Skip to main content
Download PDF
Download ePub

Identification of potential therapeutic targets for nonischemic cardiomyopathy in European ancestry: an integrated multiomics analysis

Cardiovascular Diabetology volume 23, Article number: 338 (2024) Cite this article

Abstract

Background

Nonischemic cardiomyopathy (NISCM) is a clinical challenge with limited therapeutic targets. This study aims to identify promising drug targets for NISCM.

Methods

We utilized cis-pQTLs from the deCODE study, which includes data from 35,559 Icelanders, and SNPs from the FinnGen study, which includes data from 1,754 NISCM cases and 340,815 controls of Finnish ancestry. Mendelian randomization (MR) analysis was performed to estimate the causal relationship between circulating plasma protein levels and NISCM risk. Proteins with significant associations underwent false discovery rate (FDR) correction, followed by Bayesian colocalization analysis. The expression of top two proteins, LILRA5 and NELL1, was further analyzed using various NISCM datasets. Descriptions from the Human Protein Atlas (HPA) validated protein expression. The impact of environmental exposures on LILRA5 was assessed using the Comparative Toxicogenomics Database (CTD), and molecular docking identified the potential small molecule interactions.

Results

MR analysis identified 255 circulating plasma proteins associated with NISCM, with 16 remaining significant after FDR correction. Bayesian colocalization analysis identified LILRA5 and NELL1 as significant, with PP.H4 > 0.8. LILRA5 has a protective effect (OR = 0.758, 95% CI, 0.670–0.857) while NELL1 displays the risk effect (OR = 1.290, 95% CI, 1.199–1.387) in NISCM. Decreased LILRA5 expression was found in NISCM such as diabetic, hypertrophic, dilated, and inflammatory cardiomyopathy, while NELL1 expression increased in hypertrophic cardiomyopathy. HPA data indicated high LILRA5 expression in neutrophils, macrophages and endothelial cells within normal heart and limited NELL1 expression. Immune infiltration analysis revealed decreased neutrophil in diabetic cardiomyopathy. CTD analysis identified several small molecules that affect LILRA5 mRNA expression. Among these, Estradiol, Estradiol-3-benzoate, Gadodiamide, Topotecan, and Testosterone were found to stably bind to the LILRA5 protein at the conserved VAL-15 or THR-133 residues in the Ig-like C2 domain.

Conclusion

Based on European Ancestry Cohort, this study reveals that LILRA5 and NELL1 are potential therapeutic targets for NISCM, with LILRA5 showing particularly promising prospects in diabetic cardiomyopathy. Several small molecules interact with LILRA5, implying potential clinical implication.

Introduction

Nonischemic cardiomyopathy (NISCM) encompasses a group of various myocardial diseases that are not related to coronary artery disease [1]. These diseases include hypertrophic cardiomyopathy, dilated cardiomyopathy, restrictive cardiomyopathy, diabetic cardiomyopathy, and other forms of cardiomyopathy. They have diverse and often unclear underlying causes, including genetic predisposition, autoimmune responses, and toxin exposure, ultimately leading to heart failure and arrhythmias [2]. Despite the significant incidence of NISCM and its considerable impact on morbidity and mortality, it remains a clinical challenge due to the limited availability of effective therapeutic targets. Current treatment approaches are typically symptomatic and do not address the underlying causes of the disease, highlighting the urgent need for precise and effective therapeutic interventions.

Advances in genomic technologies and bioinformatics have opened new avenues for understanding the molecular mechanisms of complex diseases such as NISCM. One promising approach is the use of circulating plasma proteins as biomarkers and potential drug targets [3]. Protein quantitative trait loci (pQTL) studies [4], which investigate the genetic regulation of protein levels in the blood, can provide valuable insights into the causal relationships between proteins and diseases. Cis-pQTLs are genetic loci located near or at the genes they regulate, offering insights into how genetic variations control protein levels and complementing genomic data with a functional perspective [5]. Mendelian randomization (MR) is a robust analytical method that uses genetic variation as an instrumental variable to infer causal relationships and identify potential therapeutic targets [6]. Recently, the strategy of combining MR and Bayesian colocalization for identifying causal biomarkers and therapeutic targets has been successfully proposed [7,8,9]. By integrating cis-pQTL data with disease phenotypes and using methods such as MR and Bayesian colocalization, causal relationships between specific proteins and disease outcomes have been discovered.

The aim of this study is to discover new drug targets for NISCM by integrating genetic, proteomic, transcriptome and toxicogenomics. Relevant cis-pQTL and single nucleotide polymorphism (SNP) data from NISCM patients were extracted from public databases. By applying a stringent MR framework and performing FDR correction, we identified circulating plasma proteins that have a causal impact on NISCM. To eliminate biases caused by SNP linkage disequilibrium, we conducted Bayesian colocalization analysis on the identified circulating plasma proteins to facilitate precise drug target localization. In conclusion, this study identifies LILRA5 and NELL1 as potential therapeutic targets for NISCM, with LILRA5 ultimately showing particularly promising prospects in diabetic cardiomyopathy. Our investigation provides a strategy for identifying therapeutic target for NISCM.

Methods

Study design and data sources

A flowchart of the study design is provided in Fig. 1. Initially, we selected cis-pQTLs as instrumental variables from a comprehensive dataset of 4,907 distinct plasma protein pQTLs derived from the Ferkingstad study of 35,559 Icelanders, predominantly of European ancestry (deCODE study, https://www.decode.com/summarydata) [10]. Subsequently, we obtained SNP data associated with NISCM as the outcome variable from the FinnGen R10 dataset (https://storage.googleapis.com/finngen-public-data-r10). The FinnGen study includes 1,754 NISCM cases and 340,815 controls, all of Finnish ancestry, representing a significant European cohort for genome-wide association studies (GWAS). Then, we estimated the association between protein levels and patient outcomes using summary statistics from cis-pQTLs and GWAS data. Proteins with significant associations underwent FDR correction, followed by Bayesian colocalization analysis. The expression of the top two proteins, LILRA5 and NELL1, was further analyzed using various NISCM datasets. Then, using the Human Protein Atlas (HPA), we analyzed the expression of target proteins in normal heart tissue. Subsequently, using Comparative Toxicogenomics Database (CTD) databases, we curated the association of target proteins with environmental small molecule compounds. Finally, molecular docking with the potential small molecules was performed. The original studies, from which these datasets were derived, adhered to the principles of the Declaration of Helsinki and received ethical consent and approval. Consequently, no further ethical approval or informed consent was required for our current analysis.

