Exploring the causality and pathogenesis of atrial fibrillation with dilated cardiomyopathy: An integrated multi-omics approach

doi:10.21203/rs.3.rs-5209187/v1

Download PDF

Research Article

Exploring the causality and pathogenesis of atrial fibrillation with dilated cardiomyopathy: An integrated multi-omics approach

https://doi.org/10.21203/rs.3.rs-5209187/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background: Atrial fibrillation (AF) is the most prevalent sustained arrhythmia, and recent evidence indicates the presence of cardiac enlargement in patients with AF. Dilated cardiomyopathy (DCM), the most common form of cardiomyopathy, is characterized by significant heart dilation and AF. However, the risk factors and underlying mechanisms linking DCM to AF remain poorly understood.

Methods: Mendelian randomization (MR) analysis was initially used to explore the potential causal relationship between AF and DCM. Data were sourced from the public database Gene Expression Omnibus (GEO), and differentially expressed genes (DEGs) and significant module genes were identified using the Limma package and weighted gene co-expression network analysis (WGCNA). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses, as well as a protein-protein interaction (PPI) network analysis, were performed on the intersected genes. Hub genes were screened using machine learning algorithms. The identification of hub genes within the DCM GSE17800 dataset was achieved using the receiver operating characteristic (ROC) curve and nomogram, which were employed to assess the diagnostic efficacy of these hub genes. Finally, the immune infiltration of DCM and the microRNA (miRNA) interaction network involving hub genes were evaluated.

Results: MR analysis demonstrated that genetic susceptibility to AF was significantly associated with an increased risk of DCM (β: 20.44, 95% CI: 15.00-25.88, p=0.0002). The AF dataset included 1850 DEGs and 572 significant module genes, and the DCM dataset included 6463 DEGs, which had a total of 209 intersected genes with module genes for AF. After correlation enrichment analysis and PPI interaction network on the intersected genes, machine learning was used to screen two hub genes (VSNL1 and ETNPPL) that had high diagnostic efficacy (area under the curve from 0.81 0.89). Immune infiltration analysis of these genes revealed a relatively normal immune status for DCM, with a wider miRNA interaction network for VSNL1.

Conclusion: MR data suggests that genetic changes in the presence of AF are significantly associated with an increased risk of DCM. The two identified hub genes (VSNL1 and ETNPPL) can be used to diagnose comorbid DCM in patients with AF.

atrial fibrillation

dilated cardiomyopathy

mendelian randomization analysis

bioinformatics analysis

machine learning

diagnosis

Atrial fibrillation (AF), a common cardiac arrhythmia, affects more than 30 million people worldwide (1, 2). AF is characterized by disturbances in the electrical activity of the atria and ineffective conduction, leading to reduced or even loss of atrial contractile function (3, 4). The clinical manifestations of AF vary and may be asymptomatic in mild cases, whereas in severe cases, it may induce serious complications, such as heart failure and stroke (5, 6). The duration and frequency of AF have a greater impact on symptoms, which tend to be more pronounced in patients with persistent and permanent AF (7).

Dilated cardiomyopathy (DCM) is a primary cardiomyopathy of undetermined cause, in the course of which damage and remodeling of cardiomyocytes may lead to structural and electrophysiological changes in the atria, thereby promoting the development of AF (8, 9). However, there is a lack of research on the mechanisms by which DCM occurs in patients with AF. The theory is that the onset of AF leads to increased myocardial injury by increasing myocardial oxygen consumption and decreasing myocardial perfusion (10). Factors such as heart failure and myocardial ischaemia accompanying persistent and permanent AF also increase the risk of DCM (7, 11). The genetic background of AF leading to DCM is unknown, and exploring common alterations in molecular expression would be beneficial for clinical diagnosis.

In the present study, we used pooled data from previous genome-wide association studies (GWASs) for Mendelian randomization (MR) analyses to explore the causal relationship between AF and DCM. The use of genetic variants as instrumental variables (IVs) in MR analyses effectively avoids reverse causation and potential confounders and allows for accurate assessment of potential causality between exposure and outcome (12). We downloaded AF and DCM datasets from the Gene Expression Omnibus (GEO) database and screened for differentially expressed genes (DEGs) using Limma. We identified important module genes using weighted gene co-expression network analysis (WGCNA) and performed functional enrichment analysis. Protein-protein interaction (PPI) networks and machine learning algorithms, including the least absolute shrinkage and selection operator (LASSO), random forests (RF), and support vector machines (SVM), were used for hub gene identification. We evaluated diagnostic biomarkers for AF combined with DCM using a nomogram and receiver operating characteristic (ROC) curves. Finally, we analyzed immune infiltration in DCM and predicted potential target miRNAs of the hub genes.

2.1 Data collection and data processing

An overview of our study is shown in Figure 1. MR data were screened from GWASs (http://www.mrbase.org/) of European ancestry (Table 1). Effect estimates of single-nucleotide polymorphisms (SNPs) associated with the risk of AF were obtained from a study including 3518 patients and 459415 controls. Genetic IVs for DCM were also extracted from a GWASs study that included 1444 patients and 353937 controls (13). Transcriptomic data were obtained from the GEO database (http://www. ncbi. nlm. nih. gov/geo/) for three gene expression datasets, GSE79768, GSE120895, and GSE17800 (14). The GSE79768 dataset consisted of 12 control samples and 14 AF samples, the GSE120895 dataset consists of 8 control samples and 47 DCM samples, and the GSE17800 dataset consisted of 8 control samples and 40 DCM samples. We used the R package “optparse” to normalize gene expression data.

