3.1 Mitochondrial Homeostasis in Necroptosis Induction

Studies have shown that the MPT pores (MPTPs) are involved in the regulation of necroptosis by mitochondria [78-80]. MPTP is an IMM transmembrane protein that opens in response to ROS and Ca²⁺ overload in the mitochondrial matrix, leading to the entry of molecules >1.5 kDa into mitochondria [81]. This abrupt increase in IMM permeability, also known as MPT, causes mitochondrial osmotic swelling, alterations in mitochondrial transmembrane potential (ΔΨM), impairment of OXPHOS, accumulation of ROS, and eventually mitochondrial rupture [82-84]. The inhibition of MPT by the compound, S-15176, significantly decreased TNF-α-induced necroptotic cell death in microvascular endothelial cells (MVECs) [79]. Cyclophilin-D (Cyp-D) regulates MPTP opening [85-88] and studies suggest that Cyp-D plays a role in necroptosis [74, 78, 79, 89]. Diarachidonoylphosphoethanolamine induced necroptosis and necrosis in malignant pleural mesothelioma cells mediated by: (1) RIPK1 and (2) Cyp-D-mediated opening of MPTP and a subsequent decrease in ATP levels [78]. In Cyp-D-deficient MEFs, there were no ultrastructural changes typical of necroptotic cell death induced by TNF-α and zVAD-fmk in wild MEFs and these cells were partially resistant to TNF-α- and zVAD-fmk-induced necroptotic cell death [74]. Furthermore, Cyp-D-deficient MEFs were resistant to caerulein-mediated necroptotic cell death [74, 79]. Similarly, the inhibition of the opening of MPTP by cyclosporin A, a Cyp-D and calcineurin inhibitor, decreased TNF-α-mediated necroptotic cell death, whereas necroptotic death occurred after incubation with FK506, a non-Cyp-D binding and calcineurin inhibitor [79]. Cyp-D-deficient MVECs (obtained from Cyp-D-deficient mice) and Cyp-D inactivated MVECs, TNF-α did not produce necroptosis, suggesting that Cyp-D is involved in mediating Cyp-D necroptotic death [79]. RIPK3-induced necroptosis in endothelial cells was accompanied by MPTP opening and an increase in the phosphorylation of Ser31 in Cyp-D [80]. However, Cyp-D deletion did not rescue embryonic lethality in caspase-8-deficient mouse, suggesting that necroptosis may not occur through MPT [46].

In contrast, bromocriptine-induced RIPK3-mediated necroptosis in MMQ prolactinoma cells increased the levels of the phosphorylated form of Cyp-D and PGAM5, and the levels of these phosphorylated proteins were significantly decreased after incubation with Necrostatin-1, an inhibitor of necroptosis [89]. PGAM5 is an IMM protein that regulates mitochondrial dynamics and programmed/regulated cell death, due to its phosphatase activity [45]. Furthermore, the knockdown of genes for Cyp-D or PGAM significantly decreased bromocriptine-induced necroptosis in MMQ cells that was characterized by a decrease in cellular ROS levels and an increase in ATP levels [89]. PGAM5L (the long splice variant of PGAM5) is the kinase substrate of RIPK1/RIPK3, which activates and phosphorylates PGAM-S on mitochondria [44]. Inactivating either PGAM5L or PGAMS in HeLa or HT-29 cells inhibited TNF-α–Smac mimetic–z-VAD-fmk-induced necroptotic cell death [44]. However, there are studies suggesting that PGAM5 do not mediate the execution of necroptosis. TNF-induced necroptosis in mice was not prevented in PGAM5-deficient mice, which might be attributed to the expression of PGAM5-L but not PGAM5-S [90].

Alternatively, melatonin inhibited RIPK3-mediated necroptosis in endothelial cells by inhibiting Cyp-D phosphorylation and decreasing PGAM5 levels [80]. Similar to the effects of melatonin, PGAM5 knockdown in cardiac microvascular endothelial cells (CMECs) inhibited necroptosis by decreasing Cyp-D phosphorylation and reversing the mitochondrial membrane potential (MMP), which is an indirect marker of the opening of MPTPs [80]. Similarly, lipopolysaccharide (LPS)-activated necroptotic cell death in cardiomyocytes involved the activation of the RIPK3/PGAM5 signaling pathway [91]. Thus, necroptosis initiated by RIPK3 activates PGAM5, which increases Cyp-D phosphorylation and MPTP opening, increasing ROS accumulation, which ultimately produced to cell death [80, 91]. Interestingly, the MLKL inhibitor, GW806742X [92], inhibited Cyp-D-mediated necroptosis, indicating that MLKL activity regulates Cyp-D-mediated necroptosis. [79]. Furthermore, the translocation of RIP3-downstream MLKL from the cytosol to mitochondria occurred in TNF-α and zVAD-induced necroptotic cell death in MEFs but was absent in RIPK3-deleted MEFs [74]. Although only a few studies support the hypothesis that mitochondria have no function in necroptosis, the majority suggested that the RIPK1–RIPK3–MLKL–PGAM5–Cyp-D–MPTP–DRP1 axis was a putative mitochondrial mechanism for necroptosis execution.

3.2 Mitochondria-Derived ROS Regulates Necroptosis

There is evidence that mtROS are involved in the regulation of necroptosis [70, 93]. For example, necroptosis initiated by TNF-α alone or in combination with BV6 (a Smac mimetic) involved the accumulation of intracellular ROS in mouse fibrosarcoma cells, mouse fibroblasts, and FADD-deficient acute lymphoblastic leukemia [93-97]. ROS scavengers, such as butylated hydroxyanisole (BHA), N-acetylcysteine (NAC), α-tocopherol, ethyl pyruvate, and amytal, an inhibitor of ETC [98], significantly inhibited necroptosis induced by TNF-α alone or in combination with BV6 in Jurkat cells, MV4-11 cells, L929 cells, HeLa cells overexpressing RIP2, U937 cells, and primary peritoneal macrophages, MEFs in combination with dexamethasone, in acute myeloid leukemia cells (Tanoue and Jurkat cells) [93, 94, 96, 97, 99, 100]. In contrast, in HT-29 cells, neither BHA nor amytal significantly affected necroptosis induced by TNF-α, zVAD, and the Smac mimetic, compound 3 [73, 101]. Another study reported that mitochondrial depletion prevented necroptosis-induced ROS generation but not TNF-α- and zVAD-induced necroptosis, in 3T3-SA and murine peripheral lymph node endothelial cells [46]. In contrast, mitochondrial depletion in L929 cells decreased necroptotic cell death induced by TNF-α [99]. Thus, the results of these above mentioned studies suggests that the regulation of necroptosis by ROS in cells is context dependent [46, 99].

TNF-α-mediated cell death was reported to be regulated by mitochondrial but not cytosolic ROS, in L929 and RAW 264.7 cells, which was confirmed by the inhibition of ROS levels by 2-thenoyltrifluoroacetone, a mitochondrial respiratory chain complex II inhibitor but not by diphenyleneiodonium, an inhibitor of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase [102].

