Differential expression profiles of plasma exosomal microRNAs in dilated cardiomyopathy with chronic heart failure

Abstract As one of the most prevalent heritable cardiovascular diseases, dilated cardiomyopathy (DCM) induces cardiac insufficiency and dysfunction. Although genetic mutation has been identified one of the causes of DCM, the usage of genetic biomarkers such as RNAs for DCM early diagnosis is still being overlooked. In addition, the alternation of RNAs could reflect the progression of the diseases, as an indicator for the prognosis of patients. Therefore, it is beneficial to develop genetic based diagnostic tool for DCM. RNAs are often unstable within circulatory system, leading to the infeasibility for clinical application. Recently discovered exosomal miRNAs have the stability that is then need for diagnostic purpose. Hence, fully understanding of the exosomal miRNA within DCM patients is vital for clinical translation. In this study, we employed the next generation sequencing based on the plasma exosomal miRNAs to comprehensively characterize the miRNAs expression in plasma exosomes from DCM patients exhibiting chronic heart failure (CHF) compared to healthy individuals. A complex landscape of differential miRNAs and target genes in DCM with CHF patients were identified. More importantly, we discovered that 92 differentially expressed miRNAs in DCM patients undergoing CHF were correlated with several enriched pathways, including oxytocin signalling pathway, circadian entrainment, hippo signalling pathway‐multiple species, ras signalling pathway and morphine addiction. This study reveals the miRNA expression profiles in plasma exosomes in DCM patients with CHF, and further reveal their potential roles in the pathogenesis of it, presenting a new direction for clinical diagnosis and management of DCM patients with CHF.


| INTRODUC TI ON
Dilated cardiomyopathy (DCM), one of the key contributors to chronic heart failure (CHF), is characterized by the dilatation of left ventricle along with systolic dysfunction. Notably, DCM is often correlated with a raised risk of serious arrhythmia, suggesting the pathological involvement of the cardiac conducting system. As the disease progresses, diastolic dysfunction and impaired right ventricular function will develop, ultimately resulting in HF and premature death. Epidemiological surveys revealed that, in the United States, the prevalence of DCM was 36 cases per 100,000 (i.e. 1:2500). Ten thousand people die from DCM each year and 46,000 are hospitalized as a result. [1][2][3][4] Additionally, population in Africa and Latin America have a higher prevalence compared with United States. 5 Alarmingly, the prevalence may be even underestimated due to the large number of asymptomatic patients.
DCM's pathogenesis has not been well understood. Familial and genetic predisposition are commonly regard as contributors.
Previous studies indicate that several genes may contribute to the initiation, progression and pathology of DCM, including titin (TTN), beta-myosin heavy chain (MYH7), Cardiac troponin T(TNNT2), Desmoplakin (DSP), lamin A/C (LMNA), type V voltage-gated cardiac Na channel gene (SCN5A), RNA-binding motif protein 20 (RBM20) et al. [6][7][8][9][10] However, although these genes are apparently associated with DCM, limited are directly contributing to the development of DCM due to variations in genetics. Fortunately, with the development of next generation sequencing, an increasing number of DCMassociated genes have been identified, 11,12 which opens up the potential for early diagnosis.
Advances in molecular biology over the past decades have revealed that non-coding RNA (ncRNA) that was once considered as 'junk' RNA plays vital roles in regulating distinct cellular processes.
Among the ncRNA, miRNAs have been widely studied and better understood. miRNA, which are single-stranded, conserved RNAs, comprised of approximately 18 ~ 25 nucleotides. It exists in both the supernatant and extracellular vesicles (EVs), 13,14 exerting regulatory effects via targeting mRNAs for cleavage or translational suppression. 15 EVs are highly heterogeneous and can be secreted by almost all cell types. Exosomes are the smallest subgroup of EVs, with a dimension of 40-160 nm. 16 Previous studies suggest that exosomal miRNAs are more stabilized, compared with in supernatants, which precisely reflect the influence on gene regulation. [17][18][19] For example, Li et al. founded that between healthy and preeclamptic patients, 7 miRNAs were differentially expressed in plasma exosomes, but only one of these can be detected in whole plasma miRNA. 17 Despite the progression of exosomal miRNAs for diagnosis, limited studies have looked into their potential for DCM early diagnosis.
Hence, we tested the expression profiling of plasma exosomal miRNAs of DCM with CHF patients. By comparative analysis, we identified the differentially expressed exosomal miRNAs. Target gene prediction and functional enrichment analysis were performed via bioinformatics methods, in order to reveal the underlying pathological changes of this disease. The result from this study may provide novel insights for the diagnosis and therapy of DCM with CHF patients.