2.2 Generation of genetic instruments

All MR approaches are based on three core assumptions to reduce the impact of bias on causal estimates (15): (i) the IVs are strongly associated with AF, (ii) the IVs affect DCM only by affecting AF, and (iii) the IVs are independent of any other confounding parameters. The violation of these three core assumptions would lead to unreliable conclusions (Figure S1). The IVs associated with AF and DCM were chosen as, first, SNPs strongly associated with AF were collected from published GWAS, with a threshold chosen to be p < 5×10^-8. Second, we adjusted for mutual interlocking disequilibrium (LD) between the SNPs (window size = 10,000 kb, r²< 0.001).

2.3 Two-sample MR analysis

Inverse variance-weighted (IVW), MR-Egger model, and weighted median analysis were used to explore the causal effect of AF on DCM risk (16, 17). The three methods were combined to provide stable estimates of the causal effects. Leave-one-out analyses and tests for horizontal pleiotropy and heterogeneity (18, 19) were then performed to extensively assess the MR results (Table S1).

2.4 Differentially expressed gene screening

DEGs between AF and the control group and between DCM and the control group were obtained from the GSE79768 and GSE120895 datasets, respectively, with a threshold adjusted p-value <0.05, and |log ₂Fold change (FC)| >1.2. The R package “Limma” was used for analyses. The Sangerbox platform (http://vip. sangerbox. com/) was used to visualize DEGs.

2.5 Weighted gene correlation network analysis

The R package “WGCNA” was used to construct gene co-expression networks to study associations between genes and phenotypes (20). First, we removed half of the genes with minimal median absolute deviation (MAD). Second, a weighted neighbor-joining matrix was constructed based on the average chaining method and weighted correlation coefficients. Adjacency were then calculated using a “soft” threshold power (β) and converted to a topological overlap matrix (TOM). Average linkage hierarchical clustering was performed based on the dissimilarity measure of the TOM, with a minimum gene group size of 30, to group genes with similar expression profiles into modules. Finally, the dissimilarity of the module feature genes was calculated, the cut-off of the module dendrogram was selected, and multiple modules were merged. The WGCNA method was used to identify the important modules in AF and build a visualized feature gene network.

2.6 Function enrichment analysis

The R package “clusterProfile” allows exploration of the biological functions of genes (21). After taking intersections of AF key module genes with DEGs of AF and DEGs of DCM, respectively, these intersected genes were subjected to Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses, with a p-value of < 0.05 (22, 23). The Sangerbox platform (http://vip. sangerbox. com/) was used to visualize the results.

2.7 Protein–protein intersection network construction

The STRING database (https: //cn.string-db.org/) was used to construct PPI networks of intersecting genes in DCM and AF key module genes to study protein-protein interactions (24). Cytoscape software was used to identify the significantly interacting genes (25).

2.8 Machine learning

LASSO, RF, and SVM analyses were used to identify the potential candidate genes for AF combined with DCM. LASSO uses regularization techniques to improve the prediction accuracy (26). As a prediction algorithm, RF does not impose constraints on variable conditions, allowing it to provide predictions without significant variations (27). SVM, on the other hand, are supervised learning algorithms that can be used for classification and regression tasks (28). The R packages “glmnet” “randomforest” and “e1071” can perform LASSO, RF, and SVM analyses, respectively (28, 29). The intersection of these three results can be used as a candidate hub gene for diagnosis.

2.9 Nomogram construction and receiver operating characteristic evaluation

A nomogram constructed using the R package “rms” was used to determine the importance of pivotal genes in the diagnosis of AF combined with DCM. The nomogram consists of “Points,” which represent the scores of candidate genes, and “Total points” Total total points, which show the sum of the scores of all genes. In addition, the diagnostic value of the candidate genes and nomogram could be further evaluated by ROC analysis, which generated the area under the curve (AUC) and 95% confidence interval (CI), with an AUC value > 0.7 being considered to have high diagnostic efficacy.

2.10 Immune infiltration analysis

The CIBERSORT software was used to estimate the infiltration of immune cells in the DCM and control groups based on gene expression profiles (30). Bar plots were used to show the proportions of various immune cells, a heatmap to depict the connections between various immune cells, and a vioplot plot plot to compare the proportions of immune cells between the DCM group and the control group (31). The Sangerbox platform (http://vip. sangerbox. com/) was used to visualize the results.

2.11 MiRNAs associated with hub genes

NetworkAnalyst (https://www.networkanalyst.ca/) was used to construct miRNA-gene interaction networks for hub genes, and the networks were visualized using Cytoscape software.

2.12 Statistical analysis

The data obtained in this study were analyzed using statistical methods. MR analysis was performed using the R software packages “Two-Sample MR” and “MRPRESSO,” and parameter estimates (β) and 95%CI were calculated. ROC curves were plotted using SPSS software (version 26.0), and the AUC and 95%CI were calculated. Statistical p-value <0.05.

3.1 Genetic susceptibility to AF and DCM risk

A total of 818,314 European populations (4962 patients and 813,352 controls) were analyzed in this MR study, and IVs significantly associated with AF and DCM were extracted from GWASs (p<5×10^-8) and LD was removed (r²<0.001, 10,000-kb). As shown in Table 2, in the European population, we observed a causal inference between genetic susceptibility to AF and DCM (IVW: β=20.44, 95%CI: 15.00-25.88, p=0.0002; weighted median: β=18.28, 95%CI: 11.37-25.19, p=0.0082) (Table 2 and Figure 2).