Similarly, BAY 87–2243, an inhibitor of mitochondrial complex I [103], induced necroptotic and ferroptotic cell death, by increasing cellular ROS and lipid peroxide levels [104]. Furthermore, shikonin-induced necroptosis in SHG-44 glioma cells was due, in part, to high levels of superoxide in mitochondria after MLKL activation, which significantly disrupted or decreased the MMP, producing mitochondrial dysfunction [105]. The inhibition of the subunit of mitochondrial complex I (NADH dehydrogenase (ubiquinone) 1 beta subcomplex 8, NDUFB8) significantly decrease cell death induced by TNF-α alone or in combination with BV6, in L929 cells [97]. Furthermore, mtROS levels regulate RIP1 by targeting three cysteines in RIP1 at positions 257, 268, and 586, and catalyze the autophosphorylation of Ser161, resulting in the recruitment of RIP3 into the necrosome, producing necroptosis [99]. TNF-α-mediated ROS generation and necroptosis are dependent on RIPK3 [106]. Alterations in RIPK3 levels produced proportional changes ROS levels caused by TNF-α plus zVAD [107]. Thus, mtROS targets downstream RIP1 oligomerization and upstream from RIP3 in TNF-α-mediated necroptosis [99]. Furthermore, ROS production occurs downstream of glucose metabolism in the mitochondria from the ETC and therefore, increases in glucose concentration increase ROS levels [108]. In hyperglycemic conditions, TNF-α and cycloheximide (CHX, an inhibitor of protein synthesis [109])-induced cell death changed from apoptosis to necroptosis, which occurred in the presence of increased mtROS and RIP1 phosphorylation and the loss of caspase-3, caspase-6, and caspase-9 [108, 110]. The suppression of glycolysis by 2-deoxyglucose inhibited hyperglycemia-induced TNF-α/CHX-induced necroptotic cell death [108]. Interestingly, the inactivation or pharmacological inhibition of enzymes involved in energy metabolism, such as glycogen phosphorylase (PYGL), glutamate-ammonia ligase (GLUL), and glutamate dehydrogenase 1 (GLUD1), inhibited ROS production caused by TNF plus zVAD, which produced cell death [106]. Overall, the results suggest that energy metabolism and ETC-associated ROS generation regulates RIPK1 and RIPK3 activity in necroptotic cell death.

PARP1 (poly (ADP-ribose) polymerase activation has also been implicated in oxidative stress and necroptosis [70, 111]. The overactivation of PARP-1 has been reported to produce dysfunction of mitochondrial respiratory chain complex I and an increase in ROS production [112]. The inhibition of PARP-1 by O-(3-piperidino-2-hydroxy-1-propyl)nicotinic amidoxime (BGP-15) maintained the ΔΨM and mitochondrial respiration and decreased ROS levels [113-115]. Glutamate-induced necroptosis occurred in the presence of increased ROS levels and PARP-1 activation and incubation with the PARP-1 and PARP-2 inhibitor, PJ34 [116] prevented glutamate-induced necroptosis in HT-22 cells [117, 118]. Thus, inhibiting PARP-1 may decrease the likelihood of necroptosis by preventing ROS formation and preserving mitochondrial function. In contrast, TRAIL-induced necroptosis in HT-29 cells was regulated by RIPK1/RIPK3-dependent PARP-1 activation; however, ROS production was not involved in TRAIL-mediated necroptotic cell death.

3.3 B-Cell Lymphoma 2 Protein Family Mediated Necroptosis

There are data suggesting that the prodeath B-cell lymphoma 2 (Bcl-2) proteins, located in the mitochondrial outer membrane, contribute to necroptotic cell death [119-121]. The Bcl-2 family is comprised of the proapoptotic pore-forming proteins (BAX, BAK, Bcl-2 related ovarian killer), antiapoptotic proteins (Bcl-2, Bcl-W, B-cell lymphoma-extra-large (BCL-X_L), myeloid leukemia-1 (MCL-1), Bcl-2-related protein A1 (Bcl-2A1 or BFL-1), and the proapoptotic BH3-only proteins (BMF), Bcl-2-associated death promoter (BAD), BH3-interacting domain death agonist (BID) Bcl-2-interacting killer, Bcl-2 interacting mediator of cell death, NOXA (Latin for damage), p53 upregulated modulator of apoptosis, Harakiri Bcl-2 interacting protein, and Mcl-1 ubiquitin ligase E3 (MULE, also known as BCL-G) [122, 123]. Bcl-2 interacting protein 3 (BNIP3), breast cancer 2, apolipoprtein6, microtubule-associated protein 1, and BECLIN-1, are proteins that contain other BH3-domains but their exact role as prodeath molecules remains to determined [122, 124].

The BMF gene was identified as a regulator of necroptosis induced by zVAD.fmk and TNF-α BMF silencing protected NIH 3T3 cells from TNF-α/CHX-induced necroptosis [125]. Similarly, the proapoptotic proteins, BAX and BAK, are involved in necroptosis [74, 126]. Green tea polyphenol (GTP) activated BAX and BAK interdependently, followed by mitochondrial translocation and subsequent cytochrome c release, to induce necroptosis in p53-deficient Hep3B cells [126]. The overexpress of BCL-2 prevented the GTP-induced necroptosis of Hep3B cells, whereas the knockdown of BAX and BAK significantly decreased cytochrome c release and necroptotic cell death in Hep3B cells [126]. The knockout of BAX and BAK produced resistance to TNF-α- and zVAD-fmk-induced necroptosis in MEFs [74]. Furthermore, Smac mimetic (BV6)/glucocorticoid (dexamethasone)-induced necroptosis involves BAK activation by RIP3, which was followed by MLKL-mediated ROS generation, that ultimately resulted in mitochondrial dysfunction and the death of acute myeloid leukemia cells (Tanoue and Jurkat cells) and colon carcinoma cells (MV4-11 and HT-29 cells), which was confirmed by a significantly decrease in necroptosis after Bak knockdown [100]. It has been reported that BCL-2 and the adenovirus E1B 19-kDa-interacting protein 3 (BNIP3) play a role in necroptosis [127, 128]. BNIP3 has been shown to mediate MPT and mitochondrial dysfunction, by activating BAX and BAK in MEFs [129, 130]. Necroptosis facilitated by TNF-α increases ROS generation and the insertion of BNIP3 in the mitochondria of A549 lung adenocarcinoma cells [127]. TNF-α increases the levels of BNIP3 and its expression level alters the susceptibility of A549 cells to TNF-α-induced necroptosis [127]. The mitochondrial matrix isoform of another BCL-2 protein, MCL-1, was significantly decreased during necroptosis; however, this loss was significantly repressed in RIPK3 deleted MEFs, indicating that necroptosis has an influence on mitochondrial death regulatory proteins and the resulting mitochondrial dysfunction [74]. Beclin-1, another BH3-domain protein, was reported to be a negative regulator of necroptosis [131, 132]. Beclin-1 suppressed MLKL oligomerization by interacting with phosphorylated MLKL, followed by its incorporation into the necrosome complex [131]. The inactivation of the BECLIN-1 gene sensitized HT-29, TC-1, L929, and Molm-13 cells, to TNF-α-induced necroptosis, by facilitating the pore-formation capacity of oligomerized MLKL [131]. However, the protein, BAD, increased the susceptibility of breast cancer cells to docetaxel-induced necroptosis [133]. The MLKL inhibitor, necrosulfonaminde [134], significantly inhibited docetaxel-mediated necroptosis in MDA-MB-231 breast cancer cells transfected with BAD, validating the role of BAD in the modulation of necroptosis [133].