Fig. 1
figure 1

Flowchart of the study design. The work conducted in this study and a summary of the main findings are presented

Full size image

Selection of instrumental variables

We selected (1) genetic variants associated with the target gene in cis positions (typically SNPs located at or near the target gene) within 1 Mb that can influence gene expression [11]; (2) variants that exhibited significant associations [12](p < 5 × 10−8); (3) variants situated outside the boundaries of the major histocompatibility complex region [11] (chromosome 6, 26–34 Mb); (4) variants that were independently associated through linkage disequilibrium clustering, with R2 < 0.1 [13]; (5) variants with F-statistic values > 10 [14]; and (6) variants with minor allele frequency > 0.01.

Mendelian randomization analysis

For two-sample MR, we used six regression models to validate the causal relationship between cis-pQTLs and NISCM: the Wald ratio method [15], MR‒Egger regression [16], random effects inverse-variance weighted regression [17], weighted median method [18], simple mode, and weighted mode. For single genetic instruments, the Wald ratio was calculated by dividing the effect size estimate of the variant–outcome association by the corresponding estimate of the variant–exposure association. When multiple SNPs were available, a meta-analysis of Wald estimates was performed using the inverse-variance weighted method. In the absence of pleiotropic heterogeneity or when heterogeneity is balanced, the inverse-variance weighted method provides an unbiased estimate. The MR‒Egger intercept test and Cochran’s Q statistic were used to detect the presence of heterogeneity or directional pleiotropy [19]. Visual inspection of forest plots, scatter plots, and leave-one-out plots was used to assess the MR assumption of “no pleiotropic heterogeneity”. MR analysis was performed using the TwoSampleMR R package (version 0.5.10;available at https://github.com/MRCIEU/TwoSampleMR/).

False discovery rate correction

To correct for multiple comparisons, the Benjamini‒Hochberg method was used for FDR correction, with the threshold set as 5% FDR. By controlling the FDR, we minimized the risk of falsely identifying significant results due to random variations in multiple comparisons, thus enhancing the reliability of our findings [20]. The analysis was conducted using R v4.3.3 software (https://cloud.r-project.org/).

Bayesian colocalization analysis

Bayesian statistical methods were used to estimate the probability that two or more phenotypes share the same causal SNPs [21]. To assess potential confounders, colocalization was assessed using the coloc R package (https://chr1swallace.github.io/coloc/). Visualization of the colocalization results was conducted using the LocusCompareR R package (https://github.com/boxiangliu/locuscomparer). To estimate the posterior probability that each genomic locus contains a single variant that affects both the protein and NISCM, all the SNPs within 1 Mb with minor allele frequency > 0.01 associated with cis-pQTLs were analyzed.

Access to GEO datasets

Using the GEO database (https://www.ncbi.nlm.nih.gov/geo/) with “cardiomyopathy” as the search keyword, the screening criteria were set as follows: (1) all samples tested must originate from human myocardial tissue; (2) the datasets must be related to NISCM and include a control group. Series Matrix File(s) and platform information that meet the above criteria were downloaded. These files contain all the statistical information for the GSE microarray and mRNA sequencing samples. The final selected datasets were GSE26887 (diabetic cardiomyopathy), GSE36961 (hypertrophic cardiomyopathy), GSE42955 (dilated cardiomyopathy), GSE4172 (inflammatory cardiomyopathy), GSE5406 (idiopathic cardiomyopathy), and GSE29819 (arrhythmogenic right ventricular cardiomyopathy). The GSE4172, GSE29819 series on the GPL570 platform (Affymetrix Human Genome U133 Plus 2.0 Array), the GSE5406 series on the GPL96 platform (Affymetrix Human Genome U133A Array), the GSE42955 and GSE26887 series on the GPL6244 platform (Affymetrix Human Gene 1.0 ST Array), and the GSE36961 series on the GPL15389 platform (Illumina HumanHT-12 V3.0 expression beadchip). Probe IDs were converted to homologous gene symbols using the platform’s annotation information.

Exploring target proteins expression and distribution in human protein atlas

The HPA (https://www.proteinatlas.org/) database was used to explore the transcriptome of LILRA5 and NELL1 in normal human heart tissues. And its expression in various cell types of normal heart was analyzed.

Comparison of differential gene expression and assessment of immune cell infiltration

The expression levels of target genes were screened across different groups. To assess the abundance of immune infiltration, the gene expression matrix data were uploaded to CIBERSORT (https://cibersort.stanford.edu/) to filter for the matrix of 22 immune cell types, obtaining the immune cell infiltration matrix. Visualization was performed using the R packages “ggplot2” or “reshape2”.

Exploring the chemical–protein interactions using comparative toxicogenomics database

Th CTD (https://ctdbase.org/) is a robust, publicly available database that aims to advance understanding about how environmental exposures affect human health. We used this database to curate information about chemical–protein interactions.

Molecular docking

Molecular docking was performed as described previously [22]. Briefly, compound structural files were downloaded from the PubChem website (https://pubchem.ncbi.nlm.nih.gov/) and converted from SDF to PDB files using Open Babel 2.3.2. Receptor proteins were obtained from the PDB database. The receptor proteins were prepared by removing water and ligands using PyMOL 2.3.4, and modifications such as adding hydrogens and balancing charges were performed using AutoDockTools. The receptor proteins and ligand molecules were converted to PDBQT format. Global molecular docking of the receptor proteins with the ligand molecules was conducted using AutoDock Vina 1.1.2. The docking results were analyzed using PLIP, and PyMol was used to visualize the docking results. The amino acid sequence LILRA5 was obtained from UniProt database (https://www.uniprot.org/). Finally, the binding sites and domains for each small molecule were identified.

