Extracellular signalling molecules in the ischaemic/reperfused heart – druggable and translatable for cardioprotection?

Abstract

In patients with acute myocardial infarction, timely reperfusion is essential to limit infarct size. However, reperfusion also adds to myocardial injury. Brief episodes of ischaemia/reperfusion in the myocardium or on organ remote from the heart, before or shortly after sustained myocardial ischaemia effectively reduce infarct size, provided there is eventual reperfusion. Such conditioning phenomena have been established in many experimental studies and also translated to humans. The underlying signal transduction, that is the molecular identity of triggers, mediators and effectors, is not clear yet in detail, but several extracellular signalling molecules, such as adenosine, bradykinin and opioids, have been identified to contribute to cardioprotection by conditioning manoeuvres. Several trials have attempted the translation of cardioprotection by such autacoids into a clinical scenario of myocardial ischaemia and reperfusion. Adenosine and its selective agonists reduced infarct size in a few studies, but this benefit was not translated into improved clinical outcome. All studies with bradykinin or drugs which increase bradykinin's bioavailability reported reduced infarct size and some of them also improved clinical outcome. Synthetic opioid agonists did not result in a robust infarct size reduction, but this failure of translation may relate to the cardioprotective properties of the underlying anaesthesia per se or of the comparator drugs. The translation of findings in healthy, young animals with acute coronary occlusion/reperfusion to patients of older age, with a variety of co-morbidities and co-medications, suffering from different scenarios of myocardial ischaemia/reperfusion remains a challenge.

Tables of Links

TARGETS
CGRPsa	Enzymesd
β-adrenoceptor	ACE
δ opioid receptor	Adenosine deaminase
κ opioid receptor	Adenosine kinase
μ opioid receptor	Adenylate cyclase
A1 receptor	Akt (PKB)
A2A receptor	Aminopeptidase M
A2B receptor	COX
A3 receptor	eNOS
AT1 receptor	ERK
AT2 receptor	GSK3β
B1 receptor	JAK
B2 receptor	PI3K
GLP-1 receptor	PKA
Ion channelsb	PKC
KATP channel	PKG
Catalytic receptorsc	sGC
gp 130 (IL-6β) receptor
Natriuretic peptide receptor
TNF receptor

LIGANDS
Acadesine	Halothane
Adenosine	Isoflurane
ATP	Ketamine
Bradykinin	Losartan
Buprenorphine	Methadone
cAMP	Metoprolol
Candesartan	Midazolam
Captopril	Morphine
cGMP	Naloxone
Cyclosporine A	Olmesartan
Diazepam	Pancuronium
Dipyridamole	Propofol
Enalapril	Prostacyclin
Enalaprilat	Ramipril
Enflurane	Remifentanil
Fentanyl	Scopolamine
Flunitrazepam	Sufentanil
	Telmisartan

These Tables list key protein targets and ligands in this article which are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Pawson et al., 2014) and are permanently archived in the Concise Guide to PHARMACOLOGY 2013/14 (^a,b,c,dAlexander et al.,2013a,b,c,d,,,).

Cardioprotection

Infarct size determines the prognosis of patients suffering an acute myocardial infarction (AMI) (Burns et al., 2002) and it is therefore important to reduce infarct size as much as possible (Braunwald, 1974). Reperfusion of the ischaemic region is mandatory to halt the progression of infarction during sustained myocardial ischaemia (Ginks et al., 1972; Maroko et al., 1972). Reperfusion, however, is a double-edged sword and not only terminates myocardial ischaemic injury, but also induces additional myocardial injury (Heusch, 2004; Yellon and Hausenloy, 2007). Therefore, there is a need for further cardioprotection beyond that by reperfusion (Heusch, 2013a).

Brief episodes of myocardial ischaemia/reperfusion before or shortly after a sustained myocardial ischaemia effectively reduce infarct size, but only when there is ultimately reperfusion. These protective phenomena are termed ischaemic pre- or postconditioning respectively. Likewise, cardioprotection can be achieved by inducing brief episodes of ischaemia/reperfusion in a myocardial territory remote from the infarcting myocardium or in an organ remote from the heart – that is remote ischaemic conditioning. The conditioning phenomena were originally reported in dogs (Murry et al., 1986; Przyklenk et al., 1993; Zhao et al., 2003), but have meanwhile been confirmed in many species, including humans (Heusch, 2013a). Apart from and beyond the reduction in infarct size, conditioning strategies also reduce arrhythmias, preserve ventricular function, prevent the development of heart failure and even improve clinical outcome (Davies et al., 2013; Heusch, 2013a; Thielmann et al., 2013; Heusch et al., 2014; Sloth et al., 2014).

Common to all conditioning phenomena is the recruitment and activation of cardioprotective signalling by the mechanical intervention which induces brief cycles of ischaemia and reperfusion in the heart or in an organ remote from the heart (Heusch et al., 2008). The underlying signal transduction of mechanically induced endogenous cardioprotection has been studied extensively with the aim to develop the identified signals into pharmacological conditioning strategies which could then be translated to patients suffering from myocardial ischaemia/reperfusion events. The different conditioning manoeuvres appear to act to a certain extent through common cardioprotective signalling steps (Downey and Cohen, 2005). However, the notion of common signalling steps for all conditioning strategies was largely derived by inference from experiments using different setups, animal models and conditioning manoeuvre. Currently, no data are available which directly compare the cardioprotective signalling under different conditioning strategies in the same animal model. With such caveats in mind, the signal transduction of the conditioning manoeuvres can be conceptually classified either by the timing of their operation (trigger, mediator and effector) or their (sub-) cellular localization (extracellular molecules, cytosolic signal transduction and target organelle/structure).