3.2 Sensitivity analyses for MR estimates

Thirteen SNPs were analyzed using leave-one-out analysis and heterogeneity tests. Leave-one-out analyses showed that no single SNP could mediate a causal relationship between AF and DCM (Figure 2). The Cochrane Q test for heterogeneity had p-values greater than 0.05 (Q value for the IVW test=10.56, p=0.5667; Q value for the MR-Egger test=10.56, p=0.4806), suggesting that there was no heterogeneity among the SNPs. The MR-PRESSO model test for horizontal pleiotropy confirmed that pleiotropy was unlikely to affect causality (p>0.05) (Table S1).

3.3 Identification of differentially expressed genes

A total of 6463 DEGs were identified in the DCM dataset with p-value <0.05, |log₂FC| >1.2, and the volcano plot and heatmap in Figure 3 demonstrate the differential expression of these DEGs. Similarly, 1850 DEGs were identified in the AF dataset with p-value <0.05, |log₂FC| >1.2, and the differential expression of these DEGs in AF is depicted in Figures 4A and B.

3.4 Weighted gene co-expression network analysis and critical module identification

A scale-free co-expression network constructed using WGCNA was used to identify the most relevant modules in AF. A “soft” threshold β of 5 was chosen based on scale independence and average connectivity (Figures 4C, D). The clustered dendrograms of AF and controls had 38 different colored gene co-expression modules with a module merging threshold of 0.25 and a minimum size of 30, as shown in Figures 4E-G. Clinical correlation analysis showed that the brown module had the strongest association with AF (r=0.72, p<0.001), as shown in Figure 4H. Therefore, we selected a brown module consisting of 572 genes for analysis. Correlation analysis between module membership and gene significance revealed a significant positive correlation (correlation coefficient=0.70, p<0.001), as shown in Figure 4I. These results suggest that genes in the brown module are most closely associated with AF.

3.5 Functional enrichment analysis of AF

The intersection of AF genes from Limma and module genes was enriched, and a total of 172 intersecting genes were obtained, as shown in Figure 5A. KEGG analysis showed that the intersected genes were involved in “ubiquitin-regulated protein degradation” and “PPAR signaling pathway,” as shown in Figure 5B. The results of GO analysis showed that the intersected genes were involved in biological processes (BPs) such as “chromosome organization”, “organelle localization, and “monovalent inorganic cation homeostasis”, as shown in Figure 5C. In the cellular component (CC), the intersected genes were involved in “Copii encapsulated ER to Golgi transport vesicles”, “autophagosome, and “LSM1-7-PAT1 complex”, as shown in Figure 5D. For molecular function (MF), the results showed that “nucleotide binding” and “guanosine binding” were more common, as shown in Figure 5E.

3.6 Enrichment analysis of AF with DCM and screening node genes via the protein– protein interaction network

The intersection of the DEGs for DCM and module genes for AF included 209 genes, as shown in Figure 6A. KEGG analysis showed that 209 genes were mainly enriched in “various infections” and “phagocytic vesicles, ” as shown in Figure 6B. GO analysis revealed genes involved in “vascular phylogeny,” “organelle localisation” (BP), “anchoring junctions,” “cellular substrate connectivity” (CC) and “binding enzymes” (MF), as shown in Figures 6C-E. The interactions between 209 genes were determined by constructing a PPI network, as shown in Figure 6F. Genes were sorted according to the number of nodes (Figure 6G).

3.7 Identification of candidate hub genes via machine learning

LASSO regression, RF, and SVM machine-learning algorithms were used to identify potential candidate genes associated with the diagnosis of AF combined with DCM. LASSO regression identified 10 genes that were closely associated with the diagnosis and plotted ROC curves, which demonstrated the high accuracy of LASSO regression (AUC 1.00, 95%CI 0.99-1.00) (Figures 7A, B, C). The top 15 significant genes from RF and SVM are shown in Figures 7D and E, respectively. Two genes (VSNL1 and ETNPPL) identified by the three methods served as the key diagnostic genes for the final validation, as shown in Figure 7F.

3.8 Diagnosis value evaluation

The model was validated using GSE17800 to construct a nomogram containing two key diagnostic genes (Figure 8B). The AUC and 95%CI of these genes were calculated by constructing ROC curves to assess the diagnostic efficacy, as shown in Figure 8A. The results were as follows: VSNL1 (AUC 0.87, 95%CI 0.74-1.00), ETNPPL (AUC 0.81, 95%CI 0.64-0.99), and nomogram (AUC 0.89, 95%CI 0.74-1.00).

3.9 Immune infiltration analysis

Figure 9 shows the correlation between the DCM gene set and immune cells in the study, which revealed that only activated CD 4+ memory T cells differed between the disease and control groups.

3.10 Target miRNAs prediction and integrated miRNAs-targets network construction

As shown in Figure 10, VSNL1 interacted with 37 target miRNAs, whereas no target miRNAs were found to interact with ETNPPL. VSNL1 interacted with ARMC1 in the PPI network, whereas no proteins were found to be associated with ETNPPL. The intracellular signalling and regulatory pathways of VSNL1 are yet to be experimentally verified.

