The Role of Pyroptosis in Ischemic and Reperfusion Injury of the Heart

Abstract

While ischemia itself can kill heart muscle, much of the infarction after a transient period of coronary artery occlusion has been found to result from injury during reperfusion. Here we review the role of inflammation and possible pyroptosis in myocardial reperfusion injury. Current evidence suggests pyroptosis’s contribution to infarction may be considerable. Pyroptosis occurs when inflammasomes activate caspases that in turn cleave off an N-terminal fragment of gasdermin D. This active fragment makes large pores in the cell membrane thus killing the cell. Inhibition of inflammation enhances cardiac tolerance to ischemia and reperfusion injury. Stimulation of the purinergic P2X7 receptor and the β-adrenergic receptor and activation of nuclear factor κ-light-chain-enhancer of activated B cells (NF-κB) by toll-like receptor (TLR) agonists are all known to contribute to ischemia/reperfusion (I/R) cardiac injury through inflammation, potentially by pyroptosis. In contrast, stimulation of the cannabinoid CB2 receptor reduces I/R cardiac injury and inhibits this pathway. MicroRNAs, Akt, the phosphate and tension homology deleted on chromosome 10 protein (PTEN), pyruvate dehydrogenase and sirtuin-1 reportedly modulate inflammation in cardiomyocytes during I/R. Cryopyrin and caspase-1/4 inhibitors are reported to increase cardiac tolerance to ischemic and reperfusion cardiac injury, presumably by suppressing inflammasome-dependent inflammation. The ambiguity surrounding the role of pyroptosis in reperfusion injury arises because caspase-1 also activates cytotoxic interleukins and proteolytically degrades a surprisingly large number of cytosolic enzymes in addition to activating gasdermin D.

Introduction

Acute myocardial infarction (AMI) is one of the main causes of death and disability worldwide. ¹ In patients suffering ST-elevation myocardial infarction (STEMI), minimizing ischemic time by urgent revascularization, either by percutaneous coronary intervention (PCI) or thrombolytic therapy, has been a successful strategy to reduce mortality. ¹ Restoration of coronary flow is an absolute requirement for survival of any cells in the ischemic zone. Surprisingly restoration of flow kills some myocardial cells by a process known as reperfusion injury. A number of interventions including ischemic postconditioning have been shown in animal models to reduce infarct size if applied during the first minutes of reperfusion. ^2,3 This observation provides clear evidence of a reperfusion injury. This was an exciting finding since ischemic or pharmacologic postconditioning should easily be translatable to the clinical arena. However, most trials of ischemic or pharmacological postconditioning in STEMI patients have failed to provide any clinical benefit despite their validation in animal models. ^{4 -7} It was reported that the RISK pathway (eNOS, NO, sGC, cGMP, PKG, PKC) and the SAFE pathway (PI3 K, PIP3, PDK, Akt, ERK, p70s6 K, GSK3β) are involved in the cardioprotective effect of ischemic or pharmacological postconditioning in animals. ⁶ Where RISK are reperfusion injury salvage kinases, eNOS is endothelial NO-synthase, NO, nitric oxide, sGC is soluble guanylyl cyclase, cGMP is cyclic guanosine monophosphate, PKG is protein kinase G, PKC is protein kinase C. Where SAFE is survival activating factor enhancement, PI3 K is phosphatidylinositol (4,5)-bisphosphate 3-kinase, PIP3 (phosphatidylinositol-3-inositole), PDK is PIP3-dependent kinase, Akt is Akt kinase (protein kinase B), ERK is extracellular regulated kinase, p70s6 K is 70-kDA ribosomal protein s6 kinase, GSK3β is glycogen synthase kinase 3β.

There have been a number of reasons proposed for this failure of clinical benefits including interference from co-morbidities in these patients, ⁸ species differences, and, most interestingly, the possibility that these patients had already been pharmacologically conditioned by one or more of the many drugs administered during the reperfusion procedure. High on this list are the loading doses of P2Y₁₂ platelet inhibitors almost universally given to these patients prior to PCI. ⁴ These agents have a strong anti-infarct effect if present during reperfusion of ischemic rabbit, rat, and monkey hearts and protect by the same signal transduction pathway as ischemic postconditioning. ^9,10 Hence, not surprisingly, adding ischemic postconditioning to either rabbits pre-treated with cangrelor, a potent P2Y₁₂ inhibitor, ¹¹ or rats pre-treated with ticagrelor, a widely prescribed P2Y₁₂ antagonist, ¹² offered no additional protection. However, inhibition of caspase-1 starting just prior to reperfusion in ticagrelor-treated rats caused a dramatic further reduction of infarct size beyond that achieved with the P2Y₁₂ inhibitor alone indicating another component of reperfusion injury exists independent of the conditioning pathway. ¹² Thus, reperfusion injury appears to involve multiple mechanisms which, if individually suppressed, can augment myocardial salvage. Caspase-1 dependent infarction appears to be one of these processes.

Pyroptosis is one of the likely manifestations of the inflammatory response to ischemia/reperfusion (I/R) of the heart. ^{13 -15} Evidence also suggests that inflammation and possibly pyroptosis play important roles in the ongoing progression of adverse post-infarction cardiac remodeling.

