Mitochondrial and mitochondrial‐independent pathways of myocardial cell death during ischaemia and reperfusion injury

Abstract Acute myocardial infarction causes lethal injury to cardiomyocytes during both ischaemia and reperfusion (IR). It is important to define the precise mechanisms by which they die in order to develop strategies to protect the heart from IR injury. Necrosis is known to play a major role in myocardial IR injury. There is also evidence for significant myocardial death by other pathways such as apoptosis, although this has been challenged. Mitochondria play a central role in both of these pathways of cell death, as either a causal mechanism is the case of mitochondrial permeability transition leading to necrosis, or as part of the signalling pathway in mitochondrial cytochrome c release and apoptosis. Autophagy may impact this process by removing dysfunctional proteins or even entire mitochondria through a process called mitophagy. More recently, roles for other programmed mechanisms of cell death such as necroptosis and pyroptosis have been described, and inhibitors of these pathways have been shown to be cardioprotective. In this review, we discuss both mitochondrial and mitochondrial‐independent pathways of the major modes of cell death, their role in IR injury and their potential to be targeted as part of a cardioprotective strategy. This article is part of a special Issue entitled ‘Mitochondria as targets of acute cardioprotection’ and emerged as part of the discussions of the European Union (EU)‐CARDIOPROTECTION Cooperation in Science and Technology (COST) Action, CA16225.


| INTRODUC TI ON
Ischaemic heart disease remains a major cause of morbidity and mortality throughout the world, and is responsible for ~20% of deaths in the European Union in both men and women. 1 Many of these deaths occur during an acute ischaemic event such as an ST-elevation myocardial infarction (STEMI). Although fatality rates immediately following acute myocardial infarction have decreased in most countries, 1 infarct size remains a major determinant of outcome and is strongly associated with all-cause mortality and hospitalization for heart failure within the following year. 2 Cardiomyocytes begin to die during exposure to prolonged ischaemia, and while reperfusion is necessary to limit this process, it causes a spike of further cell death that contributes to final infarct size. 3 Thus, finding ways to limit cardiomyocyte death during ischaemia and reperfusion (IR) has been the focus of extensive studies over the past 30 years. 3 Myocardial IR is a complex process during which the ability of physiological processes to return the cardiac cells to homeostasis is overwhelmed. A major cause of this is calcium overload which damages cellular components and drains energy (ATP) as ion pumps in the sarcolemma and sarcoplasmic reticulum (SR) are engaged to return cytosolic calcium back to appropriate levels. 4,5 Mitochondrial calcium overload causes mitochondrial damage and further depletes ATP as it is utilized to maintain mitochondrial membrane potential. In combination with oxidative stress and calcium overload, ATP levels may decrease to a critical level at which the ability of the cardiac cell to remain viable becomes compromised, and the cell undergoes uncontrolled death through a process of oncosis and necrosis, which is described in detail below. However, even before this step, programmed cell death pathways may be activated including apoptosis, necroptosis or pyroptosis. Although ultimately each of these pathways still results in the death of the cell, they can have profoundly different effects on the heart, for example in terms of the activation of an inflammatory response. Furthermore, as many of the pathways appear to overlap or utilize common cellular signalling components, modulation of one pathway may simply result in the cardiomyocyte dying by an alternative pathway. This review aims to provide insight into the different types of cell death which myocardial cells may undergo during IR, with special emphasis on the role of mitochondria in those processes, in order to understand how these processes can be targeted to protect the heart. 6 It is important to note that the initial description and definition of several of the cell death pathways (apoptosis, pyroptosis, etc) were based on experimental observations in leucocytes, and they may have different manifestations in cardiomyocytes or other non-inflammatory cell types in the heart. In this review, we focus on cell death pathways occurring in cardiomyocytes. Clearly, mitochondria are central to the function of cardiomyocytes, occupying nearly 40% of the cytosolic volume, 7 and providing the bulk of the ATP necessary for contraction as well as ion pumps and metabolic processes essential for survival. It is therefore not surprising that mitochondria appear to have a central place in the process of cardiomyocyte death.