In summary, mitochondria play a contextual role in the regulation and induction of necroptosis, mainly through mechanisms that involve ROS generation. Mitochondrial-mediated necroptosis involves MPT, BCL-2-interacting proteins, the RIPK1–RIPK3–MLKL–PGAM5–Cyp-D–MPTP–DRP1 pathway, and RIPK2–RIPK3 interactions. Mitochondrial dynamics, energy metabolism, and ROS accumulation are all critical regulators of mitochondrial dynamics, energy metabolism, and ROS accumulation.

4.1 Mitochondrial Oxidative Stress and Autophagy

Numerous studies have shown that ROS regulates autophagy and that mitochondria are the primary source of autophagy regulation [159-162]. The attenuation of autophagy by the ROS scavengers/antioxidants, such as NAC, validates the role of mtROS in the regulation of autophagy [163, 164]. For example, deoxypodophyllotoxin-induced apoptosis increased mtROS production in PC-3 cells, which further activated extracellular signal-regulated kinases/mitogen-activated protein kinases (MAPK) signaling and induced protective autophagy [165]. Similarly, the loss of solute carrier family 4 member 11 (SLC4A11), a NH3-sensitive transporter present in the plasma membrane and mitochondria of the corneal epithelium [166], increasedde mtROS, which disrupts the transcription factor EB (TFEB) signaling pathway, contributing to lysosomal dysfunction and impaired autophagy in the corneal endothelium [167]. Furthermore, mtROS produced by the overexpression of ubiquinol cytochrome c reductase binding protein (UQCRB) facilitated autophagic flux by lysosomal biogenesis through activation of the transient receptor potential cation channel, mucolipin subfamily 1 (TRPML1) Ca²⁺ channels and TFEB nuclear translocation and protects against HCT116 colorectal cancer [168]. In contrast, the inhibition of mtROS generation by the UQCRB inhibitor, A1938 [169] decreased autophagy in UQCRB-overexpressing HEK293 cells [168]. Honokiol-induced cytoprotective autophagy in PC-3 cells was partly dependent on mitochondria-derived ROS accumulation, as indicated by the partial suppression of the LC3BII protein level, after incubation with NAC [170]. NAC inhibited recombinant human arginase (rhArg)-induced autophagy in A375 human melanoma cells, indicating that rhArg-induced autophagy was mediated by ROS [171]. Another study reported that mtROS generation occurs upstream of autophagy and this produced autophagy in prostate cancer cells [172]. BIX-01294 (a histone methyltransferase inhibitor [173]) induced ROS-dependent autophagy in breast cancer cells, by increasing the levels of mitochondrial superoxide and mitochondrial and cytosolic H₂O₂ levels [163]. Furthermore, mtDNA-less human pancreatic ρ0P29 cancer cells that generated minimal ROS levels after incubation with SSHE, an ethanol extract of Syussai ginger, were resistant to autophagy, indicating that SSHE produced mtROS-mediated autophagy in human pancreatic cells [164]. The mtROS produced by damaged mitochondria has been shown to impair lysosomal function and produce aberrant autophagic flux in macrophages [174]. Interestingly, the increase in the level of the decidual protein, C10ORF10/DEPP, induced by progesterone, a transcriptional target of Forkhead box O3 (FOXO3) that is present in mitochondria [175], mediates ROS accumulation and regulates FOXO3-induced autophagy in human neuroblastoma cells [176]. Similarly, the attenuation of autophagy by ROS suppressors/antioxidants, such NAC, diphenyleneiodonium chloride (DPI), and mitochondrial superoxide dismutase 2 (SOD2), are similar to manganese (III) tetrakis (4-benzoic acid) porphyrin (MnTBAP; ROS scavenger) [177], validates the role of mtROS in the regulation of autophagy [163, 164, 176]. In stress conditions, such the exposure of human retinal pigment epithelial cells (ARPE-19) to ethanol, mtROS levels are increased and mitochondrial fission activation occurs, activating the autophagy pathway [178]. In vitro, metformin increases mtROS in CD8+ tumor-infiltrating T cells (CD8TILs), which in turn activate nuclear factor erythroid 2-related factor 2 (Nrf2), resulting in mTORC1 activation, phosphorylation of p62 and the induction of autophagy [179]. The inhibition of autophagy in prostate cancer (PC-3 cells), human retinal pigment epithelium (ARPE-19 cells), and CD8TILs, after incubation with mitochondria-targeted antioxidant drugs, such as mitoquinone (MitoQ), mitoTEMPO, or the mitochondrial SOD2 mimic, manganese (III) tetrakis (4-benzoic acid) porphyrin (MnTBAP), confirms the role of mtROS in the regulation of autophagic response [165, 178, 179].

In contrast, another study suggested that mtROS inhibits phosphatase and tensin homolog deleted on chromosome ten (PTEN) enzyme, a primary downstream effector of the PI3K–Akt–mTOR signaling pathway, which catalyzes the biodegradation of PIP3, thereby inhibiting PI3K signaling [180]. Furthermore, high levels of mtROS have been reported to activate Akt by downregulating phosphoprotein phosphatase 2A, which plays a crucial role in AKT inactivation [181] and positively regulates the PI3K–Akt–mTOR signaling pathway, thus inhibiting autophagy signaling [182].