DCM is more prone to abnormalities in electrophysiological activity during its pathogenesis due to the enlargement of the cardiac chamber structure and alterations in cardiomyocytes at the microscopic level, which provide the fundamental conditions for the development of AF (32). Simultaneously, during pathological cardiac remodeling in DCM, the original contractile function of the heart gradually decreases, which further increases the risk of AF (33). The combination of AF in patients with DCM not only aggravates the pre-existing symptoms of heart failure but also develops new symptoms such as dizziness and palpitations (5, 34). Our MR analyses aimed to assess the causal relationship between AF and DCM, and we found strong genetic susceptibility between AF and DCM. The genetic basis of AF leading to DCM is complex and not fully defined (35). Previous studies have shown that both AF and DCM are influenced by genetic factors. The presence of certain genetic variants such as mutations or SNPs may increase an individual's susceptibility to both AF and DCM (36). These genetic variants may involve multiple aspects of ion channel, myocardial structure, and signal transduction genes (37). Studies on familial AF and familial DCM have shown that there is a significant clustering of these diseases in families, suggesting an important role for genetic factors (38). Further studies may reveal the presence of common genetic variants in certain families that influence the development of both AF and DCM (39). Simultaneously, the development of AF and DCM may involve the interaction of multiple genes (40). These genes may affect the structure and function of the heart through different pathways, thus contributing to disease development. For example, certain genes may affect the electrical activity of cardiomyocytes, whereas others may affect their metabolic and contractile functions (41, 42).

When AF occurs, the abnormal electrical activity of atrial myocytes leads to uncoordinated atrial contraction, which in turn triggers atrial dilatation (43). This dilatation may be reversible in the short term, but during prolonged AF, atrial dilatation may become permanent (44). Not only does Atrial dilatation increases intra-atrial pressure and alters the anatomy of the atria, providing favorable conditions for the maintenance and progression of AF (45). Prolonged AF may also lead to myocardial fibrosis, another important aspect of atrial structural remodeling. Myocardial fibrosis causes the electrical conduction between myocardial cells to become heterogeneous, further contributing to the development and maintenance of AF (46). At the same time, AF may also cause alterations in the ion channels of atrial myocytes, such as a decrease in L-type calcium current and a decrease in transient outward current (47). These ion-channel alterations further affect the electrical activity of atrial myocytes, making them more susceptible to arrhythmias (48). Activation of the renin-angiotensin-aldosterone system (RAAS) and sympathetic excitation also contribute to the development of myocardial remodeling and cardiac insufficiency (49). Thus, AF leads to a vicious cycle of atrial structural and electrophysiological remodeling through a combination of mechanisms (50). This vicious cycle makes AF increasingly difficult to control and exacerbates myocardial damage and remodeling, ultimately leading to the development of DCM. In this study, we examined the expression of VSNL1 and ETNPPL in patients with AF and assessed the risk of combined DCM in patients with AF.

Visinin-Like Protein 1(VSNL1), an important gene encoding a neural calcium-sensing protein, is mainly expressed in the neural and cardiac tissues (51). VSNL1, an important member of the neural calcium-sensing protein (NCS) family, is mainly involved in intracellular calcium ion sensing and signaling processes (52). As an organ highly dependent on calcium ion regulation, the contractile and diastolic processes of cardiomyocytes in the heart are precisely controlled by changes in calcium ion concentration (53). Disruption of calcium ion homeostasis may lead to abnormal cardiac electrophysiological activities, thereby triggering arrhythmias, such as AF. It has been shown that VSNL1 may lead to myocardial injury and arrhythmias by activating the CNP/NPR-B signaling pathway, which in turn affects changes in calcium ion flow pathways (54). In addition, VSNL1 may be involved in the onset and progression of cardiac diseases by regulating calcium ion homeostasis in cardiac cells and affecting processes such as contraction and diastole in cardiomyocytes, which may be related to DCM development (55).

Ethanolamine-phosphate phospho-lyase (ETNPPL), an important protein kinase, is highly expressed in the liver (56). The enzyme encoded by this gene mainly catalyzes the production of ammonia, inorganic phosphate, and acetaldehyde from phosphoethanolamine (PEA), an enzymatic reaction that plays an important role in phospholipid metabolism and in maintaining intracellular metabolic homeostasis (56-58). Studies have shown that ETNPPL plays a role in the regulation of fatty acid-induced insulin resistance in hepatocytes (57). The central role of ETNPPL in phospholipid metabolism implies that its aberrant expression or dysfunction may be implicated in the onset and progression of several metabolic diseases (58). The heart is a highly metabolically demanding organ, whose normal function depends on the energy supply and efficient processing of metabolites (59). As a part of phospholipid metabolism, ETNPPL may affect the composition and stability of cell membranes, thus indirectly affecting the health and function of cardiac cells (57, 60). Phospholipid metabolism processes involving ETNPPL may affect the stability and function of cell membranes, and cell membrane abnormalities are one of the important pathological bases of cardiac arrhythmias, including AF (60).

The ROC curves in this study indicate that the diagnostic value of these two genes is relatively high, but the specific mechanism of their role of the two genes in the evolution of AF to DCM is not clear. In the future, in-depth studies on the specific expression patterns and functional roles of these two genes in the heart can be conducted to reveal the complex genetic network of cardiac disorders. miRNAs are small non-coding RNA molecules that induce mRNA degradation or inhibit mRNA translation (61). In this study, VSNL1 was found to be associated with a variety of miRNA molecules, which could pave the way for subsequent studies.

