Ability of CP-532,903 to protect mouse hearts from ischemia/reperfusion injury is dependent on expression of A₃ adenosine receptors in cardiomyoyctes

Abstract

A₃ adenosine receptor (A₃AR) agonists are effective at limiting injury caused by ischemia/reperfusion injury of the heart in experimental animal models. However, understanding of their mechanism of action, which is likely multifactorial, remains incomplete. In prior studies, it has been demonstrated that A₃AR-mediated ischemic protection is blocked by glibenclamide and is absent in Kir6.2 gene ablated mice that lack the pore-forming subunit of the ATP-sensitive potassium (K_ATP) channel, suggesting one contributing mechanism may involve accelerated activation of K_ATP channels. However, presence of A₃ARs in the myocardium has yet to be established. Utilizing a whole-cell recording technique, in this study we confirm functional expression of the A₃AR in adult mouse ventricular cardiomyocytes, coupled to activation of ATP-dependent potassium (K_ATP) channels via G_i inhibitory proteins. We further show that ischemic protection provided by the selective A₃AR agonist CP-532,903 in an isolated, buffer-perfused heart model is lost completely in Adora3^{LoxP/LoxP;Myh6-Cre} mice, which is a newly developed model developed and comprehensively described herein whereby the A₃AR gene (Adora3) is deleted exclusively in cardiomyocytes. Our findings, taken together with previously published work, are consistent with the hypothesis that A₃AR agonists provide ischemic tolerance, at least in part, by facilitating opening of myocardial K_ATP channels.

Graphical Abstract

1. Introduction

Adenosine is produced rapidly in large quantities in the myocardium during ischemia and during reperfusion where it acts in several ways to protect against injury and to promote repair [1–5]. It mediates its effects via a family of four G protein-coupled adenosine receptors designated A₁, A_2A, A_2B, and A₃ [4]. The A₃ adenosine receptor (A₃AR) is a G_i protein-coupled receptor that is widely distributed, although it is most abundantly expressed in granulocytic cells including mast cells, neutrophils, and eosinophils [6–10]. In some rodents, but not in other species, the A₃AR stimulates degranulation of mast cells and subsequent release of a number of biological mediators including histamine [6, 11].

In prior studies, selective agonists for the A₃AR have been shown to be effective at limiting injury of the heart caused by ischemia and reperfusion [5, 12–20]. These agents have high therapeutic potential because they produce less hemodynamic side effects compared to adenosine or agonists for the other AR subtypes. Several different A₃AR agonists have proven effective in both in vitro (isolated-perfused hearts) and in vivo (infarction and ‘stunning’) models of ischemia/reperfusion injury in a number of different species (mice, rats, rabbits, and dogs), including the prototypical N⁶-benzyladenosine-5’-N-methylcarboxamide derivatives IB-MECA and Cl-IB-MECA [5, 12–17]. In a recent study, we reported cardioprotective efficacy by a multivalent dendrimeric conjugate (MRS5246) bearing a member of a new class of N⁶-benzyladenosine-5’-N-methylcarboxamide derivatives that contain a bicyclic methanocarba (bicyclo[3.1.0]-hexane) ring system in place of the ribose [19].

Although efficacy has been firmly established, the mechanisms by which A₃AR signaling protect against myocardial ischemia/reperfusion injury remain unclear. While there is evidence for an anti-inflammatory effect [14, 16], A₃AR agonist-mediated cardioprotection is blocked by the ATP-sensitive potassium (K_ATP) channel antagonist glibenclamide in reductionist model systems (i.e., isolated hearts and cardiomyocytes [5]). K_ATP channels are potassium-selective ion channels expressed in the myocardium that activate during ischemia as ATP levels drop [21–23]. Activation of K_ATP channels in metabolically compromised myocardium promotes cellular survival by reducing calcium overloading and depression of force development during muscle contraction secondary to action potential shortening [21–23]. It has previously been established that K_ATP channel activity in cardiomyocytes is enhanced by A₁AR signaling through G_i inhibitory proteins and it is thought that A₁AR agonists protect against ischemic injury by potentiating K_ATP channel opening [5, 22, 23]. Similar to other modifiers of the K_ATP channel (i.e. ADP, H⁺, certain potassium channel openers), G_i proteins activated by A₁ receptors as well as certain other G_i protein-coupled receptors expressed in the myocardium such as muscarinic receptors reduce sensitivity of the channel to blockade by ATP [21–23]. Considering that A₃ARs are also G_i protein-coupled receptors, it is plausible that A₃AR agonists may act by a similar mechanism. However, this idea remains controversial because, unlike with the A₁AR [24], there is little evidence for expression of A₃ARs in adult cardiomyocytes.