Results

Causal effect estimation of plasma proteins on NISCM

We conducted a two-sample MR analysis to identify circulating plasma proteins causally associated with NISCM. We sourced cis-pQTLs from the deCODE study and, on the basis of cis-pQTL selection criteria, 58,304 SNPs representing 1,796 circulating plasma proteins were identified (Additional file 1: Table S1). NISCM data from the FinnGen R10 study (cases n = 1,754; controls n = 340,815) were evaluated as outcomes, with exposure information extracted and reverse structures removed, resulting in 55,410 SNPs (Additional file 1: Table S2). The MR analysis identified 255 plasma proteins related to NISCM (Additional file 1: Table S3 and Table S4). After FDR correction, 16 plasma proteins were found to be significantly associated with NISCM, namely, HIBCH, SAA1, BTD, SAA2, UROD, GMPR2, CNTN1, NME2, ASPN, NELL1, LCT, ACVRL1, TP53I3, C5, F7, and LILRA5 (Fig. 2). Furthermore, each instrumental variable linked to NISCM had an F-statistic value > 10 (Additional file 1 Table S2), confirming the relatively strong effect estimates of the instrumental variables. We applied Cochran’s Q statistic and MR–Egger intercept tests to detect pleiotropy [19]; neither test showed heterogeneity or pleiotropy for the 16 plasma proteins (Additional file 1: Table S5, Additional file 2: Fig. S1). Therefore, these 16 circulating plasma proteins were considered to have significant association with NISCM. Specifically, a positive correlation between genetically predicted plasma protein levels and NISCM risk was detected for 11 of the plasma proteins: HIBCH (OR: 1.282, 95% CI: 1.130–1.455, p = 1.178 × 10−4), SAA1 (OR: 1.280, 95% CI: 1.149–1.425, p = 6.99 × 10−6), BTD (OR: 1.194, 95% CI: 1.097–1.301, p = 4.59 × 10−5), SAA2 (OR: 1.297, 95% CI: 1.170–1.438, p = 7.44 × 10−7), UROD (OR: 1.741, 95% CI: 1.486–2.040, p = 6.71 × 10−12), GMPR2 (OR: 1.491, 95% CI: 1.211–1.835, p = 1.69 × 10−4), CNTN1 (OR: 1.459, 95% CI: 1.289–1.651, p = 2.33 × 10−9), NME2 (OR: 2.070, 95% CI: 1.466–2.922, p = 3.52 × 10−5), ASPN (OR: 1.203, 95% CI: 1.098–1.319, p = 7.73 × 10−5), NELL1 (OR: 1.290, 95% CI: 1.199–1.387, p = 8.15 × 10−12), and LCT (OR: 1.095, 95% CI: 1.055–1.136, p = 1.52 × 10−6). The other five plasma proteins showed a negative correlation with NISCM risk: ACVRL1 (OR: 0.642, 95% CI: 0.511–0.806, p = 1.372 × 10−4), TP53I3 (OR: 0.823, 95% CI: 0.743–0.912, p = 2.07 × 10−4), C5 (OR: 0.479, 95% CI: 0.342–0.670, p = 1.82 × 10−5), F7 (OR: 0.906, 95% CI: 0.860–0.954, p = 2.032 × 10−4), and LILRA5 (OR: 0.758, 95% CI: 0.670–0.857, p = 1.04 × 10−5) (Fig. 3).

Fig. 2
figure 2

Association between genetically predicted circulating plasma proteins and the risk of nonischemic cardiomyopathy (NISCM) visualized in a Manhattan plot. The x-axis shows the locations of genes that encode circulating plasma proteins, and the y-axis displays the − log10 of the Mendelian randomization (MR)-associated P values. Red and orange dashed lines indicate the 1% false discovery rate (FDR) and 5% FDR thresholds, respectively. Red dots indicate genes that are positively correlated with the development of NISCM; green dots indicate genes that are negatively correlated with the incidence of NISCM

Full size image
Fig. 3
figure 3

Mendelian randomization (MR) analysis of 16 plasma proteins associated with nonischemic cardiomyopathy (NISCM). Squares indicate causal estimates on the odds ratio (OR) scale, and whiskers indicate the 95% confidence intervals (CIs) for these ORs. nSNPs: number of SNPs used to estimate the causal effects; adj_P value: the P value was set at 5% false discovery rate (FDR); OR: odds ratio; CI: confidence interval

Full size image

Genetic colocalization analysis of cis-pQTLs with NISCM

To eliminate biases caused by SNP linkage disequilibrium, we conducted Bayesian colocalization analysis of the 16 circulating plasma proteins with NISCM, which also aids in precisely targeting drug candidates. The results consistently supported strong colocalization evidence for two of the circulating plasma proteins (LILRA5 and NELL1) with NISCM (PP.H4 > 0.8) as shown in Fig. 4; the 14 other plasma proteins had PP.H4 values < 0.8 (Table 1 and Additional file 3: Fig. S2).

Table 1 Details for bayesian colocalization analysis
Full size table
Fig. 4
figure 4

Genetic colocalization analysis between cis-pQTLs and nonischemic cardiomyopathy (NISCM). A) Left image: Associations from the NISCM GWAS; Bottom right image: Corresponding cis-pQTL results (LILRA5); Top right image: Colocalization of genetic associations and cis-pQTL signals. Purple diamonds indicate the SNPs. B) Left image: Associations from the NISCM GWAS; Bottom right image: Corresponding cis-pQTL results (NELL1); Top right image: Colocalization of genetic associations and cis-pQTL signals. Purple diamonds indicate the SNPs