With respect to the temporal sequence of cardioprotective signalling, a trigger is released during the repeated cycles of ischaemia/reperfusion before or after the sustained myocardial ischaemia which act as the stimulus for cardioprotection. The trigger then activates a receptor-dependent or receptor-independent signalling cascade. A mediator is activated by the trigger and actively transmits the cardioprotective signal during the sustained ischaemia/reperfusion. An effector is the target of the protective signalling which when activated during the sustained ischaemia or during early reperfusion ultimately attenuates myocardial injury (Yellon and Downey, 2003). Such temporal classification of signalling steps was originally developed for ischaemic preconditioning (Downey et al., 2008); it never really became clear where the ‘memory’ resided which remembered the trigger during the preconditioning cycles and then activated the mediator during the sustained ischaemia. Also, such temporal classification of cardioprotective signal transduction is conceptually more difficult to use during ischaemic postconditioning when trigger, mediator and effector must all be sequentially activated over a very short period during early reperfusion.

Therefore, the local classification of cardioprotective signalling is more popular now and we will use it here. Extracellular molecules are released during the brief conditioning cycles of ischaemia/reperfusion from various cellular compartments (cardiomyocytes, endothelium, nerve endings, etc.) and activate the protective signal cascade through sarcolemmal receptors or independently of receptors. Autacoids, such as adenosine, bradykinin and opioids, activate GPCRs; natriuretic peptides activate their specific receptors and cytokines activate gp130 (IL-6, β subunit) receptors. Reactive oxygen species (ROS) and NO can initiate receptor-independent protective signalling (Heusch et al., 2008; Tullio et al., 2013; Rassaf et al., 2014). Within the cardiomyocyte, a variety of proteins are activated as cytosolic signal transducers. They interact at different levels and at different time points during the conditioning and the sustained index ischaemia/reperfusion. Three major intracellular signal transduction pathways are apparent: the endothelial NOS/PKG (eNOS/PKG) pathway (Cohen and Downey, 2007), the reperfusion injury salvage kinase (RISK) pathway (Hausenloy and Yellon, 2004) and the survivor activating factor enhancement pathway (Heusch et al., 2008; Lecour, 2009) (Figure 1). These complex signalling cascades interact at different levels (Vahlhaus et al., 1996; 1998; Cohen and Downey, 2007; Cohen et al., 2007) (Figure 1). The mitochondria have been identified as a major common intracellular target structure (Heusch et al., 2008). Mitochondria are the major energy source (aerobic ATP production) of the cell and hence relevant for all cellular functions and cell survival in general. Beyond and apart from energy metabolism, mitochondria have a decisive role in apoptosis, autophagy and necrosis (Heusch et al., 2008; 2010). Various PKs activate the mitochondrial ATP-sensitive K⁺ (K_ATP) channel, thereby trigger a modest amount of ROS formation (Pain et al., 2000; Heinzel et al., 2005; Downey et al., 2007; Costa and Garlid, 2008) and inhibit the opening of the mitochondrial permeability transition pore (MPTP) (Hausenloy et al., 2002; 2004b; Argaud et al., 2004; 2005,; Juhaszova et al., 2004; Heusch et al., 2010). Nitrosation (Methner et al., 2013; Rassaf et al., 2014) and nitrosylation (Penna et al., 2013) of mitochondrial membrane proteins are causally involved in cardioprotection.

Figure 1.

Signal transduction in cardioprotection: the extracellular signalling molecules adenosine, bradykinin and opioids act as triggers, the respective receptors and downstream activated protein kinases as mediators and the mitochondria as effectors (modified from Heusch et al., 2008). eNOS/PKG, endothelial NOS/PKG pathway; gp130, glycoprotein 130; GSK3β, glycogen synthase kinase 3β; K_ATP, mitochondrial ATP-sensitive K⁺ channel: MPTP, mitochondrial permeability transition pore; NPR, natriuretic peptide receptor; P70S6K, p70 ribosomal S6 protein kinase; pGC, particulate guanylate cyclase; RISK, reperfusion injury salvage kinase pathway; ROS, reactive oxygen species; SAFE, survivor activating factor enhancement pathway; sGC, soluble guanylate cyclase; TNFR, tumour necrosis factor receptor.

In the following, we will focus on the extracellular signalling molecules which are potentially druggable and translatable for cardioprotection, and more specifically those which act through receptors and for which drugs are available that interact with them. Knowledge about such receptor ligands is derived from a multitude of studies, ranging from in vitro studies of isolated subcellular elements, cells or heart preparations to different in vivo models from different species and using a variety of techniques, ranging from immunoblotting, biochemical analyses to pharmacological agonist and antagonist approaches and molecular genetic approaches (e.g. knockout, knockdown and transgenic overexpression).

