Heart and Circulatory Physiology

Essential role of EGFR in cardioprotection and signaling responses to A1 adenosine receptors and ischemic preconditioning

Grant Williams-Pritchard, Matthew Knight, Louise See Hoe, John P. Headrick, Jason N. Peart


Transactivation of epidermal growth factor receptor (EGFR) may contribute to specific protective responses (e.g. mediated by δ-opioid, bradykinin, or muscarinic receptors). No studies have assessed EGFR involvement in cardioprotection mediated by adenosine receptors (ARs), and the role of EGFR in ischemic preconditioning (IPC) is unclear. We tested EGFR, matrix metalloproteinase (MMP), and heparin-binding EGF (HB-EGF) dependencies of functional protection via A1AR agonism or IPC. Pretreatment of mouse hearts with 100 nM of A1AR agonist 2-chloro-N6-cyclopentyladenosine (CCPA) or IPC (3 × 1.5-min ischemia/2-min reperfusion) substantially improved recovery from 25-min ischemia, reducing left ventricular diastolic dysfunction up to 50% and nearly doubling pressure development and positive change in pressure over time (+dP/dt). Benefit with both CCPA and IPC was eliminated by inhibitors of EGFR tyrosine kinase (0.3 μM AG1478), MMP (0.3 μM GM6001), or HB-EGF ligand (0.3 ng/ml CRM197), none of which independently altered postischemic outcome. Phosphorylation of myocardial EGFR, Erk1/2, and Akt increased two- to threefold during A1AR agonism, with responses blocked by AG1478, GM6001, and CRM197. Studies in HL-1 myocytes confirm A1AR-dependent Erk1/2 phosphorylation is negated by AG1478 or GM6001, and reduced with CRM197 (as was Akt activation). These data collectively reveal that A1AR- and IPC-mediated functional protection is entirely EGFR and MMP dependent, potentially involving the HB-EGF ligand. Myocardial survival kinase activation (Erk1/2, Akt) by A1AR agonism is similarly MMP/HB-EGF/EGFR dependent. Thus MMP-mediated EGFR activation appears essential to cardiac protection and signaling via A1ARs and preconditioning.

  • contractile dysfunction
  • epidermal growth factor receptor
  • ischemia-reperfusion
  • kinase signaling
  • matrix metalloproteinase
  • stunning

the heart possesses intrinsic cardioprotective responses that provide some tolerance to ischemic insult and are transduced via diverse signaling paths. Endogenous or exogenous agonism of G-protein coupled receptors (GPCRs) activates reperfusion injury salvage kinase (RISK) pathway components that appear to converge on mitochondrial targets such as KATP channels and the permeability transition pore (17, 31). Sarcolemmal receptors for adenosine, opioids, or bradykinin (among others) may initiate or mediate the benefit of pre- or postconditioning via these paths (13, 16, 40). Local TNFα may also trigger a recently identified survivor activating factor enhancement (SAFE) pathway, involving STAT-3 signal activation independently of RISK components (though both may target common end effectors), that may contribute to protection with postconditioning (23). Further complexity emerges, in that GPCR-triggered protection may involve growth factor receptor transactivation.

Growing evidence supports an important role for matrix metalloproteinase (MMP)-dependent transactivation of epidermal growth factor receptor (EGFR) in cardiac protection triggered by δ-opioid (3, 7, 12) and muscarinic receptors (21, 22), although the latter appear unimportant as endogenous protectants. While earlier analysis of bradykinin in rabbit hearts supported EGFR-independent signaling (7), data published by Methner et al. (28) in the course of the current study identify a role for MMP-8-mediated EGFR transactivation in bradykinin-triggered protection. Despite critical roles for adenosine receptors (ARs) in mediating conditioning responses (10, 11, 20, 24) and dictating intrinsic ischemic tolerance (30, 33, 36), there has been no assessment of the EGFR dependence of AR-mediated cardioprotection. The early study of Krieg et al. (21) identified a role for EGFR in responses to acetylcholine but did not test adenosine.