The Discovery of Pyroptosis and Its Mechanism

In 2001, Cookson and Brennan first coined the term “pyroptosis” from the Greek roots pyro, relating to fire, and ptosis, denoting the fall, to describe proinflammatory programmed cell death. ¹⁶ Hypothesis of the existence of pyroptosis was based on experimental data that infection per se is not sufficient to trigger macrophage death, but rather Salmonella-induced death is a programmed event. ¹⁶ Fink and Cookson found that Salmonella enterica pyroptosis kills macrophages by a caspase-1-dependent mechanism. ¹⁷ They found that pyroptosis causes DNA cleavage and fragmentation as a result of caspase-1-stimulated nuclease activity. Pyroptosis evokes formation of membrane pores in macrophages. Fink and Cookson claimed effective pore diameter to be 1.2-2.4 nm based on transit of differently sized molecules, ¹⁷ while Ding et al assessed pore diameter to be 10-14 nm based on visualization with electron microscopy. ¹⁸ These pores allow sodium ions to enter the cell resulting in osmotic swelling and subsequent overt rupture of the cell membrane. That rupture allows lactate dehydrogenase (LDH), which is too large to pass through the pores, to then be released. ¹⁷ Caspase-1 converts pro-interleukin-1β to interleukin-1β (IL-1β) and pro-interlekin-18 to interleukin-18 (IL-18). ¹⁹ Both ILs are pro-inflammatory cytokines. These cytokines were found to be released through pyroptosis from macrophages infected by S. typhimurium. ²⁰

Pyroptosis, as opposed to apoptosis, induces rapid, irreversible membrane failure. However, pyroptotic cell death does bear some similarities to apoptosis. Pyroptotic cells display DNA damage and are positive in terminal deoxynucleotidyl transferase-mediated UTP nick end-labeling (TUNEL) and annexin-5 assays. ^21,22 Chromatin condensation also occurs in pyroptosis, although in apoptosis the nucleus remains intact. ²¹ However, DNA damage does not play a key role in pyroptosis, as inhibition of DNA fragmentation with the endonuclease inhibitor aurintricarboxylic acid does not prevent cell lysis. ¹⁷

Pyroptosis in reperfused hearts is thought to be initiated by Nucleotide-binding oligomerization domain (NOD)-Like Receptor with a Pyrin domain 3, pyrin domain-containing protein 3 (NLRP3 cryopyrin) which is a molecular sensor that detects microbial motifs and endogenous danger signals trigger the formation and activation of the NLRP3 inflammasome. ²³ In 2002, the term “inflammasome” was first coined by Martinon et al. ¹³ They labeled it as “a caspase-activating complex” which consists of procaspase-1, procaspase-5, apoptosis speck-like protein (ASC), and pyrin domain-containing protein. They hypothesized that the inflammasome catalyzes the proteolytic conversion of procaspase-1 to caspase-1 and procaspase-5 to caspase-5. ¹³ The NLRP3 inflammasome is a cytoplasmic, multiprotein complex which consists of NLRP3, ASC and pro-caspase-1. ^24,25

Inflammasome assembly begins when danger-associated molecular patterns (DAMPs) or pathogen-associated molecular patterns (PAMPs) activate pattern recognition receptors (PRRs) found on Toll-like receptors (TLR) and NOD-like receptors like NLRP3. ²⁶ TLR4 recognizes lipopolysaccharide (LPS), while TLR9 recognizes certain bacterial DNAs which are similar to mitochondrial DNA (mtDNA). ²⁶ TLRs activate nuclear factor κB (NF-κB) which induces NLRP3 and pro-interleukin-1β and -18 expression. ²⁴ This is often referred to as signal 1. Signal 2 occurs when stimuli such as K⁺ efflux, lysosomal destabilization, extacellular ATP, cathepsin B release, calpain activity, increased intracellular Ca²⁺, oxidized mtDNA, or reactive oxygen species (ROS) cause NLRP3 to interact with ASC and procaspase-1 to form the inflammasome. Oligomerization causes proteolytic release of activated caspase-1 which proteolytically activates the pro-interleukins produced by signal 1. ^24,25,27 It was demonstrated that α7 nicotinic acetylcholine receptor stimulation inhibits the NLRP3 inflammasome activation in macrophages because it prevents the release of mtDNA, an NLRP3 ligand. ²⁸ During I/R of the heart reactive oxygen species (ROS) stimulate tissue inflammation and induce the activation of the NLRP3 inflammasome. ²⁹ There is evidence that damaged mitochondria may be the source of these ROS. ³⁰

Gasdermin D Is the End-Effector of Pyroptosis

In 2015, it was discovered that gasdermin D is a crucial component of pyroptosis. Deletion of the gene encoding gasdermin D prevented pyroptosis and IL-1β secretion in J774 macrophage cells ³¹ as well as in murine bone marrow macrophages. ³² Gasdermin D is proteolytically cleaved by caspase-1. Consequently, pyroptosis is defined as gasdermin-mediated programmed necrosis. Caspase-11 in mouse and its human homologs caspase-4 and 5 also can cleave gasdermin D ^15,32,33 and these caspases can mediate a non-canonical pyroptosis pathway. Caspases cleave gasdermin D at the linker between its N-terminal (gasdermin-N) and its C-terminal (gasdermin-C) domains which eliminates the autoinhibitory interactions between these 2 domains. The released gasdermin-N binds to phosphoinositides in the inner plasma membrane and oligomerizes to form membrane pores. ³³ To prove that pyroptosis is involved in any injury would require a demonstration that gasdermin D played a key role. Although there is evidence that pyroptosis is involved in many types of cell injury and, as detailed below, the inflammasome mediates cardiac I/R injury, definitive proof that pyroptosis is the cause of I/R injury must await demonstration of gasdermin’s involvement.