This study has several limitations. First, immune infiltration and inflammatory responses play a significant role in myocardial fibrosis and chamber remodeling in the pathogenesis of DCM (62). In patients with DCM, infiltration of lymphocytes and monocytes as well as increased expression of cell adhesion molecules are frequently observed (63). In contrast, the gene set we studied did not show significant findings in terms of immune infiltration, a limitation that requires further study of subsequent related gene sets. Second, the MR approach was unable to completely eliminate potential horizontal pleiotropy in the study of genetic susceptibility to AF and DCM. Third, this study only focused on a European population, and the association between AF and DCM in different populations needs to be explored.

Our MR analyses suggested a direct causal relationship between AF and DCM. Two hub genes (VSNL1 and ETNPPL) showed potential as candidate genes for the diagnosis of comorbid DCM in AF patients. Clinical detection of the expression of these two genes in patients with AF could be effective in aiding clinical decision making. In conclusion, our findings may provide new insights into the pathogenesis and diagnosis of comorbid DCM in patients with AF, although further clinical studies of both genes are necessary.

Data availability statement

The data analyzed in the present study were obtained from MR-Base (http://www.mrbase.org/) and GEO (http://www.ncbi.nlm.nih.gov/geo/) (under accession numbers GSE79768, GSE120895, and GSE17800).

Author contributions

HW and BS conceived of the project. HW, BS, JBL and HYW acquired the data and performed analyses. HW and BS drafted the manuscript with critical feedback from HYW. All authors have contributed to the manuscript and approved the submitted version.

Funding

This work was supported by the Capital's Funds for Health Improvement and Research (2020-2-6042), Key Medical disciplines/schools of Shijingshan district (Vascular Medicine); and Key clinical projects in Peking University Shougang Hospital (2019-Yuan-LC-01; SGYYZ202105).

Acknowledgments

The authors would like to appreciate the efforts of the GWAS consortium and GEO database.

Conflict of interest

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

Ethics approval and consent to participate

The data used in this study were acquired from an open source and do not require approval by any ethical committee.

Clinical trial number

Not applicable.