Interestingly, prior studies (6, 35) suggest that adenosine does not engage the same signaling harnessed by opioids, bradykinin, or acetylcholine (which also differ from each other in terms of signaling; Ref. 7). However, whether EGFR is activated by protective ARs and whether this is critical to adenosinergic protection have not been adequately assessed. The possibility is supported by evidence of EGFR involvement in AR signaling in neonatal myocytes (14) and asthmatic airways (27). Given the roles of A1ARs in determining ischemic tolerance and tissue protection (30, 33, 36), we tested whether A1AR-dependent cardioprotection and activation of the prototypical RISK element Erk1/2 exhibit MMP or EGFR dependence in intact hearts and isolated myocytes. We additionally tested for EGFR involvement in protection with ischemic preconditioning (IPC) since this may rely on ARs (10, 11, 24) and conflicting data have been acquired regarding MMP and EGFR dependence of IPC (2, 19).


All studies were approved by the Independent Review Committee, and performed in accordance with the guidelines, of the Animal Ethics Committee of Griffith University, which is accredited by the Queensland Government, Department of Primary Industries and Fisheries (under the guidelines of Animal Care and Protection Act 2001, Section 757).

Perfused heart preparation.

Hearts were removed from male C57/Bl6 mice (10–14 wk of age) anesthetized with 60 mg/kg sodium pentobarbital and were perfused as described previously (18, 34). Briefly, hearts were excised into ice-cold perfusion fluid, the aorta was cannulated, and the coronary circulation was perfused in a Langendorff mode at a constant pressure of 80 mmHg with modified Krebs-Henseleit buffer containing (in mM): 120 NaCl, 25 NaHCO3, 4.7 KCl, 2.5 CaCl2, 1.2 MgCl2, 1.2 KH2PO4, 15 d-glucose, and 0.5 EDTA. Perfusate was equilibrated with 95% O2-5% CO2 at 37°C to provide a pH of 7.4 and Po2 of ∼600 mmHg at the tip of the aortic cannula over a 1–5 ml/min flow range. Perfusate delivered to hearts was passed through an in-line 0.22-μm Sterivex-HV filter cartridge (Millipore, Bedford, MA) to remove microparticulates. The left ventricle was vented with a polyethylene apical drain, and a fluid-filled balloon constructed of polyvinyl chloride plastic film was inserted into the left ventricle via the mitral valve. Balloons were connected to a P23 XL pressure transducer (Viggo-Spectramed, Oxnard, CA) by fluid-filled tubing permitting continuous measurement of ventricular pressure. Balloon volume was increased to give a left ventricular end-diastolic pressure of 5 mmHg during stabilization and was not further adjusted. Coronary flow was monitored via a cannulating Doppler flow probe (Transonic Systems, Ithaca, NY) in the aortic perfusion line, which was connected to a T206 flowmeter (Transonic Systems). All functional data were recorded at 1 KHz on an 8-channel MacLab data acquisition system (ADInstruments, Castle Hill, Australia) connected to an Apple iMac computer. The left ventricular pressure signal was digitally processed to yield peak systolic, diastolic, and developed pressures, paotivie and negative changes in pressure over time (+dP/dt and −dP/dt), and heart rate.

After balloon placement, hearts were immersed in warmed perfusate in a water-jacketed chamber at 37°C. Temperature of the perfusion fluid was monitored by a needle thermistor probe at the entry into the aortic cannula, temperature of the water bath was assessed via a second probe, and data were recorded using a 3-channel Physitemp TH-8 digital thermometer (Physitemp Instruments, Clifton, NJ). Hearts were excluded from the study after an initial 20-min stabilization period if they met one of the following functional criteria: 1) coronary flow >5 ml/min, 2) unstable (fluctuating) contractile function, 3) left ventricular systolic pressure <100 mmHg, or 4) significant cardiac arrhythmias. More than 95% of hearts were viable for study inclusion.

