Protective effect of nicotinamide on high glucose/palmitate-induced glucolipotoxicity to INS-1 beta cells is attributed to its inhibitory activity to sirtuins

NAD+ is a component of NAD+/NADH redox pair involved in cel- lular metabolism as a cofactor for many dehydrogenases. In addi- tion, NAD+ is used as a substrate independent of its redox-carrier function by enzymes such as PARP, sirtuins, and glycohydrolases like CD38. In particular, activity of these enzymes depends on

the availability of NAD+. But, conversely, activity of these enzymes affects the intracellular pool of NAD+ because these enzymes also consume the NAD+ [50]. In fact, NAD+ was reported to belong to the fundamental common mediators of various biological pro- cesses including energy metabolism, mitochondrial functions, cal- cium homeostasis, antioxidation/generation of oxidative stress, aging, and cell death [51]. Therefore, the NAD+ level as well as the NAD+/NADH ratio is thought to constitute an important meta- bolic node involved in the control of many cellular events ranging from the regulation of energy metabolism to cell fate decision. Presently, the tNAD+ level gradually declined by HG/PA treatment in INS-1 beta cells. NAM prevented HG/PA-induced reduction of tNAD+ level. However, all attempts to replenish NAD+ through supplementation of NAD+, NMN, or kynurenine did not protect HG/PA-induced INS-1 cell death. Furthermore, the knockdown of NAMPT or NAMAT, as two key enzymes in the NAD+ salvage path- way, did not augment a HG/PA-induced cell death. In particular, the inhibition of PARP, whose activation consumes numerous NAD+, did not protect HG/PA-induced death. These data suggest that NAD+ depletion itself does not play a critical role in HG/PA-in- duced INS-1 beta cell death and that the NAD+ depletion observed in HG/PA-treated cells may be an epiphenomenon occurring during cell death. Our observation that NAD+ supplementation did not protect beta cells from HG/PA-induced glucolipotoxicity was unex- pected, since the supplementation of Nicotinamide Mononucleotide improved the glucose metabolism in the diabetic animal [37].

Sirtuins are highly conserved NAD+-dependent protein deacet- ylases and/or ADP-ribosyltransferases. In mammals, the sirtuin family comprises seven proteins (SIRT1-7), which vary in tissue specificity, subcellular localization, enzymatic activity, and targets. SIRT1 is mainly localized in the nucleus but is also present in the cytosol. SIRT2 is considered to be cytosolic. On the other hand, SIRT3, SIRT4, and SIRT5 have a mitochondrial targeting sequence, and were localized in mitochondria. Sirt6 is predominantly nuclear and Sirt7 was reported to reside in the nucleolus [52]. Several sir- tuins are known to be important regulators of cell metabolism. Activation of SIRT1 and SIRT4 was reported to stimulate oxidation metabolism of fatty acid and increase mitochondrial biogenesis [53,54]. SIRT3 was also reported to be involved in fatty acid oxida- tion during a fasting state [55]. Since lipid partitioning with re- duced fat oxidation was suggested to be a critical mediator in glucolipotoxicity [56], most trials to increase fat oxidation have been used to detoxify HG/PA-induced glucolipotoxicity [49,57,58]. Therefore, SIRT activation is supposed to be protective against HG/PA-induced INS-1 cell death whereas SIRT inhibition may augment the death. Contrary to our expectation, SIRT activa- tor resveratrol augmented HG/PA-induced death, whereas SIRT inhibitor sirtinol inhibited the death, even though the effect was not so great (Supplemental 5A and 5B). In particular, through a via- bility study and DNA fragmentation assay, it was demonstrated that the knockdown of some SIRTs (SIRT3 and SIRT4) could protect HG/PA-induced cell glucolipotoxiciyt. In addition, overexpression of SIRT3 and SIRT4 significantly augmented HG/PA-induced INS-1 cell death. It can be supposed that NAM’s protective effect on HG/PA-induced death is due to its inhibitory activity on some SIRTs such as SIRT3 and SIRT4. Although NAM as an inhibitor of SIRT2 or SIRT7 can augment HG/PA-induced cell death (Fig. 5A), the final ef- fect of NAM on HG/PA-induced death in our data was protective against HG/PA-induced death. It can easily be explained that NAM’s inhibitory effect on death-promoting SIRTs dominates its inhibitory effect on death-preventing SIRTs. Weak protective effect of sirtinol on HG/PA-induced cell death may also be due to the mixed effect on death-promoting and preventing SIRTs. On the other hand, NAM’s protective effect was greater than sirtinol’s pro- tective effect on HG/PA-induced cell death, suggesting that NAM may have another protective mechanism in addition to the inhibi- tion of SIRTs. Although the activation of SIRT1 and SIRT3 generally provides protecting signals against metabolically stressed cells [59–61], the activation of the SIRTs was also reported to be in- volved in cell death [25,62–65]. The mechanism as to how activa- tion of mitochondrial SIRT3 and SIRT4 augmented death signals in HG/PA-induced glucolipotoxicity is not clearly defined. The deple- tion of mitochondrial NAD+, by selective activation NAD+-consum- ing SIRTs in the mitochondria, may play a role in the induction of death signals in beta cell glucolipotoxicity. Target molecules by SIRT3 and SIRT4, which are involved in death signal activation, re- main to be identified. The reason as to why artificial supplementa- tion of NAD+ precursors did not protect HG/PA-induced cell death may be due to the inappropriate translocation into the mitochon- dria [66].

Apoptosis is the main mode of HG/PA-induced beta cell death. Release of mitochondrial cytochrome c into cytosol and subse- quent cascade activation of caspase have been reported in HG/ PA-induced beta cell death [67]. JNK activation and CHOP induc- tion as a damaged ER stress responses were reported to be key mediators in HG/PA-induced beta cell death [11]. In addition, down-regulation of Akt/Foxo signal was reported to be involved in the HG/PA-induced glucolipotoxicity [68,69]. However, the mechanism how the ER stress signals are activated, but how the Akt survival signals are down-regulated has not been clearly de- fined. Our recent study suggests that mitochondrial dysfunction may play a role in induction of the stress signals, but reduction of survival signals [57]. Interestingly, NAM significantly reduced death-related stress signal such as phospho-JNK and CHOP, but prevented HG/PA-induced reduction of Akt survival signal. This data suggest that metabolic alteration by NAM treatment may be an early change before induction of stress signals and reduction of survival signals and that the metabolic alteration in mitochon- dria may disturb normal cell fate-determining signals in HG/PA- treated beta cells. It can be concluded that modulation of metabo- lism is a target for protection of HG/PA-induced glucolipotoxicity. The hypothesis that NAM directly act at the level of mitochondria membrane pore formation to prevent cytochrome c release may

not be applicable to our HG/PA-induced death model because NAM protected HG/PA-induced cell death by reducing death, but promoting survival signals [20].

In conclusion, antioxidants were not protective against HG/PA- induced INS-1 cell death. Supplementation of NAD+ was also not protective against HG/PA-induced death. Pharmacologic inhibition or knockdown of PARP did not reduce death. However, knockdown of SIRT3 or SIRT4 reduced HG/PA-induced INS-1 cell death. There- fore, it is thought that NAM protects HG/PA-induced glucolipotox- icity to INS-1 beta cells through its inhibitory activity on mitochondrial Sirt3 and Sirt4.