Extracellular receptor ligands in conditioning

Ischaemic preconditioning was originally reported as an-all-or-none phenomenon. Protection by ischaemic preconditioning in rats (Barbosa et al., 1996), rabbits (Goto et al., 1995) and pigs (Schulz et al., 1998), however, depends on the strength (number and duration of ischaemia/reperfusion cycles) of the preconditioning stimulus, reflecting a dose–response relationship between the trigger/released ligand during the preconditioning cycles of ischaemia/reperfusion and the resulting infarct size reduction (Yellon and Downey, 2003). A protective response is already induced by one short (between 1.5 and 2.5 min) or more cycles of ischaemia followed by a minimum reperfusion period of about 30 s to 1 min (Alkhulaifi et al., 1993). Ligands involved in such ischaemic preconditioning protection are adenosine (Liu et al., 1991; Schulz et al., 1995), bradykinin (Goto et al., 1995; Schulz et al., 1998) and opioids (Schultz et al., 1995; Cohen et al., 2001; Schulz et al., 2001). For these three ligands, an increased release/bioavailability has been detected during/after the preconditioning procedure (Goto et al., 1995; Schulz et al., 1995; 1998,; Younes et al., 2005). A second window of protection is apparent between 12 and 72 h after the preconditioning stimulus; the ligands which are involved as trigger and/or mediator are mostly identical to those in acute ischaemic preconditioning (Bolli, 2000; Yellon and Downey, 2003; Stein et al., 2004).

Adenosine was the first signalling molecule reported to trigger and mediate cardioprotection by ischaemic preconditioning in the rabbit heart (Liu et al., 1991). Myocardial interstitial adenosine levels increase during ischaemic preconditioning in rats (Kuzmin et al., 2000), rabbits (Lasley et al., 1995), dogs (Mei et al., 1998) and pigs (Schulz et al., 1995). Adenosine is formed intracellularly through ecto-5′-nucleotidase and S-adenosylhomocysteine hydrolase in vascular cells and in cardiomyocytes. During hypoxia, the cellular formation and release of adenosine are increased. The half-life of extracellular adenosine is short, the cellular uptake of interstitial and plasmatic adenosine is fast and the enzymatic metabolism of adenosine through adenosine kinase and adenosine deaminase is high (Deussen, 2001). A decrease in interstitial adenosine with intracoronary adenosine deaminase in pigs prevents (Schulz et al., 1995), whereas an increase in interstitial adenosine with uptake inhibition by dipyridamole enhances the myocardial protection by ischaemic preconditioning in rabbits and dogs (Auchampach and Gross, 1993; Suzuki et al., 1998). Four adenosine receptor subtypes (A₁, A_2A, A_2B, A₃) have been identified which are expressed on most cells of the body. In the cardiovascular system, adenosine receptors exist on cardiomyocytes, endothelial cells, fibroblasts, smooth muscle cells and also on blood cells (Forman et al., 2006; McIntosh and Lasley, 2012). In the vasculature, all four receptor subtypes are expressed on vascular smooth muscle cells; A_2A and A_2B receptors are also expressed on endothelial cells. The A_2A receptor is the predominant receptor for coronary vasomotor effects and its activation induces coronary vasodilation (Mustafa et al., 2009). On cardiomyocytes, A₁, A_2A and A₃ receptors are expressed (Xin et al., 2012); for the A_2B receptor, an intracellular, mitochondrial localization has been demonstrated in rat cardiomyocytes which serves a cardioprotective function (Grube et al., 2011). The A₁ receptor is abundantly localized on cardiomyocytes and mediates the negative chronotropic and dromotropic effects of adenosine (Mustafa et al., 2009). All adenosine receptors are coupled to G proteins (Mustafa et al., 2009; McIntosh and Lasley, 2012). Whereas A_2A and A_2B receptors couple to Gs proteins (Xin et al., 2012), which activate PKA via adenylate cyclase activation and accumulation of cAMP, A₁ and A₃ receptors couple to Gi and Gq proteins and thereby inhibit the activity of adenylate cyclase (Mustafa et al., 2009; Xin et al., 2012). Downstream of the different G proteins, PKC is directly activated through adenosine, which appears to be a unique pathway for adenosine (Downey et al., 2007). Also, the eNOS/PKG and RISK signal transduction pathways are activated by adenosine (Heusch et al., 2008) (Figure 1). Cardioprotection during ischaemia and reperfusion is mediated through activation of different adenosine receptors (Heusch, 2010). Blockade of A₁ and A₃ receptors prevents the protection by ischaemic preconditioning. In contrast, A₂ receptor blockade does not affect infarct size in rabbits (Liu et al., 1991; 1994; McCully et al., 2001). Different from ischaemic preconditioning, A_2A and A_2B receptors are involved in the cardioprotection by postconditioning, and A₁ and A₃ receptors are irrelevant here (McIntosh and Lasley, 2012). A selective A_2A and A_2B receptor agonist, when given at early reperfusion, induces protection in different species (mouse, rat, rabbit, dog and pig) (McIntosh and Lasley, 2012). The timing and cooperation of A_2A and A_2B receptor activation are critical (Methner et al., 2010), that is they must be activated during the last minutes of ischaemia and the first minutes of reperfusion in mouse (Methner et al., 2010) and rat hearts (Xi et al., 2009) to induce cardioprotection during ischaemic postconditioning. Blockade of the A₁ receptor during ischaemic postconditioning does not affect infarct size in rats (Kin et al., 2005) and rabbits (Philipp et al., 2006). Non-selective blockade of adenosine receptors does not influence the protection by remote ischaemic preconditioning (RIPC) in pigs (Hausenloy et al., 2012). The causal involvement of adenosine in the conditioning phenomena is established in most species (Forman et al., 2006; Cohen and Downey, 2008), except the rat (Li and Kloner, 1993). However, the reduction of infarct size by exogenous adenosine or its selective agonists remains controversial. Exogenous adenosine or its selective agonists given before ischaemia in rats induced protection (Dai et al., 2009) or did not (Li and Kloner, 1993), and when given just before reperfusion in rabbits induced protection (Xu et al., 2000) or did not (Baxter et al., 2000). Notably in larger animals, adenosine reduced infarct size in pigs (van Winkle et al., 1994), but when given on top of lidocaine, adenosine either reduced infarct size in dogs (Homeister et al., 1990) or did not (Vander Heide and Reimer, 1996).