Cardiac ischemia-reperfusion.

Following 20 min of normoxic stabilization, hearts were switched to ventricular pacing at 420 beats/min (Grass S9 stimulator, Quincy, MA), normalizing rate to permit comparison of rate-dependent measures of inotropic and lusitropic state (+dP/dt and −dP/dt, respectively). Baseline measurements were made after a further 10 min before hearts were subjected to 25 min of normothermic global zero-flow ischemia followed by 45 min of aerobic reperfusion. Pacing was terminated on initiation of ischemia and resumed after 1.5 min of reperfusion (18, 34).

Groups assessed included untreated hearts (n = 12) and hearts receiving 100 nM cyclopentyladenosine (CCPA; initiated 10-min preischemia; n = 9) or IPC (3 × 1.5-min periods of ischemia separated by 2 min reperfusion; n = 9). Interventions were repeated in the presence of 300 nM of the EGFR tyrosine kinase inhibitor AG1478 (n = 7 for CCPA; n = 7 for IPC), the Zn-dependent MMP inhibitor GM6001 (n = 8 for CCPA; n = 8 for IPC), or 0.3 ng/ml of the heparin-binding (HB)-EGF ligand inhibitor CRM197 (n = 7 for CCPA; n = 6 for IPC). The high cost of the latter agent prohibited use of higher levels in perfused hearts. Inhibitor infusions were commenced 5 min before CCPA or IPC. Effects of inhibitors alone were also tested (n = 8 for AG1478; n = 6 for GM6001; n = 7 for CRM197). To specifically test effects of A1AR agonism on EGFR phosphorylation in hearts, and confirm inhibition of this process by AG1478, a series of normoxic hearts were subjected to 10-min treatment with 100 nM CCPA ± 300 nM AG1478 (n = 4 per group), before being frozen in liquid N2 for subsequent analysis of EGFR phosphorylation.

Cardiac cell (HL-1) signaling studies.

The HL-1 cell line, derived from a mouse atrial cardiomyocyte tumor and characterized by Claycomb and colleagues (5, 39) was employed to further test the EGFR dependence of A1AR-triggered signaling. Cells were kindly provided by Dr. W. C. Claycomb (Louisiana State University Health Science Center, New Orleans, LA) and were plated on fibronectin-gelatin coated plates and cultured in supplemented Claycomb medium (10% FBS, 50 U/ml penicillin, 50 mg/ml streptomycin, 0.1 mM norepinephrine, and 2.0 mM l-glutamine). Cells were incubated at 37°C in 5% CO2-95% air (95% relative humidity) with media changed daily until cultures reached 70–80% confluence, and cells were serum starved in DMEM for 12 h before experimentation.

Experiments involved transient activation of A1ARs by 5-min exposure to the selective A1AR agonist CCPA (100 nM). Phosphorylated Erk1/2 expression in response to CCPA was assessed with or without AG-1478 (1 μM), GM6001 (10 μM), or CRM197 (5 μg/ml). Inhibitory agents were added 20 min (AG-1478, GM6001) or 60 min (CRM197) before CCPA, and the effects of inhibitors alone were also tested. On completion of experiments, cells were washed twice in ice-cold PBS and lysed in ice-cold RIPA lysis buffer (with protease and phosphatase inhibitors) for 5 min. Lysates were centrifuged for 5 min (10, 000 rpm, 4°C) to pellet debris. Lysate protein was assayed using a Pierce BCA kit.

To confirm myocardial relevance of signaling responses observed in HL-1 cells, we subjected a group of normoxic perfused hearts to 100 nM CCPA for 10 min alone or following pretreatment with 300 nM GM6001, 0.3 ng/ml CRM197, or 300 nM AG1478 (n = 4 per group). Hearts were then processed for immunoblot analysis of Erk1/2 and Akt phosphorylation.

Western immunoblotting and EGFR immunoprecipitation.

