Lipotoxicity is an important contributor to cardiac dysfunction in obesity-associated heart disease18,19. Excessive uptake of fatty acids can result in either enhanced oxidation or abnormal accumulation of toxic lipid species such as ceramides and diacylglycerides, causing lipotoxicity in the heart and other solid organs, and the one of crucial underlying mechanism is ER stress and impaired UPR signaling20. Some evidence indicates that the activation of Sirt1/AMPK signaling may prevent cells from fatty-acid–induced oxidative stress and inflammation21, but the participation of Sirt1 in lipotoxicity-induced ER stress remains unclear. In the present study, we utilized mice fed a palmitate-enriched HFD and palmitate-treated H9c2 cells as in vivo and in vitro models of lipotoxicity and examined the cellular consequences, including ER stress and Sirt1 activity, as well as the crosstalk between these cellular phenomena.
There are studies showing that the lipotoxicity associated with ER stress increases ATF4 and CHOP mRNA expression, but the underlying mechanism still needs to be clarified20,22,23. Our data revealed that palmitate induces the expression of ER stress markers CHOP and ATF4 in vivo and in vitro, in line with other reports20,22,23. One study suggests that palmitate-induced cardiomyocyte dysfunction is mediated by ER stress and thereby promotes cell death3. Sirt1 may confer cardioprotection against ER stress. Alexandre et al. have reported that cardiac Sirt1 deficiency increases the contractile dysfunction caused by ER stress in a Sirt1 knockout mouse model, and the mechanism may involve eIF2α deacetylation24. According to our results, Sirt1−/− cardiomyocytes show higher expression of CHOP and ATF4 as compared to cardiomyocytes from the control mice fed either the SD or HFD, also suggesting a protective role of Sirt1 against HFD-induced ER stress in cardiomyocytes.
As an NAD+-dependent reaction, the protein deacetylation catalyzed by Sirt1 is accompanied by the hydrolysis of NAD+. Therefore, we propose that the decreased Sirt1 activity is associated with reduced NAD+ concentration in palmitate-treated H9c2 cells, in line with other reports25,26. Some research has revealed that changes in cytosolic NAD+ levels alter Sirt1 activity27–29. Additionally, one report has shown that reduced cellular NAD+ concentration, resulting from the conversion of NAD+ to NADH by the glucose metabolic pathway, leads to lower Sirt1 activity30. As a consequence of its dependence on NAD+ and therefore on the cellular NAD+/NADH ratio, Sirt1 has emerged as a key metabolic sensor with respect to various tissues31. There is evidence32,33 that in a high-energy state, such as that associated with an HFD and obesity, Sirt1 activity may decline with a decreased level of NAD+. Therefore, we propose that palmitate-induced ER stress diminishes the cytosolic level of NAD+, which in turn reduces Sirt1 enzymatic activity.
Our results indicate that the dominant negative Sirt1 H363Y mutant increases the expression of ATF4 and CHOP, also implying a protective role of Sirt1 (via its enzymatic activity) against ER stress24. Nevertheless, PA did not raise either ATF4 or CHOP expression in the H9c2 cells transfected with the plasmid encoding the Sirt1 H363Y mutant, and the same was true for H3K9Ac. We believe that the plasmid transfection of H9c2 cells attenuates the influence of PA or PA weakens the effects of the plasmid transfection. In addition, our data indicate that autophagy-related proteins beclin 1 and p63 are upregulated in the H9c2 cells treated with PA (Supplemental Data 1). PA-induced autophagy has been shown to abrogate the partial apoptosis caused by PA34. Therefore, ATF4 or CHOP expression in the H9c2 cells transfected with the Sirt1 H363Y mutant may be abrogated by the autophagy induced by PA. The increase in the H3K9Ac level in PA-treated H9c2 cells was attenuated by SRT1720 treatment, implying increased deacetylation function of Sirt1, although NAD+ concentration was not restored under these conditions. On the other hand, the PA-induced expression of ATF4 was not weakened by SRT1720 treatment. Furthermore, our findings show that the translocation of ATF4 to the nucleus increases upon SRT1720 treatment; as a consequence, the increased nuclear recruitment may enhance the gene-regulatory function of ATF4. With PA treatment, there was more binding between ATF4 and the CHOP promoter. By contrast, with PA plus SRT1720 treatment, the extent of binding between ATF4 and the CHOP promoter did not increase further, although SRT1720 made more ATF4 available in the nucleus. SRT1720 strengthened the activity of Sirt1 but had only a limited impact on the CHOP and ATF4 expression induced by PA treatment, suggesting involvement of more than a deacetylating activity. The protein–protein interaction and nuclear localization are being researched in more detail to decipher the current complicated findings.
In conclusion, PA-induced ER stress may be mediated by the upregulation of ATF4 and CHOP mRNAs and proteins. PA reduces the amount of cytosolic NAD+, which in turn suppresses Sirt1 activity. Nevertheless, the Sirt1 activator SRT1720 does not attenuate the expression of CHOP and ATF4 induced by PA but enhances nuclear translocation of ATF4.