4.2 Mitochondrial Energetics and Autophagy

Cancer cells require a significant supply of nutrients, particularly glucose and glutamine, to synthesize macromolecules and replenish the tricarboxylic acid (TCA) cycle, to meet the energy required for their proliferation [183]. It has been reported that amino acids, such as glutamine, inhibit macroautophagy before the fusion of the autophagosomes and lysosomes [184] and inhibit the starvation of glutamine induced prosurvival autophagy in colorectal carcinoma cells [185]. Glutamine deprivation has been reported to translocate and activate TFEB, producing micropinocytosis-induced autophagy in pancreatic ductal adenocarcinoma cells [186]. During amino acid deprivation, autophagy induces the reactivation of mTORC1, a major regulator of cell growth and autophagy [187, 188]. The activation of mTORC1 signaling produces an addiction to glutamine in cancer cells [189]. Amino acid deprivation depletes α-ketoglutarate (α-KG), which further inactivates HIF prolyl hydroxylases (PHDs), preventing mTORC1 activation and inducing autophagy [190]. In cells with high levels of glutaminolysis, cytosolic α-KG is transported from mitochondria to the cytosol by mitochondrial carrier protein, solute carrier family 25 member 11 (SLC25A11 [188]), activates PHDs, which further increases amino acid-dependent-mTORC1 activation in a HIF-1α-independent manner, inhibiting autophagy [188, 190, 191]. Furthermore, activating transcription factor 4 is activated upon glutaminolysis inhibition, increasing the levels of DNA damage inducible transcript 4, which inhibits mTOR and increases the likelihood of prosurvival autophagy [192]. Similarly, the enzyme, GLUD, which mediates the catalysis of glutamate to α-KG, modulates autophagy by limiting the production of ROS and activating mTORC1, based on autophagy initiation after GLUD1 knockdown, which was characterized by an increase in ROS levels and a decrease in mTORC1 activation [193]. GLUD also inhibited autophagy by the GTP-mediated activation of RAS-related GTPase and the subsequent activation of mTORC1 [193, 194]. Under conditions of amino acid deprivation, FOXO3 increased the expression of glutamine synthetase, which inhibited mTOR and increased autophagy in human colon cancer DLD1 cells [195]. The inactivation of glutamine synthetase decreased FOXO-mediated LC3II formation and decreased autophagy, indicating that glutamine synthetase, which catalyzes the conversion of glutamate and ammonia to glutamine, plays a role in the regulation of autophagy [195]. Chronic respiratory chain deficiency due to the knockdown of ubiquinol-cytochrome c reductase core protein 1 (UQCRC1), a subunit of respiratory chain complex III [196], produced an increase in folliculin expression, which downregulated AMPK signaling and decreased phosphatidylinositol 3,5-bisphosphate (PI(3,5) P2) levels [197]. The knockdown of UQCRC1 further decreased mucolipin TRP cation channel 1 activity, producing the accumulation of Ca2+ in the lysosomes, increasing lysosomal volume and the accumulation of autophagosomes, which inhibited autophagy [197]. PRKCB, a member of the protein kinase C (PRKC) family, inhibited autophagy by increasing the loss of ΔΨM and negatively regulating mitochondrial energy [198]. This finding was validated based on data indicating that PRKCB-depleted MEFs had increased autophagy with restored mitochondrial bioenergetics, based on an increase in ΔΨM [198]. These data indicate autophagy can be regulated by the bioenergetics state of mitochondria.

4.3 Mitochondria Degradation by Autophagy

The selective degradation of mitochondria by autophagy is known as mitophagy [199, 200]. Mitophagy is critical for the turnover of mitochondria and the removal of unhealthy and damaged mitochondria [199]. Mitophagy is regulated by proteins that are involved in the maintenance of mitochondrial integrity, morphology, and ubiquitination [201]. Mitochondria undergo fission to separate damaged mitochondria from the healthy mitochondria, which are subsequently removed by mitophagy [202]. In damaged mitochondria, PINK1 is accumulated on the OMM due to impaired intermembrane transport to IMM [40]. PINK1, upon phosphorylation at serine 228 and 402, recruits the enzyme, cytosolic E3 ubiquitin ligase, and PARK2, to the OMM [203]. The protein, autophagy-regulating protease 4 (ATG4), is also involved in PINK1/Parkin-mediated mitophagy. Indeed, ATG4 interacts with ATG9A and increases the interaction between the phagophore and ER, promoting mitophagy independent of LC3/the γ-aminobutyric acid receptor-associated proteins (GABARAP) lipidation system [204]. Similarly, autophagy receptors, such as Nuclear dot protein 52 kDa (NDP52) (also known as calcium-binding and coiled-coil domain 2) and the multifunctional autophagy kinase, TANK-binding kinase 1 (TBK1), activate mitophagy downstream or independent of the PINK1/Parkin axis, by recruiting and activating the ULK1 complex without binding to LC3 proteins [205].

BH3-only proteins, such as the Bcl-2 interacting protein 3 (BNIP3) and BNIP3L (BNIP3-like)/Nix, also mediate mitophagy. The phosphorylation of Ser34 and Ser35 juxtaposed at the LC3 interacting region (LIR) domain of BNIP3L, increases its interaction with ATG8 family proteins, such as LC3/GABARAP and increases mitophagy in HeLa cells [206]. The accumulation of BNIP3L on the OMM, to a certain level, induces the loss of ΔΨM, producing dysfunctional mitochondria, which activates mitophagy [207-212]. Furthermore, BNIP3 and BNIPL inhibit mTOR activity by binding with Ras homolog enriched in brain (Rheb) proteins, which facilitate mitophagy [213]. Similarly, under hypoxic conditions, the BH3-domains of BNIP3 and BNIP3L release Beclin-1 by competing with Beclin1–Bcl-2 and Beclin1–Bcl-Xl complexes, thus inducing the autophagy of mitochondria to prevent ROS and cell death [214-216]. In contrast, hypoxic-colon carcinoma-induced mitophagy was reported to be independent of BNIP3 and BNIP3L and was directed by the activation of the AMPK–ATG5 axis [217]. Furthermore, another protein involved in mitophagy is activating molecule in activating molecule in beclin-1-regulated autophagy (AMBRA1). Under basal condition, AMBRA1 is located in the mitochondria and its proautophagic activity is inhibited by Bcl-2 [218]. Mitophagy induction after mitochondrial depolarization leads to the binding of AMBRA1 to LC3 via its LIR motif, producing Parkin-mediated autophagy [219]. Intriguingly, high levels of AMBRA1 can induce mitophagy, even in cells lacking Parkin or PINK1, indicating that AMBRA1 mediates mitophagy independent of Parkin, PINK1, or p62-recruitment [219]. Under hypoxic conditions, dephosphorylated FUN14 domain containing 1 (FUNDC1), a OMM protein, binds with LC3B via its LIR and activates mitophagy [220]. The absence of mitochondria engulfment by autophagosomes and intact mitochondrial integrity with no biodegradation of mitochondrial proteins in FUNDC1-silenced cells, indicated that FUNDC1 is involved in mitophagy under hypoxic conditions [220]. The knockdown of the gene for the mitochondrial fission protein DNM1L/DRP1, prevented carbonyl cyanide-p-trifluoromethoxyphenylhydrazone- and selenite-induced mitophagy [221]. However, the knockdown of the gene for the mitochondrial fusion protein, OPA1, increased mitophagy, suggesting that FUNDC1 interacts with OPA1 and DNM1L to induce mitochondrial fission and mitophagy [221].