| ON COS IS AND NECROS IS
During myocardial ischaemia, oxygen is rapidly depleted, causing mitochondrial respiration to cease. Anaerobic metabolism is activated within seconds of flow cessation, but is unable to provide sufficient ATP for maintaining sarcolemmal ion gradients and mitochondrial membrane potential (ΔΨ m ). ATP is further depleted by the F 0 F 1 ATPase running in reverse, expending ATP in a futile attempt to maintain ΔΨ m . Eventually, the sarcolemmal ion pumps fail and the cell swells in a process called 'oncosis', which is defined as a prelethal stage following cellular injury. 8  dysfunctional proteins or even entire mitochondria through a process called mitophagy. More recently, roles for other programmed mechanisms of cell death such as necroptosis and pyroptosis have been described, and inhibitors of these pathways have been shown to be cardioprotective. In this review, we discuss both mitochondrial and mitochondrial-independent pathways of the major modes of cell death, their role in IR injury and their potential to be targeted as part of a cardioprotective strategy. This article is part of a special Issue entitled 'Mitochondria as targets of acute cardioprotection' and emerged as part of the discussions of the European Union (EU)-CARDIOPROTECTION Cooperation in Science and Technology (COST) Action, CA16225.

K E Y W O R D S
apoptosis, autophagy, cardiac, cell death, ischaemia, myocardial infarction, necroptosis, necrosis, pyroptosis, reperfusion and stain the cell ( Figure 1). However, it is important to note that membrane permeabilization, and hence necrosis, can occur secondarily to cell death by any mechanism, including the later phase of apoptosis. Likewise, it should also be noted that various cell death modalities with the features of ruptured plasma membrane can occur in parallel. These facts can make it difficult to ascertain the F I G U R E 1 The major pathways of cell death that contribute to myocardial ischaemia and reperfusion injury. During initial oncosis, cells swell-this is reversible but can proceed to necrosis. Non-cardiomyocytes can die via a processes of apoptosis or necrosis/necroptosis, in addition to other types of cell death described herein. Cardiomyocytes die primarily via a process of necrosis/necroptosis in addition to other cell-death processes such as pyroptosis, but there is a little evidence for any contribution of apoptosis. Plasma membrane rupture is the terminal event, and this is mediated either by MLKL channels during necroptosis or by GSDMD pores during pyroptosis. Multiple celldeath pathways can eventuate in plasma membrane permeabilization, as detected by dyes such as propodium iodide (resulting in red nuclei as shown) or trypan blue Mitochondria play a critical role in the process of myocardial IR injury. In particular, upon reperfusion, when the supply of oxygen to the cardiac cells is re-introduced and mitochondrial respiration recommences. The mitochondrial substrate succinate, having accumulated during ischaemia, provides a powerful source of electrons which produce oxygen radicals by reverse electron transport via complex I, resulting in oxidative stress (excess reactive oxygen species or ROS). 9 The rapid replenishment of oxygen and ATP to the cells is a double-edged sword. ATP is necessary to restore ionic homeostasis but it also reactivates the sarcoplasmic reticulum ATPASE 2A (SERCA2A) allowing it to pump Ca 2+ back into the SR. However, hyperactivation of the SR Ca 2+ release channel, RyR2, results in rapid cycles of SR Ca 2+ uptake and spontaneous SR Ca 2+ release.
The restoration of ATP levels while cytosolic Ca 2+ overload is still current also leads to hypercontraction of cardiomyocytes, which can be detected during the first few minutes of reperfusion by the appearance of 'contraction band necrosis' in haematoxylin and eosin-stained myocardial histological sections. 10 The excessive SR Ca 2+ release contributes to mitochondrial Ca 2+ uptake via the mitochondrial calcium uniporter (MCU), leading to mitochondrial Ca 2+ overload and opening of the mitochondrial permeability transition pore (MPTP). 11 With the MPTP open, the mitochondria are no longer able to maintain ΔΨ m , and shortly afterwards, ATP stores are depleted, ion pumps cease functioning, and the cells die through a process of oncosis and necrosis.
It may be possible to target necrosis during the early stages of infarction in order to protect the heart from IR injury. Mice lacking either the MPTP or the MCU have smaller infarct sizes following IR. [12][13][14] Cardioprotection against IR injury can also be achieved experimentally by activation of the MAPK/ERK1/2 or PI3K/AKT signalling pathways. 15,16 Activation of the reperfusion injury salvage kinase (RISK) pathway protects the heart by delaying opening of the MPTP. 17 At least in the isolated, perfused heart, blocking hypercontraction by lowering pH or administering inhibitors of the contractile machinery reduces infarct size, as does blocking of reverse electron flow by providing malonate. [18][19][20] However, as will be discussed later, it cannot be ruled out that such previously reported MPTP opening associated with necrosis 17 can be also a mechanism of other necrosis-like cell death modes, which have been identified more recently. 13

