System analysis based on the necroptosis-related genes in heart failure

doi:10.21203/rs.3.rs-3086572/v1

Download PDF

Research Article

System analysis based on the necroptosis-related genes in heart failure

https://doi.org/10.21203/rs.3.rs-3086572/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background: Heart failure(HF) is an emerging epidemic clinical syndrome that remains a leading cause of global morbidity and mortality. This study aimed to determine necroptosis' role in HF using bioinformatics analyses.

Methods: A total of 3 datasets, including myocardial tissues samples from 225 HF patients and 26 normal people, were acquired from the Gene Expression Omnibus (GEO). Necroptosis-related differentially expressed genes (NRDEGs) in HF were determined. A prediction model based on three NRDEGs were constructed to assess the risk of HF. Then, consensus clustering, protein-protein interactions (PPI), and identifying the top 10 hub genes were performed. The Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Gene Set Enrichment Analysis (GSEA) were carried out. Investigations were conducted into immune infiltration. Furthermore, the eight NRDEGs’ mRNA expression level was validated in blood samples by quantitative real time-PCR (RT-qPCR).

Results:

A total of eight NRDEGs, namely FLOT1, DAPK1, KLHDC10, FLOT2, FAS, UCHL1, TNFAIP3, HSPA5, were excavated and further verified by RT-qPCR in blood samples. The expression levels of TNFAIP3 and HSPA5 were lower in HF, consistent with bioinformatics analysis. The correlation analysis revealed the regulatory network between 51 NRGs. The HF risk prediction model constructed of FLOT2, FAS and FLOT1 has relatively good accuracy and reliability. The 10-hub-genes associated with necroptosis might be significantly related to HF. Among10-hub-genes, STIP1, TGFBR2, and HSPD1 might be potential markers to indicate the early stage or progression of HF. The GSEA clarified nine relevant enrichment pathways.

Conclusion: Our research supplies new information and views for investigating the underlying necroptosis-related mechanism and possible treatments of HF.

Heart Failure

Necroptosis

Diagnosis

Biomarker

Bioinformatics analysis

Heart failure(HF), the end-stage manifestation of various cardiovascular diseases, remains a leading cause of global morbidity and mortality, even today ⁽¹⁾. The prevalence of HF increases with age, which reached 4.3% among 65 to 70 years older people in 2012, even it is anticipated to rise continuously to get 8.5% by 2030 ⁽²⁾. Despite significant advances in diagnosis and pharmaceutical therapy for HF in recent years, it is worrying that the long-term prognosis of HF remains poor, with a tendency towards recurrent hospitalizations and reduced health-related quality of life ⁽³⁾. Biomarkers reflect the pathophysiological pathways involved in HF progression to assist diagnosis and treatment. The most common and established clinical biomarkers for the diagnosis of HF are brain natriuretic peptide (BNP) and its inactive N-terminal fragment (NT-proBNP), which are released by cardiomyocytes in response to pressure or volume overload ^{(4, 5)}. Nevertheless, what cannot be ignored is that in specific conditions, such as with renal disease, acute coronary syndromes, cirrhotic ascites, obesity or flash pulmonary oedema, and so on, either BNP or NT-pro BNP’s level could fluctuate ⁽⁶⁾. Hence, exploring more novel and reliable biomarkers for HF is imperative. In addition, applying a multi-biomarkers strategy could better reflect different HF pathways, helping further understand the individual phenotypic differences in patients and optimize and novel targeted therapeutic strategy ⁽⁵⁾.

Necroptosis, also termed programmed necrosis, is a newly identified mode of subtly regulated cell death characterized by necrotic morphology⁽⁷⁾. Necroptosis is principally mediated by death receptors and executed through the activation of receptor-interacting protein kinase 3(RIPK3) and its downstream substrate, mixed lineage kinase-like domain(MLKL)⁽⁸⁾. Multiple cell death programs have been linked to HF⁽⁹⁾. Studies that have recently accumulated have shown that necroptosis might play a vital role in various pathophysiological processes in HF⁽¹⁰⁾. For example, the RIPK1, RIPK3, and MLKL as critical mediators of necroptosis, as well as their phosphorylated forms, were up-regulated in HF patients⁽¹¹⁾. The caspase-8, which could inhibit necroptosis by cleaving RIPK1 and RIPK3, was significantly decreased⁽¹²⁾. In myocardial infarction mice, hydrogen sulfide protects against ischemic HF by downregulating the RIP1/RIP3/MLKL pathway to inhibit necroptosis. But propargylglycine treatment aggravated cardiac function by activating the RIP1/RIP3/MLKL pathway⁽¹³⁾. However, compared to other cell death programs involved in HF, studies on necroptosis are relatively rare. Therefore, an in-depth analysis of the role of necroptosis in HF is necessary, as well as establishing a relevant diagnostic model and exploring potential therapeutic targets of HF.

Our study utilized bioinformatics analysis to systematically examine the necroptosis-related genes(NRGs) involved in HF(See analysis process in Figure S1). We identified the necroptosis-related differentially expressed genes(NRDEGs) and hub genes in the HF database, following combination with verification in the clinical blood samples. Additionally, we constructed a necroptosis-related risk prediction model to assess the risk of HF. The Gene Ontology(GO), Kyoto Encyclopedia of Genes and Genomes(KEGG), and Gene Set Enrichment Analysis(GSEA) were applied to provide possible information. Simultaneously, the immune infiltration and the correlation between immune cell-type abundance and NRDEGs expression were investigated.