Isolated Cardiomyocytes

H9C2 cells, a cell line derived from embryonic rat cardiac muscle, exposed to 6 h of hypoxia and 18 h of reoxygenation led to elevated IL-1β and IL-18 concentrations in the culture medium (presumably the result of pyroptosis). ³⁴ There also was an increase in caspase-1 and caspase-11 expression in the H9C2 cells. In a similar study oxygen-glucose deprivation (OGD) induced LDH release from H9C2 cells. ³⁵ An increase in gasdermin D transcription and importantly its activation was seen. But, while gasdermin D activity was correlated with the LDH loss, causality was not proven. Also, it must be remembered that H9C2 are not true cardiac muscle cells. ³⁶ Therefore, they may not necessarily share the same pathways. In another study the NLRP3 inflammasome inhibitor BAY11-7082 increased tolerance of H9C2 cells to hypoxia/reoxygenation (H/R). ³⁷ The ability of H/R to induce pyroptosis-like changes in isolated cardiomyocytes is supported by many other studies. ^{34,38 -42} The above findings demonstrate that H/R can induce inflammation and exhibit pyroptosis-like LDH release in isolated cardiomyocyte models.

The Isolated Heart

Mastrocola et al studied isolated rat hearts subjected to I/R. ⁴³ At the end of reperfusion NLRP3 and caspase-1were increased 3-fold, and IL-1β content was elevated 2-fold in the myocardium. Pretreatment with the NLRP3 inflammasome inhibitor INF4E 20 min before index ischemia reduced infarct size and improved contractility at reperfusion in isolated perfused rat hearts. In untreated hearts gasdermin-D content was reduced during reperfusion (60 min) by about 30%, while the N-gasdermin-D level was elevated by about 6-fold. However, the myocardial N-gasdermin D level was not elevated after only 20 min of reperfusion.

Mastrocola et al also found that pretreatment with INF4E increased phosphorylation of extracellular receptor kinase (ERK), Akt, and glycogen synthase kinase-3β (GSK-3β) ⁴³ suggesting a similar mechanism of protection as is seen in ischemic preconditioning. ⁹ This study supports the central role of pyroptosis and reveals that all pyroptosis-associated components including gasdermin D reside in the myocardium. It should be noted that 24 h of reperfusion was needed to see protection from 16673-34-0-treated in situ mouse hearts, ^44,45 but Mastrocola et al saw a significantly reduced infarct size after only 1 h of reperfusion in the isolated rat heart treated with INF4E. ⁴³ These latter observations also argue against Kawaguchi et al’s hypothesis that inflammasomes in leukocytes were the ultimate executioners of ischemic myocardium ^46,47 (see below).

In other experiments isolated rat hearts were subjected to I/R. ¹² The highly selective caspase-1 inhibitor VRT-043198 at reperfusion decreased LDH washout during reperfusion, improved cardiac contractility, and reduced infarct size by approximately 50%. Figure 1 shows that most of the caspase-dependent reperfusion injury occurs in the first minutes of reperfusion as the plots of LDH released into the perfusate in control and VRT-043198-treated rats began to significantly diverge within 3 min of the onset of reperfusion. Such a rapid loss of membrane integrity upon reperfusion certainly argues for a role of pyroptosis, but other processes could also explain it such as cell swelling and membrane rupture caused by sodium leak during ischemia and sodium exchange with H⁺ early in reperfusion.

Figure 1.

Release of LDH from an isolated rat heart following 40 min of global ischemia. The lower trace is from hearts in which VRT-043198 was present in the reperfusate. The curves start low and rise in the first minutes of reperfusion indicating little LDH was present in the heart prior to reperfusion. Much of the LDH released appears to be caspase-1/4 dependent as evidenced by the difference between the curves. This rapid caspase-1/4-dependent membrane failure could well be caused by pyroptosis. Modified from Audia et al. ¹²

The same group subjected isolated mouse hearts to I/R. ⁴⁸ VRT-043198 added to the perfusate at reperfusion reduced infarct size by 51%. When the VRT-043198 infusion was delayed by only 10 min, protection from the drug was no longer seen. That indicated that all caspase 1/4-dependent cell killing occurred in the first 10 min of reperfusion. It was also found that emricasan, an irreversible pan-caspase inhibitor, was just as protective as VRT-043198 which is highly selective for caspases 1 and 4 suggesting that VRT-043198 was actually targeting a caspase and that no other caspases, such as those associated with apoptosis, appear to contribute to acute infarction.

Similar infarct size data were obtained by other investigators using the isolated rat heart in the presence of VX-765. Do Carmo and colleagues ⁴⁹ included VX-765 in the buffer starting before ischemia and continued it throughout the experiment. A phosphatidylinositol-3 (PI3)-kinase inhibitor abolished protection which is concordant with the findings of Mastrocola and colleagues. ⁴³ VX-765 is not protective in isolated rat hearts when given only at reperfusion, presumably because the buffer does not have the required esterases to quickly convert it to VRT-043198 which does protect isolated rat hearts when administered at reperfusion. However, if VX-765 is present in the heart during ischemia, there is apparently enough esterase in the heart tissue to activate a sufficient amount of the drug to protect the heart at reperfusion. ¹²

After isolated murine hearts were subjected to I/R, Darwesh et al ⁵⁰ documented that I/R increased the NLRP3 content 4-fold, IL-1β content increased 2.5-fold, and caspase-1 activity increased 4-fold. Because these hearts were blood-free, these pyroptosis-associated markers could not have been introduced by invading inflammatory leukocytes. These data again prove that pryroptosis-associated components do exist in heart tissue and that ischemia increases their expression.

However, it must be remembered that caspase-1 does not only trigger pyroptosis. It also activates cytotoxic interleukins ⁵¹ and it can proteolytically degrade a surprisingly large number of cytosolic enzymes. ⁵² Audia and colleagues ¹² reported that the ventricular content of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and aldolase-A, both caspase-1 targets, were reduced by more than 60% in the ventricle after I/R. VX-765 at reperfusion greatly reduced the loss of both enzymes. Caspase-1 does not target hexokinase and VX-765 had no effect on the hexokinase content in those tissue samples. Observations of changes in GAPDH are significant because GAPDH is reported to be a metabolic choke point during ischemia in the isolated rat heart. ⁵³ How much each of these potentially lethal targets of caspase-1 including release of cytotoxic interleukins, degradation of critical cytosolic enzymes, and initiation of pyroptosis contributes to reperfusion injury is currently unknown.

