Sarcolemmal dependence of cardiac protection and stress‐resistance: roles in aged or diseased hearts

Abstract

Disruption of the sarcolemmal membrane is a defining feature of oncotic death in cardiac ischaemia–reperfusion (I‐R), and its molecular makeup not only fundamentally governs this process but also affects multiple determinants of both myocardial I‐R injury and responsiveness to cardioprotective stimuli. Beyond the influences of membrane lipids on the cytoprotective (and death) receptors intimately embedded within this bilayer, myocardial ionic homeostasis, substrate metabolism, intercellular communication and electrical conduction are all sensitive to sarcolemmal makeup, and critical to outcomes from I‐R. As will be outlined in this review, these crucial sarcolemmal dependencies may underlie not only the negative effects of age and common co‐morbidities on myocardial ischaemic tolerance but also the on‐going challenge of implementing efficacious cardioprotection in patients suffering accidental or surgically induced I‐R. We review evidence for the involvement of sarcolemmal makeup changes in the impairment of stress‐resistance and cardioprotection observed with ageing and highly prevalent co‐morbid conditions including diabetes and hypercholesterolaemia. A greater understanding of membrane changes with age/disease, and the inter‐dependences of ischaemic tolerance and cardioprotection on sarcolemmal makeup, can facilitate the development of strategies to preserve membrane integrity and cell viability, and advance the challenging goal of implementing efficacious ‘cardioprotection’ in clinically relevant patient cohorts.

Linked Articles

This article is part of a themed section on Molecular Pharmacology of G Protein‐Coupled Receptors. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v173.20/issuetoc

Abbreviations

Apo

Apolipoprotein

DHA

docosahexaenoic acid

EGFR

epidermal growth factor receptor

GRK

GPCR kinase

IHD

ischaemic heart disease

I‐R

ischaemia–reperfusion

PUFA

polyunsaturated fatty acid

RTK

receptor tyrosine kinase

SR

sarcoplasmic reticulum

STZ

streptozotocin

T1D

type I diabetes

T2DM

type II diabetes

Tables of Links

TARGETS
Other ion channels a	Akt, protein kinase B
Connexins	endothelial NOS
GPCRs b	GRK, GPCR kinase
Adenosine A1 receptor	HO‐1, haem oxygenase 1
β1‐adrenoceptor	MAP kinases
β2‐adrenoceptor	mTOR, mechanistic target of rapamycin
δ‐opioid receptor	PKA
Transporters c	PKC‐β
GLUT4	Catalytic receptors e
Enzymes d	EGF receptor
Adenylyl Cyclase	Receptor tyrosine kinases

LIGANDS
Adiponectin	Linoleic acid
cAMP	NADPH
Docosahexaenoic acid	NO
HMG‐CoA	PIP2
Insulin	Rac1
α‐Linolenic acid

These Tables list key protein targets and ligands in this article that are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Southan et al., 2016), and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 (^a,b,c,d,eAlexander et al., 2015a, 2015b, 2015c, 2015d, 2015e).

The challenge of clinical cardioprotection

Cardioprotective therapy to ameliorate damage with acute myocardial infarction and surgical ischaemia–reperfusion (I‐R) (coronary artery bypass graft, valve repair and cardiac transplantation) remains a highly desirable yet unrealised clinical goal (Hassouna et al., 2006; Miura and Miki, 2008; Peart and Headrick, 2009; Przyklenk, 2011; Hausenloy et al., 2013). Although a variety of experimental interventions have been identified that are highly effective in a laboratory setting – notably ischaemic preconditioning and postconditioning stimuli – they elicit modest if any benefits in clinical trials, with outcomes insufficient to justify clinical implementation (Miura and Miki, 2008; Peart and Headrick, 2009; Przyklenk, 2011; Hausenloy et al., 2013, 2015). Evidence of benefit via conditioning stimuli may also decline as trial design is improved, and more specific measures of infarction are assessed (Pilcher et al., 2012; Abdelnoor et al., 2014; Healy et al., 2014). As has been previously emphasized (Ferdinandy et al., 1998a; Miura and Miki, 2008; Peart and Headrick, 2009; Przyklenk, 2011) and highlighted by recent working groups and reviews (Hausenloy et al., 2013; Headrick et al., 2015; Kleinbongard and Heusch, 2015; Przyklenk, 2015), disappointing clinical translation may reflect a ‘co‐morbidity conundrum’ that demands attention. While highly efficacious in young healthy tissue, evidence indicates that a broad array of cardioprotective responses are impaired or entirely negated with ageing (Przyklenk et al., 2008; Peart and Headrick, 2009; Headrick et al., 2015), diabetes (Przyklenk, 2011), obesity ± insulin‐resistance (Katakam et al., 2007; Bouhidel et al., 2008; Wagner et al., 2008), hypercholesterolaemia (Szilvassy et al., 1995; Ferdinandy et al., 1997, 2003; Tang et al., 2005; Kupai et al., 2009; Görbe et al., 2011; Zhang et al., 2012; Xu et al., 2013) and hypertension/hypertrophy (Moolman et al., 1997; Riess et al., 2005; Ebrahim et al., 2007; Ma et al., 2013, 2014). These are all key and common co‐morbidities in those ultimately requiring cardioprotective therapy.

Strategic research is thus recommended (Hausenloy et al., 2013; Ferdinandy et al., 2014; Przyklenk, 2015) to unravel and potentially manipulate abnormalities in cardioprotective signalling. It may be that conventional conditioning stimuli can be further refined (e.g. by optimizing conditioning algorithms) to improve efficacies in clinically relevant settings. Beyond attempts to ‘retrofit’ solutions revealed from research in young healthy hearts, a strategic focus on protective mechanisms and signalling (together with injury mechanisms) in aged or diseased myocardium can better rationalize development of still elusive clinical cardioprotection.

Cardioprotection in aged myocardium

Ischaemic heart disease (IHD) and myocardial infarction are highly age‐dependent, primarily affecting those over 55 years of age. It is thus an unfortunate paradox that very little research into I‐R injury, and its management focuses on older hearts. Studies reveal that ageing impairs intrinsic resistance to I‐R injury in human (Mariani et al., 2000; Liu et al., 2012; Peart et al., 2014) and animal myocardium (Headrick, 1998; Headrick et al., 2003; Peart and Gross, 2004; Peart and Headrick, 2009), together with myocardial responsiveness to diverse protective stimuli (Boengler et al., 2009; Peart and Headrick, 2009; Przyklenk, 2011). We observe age‐dependent impairment of responses to GPCRs and preconditioning in human and rodent myocardium (Headrick et al., 2003; Peart and Gross, 2004; Peart et al., 2014), and others have reported failure of preconditioning in older human (Abete et al., 1997; Bartling et al., 2003) and animal tissue (Schulman et al., 2001; Boengler et al., 2007a), together with impaired responses to remote conditioning (Hu et al., 2002), postconditioning (Boengler et al., 2007b; Przyklenk et al., 2008; Przyklenk, 2011), anaesthetic (Mio et al., 2008), helium preconditioning (Heinen et al., 2008), adenosine (Schulman et al., 2001) and opioids (Peart and Gross, 2004; Peart et al., 2007, 2014; Headrick et al., 2015), among other stimuli.

Table 1 summarizes such changes, together with highlighting the membrane microdomain dependence of these responses.

Table 1.

Caveolin‐3, age and diabetes‐dependence of protection

Protective stimuli	Ageing effect	Diabetes effect (Przyklenk, 2011)	Hypercholesterolemia	Caveolin‐3 dependence
Ischaemic PreCon	↓ (Bartling et al., 2003; Peart et al., 2014)	↓ (Ghosh et al., 2001; Hassouna et al., 2006; Peart et al., 2014)	↓ (Tang et al., 2005; Görbe et al., 2011)	YES
Ischaemic PostCon	↓ (Boengler et al., 2007b; Przyklenk, 2011)	↓ (Przyklenk et al., 2011; Wider and Przyklenk, 2014)	↓ (Kupai et al., 2009)	YES
Anaesthetic	↓ (Mio et al., 2008)	↓ (Drenger et al., 2011; Tai et al., 2012)	↓ (Zhang et al., 2012; Xu et al., 2013)	YES
Adenosine agonists	↓ (Schulman et al., 2001; Headrick et al., 2003)	??	↓ (Ueda et al., 1999)	YES
Opioid agonists	↓ (Headrick et al., 2015; Peart et al., 2014)	↓ (Gross et al., 2007; Kim et al., 2010)	??	YES
Adiponectin	??	↓ (Yi et al., 2011)	??	YES
‘SLP’ (Peart and Gross, 2004; See Hoe et al., 2014)	− (Peart and Gross, 2004)	??	??	NO

Regarding remote conditioning, two recently published large randomised, multi‐centre trials assessing the efficacy of remote preconditioning failed to show any benefit of remote preconditioning (Hausenloy et al., 2015, Meybohm et al., 2015). The authors of both studies highlighted that the use of propofol across the cohorts may have affected the remote preconditioning. Nonetheless, the age of the cohort must also be considered in interpretation of the outcomes (Meybohm et al., 2015: 66 ± 10 years, Hausenloy et al., 2015: 76 ± 6 years).

Cardioprotection in diabetes and other co‐morbid conditions

Diabetes, a major and increasingly prevalent risk factor for IHD, also appears to sensitize human (Marso et al., 2007) and animal myocardium (Hassouna et al., 2006; Fricovsky et al., 2012) to I‐R injury, while impairing cardioprotection (Table 1). Diabetes reportedly inhibits preconditioning in human (Ghosh et al., 2001; Lee and Chou, 2003; Hassouna et al., 2006) and animal tissue (Kersten et al., 2000; Przyklenk, 2011), with protection via postconditioning (Przyklenk et al., 2011; Wider and Przyklenk, 2014), opioids (Gross et al., 2007; Kim et al., 2010; Chen et al., 2013), anaesthetic (Drenger et al., 2011; Tai et al., 2012) and adiponectin (Yi et al., 2011) among other responses impaired (Peart and Headrick, 2009; Przyklenk, 2011; Headrick et al., 2015).