Ischaemic preconditioning increases the interstitial bradykinin concentration in rabbits (Goto et al., 1995) and pigs (Schulz et al., 1998). Bradykinin is cleaved from kininogens. Lys-bradykinin is the major kinin peptide in the interstitium which is then converted to bradykinin by the aminopeptidase M (Kokkonen et al., 1999). The kinin metabolism is linked to the angiotensin metabolism. Bradykinin is rapidly degraded by ACE. In parallel to ACE, neutral endopeptidase degrades bradykinin to a receptor-inactive kinin metabolite. The kinin metabolism is predominantly localized in the vascular bed; however, there is also a bradykinin synthesis in cardiomyocytes. There are two bradykinin receptors, B₁ and B₂ receptors. The B₂ receptor is localized on cardiomyocytes, endothelial cells and fibroblasts (Kokkonen et al., 2000), whereas the B₁ receptor is usually absent on all cell types and its expression is only up-regulated with inflammation and/or severe tissue damage (Burley et al., 2007). Thus, the B₂ receptor (Kokkonen et al., 2000) is the relevant one for acute protection in rats and pigs (Linz et al., 1997; Schulz et al., 1998; Kokkonen et al., 2000). The B₂ receptor is linked to Gi proteins and signals downstream to the intracellular eNOS/PKG and RISK signal transduction pathways (Heusch et al., 2008) (Figure 1). Exogenous bradykinin induces myocardial protection in rats (Linz et al., 1997), and ACE inhibitors mimic the cardioprotective effects of bradykinin (Wall et al., 1994). Increasing the plasma angiotensin II concentration by antagonism of angiotensin II type 1 (AT₁) receptors activates the AT₂ receptor, subsequently initiates a kininogen activity, which finally results in increased bradykinin formation (Jalowy et al., 1999). In this way, AT₁ receptor antagonists also mimic the cardioprotective effects of bradykinin (Jalowy et al., 1998). Bradykinin increases prostacyclin synthesis; activation of the B₂ receptor results in activation of the COX and de novo synthesis of prostacyclin, which then attenuates ischaemia/reperfusion injury (Jalowy et al., 1998). In pigs, the combination of an ACE inhibitor and an AT₁ receptor antagonist enhances the reduction of infarct size over that by monotherapy with each single drug (Weidenbach et al., 2000). Ischaemic postconditioning is also mediated through bradykinin and the respective downstream pathways, including prostacyclin synthesis (Penna et al., 2008). B₂ receptor blockade prevents remote preconditioning cardioprotection by mesenteric ischaemia/reperfusion in rats (Schoemaker and van Heijningen, 2000). In humans undergoing coronary artery bypass surgery (CABG), B₁ and B₂ receptor expression on neutrophils is down-regulated by remote ischaemic conditioning (Saxena et al., 2013), indicating altered bradykinin metabolism.

Opioid peptides are secreted from cardiac nerves or produced in the cardiomyocyte itself. Apart from and beyond their synthesis, the myocardium is capable of storage and release of opioid peptides; however, enzymatic proteolysis of the stored peptides is required before active peptides are released (Pugsley, 2002). The endogenous opioid peptides act through activation of μ, δ and κ receptors respectively. In the adult myocardium, the δ and κ receptors are expressed, whereas the μ receptor is absent (Peart et al., 2005). The opioid receptors are coupled to Gi proteins and share parts of the intracellular signalling with that of adenosine and bradykinin. Downstream of opioid receptor activation, also the activation of ERK, mediates cardioprotection (Ikeda et al., 2006; Shimizu et al., 2009) and ERK projects further downstream onto p70 ribosomal protein S6 kinase and glycogen synthase kinase 3ß (Figure 1). δ And κ receptor agonists, such as remifentanil which activates δ and κ receptors, methadone which activates the δ receptor (Gross et al., 2009) and U50488H which activates the κ receptor (Wang et al., 2014), induce myocardial protection when given before ischaemia (Wang et al., 2014) or during reperfusion (Gross et al., 2009; Wong et al., 2010b) in rats. The non-selective opioid receptor antagonist naloxone prevents the protection by ischaemic preconditioning in rats (Schultz et al., 1995), rabbits (Miki et al., 1998) and pigs (Schulz et al., 2001) and by RIPC in rabbits (Shimizu et al., 2009). The protection by ischaemic preconditioning is prevented by selective antagonism of δ and κ receptors, but not by antagonism of μ receptors (Gross, 2003; Peart et al., 2005). The most important receptor for protection is the δ receptor. A specific δ receptor antagonist reduces cardioprotection markedly, whereas a κ receptor antagonist has little effects on ischaemic preconditioning (Schultz et al., 1998; Aitchison et al., 2000) and ischaemic postconditioning in rats (Jang et al., 2008; Zatta et al., 2008; Guo et al., 2011), and during RIPC in rats (Wong et al., 2010b).

