The transcriptomic features of heart failure.
To identify the differentially expressed genes during the development of heart failure, we analyzed the expression data of heart failing and normal heart tissues from previously published GEO datasets GSE5460, GSE16499 and GSE68316 [4-6]. Totally, 252 samples were collected, including 36 normal heart tissues and 216 heart failing tissues. First, we analyzed the globe expression profiling of each dataset. Compared with the normal heart tissues, the differentially expressed genes in heart failing tissues (P<0.01) were selected for further studies. This resulted in the identification 2184 differentially expressed genes in GSE5406, 1644 differentially expressed genes in GSE16499 and 3477 differentially expressed genes in GSE68316 dataset (Fig. 1a). Among all the differentially expressed genes, only 4 genes were commonly up regulated and 86 genes were commonly down regulated in GSE5406, GSE16499 and GSE68316 datasets (Fig. 1b). In GSE16499 and GSE68316 datasets, the number of down regulated genes was for more than the up regulated genes (Fig. 1a). In GSE16499 dataset, 1407 genes were suppressed in heart failing tissues. While, only 237 genes were activated in heart failing tissues. Those results suggested that the pathological observations of depletion of cardiomyocytes and loss of mechanical functions in cardiac remodeling were induced by the suppression of heart specific genes.
Metabolism and insulin signaling pathway are inactivated in heart failing patients.
To reveal the functional relevance of the common differentially expressed genes in heart failing tissues, we performed functional signaling pathway enrichment analysis through DAVID [25] and GSEA [26] assay. Pyrimidine, purine metabolism signaling pathway and cysteine, methionine metabolism signaling pathway were highly enriched using the differentially expressed genes through DAVID analysis (Fig. 2a). Heatmap presentations showed that NME1, POLE3, POLD2, ENTPD6, PNP genes from pyrimidine, purine metabolism signaling pathway and LDHA, AHCY, AMD1 genes from cysteine, methionine metabolism signaling pathway were all down regulated in heart failing tissues in GSE5406, GSE16499 and GSE68316 datasets (Fig. 2b). Those results suggested the inactivation of those pathways in the development of heart failure.
Through GSEA analysis, we found that the insulin signaling pathway was negatively correlated with the heart failing expression profiling (Fig. 2c), suggested the inactivation of insulin signaling pathway in the development of heart failure. Fox example, MAP2K1 is a critical downstream gene of insulin signaling pathway [27, 28]. We showed that MAP2K1 was down regulated in heart failing tissues in GSE5406, GSE16499 and GSE68316 datasets (Fig. 2d).
The association among heart failure, inactivation of metabolism pathways and insulin resistance was well established [29, 30]. The cardiac metabolism, growth and survival in the heart were dependent on insulin signaling pathway [31, 32]. Loss of insulin signaling pathway induced cardiac energy deficiency and structural dysfunction accelerating the heart failure progress [33, 34]. And targeting the cardiac metabolism pathways and insulin-PI3K-Akt signaling pathway demonstrated therapeutic promise in preclinical models of heart disease [35-38]. All those observations confirmed our results derived from the GEO datasets. However, the detailed functions of genes involving the metabolism and insulin singling pathways should be further studied in heart failure developmental progress.
Transcription factors MYC and C/EBP are negatively associated with in heart failing expression profiling.
Except signaling pathways, the transcription factors enriched in heart failing tissues were also identified through DAVID analysis. We found that transcription factor MYC was highly associated with the differentially expressed genes in GSE5406, GSE16499 and GSE68316 datasets (Fig. 3a). Interestingly, TP53 and E2F were all highly enriched (Fig. 3a). E2F family genes were critical regulators of cell proliferation and cell cycle progression [39, 40]. Also, E2F family genes mediated the cardiac growth and development [41, 42].