Wang B, Lunetta KL, Dupuis J, Lubitz SA, Trinquart L, Yao L, et al. Integrative Omics Approach to Identifying Genes Associated With Atrial Fibrillation. Circ Res (2020) 126(3):350-360. doi: 10.1161/CIRCRESAHA.119.315179
Ahlberg G, Refsgaard L, Lundegaard PR, Andreasen L, Ranthe MF, Linscheid N, et al. Rare truncating variants in the sarcomeric protein titin associate with familial and early-onset atrial fibrillation. Nat Commun (2018) 9(1):4316. doi: 10.1038/s41467-018-06618-y
Hoffmann S, Clauss S, Berger IM, Weiß B, Montalbano A, Röth R, et al. Coding and non-coding variants in the SHOX2 gene in patients with early-onset atrial fibrillation. Basic Res Cardiol (2016) 111(3):36. doi: 10.1007/s00395-016-0557-2
El-Am EA, Dispenzieri A, Melduni RM, Ammash NM, White RD, Hodge DO, et al. Direct Current Cardioversion of Atrial Arrhythmias in Adults With Cardiac Amyloidosis. J Am Coll Cardiol (2019) 73(5):589-597. doi: 10.1016/j.jacc.2018.10.079
van Ouwerkerk AF, Bosada FM, van Duijvenboden K, Hill MC, Montefiori LE, Scholman KT, et al. Identification of atrial fibrillation associated genes and functional non-coding variants. Nat Commun (2019) 10(1):4755. doi: 10.1038/s41467-019-12721-5
Reiffel JA, Verma A, Kowey PR, Halperin JL, Gersh BJ, Wachter R, et al. Incidence of Previously Undiagnosed Atrial Fibrillation Using Insertable Cardiac Monitors in a High-Risk Population: The REVEAL AF Study. JAMA Cardiol (2017) 2(10):1120-1127. doi: 10.1001/jamacardio.2017.3180
Perino AC, Fan J, Askari M, Heidenreich PA, Keung E, Raitt MH, et al. Practice Variation in Anticoagulation Prescription and Outcomes After Device-Detected Atrial Fibrillation. Circulation (2019) 139(22):2502-2512. doi: 10.1161/CIRCULATIONAHA.118.038988
Gigli M, Merlo M, Graw SL, Barbati G, Rowland TJ, Slavov DB, et al. Genetic Risk of Arrhythmic Phenotypes in Patients With Dilated Cardiomyopathy. J Am Coll Cardiol (2019) 74(11):1480-1490. doi: 10.1016/j.jacc.2019.06.072
Kunamalla A, Ng J, Parini V, Yoo S, McGee KA, Tomson TT, et al. Constitutive Expression of a Dominant-Negative TGF-β Type II Receptor in the Posterior Left Atrium Leads to Beneficial Remodeling of Atrial Fibrillation Substrate. Circ Res (2016) 119(1):69-82. doi: 10.1161/CIRCRESAHA.115.307878
Opdahl A, Ambale Venkatesh B, Fernandes VRS, Wu CO, Nasir K, Choi EY, et al. Resting heart rate as predictor for left ventricular dysfunction and heart failure: MESA (Multi-Ethnic Study of Atherosclerosis). J Am Coll Cardiol (2014) 63(12):1182-1189. doi: 10.1016/j.jacc.2013.11.027
Kloosterman M, Santema BT, Roselli C, Nelson CP, Koekemoer A, Romaine SPR, et al. Genetic risk and atrial fibrillation in patients with heart failure. Eur J Heart Fail (2020) 22(3):519-527. doi: 10.1002/ejhf.1735
Davey Smith G, Ebrahim S. What can mendelian randomisation tell us about modifiable behavioural and environmental exposures? BMJ (2005) 330(7499):1076-9. doi: 10.1136/bmj.330.7499.1076
Sakaue S, Kanai M, Tanigawa Y, Karjalainen J, Kurki M, Koshiba S, et al. A cross-population atlas of genetic associations for 220 human phenotypes. Nat Genet (2021) 53(10):1415-1424. doi: 10.1038/s41588-021-00931-x
Barrett T, Wilhite SE, Ledoux P, Evangelista C, Kim IF, Tomashevsky M, et al. NCBI GEO: archive for functional genomics data sets--update. Nucleic Acids Res (2013) 41(Database issue):D991-5. doi: 10.1093/nar/gks1193
VanderWeele TJ, Tchetgen Tchetgen EJ, Cornelis M, Kraft P. Methodological challenges in mendelian randomization. Epidemiology (2014) 25(3):427-35. doi: 10.1097/EDE.0000000000000081
Burgess S, Thompson SG. Interpreting findings from Mendelian randomization using the MR-Egger method. Eur J Epidemiol (2017) 32(5):377-389. doi: 10.1007/s10654-017-0255-x
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. doi: 10.1002/gepi.21965
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):693-698. doi: 10.1038/s41588-018-0099-7
Bowden J, Del Greco M F, Minelli C, Davey Smith G, Sheehan N, Thompson J. A framework for the investigation of pleiotropy in two-sample summary data Mendelian randomization. Stat Med (2017) 36(11):1783-1802. doi: 10.1002/sim.7221
Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics (2008) 9:559. doi: 10.1186/1471-2105-9-559
Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS (2012) 16(5):284-7. doi: 10.1089/omi.2011.0118
The Gene Ontology Consortium. The Gene Ontology Resource: 20 years and still GOing strong. Nucleic Acids Res (2019) 47(D1):D330-D338. doi: 10.1093/nar/gky1055
Kanehisa M, Goto S. KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res (2000) 28(1):27-30. doi: 10.1093/nar/28.1.27
Szklarczyk D, Gable AL, Nastou KC, Lyon D, Kirsch R, Pyysalo S, et al. The STRING database in 2021: customizable protein-protein networks, and functional characterization of user-uploaded gene/measurement sets. Nucleic Acids Res (2021) 49(D1):D605-D612. doi: 10.1093/nar/gkaa1074
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 Res (2003) 13(11):2498-504. doi: 10.1101/gr.1239303
Yang C, Delcher C, Shenkman E, Ranka S. Machine learning approaches for predicting high cost high need patient expenditures in health care. Biomed Eng Online (2018) 17(Suppl 1):131. doi: 10.1186/s12938-018-0568-3
Ellis K, Kerr J, Godbole S, Lanckriet G, Wing D, Marshall S. A random forest classifier for the prediction of energy expenditure and type of physical activity from wrist and hip accelerometers. Physiol Meas (2014) 35(11):2191-203. doi: 10.1088/0967-3334/35/11/2191
Li YR, Meng K, Yang G, Liu BH, Li CQ, Zhang JY, Zhang XM. Diagnostic genes and immune infiltration analysis of colorectal cancer determined by LASSO and SVM machine learning methods: a bioinformatics analysis. J Gastrointest Oncol (2022) 13(3):1188-1203. doi: 10.21037/jgo-22-536
Alderden J, Pepper GA, Wilson A, Whitney JD, Richardson S, Butcher R, Jo Y, Cummins MR. Predicting Pressure Injury in Critical Care Patients: A Machine-Learning Model. Am J Crit Care (2018) 27(6):461-468. doi: 10.4037/ajcc2018525
Newman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, et al. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods (2015) 12(5):453-7. doi: 10.1038/nmeth.3337
Shen W, Song Z, Zhong X, Huang M, Shen D, Gao P, et al. Sangerbox: A comprehensive, interaction-friendly clinical bioinformatics analysis platform. Imeta (2022) 1(3):e36. doi: 10.1002/imt2.36
Du X, Liu M, Su J, Zhang P, Tang F, Ye P, et al. Uncoupling therapeutic from immunotherapy-related adverse effects for safer and effective anti-CTLA-4 antibodies in CTLA4 humanized mice. Cell Res (2018) 28(4):433-447. doi: 10.1038/s41422-018-0012-z
Radke MH, Polack C, Methawasin M, Fink C, Granzier HL, Gotthardt M. Deleting Full Length Titin Versus the Titin M-Band Region Leads to Differential Mechanosignaling and Cardiac Phenotypes. Circulation (2019) 139(15):1813-1827. doi: 10.1161/CIRCULATIONAHA.118.037588
Halliday BP, Gulati A, Ali A, Guha K, Newsome S, Arzanauskaite M, et al. Association Between Midwall Late Gadolinium Enhancement and Sudden Cardiac Death in Patients With Dilated Cardiomyopathy and Mild and Moderate Left Ventricular Systolic Dysfunction. Circulation (2017) 135(22):2106-2115. doi: 10.1161/CIRCULATIONAHA.116.026910