Full size image

Differential expression analysis and protein distribution

Additionally, differential expression analysis of various NISCM datasets provided further context for the roles of LILRA5 and NELL1. We screened the GEO public database and selected the following datasets: GSE26887 (diabetic cardiomyopathy), GSE36961 (hypertrophic cardiomyopathy), GSE42955 (dilated cardiomyopathy), GSE5406 (idiopathic cardiomyopathy), GSE4172 (inflammatory cardiomyopathy), and GSE29819 (arrhythmogenic right ventricular cardiomyopathy). We found that LILRA5 expression was decreased in cardiac tissues from diabetic cardiomyopathy, hypertrophic cardiomyopathy, dilated cardiomyopathy, and inflammatory cardiomyopathy (Fig. 5A), while no significant differences were observed in idiopathic cardiomyopathy and arrhythmogenic right ventricular cardiomyopathy. In contrast, NELL1 expression was increased in cardiac tissues from hypertrophic cardiomyopathy, with no significant differences observed in other cardiomyopathies (Fig. 5A). Further validation of protein expression levels was conducted using data from the HPA. This resource confirmed the high expression of LILRA5 in neutrophils, macrophages and endothelial cells within cardiac tissue (Fig. 5B), underscoring its potential importance in cardiac immune responses. On the other hand, NELL1 could not be retrieved in normal cardiac tissue (data not shown).

Fig. 5
figure 5

Expression of LILRA5 and NELL1 in subtypes of nonischemic cardiomyopathy (NISCM) and cellular correlation of LILRA5 in normal heart muscle. A) Expression of LILRA5 and NELL1 in diabetic cardiomyopathy, hypertrophic cardiomyopathy, dilated cardiomyopathy, idiopathic cardiomyopathy, inflammatory cardiomyopathy, and arrhythmogenic right ventricular cardiomyopathy (ARVC). B) Pearson’s correlation heatmap between LILRA5 and representative markers of cardiomyocytes, mitotic cells, endothelial cells, smooth muscle cells, fibroblasts, macrophages, neutrophils, T cells, and plasma cells in heart muscle

Full size image

Immune cell infiltration analysis

Considering LILRA5 is mainly expressed in immune cells within normal human heart tissues (Fig. 5B), we further analyzed the immune cell infiltration in these NISCM subtypes. By estimating the relative proportions of 22 immune cell types in blood samples from different groups compared to controls, the analysis identified notable alterations. In diabetic cardiomyopathy, neutrophil expression was found to be downregulated (Fig. 6A). In hypertrophic cardiomyopathy, there was an increase in native B cells, regulatory T cells, gamma delta T cells, resting NK cells, and M2 macrophages, alongside a decrease in monocytes and activated dendritic cells (Fig. 6B). In dilated cardiomyopathy, a reduction in activated NK cell expression was observed (Fig. 6C). No significant differences in immune cell expression were noted in inflammatory cardiomyopathy (Fig. 6D).

Fig. 6
figure 6

Immune cells infiltration. A) Diabetic cardiomyopathy. B) Hypertrophic cardiomyopathy. C) Dilated cardiomyopathy. D) Inflammatory cardiomyopathy

Full size image

Potential drug target analysis

Through the CTD, we identified the effects of environmental exposures (small molecules) on LILRA5 expression (Table 2). We selected some small molecules that could upregulate LILRA5 expression and have been shown cardiovascular protection for further molecular docking experiments, and these molecules include Estradiol, Estradiol 3-benzoate, Gardiquimod, Genistein, Testosterone, Topotecan, N,N, N’,N’-tetrakis (2-pyridylmethyl)ethylenediamine(TPEN) and Methimazole. Some molecules with binding energies less than − 5 kcal/mol, including Estradiol 3-benzoate, Estradiol, Gardiquimod, Genistein, TPEN, Testosterone, and Topotecan, were found to stably bind to LILRA5 (Fig. 7A). Methimazole exhibited a binding energy of only − 2.9 kcal/mol with LILRA5 and thus was not shown in the results. To further analyze the specific amino acid sites where these small molecules bind to LILRA5 protein, the sites obtained by molecular docking are shown in Fig. 7B, as indicated by red color. It can be seen that the VAL-15 (V, purple background) and THR-133(T, green background) sites of LILRA5 protein are relative conserved binding sites. To further clarify the potential impact of small molecule binding sites on the function of LILRA5 protein, we further constructed the topology and domain of LILRA5 protein and found that these conserved binding sites of small molecules are mainly located on the immunoglobulin-like domains (Ig-like C2-type 1 and Ig-like C2-type 2) (Fig. 7C). The function of biological processes (BP) for LILRA5 indicates that LILRA5 plays a significant regulatory role in various biological processes, particularly in immune response and inflammation regulation (Additional file 4: Fig. S3).

Table 2 Interaction between chemicals and LILRA5 in the CTD
Full size table
Fig. 7
figure 7

Structural analysis of LILRA5 protein and its interaction with various ligands. A) 3D structures of LILRA5 protein and molecular docking of LILRA5 (cyan) with selected small molecules including Estradiol 3-benzoate, Estradiol, Gardiquimod, Genistein, TPEN, Testosterone and Topotecan (yellow). The binding energy (kcal/mol) for each ligand is shown below each image. Specific binding sites and key amino acid residues are annotated in the images. B) Sequence alignment of amino acids at the ligand binding sites. Amino acid residues critical for ligand binding are marked in red. Amino acid residues on purple and green background indicate the relative conserved sites. C) Schematic representation of the molecular structure of LILRA5. This includes the signal peptide (Signal,1–41), chain (42–299), transmembrane region (Transmen, 269–289), and cytoplasmic region (290–299). Ig-like C2-type domains 1 and 2 are indicated in the diagram. Key amino acids VAL-15 and THR-133 are marked at their respective positions

Full size image

Discussion

NISCM includes myocardial diseases not caused by coronary artery disease and presents significant challenges in cardiovascular medicine. Due to improved detection and aging populations, its prevalence is increasing [23]. Studies indicate a rise in NISCM among younger individuals due to genetic predisposition and lifestyle factors [24]. Clinically, its complexity and diversity pose numerous challenges. Integrating emerging technologies and interdisciplinary research is key to new therapeutic potentials and improving patient outcomes. Research into the genetic foundations of NISCM has identified multiple critical genes involved in cardiac structure and function [25, 26]. This study identified LILRA5 and NELL1 as potential therapeutic targets, highlighting LILRA5’s role in diabetic cardiomyopathy.

