Mitochondria are also involved in other
Mitochondria are also involved in other modalities of programmed cell death, particularly apoptosis. Intriguingly, the roles of mitochondria in apoptosis and ferroptosis are fundamentally different. In mitochondria-mediated apoptosis, the mitochondrion serves as a reservoir, storing apoptosis-regulatory proteins such as cytochrome c and Smac, which, upon release to the cytoplasm, promote caspase activation (Jiang and Wang, 2004, Wang, 2001). Notably, the cytosolic apoptotic activities of these mitochondrial proteins are independent of, and can be molecularly dissected from, their mitochondrial functions. As such, a compromise of mitochondrial integrity and loss of MMP, known as MOMP, precedes the release of these proteins and subsequent caspase-9 activation. Interestingly, in CDI ferroptosis, it is the canonical metabolic function of mitochondria that actively contributes to lipid ROS generation, a prerequisite of ferroptosis, and inhibition of TCA cycle or ETC function, or disruption of MMP, can all abrogate ferroptosis under this condition. Strikingly, a transient hyperpolarization of MMP can be observed during CDI ferroptosis, followed by the eventual collapse of MMP (Figures 4 and S4; Videos S1, S2, and S3). This unique feature is indicative of the crucial role of mitochondrial metabolic activity in CDI ferroptosis. Taken together, while mitochondria play a passive role in apoptosis by storing various pro-apoptotic proteins, they play a crucial and proactive role in CDI ferroptosis by fueling metabolism and lipid ROS production.
Further, the findings presented in this work provide novel mechanistic insights into some enigmatic observations in cancer biology. FH, a metabolic enzyme in TCA cycle, has been demonstrated to be a tumor suppressor. This is counterintuitive because one would assume that defects in the TCA cycle or ETC would be disadvantageous for cancer cell growth. Indeed, under normal culture condition, reconstitution of WT FH into UOK262/268 cancer Norethindrone synthesis rendered them proliferating better (Figure S6). So, what is the molecular basis underlying the tumor suppressive activity of FH? A pseudo-hypoxic pathway was proposed to be the link between FH mutation and tumorigenesis. FH defect leads to aberrant accumulation of fumarate, which, in turn, inactivates prolyl hydroxylases and thus stabilizes hypoxia-inducible factor (HIF)-1α (Isaacs et al., 2005, Selak et al., 2005). However, both HIF-1α and HIF-2α are dispensable for the formation of hyperplastic renal cysts in mice lacking expression of FH in the kidney (Adam et al., 2011), suggesting additional tumorigenic consequences caused by FH mutation. Our finding that loss of FH function renders cells more resistant to CDI ferroptosis (Figure 6) provides a novel mechanistic explanation of how dysfunction of FH contributes to malignancy independent of hypoxic signaling—we hypothesize that stronger resistance to ferroptosis caused by FH dysfunction contributes to tumorigenesis, although this is most likely not sufficient on its own. It remains to be determined why FH mutation preferentially drives tumorigenesis only in certain specific tissue types.
Conceptually, our finding that ferroptosis contributes to the tumor suppressive function of FH is important. It has been reported recently that p53 can suppress cancer development via its ferroptosis-promoting activity even when it loses its function to promote apoptosis, senescence, and growth arrest (Jiang et al., 2015, Wang et al., 2016). Based on these novel findings about FH and p53, we hypothesize that ferroptosis is a natural tumor suppressive mechanism under diverse biological conditions. The significance of this hypothesis in cancer biology is obvious. As for the biology of ferroptosis, tumor suppression might represent an authentic physiological role of this unique modality of cell death, which, until now, remains to be defined.
Acknowledgments The authors thank Dr. W. Marston Linehan for kindly providing the FH-mutant UOK262 and UOK268 cell lines, as well as their wild-type FH reconstituted counterparts. The authors thank members of the Jiang lab and the Thompson lab for critical reading and suggestions. This work is supported by NIH R01CA204232 (to X.J.), NIH R01CA201318 (to C.B.T.), a Geoffrey Beene Cancer Research fund (to X.J.), National Natural Science Foundation of China 31871388 (to M.G.), a Leukemia and Lymphoma Society Fellowship (to J.Z.), and NIH T32 fellowship 5T32GM008539-23 (to A.M.M.). This work is also supported by NCI cancer center core grant P30CA008748 to Memorial Sloan Kettering Cancer Center.