Tissue lysate samples containing 30 μg of total protein were loaded onto precast 12% acrylamide gels and separated at 150 V for ∼1.5 h. Proteins were subsequently transferred to polyvinylidene difluoride membranes and blocked in 5% skim milk powder in TBS-Tween for 60 min. Blots were incubated with primary antibody (total or phosphorylated Erk1/2 or Akt; 1:1000; Cell Signaling) overnight at 4°C. Following three washes in TBS-Tween, the blots were incubated with the secondary antibody and visualized on a ChemiDoc XRS system (Bio-Rad).

For assessment of EGFR tyrosine residue phosphorylation in intact hearts, frozen tissue was homogenized and tissue lysate (300 μg total protein) was incubated with 2 μg of anti-EGFR antibody (Santa Cruz Biotechnology) for 2 h at 4°C with gentle rocking. Subsequently, 30 μl protein G-agarose suspension (Santa Cruz Biotechnology) was added and samples were incubated overnight on a rocking platform overnight at 4°C. Immunoprecipitates were gently washed four times with 1.0 ml RIPA buffer, resuspended in 40 μl of 2 × Laemmli sample buffer, and boiled for 5 min at 95°C. Samples were then assessed via Western immunoblotting as described above, probing with an anti-phosphotyrosine primary antibody (1:1000; Cell Signaling; cat. no. 9411). Phosphorylation of EGFR or signaling kinases is expressed as the ratio of phosphorylated to total protein levels (normalized to ratios for untreated control samples).

Statistical analysis.

All values are expressed as means ± SE. A multiway ANOVA was employed to contrast groups, with a post hoc Newman-Keuls test to identify specific time and treatment effects. Evidence of statistical significance was accepted for P < 0.05.


A1AR- and IPC-mediated protection is EGFR, MMP, and HB-EGF dependent in intact hearts.

Baseline contractile function and coronary flow were unaltered by the varied inhibitor treatments and protective stimuli (Table 1). Following ischemia, control (untreated) hearts exhibited sustained elevations in diastolic pressure to >20 mmHg (Fig. 1A) and recovered ∼50% of left ventricular pressure development (Fig. 1B) and +dP/dt (Fig. 1B). Pretreatment with CCPA improved functional outcome, significantly reducing postischemic diastolic pressure by ∼50% and increasing pressure development and dP/dt by ∼35% (Fig. 1). Neither the EGFR tyrosine kinase inhibitor AG1478 (Fig. 1), the MMP inhibitor GM6001 (Fig. 2), nor the HB-EGF inhibitor CRM197 (Fig. 3) modified responses to ischemia-reperfusion when applied alone, suggesting limited contributions of basal EGFR, MMP, or HB-EGF activities to ischemic tolerance. However, these inhibitors abolished CCPA-mediated functional protection in reperfused hearts (Figs. 13). IPC hearts with three periods of transient ischemia before the index ischemia also significantly improved functional outcome from prolonged ischemic insult (Fig. 1). This benefit was also blocked by concurrent treatment with AG1478 (Fig. 1), GM6001 (Fig. 2), or CRM197 (Fig. 3). Coronary flow rate recovered to preischemic levels in all groups and was not significantly modified by any treatments tested (data not shown).

View this table:
Table 1.

Baseline functional data for hearts stabilized under normoxic conditions

Fig. 1.

Functional protection from ischemia-reperfusion is AG1478 sensitive [epidermal growth factor receptor (EGFR) dependent]. Hearts were untreated or pretreated with A1 adenosine receptor (AR) agonist [100 nM cyclopentyladenosine (CCPA)] or ischemic preconditioning (IPC) before ischemia-reperfusion. Effects of the EGFR inhibitor AG1478 (300 nM) on these responses were assessed. Data are show for postischemic recovery of left ventricular (LV) end-diastolic pressure (A), left ventricular developed pressure (B), and positive change in pressure over time (+dP/dt; C). For n values, refer to Table 1. Data are means ± SE. *P < 0.05 vs. control; †P < 0.05 vs. in the absence of inhibitor (untreated).