In addition to age and diabetes, other IHD co‐morbidities impair cardioprotective responses and in some cases also affect intrinsic I‐R tolerance. Conditioning and other protective responses are inhibited in models of obesity ± insulin‐resistance (Katakam et al., 2007; Bouhidel et al., 2008; Wagner et al., 2008), hypercholesterolaemia (Szilvassy et al., 1995; Ferdinandy et al., 1997; 2003; Tang et al., 2005; Kupai et al., 2009; Görbe et al., 2011; Zhang et al., 2012; Xu et al., 2013) and hypertension/hypertrophy (Moolman et al., 1997; Riess et al., 2005; Ebrahim et al., 2007; Ma et al., 2013, 2014) (Table 1).

Basis of refractoriness to cardioprotection

Why does such a broad array of protective responses fail in aged and/or diseased hearts? Addressing this fundamental question can inform the development of protective therapy in a more rational manner. This question is also relevant to our understanding of the cardiac ageing process itself: a reduced capacity to respond to and withstand the effects of injurious stressors is a defining hallmark of ageing. Impairment of intrinsic cytoprotective mechanisms may thus be a critical determinant of the cardiac ageing process, akin to the ‘Green Theory’ of ageing (Gems and McElwee, 2005), which posits that the finite capacity of detoxification and repair mechanisms governs the process and rate of ageing.

Evidence to date supports age/disease‐dependent abnormalities in cardioprotection at three major subcellular levels: (i) mitochondria are critical to cell injury and survival in I‐R, and evidence is accumulating that mitochondrial architecture, dynamics (mitophagy, fission/fusion, and biogenesis) and functionality are all decreased by age and disease (Marzetti et al., 2013; Biala et al., 2015; Dorn, 2015); (ii) intracellular kinase signal cascades involved in transducing survival and death signals to mitochondria and other intracellular sites are substantially modified with both age (Tani et al., 2001; Peart et al., 2007, 2014) and disease (Gross et al., 2007; Yi et al., 2011; Whittington et al., 2013); and (iii) the sarcolemma, the site of initiation and transduction of receptor‐dependent survival and death signals (and governing ionic homeostasis, substrate availability and intercellular communication), is significantly modified with both ageing and relevant co‐morbid conditions.

Evidence of abnormalities spanning subcellular organisational levels suggests there may well be no ‘magic bullet’ for restoring I‐R tolerance and cardioprotection in clinically relevant older or diseased myocardium. Indeed, while mitochondrial effectors of protection may retain some efficacy in aged myocardium (Tani et al., 2001; Headrick et al., 2003; Huhn et al., 2012; Peart et al., 2014), direct modulation of these targets does not normalize I‐R outcomes, while targeting survival kinase pathways may confer protection yet also fails to normalize I‐R tolerance across age groups (Tani et al., 2001; Headrick et al., 2003; Peart et al., 2014). From a theoretical standpoint, addressing the most proximal causes of cardioprotective dysfunction might offer the greatest potential for improving cardioprotective phenotypes in target cohorts. Attention to abnormalities within the sarcolemma can not only improve the functionality of receptors governing cell survival versus death signalling but also influence subsequent signal transduction, ionic homeostasis, substrate uptake, intercellular communication and ultimately mitochondrial function and viability.

Sarcolemmal makeup changes with age and disease

Broad‐spectrum dysfunction in cardioprotection, encompassing multiple receptors and effectors (Table 1), is consistent with abnormalities in overarching determinants of cardioprotection, including membrane‐dependent control of cell signalling. Caveolae and caveolins, for example, have been picked out as critical determinants of stress responses and signalling (Roth and Patel, 2011; Fridolfsson et al., 2014; Peart et al., 2014; Schilling et al., 2015). Such a possibility is consistent with evidence of the following: essential roles for caveolae and caveolin‐3 in governing I‐R tolerance (See Hoe et al., 2014) and in expression of protection via preconditioning, opioid, adiponectin and anaesthetic stimuli (Horikawa et al., 2008; Tsutsumi et al., 2010; Roth and Patel, 2011; Sun et al., 2012; Wang et al., 2012a; Schilling et al., 2015); depletion of cardiac caveolin‐3 and caveolae with ageing (Peart et al., 2014) and diabetes (Lei et al., 2013; Liu et al., 2015); and differential efficacies of caveolin‐3 independent versus dependent interventions in aged versus young myocardium (Roth and Patel, 2011; Schilling et al., 2015) (Table 1). The influences of age and disease on membrane cholesterol and phospholipid profiles, and caveolar microdomains, may be particularly relevant to changes in myocardial I‐R tolerance and cardioprotection with age and disease.

Effects of ageing

Numerous studies have examined cell membrane ‘ageing’, with associated structural and functional changes attributed to modifications in membrane lipid structure, levels and distribution. The process of cellular ageing itself, together with organism longevity, may be governed by the cell membrane: replicative senescence is postulated to involve membrane‐related changes (elevated caveolin‐1/reduced amphiphysin‐1) that suppress growth factor receptor tyrosine kinase (RTK) signalling (Park et al., 2002), while organism lifespan is influenced by membrane phospholipid fatty acid composition (Hulbert, 2005; Mitchell et al., 2007; Hulbert, 2008; Valencak and Ruf, 2013). Hulberts ‘membrane pacemaker’ theory suggests dependence of ageing or longevity on membrane lipid saturation, with extent of acyl chain polyunsaturation increasing with metabolic rate and ROS production, promoting susceptibility to membrane peroxidative damage and reducing cellular and organism lifespan. There is experimental support for such a scheme, and age‐dependent increases in membrane free radical production and in lipid peroxidation are observed in different animal models, in association with reductions in membrane fluidity (Sawada et al., 1992; Yu et al., 1992; Sawada et al., 1993; Choe et al., 1995). However, recent work in Drosophila suggests susceptibility to lipid peroxidation is not a major factor in organism longevity. Nonetheless, the physiochemical effects of membrane phospholipid fatty acids may retard the cellular ageing process (Moghadam et al., 2013).

Ageing is associated with significant alterations in lipid composition of the cell membrane (Prisco et al., 1991; Laganiere and Yu, 1993), resulting in shifts in physical properties – notably reduced fluidity. This particular age‐dependent change is widely observed across cell types that include erythrocytes (Levin et al., 1992), platelets (Hossain et al., 1999; Noble et al., 1999), hepatocytes (Benedetti et al., 1988), macrophages (Alvarez et al., 1993), lymphocytes (Rivnay et al., 1980; Noble et al., 1999), endothelium (Hashimoto et al., 1999), synaptic membranes (Igbavboa et al., 1996), bile canaliculi (Hashimoto et al., 2001) and intestinal microvilli (Wahnon et al., 1989). This fluidity change correlates with alterations in cholesterol and phospholipid contents. A caveat is that much of this evidence stems from non‐cardiac cells, and important cell‐ and organ‐specific responses to age (and disease) may emerge. Our work, and that of others, supports decreased cardiac membrane cholesterol, caveolae and caveolin‐3 with age (Kawabe et al., 2001; Peart et al., 2014). While this decrease in cholesterol might be predicted to increase sarcolemmal fluidity, other factors also govern fluidity, including processes of glycation and glycoxidation (Waczulíková et al., 2010). Membrane changes do appear to differ between organs (and cell types). For example, the expression profile for caveolins from neonate, young and aged tissue varies significantly between heart and lung (Kawabe et al., 2001). Indeed, as noted in a recent review (Schilling and Patel, 2015), there appear key differences in age‐associated caveolae/caveolin expression between proliferative and terminally differentiated cell types. Assuming links between caveolae, cholesterol and fluidity, then one must consider organ‐ and cell‐specific responses when interpreting data on the effects of age and disease.

Ageing also results in important shifts in mitochondrial membrane lipid makeup that may be important to stress‐resistance (Paradies et al., 1992; Pepe et al., 1999). For example, advancing age is associated with loss of mitochondrial cardiolipin (Pepe et al., 1999), an inner mitochondrial membrane phospholipid required for effective electron transport and ATP synthesis. The high linoleic acid content of its acyl side chains renders it susceptible to oxidation (Shi, 2010), with depletion of linoleic acid associated with I‐R, heart failure and diabetes (Sparagna and Lesnefsky, 2009). Indeed, remodelling of cardiolipin may be critical in the aetiology of mitochondrial dysfunction in conditions characterized by oxidative stress (ageing, obesity/diabetes and heart failure) (Shi, 2010). In terms of ischaemic tolerance, cardiolipin regulates the activity of the mitochondrial permeability transition pore (MPTP), with its oxidation key to associated release of pro‐apoptotic cytochrome c and resultant cell death. Further detailed discussion of mitochondrial changes are beyond the scope of this review.

Early microscopic analyses identified ultrastructural changes to the cardiac sarcolemma with age, including the presence of large membrane‐bound spaces adjacent to or communicating with, gap regions of the intercalated disc (while desmosomes and fascia adherens were unaltered), together with enlargement and rounding of T‐tubules in the intercalated disc region. Subsequent biochemical analyses in rat hearts revealed shifts in fatty acid composition, including consistently increased saturated versus reduced polyunsaturated fatty acid (PUFA) contents (particularly within the two major phosphatidylcholine and phosphatidylethanolamine fractions), while cholesterol and total phospholipid levels were stable (Awad and Clay, 1982; Awad and Chattopadhyay, 1983). Studies in cultured cell models of ageing report increased sphingomyelin and cholesterol, and reduced phosphatidylcholine contents in both cardiac myocytes (Yechiel and Barenholz, 1985; Yechiel et al., 1985) and fibroblasts (Yechiel et al., 1986), consistent with reductions in membrane fluidity. In the cardiac membrane of miniature pigs ageing results in greater lipid peroxidation and significantly increased vitamin E content, with a differential increase in mid‐bilayer membrane fluidity versus decrease in polar head‐group domain fluidity (Banks et al., 1996). Examining more moderate ‘ageing’, Brochot et al. (2009) observed a significant increase in sarcolemmal stearic acid versus reduction in PUFA levels in 2‐ versus 6‐month rat hearts. Ageing is also linked with reduced phospholipid methylation (Heyliger et al., 1988), a process influencing mechanisms critical to stress‐resistance, including Ca²⁺ and Na⁺ handling (Panagia et al., 1986, 1987), and for which defects have also been observed in cardiac hypertrophy and different cardiomyopathies.