2.1 Data acquisition and pre-processing

GSE3585⁽¹⁴⁾, GSE5406⁽¹⁵⁾, and GSE42955⁽¹⁶⁾ expression profile datasets with reliable sample sources for HF were obtained from the Gene Expression Omnibus(GEO) database (Table S1) (https://www.ncbi.nlm.nih.gov/geo/) using the GEOquery package⁽¹⁷⁾ in R software (version 4.1.1, http://r-project.org/). All dataset samples came from Homo sapiens. The gene expression profiles were respectively based on platform GPL96 [HG-U133A] Affymetrix Human Genome U133A Array and GPL6244 [HuGene-1_0-st] Affymetrix Human Gene 1.0 ST Array [transcript (gene) version]. The GSE3585 contained seven myocardial tissue samples with dilated cardiomyopathy(DCM) causing HF and five normal myocardial tissue samples; GSE5406 contained 86 myocardial tissue samples with DCM causing HF, 108 myocardial tissue samples with ischemic cardiomyopathy(ICM) causing HF and 16 normal myocardial tissue samples; and GSE42955 contained 12 myocardial tissue samples with DCM causing HF, 12 myocardial tissue samples with ICM causing HF and five normal myocardial tissue samples. To acquire the individual gene expression matrices, we utilized the GEOquery package to read the expression files, conducting background correction and normalization further. The combined dataset was removed batch effects using the sva package⁽¹⁸⁾, and the impact of inter-sample correction was demonstrated by the boxplot. Meanwhile, a list of NRGs was downloaded from the GeneCards: The Human Gene Database (GeneCards) (https://www.genecards.org/).

2.2 Model building

The Least Absolute Shrinkage and Selection Operator(LASSO) regression model⁽¹⁹⁾ as a machine learning algorithm was used to construct diagnostic models generally, which utilized regularization to address the overfitting in order to improve the prediction accuracy. Glmnet package⁽²⁰⁾ in R was applied to construct a model of NRGs using parameters set.seed(1234), and family = "binomial". The optimal hyperparameter λ value was derived by feature downscaling through 10-fold cross-validation. In order to determine more accurate independent prognostic factors, we performed multivariable logistic regression analysis and carried out the final screening by stepwise regression.

All samples were next randomly assigned into a training set and a test set according to the ratio of 7:3, then the receiver operating characteristic curve(ROC) analysis to test the accuracy and reliability of the logistic regression models in two sets, respectively. Afterwards, nomograms were generated using logistic regression and evaluated using ROC analysis and calibration curves.

2.3 Cluster analysis

Consensus clustering is an approach for quantifying the number and components of potential clusters within a dataset (such as microarray). To facilitate better differentiation between HF molecular subtypes, the ConsensusClusterPlus package⁽²¹⁾ in R was utilized for consistent clustering of the joint dataset standing on prognostic genes of the constructed model. The 80% of the entire sample was drawn 50 times during the procedure, with the number of clusters set between 2 and 10, distances set to "euclidean", and clusterAlg set to "km".

2.4 Expression distribution of regulatory factors of necroptosis genes

Based on the molecular subtype distribution of all samples, the HF samples in the joint dataset were separated into two subgroups (Cluster 1 and Cluster 2). Concerning the two subtype expression matrices, the limma package was used to screen the differentially expressed genes(DEGs)⁽²²⁾. The volcano maps and heatmaps of DEGs were generated using the ggplot2 package⁽²³⁾ and pheatmap package (https://CRAN.R-project.org/package=pheatmap), to visualize the differential expression of DEGs. The following criteria were used to determine differential expression: P value 0.05, logFC > = 0.5 for DEGs, with logFC > = 0.5 for upregulated genes and logFC = 0.5 for downregulated genes.

2.5 Protein-protein interaction network analysis

The STRING protein-protein interaction(PPI) database⁽²⁴⁾ was used to analyze the interaction relationships of necroptosis gene regulatory factors. The results were visualized by Cytoscape software⁽²⁵⁾ and Hub genes were identified by the cytoHubba plugin⁽²⁶⁾.

2.6 Enrichment Analysis

Large-scale functional enrichment studies, encompassing biological process(BP), molecular function(MF), and cellular component(CC), are frequently carried out using the Gene Ontology (GO) analysis⁽²⁷⁾. KEGG, or the Kyoto Encyclopedia of Genes and Genomes, is a popular database for storing data about genomes, biological process, illnesses, and drugs⁽²⁸⁾. For GO functional annotation analysis and KEGG pathway enrichment analysis of necroptosis gene regulators, the ClusterProfiler package was utilized in R⁽²⁹⁾. Statistical significance is regarded to exist at the FDR threshold value of 0.05.

Gene Set Enrichment Analysis(GSEA) was used to examine the biological processes that differ between various groupings⁽³⁰⁾, based on the expression profiles of Cluster 1 and 2 data sets of HF samples from the combined data sets GSE3585, GSE5406, and GSE42955. GSEA is a computational method that assesses whether a specific gene set statistically differs between two biological states. It is frequently applied to samples expression datasets to estimate changes in pathways and biological process activity. The gene set “c2.cp.v7.0.entrez.gmt” was transferred from the MSigDB database for GSEA, with FDR < 0.25 as the threshold value.

2.7 Immune Infiltration Analyses

To farther quantify the percentage of various immune cells in Cluster 1 and 2 groups, we employed a single-sample GSEA(ssGSEA) to distinguish between different human immune cells phenotypes in the immune microenvironment with excellent sensitivity and specificity. The relative abundance of infiltrating immune cells was calculated using the ssGSEA. The algorithm was utilized to obtain a set of 28 genes that permitted the content of Bindea et al. to tab the diverse infiltrating immune cell kinds⁽³¹⁾. The gene set involves several subtypes of human immune cells, namely, CD8 + T cells, macrophages, dendritic cells, regulatory T cells, etc. The infiltration level of each immune cell type in each sample can be represented by the enrichment scores generated by ssGSEA via the GSVA package⁽³²⁾ in R software.

