐oriented resource for the analysis of systems‐level datasets. Nat Commun 2019;10:1523. doi:10.1038/s41467‐019‐09234‐6

Zhao X, Xu M, Cai Z, Yuan W, Cui W, Li MD. Identification of LIFR, PIK3R1, and MMP12 as novel prognostic signatures in gallbladder cancer using network‐based module analysis. Front Oncol 2019;9:325. doi:10.3389/fonc.2019.00325

Zhuang X, Sun X, Zhou H, Zhang S, Zhong X, Xu X, et al. Klotho attenuated doxorubicin‐induced cardiomyopathy by alleviating dynamin‐related protein 1‐mediated mitochondrial dysfunction. Mech Ageing Dev 2021;195:111442. doi:10.1016/j.mad.2021.111442

Xia Y, Chen Z, Chen A, Fu M, Dong Z, Hu K, et al. LCZ696 improves cardiac function via alleviating Drp1‐mediated mitochondrial dysfunction in mice with doxorubicin‐induced dilated cardiomyopathy. J Mol Cell Cardiol 2017;108:138‐148. doi:10.1016/j.yjmcc.2017.06.003

Zhu M, Chen Y, Cheng L, Li X, Shen Y, Guo G, et al. Calsyntenin‐1 promotes doxorubicin‐induced dilated cardiomyopathy in rats. Cardiovasc Drugs Ther 2024;38:237‐252. doi:10.1007/s10557‐022‐07389‐x

Singal PK, Iliskovic N. Doxorubicin‐induced cardiomyopathy. N Engl J Med 1998;339:900‐905. doi:10.1056/NEJM199809243391307

Rosen GM, Halpern HJ. Spin trapping biologically generated free radicals: Correlating formation with cellular injury. Methods Enzymol 1990;186:611‐621. doi:10.1016/0076–6879(90)86156‐p

Dempsey EC, Newton AC, Mochly‐Rosen D, Fields AP, Reyland ME, Insel PA, et al. Protein kinase C isozymes and the regulation of diverse cell responses. Am J Physiol Lung Cell Mol Physiol 2000;279:L429‐L438. doi:10.1152/ajplung.2000.279.3.L429

Ma Y, Usuwanthim K, Munawara U, Quach A, Gorgani NN, Abbott CA, et al. Protein kinase Cα regulates the expression of complement receptor Ig in human monocyte‐derived macrophages. J Immunol 2015;194:2855‐2861. doi:10.4049/jimmunol.1303477

Liu Q, Chen X, Macdonnell SM, Kranias EG, Lorenz JN, Leitges M, et al. Protein kinase Cα, but not PKCβ or PKCγ, regulates contractility and heart failure susceptibility: Implications for ruboxistaurin as a novel therapeutic approach. Circ Res 2009;105:194‐200. doi:10.1161/CIRCRESAHA.109.195313

Bowling N, Walsh RA, Song G, Estridge T, Sandusky GE, Fouts RL, et al. Increased protein kinase C activity and expression of Ca2+‐sensitive isoforms in the failing human heart. Circulation 1999;99:384‐391. doi:10.1161/01.CIR.99.3.384

Zhang D, Li L, Zhu Y, Zhao L, Wan L, Lv J, et al. The NFKB1‐94 ATTG insertion/deletion polymorphism (rs28362491) contributes to the susceptibility of congenital heart disease in a Chinese population. Gene 2013;516:307‐310. doi:10.1016/j.gene.2012.12.078

Frantz S, Tillmanns J, Kuhlencordt PJ, Schmidt I, Adamek A, Dienesch C, et al. Tissue‐specific effects of the nuclear factor κB subunit p50 on myocardial ischemia‐reperfusion injury. Am J Pathol 2007;171:507‐512. doi:10.2353/ajpath.2007.061042

Kawamura N, Kubota T, Kawano S, Monden Y, Feldman A, Tsutsui H, et al. Blockade of NF‐κB improves cardiac function and survival without affecting inflammation in TNF‐α‐induced cardiomyopathy. Cardiovasc Res 2005;66:520‐529. doi:10.1016/j.cardiores.2005.02.007

Wang X, Zhang Y, Nilsson CL, Berven FS, Andren PE, Carlsohn E, et al. Association of chromosome 19 to lung cancer genotypes and phenotypes. Cancer Metastasis Rev 2015;34:217‐226. doi:10.1007/s10555‐015‐9556‐2

Yuan Z, Murakoshi N, Xu D, Tajiri K, Okabe Y, Aonuma K, et al. Identification of potential dilated cardiomyopathy‐related targets by meta‐analysis and co‐expression analysis of human RNA‐sequencing datasets. Life Sci 2022;306:120807. doi:10.1016/j.lfs.2022.120807

Puigserver P, Spiegelman BM. Peroxisome proliferator‐activated receptor‐γ coactivator 1α (PGC‐1α): Transcriptional coactivator and metabolic regulator. Endocr Rev 2003;24:78‐90. doi:10.1210/er.2002–0012

Daley D, Park JE, He JQ, Yan J, Akhabir L, Stefanowicz D, et al. Associations and interactions of genetic polymorphisms in innate immunity genes with early viral infections and susceptibility to asthma and asthma‐related phenotypes. J Allergy Clin Immunol 2012;130:1284‐1293. doi:10.1016/j.jaci.2012.07.051

Lerebours F, Vacher S, Andrieu C, Espie M, Marty M, Lidereau R, et al. Nf‐kappa B genes have a major role in inflammatory breast cancer. BMC Cancer 2008;8:41. doi:10.1186/1471‐2407‐8‐41

Li N, Li X, Deng L, Yang H, Gong Z, Wang Q, et al. 6‐Shogaol inhibits the proliferation, apoptosis, and migration of rheumatoid arthritis fibroblast‐like synoviocytes via the PI3K/AKT/NF‐κB pathway. Phytomedicine 2023;109:154562. doi:10.1016/j.phymed.2022.154562