We conducted MR analysis on the circulating plasma proteome to identify proteins causally related to NISCM. Using multiple cis-pQTLs as instrumental variables, we performed a comprehensive analysis. After FDR correction, we identified proteins closely associated with NISCM pathology. Eleven proteins (HIBCH, SAA1, BTD, SAA2, UROD, GMPR2, CNTN1, NME2, ASPN, NELL1, LCT) were positively correlated with NISCM risk, while five (ACVRL1, TP53I3, C5, F7, LILRA5) were negatively correlated. Drug development and clinical translation often require significant time and financial investment, leading to lower success rates. Population-based genetic studies have successfully identified new drug targets or repurposed existing drugs; two-thirds of FDA-approved drugs in 2021 had genetic evidence support [27]. Next-generation sequencing of 208 patients identified NISCM-related genes [28], broadening our understanding of NISCM. RNA sequencing described the transcriptome of 880,000 nuclei from NISCM hearts, identifying genotype-related pathways, cell interactions, and gene expression differences at the single-cell level [29]. Although circulating plasma proteins do not have a direct histological relationship with NISCM like atrial or ventricular tissues, they are valuable for drug target screening due to their detectability and accessibility. Bayesian colocalization analysis of the 16 proteins identified by MR showed strong colocalization of LILRA5 and NELL1 with NISCM (PP.H4 > 0.8). High posterior probabilities indicate a shared genetic architecture between protein expression and NISCM risk, suggesting a direct role in disease pathogenesis. This strong genetic association highlights the potential of LILRA5 and NELL1 as therapeutic targets. Utilizing MR and colocalization analysis minimizes confounders and enhances the reliability of causal relationship identification, providing a solid foundation for subsequent therapeutic strategies.

LILRA5, a member of the leukocyte immunoglobulin-like receptor A subfamily, may trigger innate immune responses [30, 31]. GO enrichment analysis indicates LILRA5 plays a crucial role in immune response and inflammation regulation. It may influence the pathophysiology of NISCM through cytokine production, calcium ion transport, and protein tyrosine phosphorylation, supporting further research on its function and therapeutic potential. NELL1, a protein kinase C-binding protein involved in cell growth and differentiation, is a biomarker for idiopathic pulmonary fibrosis and is consistently expressed in both rats and humans [32]. Differential expression analysis shows LILRA5 is significantly reduced in diabetic, hypertrophic, dilated, and inflammatory cardiomyopathies, suggesting its broad involvement in these diseases. Interestingly, a recent study has also reported a negative correlation between LILRA5 and severe insulin-resistant diabetes with relative insulin deficiency; furthermore, this study suggests a significant association between LILRA5 and heart failure [33], which is highly consistent with our current findings. Conversely, NELL1 expression increases in hypertrophic cardiomyopathy, indicating a role in hypertrophic remodeling. HPA validation revealed high LILRA5 expression in neutrophils, macrophages and endothelial cells within cardiac tissue, highlighting its importance in cardiac immune responses. Limited NELL1 expression suggests a specific role in hypertrophic remodeling. Future research should establish multicenter cohorts and use large-scale proteomics and clinical samples to validate these proteins’ roles in NISCM.

Further immune infiltration analysis revealed significant changes in immune cell populations in NISCM. In diabetic cardiomyopathy, downregulation of neutrophils correlates with reduced LILRA5 expression, suggesting a link between neutrophil activity and LILRA5 [30]. In hypertrophic cardiomyopathy, increased activation of native B cells, regulatory T cells, gamma delta T cells, resting NK cells, and M2 macrophages, along with decreased in monocytes and activated dendritic cells, highlights a complex immune environment, aligning with recent findings [34]. In dilated cardiomyopathy, activated NK cell expression is reduced. No significant immune cell changes were observed in inflammatory cardiomyopathy. Studies show immune cells play a critical role in NISCM [33,34,35], consistent with our findings. Different NISCM subtypes exhibit distinct immune characteristics, suggesting precise immunomodulatory therapies as a potential treatment approach, highlighting the potential of immune-targeted therapies.

CTD analysis identified several small molecules affecting LILRA5 mRNA expression. Due to the negative correlation between LILRA5 and NISCM, we screened small molecules that upregulate LILRA5 and conducted molecular docking. Results showed that Benzoate estradiol, Estradiol, Gadodiamide, Genistein, TPEN, Testosterone, and Topotecan stably bind to LILRA5, mainly at its Ig-like domain, suggesting the domain’s role in binding and function. Conserved sites VAL-15 and THR-133 are crucial for these interactions. Pan-assay interference compounds (PAINs) are known to hinder drug development [35, 36]. Among our candidates, Genistein is a PAINs molecule due to nonspecific binding and false positives. TPEN, while not typically a PAINs molecule, may cause nonspecific interference due to metal-chelation. Other molecules (Topotecan, Estradiol-3-benzoate, Estradiol, Gadodiamide, and Testosterone) did not exhibit PAINs characteristics, making their binding to LILRA5 more reliable. Our findings suggest the Ig-like domain is key for ligand binding. Despite Genistein and TPEN’s potential nonspecific binding, other small molecules showed high specificity and stability. Studies have indicated these molecules benefit the cardiovascular system [37,38,39,40,41,42,43]. Our current findings provide a foundation for further research on LILRA5 function and LILRA5-targeted drug development. Considering the fact that these small molecules have the capacity to upregulate the expression of LILRA5 mRNA on the one hand, and on the other, they are able to interact with LILRA5 protein, thereby influencing its crucial domains, it implies that these molecules modulate the expression and functionality of LILRA5 via multifaceted mechanisms, but the precise mechanisms still need further research.