CP-532,903 is an A₃AR agonist with greater than 100-fold selectivity versus the A₁AR and little if any activity versus A_2A and A_2BARs [18, 20, 25, 26]. Structurally, CP-532,903 is an N⁶-dichlorobenzylsubstituted adenosine derivative modified at the 5’ position of the ribose moiety as a methyluronamide and at the 3’ position with an amino group in place of the hydroxyl group [18, 25, 26]. In a previous study, we reported that CP-532,903 was highly cardioprotective in an isolated mouse heart model of global ischemia/reperfusion injury and that protection provided by CP-532,903 was absent in experiments using hearts from Adora3^−/− mice or from mice lacking the pore-forming subunit (Kir6.2) of the K_ATP channel, implicating involvement of the A₃AR and K_ATP channels [18]. Given remaining questions regarding the cellular expression of A₃ARs in the heart, however, whether the cardiomyocyte is the primary site of action of CP-532,903 remains uncertain.

The goal of this study was to directly address these issues by creating a new mouse model described herein allowing for conditional deletion of the A₃AR gene (Adora3) specifically in cardiomyocytes using the LoxP/Cre recombinase strategy. Utilizing this new tool and an isolated buffer-perfused mouse heart model, we describe complete loss of cardioprotective efficacy by CP-532,903. In parallel electrophysiological studies utilizing cardiomyocytes isolated from global Adora1 or Adora3-deleted mice, we demonstrate that CP-532,903 activates K_ATP channels in adult mouse ventricular cardiomyocytes via activation of A₃ARs and pertussis toxin-sensitive inhibitory G proteins. Collectively, these results confirm the presence of functional A₃ARs in ventricular cardiomyocytes of mice and confirm cardiomyocyte A₃ARs as an important target for A₃AR agonist-mediated cardioprotection.

2. Materials and Methods

2.1. Materials

Cell culture reagents and TRIzol reagent were from Invitrogen (Carlsbad, CA). CP-532,903 was provided by Pfizer Inc. (Groton, CT), Liberase Blendzyme I was from Roche Applied Biosciences (Indianapolis, IN), E-4031 was purchased from Wako (Osaka, Japan), nisoldipine was a gift from Miles-Pentex (West Haven, CT), SYBR Green Supermix was from Bio-Rad Laboratories (Hercules, CA), GoTaq G2 Green Mastermix was from Promega, ZM 241385 and PSB-603 were from Tocris Bioscience (Ellisville, MO), histamine radioimmunoassays were obtained from Immunotech (Marseille, France), and all other drugs and reagents were purchased from Sigma-Aldrich (St. Louis, MO).

2.2. Animal Care and Use

This investigation adhered to the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals (NIH Pub. Nos. 85–23, Revised 1996). All protocols utilizing mice were preapproved by the Medical College of Wisconsin’s Institutional Animal Care and Use Committee (IACUC; Animal Welfare Assurance #A3102–01).

2.3. Source of Mice

Wild-type C57Bl/6J, Tg(Myh6-cre)2182Mds/J (#011038), and Tg(ACTFLPe)9205Dym (#003800) mice were purchased from The Jackson Laboratory. Global Adora1 null mice were a gift from Dr. Bertil Fredholm (Karolinska Institute) [27] and global Adora3 null mice were from Merck Research Laboratories [28].