The popular hypothesis is that cardiomyocytes are being killed by direct membrane failure related to N-gasdermin Dinduced pyroptosis. But it may not be that simple. Pyroptosis may indirectly kill cardiomyocytes. Pyroptosis may just occur in certain cells in the heart such as resident macrophages, mast cells or fibroblasts which allows them to secrete their toxic interleukins that then attack the cardiomyocytes. In such a scenario the infarction would still appear to be gasdermin D-dependent even though gasdermin D acted as a trigger rather than the end-effector. Blockade of IL-1 receptors has been reported to protect ischemic heart muscle. ^51,54 Alternatively we may find that knocking out gasdermin D would offer no protection against I/R and that caspase-1 activity may be injurious to cardiac tissue independent of gasdermin D.

Myocardial Ischemia and NLRP3 Inflammasome

It was demonstrated that the caspase-1, ASC, and cryopyrin levels are increased in cardiomyocytes, fibroblasts, endothelial cells, and leukocytes in the border zone between infarct and surviving myocardium of mice 3 days after permanent coronary artery occlusion (CAO). ⁵⁵ Increased expression of these markers of inflammation in the border zone persisted for at least 7 days after CAO. These markers were not found in the non-ischemic zone of the heart. Whether these components were in cardiac tissue or invading leukocytes was not determined.

One of the earliest studies of the role of the innate immune system in inflammation was a study by Pomerantz using a human atrial muscle segment. ⁵¹ They found that loss of contractile force after H/R could be attenuated by blocking interleukin-1 receptors or caspase-1. This study set the stage for investigation of inflammation in myocardial ischemia.

Subsequently Kawaguchi and colleagues inhibited NLRP3 inflammasome formation by knocking out ASC. ⁴⁶ In the absence of ASC, NLRP3 inflammasomes would be unable to form and activate caspase-1. Infarct size in murine hearts subjected to 30 min of regional CAO and 2 days of reperfusion was smaller in ASC^-/- than in wild type mice. The same was true of caspase-1^-/- mice. They concluded that inflammasomes were contributing to infarction by activating caspase-1. They investigated cultured neonatal cardiomyocytes and fibroblasts. When challenged with LPS, only fibroblasts expressed pro-IL-1β. Wild type fibroblasts showed robust conversion of pro-IL-1β to mature IL-1β, while ASC^-/- fibroblasts had attenuated conversion. The authors concluded that only fibroblasts were capable of forming inflammasomes. Finally they knocked out ASC in either just the bone marrow or in all tissue except the bone marrow in mice to determine the contribution of leukocytes. To their surprise infarct size was smaller in both models. Kawaguchi et al concluded that interleukins released from ischemic fibroblasts signaled leukocytes to migrate to the heart, and the invading leukocytes then killed the cardiomyocytes. Interestingly treatment with IL-1 neutralizing antibody (canakinumab) in AMI patients attenuates the development of post-infarction cardiac remodeling and heart failure which supports the role of interleukins. ⁵⁶ IL-1β and IL-18 both bind the IL-1 receptor.

In a study several years later by Marchetti and colleagues, mice underwent CAO and 24 h of reperfusion. ⁵⁷ The NLRP3 inflammasome formation inhibitor 16673-34-0 was administered before I/R. Pretreatment with 16673-34-0 promoted a reduction in infarct size by 40% and inhibited caspase-1 activity in the heart by about 90%. While the cellular location of the caspase-1 was unknown, the fact that inhibition of NLRP3 formation was associated with a reduction of infarct size was strong evidence that inflammation involving the NLRP3 inflammasome was involved.

In a follow-up study mice were subjected to CAO and 24 h of reperfusion. ⁴⁴ When 16673-34-0 was administered at reperfusion infarct size was decreased by 56%. Continuous treatment with the NLRP3 inhibitor improved cardiac contractility measured 7 days after reperfusion. When CAO was permanent, as expected, 16673-34-0 had no effect on infarct size, but cardiac dysfunction was still limited suggesting a role in post-infarction remodeling to account for improved left ventricular function. These findings support NLRP3 inflammasome’s involvement in I/R cardiac injury.

In subsequent experiments mice again experienced CAO followed by reperfusion. ⁴⁵ These investigators found a time-dependent increase in infarct size in untreated mice when measured after 1, 3, or 24 h of reperfusion (11%, 30% and 43%). Giving the NLRP3 inhibitor at reperfusion or 1 h thereafter decreased caspase-1 activity and infarct size measured at 24 h when the infarct was fully established. The inhibitor did not significantly affect infarct size or caspase-1 activity when it was injected 3 h after the onset of reperfusion. These data suggest that NLRP3 inflammasomes kill heart muscle sometime during the first 3 h of reperfusion.

Pigs were subjected to CAO and 7 days of reperfusion. ⁵⁸ The NLRP3-inflammasome inhibitor MCC950 was injected 5 min before reperfusion. Infarct size was evaluated 7 days after reperfusion. MCC950 promoted the infarct size reduction by 17%. Based on the observation of Marchetti et al, ⁴⁴ the rather modest reduction of infarct size might be attributed to use of only a single bolus of the inhibitor at reperfusion. This NLRP3-inflammasome inhibitor may require continuous treatment to be effective. Alternatively increasing the duration of ischemia to 75 min will also cause more ischemic cell death which would be unresponsive to treatment at reperfusion.