Strengths and limitations

A key advantage of this study is the use of the most comprehensive and high-quality pQTL database for circulating plasma proteins. Using cis-pQTLs as instrumental variables reduces the impact of environmental confounders typical in observational studies, minimizing pleiotropy. This strengthens causal inferences between gene expression and disease outcomes, helping identify genetic mechanisms and pathways in disease pathogenesis. The cis-pQTL data used SomaLogic SOMAscan technology, ensuring high selectivity, specificity, and sensitivity, which reduces false pQTL results. This enables large-scale pQTL studies to explore genetic variations in protein levels and their disease associations comprehensively. We also applied multiple testing corrections (FDR control) to minimize false positives from multiple comparisons.

There are some limitations in our study. This study is restricted to individuals of European ancestry and a single cohort, which may limit the generalizability of the findings. Integrating more databases would enhance the credibility of our conclusions. We used the GEO database to analyze and validate LILRA5 and NELL1 expression in different NISCM subtypes, future research should involve multicenter collaborations and large-scale proteomics data to further validate these proteins’ roles in NISCM. Finally, we used an R²<0.1 threshold for instrumental variable selection, increasing sample size and improving detection of causal relationships but introducing potential collinearity.

Conclusion

This study identifies LILRA5 and NELL1 as potential therapeutic targets for NISCM based on a European ancestry cohort, with LILRA5 particularly promising for diabetic cardiomyopathy. Integrating genetic, proteomic, transcriptomic, and toxicogenomic data provides a comprehensive understanding of NISCM pathogenesis and potential treatments. These findings underscore the importance of personalized medicine in NISCM treatment, targeting specific molecular profiles based on patients’ unique genetic and molecular characteristics.

Data availability

The datasets used in MR analyses and supported the conclusions of this article are included within the article and its additional files. The main R scripts and code have been made publicly available on GitHub and archived on Zenodo (DOI: 10.5281/zenodo.12977418). If further code is required, please contact the authors.

Abbreviations

CTD:

Comparative toxicogenomics database

FDR:

False discovery rate

HPA:

Human protein atlas

MR:

Mendelian randomization

NISCM:

Nonischemic cardiomyopathy

PAINs:

Pan-assay interference compounds

pQTL:

Protein quantitative trait loci

SNP:

Single nucleotide polymorphism

TPEN:

N, N,N’,N’-tetrakis (2-pyridylmethyl)ethylenediamine

References

  1. Maron BJ, Towbin JA, Thiene G, Antzelevitch C, Corrado D, Arnett D, Moss AJ, Seidman CE, Young JB. Contemporary definitions and classification of the cardiomyopathies: an American Heart Association Scientific Statement from the Council on Clinical Cardiology, Heart Failure and Transplantation Committee; quality of Care and Outcomes Research and Functional Genomics and Translational Biology Interdisciplinary Working Groups; and Council on Epidemiology and Prevention. Circulation. 2006;113(14):1807–16.

    Article  PubMed  Google Scholar 

  2. Zhao Y, Quan E, Zeng T, Huang Z, Luo Y, Peng L, Li S, Liu J, Chong Y, Cao H. Type 1 diabetes, its complications, and non-ischemic cardiomyopathy: a mendelian randomization study of European ancestry. Cardiovasc Diabetol. 2024;23(1):31.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. Nikolova AP, Hitzeman TC, Baum R, Caldaruse AM, Agvanian S, Xie Y, Geft DR, Chang DH, Moriguchi JD, Hage A, et al. Association of a Novel Diagnostic Biomarker, the plasma Cardiac bridging integrator 1 score, with heart failure with preserved ejection Fraction and Cardiovascular hospitalization. JAMA Cardiol. 2018;3(12):1206–10.

    Article  PubMed  PubMed Central  Google Scholar 

  4. Sun BB, Chiou J, Traylor M, Benner C, Hsu YH, Richardson TG, Surendran P, Mahajan A, Robins C, Vasquez-Grinnell SG, et al. Plasma proteomic associations with genetics and health in the UK Biobank. Nature. 2023;622(7982):329–38.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Zhang J, Dutta D, Köttgen A, Tin A, Schlosser P, Grams ME, Harvey B, Yu B, Boerwinkle E, Coresh J, et al. Plasma proteome analyses in individuals of European and African ancestry identify cis-pQTLs and models for proteome-wide association studies. Nat Genet. 2022;54(5):593–602.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Larsson SC, Butterworth AS, Burgess S. Mendelian randomization for cardiovascular diseases: principles and applications. Eur Heart J. 2023;44(47):4913–24.

    Article  PubMed  PubMed Central  Google Scholar 

  7. Zuber V, Grinberg NF, Gill D, Manipur I, Slob EAW, Patel A, Wallace C, Burgess S. Combining evidence from mendelian randomization and colocalization: review and comparison of approaches. Am J Hum Genet. 2022;109(5):767–82.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  8. Zheng J, Haberland V, Baird D, Walker V, Haycock PC, Hurle MR, Gutteridge A, Erola P, Liu Y, Luo S, et al. Phenome-wide mendelian randomization mapping the influence of the plasma proteome on complex diseases. Nat Genet. 2020;52(10):1122–31.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  9. Yang C, Fagan AM, Perrin RJ, Rhinn H, Harari O, Cruchaga C. Mendelian randomization and genetic colocalization infer the effects of the multi-tissue proteome on 211 complex disease-related phenotypes. Genome Med. 2022;14(1):140.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Ferkingstad E, Sulem P, Atlason BA, Sveinbjornsson G, Magnusson MI, Styrmisdottir EL, Gunnarsdottir K, Helgason A, Oddsson A, Halldorsson BV, et al. Large-scale integration of the plasma proteome with genetics and disease. Nat Genet. 2021;53(12):1712–21.

    Article  CAS  PubMed  Google Scholar 

  11. Fauman EB, Hyde C. An optimal variant to gene distance window derived from an empirical definition of cis and trans protein QTLs. BMC Bioinformatics. 2022;23(1):169.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Hemani G, Zheng J, Elsworth B, Wade KH, Haberland V, Baird D, Laurin C, Burgess S, Bowden J, Langdon R, et al. The MR-Base platform supports systematic causal inference across the human phenome. eLife. 2018;7:e34408.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Bowden J, Del Greco MF, Minelli C, Davey Smith G, Sheehan NA, Thompson JR. Assessing the suitability of summary data for two-sample mendelian randomization analyses using MR-Egger regression: the role of the I2 statistic. Int J Epidemiol. 2016;45(6):1961–74.