The available evidence suggests that the above-mentioned autacoids are involved in the different ischaemic conditioning manoeuvres and possibly share parts of their signalling. However, that notion does not imply that the underlying mechanism(s) are indeed identical. Autacoids appear to initiate protection in an interactive manner (Figure 1). At weaker stimuli (shorter duration or less cycles of preconditioning ischaemia), bradykinin is more important; at stronger stimuli (longer duration or more cycles of preconditioning ischaemia), adenosine becomes more important (Goto et al., 1995; Barbosa et al., 1996; Schulz et al., 1998). Fentanyl-induced myocardial protection is abolished by pretreatment with an adenosine receptor antagonist in rats (Kato et al., 2000), emphasizing the interactive signalling of cardioprotection.

Drugs based on receptor ligands

Numerous experimental studies suggest that adenosine, bradykinin and endogenous opioids participate in myocardial protection during ischaemia/reperfusion. These studies encouraged the use of the respective drugs in clinical settings. Here, we focus on studies reporting settings of acute myocardial ischaemia/reperfusion, either spontaneous or elective, during percutaneous coronary intervention (PCI) or cardiac surgery involving ischaemic cardioplegic arrest (CABG and/or valve replacement). We selected studies using the autacoids as such, specific agonists of the autacoids or drugs increasing their bioavailability respectively. We considered only placebo-controlled (or for opioids: alternative drug) studies using myocardial infarct size as the primary and most robust endpoint for cardioprotection. Infarct size has been characterized in these studies by biomarkers reflecting myocardial damage (such as creatine kinase or troponins) or by imaging techniques (such as thallium or sestamibi single-photon emission CT or gadolinium contrast MRI). We used the following terms for systematic search in Medline, Current Contents and PubMed: ‘adenosine/bradykinin, ACE inhibitor (specifically: captopril, enalapril, ramipril), AT₁ receptor antagonist (specifically: candesartan, losartan, olmesartan, telmisartan)/opioid, opioids (specifically: buprenorphine, butorphanol, fentanyl, levorphanol, morphine, remifentanil, sufentanil)’ and ‘myocardial infarction/coronary artery bypass surgery/percutaneous coronary intervention/reperfusion/revascularisation’ and ‘infarct size/creatine kinase/troponin’ and ‘human/patient’. We included those reports published up to May 2014 in English, or with an abstract in English, which provided data for infarct size in humans. Studies analysing only subgroups of larger trials were not systematically considered.

Adenosine has been tested in several clinical trials in the scenario of AMI, elective PCI, CABG or other cardiac surgery. Mode (bolus vs. infusion), dose and timing (before or during ischaemia and reperfusion) of administration differed (Table 1). Taken together and considering such differences, there are several positive studies, but there is no consistent evidence that adenosine reduces infarct size in clinical settings of myocardial ischaemia/reperfusion (Table 1). However, even in studies with reduced infarct size (e.g. AMISTAD II), there was no benefit in clinical outcome with adenosine or its selective agonists (Ross et al., 2005). Exogenous bradykinin or acutely increasing bradykinin's bioavailability by ACE inhibitors or angiotensin receptor blockers has also been tested in several clinical scenarios (Table 2). Independent of mode, dose or timing, infarct size was consistently reduced in these studies (Table 2). Data on clinical outcome from these studies are lacking. Different synthetic opioid agonists have been tested in several clinical scenarios of myocardial ischaemia/reperfusion (Table 3). Drugs, dose, mode (bolus vs. infusion) and timing (before or during ischaemia) of administration differed (Table 3). From these studies, there is no evidence that synthetic opioid agonists reduce infarct size in clinical settings of myocardial ischaemia/reperfusion (Table 3).

Table 1.