4.4 Mitochondrial Membrane Contribution to Autophagy

Mitochondria are one of the organelles, besides the ER and Golgi bodies, that are source of membrane and lipids for the formation, expansion, and fusion of autophagosomes [222]. The membrane contact sites (MSCs) for autophagosome biogenesis include ER–phagophore MSCs, ER–mitochondria MSCs, mitochondria-associated ER membranes (MAMs), ER–plasma membrane MSCs, ER–Golgi intermediate compartments (ERGIC), and plasma-membrane-derived vesicles [205, 223–225]. During nutrient deprivation, mitochondria are involved in autophagosome biogenesis, as they transfer lipids from its outer membrane to newly formed autophagosomes, following the recruitment of the early and late autophagosome markers, ATG5 and LC3, on mitochondria [222]. The ER-resident SNARE protein, syntaxin 17, translocates to MAMS after nutrient deprivation and recruits the preautophagosome/autophagosome marker, ATG14 and the autophagosome-formation marker, ATG5, to the MAMs, to produce autophagosomes [226]. Furthermore, the formation of MAMs occurs by the tethering of the ER protein, VAMP-associated protein B (VAPB), to the mitochondrial protein, protein tyrosine phosphatase-interacting protein 51 (PTPIP51), followed by ER Ca²⁺ delivery to mitochondria [227-230]. The inactivation of the gene for VAPB and PTPIP51 decreases the mitochondria contact sites and increases autophagosome formation, inducing autophagy [227]. The overexpression of VAPB and PTPIP51 increases the ER–mitochondria contact sites and inhibits rapamycin- and torin 1-induced autophagosome formation, which impairs autophagy [227]. Autophagy was restored by decreasing the IP3-receptor-mediated Ca²⁺ supply to mitochondria in cells overexpressing VAPB and PTPIP51 [227], indicating that autophagy is activated when Ca²⁺ transport from MAM-located IP3 receptors to mitochondria is disrupted [227, 231–233]. Another ER protein involved in autophagy regulation is etoposide-induced protein 2.4 (EI24). EI24 interacts with glucose-regulated protein 75, an OMM chaperone protein, that mediates MAM integrity by forming a complex with VDAC1 and the inositol 1,4,5-trisphosphate receptor (IP₃R) [234, 235]. A deficiency of EI24 affects Ca²⁺ transfer from the ER to mitochondria and inhibits autophagy [234]. Similarly, the sigma-1 receptor (Sig-1R), an ER chaperone protein located primarily at the MAM, also regulated Ca²⁺ signaling between ER and mitochondria and its ablation or suppression impaired the assembly and clearance of autophagosome [236, 237]. Interestingly, under energy stress, cytosolic AMPKα1 was translocated to MAMs tethered to MFN2, where it interacted with and phosphorylated MFN2 [238]. This, in turn, increased MAMs levels, inducing autophagy. Furthermore, MFN2-depleted cells had a significant decrease in autophagic capacity and defective MAM production, indicating the importance of the AMPK–MFN2 axis in the induction of autophagy [239]. Lipid rafts in MAMs containing ganglioside GD3, also known as the brick of lipid rafts [240], interacts with AMBRA1 and WD repeat domain phosphoinositide-interacting protein 1 (WIPI1), to initiate autophagy [241]. Furthermore, the depletion of the enzyme, ST8 alpha-N-acetyl-neuraminide alpha-2,8-sialyltransferase 1 (ST8SIA1), which is involved in ganglioside production, decreased ER–mitochondria interaction and impaired the nucleation of autophagosomes [241, 242]. In conclusion, MAM plays a role in autophagosome biogenesis and disruption of MAMs or ER–mitochondria contact inhibits autophagy.

Mitochondria regulate autophagy via diverse mechanisms, including ROS signaling, energy metabolism, and mitochondrial mitophagy among others. For example, mitochondria-generated ROS can induce or inhibit autophagy, whereas mitochondrial bioenergetics modulate autophagic responses under stress or nutrient-depleted conditions. To maintain cellular homeostasis, mitochondria are specifically targeted for degradation by mitophagy. Furthermore, mitochondria-associated membranes (MAMs) facilitate autophagosome formation by providing structural components and mediating Ca²⁺ transfer from mitochondria to the ER. Mitochondria play a key role in autophagy regulation by integrating metabolic and signaling cues. Thus, mitochondria regulate autophagy by integrating metabolic and signaling cues to control cellular health.

5.1 Mitochondrial Lipids Affect Ferroptosis

Cardiolipins (CLs), a unique PL found exclusively in the inner mitochondrial membrane (IMM) and other PLs, such as phosphatidylcholine (PC) and PE, undergo oxidative reactions in the kidneys of Gpx4^−/− mice and in HT-1080 cells incubated with FINO2, an inducer of ferroptosis [259, 267]. The knockdown of SMPD1 and TAZ (tafazzin) genes, which mediate CLs remodeling [268], exacerbated acetaminophen-induced mitochondrial dysfunction and ferroptosis in hepatocytes [262]. The oxidized PLs and CLs primarily accumulated in the IMM of ferroptotic cells [259]. According to Gao et al. [269], lipid peroxides are co-localized in the mitochondria at early time points and in the plasma membrane at later time points. Similarly, doxorubicin (DOX)-induced ferroptosis is characterized by increased lipid peroxidation in mitochondria but not in the cytoplasm of cardiomyocytes [270, 271]. MitoTEMPO, an antioxidant that targets mitochondria [272] but not TEMPO, a nonmitochondria targeted antioxidant, abolished DOX-induced lipid peroxidation and subsequent ferroptosis, indicating that the oxidative damage induced by mitochondria is a major mechanism of ferroptosis induction [271]. The efficacy of the mitochondria-targeted nitroxide antioxidant, XJB-5-131, to rescue FINO2 (endoperoxide-containing 1,2-dioxolane)-induced ferroptotic death of HT-1080 cells, was 39-fold greater than that of the cytosolic targeted antioxidant, JP4-039 [273], highlighting the critical role of mitochondrial lipid peroxidation in ferroptosis [267]. However, only a few studies have shown the accumulation of lipid peroxides to be extra-mitochondrially, and it occurs primarily in the ER [253].

5.2 Mitochondrial Dynamics in Ferroptosis

Mitochondria establish an electrochemical proton gradient across the mitochondrial membrane, commonly known as MMP, by utilizing oxidizable substrates, which produces ATP [274]. Several ferroptosis inducers, such as erastin, an amino acid-free medium, glutamate, and cysteine deprivation, resulted in MMP hyperpolarization [269]. The incubation of human fibrosarcoma HT1080 cells with the mitochondrial OXPHOS uncoupler, carbonyl cyanide m-chlorophenyl hydrazone [275], impaired MMP, depolarized mitochondria and prevented the accumulation of lipid peroxides, due to cysteine deprivation and ferroptosis [269]. Erastin binds to the VDACs, VDAC2, and VDAC3 [276], whereas RSL3 produced a conformational change in VDAC2 [277] by inducing carbonylation [278]. VDAC facilitated the flux of ions and metabolites across the OMM [279]. The binding of erastin and RSL3 prevented the tubulin-dependent closure of VDAC, thus producing mitochondrial hyperpolarization [278]. The knockdown of the VDAC2/3 genes prevented erastin-induced ferroptosis [280]. In contrast, acetaminophen-induced ferroptosis in hepatocytes, as well as erastin- or RSL3-induced ferroptosis, in MEFs, HT-22, HepG2, and HA22T/VGH cancer cells, resulted in the loss of the MMP [262, 263, 265, 281]. Ferroptosis is also regulated by the mitochondrial outer membrane protein, FUN14 domain-containing 2 (FUNDC2). FUNDC2 interacts with and destabilizes the mitochondrial GSH transporter, SLC25A11, decreasing mitoGSH levels, thus, contributing to DOX-induced ferroptosis [282]. The ablation of the FUNDC2 gene prevented DOX-induced cardiomyopathy by ferroptosis and the knockdown of the SLC25A11 gene in FUNDC2-KO cells significantly decreased mitoGSH levels and augmented erastin-mediated ferroptosis [282].