Fig. 2.

Functional protection from ischemia-reperfusion is GM6001 sensitive [matrix metalloproteinase (MMP) dependent]. Hearts were untreated or pretreated with A1AR agonist (100 nM CCPA) or IPC before ischemia-reperfusion. Effects of the MMP inhibitor GM6001 (300 nM) on these responses were assessed. Data are show for postischemic recovery of left ventricular end-diastolic pressure (A), left ventricular developed pressure (B), and +dP/dt (C). For n values, refer to Table 1. Data are means ± SE. *P < 0.05 vs. control; †P < 0.05 vs. in the absence of inhibitor (untreated).

Fig. 3.

Functional protection from ischemia-reperfusion is CRM197 sensitive [heparin-binding (HB)-EGF dependent]. Hearts were untreated or pretreated with A1AR agonist (100 nM CCPA) or IPC before ischemia-reperfusion. Effects of the HB-EGF inhibitor CRM197 (0.3 ng/ml) on these responses were assessed. Data are show for postischemic recovery of left ventricular end-diastolic pressure (A), left ventricular developed pressure (B), and +dP/dt (C). For n values, refer to Table 1. Data are means ± SE. *P < 0.05 vs. control; †P < 0.05 vs. in the absence of inhibitor (untreated).

Changes in myocardial EGFR phosphorylation mirrored the cardioprotective response to CCPA (Fig. 4). Agonism of A1ARs substantially increased cardiac EGFR phosphorylation, and this response was negated by AG1478. Total expression of EGFR was unchanged in response to CCPA (data not shown). Preliminary analysis employing an anti-Tyr-1172 antibody supports phosphorylation of this specific residue (data not shown). While we also attempted to assess EGFR phosphorylation in HL-1 experiments, the levels were too low for reliable detection via Western immunoblot. Subsequent quantitative real-time PCR analysis (using a Taqman Gene Expression Assay from Applied Biosystems: Assay ID Mm01187858_m1) revealed very low abundance of EGFR transcript in HL-1 cells vs. cardiac tissue: the Ct (threshold cycle) value ranged from 36 to 38 for HL-1 cells vs. 32 for heart homogenate samples (5.0 μg total RNA from untreated HL-1 cells was used in cDNA synthesis).

Fig. 4.

Myocardial EGFR phosphorylation in perfused hearts. Hearts were untreated or pretreated with the A1AR agonist CCPA (100 nM) alone or in the presence of the EGFR inhibitor AG1478 (300 nM). Degree of EGFR phosphorylation is expressed as the ratio of phosphorylated to total protein (P:T), normalized to untreated controls. A representative immunoblot is shown at top. Data are means ± SE. *P < 0.05 vs. control; †P < 0.05 vs. in the absence of inhibitor.

A1AR activation of signaling kinases is EGFR, MMP, and HB-EGF dependent in HL-1 cells.

The cytoprotective effects of ARs involve Erk1/2 activation (15, 25, 38, 41), and this kinase is also implicated in pre- and postconditioning responses (8, 10). Activation of Erk1/2 by A1AR agonism was confirmed in HL-1 cells: 5-min exposure to 100 nM CCPA triggered 200–250% elevations in Erk1/2 phosphorylation (Fig. 5). Initial time-course analysis showed that phosphorylation peaked after 5–10 min of A1AR stimulation (data not shown). Phosphorylation of Erk1/2 in response to CCPA was reduced by ∼70% by AG1478, eliminated by GM6001, and reduced by 50% with CRM197 (Fig. 5). Inhibitors alone did not alter Erk1/2 phosphorylation in the absence of CCPA. The effects of CCPA on kinase activation were predictably blocked by application of the A1AR-selective antagonist 1,3-dipropyl-8-cyclopentylxanthine (data not shown). We also observed Akt phosphorylation in response to CCPA, a response similarly sensitive to CRM197 (Fig. 6). Thus ligation of A1ARs triggers both Erk1/2 and Akt phosphorylation in murine myocytes in what is predominantly an MMP/EGFR-dependent process involving HB-EGF. Confirming commonality of this signaling in both intact hearts and HL-1 cells, Fig. 7 depicts data for CCPA-dependent phosphorylation of Erk1/2 and Akt in perfused hearts and confirms abolition of these signaling response by GM6001, CRM197, and AG1478. Total expression levels for Erk1/2 and Akt in cells and hearts were unaltered by A1AR agonist or different inhibitor treatments (data not shown).