Effects of adenosine or its agonists on infarct size in patients

Study	Clinical scenario	Method to assess endpoint	Dosage	Time of application	Intervention [n]	Placebo [n]	Infarct size [Δ% of placebo]
Mahaffey et al., 1999 AMISTAD I	STEMI AMI	SPECT (at day 5–9)	Adenosine, i.v. (70 μg·kg−1·min−1)		101	96	↓33*
Marzilli et al., 2000	AMI	CK (max within 24 h)	Adenosine, i.v. (4 mg)		27	27	↓22
		CK-MB (max within 24 h)					↓55
Kopecky et al., 2003 ADMIRE	STEMI AMI	SPECT (at day 5–9)	AMP579, i.v. (15 μg·kg−1)		78	74	↑17
			AMP579, i.v. (30 μg·kg−1)		83		↓22
			AMP579, i.v. (60 μg·kg−1)		76		↓28
Ross et al., 2005 AMISTAD II	STEMI AMI	SPECT (at day 5–9)	Adenosine, i.v. (50 μg·kg−1·min−1)		∼60	∼60 (243)	↓15
			Adenosine, i.v. (70 μg·kg−1·min−1)		∼60		↓61*
Fokkema et al., 2009	STEMI AMI	CK (max within 48 h)	Adenosine, i.c. (2 × 120 μg)		226	222	↑2
		CK-MB (max within 48 h)					↑15
Desmet et al., 2011	STEMI AMI	MRI (at day 5–9)	Adenosine, i.c. (4 mg)		51	49	↑15
		CK (max within 24 h)					↑15
		CK (AUC 24 h)					↑31
		CK-MB (max within 24 h)					↑7
		CK-MB (AUC 24 h)					↑2
		Troponin I (max within 24 h)					↑16
		Troponin I (AUC 24 h)					↑16
Grygier et al., 2011	STEMI AMI	CK (max within 24 h)	Adenosine, i.c. (1–2 mg)		35	35	↓21
		CK-MB (max within 24 h)					↑13
		Troponin I (max within 24 h)					↑8
Zhang et al., 2014	STEMI AMI	SPECT (at day 1)	Adenosine, i.v. (50 μg·kg−1·min−1)		19	17	↓15
			Adenosine, i.v. (70 μg·kg−1·min−1)		18		↓21
		SPECT (after 6 months)	Adenosine, i.v. (50 μg·kg−1·min−1)		19	17	↓13
			Adenosine, i.v. (70 μg·kg−1·min−1)		18		↓25*
		CK-MB (max within 24 h)	Adenosine, i.v. (50 μg·kg−1·min−1)		32	31	↓18
			Adenosine., i.v. (70 μg·kg−1·min−1)		27		↓35*
Lee et al., 2007	Elective PCI	CK-MB (max within 24 h)	Adenosine, i.c. (50 μg)		31	31	↓50*
De Luca et al., 2012	Elective PCI	Troponin I (max within 24 h)	Adenosine, i.c. (120 μg and 180 μg)		130	130	↑11
Kim et al., 2012	Elective PCI	CK-MB (max within 12 h)	Adenosine, i.c. (50 μg)		54	55	↑89
Belhomme et al., 2000	CABG	Troponin I (AUC 48 h)	Adenosine, i.v. (140 μg·kg−1·min−1)		22	22	↑22
Rinne et al., 2000	CABG	CK-MB (max within 48 h)	Adenosine, i.v. (12 mg)		20	20	↑36
Teoh et al., 2002	CABG	Troponin T (mean of 72 h)	GR79236X, i.v. (100 μg·mL−1)		10	10	↓7
Jakobsen et al., 2013	CABG	Troponin T (max within 48 h) CK-MB (max within 48 h)	Adenosine, with cardioplegic solution (1.2 mmol·L−1)		30	30	–
Jin et al., 2007	Valve surgery	Troponin I (AUC 24 h)	Adenosine, i.c. (1.5 mg·kg−1)		30	30	↓31*

^*P < 0.05 versus placebo.

AMI, acute myocardial infarction; CABG, coronary artery bypass surgery; CK-MB, creatine kinase – muscle, brain; i.c., intracoronary; isch, ischaemia; max, maximum; PCI, percutaneous coronary intervention; rep, reperfusion; SPECT, thallium or sestamibi single-photon emission CT; STEMI AMI, ST elevation myocardial infarction, acute myocardial infarction.

Table 2.

Effects of acute ACE inhibitor, ARB or bradykinin on infarct size in patients

Study	Clinical scenario	Method to assess endpoint	Dosage	Time of application	Intervention [n]	Placebo [n]	Infarct size [Δ% of placebo]
Bussmann et al., 1995	AMI	CK (max within 48 h)	Captopril, i.v. (2.5–5 mg and 1.5–2 mg·h−1)		22	24	↓13
		CK-MB (max within 48 h)			22	24	↓1
Kurz et al., 2001	AMI	CK (max within 36 h)	Enalaprilat, i.c. (50 μg)		11	11	↓18
		CK-MB (max within 36 h)			11	11	↓17
Shariff et al., 2010	STEMI AMI	Troponin I (max within 24 h)	ACE inhibitor or ARB, oral (drugs and dosage not specified)		66	445	↓33*
Mangiacapra et al., 2013	Elective PCI	hs troponin T (mean of 24 h)	Enalaprilat, i.c. (50 μg)		20	20	↓63*
Boldt et al., 1996	CABG	CK-MB (max within 72 h)	Enalaprilat, i.v. (5 μg·kg−1·min−1)		22	22	↓62*
		Troponin T (max within 72 h)					↓87*
Walter et al., 2002	CABG	CK (within 120 h)	Enalaprilat, oral (20 mg·day−1)		22	21	–
		CK-MB (within 120 h)			22	21	–
		troponin T (within 120 h)			22	21	–
Wei et al., 2004	CABG	CK-MB (within 48 h)	Bradykinin, i.v. (4 μg·min−1)		21	20	↓37*
		Troponin I (within 48 h)			21	20	↓32
Benedetto et al., 2008	CABG	Troponin I (max within 48 h)	ACE inhibitor, oral (drugs and dosage not specified)		245	236	↓33*
Wang et al., 2009	CABG	CK-MB (max within 48 h)	Bradykinin, i.v. (4 μg·min−1)		19	19	↓40*
		Troponin T (max within 48 h)			19	19	↓16

^*P < 0.05 versus placebo.