| Participants
Approval of this research was obtained from the institutional review board for the First Affiliated Hospital of Zhengzhou University (2020-KY-142). The DCM with CHF patients were recruited from the First Affiliated Hospital of Zhengzhou University. DCM was diagnosed according to the diagnostic criteria proposed by the American Heart Association expert consensus panel and the European Society of Cardiology, 20 and the New York Heart Association (NYHA) criteria were applied to assess the cardiac function grade. 21 The inclusion criteria for the DCM with HF patients were as follows: (1) left ventricular ejection fraction (LVEF) is equal or lesser than 45% and had heart function grades ranging from II to IV; (2) patients with a history of congestive HF more than 6 months; and (3) informed consent was signed by all participates. The exclusion criteria were as follows: (1)

| Plasma sample collection
Participants diagnosed with DCN and HF, as well as healthy controls were recruited for the study. Venous blood samples, collected using Na-EDTA tubes to prevent coagulation, amounted to a minimum volume of 6 mL per participant. The blood samples were subjected to centrifugation at 2000 × g for 15 min at room temperature.
Subsequently, the upper plasma fraction, with an ideal volume of 2.5 mL per sample, was promptly frozen at −80°C to maintain the stability and integrity of the plasma biomolecules.

| Exosome isolation
Plasma exosomes were extracted using Exoquick reagent (EXOQ5A1; System Biosciences, USA) in accordance with the manufacturer's protocol. Briefly, 250 μL of plasma sample was supplemented with 36 μL of ExoQuick Exosome Precipitation Solution. The mixed solution was set for 30 min at 4°C, followed by centrifugation at 1500 × g for 30 min. After the centrifugation, the supernatant was discarded before additional centrifugation (1500 × g for 5 min). Finally, 100 μL of sterile phosphate-buffered saline (PBS) was add to resuspend exosome pellet. The solution was stored at −80 °C.

| Transmission electron microscopy
The transmission electron microscopy (TEM) was used to characterize the morphology character of isolated exosomes. Briefly, PBS was used to dilute extracted exosomes, before being dripped onto a copper mesh grid coated in holey carbon. The solution was set for 10 min, before the excess solution was blotted away with filter paper.
The sample was further soaked in 3% of glutaraldehyde for 5 min, prior to the wash process with 10 times washing using de-ionized water (2 min each time). Afterwards, 4% uranylacetate solution was dropped on the samples for 10 min, followed by 1% methylcellulose fixation for 5 min. Finally, the sample was air-dried at room temperature for 30 minters and examined by TEM later.

| Flow nanoanalyzer
The concentration and diameter distribution of extracted exosomes were determined by Flow NanoAnalyzer (FL Sciences).

| Western blot
The extracted exosomes were lysed with precooled RIPA lysis buffer

| RNA extraction
The total RNA that containing miRNA in plasma exosomes was ex-  The raw data were analysed via using ACGT101-miR (LC Sciences, USA) to remove adapter dimers, low complexities, junk, repeats (http://www.girin st.org/repbase), common RNA families, including tRNA, rRNA, snRNA and snoRNA (http://rfam.sanger.ac.uk/), and sequence length <18 nucleotide (nt) or >26 nt. Based on miRBase 22.0 (http://www.mirba se.org/), the unique sequences with the length of 18 ~ 26 nt were mapped to miRNA sequences, to recognize confirmed miRNAs and pre-miRNAs (including novel 3p-and 5p-derived miR-NAs). Also, after the procedure, data normalization was applied based on the method described previously. [22][23][24] The remaining unmapped sequences were further screened against Homo sapiens genomic sequences to identify potential novel miRNAs. RNAfold was used to predict the secondary structure of miRNAs (http://rna.tbi.univie. ac.at/cgi-bin/RNAWe bSuit e/RNAfo ld.cgi), in order to confirm the results of putative miRNAs in Homo sapiens. Raw sequencing data from this study are uploaded to the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/).

