Background:
Heart failure (HF) is a major cause of mortality after coronary artery diseases in the world. Chronic hypoxia is an essential pathophysiological component in many cardiac diseases. The heart switches energy metabolism efficiently to maintain the balance of energy supply and demand in the adaptation to chronic hypoxia. However, impaired metabolic efficiency leads to metabolism disorder, which ultimately contributes to the severity of heart failure. Therefore, targeting of the cardiac metabolism is an emerging therapeutic strategy to improve cardiac function. Circular RNAs (circRNAs) – a subgroup of the non-coding RNA family – are formed via a non-canonical splicing event called backsplicing. Alterations in expression levels of several circRNAs have been shown in the progression of cardiac diseases. However, the exact function of circRNAs in HF - especially during mitochondrial dysfunction - still remains unknown.

Methods and Results:
To identify circRNAs which play an essential role in cardiac metabolism during heart failure progression, RNA sequencing was performed from heart biopsy samples of patients with endstage HF and healthy donor samples, serving as a control group. 15,255 out of 86,406 circRNAs were found to be differentially expressed in hearts of HF patients. The function of host gene, the level of sequence conservation between human and mouse genome, and resistance against RNase R-mediated digestion were used potential mechanistically important circRNAs. By these stringent selection criteria two circRNAs, namely circAS1 and circAS4, could be identified as the most promising candidates. Both circRNAs are single-exon-encoded and highly conserved between human, mouse, and rat. Their expression in cardiac tissue is highly upregulated in HF patients compared to healthy donor samples. In line, our in vitro study also recaptured the upregulation of both circRNAs upon hypoxic stress in cardiomyocytes and cardiac fibroblasts. Next, to further determine the potential roles of circRNAs, we applied gene-silencing mediated by RNA interference (RNAi) molecules, which specifically target the backsplice junction of both circRNAs. The knockdown of both circRNAs resulted in decreased metabolic activity as well as increased oxidative stress in cardiac fibroblasts and cardiomyocytes. More interestingly, specific overexpression of both circRNAs rescued the impaired metabolic activity, suggesting protective role of these circRNAs in metabolic dysfunction. Furthermore, reduced circAS4 levels led to inadequate wound healing as well as enhanced proliferation in cardiac fibroblasts. Collectively, these results indicate a possible mode of action of circAS1 and circAS4 in the regulation of cardiac energy metabolism as well as fibroblast features, in particular proliferation and migration. Therefore, both circRNAs could serve as potential therapeutic targets in cardiovascular disease and should be further explored in detail.