5.3 Mitochondrial Energetics in Ferroptosis

The major function of mitochondria is to produce ATP by the process of OXPHOS that involves electron transport and chemiosmosis [283]. During this process, the transfer of electrons during the production of ATP leads to the generation of ROS, such as hydrogen peroxide (H₂O₂), hypochlorous acid (HOCl), hydroxyl radical (HO^•), alkoxyl radical (RO^•), superoxide anion (O₂^•−), peroxyl radical (RO₂^•), hydroperoxyl radical (HO₂^•), and singlet oxygen (¹O₂) [284]. The enzyme, NADPH oxidase (NOX), and the family of superoxide-producing enzymes, such as NOX1-5 and DUOX1-2, have been shown to produce lethal levels of ROS that activated erastin-mediated ferroptosis [280]. The erastin-induced ferroptotic cell death in human non-small-cell lung cancer Calu-1 cells was inhibited by DPI (a NOX inhibitor [285]), GKT137831 (a NOX1/4-specific inhibitor [286]), and 6-aminonicotinaminde (6-AN, a NADPH-producing pentose phosphate pathway (PPP) inhibitor [287]), suggesting that ferroptosis is induced by NADP-dependent ROS production [280]. Similarly, the knockdown of the genes for the PPP enzymes, glucose-6-phosphate dehydrogenase and phosphoglycerate dehydrogenase, which are involved in NADPH production [288], rescued the erastin-induced ferroptosis in Calu-1 cells [280]. In contrast, erastin-induced ferroptosis in HT-1080 fibrosarcoma cells was only moderately inhibited by the NOX inhibitors, DPI and GKT137831, and 6-AN [258]. This finding suggested that in addition to the PPP/NOX pathway, other pathways can mediate ferroptosis after the initial inhibition of system X_c⁻ [280]. High levels of mtROS were reported in ferroptotic death induced by glutamate in HT-22 mouse hippocampal neurons cells, acetaminophen in Hepa1-6 murine hepatoma cells, and RSL3-challenged MEF and HT-22 cells [262, 265, 289].

Studies indicate that ferroptosis is dependent on cellular metabolism [290]. Numerous studies have reported that mitochondrial dysfunction causes a switch from aerobic respiration to glycolysis for ATP production in cancer cells, known as the Warburg effect [291]. However, mitochondrial failure and increased glycolysis in ferroptosis appear to be unrelated [258, 292]. Erastin-induced ferroptosis increased ATP synthesis and decreased the rate of glycolysis [258, 292]. Erastin-induced ferroptotic cell death involved the opening of VDAC in HepG2 and Huh7 human hepatomacarcinoma cells [278]. This erastin-mediated effect permitted the flux of ATP, ADP, Pi, and metabolite exchange across the OMM, thus promoting mitochondrial metabolism, increased MMP, and increased mtROS levels, followed by mitochondrial dysfunction and the ferroptotic death of HepG2 cells [278, 293]. Cells with greater mitochondrial respiration and lower glycolytic flux, such as HepG2 cancer cells, were more vulnerable to RSL3-induced ferroptosis, highlighting the specific function of mitochondria in RSL3-induced ferroptosis [281]. Interestingly, mitochondrial-deficient cells are less vulnerable to ferroptosis inducers, such as erastin and cysteine deprivation, with cysteine-deprived mitochondrial depleted cells being more resistant to ferroptosis than cells incubated with erastin [269]. Recent studies have shown that the glutaminolysis–TCA cycle, in combination with the ETC, play an essential role in ferroptosis regulation [294, 295]. Glutaminolysis is the process by which cellular glutamine is biotransformed to glutamate, which is biotransformed to α-KG, to drive the TCA cycle by anaplerosis [296-298]. Glutamine is a carbon source for the TCA cycle and a nitrogen source for the de novo synthesis of amino acids, nucleotides and hexosamines, which is used by cells that are growing and proliferating [299, 300]. Glutamine absorption occurs via the glutamine importers, SLC1A5/SLC38A1, and glutamine is biotransformation to glutamate by glutaminase and eventually to α-KG, either by glutamic-oxaloacetic transaminase 1 (GOT1) or GLUD1-mediated glutamate deamination [301]. The glutaminolysis regulating genes, SLC38A1 and glutaminase 2 (GSL2), positively regulated ferroptosis [294]. Similarly, miR-137, a tumor suppressor microRNA in several cancers, such as ovarian cancer, pancreatic cancer [302], inhibited ferroptosis by directly inhibiting SLC1A5 [303]. Furthermore, l-gamma-glutamyl-p-nitroanilide, an inhibitor of the SLC1A5 (ASCT2) and SLC38A1 (SNAT1) transporters [304], decreased lipid peroxide levels and rescued erastin-induced ferroptosis in melanoma cells [303]. Furthermore, the knockdown of mitochondrial enzyme glutaminase 2 (GLS2) (catalyzes the biotransformation of glutamine to glutamate) but not GLS1 or pharmacological inhibition of GLS, by compound 968 [305], significantly inhibited ferroptosis [299]. The inhibition of transaminases by amino-oxyacetate or the inactivation of the gene for of GOT1 but not GLUD1, prevented ferroptosis [295]. Interestingly, the supplementation of α-KG with downstream TCA cycle metabolites, such as succinate, fumarate, and malate, substituted for glutamine and induced ferroptosis in MEFs and HT1080 human fibrosarcoma cells in the presence of cysteine deprivation of inhibition of system Xc− [269, 295, 303]. Furthermore, inhibitors of ETC components, such as rotenone (mitochondrial complex I inhibitor), diethyl butylmalonate (complex II inhibitor), antimycin (complex III inhibitor), and sodium azide (NaN₃, complex IV inhibitor), prevented ferroptosis by erastin or cysteine deprivation, suggesting that the ETC mediated the induction of ferroptosis [269]. In contrast, recent studies have shown that energy stress activates AMPK, an enzyme that is an energy crisis sensor in tumor cells, to produce ATP, and this inhibited ferroptosis [290, 306–308]. Furthermore, AMPK phosphorylates and downregulates acetyl-CoA carboxylase 1 and 2 (ACC1/2), thereby decreasing the biotransformation of acetyl-CoA to malonyl Co-A, inhibiting the production of PUFAs, ultimately blocking cysteine deprivation and GPX4 inhibition-mediated ferroptosis [307]. The inhibition of the protein liver kinase B1 (also known as serine/threonine kinase 11), an upstream kinase that activates AMPK [309], increased the likelihood of ferroptosis in non-small cell lung cancer cells [308]. Overall, these findings show the importance of mitochondrial metabolism in regulating lipid peroxidation and ferroptosis via metabolic processes, such as the glutaminolysis–TCA cycle–ETC axis.

