Ceramide Synthase 2 Promotes Cardiac Very-Long-Chain Dihydroceramide Accumulation and Is Linked to Arrhythmias and Heart Failure in Humans

Abstract

Acute myocardial hypoxia/ischemia is associated with abnormal accumulation of myocardial lipids, including dihydroceramides. Here, we characterized how dihydroceramides are remodeled in response to hypoxia and assessed how dihydroceramide remodeling correlates to human cardiac pathophysiology. Hypoxia resulted in a marked accumulation of very-long-chain (VLC)-dihydroceramides in cultured HL-1 cardiomyocytes. In humans, we identified a correlation between the abundance of VLC-dihydroceramides in myocardial biopsies and arrhythmias and heart failure and showed that cardiac expression of CERS2, coding for an enzyme that promotes synthesis of VLC-dihydroceramides, was associated with signaling pathways linked to cardiac arrhythmia and cardiomyopathy. In cultured HL-1 cardiomyocytes, we showed that CerS2 knockdown reduced accumulation of VLC dihydroceramides and altered the expression of mediators regulating Ca²⁺ cycling and electrical conduction. In conclusion, our findings indicate that increased abundance of VLC-dihydroceramides, promoted by increased activity of CerS2 in response to hypoxia, could play a role in cardiac arrhythmias and heart failure.

Keywords: bioactive lipids; cardiac sphingolipids; ceramide synthase; dihydroceramide; hypoxia.

Figure 1

Hypoxia induces marked accumulation of VLC dihydroceramides in HL-1 cardiomyocytes. (A) Concentration (left) and relative abundance (right) of dihydroceramide species in HL-1 cardiomyocytes after incubation in hypoxia (1% oxygen) or normoxia for 8 h, measured by HPLC-MS (n = 3). (B) Change in dihydroceramide composition in hypoxia over time, measured by HPLC-MS, (n = 3). Data are shown as fold over normoxia. (C) Levels of de novo synthesized (d3-labelled) dihydroceramide species in HL-1 cells after incubation with d3-serine in normoxia and hypoxia for the indicated time. Data are presented as mean ± SEM. ** p < 0.01, *** p < 0.001 vs. normoxia, t-test.

Figure 2

Dihydroceramide chain length correlates with cardiac arrhythmias and heart failure in humans. (A) Correlation matrix of the abundance of dihydroceramide species of different chain lengths in human myocardial biopsies and cardiac-related outcome variables (n = 85). (B) Box plots of dihydroceramide species of different chain lengths (from A) stratified by atrial fibrillation status, other arrythmia status and/or heart failure. Analysis is based on n = 85, of which 6 patients had AF, 3 patients had other arrhythmias, and 24 patients had heart failure. (C) Association between dihydroceramide chain length and the relative abundance of each lipid species in individuals with arrythmia or heart failure compared with controls. Each data point indicates one lipid species from (A). Correlation coefficients were calculated using the Spearman method; the variables Atrial fibrillation, Other arrhythmia, Heart failure, Myocardial infarction, Aortic valve replacement, CABG and Beta blocker medication are binary and coded as 0/1, whereas NT-pro-BNP and Ejection fraction are continuous variables.

Figure 3

CERS2 expression in human hearts is positively associated with regulation of genes involved in arrhythmogenic cardiomyopathy. (A) Log2 read counts of CERS2, 4 and 5 in the left ventricle from donors with values in the bottom and top quartiles, indicating low and high expression of each gene (108 samples per group). (B) Significantly regulated functional terms related to cardiomyopathy from the KEGG pathway enrichment analysis of genes that were differentially expressed in heart tissues from humans with high versus low expression of each gene. (C) Heat map of significant differentially expressed genes (DEGs) in the arrhythmogenic right ventricular cardiomyopathy KEGG functional term pathway (49/64) from the analysis in heart tissues from humans with high versus low CERS2 expression. Log2 fold change of those genes are also shown for the analysis in heart tissues from humans with high versus low CERS4 and CERS5 expression. Gray boxes show non-significant DEGs. Source data are available for this figure.

Figure 4

Hypoxia-induced accumulation of VLC dihydroceramides is reduced by depletion of CerS2 in HL-1 cardiomyocytes. Total concentration of dihydroceramides in HL-1 cardiomyocytes treated with (A) CerS2 siRNA, (B) CerS4 siRNA or (C) CerS5 siRNA after incubation in hypoxia or normoxia for 8 h (A,B) or 4 h (C), measured by HPLC-MS (n = 3). (D–F) Concentration (left) and relative abundance (right) of dihydroceramide species in (D) CerS2-depleted, (E) CerS4-depleted and (F) CerS5-depleted HL1 cardiomyocytes after incubation in hypoxia or normoxia for 8 h (D,E) or 4 h (F), measured by HPLC-MS (n = 3). Data are presented as mean ± SEM, * p < 0.05, ** p < 0.01, *** p < 0.001, ns = not significant, one-way ANOVA followed by Tukey’s multiple comparison test.

Figure 5

CerS2 depletion alters the expression profile of functional markers in HL-1 cardiomyocytes. mRNA expression of Cacna1c, Dsg2, Dsp, Slc8a1, Atp2a2, Ryr2 and Pkp2 in CerS2-deficient HL1 cardiomyocytes and in cells treated with scrambled control (siNC), 48 h after transfection with CerS2 siRNA (n = 6). Data are presented as mean ± SEM, * p < 0.05, ** p < 0.01, *** p < 0.001, ns = not sigificant t-test.