Shifts in sarcolemmal PUFA levels may contribute to alterations in electrophysiology, stress‐resistance and disease susceptibility with age (Pepe, 2007), with changes in mitochondrial membrane PUFAs also implicated in age‐related shifts in mitochondrial function (Pepe et al., 1999; Pepe, 2005). Because many PUFAs cannot be synthesized de novo in mammals, tissue levels are governed by dietary intake, and Western diets low in α‐linolenic acid and high in arachidonic acid may reduce n‐3 PUFA contents (exacerbated by reduced desaturase enzyme function). McLennan et al. (1989) presented evidence for increased n‐6 PUFA content of cardiac membranes, while n‐3 PUFA levels declined with ageing in rats, in association with abnormalities in Ca^{2 +} handling and arrhythmogenesis, while more moderate age differences (between 2‐ and 6‐months) were associated with reductions in both n‐3 and n‐6 PUFAs, with increased n‐6/n‐3 and decreased PUFA/saturated fatty acid ratios. Valencak and Ruf (2011) reported that dietary n‐3 and n‐6 PUFA enrichment enhances cardiac monounsaturated and PUFA levels in aged hearts with minimal changes in saturated fats; however, these profiles were not compared with those for young hearts.

Ageing also modifies myocardial caveolar microdomains. We recently identified depression of caveolin‐3 expression and caveolar density in aged murine myocardium (Peart et al., 2014), consistent with evidence of reduced caveolin‐3 in middle‐aged, compared with young, rat hearts (Kawabe et al., 2001), and disrupted caveolin‐3 localization in senescent rat heart (Ratajczak et al., 2003). Skeletal muscle caveolin‐3 is also depressed with ageing (Barrientos et al., 2015). However, it is important to note that the age‐dependence of caveolin expression is highly tissue‐specific, as is caveolin function. For example, in vitro studies demonstrated a role for elevated caveolin‐1 in senescence of replicating cells (Park et al., 2002), while in the brain ageing results in reductions in caveolin‐1, with its re‐expression reversing neurodegeneration (Kawabe et al., 2001; Head et al., 2010; Mandyam et al., 2015). Such findings highlight the importance of assessing the expression and roles of caveolins in vivo, within specific tissue types.

Effects of diabetes and metabolic disorders

Studies confirm shifts in plasma membrane structure and function in diabetic patients and also in animal models. Alterations in fatty acid and cholesterol profiles of erythrocytes and leukocytes from diabetic patients, including increased cholesterol and saturated fatty acid levels and reduced PUFA levels (Barnes et al., 1977; Garnier et al., 1990; Bakan et al., 2006), generally favour reductions membrane fluidity. In terms of cardiac membranes, Huang et al. (1984) found that diabetes in rats increased myocardial proportions of linoleic and dihomo‐γ‐linolenic acid acids, whereas decreased arachidonic acid levels were decreased. Gudbjarnason et al. (1987) also observed declining fatty acid desaturation (particularly altering the composition of phosphatidylcholine) in type I diabetic (T1D) rat hearts, together with elevations in cardiac triglycerides and linoleic acid. More recently, Wang et al. reported reductions in free cholesterol in several tissues in a rat T1D model, including heart (2012b). Diabetes has also been associated with mechanistically relevant changes in N‐methylation of cardiac membrane phospholipids (Panagia et al., 1990).

The handful of studies assessing myocardial caveolae and caveolins in diabetes, almost exclusively streptozotocin (STZ) dependent rat models of T1D, generally support hyperglycaemic depression of caveolin‐3 (Penumathsa et al., 2008a; Sharma et al., 2011; Lei et al., 2013) along with augmentation of caveolin‐1 (Penumathsa et al., 2008b; Lei et al., 2013). This suppression of caveolin‐3 involved excess PKCβ2 activity (Lei et al., 2013) and was recently validated by the beneficial effects of PKCβ2 inhibition on cardiac caveolin‐3 expression (and Akt signalling) reported in a T1D model (Liu et al., 2015). More recently, Li et al. (2015a) report reduced cavelin‐3 expression in T1D, which may disrupt protective adiponectin/STAT signalling to block postconditioning. A study of cardiac caveolin in the non‐obese Goto‐Kakizaki rat model of type II diabetes (T2DM), limited to mRNA expression, reported a non‐significant 30% depression of Cav3 transcript (Salem et al., 2013). Our recent studies have found more substantial (40–50%) depression of cardiac caveolin‐3 protein in a murine model of T2DM (unpublished data).

Other relevant diseases and metabolic disorders are likely to alter cardiac sarcolemmal structure. For example, hypertension has been shown to influence membrane makeup (Zicha et al., 1999), although specific effects on the cardiac sarcolemma remain to be detailed. Hypertrophy, heart failure and chronic hypoxia were all associated with reductions in caveolar proteins caveolin‐1 and ‐3 (Hare et al., 2000; Shi et al., 2000; Piech et al., 2002), whereas elevations in caveolin‐1 and ‐3 were observed in pressure‐overloaded and failing hearts (Uray et al., 2003; Kikuchi et al., 2005).

Effects of ischaemia and reperfusion

In addressing the effects of the sarcolemmal changes on I‐R tolerance and cardioprotection, it is important to appreciate the consequences of I‐R itself on the membrane. Ischaemia and reperfusion both profoundly influence the contents and organization of sarcolemmal cholesterol, phospholipids and lipid microdomains in reversibly and irreversibly injured tissue. Studies of ex vivo myocardium and myocytes indicate loss of sarcolemmal integrity after ~20 min of ischaemia/anoxia. However, membrane architecture is substantially modified prior to overt disruption, including loss of membrane lipids and aggregation of intra‐membrane protein particles, preceded by more subtle shifts in phospholipid distribution and lateral phase separation, and sarcolemma detachment from the sub‐sarcolemmal lattice and cytoskeleton (Post et al., 1985; Musters et al., 1993; Post et al., 1995a, 1995b).

Altona and Van Der Laarse showed in 1982 that anoxia and substrate deprivation reduced cardiac sarcolemmal cholesterol (prior to changes in membrane proteins), consistent with subsequent observations (Bester and Lochner, 1988). Ischaemia appears to redistribute cholesterol from sarcolemma and sarcoplasmic reticulum (SR) to mitochondria, resulting in a reduction in sarcolemmal fluidity (Rouslin et al., 1980, 1982; Venter et al., 1991). On the other hand, Netticadan et al. (1997) reported no change in canine cardiac sarcolemmal cholesterol levels with ischaemia, while phospholipids (primarily phosphatidylcholine and phosphatidylethanolamine) were significantly depleted. They also observed a general increase in fatty acid unsaturation following myocardial ischaemia. Reasons for these differing observations are unclear.

Cholesterol‐rich caveolar microdomains appear to be significantly modified by I‐R. Ballard‐Croft et al. (2006) report redistribution of caveolin‐3, and reductions in caveolin‐1 and cholesterol levels in cardiac membrane fractions. Chaudhary et al. (2013) found that I‐R depleted myocardial caveolae (and caveolin‐1) without substantially altering caveolin‐3 expression, while Gao et al. (2014) also reported reduced caveolin‐1 levels in infarcted myocardium and hypoxic fibroblasts. These observations are consistent with caveolin and caveola depletion with ischaemic insult in other organs (e.g. Zager et al., 2002; Kang and Lee, 2014).

Several studies also reveal ischaemia‐dependent reductions in sarcolemmal phospholipid levels and saturation (phosphatidylcholine and phosphatidylethanolamine primarily modified) (Vasdev et al., 1980; Bester and Lochner 1988), together with disruption of phospholipid distribution (Musters et al., 1993; Post et al., 1993; Post et al., 1995a, 1995b). Loss of cardiac phospholipids and accumulation of phospholipid degradation products are also documented during reperfusion (Van Der Vusse et al., 1992). Mitochondrial membrane phospholipid composition and microviscosity are also significantly modified with ischaemia (Victor et al., 1985). These reductions in membrane phospholipid content with I‐R may involve both enhanced degradation (e.g. via phospholipase activities) together with impaired resynthesis. De novo synthesis of phosphatidylcholine (Lochner and De Villiers, 1989) and reacylation of lysophospholipids (Kajiyama et al., 1987) are depressed in ischaemic myocardium, and myocardial capacity for phospholipid synthesis and reacylation is depressed in reperfused tissue (Das et al., 1986; Otani et al., 1989).

These sarcolemmal changes may not only limit intrinsic I‐R tolerance but also impair the initiation and transduction of cardioprotective signals, whether triggered via exogenous or intrinsic protective stimuli. There is some evidence for derangement of cardiac GPCR signalling with age can be countered by targeting membrane lipids ( see Moscona‐Amir et al., 1989). While prolonged ischaemia (and reperfusion) exert detrimental effects on the cardiac sarcolemma, milder forms of ischaemic stress may actually benefit the heart through influences on membrane domains: preconditioning may involve ROS‐dependent glucose transporter type 4 (GLUT‐4) translocation and increased expression of caveolin‐3 (and phosphorylated endothelial NO synthase (eNOS) and Akt) along with decreased expression of caveolin‐1 (Koneru et al., 2007). The following section details the potential influences of age/disease‐dependent shifts in cholesterol, caveolae/caveolin‐3, phospholipids, PUFAs and protein lipidation on cardiac stress signalling and cardioprotection.