In rats subjected to regional I/R the NLRP3 inhibitor BAY11-7082 was administered prior to the onset of reperfusion. ³⁷ BAY11-7082 reduced infarct size by approximately 20%. Importantly, BAY11-7082 also exhibited the same infarct-sparing effect in rats with streptozotocin-induced diabetes. This might be important because many acute coronary syndrome (ACS) patients have diabetes as a co-morbidity which has been proposed to be an impediment to cardioprotective interventions such as ischemic postconditioning. ⁸ Investigators found significant elevations of NLRP3, ASC, caspase-1, and IL-1β levels in untreated ischemic myocardial tissue, while BAY11-7082 decreased these inflammation markers in the myocardium. BAY11-7082’s infarction-limiting effect was modest, perhaps because death from inflammation may have its maximum effect many hours after reperfusion. It should be noted, however, that the progressive increase in infarct size beyond 2 h of reperfusion reported for mice ⁴⁵ has not been described in rat hearts. Audia et al noted that infarct size in untreated rats or those receiving a caspase-1 inhibitor measured after 2 h of reperfusion was no different from that seen after 3 days of reperfusion with either treatment. ¹²

Myocardial Ischemia and Caspase-1

A caspase-1 inhibitor can also be used to prevent I/R cardiac injury. VRT-043198 is a highly selective inhibitor of caspase-1 and 4. ⁵⁹ It has been approved for testing in patients in the form of its pro-drug VX-765 ^60,61 which is rapidly converted to VRT-043198 by esterases in the blood. VX-765 is favored for its high bioavailability when used orally. It should be cautioned, however, that VX-765 may not be activated in blood-free models. For those studies VRT-043198 should be used. Emricasan, unlike VRT-043198, is a potent, irreversible pan-caspase inhibitor (inhibits all caspases including 1/4) and it also has been used in clinical trials. ⁶²

The earliest report we could find that demonstrated a role for caspase-1 in myocardial infarction was that by Pomerantz et al. ⁵¹ In a human atrial muscle strip model exposed to H/R inhibition of caspase-1 with the peptide YVAD improved post-ischemic contractility. Kawaguchi and colleagues reported that infarct sizes were smaller in caspase-1 deficient mouse hearts after in situ I/R. ⁴⁶ Audia and colleagues administered VX-765 (32 mg/kg) to open-chest, anesthetized rats 5 min prior to the onset of 2 h of reperfusion following 60 min of CAO. ¹² VX-765 reduced infarct size from 60% of the risk zone in vehicle-treated rats to only 29%. VX-765 also decreased serum IL-1β. Infarct size was reduced in rats treated with the P2Y₁₂ inhibitor ticagrelor given intraperitoneally 10 min prior to reperfusion (44% infarction). Addition of ischemic postconditioning to ticagrelor offered no further protection. However, adding VX-765 to ticagrelor reduced infarct size to 18%, nearly half of that seen with either VX-765 or ticagrelor alone. P2Y₁₂ inhibitors are known to protect from I/R injury by a signal transduction pathway similar to that used by postconditioning. ⁹ But blocking caspase-1 apparently protects by a different mechanism, likely by preventing pyroptosis. Audia et al, ¹² as discussed above, found no evidence of either delayed infarction or protection from VX-765 beyond 2 h of reperfusion and a single bolus of VX-765 given just prior to reperfusion was enough to produce sustained salvage observable 3 days later indicating that it had protected against an injury occurring soon after reperfusion with no additional infarction occurring after the drug was washed out. Pretreating rats with VX-765 prior to I/R did not result in infarct sizes any smaller than those in rats treated with VX-765 at reperfusion indicating that all caspase-1 dependent killing occurs at reperfusion in the open-chest rat model. ⁶³

These observations provide overwhelming evidence that the innate immune system contributes to reperfusion injury in the heart by involvement of inflammasomes and caspase activation, and its manipulation likely has impressive clinical potential. All results support pyroptosis being one of the end-effectors of reperfusion injury. However, as of this writing, all we can find are studies merely correlating activation of components in the pyroptosis pathway including gasdermin D. But correlation does not prove cause and effect. What is lacking is a clear demonstration that myocardial infarction is actually dependent on the presence of N-gasdermin D.

The ultimate infarct size is achieved soon after reperfusion. The timing varies from as early as the first 2 h of reperfusion ⁴⁸ to as late as 24 h. ⁴⁵ Thus detection of pyroptosis-related markers beyond the time when the ultimate infarct size is reached would have nothing to do with pyroptosis killing cardiomyocytes. The infarcted heart soon becomes flooded with leukocytes, and, of course, these cells have a high inflammasome content. ⁵⁵ In some models exposure of reperfused myocardium to leukocyte migration may continue to aid in beneficial scar formation. Hence only sampling of pyroptosis-related markers prior to establishment of the ultimate infarct size can be used to support evidence of possible involvement of the markers in the infarction. Invading leukocytes actually have a beneficial role in the healing process following myocardial infarction and their presence does not always indicate injury. Isolated, blood-free hearts also have inflammation component to their reperfusion injury, and that model is not complicated by invading leukocytes making it a useful model for biochemical study, although there are likely some resident macrophages. ⁴⁷

Triggers of Inflammation in the Heart

Ischemic preconditioning (IPC) in isolated mouse hearts improved the recovery of cardiac contractility and reduced LDH release during reperfusion following an ischemic insult in wild type and ASC^-/- mice. ⁶⁴ However, IPC did not exhibit the same cardioprotective effect in NLRP3^-/- mice. It was found that the hearts of NLRP3^-/- mice had decreased IL-6 content. An antibody against IL-6 abrogated the cardioprotective effect of IPC in wild type hearts. Decreased signal transducer and activator of transcription 3 (STAT3) expression was also found in NLRP3^-/- hearts. It is well known that STAT3 is involved in IPC. ⁶⁵ Zuurbier and colleagues concluded that the NLRP3-inflammasome may be an integral component of IPC possibly through an IL-6/STAT3-dependent mechanism. ⁶⁴