Similar results were obtained using GSEA assay. We found that transcription factor MYC was negatively associated with the heart failing expression profiling in all three GEO datasets (Fig. 3b). Additionally, we showed that transcription factor C/EBP was also negatively correlated with the heart failing expression profiling (Fig. 3c).
C/EBP is a CCAAT/enhancer-binding protein transcription factor which regulates cell growth and differentiation [43]. Previous results suggested that C/EBPβ was a critical regulator of exercise induced cardiac growth and protected against pathological cardiac remodeling [10]. C/EBPβ was also a master regulator of metabolism pathways [13] and insulin resistance [44, 45]. All those reports highlighted the critical roles of C/EBPβ in the development of heart failure. However, the functions of MYC in the development of heart failure are unclear.
Transcription factors MYC and C/EBPβ are down regulated in heart failing tissues.
Next, we detected the expression of MYC and C/EBPβ in heart failing and normal heart tissues. Previous report showed that MYC was increased in pathological hypertrophy [46]. Inhibition of MYC was a potential therapeutic approach in the treatment of hypertrophic cardiomyopathy [47]. On the contrary, we found the down regulation of MYC expression in heart failing tissues in GSE5406 and GSE16499 datasets (Fig. 4a). Similarly, we found that C/EBPβ gene expression was particularly down regulated in heart failing tissues, compared with normal heart tissues in all GSE5406, GSE16499 and GSE68316 datasets (Fig. 4b).
Since MYC and C/EBPβ were both down regulated in heart failure tissues, we tested the correlation between MYC and C/EBPβ expression in GSE5406 and GSE16499 datasets. We found that C/EBPβ expression was positively correlated with MYC expression. Heart tissues with high C/EBPβ expression were also with high MYC expression (Fig. 4c). All those results emphasized the importance of MYC and C/EBPβ in heart failure development.
MYC and C/EBPβ target genes are down regulated in heart failing tissues.
Transcription factors are usually the master regulators of disease and regulate multiple target genes by binding to a specific region of the DNA sequence [8, 9]. In the GSEA assay, we identified 62 MYC target genes and 22 C/EBPβ target genes. Consistent with the decreased expression of MYC and C/EBPβ in heart failing tissues (Fig. 4a and 4b), MYC target genes were down regulated in heart failing tissues, compared with normal heart tissues (Fig. 5a). As demonstrated in the heatmap, C/EBPβ target genes were also particularly down regulated in heart failing tissues in GSE16499 dataset (Fig. 5b).
Interestingly, we found that some genes, for example, EIF4A1, SYNCRIP, ARF6 and C/EBPβ, were both MYC and C/EBPβ downstream target genes (Fig. 5c). We showed that SYNCRIP gene expression was particularly down regulated in heart failing tissues in all GSE5406, GSE16499 and GSE68316 datasets (Fig. 5d). However, whether SYNCRIP was involving in the heart failure development was unclear.
The MYC and C/EBPβ mediated transcriptional networks.
To further explore MYC and its connection to downstream target genes, the MYC mediated regulatory network was constructed using Cytoscape. As expected, as a MYC target gene, C/EBPβ was connected with MYC through the transduction of multiple genes (Fig. 6a). Furthermore, through literature research, we found that 9 MYC target genes were previously reported involving the development of heart failure, including STAT3 [48], PRMT1 [49], PRKCH [50], HSPA4 [51], DDX3X [52], GYG1 [53], GADD45B [54] and PKN1 [55] (Fig. 6a, the red ones).
Similarly, the C/EBPβ mediated regulatory network was constructed (Fig. 6b). Some C/EBPβ target genes, for example, OSMR [56], MAP2K3 [57], CDKN1B [58], PDE4D [59] and DDAH2 [60] also have been studied in heart failure developmental progress (Fig. 6b, the red ones). All those results highlighted the importance of MYC, C/EBPβ and their downstream target genes in heart failure development. The functions of other MYC and C/EBPβ target genes should be further studied to reveal their connections with heart failure.