    PubMed  PubMed Central  Google Scholar 

  14. Burgess S, Thompson SG. Avoiding bias from weak instruments in mendelian randomization studies. Int J Epidemiol. 2011;40(3):755–64.

    Article  PubMed  Google Scholar 

  15. Burgess S, Butterworth A, Thompson SG. Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol. 2013;37(7):658–65.

    Article  PubMed  PubMed Central  Google Scholar 

  16. Bowden J, Davey Smith G, Burgess S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol. 2015;44(2):512–25.

    Article  PubMed  PubMed Central  Google Scholar 

  17. Zhang N, Li Y, Sundquist J, Sundquist K, Ji J. Identifying actionable druggable targets for breast cancer: mendelian randomization and population-based analyses. EBioMedicine. 2023;98:104859.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Bowden J, Davey Smith G, Haycock PC, Burgess S. Consistent estimation in mendelian randomization with some Invalid instruments using a weighted median estimator. Genet Epidemiol. 2016;40(4):304–14.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Verbanck M, Chen CY, Neale B, Do R. Detection of widespread horizontal pleiotropy in causal relationships inferred from mendelian randomization between complex traits and diseases. Nat Genet. 2018;50(5):693–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. Sun L, Craiu RV, Paterson AD, Bull SB. Stratified false discovery control for large-scale hypothesis testing with application to genome-wide association studies. Genet Epidemiol. 2006;30(6):519–30.

    Article  PubMed  Google Scholar 

  21. Giambartolomei C, Vukcevic D, Schadt EE, Franke L, Hingorani AD, Wallace C, Plagnol V. Bayesian test for colocalisation between pairs of genetic association studies using summary statistics. PLoS Genet. 2014;10(5):e1004383.

    Article  PubMed  PubMed Central  Google Scholar 

  22. Xu X, Yu Y, Yang L, Wang B, Fan Y, Ruan B, Zhang X, Dai H, Mei W, Jie W, et al. Integrated analysis of Dendrobium nobile extract Dendrobin a against pancreatic ductal adenocarcinoma based on network pharmacology, bioinformatics, and validation experiments. Front Pharmacol. 2023;14:1079539.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. Martin SS, Aday AW, Almarzooq ZI, Anderson CAM, Arora P, Avery CL, Baker-Smith CM, Barone Gibbs B, Beaton AZ, Boehme AK, et al. 2024 Heart Disease and Stroke statistics: a report of US and Global Data from the American Heart Association. Circulation. 2024;149(8):e347–913.

    Article  PubMed  Google Scholar 

  24. Finocchiaro G, Westaby J, Sheppard MN, Papadakis M, Sharma S. Sudden Cardiac death in young athletes: JACC State-of-the-art review. J Am Coll Cardiol. 2024;83(2):350–70.

    Article  PubMed  Google Scholar 

  25. Jansweijer JA, Nieuwhof K, Russo F, Hoorntje ET, Jongbloed JD, Lekanne Deprez RH, Postma AV, Bronk M, van Rijsingen IA, de Haij S, et al. Truncating titin mutations are associated with a mild and treatable form of dilated cardiomyopathy. Eur J Heart Fail. 2017;19(4):512–21.

    Article  CAS  PubMed  Google Scholar 

  26. Gray MP, Fatkin D, Ingles J, Robertson EN, Figtree GA. Genetic testing in cardiovascular disease. Med J Australia. 2024;220(8):428–34.

    Article  PubMed  Google Scholar 

  27. Ochoa D, Karim M, Ghoussaini M, Hulcoop DG, McDonagh EM, Dunham I. Human genetics evidence supports two-thirds of the 2021 FDA-approved drugs. Nat Rev Drug Discovery. 2022;21(8):551.

    Article  CAS  PubMed  Google Scholar 

  28. Vokač D, Stangler Herodež Š, Krgović D, Kokalj Vokač N. The role of Next-Generation sequencing in the management of patients with suspected non-ischemic cardiomyopathy after Syncope or termination of sudden arrhythmic death. Genes. 2024;15(1):72.

    Article  PubMed  PubMed Central  Google Scholar 

  29. Reichart D, Lindberg EL, Maatz H, Miranda AMA, Viveiros A, Shvetsov N, Gärtner A, Nadelmann ER, Lee M, Kanemaru K, et al. Pathogenic variants damage cell composition and single cell transcription in cardiomyopathies. Science. 2022;377(6606):eabo1984.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  30. Borges L, Kubin M, Kuhlman T. LIR9, an immunoglobulin-superfamily-activating receptor, is expressed as a transmembrane and as a secreted molecule. Blood. 2003;101(4):1484–6.

    Article  CAS  PubMed  Google Scholar 

  31. Li Y, Sun Y, Liu Y, Wang B, Li J, Wang H, Zhang H, Wang X, Han X, Lin Q, et al. Genome-wide meta-analysis identifies susceptibility loci for autoimmune hepatitis type 1. Hepatology. 2022;76(3):564–75.

    Article  CAS  PubMed  Google Scholar 

  32. Bauer Y, Tedrow J, de Bernard S, Birker-Robaczewska M, Gibson KF, Guardela BJ, Hess P, Klenk A, Lindell KO, Poirey S, et al. A novel genomic signature with translational significance for human idiopathic pulmonary fibrosis. Am J Respir Cell Mol Biol. 2015;52(2):217–31.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Gurung RL, Zheng H, Lee BTK, Liu S, Liu JJ, Chan C, Ang K, Subramaniam T, Sum CF, Coffman TM, et al. Proteomics profiling and association with cardiorenal complications in type 2 diabetes subtypes in Asian population. Diabetes Res Clin Pract. 2024;214:111790.