Fig. 5.

A1AR-mediated Erk1/2 phosphorylation in HL-1 cells. HL-1 myocytes were subjected to 5 min stimulation with the A1AR agonist CCPA (100 nM) alone or in the presence of an EGFR inhibitor (A; 1 μM AG1748); in the presence of an MMP inhibitor (B; 10 μM GM6001); or in the presence of a HB-EGF inhibitor (C; 5 μg/ml CRM197). Inhibitors were also applied alone. Degree of phosphorylation is expressed as the ratio of phosphorylated to total protein, normalized to untreated controls. Representative immunoblots for phosphorylated (top) and total (bottom) Erk1/2 are shown. Data are means ± SE. *P < 0.05 vs. control; †P < 0.05 vs. in the absence of inhibitor.

Fig. 6.

Effects of HB-EGF inhibition on A1AR-triggered Akt activation in HL-1 myocytes. HL-1 myocytes were subjected to 5-min stimulation with the A1AR agonist CCPA (100 nM) alone or in the presence of HB-EGF inhibitor 5 μg/ml CRM197. Degree of phosphorylation is expressed as the ratio of phosphorylated total protein, normalized to untreated controls. Representative immunoblots for phosphorylated (top) and total (bottom) Akt are also shown. Data are means ± SE. *P < 0.05 vs. control; †P < 0.05 vs. in the absence of inhibitor.

Fig. 7.

Activation of cardiac Erk1/2 and Akt in hearts is MMP and HB-EGF dependent. Hearts were untreated or pretreated with the A1AR agonist CCPA (100 nM) alone or in the presence of the MMP inhibitor GM6001 (300 nM) or HB-EGF inhibitor CRM197 (0.3 ng/ml). Degree of phosphorylation is expressed as the ratio of phosphorylated to total protein, normalized to untreated controls. Data are means ± SE. *P < 0.05 vs. control; †P < 0.05 vs. in the absence of inhibitor.


The present study is the first to identify an essential role for EGFR signaling in AR-mediated cardiac protection. Our data reveal that A1AR-dependent functional protection (and also protection with IPC) is associated with myocardial EGFR phosphorylation and is abrogated by inhibitors of EGFR, MMP, and HB-EGF. Analysis of signaling confirms that cardiac A1ARs activate Erk1/2 and Akt in a MMP- and EGFR-dependent manner, likely involving HB-EGF ligand. Data collectively demonstrate that MMP-dependent EGFR transactivation is critical to A1AR- and IPC-mediated cardioprotection, as it appears to be in opioid (3, 7, 12)- and bradykinin (28)-triggered responses. Transactivation of EGFR thus plays a central (and essential) role in cardioprotection mediated by multiple GPCRs and conditioning stimuli.

EGFR involvement in A1AR- and IPC-triggered cardioprotection.

Transactivation of EGFR is involved in mediating responses to specific GPCRs in different tissues (1, 3, 7, 9, 12). Transactivation may involve MMP-dependent cleavage of ligand (e.g.. HB-EGF) from membrane-associated proligand complexes, permitting interaction with the EGFR binding site (22). Ligand cleavage may also involve specific ADAMs (a disintegrin and metalloproteinase) (19). Bound EGFR dimerizes, transactivating tyrosine kinase to autophosphorylate tyrosine residues and recruit Src kinase and PI3-K. This in turn leads to activation of Erk1/2, Akt/PKB, NO synthase, and protein kinase G (10), among other targets. These latter mediators are all implicated in GPCR-triggered cardioprotection (10, 13, 17, 31), including AR responses (14, 15, 32, 38, 41). Transactivation of EGFR has been implicated in cytoprotection in response to δ-opioid agonism (7, 12) and acetylcholine (22). However, there are no studies of EGFR involvement in adenosinergic protection.