| Protein-protein interaction network construction and hub genes identification
Protein-protein interaction (PPI) network analysis was done using The Search Tool for the Retrieval of Interacting Genes (STRING) database. 28 PPI pairs with a combined score >0.4 (medium confidence score) were applied for the PPI network establishment.
Subsequently, the Cytoscape software (http://www.cytos cape. org/) was used to visualize the PPI network. As a commonlyused plugin, CytoHubba calculates the degree of each node in Cytoscape. Nodes with a higher degree or connectivity indicated the highly interacting proteins or genes in the network. In our study, nodes with a degree >5 were recognized as hub genes in the network.

| Establishment of the miRNA-mRNA regulatory network
The miRNet database (https://www.mirnet.ca) is an easy-to-use platform for the visualization and analysis of miRNA-centric network, designing to help unravel microRNA functions through consolidating existing knowledge with users' data. 29 The miRNet database was conducted to recognize miRNAs targeting hub genes in this study. The miRNA-mRNA network of DCM accompanied with CHF was constructed by Cytoscape software.

| Calculation of the global difference between a pair of expression profiles
We applied two different methods to calculate the global difference between a pair of expression profiles 30 : The Euclidean distance, where x i and y i are the expression of hub miRNA i over two expression profiles (DCM-Exo and Health-Exo) with p and q samples (x 1 , x 2 , …, x p ), (y 1 , y 2 , …, y q ).

| Dimensional reduction analysis
The uniform manifold approximation and projection (UMAP) using the umap-learn package (https://umap-learn.readt hedocs.io/en/ lates t/) was applied for dimensional reduction of internal transcriptomics data based on DE miRNAs. To test the discrimibility of identified hub miRNAs, we performed principal coordinates analysis (PCoA) and analysis of similarities (ANOSIM) on them. Bray-Curtis dissimilarity matrix was calculated by beta_diversity.py, and Bray-Curtis diversity was calculated using the R package Vegan with the function vegdist.

| Statistical analysis
All statistical tests were two-sided. p-value <0.05 and FDR <0.05 were suggested to be statistically significant. The mean ± standard deviation for descriptive statistics was used for continuous variables with a normal distribution. The Wilcoxon rank-sum test or Student's t-test was applied to compare continuous variables, and categorical variables were compared through the chi-squared or Fisher exact test. 31 All data processing, statistical analysis and plotting were conducted with R 4.1.3 software.

| Identification of isolated plasma exosomes
An overall flowchart of the study design is shown in Figure 1A.
Plasma exosomes were characterized for their diameter, morphology and the exosome surface proteins such as CD63 and CD81. TEM images ( Figure 1B) demonstrated that isolated plasma exosomes were cuplike constructs. High sensitivity flow cytometry for nanoparticle analysis revealed that both DCM-Exo and Nor-Exo had a similar median diameter about 81 nm ( Figure 1C). Moreover, these exosomes were all positive expression of CD63 and CD81 proteins but negative expression of Calnexin proteins on Western blotting ( Figure 1D). As a result, we have accurately identified plasma exosomes for the further high-throughput sequencing.

| Exosomal miRNA profile of DCM with CHF patients by RNA sequencing
A total of 3687 miRNAs were detected in the plasma exosome in the plasma exosome of the DCM patients with CHF patients and healthy controls (Table S1). The heatmap demonstrated the landscape of miRNA profile (Figure 2A). Scatter plot demonstrated all expressed exosomal miRNAs between DCM with CHF and controls ( Figure 2B).
The conservation analysis of all expressed exosomal miRNAs also had been performed ( Figure 2C).   (Table S1). Based on p-value less than 0.05 and |log fold change (FC)| more than 0.5, we determined 92 DE-miRNAs for subsequent analysis, including 48 upregulated and 44 downregulated miRNAs ( Figure 3A, Table 1).
The details of downregulated miRNAs were further demonstrated in the volcano map and bar plot ( Figure 3B,C). We further performed Pearson correlation analysis and UMAP analysis on the miRNA profiles to evaluate the differences and similarities in miRNA expression between the samples of DCM with CHF patients and healthy control participants ( Figure 3D,E). These results have revealed a significant difference between DCM and healthy individuals on their plasma exosomal miRNA landscapes.