Sarcolemmal makeup and cardioprotective signalling

Ischaemic preconditioning and postconditioning responses, and responses to cardioprotective adenosine, opioids, bradykinin, catecholamines and growth factor ligands, are all initiated by sarcolemmal GPCRs and growth factor RTKs. The conformation, stability, dimerisation and effector coupling of these receptors are strongly dependent upon membrane lipid composition and regulated by both shifts in the membrane environment and their own lipidic modifications (Mitchell et al., 2003; Inagaki et al., 2012; Mahmood et al., 2013).

Membrane cholesterol and protective signalling

Amphiphilic cholesterol is essential to optimal membrane structure and function (Lee, 2004) (Figure 1). In terms of cardioprotective signalling, the localization and interaction of protective transmembrane GPCRs with high local levels of cholesterol and glycosphingolipids critically influences receptor conformation and function (Burger et al., 2000). Cholesterol directly affects receptor proteins via molecular interactions (Gimpl et al., 2002), and indirectly affects receptor function via modulation of the physico‐chemical properties of the membrane bilayer (Yeagle, 1985; Ohvo‐Rekilä et al., 2002; Lee, 2004; Bandari et al., 2014).

Figure 1.

Diagram outlining hypothetical changes that may occur in the sarcolemma in states of pathological depletion of cholesterol. This figure is adapted and modified from Stary et al., 2012; Balse et al., 2012; and Campbell and Reece, 2007. Panels A, C, E and G show physiological sarcolemma structure, with ion channels, GPCRs and caveolae respectively. Ordering of phospholipids (B) between ‘tight’ and ‘loose’ ordering may depend on tissue type, temperature and saturation of phospholipid tails in states of significantly reduced membrane cholesterol. With ion channels (D), reduction in membrane cholesterol may induce compression and/or bending of the surrounding sarcolemmal structure. This may induce a more open probability of the ion channel. GPCR derangements (F) are likely to be affected by the direct modification of the sarcolemma (potentially differing under different states). These derangements are likely to alter receptor binding, including modification of allosteric sites. Reduced cholesterol appears to reduce caveolae expression in the myocardial plasma membrane (H). Given the role of the caveolae as a signalling scaffold (G), including the caveolin oligomers (which possess specific caveolin scaffolding domains, the binding site for numerous pro‐survival and pro‐growth molecules, such as PKA, PKC, AC, MAPKs, Src family kinases and endothelial NOS), loss of caveolae can significantly change pro‐survival and pro‐growth kinase signalling.

Through modifying sarcolemmal fluidity and compartmentation, membrane cholesterol (which reduces fluidity via increasing phospholipid packing and lateral assembly) and phospholipid fatty acids (the unsaturation of which generally increases fluidity) govern the probability of membrane receptor‐effector interactions. Cholesterol also appears to stabilize GPCRs and promote functionally crucial dimerisation (Mondal et al., 2014; Sengupta and Chattopadhyay, 2015), while directly interacting with some GPCRs. For example, almost half of family A GPCRs contain consensus cholesterol binding sites (Oates and Watts, 2011). Recent studies confirm the regulatory importance of direct cholesterol‐GPCR interactions (Khelashvili et al., 2010; O'Malley et al., 2011; Oates et al., 2012; Prasanna et al., 2014; Wu et al., 2014; Mondal et al., 2014; Sengupta and Chattopadhyay, 2015). Cholesterol oxidation may also determine the pathway (caveolar or clathrin dependent) of GPCR internalization (Okamoto et al., 2000). Cholesterol additionally influences protective epidermal growth factor receptor (EGFR) signalling through modulation of membrane fluidity and compartmentation, and formation of caveolar domains critical in spatio‐temporal control of EGFR signalling (Chen and Resh, 2002; Westover et al., 2003; Pike, 2005; Balbis and Posner, 2010).

Shifts in lateral mobility and organization of membrane components are reported for cardiomyocytes aged in culture (Yechiel et al., 1985), and aged cardiac fibroblasts (Yechiel et al., 1986), while Banks et al. (1996) report differential shifts in mid‐bilayer as distinct from polar head‐group domain fluidity in myocardial membranes from aged rats. As already noted, ageing‐dependent reductions in membrane fluidity correlate with shifts in cholesterol or phospholipid content and distribution (Yechiel et al., 1985; Alvarez et al., 1993; Choe et al., 1995; Igbavboa et al., 1996; Hashimoto et al., 2005).

Age‐related reductions in membrane fluidity generally correlate with alterations in cholesterol or phospholipid content (Alvarez et al., 1993; Choe et al., 1995; Hashimoto et al., 2005) and distribution (Igbavboa et al., 1996), although ageing has been linked with other relevant changes including cell‐specific shifts in phospholipid content and membrane asymmetry (Noble et al., 1999), reductions in n‐3 PUFA content and n‐3, relative to n‐6, PUFA levels (Hashimoto et al., 2001), and increased oxidative modification (Sawada et al., 1992; Yu et al., 1992; Sawada et al., 1993; Choe et al., 1995).

Reduced membrane fluidity is also widely observed in animal models of diabetes (Masuda et al., 1990), and cardiac sarcolemma from diabetic rats exhibit reduced fluidity in association with decreased Ca^{2 +} influx and Na⁺/K⁺‐ATPase activity (Ziegelhöffer et al., 1996, 1997; Ziegelhöffer‐Mihalovicová et al., 2003). Reductions in fluidity have been linked to impaired receptor‐G protein coupling (Gurdal et al., 1995), although age‐related shifts in receptor and G protein coupling in other tissues are inconsistent with a primary role for altered membrane fluidity (Fraeyman et al., 1993).

Membrane cholesterol also strongly influences ion channel function and ionic homeostasis in myocytes, which is highly relevant to I‐R outcomes and cardioprotection. Early work from Kutryk and Pierce (1988) showed that Na⁺‐Ca^{2 +} exchange in cardiac sarcolemmal vesicles is enhanced and reduced via cholesterol enrichment and depletion, respectively; intracellular Ca^{2 +} levels increase with cardiomyocyte cholesterol content (Bastiaanse et al., 1994); and hypercholesterolaemia reduces Na⁺ current magnitude and the threshold for excitation in cardiomyocytes (Wu et al., 1997). While the last study reported some similarity between effects of ageing and hypercholesterolaemia, the cholesterol‐dependent increase in excitability was distinct from age changes. More recent studies confirm and clarify the effects of cholesterol on ion handling (Balse et al., 2012; Dopico et al., 2012; Rosenhouse‐Dantsker et al., 2012). Membrane components critical to electromechanical coupling may also be cholesterol‐dependent, including T‐tubule formation (Carozzi et al., 2000), and gap junctions and connexin‐43 (Zou et al., 2014)

Myocardial I‐R tolerance does appear to be significantly cholesterol dependent, beyond the pro‐atherosclerotic influences of cholesterol. Bastiaanse et al. (1994) found that tolerance of cardiomyocytes to anoxia was linked to sarcolemmal cholesterol content, and subsequent studies confirm the effects of cholesterol on myocardial stress‐resistance. We recently reported that cholesterol depletion with methyl‐β‐cyclodextrin significantly impairs ischaemic tolerance in the murine heart and alters cardiac function under normoxic conditions (See Hoe et al., 2014). This contrasts with reports that cholesterol depletion disrupts cardiac architecture and abolishes preconditioning without altering contractile function or intrinsic stress‐resistance (Das et al., 2008, Sun et al., 2012). On the other hand, acute hypercholesterolaemia (3–4 day cholesterol‐enriched diet) increases myocardial infarct size (Golino et al., 1987; Hoshida et al., 1996), and longer term high‐cholesterol feeding also worsens myocardial I‐R tolerance and infarction (Ma et al., 1996; Jung et al., 2000; Xu et al., 2013). Dual ApoE/LDL receptor knockout mice fed atherogenic diets also exhibit worsened I‐R tolerance (Li et al., 2001). Nonetheless, not all studies report repression of myocardial I‐R tolerance with hypercholesterolaemia, and there is some evidence for differences between acute and chronic effects (Girod et al., 1999; Jones et al., 2001; Lauzier et al., 2009). For example, the recent study by Haar et al. (2014) report that acute consumption of a high‐fat diet preceding ischaemia induced a cardioprotective state associated with NF‐κB‐dependent modulation of autophagy and apoptosis.

Several forms of cardioprotection appear to be cholesterol‐sensitive. Hypercholesterolaemia impairs preconditioning responses to ischaemia, pacing and anaesthetic (Szilvassy et al., 1995; Ferdinandy et al., 1997, 2003; Tang et al., 2005; Görbe et al., 2011; Zhang et al., 2012; Xu et al., 2013), together with ischaemic postconditioning (Lauzier et al., 2009). These inhibitory effects of hypercholesterolaemia have been linked to impairment of NO‐cGMP signalling (Ferdinandy et al., 1997; Howitt et al., 2012), reduced expression of protective heat shock proteins (Csont et al., 2002) and altered sarcolemmal and mitochondrial distribution of connexin‐43 (Görbe et al., 2011). Repression of cardioprotective signalling may also reflect the specific importance of cholesterol to caveolae and caveolin‐dependent cell signalling.