CIBERSORT is a tool that deconvolutes the transcriptome expression matrix to estimate the composition and abundance of immune cells in the mixed cells based on the principle of linear support vector regression.⁽³³⁾. We analyzed the immune-infiltration situation by CIBERSORT, which comprises 22 types of immune cells and filtered the output (P value < 0.05) to derive the immune cell infiltration matrix. The heat maps were plotted to visualize the correlation of 22 immune cell infiltrations by corrplot package⁽³⁴⁾. The immune cells expression difference between the high- and low-risk groups was plotted by ggpubr package (https://CRAN.R-project.org/package=ggpubr).

2.8 Sample collection

We enrolled ten patients diagnosed with HF and ten matched normal subjects at Beijing hospital. All blood samples were collected via a direct venous puncture. Then the blood was immediately transferred into RNase/DNase-free 15ml conical tubes after adding three times RNALock Reagent (no. DP440, TIANGEN BIOTECH (Beijing) CO., LTD) and stored at -80℃. This research was conducted in accordance with the Declaration of Helsinki and approved by the Medical Ethics Committee of Beijing hospital (2023BJYYEC-069-01).

2.9 RT-qPCR

Total RNA was extracted from the blood samples and reversed transcribed into cDNA using RNAprep Pure Hi-Blood Kit (no. DP443, TIANGEN BIOTECH (Beijing) CO., LTD) for 15min at 42°C. Heating at 95°C for 3 minutes terminated the reaction. The amplification of the hub genes was carried out by RT-qPCR in the Light Cycler 480II (ROCHE) using a specific set of primers. Table S2 displays the employed primers. The following were the thermocycling conditions: 95°C for 10 seconds, 60°C for 20 seconds, and followed by 72°C for 20 seconds (a total of 40 cycles). Each sample's relative quantification units, are computed as a fold change in gene expression in relation to the normal group (relative quantification = 2 − ΔΔCt, where Ct represents quantification cycle values). GAPDH served as an endogenous control.

2.10 Statistics analysis

R software (version 4.1.1, http://r-project.org/) was executed all statistical analysis and data processing. Independent Student t-tests were employed to determine the statistical significance of normally distributed variables, and Mann-Whitney U-tests(i.e., Wilcoxon rank sum test) was utilized to determine differences between non-normally distributed variables when comparing two groups of continuous variables. All statistical P values were two-sided and statistically significant at P value < 0.05.

3.1 Data pre-processing

The gene expression matrices of GSE3585, GSE5406, and GSE42955 datasets were first normalized and processed (Figure S2A-S2C), and the datasets were combined for de-batching effects, which were presented as boxplot plots before and after de-batching effects (Figure S2D, S2E).

3.2 Screening and correlation analysis of NRGs

A total of 611 NRGs were obtained from the GeneCards database, of which 91 NRGs had a Relevance score > 1. To identify differentially expressed NRGs in HF samples, 91 NRGs were matched with a combined dataset to obtain 51 NRGs. Boxplots displaying that only eight genes were differentially expressed between the HF and normal groups, namely FLOT1, DAPK1, KLHDC10, FLOT2, FAS, UCHL1, TNFAIP3, and HSPA5, which expressions were all downregulated in the HF group (Figure S3A, S3B). Combined with the chromosomal locations of these NRGs, the chromosomal Circos plot was mapped with the finding that the distribution of these NRGs was scattered on the chromosomes (Figure S3C).

Subsequently, the expression correlation between 51 NRGs was further analyzed (Fig. 1A). In the correlation analysis of these NRGs, we found positive correlations between CFLAR and FLOT1, PTGES3 and FMR1, NFE2L2 and PTGES3, as well as NFE2L2 and FMR1 (Top four), suggesting that they may have positive interactions (Fig. 1B); SIRT2 and FMR1, NFE2L2, RB1 and DNM1L were negative correlation (Top four), indicating that they may have a negative regulatory role (Fig. 1C).

3.3 Model construction based on NRGs

We analyzed the feasibility of constructing a HF model based on NRGs by the LASSO algorithm. Based on 10-fold cross-validation, the optimal λ = 0.01804121 and 18 NRGs with their corresponding coefficients were obtained (Fig. 2A-2B). Then a univariate logistic analysis was performed to screen 7 NRGs with P < 0.05 (Fig. 2C); the multivariate logistic analysis was performed to gain 3 NRGs with P value < 0.05, namely FLOT2, FAS, and FLOT1(Fig. 2D).

We subsequently constructed models and plotted nomograms by FLOT2, FAS, and FLOT1 (Fig. 3A), which were used to better assess the risk of HF in patients. The Calibration curve showed that the models had good stability (Fig. 3B). The accuracy and reliability of the training and test set models were also evaluated by receiver operating characteristic curve(ROC). In the training set, there were differences in scores between the normal and HF groups, and the area under the curve(AUC) of the ROC was 0.749 (Fig. 3C, 3D). In the test set, differences in scores between the normal and HF groups were likewise found, with an AUC of 0.787 (Fig. 3E, 3F). All the results above suggested that the constructed model based on three NRGs, namely FLOT2, FAS, and FLOT1, has relatively good accuracy and reliability.