Olenchock BA, Moslehi J, Baik AH, Davidson SM, Williams J, Gibson WJ, et al. EGLN1 Inhibition and Rerouting of α-Ketoglutarate Suffice for Remote Ischemic Protection. Cell (2016) 164(5):884-95. doi: 10.1016/j.cell.2016.02.006
Savoji H, Mohammadi MH, Rafatian N, Toroghi MK, Wang EY, Zhao Y, et al. Cardiovascular disease models: A game changing paradigm in drug discovery and screening. Biomaterials (2019) 198:3-26. doi: 10.1016/j.biomaterials.2018.09.036
Gao R, Liang X, Cheedipudi S, Cordero J, Jiang X, Zhang Q, et al. Pioneering function of Isl1 in the epigenetic control of cardiomyocyte cell fate. Cell Res (2019) 29(6):486-501. doi: 10.1038/s41422-019-0168-1
Giudicessi JR, Ackerman MJ. Potassium-channel mutations and cardiac arrhythmias--diagnosis and therapy. Nat Rev Cardiol (2012) 9(6):319-32. doi: 10.1038/nrcardio.2012.3
Halliday BP, Wassall R, Lota AS, Khalique Z, Gregson J, Newsome S, et al. Withdrawal of pharmacological treatment for heart failure in patients with recovered dilated cardiomyopathy (TRED-HF): an open-label, pilot, randomised trial. Lancet (2019) 393(10166):61-73. doi: 10.1016/S0140-6736(18)32484-X
Yi N, Shriner D, Banerjee S, Mehta T, Pomp D, Yandell BS. An efficient Bayesian model selection approach for interacting quantitative trait loci models with many effects. Genetics (2007) 176(3):1865-77. doi: 10.1534/genetics.107.071365
Thompson SA, Blazeski A, Copeland CR, Cohen DM, Chen CS, Reich DM, Tung L. Acute slowing of cardiac conduction in response to myofibroblast coupling to cardiomyocytes through N-cadherin. J Mol Cell Cardiol (2014) 68:29-37. doi: 10.1016/j.yjmcc.2013.12.025
Naumenko N, Huusko J, Tuomainen T, Koivumäki JT, Merentie M, Gurzeler E, et al. Vascular Endothelial Growth Factor-B Induces a Distinct Electrophysiological Phenotype in Mouse Heart. Front Physiol (2017) 8:373. doi: 10.3389/fphys.2017.00373
Lowenstern A, Al-Khatib SM, Sharan L, Chatterjee R, Allen LaPointe NM, Shah B, et al. Interventions for Preventing Thromboembolic Events in Patients With Atrial Fibrillation: A Systematic Review. Ann Intern Med (2018) 169(11):774-787. doi: 10.7326/M18-1523
Cleland JGF, Lyon AR, McDonagh T, McMurray JJV. The year in cardiology: heart failure. Eur Heart J (2020) 41(12):1232-1248. doi: 10.1093/eurheartj/ehz949
Darbar D, Hardy A, Haines JL, Roden DM. Prolonged signal-averaged P-wave duration as an intermediate phenotype for familial atrial fibrillation. J Am Coll Cardiol (2008) 51(11):1083-9. doi: 10.1016/j.jacc.2007.11.058
Patel VB, Zhabyeyev P, Chen X, Wang F, Paul M, Fan D, et al. PI3Kα-regulated gelsolin activity is a critical determinant of cardiac cytoskeletal remodeling and heart disease. Nat Commun (2018) 9(1):5390. doi: 10.1038/s41467-018-07812-8
Veldkamp MW, Geuzebroek GSC, Baartscheer A, Verkerk AO, Schumacher CA, Suarez GG, et al. Neurokinin-3 receptor activation selectively prolongs atrial refractoriness by inhibition of a background K+ channel. Nat Commun (2018) 9(1):4357. doi: 10.1038/s41467-018-06530-5
Inciardi RM, Giugliano RP, Claggett B, Gupta DK, Chandra A, Ruff CT, et al. Left atrial structure and function and the risk of death or heart failure in atrial fibrillation. Eur J Heart Fail (2019) 21(12):1571-1579. doi: 10.1002/ejhf.1606
Pellegrino PR, Schiller AM, Haack KK, Zucker IH. Central Angiotensin-II Increases Blood Pressure and Sympathetic Outflow via Rho Kinase Activation in Conscious Rabbits. Hypertension (2016) 68(5):1271-1280. doi: 10.1161/HYPERTENSIONAHA.116.07792
Pabel S, Ahmad S, Tirilomis P, Stehle T, Mustroph J, Knierim M, et al. Inhibition of NaV1.8 prevents atrial arrhythmogenesis in human and mice. Basic Res Cardiol (2020) 115(2):20. doi: 10.1007/s00395-020-0780-8
Zhang H, Du ZQ, Dong JQ, Wang HX, Shi HY, Wang N, Wang SZ, Li H. Detection of genome-wide copy number variations in two chicken lines divergently selected for abdominal fat content. BMC Genomics (2014) 15:517. doi: 10.1186/1471-2164-15-517
Braunewell KH, Dwary AD, Richter F, Trappe K, Zhao C, Giegling I, Schönrath K, Rujescu D. Association of VSNL1 with schizophrenia, frontal cortical function, and biological significance for its gene product as a modulator of cAMP levels and neuronal morphology. Transl Psychiatry (2011) 1(7):e22. doi: 10.1038/tp.2011.20
Quick AP, Wang Q, Philippen LE, Barreto-Torres G, Chiang DY, Beavers D, et al. SPEG (Striated Muscle Preferentially Expressed Protein Kinase) Is Essential for Cardiac Function by Regulating Junctional Membrane Complex Activity. Circ Res (2017) 120(1):110-119. doi: 10.1161/CIRCRESAHA.116.309977
Dickey DM, Flora DR, Bryan PM, Xu X, Chen Y, Potter LR. Differential regulation of membrane guanylyl cyclases in congestive heart failure: natriuretic peptide receptor (NPR)-B, Not NPR-A, is the predominant natriuretic peptide receptor in the failing heart. Endocrinology (2007) 148(7):3518-22. doi: 10.1210/en.2007-0081
Liang D, Xue J, Geng L, Zhou L, Lv B, Zeng Q, et al. Cellular and molecular landscape of mammalian sinoatrial node revealed by single-cell RNA sequencing. Nat Commun (2021) 12(1):287. doi: 10.1038/s41467-020-20448-x
Raffler J, Friedrich N, Arnold M, Kacprowski T, Rueedi R, Altmaier E, et al. Genome-Wide Association Study with Targeted and Non-targeted NMR Metabolomics Identifies 15 Novel Loci of Urinary Human Metabolic Individuality. PLoS Genet (2015) 11(9):e1005487. doi: 10.1371/journal.pgen.1005487
Chen X, Liu P, Zhang W, Li X, Wang C, Han F, et al. ETNPPL modulates hyperinsulinemia-induced insulin resistance through the SIK1/ROS-mediated inactivation of the PI3K/AKT signaling pathway in hepatocytes. J Cell Physiol (2023) 238(5):1046-1062. doi: 10.1002/jcp.30993
Wang C, Li X, Zhang W, Liu W, Lv Z, Gui R, et al. ETNPPL impairs autophagy through regulation of the ARG2-ROS signaling axis, contributing to palmitic acid-induced hepatic insulin resistance. Free Radic Biol Med (2023) 199:126-140. doi: 10.1016/j.freeradbiomed.2023.02.017
Karlstaedt A, Khanna R, Thangam M, Taegtmeyer H. Glucose 6-Phosphate Accumulates via Phosphoglucose Isomerase Inhibition in Heart Muscle. Circ Res (2020) 126(1):60-74. doi: 10.1161/CIRCRESAHA.119.315180
Xu M, Guo YY, Li D, Cen XF, Qiu HL, Ma YL, Huang SH, Tang QZ. Screening of Lipid Metabolism-Related Gene Diagnostic Signature for Patients With Dilated Cardiomyopathy. Front Cardiovasc Med (2022) 9:853468. doi: 10.3389/fcvm.2022.853468
Zinkstok JR, Boot E, Bassett AS, Hiroi N, Butcher NJ, Vingerhoets C, Vorstman JAS, van Amelsvoort TAMJ. Neurobiological perspective of 22q11.2 deletion syndrome. Lancet Psychiatry (2019) 6(11):951-960. doi: 10.1016/S2215-0366(19)30076-8
Frantz S, Falcao-Pires I, Balligand JL, Bauersachs J, Brutsaert D, Ciccarelli M, et al. The innate immune system in chronic cardiomyopathy: a European Society of Cardiology (ESC) scientific statement from the Working Group on Myocardial Function of the ESC. Eur J Heart Fail (2018) 20(3):445-459. doi: 10.1002/ejhf.1138
Kaya Z, Göser S, Buss SJ, Leuschner F, Ottl R, Li J, et al. Identification of cardiac troponin I sequence motifs leading to heart failure by induction of myocardial inflammation and fibrosis. Circulation (2008) 118(20):2063-72. doi: 10.1161/CIRCULATIONAHA.108.788711