Agonism of A1ARs was without effect on postischemic outcomes in the presence of inhibitors of EGFR, MMP, or HB-EGF. Moreover, A1AR agonism was shown to phosphorylate cardiac EGFR, a response negated by AG1478 (Fig. 4). Consistent with transactivation of EGFR by receptors implicated in preconditioning (3, 7), inhibition of MMP, HB-EGF, or EGFR also abrogated the benefit with IPC (Figs. 13). We did not directly assess EGFR or kinase phosphorylation responses in the setting of IPC. Nonetheless, the effects of MMP, HB-EGF, and EGFR inhibitors on signaling responses and protection are consistent across the HL-1 and heart models. These findings agree with the data of Benter et al. (2) in rats, although they are counter to the study of Ichikawa et al. (19) in rabbits. These differences may reflect species-specific mechanisms of IPC, with the rabbit also the model in which differing signaling was observed for adenosine vs. acetylcholine or opioids (6). It is interesting that Ichikawa et al. (19) did detect a role for ligand cleavage by an ADAM in rabbits, although the receptor subsequently targeted was not the EGFR.

Complete elimination of either A1AR- or IPC-mediated protection suggests these responses are entirely MMP and EGFR dependent. If protective signaling involved both GPCR- and EGFR-coupled kinase signals (e.g.. Erk1/2, PI3-K, and Akt), inhibition of the EGFR arm alone would incompletely limit protection with A1ARs or IPC. Thus triggering receptors appear to act solely through MMP-dependent generation of EGFR ligand (likely HB-EGF), with subsequent EGFR activation. Why an acute protective response to GPCR agonism involves this somewhat indirect signaling sequence is not immediately clear. However, interactions between triggering GPCRs, MMP-dependent ligand shedding, and EGFR activation provide considerable scope for regulation of tissue response and may also couple acute responses to subsequent tissue growth/remodeling. For example, recent data suggest that Erk1/2 phosphorylation by EGFR promotes ADAM-dependent EGFR activation in a positive feedback manner (42). Such effects, together with shifts in MMP expression, could prime longer-term growth signaling subsequent to the initial retaliatory or protective stimulus.

EGFR involvement in A1AR-dependent kinase signaling.

Since the cardiac protective effects of ARs (15, 25, 38, 41), preconditioning (10), and postconditioning (8) may involve Erk1/2 activity, we analyzed the activation of this MAPK and additionally tested for Akt regulation. Data from HL-1 cells (Figs. 5 and 6) further support A1AR-triggered growth factor shedding (likely HB-EGF), leading to EGFR activation and subsequently Erk1/2 and Akt phosphorylation. This signaling model seems to apply in both HL-1 cells and intact hearts (Fig. 7). Thus the A1AR appears to engage both Erk1/2 and Akt via essential MMP/EGFR engagement.

There are surprisingly few prior studies of EGFR involvement in adenosine signaling: Germack et al. (14) found that A1 and A3AR activation of Akt in neonatal myocytes is blocked by AG1478, while McNamara et al. (27) recently showed that A1AR control of mucin production in asthmatic airways requires EGFR transactivation. Other work identifies A2AAR transactivation of Trk receptor tyrosine kinases, although not targeting of EGFR (38). Differing effects of AR subtypes may reflect tissue-specific responses and/or G-protein dependence of EGFR transactivation: the A1AR signals via Gi or Go, for example, whereas A2AARs are coupled to Gs.