2.4. Cardiomyocyte and Cardiac Fibroblast Isolation

Cardiomyocytes from the left ventricles of 10–14 week-old male or female adult mice were isolated as described previously (www.signalinggateway.org/data/ProtocolLinks.html; protocol no. PP000000125). In brief, hearts were excised from pentobarbital-anesthetized mice (75 mg/kg i.p.), cannulated via the aorta onto a blunted needle, and perfused for 10 min with warmed (37°C) perfusion buffer (in mM: 113 NaCl, 4.7 KCl, 0.6 KH₂PO₄, 0.6 Na₂HPO₄, 1.2 MgSO₄-7H₂O, 0.032 phenol red, 12 NaHCO₃, 10 KHCO₃, 10 HEPES [pH 7.4], 30 taurine, 10 2,3-butanedione monoxime, 5.5 glucose) containing 0.25 mg/ml Liberase Blendzyme I, 0.14 mg/ml trypsin, and 12.5 μM CaCl₂. Following perfusion, the left ventricle was dissected free from the atria and right ventricle, and repeatedly passed through a plastic transfer pipette to disaggregate the cells into a single-cell suspension. Subsequently, myocytes were enriched by sedimentation in perfusion buffer containing 5% bovine calf serum while slowly exposing the cells to increasing concentrations of CaCl₂ to achieve a final concentration of 1.2 mM. The final cell pellet containing calcium-tolerant ventricular cardiomyocytes was re-suspended in minimal essential medium containing Hanks’ salts, 2 mM L-glutamine, 5% bovine calf serum, 10 mM 2,3-butanedione monoxime, and 100 U/ml penicillin, which were used in electrophysiology studies. To further increase purity for the qPCR studies, the cardiomyocytes were allowed to attach to laminin-coated tissue culture plates for 1 h and then washed extensively with cell culture media to remove non-adherent cells. Cardiomyocyte purity after selective plating averaged 90–95%. Supernatants containing fibroblasts from the cardiomyocyte sedimentation procedure were plated onto plastic cell culture dishes and cultured in DMEM containing 10% fetal bovine serum, 100 units/ml penicillin, and 100 μg/ml streptomycin until confluent.

2.5. Quantitative RT-PCR (qPCR)

qPCR was performed, as described previously [8, 29], to assess mRNA levels of AR transcripts in isolated cardiomyocytes and cultured cardiac fibroblasts. Total RNA was obtained using TRIzol reagent. Subsequently, 1 μg of total RNA was reverse-transcribed using a mixture of random and poly-T primers, according to the manufacturer’s protocol (Invitrogen). Primers were designed for the mouse A₁ (forward, 5′-TGGCTCTGCTTGCTATTG-3′; reverse, 5′-GGCTATCCAGGCTTGTTC-3′), A_2A (forward, 5′ TCAGCCTCCGCCTCAATG-3′; reverse, 5′-CCTTCCTGGTGCTCCTGG-3′), A_2B (forward, 5′-TTGGCATTGGATTGACTC-3′; reverse, 5′-TATGAGCAGTGGAGGAAG-3′), and A₃AR (forward, 5′-CGACAACACCACGGAGAC-3′; reverse, 5′-GCTTGACCACCCAGATGAC-3′) using Beacon Design software (Bio-Rad Laboratories). PCR amplification (in SYBR Green Supermix) was performed using an iCycler iQ thermocycler (Bio-Rad Laboratories) for 40 cycles of 25 s at 95°C followed by 45 s at an optimized annealing temperature for each AR. The cycle threshold, determined as the initial increase in fluorescence above background, was ascertained for each sample. Melt curves were performed upon completion of the cycles to ensure that nonspecific products were absent. For quantification of AR transcripts, a standard curve plotting cycle threshold versus copy number was constructed for each receptor subtype by analyzing 10-fold serial dilutions of plasmids containing the full-length mouse AR cDNA clones. AR transcript levels were expressed as copies per 100 ng of total RNA.

2.6. Electrophysiology

Whole-cell K_ATP currents (I_KATP) were recorded from isolated cardiomyocytes at room temperature in a sodium-free external solution (5 mM KCl, 132 mM N-methyl-D-glucamine, 1 mM CaCl₂, 2 mM MgCl₂, 10 mM HEPES [pH 7.4]) containing 4-aminopyridine (5 mM) and E-4031 (5 μM) to block voltage-activated potassium currents, nisoldipine (200 nM) to block voltage-activated calcium currents, ZM 241385 (100 nM) to block A_2AARs, and PSB 603 (100 nM) to block A_2BARs. For the pertussis toxin experiments, cardiomyocytes were incubated in external solution containing 2 μg/ml pertussis toxin for a minimum of 1 h and up to 6 h during the course of the day while experiments were being conducted. The pipette solution was also sodium-free (60 mM K-glutamate, 50 mM KCl, 1 mM MgCl₂, 1 mM CaCl₂, 11 mM EGTA, and 10 mM HEPES [pH 7.4]) and contained K₂-ATP at a low concentration (0.2 mM) to reduce inhibition of K_ATP channels and 0.5 mM Na⁺-GTP to permit G protein activation. Recording pipettes were pulled from borosilicate glass capillary tubes (Garner Glass, Claremont, CA) using a horizontal two-stage puller (Sachs-Flaming P-97; Sutter Instruments, Novato, CA) and heat polished (microforge MF-830; Narishige, Tokyo, Japan). In standard solutions, the pipette resistance ranged from 3 to 5 MΩ. Current was measured using a patch clamp amplifier (Axopatch 200B; Axon Instruments, Foster City, CA) interfaced to a computer via a digitizer (Digidata 1322 A; Axon Instruments). Data acquisition and analysis were conducted using the pClamp software package version 9.0 (Axon Instruments). Additional analyses were performed on Origin version 7 (OriginLab, Northampton, MA).