Wang J, Uddin MN, Li Q, Aierken A, Li MY, Wang R, et al. Identifying potential mitochondrial proteome signatures associated with the pathogenesis of pulmonary arterial hypertension in the rat model. Oxid Med Cell Longev 2022;2022:8401924. doi:10.1155/2022/8401924

Joshi KK, Chen L, Torres N, Tournier V, Madura K. A proteasome assembly defect in rpn3 mutants is associated with Rpn11 instability and increased sensitivity to stress. J Mol Biol 2011;410:383‐399. doi:10.1016/j.jmb.2011.05.005

Rubio AJ, Bencomo‐Alvarez AE, Young JE, Velazquez VV, Lara JJ, Gonzalez MA, et al. 26S proteasome non‐ATPase regulatory subunits 1 (PSMD1) and 3 (PSMD3) as putative targets for cancer prognosis and therapy. Cells 2021;10:2390. doi:10.3390/cells10092390

Bencomo‐Alvarez AE, Rubio AJ, Olivas IM, Gonzalez MA, Ellwood R, Fiol CR, et al. Proteasome 26S subunit, non‐ATPases 1 (PSMD1) and 3 (PSMD3), play an oncogenic role in chronic myeloid leukemia by stabilizing nuclear factor‐kappa B. Oncogene 2021;40:2697‐2710. doi:10.1038/s41388‐021‐01732‐6

Pisano E, Pacifico L, Perla FM, Liuzzo G, Chiesa C, Lavorato M, et al. Upregulated monocyte expression of PLIN2 is associated with early arterial injury in children with overweight/obesity. Atherosclerosis 2021;327:68‐75. doi:10.1016/j.atherosclerosis.2021.04.016

Dorsett D, Strom L. The ancient and evolving roles of cohesin in gene expression and DNA repair. Curr Biol 2012;22:R240‐R250. doi:10.1016/j.cub.2012.02.046

Nasmyth K, Haering CH. Cohesin: Its roles and mechanisms. Annu Rev Genet 2009;43:525‐558. doi:10.1146/annurev‐genet‐102108–134233

Zhu T, Gao Z, Yuan K, Wang Y. High expression of RAD21 predicts poor survival in patients with operated non‐small‐cell lung cancer. Tumori 2020;106:300891620910805. doi:10.1177/0300891620910805

Deng P, Wang Z, Chen J, Liu S, Yao X, Liu S, et al. RAD21 amplification epigenetically suppresses interferon signaling to promote immune evasion in ovarian cancer. J Clin Invest 2022;132:e159628. doi:10.1172/JCI159628

Schuster K, Leeke B, Meier M, Wang Y, Newman T, Burgess S, et al. A neural crest origin for cohesinopathy heart defects. Hum Mol Genet 2015;24:7005‐7016. doi:10.1093/hmg/ddv402

Matsuzono K, Honda H, Sato K, Morihara R, Deguchi K, Hishikawa N, et al. ‘PrP systemic deposition disease’: Clinical and pathological characteristics of novel familial prion disease with 2‐bp deletion in codon 178. Eur J Neurol 2016;23:196‐200. doi:10.1111/ene.12905

Kim DY, Shim KH, Bagyinszky E, An SSA. Prion mutations in Republic of Republic of Korea, China, and Japan. Int J Mol Sci 2022;24:625. doi:10.3390/ijms24010625

Armocida D, Busceti CL, Biagioni F, Fornai F, Frati A. The role of cellular prion protein in glioma tumorigenesis could be through the autophagic mechanisms: A narrative review. Int J Mol Sci 2023;24:1405. doi:10.3390/ijms24021405

Villarreal‐Calderon R, Franco‐Lira M, Gonzalez‐Maciel A, Reynoso‐Robles R, Harritt L, Perez‐Guille B, et al. Up‐regulation of mRNA ventricular PRNP prion protein gene expression in air pollution highly exposed young urbanites: Endoplasmic reticulum stress, glucose regulated protein 78, and nanosized particles. Int J Mol Sci 2013;14:23471‐23491. doi:10.3390/ijms141223471

Wood AR, Foliaki ST, Groveman BR, Walters RO, Williams K, Yuan J, et al. Hereditary E200K mutation within the prion protein gene alters human iPSC derived cardiomyocyte function. Sci Rep 2022;12:15788. doi:10.1038/s41598‐022‐19631‐5

Philips RL, Wang Y, Cheon H, Kanno Y, Gadina M, Sartorelli V, et al. The JAK‐STAT pathway at 30: Much learned, much more to do. Cell 2022;185:3857‐3876. doi:10.1016/j.cell.2022.09.023

Darnell JE Jr. STATs and gene regulation. Science 1997;277:1630‐1635. doi:10.1126/science.277.5332.1630

Ng DC, Court NW, Dos Remedios CG, Bogoyevitch MA. Activation of signal transducer and activator of transcription (STAT) pathways in failing human hearts. Cardiovasc Res 2003;57:333‐346. doi:10.1016/S0008‐6363(02)00664–8

Hu X, Li J, Fu M, Zhao X, Wang W. The JAK/STAT signaling pathway: From bench to clinic. Signal Transduct Target Ther 2021;6:402. doi:10.1038/s41392‐021‐00791‐1

Bolivar S, Espitia‐Corredor JA, Olivares‐Silva F, Valenzuela P, Humeres C, Anfossi R, et al. In cardiac fibroblasts, interferon‐beta attenuates differentiation, collagen synthesis, and TGF‐β1‐induced collagen gel contraction. Cytokine 2021;138:155359. doi:10.1016/j.cyto.2020.155359