3.4 Molecular typing based on NRGs

Consensus clustering methods were utilized to cluster the samples. The interference between subgroups can be minimized when K = 2 is chosen for consensus clustering analysis (Fig. 4A). As a result, two subgroups Cluster 1 (n = 146) and Cluster 2 (n = 79) were identified and their expression profiles were compared (Fig. 4B). A total of 112 genes were discovered to be differentially expressed between Cluster 1 and Cluster 2 by analyzing the DEGs between the two groups, of which four were up-regulated and 108 were down-regulated in Cluster 2 (Fig. 4C, 4D).

3.5 Regulatory network of characteristic NRGs

Using the STRING database, 112 NRDEGs were subjected to PPI network analysis (Figure S4A). The degree of freedom was calculated in the Cytoscape software, followed by ranking for the above genes according to the degree value with a total of 91 points and 147 edges obtained (Figure S4B). The CytoHubba plugin's degree algorithm selected the top 10 hub genes, which were CALR, CCT2, PSMC3, RAB5A, FH, STIP1, OGDH, TGFBR2, SYNE2, and HSPD1 (Figure S4C).

3.6 Analysis of the genetic biology of HF in different typing groups

The 112 NRDEGs were subjected to GO/KEGG analysis with a P value < 0.05, and the results are shown in Figure S5 and Table S3. GO analysis reveals that NRGs regulators are closely associated with many biological processes, such as metal ion export, export across the plasma membrane, cell-substrate junction, sarcolemma, and chaperone binding. In the meantime, KEGG analysis reveals that NRGs regulators could influence proximal tubule bicarbonate reclamation, protein digestion and absorption, endocrine, cardiac muscle contraction, and other factor-regulated calcium reabsorption mineral absorption, and so on.

Based on the differential analysis of different molecular typing groups, all genes were included in the GSEA to clarify the relevant enrichment pathways in different molecular typing groups. The pathways were mainly enriched in 9 pathways, including muscle contraction, pyruvate metabolism, citric acid tricarboxylic acid (TCA) cycle, cardiac conduction, ALK1 pathway, PI3K PLC TRK pathway, ERBB1 downstream pathway, insulin receptor pathway in cardiac myocytes, thrombin PAR1 pathway, and PID IGF1 pathway (Fig. 5 and Table S4).

3.7 Correlation analysis between NRGs’ expression and immune characteristic subtypes and infiltration level

HF was divided into two subtypes according to molecular characteristics. Afterwards, ten hub genes were found to be somewhat different between the two subtypes, and all of them were expressed at higher levels in the Cluster 1 subtype compared to the Cluster 2 subtype (Fig. 6A-6J). The correlation analysis of these 10 hub genes with immune cells (R ≥ 0.45, P value < 0.05) revealed that CCT2 and TGFBR2 genes were associated with activated dendritic cells, CCT2, FH, and RAB5A genes were associated with effector memory CD4 + T cells, and TGFBR2 gene was associated with plasmacytoid dendritic cells (Fig. 6K-6P).

3.8 Immune Infiltration Analyses

The altered levels of immune cell infiltration in HF were evaluated by CIBERSORT method, with T cells follicular helper and T cells regulatory (Tregs) subpopulations accounting for a large proportion of infiltrating immune cells (Fig. 7A). Correlation analysis showed correlations between the levels of infiltration of multiple immune cells, where orange-red represents positive correlations and purple represents negative correlations (Fig. 7B). Meanwhile, immune cells were analyzed differently between different molecular typing groups. B cells memory, macrophages M2, mast cells activated, mast cells resting, T cells CD4 memory resting, and T cells CD4 naive immune cells were found to be existing differences between Cluster 1 and Cluster 2 (Fig. 7C-7H).

3.9 Validation of the NRDEGs in HF patients

To validate the mRNA expression levels of NRDEGs, we further identified eight NRDEGs by RT-qPCR in collected clinical blood samples. The blood samples come from 10 HF patients, including 5 ICM causing HF, 3 hypertrophic cardiomyopathy causing HF and 2 DCM causing HF (clinical characteristics of HF and non-HF see Table S5). The expression levels of TNFAIP3, and HSPA5 were significantly lower in HF blood samples than in normal blood samples (Fig. 8G, 8H). However, the expression levels of FLOT1, DAPK1, KLHDC10, FLOT2, FAS, UCHL1, and TNFAIP3 showed no significant difference between the two groups (Fig. 8A-8F).

HF is an emerging epidemic that is increasing strongly in prevalence as the population ages, with severe morbidity and a 5-year mortality rate of 50%⁽³⁵⁾. The burden of HF as a population-wide burden is highlighted by the fact that lifetime risks are pretty high regardless of sex, ethnicity, or region⁽²⁾. In recent years, the multiple roles of biomarkers for HF have been validated to predict the future occurrence of HF, to identify its presence when fully developed, to stratify the risk of affected patients, and to serve as biological tools to guide treatment potentially. Although the pivotal role of necroptosis in various heart diseases has been emphasized, there are fewer studies of necroptosis in prognostic, diagnostic, and therapeutic value in the HF field. This study was a systematic bioinformatics analysis to find biomarkers of necroptosis in HF, which could offer reliable directions for future specific studies of HF and novel opportunities for developing effective diagnoses and therapies.

Based on the public GEO database, we have first screened 51 NRDEGs. Only eight genes were differentially expressed between the HF and normal groups, namely FLOT1, DAPK1, KLHDC10, FLOT2, FAS, UCHL1, TNFAIP3, and HSPA5, which expressions were all down-regulated in the HF group. Furthermore, we utilized FLOT2, FAS, and FLOT1 to construct a risk prediction model by the LASSO algorithm and multiple logistic analysis, which aims to assess the risk of HF in patients. The AUC was 0.749 in the training set and the AUC was 0.787 in the test set, suggesting that the model has relatively good accuracy and reliability. The AUC values for BNP and NT-proBNP to predict ischemic HF were 0.770 and 0.754 respectively^{(36, 37)}. To compare with such established and mature biomarkers in related HF risk assessment, like BNP and NT-proBNP, this risk prediction model based on three necroptosis-related genes has a certain clinical potential. Overall, a multi-biomarkers strategy could better reflect different HF pathways to improve clinical predictive performance.