Calpain may be an important stimulus of inflammation. Isolated murine cardiomyocytes were subjected to H/R. ²⁷ NLRP3 small interfering RNA (siRNA) or the non-selective caspase-1 inhibitor peptide zVAD increased the survival rate of cardiomyocytes. Calpain siRNA also increased the survival rate of cardiomyocytes. In addition, suppressing calpain with calpain siRNA inhibited the increase in NLRP3, ASC, and caspase-1 content in cardiomyocytes after H/R. Thus, it would appear that calpain can actually stimulate the formation of the NLRP3/ASC/caspase-1 complex during H/R of cardiomyocytes. Zhang et al noted that calpain proteolytically releases procaspase-1 bound to the cellular cytoskeleton of macrophages by a protein called flightless. ⁶⁶ The liberated procaspase-1 was then able to join NLRP3 and ASC to form the NLRP3 inflammasome. The same may occur in cardiomyocytes.

Calpain’s involvement could explain why caspase-1-dependent injury only occurs at reperfusion. Calcium enters the reperfused cardiomyocytes upon reperfusion and activates calpain at that time. Calpain inhibition has been reported to be cardioprotective over the years, but its mechanism has never been established. The calpain inhibitor calpeptin administered to isolated mouse hearts exposed to I/R was as protective as the caspase inhibitor emricasan and combining the two produced no additional protection suggesting a common mechanism. ⁴⁸

ATP is a well known DAMP that stimulates inflammation by activation of its purinergic P2X7 receptor. Activation of that receptor was found to be an important trigger of inflammasome formation in the border zone of the murine heart after permanent CAO. Both P2X7 siRNA and the purinergic P2X7 receptor inhibitor PPADS limited inflammasome formation in that model and reduced infarct size. ⁵⁵ Stimulation of endothelial nitric oxide synthase (eNOS) is known to be cardioprotective. The NO synthase activator serelaxin decreases infarct size in mice experiencing CAO (30 min) and reperfusion (24 h) and that protection is lost in eNOS^-/- mice. Interestingly caspase-1 activity was diminished in the serelaxin-treated mice. ⁶⁷ Whether eNOS protected by directly blocking inflammation components or eNOS-protected hearts simply by release of less proinflammatory DAMPs was not determined.

Nuclear factor κ-light-chain-enhancer of activated B cells (NF-κB) plays an important role in the activation of pyroptosis. Inhibition of NF-κB with pyrrolidine dithiocarbamate (PDTC) suppressed death of H9C2 cells induced by oxygen-glucose deprivation (OGD). ³⁵ N-acetylcysteine, a non-selective ROS scavenger, also inhibited OGD-induced cell death. ³⁵ A high dose of the β-adrenergic receptor agonist isoproterenol has long been known to induce cardiac injury. Intravenous injection of isoproterenol at a dose of only 2 µg/kg is sufficient to trigger the chronotropic response to isoproterenol. ⁶⁸ Isoproterenol at a dose of 5 mg/kg induced an increase in IL-1β, IL-18, and caspase-1 levels in the myocardium along with infarction. ⁶⁹

In mice with permanent CAO the selective cannabinoid receptor-2 (CB2) agonist JWH-133 (10 mg/kg) decreased infarct size by about 33% and improved cardiac contractility 6 h after myocardial infarction. ⁷⁰ The cardioprotective effect was associated with a reduction in the serum IL-1β and IL-18 levels and a decrease in IL-1β, IL-18, NLRP3 and caspase-1 expression in the myocardium. The investigators reported that the selective CB2 receptor antagonist AM630 abolished the cardioprotective effect of JWH-133. The correlation between CB2 activation and inflammasome activity suggests that the cardioprotection might have involved suppression of inflammation.

Protein kinases may also be involved in the regulation of pyroptosis. The phosphate and tension homology deleted on chromosome 10 protein (PTEN) is a phosphatase which dephosphorylates phosphatidylinositol-3,4,5-trisphosphate, a potent activator of 3-phosphoinositide-dependent kinase (PDK) and Akt kinase. ⁷¹ After PTEN deletion cardiac tolerance to I/R was enhanced and the NLRP3, caspase-1, and IL-1β protein levels in the myocardium were decreased. ⁷² This correlation is indirect evidence of involvement of Akt and PTEN in the regulation of pyroptosis of cardiomyocytes during I/R.

Oxidized mitochondrial DNA is a known DAMP that can activate the NLRP3 inflammasome. ⁷³ Yang and colleagues observed that ENDO III, a fusion molecule of Endonuclease III (a DNA glycosylase/AP lyase) and molecular moieties that traffic the enzyme to the mitochondria, reduced myocardial infarct size when administered at reperfusion in rats. ⁷⁴ ENDO III would act to remove oxidized bases from DNA in the mitochondria. It was proposed that it prevents release of oxidized mtDNA into the cytosol which would otherwise stimulate NLRP3 inflammasome formation in the ischemic zone. ENDO III was as potent as a caspase-1 inhibitor in limiting infarct size and, like the caspase-1 inhibitor VX-765, its protection against infarction could be added to that from a P2Y₁₂ inhibitor suggesting that VX-765 and ENDO III might protect by similar mechanisms. ⁷⁴

Many of these interventions are summarized in Table 1.

Table 1.

Effect of Different Compounds on Tolerance of Hearts to Ischemia/Reperfusion or Cardiomyocytes to Hypoxia/Reoxygenation and Pyroptosis-Related Markers.