Caveolae, caveolins and protective signalling

Caveola microdomains and associated caveolins are critical determinants of signalling via protective GPCRs and EGFR and influence other relevant processes including ion channel function, insulin signalling and substrate metabolism. Caveolae are relatively rich in cholesterol (Pike et al., 2002), which is essential to their formation, stability and functionality (Rothberg et al., 1992; Pol et al., 2005). Anderson and Jacobson (2002) propose a ‘lipid shell’ model in which caveolar proteins are encased in shells of cholesterol and other lipids. Caveolin binds cholesterol at a 1:1 ratio with high affinity (Murata et al., 1995), with high caveolar cholesterol likely to stem from high levels of oligomeric caveolin complexes. Cholesterol is essential to formation of stable caveolar domains, although not for transport of caveolins to the plasma membrane. Work in MDCK cells shows cholesterol levels strongly influence caveolar synthesis and caveolins and supports a threshold effect whereby caveola formation only occurs when cholesterol levels are ≥50% of normal (Hailstones et al., 1998). Caveolae are, in turn, important in cellular cholesterol transport, although hypercholesterolaemia and LDL exert negative effects on caveola and caveolins. For example, elevated cholesterol disturbs inhibitory control of vascular adhesion via caveolae and caveolin‐1 (Fu et al., 2010), while oxidized LDL (oxLDL) may induce translocation of caveolin‐1 and eNOS from caveolae to suppress NOS activity (Blair et al., 1999; Shaul, 2003). Caveolin‐1 inhibits NOS, and increased co‐localization (whether within or outside of caveolae) is predicted to suppress NOS signalling. Interestingly, serum isolated from hypercholesterolaemic patients has been shown to promote the caveolin‐1/eNOS interaction (Feron et al., 1999). Conversely, treatment with statins may deplete myocyte caveolin‐3, caveolae and cholesterol (Pugh et al., 2014), potentially influencing protective signalling.

Myocyte‐specific caveolin‐3 is essential to diverse protective responses (Table 1) and maintains mitochondrial integrity (Fridolfsson et al., 2012; Wang et al., 2014a) and cardiac resistance to I‐R injury (See Hoe et al., 2014). These dependencies may involve influences of caveolin‐3 on protective GPCR or EGFR function and survival kinase and NOS activities (Roth and Patel, 2011; Fridolfsson et al., 2014; Schilling et al., 2015), together with death receptor and caspase‐3 signals (Sharma et al., 2011; Carotenuto et al., 2013). Additionally, caveolae function as Ca^{2 +} signalling microdomains (Pani and Singh, 2009) and govern the functionality of important ion channels (Patel et al., 2008). Recent work also reveals a role for caveolin‐1 in gap junction homeostasis and arrhythmogenesis (Yang et al., 2014).

The influence of caveolae (and caveolins) on the functionality of protective GPCR signalling is highly variable. Signalling via the β₂‐adrenoceptor, attributed with protective actions and potential roles in cardiac conditioning, is caveola‐dependent. Studies in HEK293 cells support differential compartmentation of β₂‐adrenoceptors to non‐caveolar domains and localization of G proteins and adenylate cyclase to caveolar domains (potentially limiting basal activity and β₂‐adrenoceptor signalling responses) (Pontier et al., 2008). However, studies in ventricular myocytes support co‐localization of β₂‐adrenoceptors, Gα_s, L‐type Ca^{2 +} channels, PKA and cardiac type V/VI AC within cholesterol‐rich caveolar domains (Rybin et al., 2000; Ostrom et al., 2001; Steinberg and Brunton, 2001; Xiang et al., 2002), with receptor activity triggering β₂‐adrenoceptor translocation out of this compartment (Rybin et al., 2000; Ostrom et al., 2001; Steinberg and Brunton, 2001). This localization of β₂‐adrenoceptors and downstream signal elements suggests caveolae may function as signal platforms to spatially limit and control β‐adrenoceptor‐dependent cAMP signals, explaining the apparently greater ‘efficiency’ of β₂‐adrenoceptors over non‐caveolar β₁‐adrenoceptor signalling, and constraining β₂‐adrenoceptor‐dependent cAMP signalling to membrane events while avoiding non‐selective effects on the SR or myofilaments (Rybin et al., 2000). Indeed, an enhanced caveolar density of β₂‐adrenoceptors promotes AC activation, dependent on these receptors (Xiang et al., 2002), while caveolar disruption impairs localized β₂‐adrenoceptor control of AC activity and cAMP, leading to non‐selective modulation of SR and myofilament function (Calaghan et al., 2008).

The cardioprotective δ‐opioid receptor appears to possess a caveolin binding motif, although has been localized to lipid‐poor fractions in HEK293 cells (Levitt et al., 2009). Nonetheless, analysis of cardiac tissue supports localization of protective δ‐opioid receptors with caveolin‐3 and caveolae (Patel et al., 2006), and both cholesterol depletion and caveolin‐3 knockout negate the cardioprotective effects of opioid receptor agonists (Tsutsumi et al., 2010; See Hoe et al., 2014). This caveolar dependence may involve co‐localization of receptor and signal elements within these domains (Head et al., 2005; Patel et al., 2006). The protective adenosine A₁ receptor is also implicated in conditioning responses and intrinsic I‐R tolerance (Cohen and Downey, 2008; Headrick et al., 2013), and Lasley et al., report that a major fraction of myocyte A₁ receptors reside within caveolar domains in un‐stimulated cells and translocate to non‐caveolar fractions upon stimulation (Lasley et al., 2000). They later presented evidence of A₁ receptor‐dependent shifts of MAPKs from cardiomyocyte caveola domains (Ballard‐Croft et al., 2008). Conversely, Yang et al. (2009) report A₁ receptor‐mediated translocation of PKC isoforms into caveola in cardiac myocytes. These data collectively suggest complex caveola‐dependence of A₁ receptor signalling, with evidence for differential translocation of activated kinases into and out of these domains in cardiac myocytes.

The functionality of protective EGFR signalling may also be influenced by caveolar proteins; however, the specific roles of caveolae in governing EGFR signalling remain unclear (Balbis and Posner, 2010). Early evidence suggested that membrane domains promote efficient EGFR signalling, in part because of co‐localization of effector molecules (Chen and Resh 2002; Pike, 2005). Disruption of the microdomains results in a shift of EGFR to bulk membrane regions, impairing EGFR activation (Pike and Casey, 1996, Pike and Miller, 1998; Chen and Resh, 2002; Ringerike et al., 2002; Roepstorff et al., 2002; Westover et al., 2003). Caveolin‐1 has also been implicated in the negative regulation of EGFR function, potentially contributing to age‐dependent shifts in EGFR signalling and cellular replicative senescence (Park et al., 2002). The cardiac EGFR has been shown to mediate cardioprotection triggered by caveola‐dependent A₁ receptors and ischaemic preconditioning (Williams‐Pritchard et al., 2011).

Downstream signalling relevant to myocardial I‐R tolerance/cardioprotection may also be dependent upon the lipid micro‐environment of cardiac caveolae. In studies of non‐cardiac cells, Pike and co‐workers found a major fraction of phosphatidylinositol‐4,5‐bisphosphate (PIP₂) is located within caveolin‐rich buoyant membrane fractions (Hope and Pike, 1996; Pike and Casey, 1996; Liu et al., 1998), with this pool specifically sensitive to EGFR and GPCR activity (sensitivities negated by caveolar disruption) (Pike and Miller, 1998). In cardiac myocytes, α‐adrenoceptor‐dependent PLC activity has also been shown to specifically regulate caveola PIP₂ levels (Morris et al., 2006).

A number of important ion channels specifically localize with cardiac sarcolemmal caveola: the L‐type Ca^{2 +} channel Ca_v1.2 (Balijepalli et al., 2006; Shibata et al., 2006), the voltage‐gated Na⁺ channel Na_v1.5 (Yarbrough et al., 2002), HCN4 pacemaker channels (Ye et al., 2008), and potentially the Na⁺/Ca^{2 +} exchanger NCX1 (Bossuyt et al., 2002), the K_v1.5 K⁺ channel (responsible for the ultra‐rapid component of the delayed rectifier K⁺ current, I _Kur), the K_v7.1 pore forming subunit of the K⁺ channel (responsible for the slow component of the delayed rectifier current, I _Ks) and the K_ir6.2/SUR2A components of the sarcolemmal K_ATP channel (Garg et al., 2009). Molecular complexes regulating the function of these channels are also co‐localized within caveolar domains. Thus, caveola L‐type Ca^{2 +} channels may form a localized signalling complex with caveolin‐3, β₂‐adrenoceptors, AC, Gα_s, PKARII and PP2A that is essential to β₂‐adrenoceptor control of cardiomyocyte channel activity (Balijepalli et al., 2006; Tsujikawa et al., 2008).

Importance of caveolar changes in ageing

That age‐dependent depletion of cardiac caveolin‐3 and caveola may underlie dysfunctional cardioprotection, and I‐R intolerance is internally consistent with failure of caveolin‐3 dependent responses and exaggerated I‐R injury in aged tissue from humans and rodents (Peart et al., 2014). In contrast, a novel stimulus shown to be caveolin‐3 independent (See Hoe et al., 2014) is uniquely effective in aged heart (Peart and Gross, 2004), suggesting a potential strategy of ‘caveolin‐3 independent’ stimuli. Intriguingly, cardiac ageing itself may be governed by caveolin‐3, with declining stress‐resistance a defining feature of ageing, and several mechanisms influencing ageing, such as mitochondrial dysfunction and oxidative stress, autophagy, mitochondrial fusion : fission, and EGFR, mTOR and Akt signals, all being caveolin‐3 dependent (Fridolfsson et al., 2012; Fridolfsson et al., 2014). Caveolin‐3 depletion may thus limit cardioprotection and stress‐resistance and accelerate the process of cardiac ageing. Such a role is consistent with the so‐called ‘Green Theory’ of ageing (Gems and McElwee, 2005).