FLOT1(flotillin-1) and FLOT2(flotillin-2), which belong to the SPFH superfamily, are ubiquitously expressed, highly conserved raft-associated integral membrane proteins⁽³⁸⁾. FLOT1 and FLOT2 share the same domain architecture and show a sequence identity of about 50% for the human protein⁽³⁹⁾. Even though flotillins are known to involve various cellular events, for instance, cell adhesion, cargo endocytosis, protein sorting and recycling, and cell migration, studies of their roles in cardiovascular disease are rare^{(40, 41)}. Flotillins are primarily known for their role in cancer by promoting tumorigenesis and metastasis⁽⁴²⁾. FLOT2 has altered in cardiac intercalated discs of patients with arrhythmogenic cardiomyopathy and DCM, thus suggesting its potential role in cardiac remodeling⁽⁴³⁾. FLOT1 with an excellent diagnostic value in DCM-caused HF has been confirmed in another study⁽⁴⁴⁾. Moreover, FLOT1 and FLOT2 were closely associated with cholesterol absorption. FLOT2-dependent exosomal cholesterol secretion was described by Strauss et al. as a novel cellular cholesterol homeostasis mechanism⁽⁴⁵⁾, while cholesterol absorption is strongly linked with various cardiovascular diseases. In our study, further correlation analysis showed that CFLAR is positively correlated with FLOT1, which might suggest that FLOT1 and CFLAR are both potential regulatory targets for HF. Fas as the tumor necrosis factor (TNF) receptor family member, which is encoded by the FAS gene, is a type I membrane protein that is produced by a variety of cells⁽⁴⁶⁾. The binding of Fas to its ligand FasL to form the FasL/Fas pathway plays a major role in immune function, particularly in activation-induced cell death⁽⁴⁷⁾. FasL/Fas pathway has been shown to provide an important mechanism for cardiomyocyte death following ischemia/reperfusion. Existing evidence strongly suggests FasL/Fas signaling is associated with progressive DCM, congestive HF, and ischemic HF in humans^(48–50). Fas/FasL signaling represents a novel therapeutic target for HF.

We have constructed a PPI network and hub genes screening to identify more potential biomarker candidates related to necroptosis in HF. The top ten Hub genes were CALR, CCT2, PSMC3, RAB5A, FH, STIP1, OGDH, TGFBR2, SYNE2, and HSPD1. Stress-inducible phosphoprotein 1(STIP1) is an adaptor protein that bridges the gap between the folding of HSP70 and HSP90 and a secreted protein that regulates the growth of malignant cells⁽⁵¹⁾. Oxidative stress has been directly linked to the pathophysiological mechanisms of HF. Nrf2 ablated mice have pathological susceptibility to exogenous oxidative, and STIP1 was reduced in its heart⁽⁵²⁾. We speculated that the downregulating of STIP1 caused increased oxidative stress to accelerate HF progress. In addition, glucocorticoids have been used to improve the curative effect of diuretics and natriuretic peptides in advanced HF. In the glucocorticoid pathway, STIP1 is a novel therapeutic target that is involved in the activation of the glucocorticoid receptor⁽⁵³⁾. Together the results hint that STIP1 downregulation may indicate early HF. On the other hand, STIP1 may become an indicator of oxidative stress levels in HF progress to apply in treatment.

TGFBR2 encodes the TGF-beta II receptor, which can activate the TGF-beta signaling pathway by specific ligand binding ⁽⁵⁴⁾. HF patients had elevated plasma concentrations of TGF-beta1⁽⁵⁵⁾. TGF-beta1 signal and downstream Smads are essential for tissue fibrosis and matrix remodeling in multiple etiologies of HF. Direct evidence is provided by Rojas et al. that the specificity of the TGF-beta1 response can be influenced by the regulation of the expression level of TGFBR2⁽⁵⁶⁾. Hence, TGFBR2 may be a key target for alleviating ventricular remodeling in HF.

The HSPD1 gene encodes for a constitutively expressed mitochondrial heat shock protein Family D Member 1(HSPD1, also named Hsp60) chaperonin protein, which is responsible for the folding, transport, and quality control of mitochondrial matrix proteins and is essential for maintaining life⁽⁵⁷⁾.

The levels of circulatory HSPD1 increase during HF⁽⁵⁸⁾. Enomoto et al. demonstrated that the upregulation of the HSPD1 protein inhibited the activity of mitochondrial complex IV, increasing reactive oxygen species(ROS) concentration. Elevated ROS caused HF by significantly increasing the amount of mitophagy in the heart⁽⁵⁹⁾. This may suggest that the elevated circulating concentrations of HSPD1 may indicate early HF or exacerbation of HF.