| Analysis of DE-miRNAs targeted genes
TargetScan and miRanda, two target prediction software, were used  (Table S2). These 75 targets genes could play majors roles in the pathogenic regulatory mechanism of DCM.

| Overrepresentation gene set analysis of DE-miRNAs targeted genes
To confirm the accuracy and reliability of targeted genes prediction, we performed gene set analysis on 75 targets of DE-miRNAs by using MsigDB. As shown in Figure 4A, consistent with our data sources, these targets significantly enriched in C3 regulatory target gene sets and C8 cell type signature gene sets. This analysis indicates that the predicted target genes belong to gene sets that were closely correlated with the plasma exosomal miRNAs of DCM with CHF.

| GO Enrichment analysis of DE-miRNAs targeted genes
GO software was used to annotate the functions of the 75 pre-  Figure 4D. To further obtain the relationships between these top 20 terms, a functional annotation network was constructed ( Figure 4E,F).

| Protein-protein interaction network analysis
The 1176 matched genes of DE-miRNAs through TargetScan and miRanda were imported into the STRING database. Excluding the isolated targeted genes without interaction, there was a total of 156 target genes were mapped into the PPI network (confidence score cut-off value 900), which comprised with 156 nodes and 263 edges.
We also used Cytoscape software to visualize the interactions between the genes ( Figure 5A). According to the CytoHubba plugin compute, 23 nodes with a degree >5 were confirmed as hub genes in the PPI network. These 23 hub genes are presented in Figure 5B and Table 2.
The specific roles of DE-miRNAs and their target hub genes were further investigated using the comprehensive biological annotation analysis ( Figure S1). Based on WikiPathway annotation Furthermore, BioPlanet database annotation confirmed they were significantly associated with the activation of signalling mediated by EGFR, VEGF and VEGFR.

| miRNA-mRNA interaction network analysis
Aiming to understand the molecular mechanisms of previously identified exosomal DE-miRNAs in DCM accompanied with CHF, miRNet was used to recognize miRNAs targeting hub genes. As a result, a miRNA-mRNA regulatory network with 55 nodes (32 miRNAs and 23 hub genes) and 59 edges was constructed ( Figure 5C).

| Validation of hub miRNAs performance in an external cohort
The top 10 miRNAs that were significantly dysregulated between DCM and healthy individuals in silico were identified as hub miR-  silico ( Figure 6A). We then applied one measure of divergence between a pair of expression profiles, the Euclidean distance to investigate the global shifts in the hub miRNA expression profile between and within DCM and healthy controls in the external cohort. Relative differences between the distributions were consistent for this metric of expression divergence. The expression distance between DCM and healthy controls or within the DCM samples was significantly larger than the distance within normal healthy controls ( Figure 6B).
Principal coordinates analysis (PCoA) of hub miRNA profile in the internal silico cohort showed that significant shifts separated DCM from healthy controls. Similar results were observed from our external independent cohort, where there was a significant difference in the hub miRNA profiles between them ( Figure 6C).

| DISCUSS ION
In the present study, a specific miRNA signature of plasma exosomes was identified in DCM accompanied with CHF patients. We Circadian entrainment is endogenous oscillations, widely found in biological species, that possess the ability of entraining to the 24 h light-dark cycle. Studies have revealed that circadian dysregulation is closely associated with increased risk of cardiovascular diseases, including hypertension, atherosclerosis, stroke and myocardial infarction. 39 The nuclear receptor Rev-erbα/β, the key constituent of the circadian clock, emerges as a drug target for cardiovascular diseases.

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Any additional information required to reanalyse the data reported in this paper is available from the corresponding author contact upon request.