2.7. Generation of Mice with LoxP-Flanked Adora3 Alleles

To generate mice with LoxP-flanked alleles, a targeting vector was prepared from a 13 kb bacterial artificial chromosome (BAC) clone containing Adora3 (Figure 1). The vector contained a LoxP site in the 5’ non-coding region, which was created by subcloning an insert extended with the LoxP sequence between XhoI and BlpI sites. In addition, the vector contained a neomycin phosphotransferase (neo) cassette (a gift from Stephen Duncan, PhD, Medical University of South Carolina) inserted into intron 1 by recombineering. The neo cassette was flanked on either side by Frt sites and on the 3’ end with a LoxP site. Correctly targeted embryonic stem cells (ESCs; line D3), confirmed by Southern blotting (Figure 1), were injected into C57Bl/6NCrl blastocysts, which implanted and transmitted the targeted allele via germline. The neo cassette in intron 1 was subsequently removed by mating with a mouse ubiquitously expressing a Flp recombinase leaving behind the LoxP site. Mice heterozygous for the targeted allele (Adora3^LoxP/+) were back-crossed with wild-type C57Bl/6J mice for nine generations, and then bred to homozygosity resulting in the genotype Adora3^LoxP/LoxP. Mice (C57Bl/6) containing the cardiac-specific Myh6-driven Cre-recombinase transgene, which is briefly activated during early development followed by strong and persistent expression in the ventricles following birth [30], were bred into the heterozygous Adora3^LoxP/+ background. Adora3^{LoxP/LoxP;Myh6-Cre+} were bred with Adora3LoxP/LoxP mice to obtain the experimental (Adora3^{LoxP/LoxP;Myh6-Cre}) and control (Adora3^LoxP/LoxP) littermate genotypes that were compared in the study.

Figure 1. Gene targeting strategy.

The targeting vector, containing a 13.0 kb fragment of Adora3 isolated from a mouse BAC clone library, was prepared to introduce two LoxP sites into Adora3, one into the 5’ non-coding region proximal to exon 1 and the second into intron 1. This permits Cre recombinase-mediated excision of exon 1 including the ATG start site along with ~2 kb of upstream sequence containing promoter elements. This targeting strategy avoids depletion of the transmembrane and immunoglobulin domain containing 3 splice variant (TMIGD3 or Adora3i3) that shares exon 2 of Adora3 with the A₃AR [42]. Correctly targeted ESCs (line D3 derived from blastocysts of a 129S2/SvPas mouse) were injected into C57Bl/6 blastocysts, generating 60–90% chimeric male mice estimated from coat color that transmitted the floxed allele via germline, providing the genotype Adora3^{LoxP(FrtneoFrt)LoxP/+)}. The neo gene in intron 1 was removed by mating with a mouse expressing the Flp recombinase transgene while leaving the LoxP site behind. A. Restriction map of the LoxP-targeted Adora3 allele. Bars denote Southern blot probes named 5’ probe and 3’ probe. B. Southern blotting of a correctly targeted ESC line used for blastocyst injections. Left panel. Digestion with NheI followed by hybridization with the 5’ probe reveals a 10.8 kb band diagnostic of the floxed allele due to presence of a NheI site within the neo cassette and a 16.6 kb band diagnostic of the wild-type allele. Right panel. Digestion with BglI and hybridization with the 3’ probe yields a 19.5 kb band from the wild-type allele and a 16.6 kb band from the floxed allele created by a BglI site within the neo cassette. An additional Southern blot was performed by hybridizing a BglI digest with a probe corresponding to the neo cassette, which revealed a single 16.6 kb band ensuring that the targeting vector did not randomly insert into non-endogenous targets (not shown). Appropriate targeting was further confirmed by PCR amplification to detect presence of the LoxP sites, the neo cassette, and site-specific insertion (not shown).