Importance of caveolar changes in diabetes

Shifts in membrane caveolae and caveolins have been implicated in tissue‐specific sequelae of diabetes. For example, enhanced fibroblast caveolin‐1 may underlie impaired wound healing (Bitar et al., 2013), while depressed endothelial caveolin‐1, caveolar disruption and eNOS uncoupling may contribute to coronary vascular dysfunction in diabetes (Cassuto et al., 2014). Such findings again highlight the tissue specificities of age or disease effects on caveolins. Relatively little is known regarding mechanistic involvement of caveolins in the cardiac abnormalities of diabetes, with data deriving solely from rat models of T1D and no studies yet undertaken in T2DM models. Certainly, the effects of caveolins and caveolae on insulin receptor signalling and glucose handling in other tissues (Cohen et al., 2003; Ishikawa et al., 2005; Strålfors, 2012) suggest involvement of caveolin changes in progression of the disease. Potential roles for caveolin‐3 in T1D‐dependent diastolic dysfunction (Lei et al., 2013), and impaired GLUT4 translocation (Penumathsa et al., 2008a) have been revealed in these rat studies. In terms of cardioprotection, one study indirectly implicates a role for elevated caveolin‐1 expression (Ajmani et al., 2011), although caveolin levels were not measured and failure of preconditioning not confirmed (with omission of non‐preconditioned diabetic comparators). On the other hand, Li et al. (2015a recently presented evidence for depression of caveolin‐3 in T1D, which may impair cardioprotective signalling and postconditioning via disruption of adiponectin receptor and STAT3 signalling. Caveolae and caveolins not only regulate protective signalling, glucose transport and insulin signalling but may also influence hyperglycaemia‐dependent cell death. Hyperglycaemia suppressed macrophage caveolin‐1 expression and relocation of NADPH oxidase from cytosol to caveolae where it promoted superoxide formation and cellular damage (Hayashi et al., 2007). Whether caveola play direct roles in specific forms of cardiomyocyte or fibroblast death remains to be clarified.

Influences of oxLDL and saturated fatty acids on caveolae

Cholesterol oxidation may lower fibroblast caveolar density and translocate caveolin‐1 from membrane to Golgi (Smart et al., 1994), with studies in endothelium also demonstrating reductions in membrane caveolin‐1 (and reduced NOS localization and activity) in response to oxLDL (Blair et al., 1999; Uittenbogaard et al., 2000). Endothelial surface expression of the lipid raft marker GM1 is also decreased by oxLDL (Byfield et al., 2006). Oxidized phospholipids can also deplete membrane cholesterol (Yeh et al., 2004). In contrast, HDL exposure preserves caveolar cholesterol in oxLDL treated endothelium, suggesting the vascular effects of altered LDL : HDL ratios could involve caveolar modulation. Functionally, oxidation of membrane cholesterol has been shown to modify ionic interactions with the sarcolemma (Kutryk et al., 1991), reducing Na⁺‐Ca^{2 +} exchange (and to a lesser extent ATP‐dependent Ca^{2 +} uptake), while passive Ca^{2 +} efflux was greatly enhanced.

There is also evidence that saturated fats lead to depletion of myocyte (Knowles et al., 2011) and myocardial (Knowles et al., 2013) caveolin‐3. Conversely, n‐3 PUFAs may preserve caveolin‐3 levels and cardiac stress‐resistance (Carotenuto et al., 2013). Dietary PUFAs exert significant effects on the makeup of lipid rafts and caveolae in multiple cell types, reducing cholesterol and sphingomyelin content and altering caveolins and caveolar function (Ma et al., 2004a; Chapkin et al., 2008a, 2008b). However, docosahexaenoic acid (DHA; the longest and most highly unsaturated fatty acid) has been shown to exclude EGFR from caveolin‐rich lipid raft fractions, decreasing association with effector molecules and thus ERK activation (Rogers et al., 2010). Others confirm DHA impairs EGFR transactivation and alter internalization or degradation as a result of changes in lateral and subcellular localisations of EGFR (Turk et al., 2012).

Other proteins critical to caveola formation may also shift in response to cholesterol feeding or dyslipidemias. Uyy et al. (2013) showed that high‐fat feeding in ApoE knockout (KO) mice increases lung endothelial caveolin‐1 (and phospho‐Akt), while essential cavin‐1 declines (suggesting a cavin‐1 dependent reduction in caveolae and associated signalling, despite elevations in caveolin‐1). High‐fat feeding in ApoE KO mice also results in significant shifts in vascular caveolar cholesterol, potentially in both endothelial and smooth muscle (Kincer et al., 2002). Further work is warranted to assess possible effects on myocardial membranes.

Phospholipids and cardioprotection

The levels and physico‐chemical properties of sarcolemmal phospholipids are also important determinants of myocyte I‐R damage (Verkleij and Post, 1987). For example, improved tolerance to ischaemia and metabolic inhibition is observed in myocytes possessing higher ratios of bilayer to non‐bilayer preferring phospholipids (Post et al., 1995a, 1995b). Optimal functionality of protective receptors depends on the lipid environment, and studies confirm important influences of phospholipids on GPCR signalling (Mitchell et al., 2003; Alves et al., 2005). Receptor activation is modulated for example by the presence of phosphatidylethanolamine (Alves et al., 2005), while phospholipids with unsaturated fatty acids have been shown to optimize rhodopsin signalling (Mitchell 2003), and negatively charged lipids also modulate G protein coupling for non‐visual GPCRs (Inagaki et al., 2012). Other work shows GPCR interactions with membrane lipids can increase phospholipid motion and decrease phospholipid order, suggesting the membrane bilayer may not only influence receptor function but also operate as a mechano‐chemical mediator of GPCR signalling (Tiburu et al., 2011).

Initiation of GPCR deactivation via GPCR kinase (GRK) mediated phosphorylation is additionally regulated by phospholipids. Recruitment of GRKs to lipid bilayers and their activation depend upon specific anionic phospholipids (Homan et al., 2013). While GRKs possess common regulator of G protein signalling homology and kinase domains, differing C‐terminal regions uniquely govern membrane recruitment. Thus, regulation of the GRK1 subfamily involves farnesylation or geranylgeranylation, while domains in the GRK2 subfamily interact with heterotrimeric Gβγ subunits and phospholipids. Activation by Gβγ is dependent on anionic phospholipids, particularly PIP₂. Other negatively charged phospholipids, such as cardiolipin, phosphatidylglycerol, phosphatidylserine and phosphatidylinositol, may promote GRK2 activity at high concentrations, although only phosphatidylserine and PIP₂ are likely to be present at levels appropriate to physiological control. In contrast, Gβγ‐dependent activity is unaltered by uncharged lipids such as phosphatidylcholine and phosphatidylethanolamine (Homan et al., 2013). The GRK4 subfamily members possess a positively charged amphipathic helix in the C‐terminal region, with some also containing palmitoylation sites, and membrane interactions appear to be PIP₂‐dependent (Pitcher et al., 1996). Membrane recruitment is governed by electrostatic interactions between anionic phospholipids and positively charged zones in both N‐ and C‐termini, with palmitoylation contributing in some cases.

Membrane phospholipid changes may also govern cellular insulin sensitivity and glucose uptake, which are crucial to cellular responses to I‐R. For example, levels of long‐chain PUFAs within skeletal muscle membranes correlate with insulin sensitivity (Borkman et al., 1993), suggesting a shift to unsaturated phospholipids enhances membrane GLUT4 transporter density. Increased plasma free fatty acids in diabetes could thus promote a shift from unsaturated to saturated fatty acyl chains of membrane phospholipids, increasing membrane stiffness and reducing fluidity. This, in turn, may limit membrane insertion of insulin‐independent GLUTs and fusion of insulin‐dependent GLUT4 containing vesicles with the membrane. Manipulation of the fatty acid composition of cell membranes does significantly modify insulin responsiveness (Grunfeld et al., 1981; Ginsberg et al., 1987). Effects specific to myocardial tissue await investigation.

PUFAs and cardioprotection

Cell membrane PUFA content shifts with both ageing and disease and is sensitive to dietary PUFA intake (Pepe 2007; Borsonelo and Galduróz, 2008). In the heart, dietary PUFA supplementation modifies sarcolemmal, mitochondrial and SR membrane makeups (Swanson and Kinsella, 1986; Croset et al., 1989; Croset and Kinsella, 1989; Taffet et al., 1993; Pepe and McLennan 1996; Brochot et al., 2009; Giroud et al., 2013; Ting et al., 2015), with these changes affecting processes crucial to I‐R tolerance and cardioprotection, including Ca^{2 +} and Na⁺ fluxes, membrane ATPase activities (Croset et al., 1989; Croset and Kinsella, 1989 Goel et al., 2002), and other membrane currents (Verkerk et al., 2006). The cardioprotective effects of PUFAs may not only involve modulation of sarcolemmal ion channels but also shifts in membrane microdomains and associated receptor signalling together with inhibitory effects on inflammation and thrombosis (Adkins and Kelley, 2010).

Dietary PUFAs can uniquely modulate the biochemical makeup of lipid rafts and caveola microdomains (thus protective GPCR and EGFR signalling). For example, n‐3 PUFAs are incorporated into fatty acyl groups of caveolae phospholipids and reduce caveolar cholesterol and caveolin‐1 levels (Ma et al., 2004a, 2004b; Chapkin et al., 2008a; Turk and Chapkin, 2013). Chapkin et al. (2008b) found PUFA supplementation increased protein clustering in cholesterol‐dependent microdomains whereas non‐raft microdomains were insensitive to n‐3 PUFA. This selective remodelling of membrane microdomains, with reductions in cholesterol or sphingomyelin contents, may alter the assembly of a liquid‐ordered phase structure and caveolae protein localization/functionality, thereby modifying GPCR/EGFR signalling and ion handling. Curiously, studies employing quantitative imaging with polarity‐sensitive probes unexpectedly reveal (in non‐cardiac cell types) that n‐3 PUFA increases rather than decreases the molecular order of lipid microdomains (Kim et al., 2008, 2014; Rockett et al., 2012). The mechanisms by which n‐3 PUFA modifies molecular order and formation of cholesterol‐enriched microdomains remain to be more fully defined.