5.4 Mitochondrial Iron Metabolism in Ferroptosis Regulation

Mitochondria play an important role in the metabolism, utilization, and homeostasis of cellular iron [310]. Extracellular iron is internalized into cells by the transferrin receptor (TFR)-mediated endocytosis of transferrin-bound iron (Fe³⁺), followed by the six-transmembrane epithelial antigen of prostate 3 (STEAP3)-mediated reduction of Fe³⁺ to Fe²⁺, direct absorption of Fe²⁺ through divalent metal transporter 1 (DMT1), zinc transporter 8, or heme import and its biotransformation to iron by heme oxygenase (HO) [310-314]. The efflux or export of Fe²⁺ is regulated by the transporter, ferroportin 1 (FPN), the only known transporter that exports elemental iron from cells [315]. Typically, cellular iron is either bound to heme or is present in iron–sulfur cluster (ICS) or stored as Fe³⁺ in the iron storage protein, ferritin [316]. Ferritin releases Fe³⁺ upon its biodegradation through proteasome-mediated or autophagic process, also known as ferritinophagy [317]. Ferroptosis occurs in the presence of dysfunctional iron metabolism. The upregulation of the iron regulatory protein 2 (IRP-2) increased ferroptosis by producing an increase in cellular iron intake by increasing the expression of TFR-1 [318]. The immuno-induced depletion of transferrin in serum or by inactivating the TFR gene, prevented MEFs ferroptotic cell death [295]. Furthermore, the knockdown of the gene that codes for the protein, IRP-2 and the incubation of d-gal-induced aged PC12 rat adrenal medullary tumor cells with the iron chelators, deferoxamine, ciclopiroxolamine, 2, 2-bipyridyl, or iron-free bovine apo-transferrin, suppressed ferroptotic cell death [295, 318]. Overall, these findings suggested that the iron carrier protein, transferrin, is required for ferroptotic cell death.

Iron is transported into the mitochondrial matrix by the DMT1, siderofexin (SFXN1), or mitoferrin, where it is utilized for the biosynthesis of heme or ISCs or is stored as mitochondrial ferritin [6]. Interestingly, an increase in cellular Fe²⁺ levels via ferritinophagy, increased the expression of SFXN1 (has mitochondrial serine transporter activity [319]) on the mitochondrial membrane [320]. This process increased the entry of cytoplasmic Fe²⁺ to mitochondria, increasing mtROS levels and the induction of ferroptosis, in sepsis-induced cardiac injury [320]. The incubation of cardiomyocytes with apelin-13 (an endogenous ligand for the APJ receptor involved in regulating fluid homeostasis, cardiovascular function, and insulin sensitivity [321]) produced excessive ferritinophagy and subsequent SFXN1-mediated mitochondrial iron accumulation, producing mtROS formation and ferroptosis [322]. In contrast, HO-1 has been reported to have a dual (protective as well as detrimental) role in the regulation of ferroptosis [323]. The RNA interference silencing of HO-1 facilitated ferroptosis by erastin in hepatocellular carcinoma cells (HCC) and renal proximal tubular epithelial cells and ferroptosis produced by sorafenib, an inhibitor of system Xc− in HCCs [324, 325]. However, in HT-1080 fibrosarcoma cells, hemin, an HO-1 substrate that increases the expression and enzymatic activity of HO-1, increased erastin-induced ferroptotic cell death [326]. However, the incubation of HCC with bilirubin or biliverdin (a metabolite produced by the biotransformation of heme to Fe²⁺ [327]) did not retain erastin-induced ferroptosis [328]. The induction of ferroptotic death by BAY 11–7085, an inhibitor of IκBα phosphorylation that produces NF-κB inactivation [329], in triple-negative breast cancer and glioblastoma cells, involved the upregulation of HO-1, and HO-1's effect was decreased by the specific HO-1 inhibitor, zinc protoporphyrin-9 [330]. Furthermore, HO-1 upregulation was the major mechanism involved in DOX-induced ferroptosis in cardiomyocytes [271]. In these studies, the biodegradation of heme by HO-1 increased free iron levels and significantly increased lipid peroxidation in mitochondria, producing ferroptosis [271, 328, 330]. ISCs have been reported to regulate ferroptotic cell death [331]. ISCs are located in the mitochondrial matrix and are assembled by the ISC assembly enzyme, composed of cysteine desulfurase (NFS1), small accessory protein (ISD11), and the iron chaperone, frataxin (FXN) [332]. A deficiency of ISCs increased iron loading, activating IRP-2 and increasing the likelihood of ferroptosis [333]. Recently, it was reported that FXN controls iron homeostasis and mitochondrial activity and it was a crucial regulator of ferroptosis. The expression of FXN inhibits ISCs assembly, induces iron deprivation stress, increases the accumulation of free iron, and significantly increases erastin-induced lipid peroxidation and ferroptosis in human fibrosarcoma HT-1080 cells [334]. The inhibition of cysteine desulfurase NFS1, an enzyme that removes sulfur from cysteine [335], induced ferroptosis in lung adenocarcinoma cells [336]. Furthermore, the mitochondrial iron–sulfur (Fe–S) protein, CDGSH iron sulfur domain 1 (CISDs), has been reported to play a role in ferroptosis [337-340]. The inhibition of CISD1 and CISD2 increased iron-mediated mitochondrial lipid peroxidation and induced erastin and sorafenib-induced ferroptosis, in HCC cells [337-339]. The ablation of CISD3 increased the sensitivity of HT-1080 cancer cells to ferroptosis caused by cysteine deprivation [340]. Overall, mitochondria are a hub of cellular iron metabolism and regulate mitochondrial and cellular iron status to modulate ferroptosis.

In summary, mitochondrial lipid peroxidation can induce ferroptotic cell death, and antioxidants that target mitochondria could mitigate it. Ferroptosis is also dependent on mitochondrial energy pathways, such as the glutaminolysis–TCA cycle–ETC axis, to produce ROS. Furthermore, the MMP has a dual role, as hyperpolarization facilitates ferroptosis and disruption prevents it. In summary, mitochondrial regulation of ferroptotic cell death could be therapeutic targets for diseases associated with ferroptosis.

6.1 mtDNA Association with Pyroptosis

MtDNA, if leaked into the cytosol, is a DAMP and it can activate the NLRP3 inflammasome [348]. Mitochondrial damage produced by LPS results in the leakage of mtDNA into the cytoplasm of the cell [348]. The leaked DNA activates the NLRP3 inflammasome, which activates caspase-1 and gasdermin D [343]. The intracellular cyclic guanosine monophosphate-adenosine monophosphate synthase (cGAS)–STING pathway is the main signaling pathway for DNA sensing, and after the knockdown of the gasdermin D gene, it is activated, indicating the presence of a negative feedback mechanism [349]. Therefore, gasdermin D facilitates the release of mtDNA into the extracellular space to prevent the accumulation of DNA in the cytosol and the inflammatory response caused by the cGAS–STING pathway [349, 350].

6.2 Mitochondria-Derived ROS Impacts Pyroptosis

mtROS can be formed from an overload of calcium ions in the mitochondria, and from mitochondrial dysfunction [351, 352]. Increased mtROS generation can activate the NLRP3 inflammasome [353]. NLRP3 activation results in lysosomal damage, which can further activate the inflammasome [341]. The upstream molecular mechanisms are still not understood but it has been hypothesized that the mitochondrial antiviral signaling protein and the serine–threonine kinase, NIMA-related kinase 7, are involved in pyroptosis [354]. Pyroptosis activation can produce irreversible mitochondrial damage [355]. mtROS can localize around the damaged mitochondria and signal for gasdermin D oxidation [355]. Therefore, gasdermin D is the key ROS-targeting protein during oxidative stress. Gasdermin D also contributes to the fragmentation of the mitochondrial network. In gasdermin D-deficient models, morphological changes were not observed; however, in gasdermin D pyroptosis, there were rounded, rather than filamentous mitochondrial networks [349].