2.8. Genotyping

Genotyping of Adora3^LoxP mice was performed by PCR in 20 μl reactions that included 1x GoTaq G2 Green Mastermix (Promega), primers (0.5 μM), MgCl₂ at the concentrations listed in Table 1, and 4 μl template. Templates consisted of DNA extracted from ear samples that had been boiled for 10 minutes in 0.3 ml 10 mM NaOH/1 mM EDTA. PCR products were amplified as indicated in Table 1, followed by separation and imaging of amplicons on ethidium bromide-stained 1.5% agarose gels. An example of genotyping results is shown in Figure 2.

Table 1.

Primer pairs for PCR genotyping.

Primer Set	Primer Sequences (5’-3’)		Amplicon (bp)		Purpose
5’ LoxP	FWD REV	GGTTTGAATGTCAGTTTGTCAC GCATACATTATACGAAGTTATCCC	wt allele	none	detect 5’ LoxP
			floxed allele	820
3’ LoxP	FWD REV	GGAAGGTTTGTTTAAGCAGAAAT GCCTCCACTCAACCTTAGC	wt allele	581	distinguish het. vs homozygous
			floxed allele	615
Recombined LoxP	FWD REV	CCACACCTTCTAATAGTGCCACT CTATCTGTCCCTGTCCCCTTTC	intact floxed allele	4,317	detect Cre-mediated recombination
			recomb floxed allele	190
Myh6-Cre	FWD REV	ATGACAGACAGATCCCTCCTATCTCC CTCATCACTCGTTGCATCATCGAC	wt allele	none	detect Myh6-Cre transgene
			Myh6-Cre	300

Reactions (20 μl) contained 1x GoTaq G2 Mastermix (Promega), primers (0.5 μM), 1.5 mM MgCl₂ (2.0 mM for Myh6-Cre amplification), and 4 μl DNA template prepared as described in METHODS. Cycling conditions were as follows: 95 °C for 15 min followed by 35 cycles of 94 °C for 30 s/60 °C for 45 s/72 °C for 30 s, and then concluding with 72 °C for 10 min.

Figure 2. Genotyping protocol.

Mice are genotyped using two sets of primers to identify the 5’ LoxP site upstream of exon 1 and the 3’ LoxP site within intron 1. A third primer set is used to detect presence of the Myh6-Cre transgene. Lane 1: The reverse primer of the 5’ LoxP primer set anneals to the LoxP sequence thereby generating an 820 bp amplicon from the floxed allele or no product with the wild-type allele. Lane 2: The 3’ LoxP primer set flanks the 3’ LoxP site thereby generating a 581 bp amplicon diagnostic of the wild-type allele and a 615 bp amplicon diagnostic of the floxed allele. The presence of both amplicons with the 3’ LoxP primer set indicates heterozygosity. Lane 3: The Myh6-Cre primer set amplifies a 300 bp product from the Myh6-Cre transgene.

2.9. Assessment of Cre-Mediated Adora3 Recombination

Cre-mediated recombination of the Adora3 allele was assessed by PCR on DNA samples of the spleen, lung, brain, liver, kidney, small intestine, large intestine, skeletal muscle (quadriceps), and left ventricle using the primer pairs and reaction conditions listed in Table 1. These primers, which anneal to the 5’ non-coding region and intron 1 amplify a 190 bp segment following recombination.

2.10. Measurement of Plasma Histamine Levels

To ensure that placement of LoxP sequences did not interfere with normal expression and function of the A₃AR in Adora^LoxP mice, Cl-IB-MECA-induced increases in plasma histamine levels was assessed. In prior studies we have shown that both Cl-IB-MECA and CP-532,903 increase plasma histamine levels in mice due to A₃AR-mediated mast cell degranulation [13, 14, 18]. We used Cl-IB-MECA in this study to conserve our limited supply of CP-532,903. Mice were anesthetized with sodium pentobarbital (75 mg/kg i.p.), intubated, respirated with room air supplemented with 100% oxygen (tidal volume 225 μl; rate ~100 strokes/minute; model 845 Hugo Sachs Elektronic), and subjected to a thoracotomy to expose the heart. A saline-filled cannula (stretched PE-10 tubing) was inserted into the left carotid artery for drug administration. Blood samples (~150 μl) were collected by cardiac puncture into EDTA-coated syringes containing the histamine N-methyltransferase inhibitor SKF-91488 (final concentration = 20 μM). The blood samples were centrifuged (1000 x g for 10 min at 4° C) and the plasma was analyzed for histamine content by enzyme-linked immunosorbant assay (Immuno-Biological Laboratories).