Modulation of sarcolemmal and mitochondrial membranes via PUFA incorporation does improve myocardial I‐R tolerance (Hock et al., 1987; McLennan et al., 1988; al Makdessi et al., 1995; Nageswari et al., 1999; Pepe et al., 1999; Pepe and McLennan 2002; Abdukeyum et al., 2008; Zeghichi‐Hamri et al., 2010), reduces injury during cardiac cold‐storage (Ku et al., 1999) and may also influence responsiveness to cardioprotective stimuli. Supplementation with PUFAs and resultant membrane modifications are associated with improved post‐ischaemic contractile recoveries (Pepe and McLennan, 2002), reduced cellular disruption and cardiac infarction (Hock et al., 1987; Abdukeyum et al., 2008; Zeghichi‐Hamri et al., 2010), improved mitochondrial energy metabolism (Demaison et al., 1994; Pepe et al., 1999) and a reduced propensity to MPTP activation (Khairallah et al., 2012; Galvao et al., 2013), and electrophysiological changes including protection against arrhythmogenesis with I‐R (McLennan et al., 1988; Pepe and McLennan, 1996; al Makdessi et al., 1995), atrial conduction slowing and reduced inducibility of atrial fibrillation (Kumar et al., 2013).

Myocardial PUFAs are also modified in diabetes, where there is evidence that PUFA supplementation may be beneficial. In rat models of STZ‐dependent T1D, n‐3 PUFAs improve cardiac function and inhibit myocardial damage, cardiomyopathy and accumulation of periodic acid Schiff‐positive glycoproteins (Black et al., 1989, 1993; Christopher et al., 2003). Dietary PUFA supplementation also improves glucose handling and insulin sensitivity in models of diabetes (Islin et al., 1991; Steerenberg et al., 2002; Liu et al., 2013), and elevated n‐3 PUFA status limits age‐ and obesity‐related insulin‐resistance (Romanatto et al., 2014), although the basis of these effects remains to be delineated. For example, improvements in models of obesity and diabetes have been linked with suppression of TLR4 activity and inflammatory signalling (Liu et al., 2013). Fish oil supplementation and membrane incorporation of n‐3 PUFAs also improve cardiac function and limit progression of hypertrophy in other settings (McLennan et al., 2007; McLennan et al., 2012).

Lipid modifications and protective GPCR signalling

Sarcolemmal GPCRs are subject to lipidic modification, including palmitoylation, myristoylation or isoprenylation (addition of farnesyl or geranylgeranyl isoprenoids) (Escriba et al., 2007; Vögler et al., 2008). Palmitoylation involves addition of palmitic acid at one or more cysteines, for example, to the intracellular side of GPCRs. Because the thioester bond linking palmitate to cysteine is readily cleaved, this allows for reversible regulation of receptor function. Palmitoylation can also occur on residues other than cysteine. Importantly, all aspects of GPCR signalling may be modified via palmitoylation (Escriba et al., 2007; Vögler et al., 2008): modification of carboxy‐terminal tail sites may select for specific G protein interactions (promoting biased agonism); receptor phosphorylation governing desensitization or internalization is sensitive to palmitoylation; and receptor trafficking may also be governed by palmitoylation independently of phosphorylation, for example, targeting some GPCRs to lipid rafts. Palmitoylation generally promotes membrane targeting and signalling of GPCRs (Escriba et al., 2007). Palmitoylation‐dependent receptor‐G protein interactions have been observed for β₂‐adrenoceptor and M₂ muscarinic cholinoceptors (O'Dowd et al., 1989; Hayashi and Haga, 1997), and there is evidence a cholesterol‐palmitoyl interaction facilitates receptor homodimerisation and G protein coupling for the μ‐opioid receptor (Zheng et al., 2012). While GPCR palmitoylation rate is itself controlled by receptor agonism, other mechanisms may contribute – high NO levels can reduce basal and agonist‐dependent palmitoylation of β₂‐adrenoceptors (Adam et al., 1999).

Effector G proteins are also influenced by lipids and lipidic modification (Escriba et al., 2007; Vögler et al., 2008). Thus, the physico‐chemical properties of the membrane modify G protein localization and function; membrane interaction of G proteins enables them to also modulate their lipid environment and influence binding; and G protein subunits undergo post‐translational lipid modifications to ensure appropriate membrane attachment and trafficking. The Gα subunits are modified within their N‐terminus by myristate and/or palmitate, while γ subunits of Gβγ dimers are isoprenylated within their C‐terminus regions (Escriba et al., 2007; Vögler et al., 2008). Studies of amino acid mutation to prevent such modifications confirm the essential role of these changes in G protein function and reveal impaired attachment to the cell membrane and redistribution to the cytosol. Trafficking is critically dependent on these modifications: isoprenylation of γ subunits of Gβγ dimers leads to endoplasmic reticular localization, which may be followed by G protein heterodimerization and palmitoylation to target the proteins to the plasma membrane (Escriba et al., 2007; Vögler et al., 2008).

The signalling function/activity of Ras GTPase proteins is also altered by plasma membrane lipid interactions and their own lipidic modifications (Vögler et al., 2008). Correct membrane targeting requires post‐translational modifications at the protein's C‐terminus: farnesylation and geranylgeranylation are both important in Ras targeting, and palmitoylation promotes plasma membrane localization. Different Ras isoforms interact in different ways with the plasma membrane: H‐Ras interacts transiently with lipid rafts when bound to GDP and accumulates within non‐raft domains; K‐Ras is clustered in cholesterol‐insensitive, non‐raft domains that differ from the activated H‐Ras domains. Recent work suggests Ras promotes the assembly/stability of signalling microdomains, and there is evidence that the palmitate anchors of Ras proteins facilitate formation of signalling clusters (Hancock and Parton, 2005). Shifts in the lipid environment thus have the capacity to modify cardioprotective GPCR signalling at the level of the receptors themselves (ligand binding, trafficking and coupling), together with shifts in G protein and GTPase function and localization.

The sarcolemma in signal bias and allosteric modulation with age and disease

As already detailed, changes in membrane lipid profiles, and thereby the local physico‐chemical environment, can subtly but importantly modify the protective functionality of GPCRs and RTKs. Membrane changes may act to effectively ‘bias’ receptor signalling and allosterically manipulate GPCR function. Transmembrane GPCRs are dynamic proteins, in flux between different conformations influencing cellular signalling activation (Kenakin and Miller, 2010). Newly emerging paradigms of allosteric modulation and biased agonism are revolutionizing modern drug design, exploiting the dynamic nature of GPCRs and stabilizing receptor conformations that yield specific receptor : ligand interactions (allosteric modulation) or effector modulation (biased agonism) (Keov et al., 2011). Since ligand interactions with orthosteric or allosteric sites may be differentially sensitive to membrane influences on receptor‐effector complexes, membrane changes with age or disease may promote or inhibit distinct subsets of GPCR‐mediated signal transduction (Leach et al., 2007; Mailman, 2007; Violin and Lefkowitz, 2007).

There is some evidence for the emergence of system bias with age or disease, although underlying mechanisms await delineation. For example, ageing has been shown to differentially inhibit the cytoprotective but not the electrophysiological actions of the myocardial A₁ receptors, without apparent changes in receptor transcription or expression (Headrick et al., 2003). Indeed, negative chronotropic actions of this receptor may actually be augmented with age (Hinschen et al., 2000). Such biased signalling effects could arise via shifts in GRK2 function, which as detailed above, is highly dependent upon specific anionic phospholipids. While A₁ receptor‐dependent inhibition of cAMP formation was insensitive to GRK2 overexpression, Gβγ‐dependent activation of MAPKs (implicated in stress signalling) was selectively suppressed (Iacovelli et al., 1999). Thus, age‐ or disease‐dependent shifts in GRK sub‐type expression and/or membrane levels of phosphatidylserine, PIP₂ or other negatively charged phospholipids could effectively bias protective GPCR function. Indeed, GRK and β‐arrestin1 are altered in congestive heart failure (Vinge et al., 2001), and GRK2/3 in the failing human heart (Ungerer et al., 1994). Studies in other tissues support age‐dependent shifts in GRK expression: GRK2 and 6 (and β‐arrestin‐2) are suppressed with age in the human brain (Grange‐Midroit et al., 2002); and age switches the liver isoform profile from GRK2 to GRK3 and down‐regulates β‐arrestin (Kim et al., 2009). In contrast, in diabetic mice, hyperinsulinaemia up‐regulates GRK (and down‐regulates β‐arrestin2) in vascular smooth muscle, inhibiting activation of the Akt/eNOS pathway involved in cytoprotection in other cell types (Taguchi et al., 2011).

Changes in membrane cholesterol and PUFA contents may also confer GPCR signalling bias via changes in receptor confirmation and effector interactions, as detailed earlier. Cholesterol may stabilize receptors in defined conformations relevant to select biological functions (Burger et al., 2000), modulate cell signalling through interaction with scaffold proteins (Sheng et al., 2012), influence GPCR localization, and modulate agonist‐induced trafficking and agonist‐dependent signalling through controlling recruitment of β‐arrestins (Qiu et al., 2011). Moreover, the compartmentalization of receptor and effector molecules in membrane microdomains such as caveolae (see above) may also promote biased signalling. The notion that membrane lipid changes may facilitate the allosteric modulation of GPCR function through altering the local environment has important implications for drug design and development of cardioprotective interventions. If the membrane environment in aged or diseased hearts does influence these processes then it is crucial to unravel these molecular mechanistic changes and strategically target GPCR conformations that can promote desired signalling within the context of altered sarcolemmal environments of aged or diseased hearts.

Strategic manipulation of membrane‐dependent protection

Given the considerable body of evidence supporting involvement of sarcolemmal and specific caveolar abnormalities in failure of conventional cardioprotective interventions, is it possible to specifically counter such changes to promote effective cardioprotection and/or develop interventions that effectively bypass such limitations?

Dietary manipulations have the capacity to influence protective machinery via shifts in membrane makeup and properties. Overall reductions in caloric intake are reported to improve preconditioning responses in aged human and animal myocardium (Long et al., 2002; Abete et al., 2002, 2005), and in diabetic mice (Van der Mieren et al., 2013), although underlying mechanism remain contentious. Nonetheless, Tamburini et al. (2004) have shown that caloric restriction appears to broadly counteract the age‐related changes in membrane lipids.