It is believed that mtROS and its relationship to pyroptosis may be linked to a negative feedback regulation between pyroptosis and mitophagy [356]. Caspase-1 activation increases mtROS levels and mitochondria fragmentation [357]. Mitochondrial damage is increased by caspase-1 cleaving of the protein, Parkin, a key regulator of mitophagy [357]. This biotransformation resulted in the inhibition of mitophagy, release of DAMPs, and increased plasma membrane permeabilization [357]. IL-1β secretion in macrophages increased in tandem with the inhibition of mitophagy; however, it remains to be determined if mitophagy interferes with the biotransformation of cytokines of IL-1 family for secretion into specialized secretory vesicles [358]. In contrast, the activation of mitophagy can remove damaged mitochondria, which decreases the magnitude of inflammasome-mediated effects [359].

In summary, mitochondria play an important role in regulating pyroptosis. For example, as a result of mtDNA release, ROS production, and/or interactions with inflammasome components. The release of mtDNA into the cytosol acts as a DAMP that activates the NLRP3 inflammasome and gasdermin D, releasing proinflammatory cytokines, inducing pyroptosis. It has also been shown that mitochondrial dysfunction also influences pyroptosis through a complex feedback loop with mitophagy. Although pyroptosis inhibits mitophagy, increasing inflammation, mitophagy can attenuate pyroptosis by clearing damaged mitochondria and limiting inflammasome activation. These insights provide information about mitochondria as a key mediator of inflammation through pyroptosis, highlighting them as a potential target for inflammatory diseases.

7.1 Mitochondrial Dysfunction in Paraptosis

In paraptosis, early-stage mitochondrial swelling is thought to be an adaptive response that allows for the removal of calcium released by the ER, and the maintenance of mitochondrial function, to postpone cell death [371]. During proteasomal inhibition and an excess of calcium in the mitochondria, vacuoles are formed from the ER and the mitochondria [372, 373]. Due to the swelling and extensive cytoplasmic vacuolization that occurs during paraptosis, inducing paraptosis could be used as a therapeutic strategy to kill cancer cells that are resistant to apoptosis [374]. Ca²⁺ ions released from the ER by ryanodine receptors (RYRs) and inositol IP₃Rs, flow across the OMM, primarily via the VDAC [375]. Ca²⁺ ions enter the intermembranous space and pass through the IMM, primarily via the mitochondrial Ca²⁺ uniporter (MCU) complex, which allows Ca²⁺ flux into the mitochondria, resulting in Ca²⁺ overload [376]. The Na⁺/Ca²⁺ exchanger in mitochondria are also inhibited by Ca²⁺ overload [377, 378]. Subsequently, the MMP is lost, leading to proton accumulation [377, 378]. Mitochondrial and ER swelling occur, which is caused by mitochondrial and ER dilation [225]. Ginger extract induces paraptosis in MDA-MB-231 and A549 cancer cells by producing mitochondrial dysfunction, characterized by a loss of the MMP, decreased ATP production, increased ROS levels and DNA fragmentation, due to the translocation of apoptosis-inducing factor (AIF) from mitochondria to the nucleus [379].

7.2 Mitochondrial Calcium Overload in Paraptosis

Natural and chemically induced paraptosis-mediated vacuolization occurs through the dilation of mitochondria and ER, by producing calcium overload in the mitochondria, leading to cell death [365, 378]. Yoon et al.[369] reported that curcumin (a natural polyphenol isolated from rhizomes of Curcuma longa [380]), induced paraptosis, by inducing the release of Ca²⁺ from ER, in tandem with MCU-mediated mitochondrial Ca²⁺ uptake, leading to mitochondrial Ca²⁺ overload and subsequent dilation of mitochondria and ER in the breast cancer cell lines, MDA-MB-435S and MDA-MB-231 [373, 381]. Yoon et al. later reported that celastrol, a compound present in Tripterygium wilfordii [382, 383], at 1.2, 0.5, and 4 µM, caused the IP₃R-mediated release of Ca²⁺ from the ER and MCU-mediated mitochondrial Ca²⁺ uptake, producing paraptosis in MDA-MB-435S cells [369]. IP₃R and RyR mediated the release of Ca²⁺ from ER and its MCU-mediated mitochondrial Ca²⁺ uptake, contributing to hesperidin (a flavanone glycoside present in many citrus fruits and vegetables [384, 385])-induced paraptosis in human hepatoblastoma, HepG2 cells [386]. Morusin, a prenylated flavonoid isolated from Morus australis root bark, induced paraptosis, at 30 µM, through VDAC-mediated mitochondrial Ca²⁺ overload and MMP (ΔΨM) depletion, in the human epithelial ovarian cancer cells, A2780 and SKOV-3 [387]. Fontana et al. [388] reported that mitochondrial Ca²⁺ overload is required for δ-tocotrienol-mediated paraptosis in the castration resistance prostate cancer cells, PC-3 and DU145. In addition to Ca²⁺ overload, δ-tocotrienol also induced mitochondrial dysfunction in A375 and BLM cell lines, by inhibiting the expression of the OXPHOS complex I, producing an accumulation of ROS, which played a role in inducing paraptosis [389]. Chalcomoracin, a major secondary metabolite found in fungus-infected mulberry leaves, at 6 µM, abolished the MMP and disrupted Ca²⁺ homeostasis, producing paraptosis cell death in MDA-MB-231 and PC-3 cancer cells [390]. The incubation of MCF-7 and MDA-MB-231 cancer cells with 4 µM of withaferin-A, isolated from the root of Withania somnifera Dunal [391], caused paraptosis that was characterized by swollen mitochondria associated with hyperpolarization of the mitochondrial membrane [392]. The incubation of human lung carcinoma, ASTC-a-1 cells, with 70 µM of taxol, an anticancer compound [393], produced the loss of the MMP before the formation of vacuoles and the collapse of the microtubule cytoskeleton [394]. Interestingly, a cyclometalated iridium (Ir) complex-peptide hybrid and CGP37157, an inhibitor of a mitochondrial sodium (Na⁺)/Ca²⁺ exchanger [395], induced paraptosis by mitochondria and ER membrane fusion or by tethering via the MAMs, which allowed the transport of Ca²⁺ transport between ER to mitochondria, producing a decrease in ΔΨM and subsequent paraptosis death in Jurkat cells [396].

In summary, paraptosis occurs when calcium overload and disruption of ion homeostasis are mediated by mitochondria. This results in ER swelling and mitochondrial damage, producing cell death. During paraptosis, cellular features such as vacuolation and mitochondrial dilation occur, which differ from classical apoptosis. The accumulation of calcium in mitochondria can produce changes in membrane potential, which are frequently induced by natural compounds or chemical inducers. In addition to swelling, calcium overload results in increased ROS levels, which drive paraptotic cell death.