Although the valuable NRDEGs were mined from the myocardial tissue databases, myocardial tissue samples from most HF patients are not readily available clinically, but blood samples are available for measuring the expression levels of NRDEGs. To validate the results further, we preliminarily explored the expression levels of these eight NRDEGs in collected human blood samples and found that the expression of TNFAIP3 and HSPA5 was lower in HF compared to the normal group, which is consistent with our previous bioinformatics analysis. Obviously, the differences in other NRDEGs need to be validated and explored by large-scale clinical trials. The TNFAIP3 gene encodes the ubiquitin-binding protein A20, a negative regulator of NF-κB⁽⁶⁰⁾. A host of studies confirmed that NF-κB activity was positively correlated with HF progression, and inhibition of NF-κB activity limits HF progression^{(61, 62)}. However, the direct pathological changes related to TNFAIP3/A20 in HF have been sparsely examined. Binding immunoglobulin protein (BiP), an endoplasmic reticulum chaperone from the Hsp70 family, is encoded by the HSPA5 gene⁽⁶³⁾. In an integrative meta-analysis of cellular and molecular transcriptomic landscape in nonfailing and HF human hearts, Luo et al. identified an altered mRNA transcription of HSPA5, associated with mitochondrial transaminases, in the left ventricle of the failing heart ⁽⁶⁴⁾. Our initial findings indicated that TNFAIP3 and HSPA5 could be promising biomarkers and further increase TNFAIP3 and HSPA5 expression might be implicated in HF progression.

There are still some limitations to this research. First, the datasets used in the mining strategy only include DCM and ICM causing HF. Hence, there are a limited number of background diseases causing HF. Second, the animal experiment to further comprehensively verify the conclusions is inadequate. Third, the data were mined from a database of myocardial tissue, but for obvious reasons, only blood samples were collected for further validation. Moreover, the collected blood sample size was small, and the analysis originated from a single center. To sum up, a multi-center, randomized, controlled trial study and basic follow-up experiments are necessary for the future.

In summary, 51 NRDEGs were identified and eight were further verified using RT-qPCR. The hub genes CALR, CCT2, FH, HSPD1, OGDH, PSMC3, RAB5A, STIP1, SYNE2 and TGFBR2 genes might be significantly related to HF. In addition, we developed a risk prediction model based on FLOT2, FAS, and FLOT1 to assess the risk of HF in patients. Our study supplies fresh evidence and ideas for investigating the underlying necroptosis-related mechanisms in HF.

Ethics approval and consent to participate

The study was performed in accordance with the ethical standards of the Declaration of Helsinki (1964) and its subsequent amendments. The experiments were approved by the Medical Ethics Committee of Beijing hospital (2023BJYYEC-069-01). Informed consent was obtained from all subjects and/or their legal guardian(s).

Consent for publication

Not Applicable

Availability of data and materials

The datasets generated and analysed during the current study are available in this published article and the Supplementary information files

Competing interests

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

Funding

This work was supported by the National Program on Key Research Project of China [2020YFC2008106].

Authors' contributions

YK performed the experiments, analyzed the data and drafted the manuscript. YG and TX collected the samples. FW and JZ revised the manuscript. FW conceived the whole idea for the study.

Acknowledgement

The authors gratefully acknowledge the data generated by GEO database used in this study.