It should be noted that while A1AR-mediated protection and Erk/Akt signaling were both sensitive to inhibitors of EGFR, MMP, and HB-EGF, specific involvement of Erk1/2 and Akt in cardioprotection was not tested and has been questioned. Germack et al. (14) found that despite EGFR-dependent Akt activation by A1 or A3ARs, associated protection was independent of Akt activity. Similarly, Skyschally et al. (37) reported that protection via postconditioning is unrelated to associated Erk1/2 or Akt phosphorylation, and Clarke et al. (4) presented evidence of dissociation between RISK activation and cardiac protection. Thus, while A1AR-mediated cardioprotection and Erk1/2 and Akt activation all appear MMP/EGFR dependent, we can make no conclusions here regarding the importance of Erk1/2 or Akt in A1AR cardioprotection.

Study limitations.

One limitation of this work is reliance on pharmacological manipulation of signaling elements. Inhibitory agents may exert nonspecific actions. However, we show no effects of any agent alone on baseline function, kinase signaling, or ischemic responses, the effects of different inhibitors are internally consistent (i.e.. inhibiting 3 different pathway components generates consistent outcomes), parallel effects arise in kinase signaling and functional recoveries, and we also confirm AG1478 inhibits EGFR phosphorylation. Since inhibitors may also be competitive, it is possible to underestimate the importance or contribution of targeted proteins. Thus, while AG1478 and GM6001 largely eliminated protection and signaling responses, partial blockade of signaling with CRM197 may either reflect incomplete antagonism or contribution from alternate EGFR ligands.

A second limitation relates to analysis of functional outcomes from ischemia. All functional benefit with A1ARs or IPC was eliminated by inhibitors of MMP and EGFR signaling. We thus conclude with some confidence that MMP/EGFR signaling is necessary for A1AR- and IPC-mediated protection against postischemic dysfunction. Nonetheless, ischemia-reperfusion injury entails a mix of reversible mechanical/electrophysiological dysfunction together with irreversible cell death/infarction (18, 34). While contractile function is a major determinant of outcome and thus an important clinical or experimental end point, a limitation is that we do not delineate relative effects of protective stimuli on differing components of acute injury. We have previously established that A1ARs and IPC limit infarction in this model (18, 29, 34) and have shown that contractile recoveries correlate strongly with cell death (34). Since all functional benefit with A1AR agonism (Fig. 3) or IPC (Fig. 5) was abrogated by MMP/EGFR inhibition, it is thus highly likely that changes in cell death and dysfunction are both involved, although their relative contributions cannot be directly ascertained.


In summary, our findings support essential roles for MMP-dependent ligand shedding and EGFR activation in A1AR- and IPC-mediated cardiac protection: neither stimulus provides functional benefit in the presence of MMP, HB-EGF, or EGFR inhibition. Furthermore, A1AR-triggered Erk1/2 and Akt activation in myocytes is similarly MMP/HB-EGF/EGFR dependent. Thus despite evidence of distinct signaling in adenosinergic responses (6, 35), our data support a common and essential role for EGFR transactivation in cardioprotection in response to ARs and IPC, as has been observed for opioid (3, 7, 12), muscarinic (22), or bradykinin (28) receptors. A general scheme of GPCR agonism, MMP-dependent shedding of receptor ligand, and engagement of EGFR signaling may apply for the major endogenous agonists that trigger cardioprotection and are implicated in conditioning responses. Given the importance of these signals to myocardial stress resistance, the protective functions and mechanisms of myocardial EGFR signaling warrant further investigation.


We gratefully acknowledge grant support from the National Heart Foundation of Australia (G 08B 3971; G 05B 2029) and the National Health and Medical Research Council of Australia (NHMRC). G. Williams-Pritchard was supported by a postgraduate scholarship from the Heart Foundation Research Centre, Griffith University, and J. N. Peart was the recipient of CDA fellowship support from the NHMRC.


No conflicts of interest, financial or otherwise, are declared by the author(s).


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
View Abstract