AMI, acute myocardial infarction; ARB, angiotensin receptor blocker; CABG, coronary artery bypass surgery; CK-MB, creatine kinase – muscle, brain; i.c., intracoronary; isch, ischaemia; max, maximum; PCI, percutaneous coronary intervention; rep, reperfusion; STEMI AMI, ST elevation myocardial infarction, acute myocardial infarction.

Table 3.

Effects of opioids on infarct size in patients

Study	Clinical scenario	Method to assess endpoint	Drug and dosage	Time of application	Opioid [n]	Alternative anaesthesia or placebo [n]	Infarct size [Δ% of placebo]
Rentoukas et al., 2010	STEMI AMI	Troponin I (max during hospitalization)	Morphine, i.v. (5 mg) + RIPC vs. RIPC, no further medication		33	33	↓38
Abdel-Wahab et al., 2008 PROFIT	Elective PCI	Troponin T (max within 12 h)	Fentanyl (0.05 mg) vs. diazepam, both i.v.		94	90	↓31
		CK-MB (max within 12 h)					↓5
		Troponin T (max within 12 h)	Fentanyl (0.1 mg) vs. diazepam, both i.v.		92		↓8
		CK-MB (max within 12 h)					↑5
Slogoff et al., 1989	CABG	CK-MB (mean within 16 h)	Sufentanil (15–25 μg·kg−1 and 5 mg·kg−1) vs. enflurane or halothane or isoflurane on top of diazepam, pancuronium, fentanyl, all i.v.		254	257/253/248	–
Tuman et al., 1989	CABG	CK-MB (max within 72 h)	Fentanyl (>50 μg·kg−1) vs. diazepam (0.4–1 mg·kg−1) or halothane (0.5–2.5%) on top of benzodiazepine, scopolamine, isoflurane, enflurane, halothane all i.v. – except halothane volatile		240	250	↑67
							vs. diazepam
							↑3
							vs. halothane
			Fentanyl (<50 μg·kg−1) vs. diazepam (0.4–1 mg·kg−1) or halothane (0.5–2.5%) on top of benzodiazepine, scopolamine, isoflurane, enflurane, halothane all i.v. – except halothane volatile		345		↑57
							vs. diazepam
							↑9
							vs. halothane
			Sufentanil (3–8 μg·kg−1) vs. diazepam (0.4–1 mg·kg−1) or halothane (0.5–2.5%) on top of benzodiazepine, scopolamine, isoflurane, enflurane, halothane all i.v. – except halothane volatile		212		↑53
							vs. diazepam
							–
							vs. halothane
Neuhäuser et al., 2008	CABG	Troponin T (max within 24 h)	Sufentanil (0.25–1 μg·kg−1) on top of flunitrazepam, morphine, midazolam, pancuronium vs. ketamine (1–3 mg·kg−1) on top of flunitrazepam, morphine, midazolam, propofol, pancuronium all i.v.		108	101	↑50*
Wong et al., 2010a	CABG	Troponin I (max within 24 h)	Remifentanil (1 and 0.5 μg·kg−1) vs. placebo on top of morphine, scopolamine, entomidate, fentanyl, pancuronium, propofol all i.v.		20	20	↓38*
		CK-MB (max within 24 h)					↓39*

^*P < 0.05 versus alternative anaesthesia/placebo.

AMI, acute myocardial infarction; CABG, coronary artery bypass surgery; CK-MB, creatine kinase – muscle, brain; isch, ischaemia; max, maximum; PCI, percutaneous coronary intervention; rep, reperfusion; RIPC, remote ischaemic preconditioning; STEMI AMI, ST elevation myocardial infarction, acute myocardial infarction.

Cardioprotection by other receptor-dependent and non-receptor-dependent signalling molecules and its clinical translation

Activation of brain natriuretic peptide receptors recruits a cardioprotective signal transduction cascade which involves increased myocardial cGMP and activation of mitochondrial K_ATP channels to reduce infarct size in isolated rat hearts (D'Souza et al., 2003). In the J-WIND trial in more than 500 patients with reperfused AMI, i.v. infusion of atrial natriuretic peptide (ANP) for 3 days after reperfusion reduced infarct size, as assessed from the AUC of creatine kinase (Kitakaze et al., 2007).

Recently, i.v. infusion of the ß-blocker metoprolol just before reperfusion in patients with AMI undergoing primary PCI reduced infarct size, as assessed by MRI (Ibanez et al., 2013), and the benefits for ventricular function and survival persisted more long-term (Pizarro et al., 2014). It is currently unclear, whether the observed benefit from metoprolol is shared by other ß-blockers and to which particular property of metoprolol or ß-adrenoceptor blockade the protection might relate.

Exenatide is a mimetic of human glucagon-like peptide (GLP)-1, activates the G-protein coupled GLP-1 receptor, recruits a protective signal transduction and reduces infarct size in an anaesthetized pig model of myocardial ischaemia/reperfusion (Timmers et al., 2009). Such cardioprotection was also recently reported in patients with AMI undergoing primary PCI who had reduced infarct size with i.v. exenatide just before reperfusion, as assessed by MRI (Lønborg et al., 2012).