| AP OP TOS IS
Apoptosis is a form of cell death that can be distinguished microscopically from oncosis by cell shrinkage, chromatin condensation and distinctive blebbing (budding) of the plasma membrane.
Apoptosis can occur via intrinsic or extrinsic mechanisms but both result in mitochondrial outer membrane permeabilization (MOMP), Believed to be similar to the process in other cell types Note: As some processes can overlap, not all features are necessarily diagnostic of the type of cell death and may depend on the time-point being examined (eg: most forms of cell death will ultimately result in plasma membrane permeability).

TA B L E 1
The key characteristics of the main cell death processes discussed in this review, and their manifestation in cardiomyocytes (if known) mitochondrial cytochrome c release, caspase activation, DNA fragmentation and cell blebbing. 21 Early studies detected evidence of apoptosis along with necrotic cell death following myocardial IR. 22 However, the relative contribution of apoptosis to the extent of cardiac damage is still debated due to the large differences in its magnitude as reported by different investigators. DNA laddering, one of the hallmarks of apoptosis, was not detected in myocardium subjected to ischaemia alone, but was only observed after reperfusion, suggesting that the apoptotic component of cell death in the myocardium is triggered at the time of reperfusion and does not manifest during the ischaemic period. 23,24 In contrast, other studies have shown that apoptosis begins either after prolonged myocardial ischaemia without reperfusion or after a brief period of ischaemia followed by reperfusion. 25,26 Detection of pro-apoptotic factors and caspase activation during ischaemia in the absence of DNA fragmentation followed by a more massive increase during reperfusion indicates that the apoptotic cascade is initiated during ischaemia, but is fully executed during reperfusion. 27,28 More supportive evidence for the acceleration of apoptosis during reperfusion comes from studies showing a reduction in infarct size using inhibitors of pro-apoptotic mediators at early reperfusion. 29,30 Studies in humans have also demonstrated the detection of apoptotic cardiomyocytes in the border zone of the infarcted myocardium within hours to days of infarction. 31 In contrast to the above, other studies have argued against the significant role of apoptosis in IR-induced cell death, based on the fact that there is minimal expression of most proteins required for the apoptotic program in adult cardiomyocytes. 32,33 In addition, using cardiac-specific knockout mice, it was conclusively shown that the executioner caspase-3 and caspase-7 do not significantly contribute to the acute effects of myocardial IR injury. 34 Even forced overexpression of caspase in cardiomyocytes is not able to trigger a full apoptotic response in cardiomyocytes during IR, although it does result in increased infarct sizes. 35 This raises the possibility that the previous observations of apoptosis in the heart are likely to have been due to apoptosis of non-cardiomyocytes. 21  Mitophagy can be considered a beneficial cellular process that enhances cell viability following stressful stimuli by eliminating dysfunctional mitochondria. Mitophagy is therefore essential for cardiomyocyte survival. 52 Accumulated evidence suggests that IR causes an imbalance in the mitophagy process, and one could easily imagine how this would allow dysfunctional mitochondria to accumulate in the cell, causing further cytotoxic damage and potentially leading to cell death. However, it remains controversial whether it is excessive mitophagy or large-scale accumulation of autophagosomes that is the main mechanism underlying 'autophagic cell death'. 40,53,54 The emerging consensus is that cellular insults induce changes consistent with autophagosome formation and the initiation of autophagy in cardiac cells, and that these processes can lead to cell death.
The results of early studies of the role of autophagy in IRinduced cardiomyocyte death were somewhat contradictory, indicating that autophagy could be cyto-protective, but could also direct cells towards apoptosis. 55 These observations are in contrast with the notion that autophagy may trigger cell death in a caspase-independent way as assessed in vitro 56 and in vivo, 57 where it has been shown that impairing the expression of ATG genes leads to reduction of cell death. However, rigorous kinetic analyses are required to establish whether autophagic cell death is independent from apoptotic or necrotic processes and whether it represents a step through which these processes culminate with cell disruption. 53 Evidence from Sadoshima's group suggested that autophagy is beneficial during ischaemia but harmful during reperfusion. 58 However, more recently, the balance of evidence favours a beneficial role for autophagy in the heart under most conditions. 59