TABLE 1 Details of GWASs analyzed in the present MR analyses.

Phenotype	First author	Sample size	Number of patients	Number of controls	Number of variants	Ethnicity	Pubmed ID	Trait ID in GWAS	Year
Exposure
Atrial fibrillation	Ben Elsworth	462933	3518	459415	9851867	European	NA	ukb-b-11550	2018
Outcome
Dilated cardiomyopathy	Sakaue S	355381	1444	353937	19080278	European	34594039	ebi-a-GCST90018834	2021
MR, Mendelian Randomization; GWAS, genome-wide association studies; NA, not available.

TABLE 2 Mendelian randomization estimates of the casual relationships between AF and DCM risks.

Phenotype	nSNPs	IVW method		Weighted median method		MR-Egger
		β(95%CI)	p value	β(95%CI)	p value	β(95%CI)	p value
DCM	13	20.44(15.00-25.88)	0.0002	18.28(11.37-25.19)	0.0082	20.55(9.59-31.51)	0.0874
AF, Atrial fibrillation; DCM, Dilated cardiomyopathy; IVW, inverse-variance weighted; MR, mendelian randomization; OR, odds ratio; CI, confidence interval.

No competing interests reported.

SupplementarymaterialAFDCM.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Exploring the causality and pathogenesis of atrial fibrillation with dilated cardiomyopathy: An integrated multi-omics approach

Status:

Version 1

Abstract

Figures

1 Introduction

2 Materials and Methods

3 Results

4 Discussion

5 Conclusion

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1