Compound	Effect	Pyroptosis Markers	Reference
NLRP3 inflammasome formation inhibitor16673-34-0	Before Ischemia Infarct Size ↓	Inhibition	57
16673-34-0	Before Reperfusion Infarct Size ↓	Inhibition	44
16673-34-0	Permanent CAO	No effect	44
NLRP3 inflammasome formation inhibitor	Before Reperfusion Infarct Size ↓	Inhibition	45
NLRP3 inflammasome formation inhibitor MCC950	Before Reperfusion Infarct Size ↓	Inhibition	58
NLRP3 inflammasome formation inhibitor BAY11-7082	Before Reperfusion Infarct Size ↓	Inhibition	37
NLRP3 inflammasome formation inhibitor INF4E	Before Ischemia Infarct Size ↓	Inhibition	43
Caspase-1 inhibitor VX-765	Before Reperfusion Infarct Size ↓	Not evaluated	12
Caspase-1 inhibitor VRT-043198	Before Reperfusion Cardioprotection	Not evaluated	12
Caspase-1 inhibitor VRT-043198	Before Reperfusion Infarct Size ↓	Not evaluated	48
Caspase-1 inhibitor VX-765	Before Ischemia Infarct Size ↓	Not evaluated	49
NLRP3 inflammasome formation inhibitor BAY11-7082	Before Hypoxia Increase in tolerance of H9C2 cells to H/R	Inhibition	35, 37
NLRP3 siRNA	Before Hypoxia Increase in tolerance of Cardiomyocytes to H/R	Inhibition	27
Calpain siRNA	Before Hypoxia Increase in tolerance of Cardiomyocytes to H/R	Inhibition	27
P2X7 siRNA	Before Ischemia Infarct Size ↓	Inhibition	55
CB2 receptor agonist JWH-133	Permanent CAO Infarct Size ↓	Inhibition	70
Endonuclease III	Before Reperfusion	Not evaluated	74

Inflammation in ACS Patients

Information about the involvement of pyroptosis in I/R injury of the human heart is limited. It was noted that the plasma IL-18 level in patients with STEMI was elevated 2.4-fold compared with control subjects. ⁷⁵ A positive correlation was demonstrated between elevated circulating IL-18 and cardiac troponin-I (an index of infarct size). IL-18, like IL-1β, binds to the IL-1 receptor. Elevated plasma IL-18 on admission also appears to predict an adverse 60-day clinical outcome. ⁷⁵ Patients with ACS showed 2-fold higher serum IL-18 titer than healthy control subjects. ⁷⁶ Long-term cardiovascular mortality was correlated with an increase in the serum IL-18 concentration in patients with ACS. ⁷⁷ The plasma IL-18 level was increased 3.9-fold in patients with AMI compared to levels in patients with stable angina pectoris. ⁷⁸ In another study the serum IL-18 level was 2-fold higher in patients with AMI than in healthy volunteers. ⁷⁹ The plasma IL-18 concentration in ACS patients was reported to become elevated after the acute event and remained increased for 6 months. ⁸⁰

Plasma IL-1β was reported to be elevated about 2-fold above normal in patients with AMI. ⁸¹ Increased caspase-1 and IL-1β plasma levels are associated with left ventricular hypertrophy and cardiac remodeling in patients with STEMI and PCI. ⁸² However, in 2015 Wang and colleagues reported a paradoxical result. ⁸³ They found that plasma IL-18 measured 12 h after PCI was 49% lower in patients with STEMI than in healthy volunteers. It remained low for at least 72 h after PCI. The authors did not provide any convincing explanation for these contrary data. Some investigators argue that monocytes are a major source of serum IL-18. ⁷⁹ Therefore, it is unclear whether a decrease in the IL-18 concentration is a result of less pyroptosis of cardiomyocytes or a consequence of less monocyte activation.

Not All Data Support the Role of the NLRP3 Inflammasome in I/R Injury

Some of the results are confusing. Kawaguchi et al ⁴⁶ blocked NLRP3 inflammasome formation by knocking out ASC, and in isolated hearts infarct size was smaller in the ASC^-/- hearts. Sandanger and colleagues also studied isolated, perfused murine hearts. ⁸⁴ An improvement of post-ischemic contractility and a reduced infarct size was seen in the NLRP3^-/- isolated hearts, but, surprisingly, not in ASC^-/- deficient hearts. While these data partially support the hypothesis that NLRP3-induced pyroptosis may be involved in I/R cardiac injury, a follow-up study by Sandanger et al did not. ⁸⁵ They measured infarcts this time in in situ mouse hearts, and, unexpectedly, infarcts were significantly bigger in the NLRP^-/- hearts. Failure to see protection against infarction in the NLRP3^-/- mice does not prove that NLRP3 is not involved since animals lacking the NLRP3 protein from birth may have compensated for the loss with an altered phenotype. Recently, it has been suggested that the negative effect of NLRP3 deficiency may be dependent on TLR4 overexpression in NLRP3^-/- hearts. ⁸⁶

Along these same lines, Jong et al observed that there was no difference in IL-1β levels in the myocardium in the AAR of NLRP3^-/- and wild type mice after CAO and 3 h of reperfusion. ⁸⁷ They then prepared NLRP3^-/- and wild type mice with an exteriorized ligature occluder around the left anterior descending coronary artery. After waiting 10 days to allow the stress of surgery to subside, the ligature was pulled to occlude the coronary artery for 30 min followed by reperfusion for 3 h. Average infarct size was not different between the 2 groups. No NLRP protein was detectable in the AAR myocardium of either knockout or wild type animals. The investigators examined NLRP3 levels in AAR tissue in open-chest animals undergoing the same I/R protocol. As expected, NLRP3 was undetectable in knockout mice. But the protein was present in the wild type mice. The authors concluded that in the absence of surgical trauma NLRP3 protein is not present in the heart and therefore would play no role in the development of myocardial infarction in an ACS patient. There has been no independent corroboration of this hypothesis.