| NECROP TOS IS
Necroptosis, a regulated mode of cell death with a necrotic appearance, has been identified in various cardiac pathologies, including myocardial IR (reviewed in 60,61 ). The precise cytotoxic mechanisms of necroptosis are not fully understood; however, the activation of Necroptosis has been seen to occur in both H9c2 cardiomyoblasts subjected to hypoxia/reoxygenation (HR) and in vivo rat hearts subject to IR, as evidenced by increased levels of RIP1, RIP3 and MLKL. 72 An increase was also seen in the mitochondrial membrane Nonetheless, at early stages of SMAC-mimetic induced necroptosis in various cancer cell types, a disruption of mitochondrial membrane potential is observed that presumably depends on BAK-BAX activity. 78 In accordance with this, the pro-apoptotic protein PUMA could act as an amplifier of necroptosis by exposing mitochondrial DNA to the cytosolic sensors, which further stimulates the necrosome formation. 79 Moreover, at the inner mitochondrial membrane, MPTP response could be an important mediator of this kind of cell death, as either cyclosporine A treatment or cyclophilin D deficiency confer resistance to necroptosis prompted by TNFα exposure of endothelial cells. 80 In summary, therefore, while specific mitochondrial mechanisms in the membrane canonical pathway of necroptosis cannot be ruled out, the details remain to be elucidated.