Heidenreich P, Bozkurt B, Aguilar D, Allen L, Byun J, Colvin M, et al. 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2022;145(18):e895-e1032.
Roger V. Epidemiology of Heart Failure: A Contemporary Perspective. Circulation research. 2021;128(10):1421-34.
McDonagh T, Metra M, Adamo M, Gardner R, Baumbach A, Böhm M, et al. 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure. European heart journal. 2021;42(36):3599-726.
Chow, Sheryl, L., Bozkurt, Biykem, Liu, et al. Role of Biomarkers for the Prevention, Assessment, and Management of Heart Failure A Scientific Statement From the American Heart Association. Circulation: An Official Journal of the American Heart Association. 2017;135(22):E1054-E91.
Castiglione V, Aimo A, Vergaro G, Saccaro L, Passino C, Emdin M. Biomarkers for the diagnosis and management of heart failure. Heart failure reviews. 2022;27(2):625-43.
Christenson, R. H. National Academy of Clinical Biochemistry Laboratory Medicine Practice Guidelines for Utilization of Biochemical Markers in Acute Coronary Syndromes and Heart Failure. Circulation. 2007;115(13):e350-e1.
Galluzzi L, Kroemer G. Necroptosis: a specialized pathway of programmed necrosis. Cell. 2008;135(7):1161-3.
Galluzzi L, Vitale I, Aaronson SA, Abrams JM, Kroemer G. Molecular mechanisms of cell death: recommendations of the Nomenclature Committee on Cell Death 2018. Cell Death and Differentiation. 2018.
Del Re DP, Amgalan D, Linkermann A, Liu Q, Kitsis RN. Fundamental mechanisms of regulated cell death and implications for heart disease. Physiological reviews. 2019;99(4):1765-817.
Guo X, Chen Y, Liu Q. Necroptosis in heart disease: Molecular mechanisms and therapeutic implications. Journal of molecular and cellular cardiology. 2022;169:74-83.
Szobi A, Gonçalvesová E, Varga ZV, Leszek P, Kuśmierczyk M, Hulman M, et al. Analysis of necroptotic proteins in failing human hearts. Journal of translational medicine. 2017;15(1):1-7.
Feng S, Yang Y, Mei Y, Ma L, Zhu D-e, Hoti N, et al. Cleavage of RIP3 inactivates its caspase-independent apoptosis pathway by removal of kinase domain. Cellular signalling. 2007;19(10):2056-67.
Ma F, Zhu Y, Chang L, Gong J, Luo Y, Dai J, et al. Hydrogen sulfide protects against ischemic heart failure by inhibiting RIP1/RIP3/MLKL-mediated necroptosis. Physiological research. 2022.
Barth AS, Kuner R, Buness A, Ruschhaupt M, Merk S, Zwermann L, et al. Identification of a common gene expression signature in dilated cardiomyopathy across independent microarray studies. Journal of the American College of Cardiology. 2006;48(8):1610-7.
Hannenhalli, S. Transcriptional genomics associates FOX transcription factors with human heart failure. Circulation. 2006;114(12):1269-76.
Micaela M, Esther RL, Ana O, Estefanía T, Manuel O, Luis M, et al. Differential Gene Expression of Cardiac Ion Channels in Human Dilated Cardiomyopathy. Plos One. 2013;8(12):e79792.
Meltzer DPS. GEOquery: a bridge between the Gene Expression Omnibus (GEO) and BioConductor. Bioinformatics. 2007;23(14):1846-7.
Leek JT, Johnson WE, Parker HS, Jaffe AE, Storey JD. The SVA package for removing batch effects and other unwanted variation in high-throughput experiments. Bioinformatics. 2012;28(6):882-3.
Hua, Zhou, Sehl, Mary, E., Sinsheimer, et al. Association screening of common and rare genetic variants by penalized regression. Bioinformatics. 2010.
Friedman JH, Hastie T, Tibshirani R. Regularization Paths for Generalized Linear Models via Coordinate Descent. Journal of Statistical Software. 2010;33.
Wilkerson, Matthew D, Hayes, Neil D. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics. 2010.
Ritchie ME, Belinda P, Wu D, Hu Y, Law CW, Wei S, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic acids research. 2015;43(7):e47.
Wickham H. ggplot2: elegant graphics for data analysis Springer-Verlag New York; 2009. Book; 2016.
Mering Cv, Huynen M, Jaeggi D, Schmidt S, Bork P, Snel B. STRING: a database of predicted functional associations between proteins. Nucleic acids research. 2003;31(1):258-61.
Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome research. 2003;13(11):2498-504.
Chin C-H, Chen S-H, Wu H-H, Ho C-W, Ko M-T, Lin C-Y. cytoHubba: identifying hub objects and sub-networks from complex interactome. BMC systems biology. 2014;8(4):1-7.
Collaborators. Gene Ontology Consortium: going forward. Nucleic Acids Res. 2015;43:D1049-D56.
Kanehisa M, Goto S. Comprehensive gene and pathway analysis of cervical cancer progression. Nucleic Acids Res. 2000;28:27-30.
Wu T, Hu E, Xu S, Chen M, Guo P, Dai Z, et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. The Innovation. 2021;2(3):100141.
Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proceedings of the National Academy of Sciences of the United States of America. 2005;102(43):P.15545-50.
Bindea G, Mlecnik B, Tosolini M, Kirilovsky A, Waldner M, Obenauf AC, et al. Spatiotemporal dynamics of intratumoral immune cells reveal the immune landscape in human cancer. Immunity. 2013;39(4):782-95.
Hänzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC bioinformatics. 2013;14(1):1-15.
Newman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, et al. Robust enumeration of cell subsets from tissue expression profiles. Nature methods. 2015;12(5):453-7.
Wei T, Simko V, Levy M, Xie Y, Jin Y, Zemla J, et al. Package ‘corrplot’. Statistician. 2021;56:316-24.
Virani S, Alonso A, Aparicio H, Benjamin E, Bittencourt M, Callaway C, et al. Heart Disease and Stroke Statistics-2021 Update: A Report From the American Heart Association. Circulation. 2021;143(8):e254-e743.
Grewal J, Mckelvie R, Lonn E, Tait P, Carlsson J, Gianni M, et al. BNP and NT-proBNP predict echocardiographic severity of diastolic dysfunction. European journal of heart failure: journal of the Working Group on Heart Failure of the European Society of Cardiology. 2008(10-3).
Seino Y, Ogawa A, Yamashita T, Fukushima M, Ogata KI, Fukumoto H, et al. Application of NT-proBNP and BNP measurements in cardiac care: a more discerning marker for the detection and evaluation of heart failure. European Journal of Heart Failure. 2014;6(3):295-300.
Langhorst MF, Reuter A, Stuermer C. Scaffolding microdomains and beyond: the function of reggie/flotillin proteins. Cellular & Molecular Life Sciences. 2005;62(19-20):2228.