    Article  CAS  PubMed  Google Scholar 

  34. Zheng X, Yang Y, Huang Fu C, Huang R. Identification and verification of promising diagnostic biomarkers in patients with hypertrophic cardiomyopathy associate with immune cell infiltration characteristics. Life Sci. 2021;285:119956.

    Article  CAS  PubMed  Google Scholar 

  35. Baell J, Walters MA, Chemistry. Chemical con artists foil drug discovery. Nature. 2014;513(7519):481–3.

    Article  CAS  PubMed  Google Scholar 

  36. Baell JB. Feeling Nature’s PAINS: Natural products, natural product drugs, and Pan Assay Interference Compounds (PAINS). J Nat Prod. 2016;79(3):616–28.

    Article  CAS  PubMed  Google Scholar 

  37. Dworatzek E, Mahmoodzadeh S, Schriever C, Kusumoto K, Kramer L, Santos G, Fliegner D, Leung YK, Ho SM, Zimmermann WH, et al. Sex-specific regulation of collagen I and III expression by 17β-Estradiol in cardiac fibroblasts: role of estrogen receptors. Cardiovasc Res. 2019;115(2):315–27.

    Article  CAS  PubMed  Google Scholar 

  38. Prossnitz ER, Barton M. The G-protein-coupled estrogen receptor GPER in health and disease. Nat Rev Endocrinol. 2011;7(12):715–26.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Cho KHT, Fraser M, Wassink G, Dhillon SJ, Davidson JO, Dean JM, Gunn AJ, Bennet L. TLR7 agonist modulation of postasphyxial neurophysiological and cardiovascular adaptations in preterm fetal sheep. Am J Physiol Regul Integr Comp Physiol. 2020;318(2):R369–78.

    Article  CAS  PubMed  Google Scholar 

  40. Wei TT, Chandy M, Nishiga M, Zhang A, Kumar KK, Thomas D, Manhas A, Rhee S, Justesen JM, Chen IY, et al. Cannabinoid receptor 1 antagonist genistein attenuates marijuana-induced vascular inflammation. Cell. 2022;185(10):1676–e16931623.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  41. Shmist YA, Kamburg R, Ophir G, Kozak A, Shneyvays V, Appelbaum YJ, Shainberg A. N,N,N’,N’-tetrakis(2-pyridylmethyl)-ethylenediamine improves myocardial protection against ischemia by modulation of intracellular Ca2 + homeostasis. J Pharmacol Exp Ther. 2005;313(3):1046–57.

    Article  CAS  PubMed  Google Scholar 

  42. Lincoff AM, Bhasin S, Flevaris P, Mitchell LM, Basaria S, Boden WE, Cunningham GR, Granger CB, Khera M, Thompson IM Jr, et al. Cardiovascular Safety of testosterone-replacement therapy. N Engl J Med. 2023;389(2):107–17.

    Article  CAS  PubMed  Google Scholar 

  43. Zhang Y, Xu Y, Zhou K, Kao G, Yan M, Xiao J. Hypoxia-inducible transcription factor-1α inhibition by topotecan protects against lipopolysaccharide-induced inflammation and apoptosis of cardiomyocytes. Biomed Eng Online. 2021;20(1):88.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We appreciate all the participants in the study. We thank Margaret Biswas, PhD, from Liwen Bianji (Edanz) (www.liwenbianji.cn/) for editing the English text of a draft of this manuscript.

Funding

This work was supported by the Hainan Key Research and Development Project (ZDYF2020122, ZDYF2022SHFZ038), the Innovational Fund for Scientific and Technological Personnel of Hainan Province (KJRC2023B07). the National Natural Science Foundation of China (82060053, 82260083, U220A20270), the Cardiovascular Disease Research Science Innovation Group of Hainan Medical University, and the Innovative Research Projects for Graduate Students in Hainan Province (Qhyb2022-140).

Author information

Author notes

  1. Kaijia Shi and Xu Chen contributed equally to this work.

Authors and Affiliations

  1. Key Laboratory of Tropical Translational Medicine of Ministry of Education & Key Laboratory for Tropical Cardiovascular Diseases Research of Hainan Province, School of Public Health, Hainan Medical University, Haikou, 571199, China

    Kaijia Shi, Yangyang Zhao, Peihu Li, Jinxuan Chai, Jianmin Qiu, Junli Guo & Wei Jie

  2. Department of Intensive Care Unit, Yuhuangding Hospital, Yantai, 264000, China

    Xu Chen

  3. Department of Pathology and Pathophysiology, School of Basic Medicine Sciences, Guangdong Medical University, Zhanjiang, 524023, China

    Zhihua Shen

Authors
  1. Kaijia Shi
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Xu Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Yangyang Zhao
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Peihu Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Jinxuan Chai
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Jianmin Qiu
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Zhihua Shen
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Junli Guo
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Wei Jie
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

K.S. and X.C. significantly contributed to the conceptualization, investigation, visualization, formal analysis, data curation, and methodology development of the study. Y.Z., J.C., P.L. and J.Q. assisted in enhancing the visualization, optimizing software utilization, and managing data curation. K.S. and X.C. were primarily responsible for drafting the manuscript. Additionally, Z.S., J.G., and W.J. were involved in project administration and provided supervision. They also played a critical role in reviewing and editing the manuscript to ensure its academic rigor and coherence. Moreover, K.S., J.G., and W.J. secured the funding necessary to support this project. All authors critically reviewed the manuscript.

Corresponding authors

Correspondence to Zhihua Shen, Junli Guo or Wei Jie.

Ethics declarations

Competing interests

The authors declare no competing interests.

Ethical approval

Ethical approval was waived because this study used the data from publicly available databases.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Shi, K., Chen, X., Zhao, Y. et al. Identification of potential therapeutic targets for nonischemic cardiomyopathy in European ancestry: an integrated multiomics analysis. Cardiovasc Diabetol 23, 338 (2024). https://doi.org/10.1186/s12933-024-02431-8

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12933-024-02431-8

Keywords

Cardiovascular Diabetology

ISSN: 1475-2840