| PYROP TOS IS
Pyroptosis, meaning 'fire' and 'falling', is a pro-inflammatory celldeath program occurring after cytosolic receptor-mediated recognition of pathogen-associated molecular patterns (PAMPs), or host-derived, danger signals such as damage-associated molecular patterns (DAMPs). Some well-characterized DAMPs include glucose-regulated proteins (GRPs), high-mobility group box 1 (HMGB-1), IL-1β, S100 family proteins and some heat shock proteins Of note, bidirectional crosstalk between pyroptotic and apoptotic cell death mechanisms has been described: caspase-1 can cleave and activate caspase-3 and caspase-7 to start apoptosis, which in turn can cleave and inactivate GSDMD to limit pyroptosis. 85 Likewise, a crosstalk between pyroptosis and necroptosis being associated with RIP3 and MLKL activation has also been suggested. 86 The The mechanisms of myocardial IR damage by DAMPs may involve binding to receptor for AGE (advanced glycation end products, RAGE) or to TLR4 (toll-like receptor 4), thus activating NFκB and exacerbating myocardial damage. Indeed, among pro-inflammatory genes, NFκB promotes the transcription of components of the NLRP3 inflammasome in cardiac cells. 95 Therefore, the NLRP3 inflammasome may be considered a sensor that links myocardial damage to inflammation, thereby contributing to the progression of the wavefront of IR injury. 96,97 After activation, the NLRP3 inflammasome promotes cardiomyocyte death and infarct size progression in the first hours of reperfusion likely through pyroptosis and then through production of IL-1β and inflammation. 91,96 Despite the fact that IR-damaged myocardium releases a combination of priming and triggering factors of the NLRP3 inflammasome, it has been proposed that the NLRP3 inflammasome in the heart is not sufficient to respond to a trigger signal in the absence of a priming. 98 Actually, after IR, the size of the infarct is found to increase more in the presence of an active NLRP3, 93 and caspase-7), in a K + efflux-dependent way. Furthermore, the serine threonine kinase NIMA-related kinase 7 (NEK7) is a promoter of NLRP3 inflammasome assembly downstream of ROS and K + efflux. 106,107 In fact, the catalytic domain of NEK7 interacts with the NACHT/LRR domain to favour NLRP3 inflammasome activation. 106,107 Intriguingly, this interaction may be disrupted by cytochrome c, an intrinsic/mitochondrial regulator of apoptosis, to limit pyroptotic cell death, in a sort of yin yang, bidirectional process between apoptosis and pyroptosis. 85,107 Finally, it has also been proposed that mitophagy may dampen NLRP3 activation by removing injured mitochondria. Actually, NLRP3 activity and metabolic disease progression have been associated with low levels of mitochondrial mitofusin proteins and elevated levels of DRP1, especially in hyperglycaemic conditions. 108,109 Yet, deletion of DRP1 may lead to an increase in NLRP3-dependent caspase-1 activation and IL-1 secretion. 110 Recent evidence has shown that autophagy is necessary to reduce myocardial damage after acute myocardial infarction, thus confirming that the autophagic process can limit the activation of the NLRP3 inflammasome by removing damaged mitochondria and that impaired mitophagy may contribute to adverse cardiac remodelling in myocardial infarction. 111,112 Arguing

| OTHER T YPE S OF CELL DE ATH
Although the involvement of mitochondria in extrinsic and intrinsic apoptotic pathways is undeniable, and their role in necroptosis and pyroptosis is being elucidated, the degree to which mitochondria are involved in other, more newly recognized, forms of cell death has just begun to be disentangled. In this respect, 'parthanatos' entails cell death characterized by excessive activation of poly(ADP-ribose) polymerase-1 (PARP-1). 117

| CON CLUS IONS
An important limitation in all investigations of cell death pathways is the methodology used to evaluate cell death in cardiomyocytes (Table 1). TUNEL staining can be misleading as it alone does not distinguish between cells undergoing apoptosis, necrosis or DNA repair. 62 Caspase activity can be measured but is also not specific to apoptosis as caspases can be involved in other pathways of cell death. Cell-based assays for more recently described forms of cell death such as pyroptosis and necroptosis have not yet been developed and many used assays are rather nonspecific being limited to the detection of cell damage due to membrane rupture and impairment of cell metabolic activity. Distinguishing between different forms of cell death is further complicated by the potential overlap between them, and sharing of common signalling components. In this regard, recently published guidelines offer useful advice to the use of different assays of cell death. 62 However, complicating this analysis is the fact that different types of cell can exhibit different characteristics during cell death, and cardiomyocytes may exhibit some particular differences to other cell types. Using observations taken from immune cells to interpret the response of cardiomyocytes to injury may not always lead to the correct interpretation and must be performed with care.
In summary, mitochondria seem to be a convergent point between various regulated cell death processes. In some cases, such as MPTP-mediated necrosis, apoptosis and parthanatos, their participation is clear, whereas in other cases, such as necroptosis, the extent of their involvement might be context-dependent. In either case, increasing evidence points to a crosstalk between the diverse pathways of death, with mitochondria likely to be a central node of all such pathways. They are therefore a key target for maintaining the health of cardiomyocytes during IR and thereby protecting the heart from injury.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

AUTH O R CO NTR I B UTI O N S
Each author drafted and critically revised a section the paper; SD