While there has been some controversy concerning the role of NLRP3 in myocardial ischemia, it should be noted that there are other inflammasomes that can activate caspase-1 and could cause pyroptosis. They include NLR family CARD domain-containing protein 4 (NLRC4) and interferon-inducible protein 2 (AIM2). ^21,22 AIM2 is of particular interest since it is found in the heart and is activated by extra-nuclear DNA. ²¹ Injection of DNase I at reperfusion is as protective as a caspase-1 inhibitor. ⁷⁴ As was found with the caspase-1 inhibitor ¹² the cardioprotection afforded by DNase I can also be added to that from a P2Y₁₂ inhibitor, suggesting a possible mechanism of protection similar to that from a caspase-1 inhibitor.

Conclusions and Clinical Perspective

Figure 2 shows a cartoon that summarizes how we think inflammation may be contributing to death of myocardium during I/R based on the reports that we have reviewed. Inflammation arising from the innate immune system within the heart causes ischemic and reperfusion injury of the heart. The leading hypothesis is that this injury is the result of destruction of the cell membrane by pyroptosis. However, pyroptosis is strictly defined as a process of cell death from N-gasdermin D-induced permeabilization of the cell membrane. At the time of this writing we are unable to find published proof that gasdermin D is directly responsible for any ischemia-induced myocardial infarction. While inflammasomes in heart tissue cause reperfusion injury in both the isolated heart and cardiomyocytes, the possibility cannot be excluded that bone marrow-derived leukocytes also contribute to the inflammatory injury of the in situ heart.

Figure 2.

A proposed diagram of NLRP3's role in the ischemic heart. During ischemia ligands for Toll-Like Receptor (TLR) cause expression of NLRP3 proteins and 2 inactive pro-interleukins-1β and -18 (Pro IL-1β and -18). At the same time chemical signals such as oxidized DNA cause NLRP3, ASC and Procaspase-1 to form an inflammasome during which caspase-1 is activated. Caspase-1 proteolytically activates the interleukins. It also cuts off the inhibitory domain (ID) of gasdermin D. Gasdermins N-terminal segment forms lethal pores in the membrane through which interleukins can be secreted, a process called pyroptosis. In some cells pro-caspase-1 is bound to the actin cytoskeleton by a protein called flightless (FL). It is possible that bound procaspase-1 is released by the calcium-activated protease calpain when the ischemic heart is reperfused, thus resulting in absence of assembly of the inflammasome until reperfusion. Definitions not in the text: Nuclear Factor Kappa B (NFkB), inhibitor of NFkB (IkB), myeloid differentiation primary response 88 (MyD88), p50 and p65 are 2 proteins involved in the activation of NFkB (the small p and numbers simply refer to their molecular weights), reactive oxygen species (ROS), cytochrome C (Cyto-C), mitochondrial DNA (mDNA), purinergic type 2 receptor family receptor X7 (P2X7) (a receptor selective for ATP), lipopolysaccharide (LPS). Modified from Mishra et al. ⁸⁸

While some triggers of inflammasome activation have been identified, it is unknown what effect ischemic preconditioning and postconditioning have on inflammation’s component of injury. Some aspects of inflammation are still unknown. For example, what is the role of kinases in inflammation and what are the endogenous regulators of the heart’s vulnerability to inflammation? It is known that adaptation to hypoxia and cold increases cardiac tolerance to I/R. ^89,90 However, it is unknown whether reduced inflammation has a role in the infarct-reducing effect of adaptation.

In a recent clinical trial a quarter of patients with an anterior infarct treated with percutaneous intervention alone suffered death or heart failure in the following year, presumably the result of infarction of too much ventricular muscle. ⁹¹ NLRP3 inhibitors and caspase-1 inhibitors clearly reduce infarct size following I/R. Importantly, in an animal model of PCI patients receiving a P2Y₁₂ inhibitor, blocking caspase-1 at reperfusion caused further reduction of infarct size while ischemic postconditioning had no effect. ¹² The lack of clinical effectiveness of ischemic postconditioning in STEMI patients ^{4 -7} should not be interpreted to preclude the possible protective effect of targeting inflammation. In a recent preclinical study infarct size in rats receiving the same opioid, heparin, and P2Y₁₂ inhibitor treatment that AMI patients currently receive could not be reduced with added remote conditioning but could be with the addition of the caspase inhibitor emricasan. ⁹² These observations all suggest that blocking inflammation is likely to be an effective way to further salvage ischemic heart muscle in today’s AMI patients.

Additionally inflammation appears to be involved in post-infarction cardiac remodeling. ^93,94 It has been demonstrated that M1 macrophages promote cardiac fibrosis by inducing inflammasome formation. ⁹⁵ Inhibition of pyroptosis effectively protected against post-infarction cardiac remodeling in diabetic models as well. ⁹⁶ Since inhibition of the inflammasome pathway has the potential to block early reperfusion injury and late post-infarction remodeling, the likelihood that this approach would have significant salutary clinical benefits is high.

Our manuscript points out that while there is much correlative evidence of pyroptosis in myocardial ischemia/reperfusion injury, direct evidence of the involvement of pyroptosis in infarction has been lacking. Since acceptance of our manuscript, a paper accepted by Circulation Research offers at least some of the required direct evidence. ⁹⁷ The investigators genetically deleted gasdermin D, the primary effector of pyroptosis, exclusively in cardiac muscle cells. Those mice had a reduction in infarct size following an ischemia/reperfusion insult by an amount similar to that seen with a caspase-1 inhibitor. ¹² That would indicate that pyroptosis was responsible for all of the caspase-1-dependent infarction.