There is currently a notion that cardioprotective signal transduction converges onto the mitochondria, more specifically the MPTP (Heusch et al., 2008; 2010,). c inhibits MPTP opening and reduces infarct size in mice (Boengler et al., 2010) and pigs (Skyschally et al., 2010; Gedik et al., 2013), but not in rats (De Paulis et al., 2013). Also, in patients with AMI undergoing primary PCI, an i.v. bolus of cyclosporine A just before reperfusion reduced infarct size, as assessed from the AUC of creatine kinase release (Piot et al., 2008); similarly, protection by cyclosporine A was also reported for patients undergoing CABG (Hausenloy et al., 2014).

Conclusion and perspectives for the future of pharmacological cardioprotection

The translation of extracellular signalling molecules with established cardioprotective potential in experimental animal models to cardioprotection in clinical scenarios of myocardial ischaemia and reperfusion has been largely disappointing, so far.

The obstacles to successful translation have often been emphasized (Ovize et al., 2010; Schwartz Longacre et al., 2011; Bell et al., 2012; Heusch et al., 2012; Hausenloy et al., 2013; Heusch, 2013a; Kloner, 2013). Most of the experimental studies were performed in young and healthy animals, often only in rodents and not in larger mammals. Apart from the obvious species differences in cardiac size and geometry, as well as in haemodynamics, notably heart rate (Heusch, 2008), there are substantial species differences in cardioprotective signal transduction: for example, adenosine appears to be not important for ischaemic preconditioning in rats (Li and Kloner, 1993) and the RISK pathway is important for ischaemic postconditioning in rodents (Hausenloy and Yellon, 2004; Hausenloy et al., 2004a), but not in pigs (Skyschally et al., 2009). Age (Boengler et al., 2008), co-morbidities and co-medications (Heusch, 2012; Ferdinandy et al., 2014) are additional confounders of translation of cardioprotection to the clinical scenario. Also, the clinical scenario per se differs between CABG where there is controlled global myocardial ischaemia and reperfusion under cardioplegic protection (Thielmann et al., 2013) and primary PCI where there is plaque rupture with release of atherosclerotic debris, thrombotic material and soluble vasoconstrictor, thrombogenic and inflammatory factors (Kleinbongard et al., 2011).

Adenosine was the first signalling molecule of ischaemic preconditioning to be identified (Liu et al., 1991), and ever since there has been great enthusiasm to recruit its cardioprotective potential in patients with ischaemic heart disease. As detailed earlier, a consistent reduction in infarct size was not observed with adenosine or its agonists in various clinical trials, and even those with reduced infarct size did not report improved clinical outcome. In retrospect, not all studies using adenosine, notably not those in larger animals, were positive. It is therefore surprising that apparently people have not given up on adenosine. Preliminary data of the PROMISE trial again just report reduced infarct size in a subgroup of patients with ischaemia of less than 200 min duration (Garcia-Dorado et al., 2013). Acadesine, a substance that increases adenosine's bioavailability, again did not improve all-cause mortality, non-fatal stroke and severe left ventricular dysfunction in patients undergoing CABG (Newman et al., 2012).

As compared with adenosine, only few clinical studies investigated the acute use of bradykinin or of drugs increasing bradykinin's bioavailability in scenarios of myocardial ischaemia/reperfusion. ACE inhibitors and AT1 antagonists, substances which increase bradykinin's bioavailability, are typically used chronically for the treatment of hypertension or heart failure. With chronic use of ACE inhibitors and AT₁ receptor antagonists, the incidence of myocardial infarction is reduced (McAlister, 2012; Savarese et al., 2013), and apart from more favourable haemodynamics, bradykinin may contribute to such benefit.

There is no consistent evidence for cardioprotection by opioids in clinical settings. This lack of evidence for cardioprotection may relate to the underlying background anaesthesia, as most anaesthetics are per se cardioprotective (e.g. halothane, isoflurane, ketamine, propofol, sevoflurane, sufentanil) (Kato and Foex, 2002; Zaugg et al., 2014). Also, in studies where the opioid was not compared with strict placebo but to another drug [e.g. diazepam (Obame et al., 2007)], the potential cardioprotection by the comparator drug may obscure the cardioprotection by the opioid.

Given the recent evidence that remote ischaemic preconditioning and preconditioning by repeated inflation/deflation of a blood pressure cuff around a limb reduces infarct size during elective interventional and surgical coronary revascularization (Hausenloy et al., 2007; Thielmann et al., 2010) as well as in patients with reperfused AMI (Bøtker et al., 2010) and may even improve clinical outcome (Davies et al., 2013; Thielmann et al., 2013; Sloth et al., 2014), the question arises whether we should abandon the search for cardioprotective drugs and embark on remote ischaemic conditioning as a simple, safe, cheap and effective cardioprotective strategy (Heusch, 2013b).

Glossary

AMI

acute myocardial infarction

AT₁ receptor

angiotensin II type 1 receptor

CABG

coronary artery bypass surgery

eNOS/PKG

endothelial NOS/PKG

GLP

glucagon-like peptide

K_ATP

mitochondrial ATP-sensitive K⁺ channel

MPTP

mitochondrial permeability transition pore

PCI

percutaneous coronary intervention

RIPC

remote ischaemic preconditioning

RISK

reperfusion injury salvage kinase

ROS

reactive oxygen species

Conflict of interest

None.