Rivera-Milla E, Stuermer C, Málaga-Trillo E. Ancient origin of Reggie (Flotillin), Reggie-like, and other lipid-raft proteins: convergent evolution of the SPFH domain. Cellular and Molecular Life Sciences CMLS. 2006;63(3):343-57.
Schaffer AM, Minguet S. Caveolin-1, tetraspanin CD81 and flotillins in lymphocyte cell membrane organization, signaling and immunopathology. Biochemical Society Transactions. 2020;48(6):2387-97.
Antje B, Tanja B, Nina K, Melanie M, Mika R, Ritva T. Flotillins Regulate Focal Adhesions by Interacting with α-Actinin and by Influencing the Activation of Focal Adhesion Kinase. Cells. 2018;7(4):28-.
Song L, Gong H, Lin C, Wang C, Liu L, Wu J, et al. Flotillin-1 promotes tumor necrosis factor-α receptor signaling and activation of NF-κB in esophageal squamous cell carcinoma cells. Gastroenterology. 2012;143(4):995-1005.e12.
Soni S, Raaijmakers A, Raaijmakers L, Damen J, van Stuijvenberg L, Vos M, et al. A Proteomics Approach to Identify New Putative Cardiac Intercalated Disk Proteins. PloS one. 2016;11(5):e0152231.
Wang X, Guan H, Liu W, Li H, Ding J, Feng Y, et al. Identification of Immune Markers in Dilated Cardiomyopathies with Heart Failure by Integrated Weighted Gene Coexpression Network Analysis. Genes. 2022;13(3).
Strauss K, Goebel C, Runz H, Möbius W, Weiss S, Feussner I, et al. Exosome secretion ameliorates lysosomal storage of cholesterol in Niemann-Pick type C disease. Journal of Biological Chemistry. 2010;285(34):26279-88.
Ferguson TA, Griffith TS. A vision of cell death: Fas ligand and immune privilege 10 years later. Immunological reviews. 2006;213(1):228-38.
Kavurma M, Khachigian L. Signaling and transcriptional control of Fas ligand gene expression. Cell Death & Differentiation. 2003;10(1):36-44.
Huby A, Turdi S, James J, Towbin J, Purevjav E. FasL expression in cardiomyocytes activates the ERK1/2 pathway, leading to dilated cardiomyopathy and advanced heart failure. Clinical science (London, England : 1979). 2016;130(4):289-99.
Fan Q, Huang Z, Boucher M, Shang X, Zuo L, Brinks H, et al. Inhibition of Fas-associated death domain-containing protein (FADD) protects against myocardial ischemia/reperfusion injury in a heart failure mouse model. PloS one. 2013;8(9):e73537.
Feng Q, Zhao Y, Abdelwahid E. The role of Fas in the progression of ischemic heart failure: prohypertrophy or proapoptosis. Coronary artery disease. 2008;19(7):527-34.
Blair LJ, Genest O, Mollapour M. The multiple facets of the Hsp90 machine. Nature Publishing Group; 2019.
Quiles JM, Narasimhan M, Shanmugam G, Milash B, Hoidal JR, Rajasekaran NS. Differential regulation of miRNA and mRNA expression in the myocardium of Nrf2 knockout mice. BMC genomics. 2017;18(1):1-16.
Hawkins GA, Lazarus R, Smith RS, Tantisira KG, Meyers DA, Peters SP, et al. The glucocorticoid receptor heterocomplex gene STIP1 is associated with improved lung function in asthmatic subjects treated with inhaled corticosteroids. Journal of allergy and clinical immunology. 2009;123(6):1376-83. e7.
Vander Ark A, Cao J, Li X. TGF-β receptors: In and beyond TGF-β signaling. Cellular signalling. 2018;52:112-20.
Dixon D-L, Griggs KM, Bersten AD, De Pasquale CG. Systemic inflammation and cell activation reflects morbidity in chronic heart failure. Cytokine. 2011;56(3):593-9.
Rojas A, Padidam M, Cress D, Grady WM. TGF-β receptor levels regulate the specificity of signaling pathway activation and biological effects of TGF-β. Biochimica et Biophysica Acta (BBA)-Molecular Cell Research. 2009;1793(7):1165-73.
Lin KM, Lin B, Lian IY, Mestril R, Scheffler IE, Dillmann WH. Combined and Individual Mitochondrial HSP60 and HSP10 Expression in Cardiac Myocytes Protects Mitochondrial Function and Prevents Apoptotic Cell Deaths Induced by Simulated Ischemia-Reoxygenation. Lippincott Williams & Wilkins. 2001(13).
Sidorik L, Kyyamova R, Bobyk V, Kapustian L, Rozhko O, Vigontina O, et al. Molecular chaperone, HSP60, and cytochrome P450 2E1 co-expression in dilated cardiomyopathy. Cell biology international. 2005;29(1):51-5.
Enomoto H, Mittal N, Inomata T, Arimura T, Izumi T, Kimura A, et al. Dilated cardiomyopathy-linked heat shock protein family D member 1 mutations cause up-regulation of reactive oxygen species and autophagy through mitochondrial dysfunction. Cardiovascular research. 2021;117(4):1118-31.
Momtazi G, Lambrecht B, Naranjo J, Schock B. Regulators of A20 (TNFAIP3): new drug-able targets in inflammation. American journal of physiology Lung cellular and molecular physiology. 2019;316(3):L456-L69.
Gupta S, Young D, Maitra R, Gupta A, Popovic Z, Yong S, et al. Prevention of cardiac hypertrophy and heart failure by silencing of NF-kappaB. Journal of molecular biology. 2008;375(3):637-49.
Gupta S, Sen S. Role of the NF-kappaB signaling cascade and NF-kappaB-targeted genes in failing human hearts. Journal of molecular medicine (Berlin, Germany). 2005;83(12):993-1004.
Jie W, Lee J, Liem D, Ping P. HSPA5 Gene encoding Hsp70 chaperone BiP in the endoplasmic reticulum. Gene. 2017;618:14-23.
Luo X, Yin J, Dwyer D, Yamawaki T, Zhou H, Ge H, et al. Chamber-enriched gene expression profiles in failing human hearts with reduced ejection fraction. Scientific reports. 2021;11(1):11839.

No competing interests reported.

SupplementaryMaterial.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

System analysis based on the necroptosis-related genes in heart failure

Status:

Version 1

Abstract

Figures

1 Introduction

2 Materials and Methods

2.1 Data acquisition and pre-processing

2.2 Model building

2.3 Cluster analysis

2.4 Expression distribution of regulatory factors of necroptosis genes

2.5 Protein-protein interaction network analysis

2.6 Enrichment Analysis

2.7 Immune Infiltration Analyses

2.8 Sample collection

2.9 RT-qPCR

2.10 Statistics analysis

3 Results

3.1 Data pre-processing

3.2 Screening and correlation analysis of NRGs

3.3 Model construction based on NRGs

3.4 Molecular typing based on NRGs

3.5 Regulatory network of characteristic NRGs

3.6 Analysis of the genetic biology of HF in different typing groups

3.7 Correlation analysis between NRGs’ expression and immune characteristic subtypes and infiltration level

3.8 Immune Infiltration Analyses

3.9 Validation of the NRDEGs in HF patients

4 Discussion

5 Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1