Dietary fats

Select manipulation of dietary fats has potential in modulating myocardial stress‐tolerance and cardioprotective efficacies. Certainly, saturated fat supplementation suppressed (Knowles et al., 2013) while n‐3 PUFAs preserved (Carotenuto et al., 2013) caveolin‐3 and stress‐resistance, supporting the potential of dietary manipulation of microdomain‐dependent responses. However, this requires further study. Abdukeyum et al. (2008) report lack of additivity between the protective effects of PUFAs and ischaemic preconditioning in young healthy hearts, while Hlavácková et al. (2007) observed improved anti‐arrhythmic effects of hypoxic preconditioning following PUFA supplementation. Interestingly, Esterhuyse et al. (2005a, 2005b) reported improved post‐ischaemic functional outcomes following red palm oil supplementation in association with improved PUFA content in normo‐cholesterolaemic animals. Thus, I‐R tolerance may be improved via increased membrane PUFA levels in control animals, although the authors report no change in post‐ischaemic PUFA content in hypercholesterolaemic rats (Esterhuyse et al., 2005b). Cardioprotection with red palm oil has also been linked with differential changes in pre‐ and post‐ischaemic MMP2 expression in hypercholesterolaemic rats (Szucs et al., 2011), with MMP2 known to co‐localize with caveolin‐1 and caveolin‐3 in the cardiac sarcolemma (Cho et al., 2007).

Resveratrol

Multi‐potent resveratrol improves cardioprotective efficacies and myocardial I‐R tolerance in models of ageing, diabetes and obesity (Thirunavukkarasu et al., 2007; Lekli et al., 2008; Zheng et al., 2015). While conventionally attributed with activation of anti‐ageing sirtuin‐1 (and mitochondrial sirtuin‐3), resveratrol induces other changes that may be relevant. Indeed, depression of sirtuin‐1 itself does not appear to underlie effects of age on cardioprotective preconditioning (Adam et al., 2013). Thus, resveratrol may restore protective responses via other mechanisms, including modulation of caveolar domains and associated signalling: Penumathsa et al. (2008a) report resveratrol induces cardioprotection in a rat model of T1D in association with improved caveolin‐1 and ‐3 expression, and caveolar localization of GLUT4. Additionally, resveratrol elevates myocardial haem oxygenase‐1 (HO‐1) and phospho‐eNOS expression in high cholesterol‐fed rats (Penumathsa et al., 2008b), with protective effects of HO‐1 induction also linked to modulation of caveolin‐1 and eNOS (Issan et al., 2014).

Exercise

Exercise not only induces powerful cardioprotection (Powers et al., 2004, 2014) but also improves preconditioning responses in aged patients and animal models. Abete et al. (2000, 2001) have demonstrated that exercise training partially restores, while tandem training and caloric limitation completely preserves ischaemic preconditioning in aged rats (Abete et al., 2005). Exercise has also been shown to restore cardioprotective signalling and preconditioning in obese mice, independently of associated co‐morbidities (Pons et al., 2013; Li et al., 2015b). Whether these exercise effects involve sarcolemmal modulation is presently unclear. However, exercise can improve ß‐adrenoceptor signalling and reduce oxidative stress, which may limit lipid modifications in sarcolemmal and other membranes (Corbi et al., 2012). Moreover, Wang et al. (2014b) recently suggested exercise‐dependent restoration of ischaemic preconditioning in aged hearts involves enhanced polyamine synthesis, which may stabilize the cell membrane, limit Ca^{2 +} overload and enhance mitochondrial function.

These findings highlight the importance of modifiable lifestyle factors (caloric intake, PUFA or fatty acid intake and physical activity level) that can positively influence the membrane, thus improving stress‐resistance and sensitivity to protective intervention in older or diseased hearts. Data lend support to the notion of membrane therapy, via manipulation of PUFA and other fat intakes, as a means of reducing the development of IHD, limiting the effects of myocardial I‐R, and enhancing the ability of the heart to respond to protective stimuli.

Other interventions

Apolipoprotein A‐1 (ApoA1) increased caveolin‐1 expression and cholesterol transport to cholesterol‐rich caveolae (Sviridov et al., 2001), and ApoA1 mimetics may restore preconditioning in acute hyperglycaemia via shifts in caveolin‐1 and caveolar eNOS localization (Baotic et al., 2013). Intriguingly, Liu et al. (2015), building on evidence that hyperglycaemic suppression of caveolin‐3 might involve enhanced PKCβ2 activity, recently showed that targeting PKCβ2 improved caveolin‐3 expression in T1D hearts, in association with reduced cell death, improved mitochondrial function and enhanced I‐R tolerance. Targeting of this path could thus promote cardioprotective efficacies in diabetes. Curiously, left ventricular assist devices in heart failure patients also increase caveolin expression (Uray et al., 2003, Fridolfsson and Patel, 2013). The basis of this effect is unclear, and whether it reflects the direct effects of mechanical unloading and limitation of adverse remodelling or other processes remains to be established.

The ability of farnesol, a metabolite of the mevalonate pathway, to restore pacing‐dependent preconditioning in hearts of hypercholesterolaemic rats suggests that excess cholesterol may inhibit formation of polyprenyl derivatives and the prenylation of G proteins, impairing signal transduction (Ferdinandy et al., 1998b). Other strategies to limit this process might thus be of benefit in optimizing cardioprotection in relevant dyslipidemias.

Interestingly, cardioprotection via a sustained opioid receptor stimulus is retained in aged and diabetic myocardium, inducing a prolonged protective window that appears to be caveolin‐3 independent in young healthy myocardium (See Hoe et al., 2014). This novel δ‐opioid receptor mediated response supports the potential value of caveolin‐3 independent stimuli (although most major conditioning responses appear caveolin‐3 dependent). More intriguingly, this stimulus has been shown in pilot studies to augment caveolin‐3 expression in both aged and T2DM myocardium (each exhibiting depressed baseline levels of caveolin‐3), in association with improved caveolar density and ischaemic tolerance (data not shown). Thus, while this novel response protects young healthy hearts independently of caveolin‐3 and caveolae, it is possible that the stimulus confers benefit through their up‐regulation in conditions such as ageing and disease,that involve significant depression of caveolin‐3 and caveolae. The basis of this stimulatory effect expression awaits delineation.

In discussing the sarcolemma, and in particular, cholesterol content of the sarcolemma, it is relevant to consider potential influences of widely employed statins. Statins are 3‐hydroxy‐3‐methylglutaryl CoA (HMG‐CoA) reductase inhibitors employed in management and prevention of hypercholesterolaemia. Although experimental and clinical studies clearly establish beneficial cardiovascular effects of statins, involving both lipid‐lowering and pleiotropic actions, surprisingly, few studies have detailed the effects of statins on sarcolemmal composition, and contradictory findings are reported. For example, Pogue et al. (1995) reported that 13 day lovastatin treatment failed to modify membrane cholesterol content. In contrast, Gray et al. (2000) report a significant reduction in sarcolemmal Na⁺‐K⁺ pump activity and abundance with lovastatin, with Na⁺‐K⁺ pump function highly dependent upon sarcolemmal cholesterol. The effects of statins on myocardial ischaemic tolerance appear to primarily involve a range of effects, including modulation of mitochondrial K_ATP channels and the permeability transition pore, NOS, RhoA and Rac1, and the processes of autophagy and apoptosis, with longer term outcomes additionally involving effects on fibrosis and remodelling. These actions may predominate over potentially detrimental influences on sarcolemmal cholesterol, although this requires more specific study.

While there is a paucity of information regarding statins and cardiac sarcolemmal makeup and function, a number of studies have focused on potential changes in brain membranes and function (the brain being the most cholesterol‐rich organ). For example, simvastatin significantly reduces cellular cholesterol content in hypothalamic cells (Fukui et al., 2015), in association with impaired insulin‐dependent Akt phospho‐activation (Fukui et al., 2015). Such effects may arise throughout the CNS, as simvastatin reduced cholesterol content across brain regions, coupled with increased membrane fluidity, decreased serotonin transporter activity and behavioural changes (Vevera et al., 2016). Atorvastatin treatment is also associated with alterations in membrane lipid raft proteins and behaviour or cognition (Schilling et al., 2014). Lovastatin, simvastatin and atorvastatin are all lipophilic statins, limiting their capacity to cross the blood–brain barrier. Hydrophilic statins, which require carrier‐mediated uptake, may have more limited effects in the CNS.

Conclusions

Over recent decades, there has been a substantial global effort to develop cardioprotective strategies to limit myocardial injury, as a consequence of accidental or surgical I‐R. However, there is still no widely applicable cardioprotective intervention, although a number of recently trialled strategies offer some encouragement. That said, there remains a significant dissociation between the efficacies of protective stimuli in experimental models (in which infarction can be profoundly limited) those in clinical cohorts (in which reductions in injury markers are modest at best). Unravelling the basis of this dissociation is an important step to developing truly effective interventions. While considerable attention has focused on changes to the intracellular milieu with age and disease (particularly mitochondria and intracellular mediators of survival and death signalling), less attention has been paid to the critically important role of the sarcolemma in I‐R injury/infarction and cardioprotection. Abundant evidence implicates significant modifications in membrane makeup and function with age and disease, with these sarcolemmal changes predicted on theoretical grounds (and supported by experimentation) to significantly modify the initiation and transduction of protective signals in cardiac cells. The potential value of addressing these changes is supported by the ability of specific membrane‐targeted interventions (together with several less targeted manipulations that may nonetheless improve membrane makeup and function) to enhance or restore cardioprotection in aged and diseased hearts. A focus on sarcolemmal changes can not only more effectively and rationally advance the quest for clinical cardioprotection but also fundamentally increase our understanding of cardiovascular injury processes and the still ill‐defined basis of cellular ageing.

Conflict of interest

The authors declare no conflicts of interest.

See Hoe, L. E. , May, L. T. , Headrick, J. P. , and Peart, J. N. (2016) Sarcolemmal dependence of cardiac protection and stress‐resistance: roles in aged or diseased hearts. British Journal of Pharmacology, 173: 2966–2991. doi: 10.1111/bph.13552.