2.11. Langendorff-Perfused Isolated Heart Model

Male or female mice (10–12 weeks of age) were anesthetized with sodium pentobarbital (75 mg/kg i.p.). As soon as deep anesthesia was achieved, the hearts were removed, arrested in ice-cold perfusion solution, and then rapidly cannulated via the aorta and perfused by the Langendorff method at continuous pressure (80 mm Hg) with Krebs-Henseleit (KH) buffer containing 118 mM NaCl, 4.7 mM KCl, 1.2 mM MgCl₂, 2.5 mMCaCl₂, 1.2 mM KH₂PO₄, 0.5 mM EDTA, 25 mM NaHCO₃, and 11 mM glucose. The KH buffer was equilibrated with 95% O₂/5% CO₂ at 37°C to achieve a pH of 7.4 and filtered through an in-line Sterivex 0.22 μm filter unit (Millipore, Bedford, MA) to remove particulate matter. With each heart, polyethylene tubing was inserted into the apex to drain the left ventricle, after which a custom-prepared fluid-filled plastic film balloon connected to a pressure transducer (ADInstruments, Colorado Springs, CO) was inserted into the left ventricle via the left atria for continuous measurement of left ventricular pressure. The hearts were immersed in perfusion buffer maintained at 37°C, and the balloons were inflated to achieve end-diastolic pressures of 5–10 mm Hg. The left ventricular pressure signals were acquired continuously using a PowerLab data acquisition system (ADInstruments) and processed (Chart software) to yield heart rates and left ventricular dP/dts. Coronary flows were monitored by an ultrasonic in-line flow-probe connected to a T206 flowmeter (Transonic Systems, Inc., Ithaca, NY). To allow for pacing, the left ventricles were pierced with 0.125 mm silver electrodes connected to a Grass SD9 stimulator (Astro-Med, Inc., West Warwick, RI).

The experimental protocol was as follows. Once prepared, the hearts were perfused for 20 min while being paced at 420 bpm (2 ms square waves 20% above threshold) to allow for stabilization and acquisition of baseline parameters. Subsequently, they were subjected to 20 minutes of no-flow global ischemia followed by 45 minutes of reperfusion, achieved by closing and opening an in-line stopcock. Hearts were randomly assigned to be perfused with buffer containing either CP-532,903 (100 nM) or equivalent vehicle (less than 0.1 %) for 10 min before the induction of ischemia. Left ventricular function was continually assessed prior to ischemia and for 45 minutes after flow was resumed. All experiments and data analyses were conducted in a blinded manner without knowledge of drug treatment or mouse heart genotype. Headrick and colleagues have previously documented that adenosine efflux increases ~150-fold following global, normothermic ischemia (20 min) using a similar isolated mouse heart model [31].

2.12. Data Analysis

All data are reported as means ± SEM. Left ventricular functional parameters and I_KATP densities were compared by an unpaired Student’s t-test or a one-way ANOVA, as appropriate. P<0.05 was considered statistically significant.

3. Results

3.1. Identification of A₃ARs in Cardiomyocytes: mRNA Expression and Electrophysiology

Due to lack of useful commercially available antibodies, it is currently not possible to assess expression of ARs in cells/tissues by conventional immunological methods. By qPCR, A₃AR transcripts were detected in total RNA prepared from isolated cardiomyocyte preparations, although the level of expression was very low compared to that of A₁ (150-fold less) and A_2AAR (28-fold less) mRNAs (Table 2). Low-level expression of A_2BAR transcripts was also detected in cardiomyocytes.

Table 2.

Adenosine Receptor mRNA content in isolated cardiomyocytes and cardiac fibroblasts determined by qPCR.

	Cardiomyocytes (n=4)	Cardiac Fibroblasts (n=3)
A1	12,830 ± 2,534	8 ± 4
A2A	2,388 ± 304	4,083 ± 141
A2B	330 ± 45	40,940 ± 2,123
A3	85 ± 5	19 ± 3

copy number (mean ± SEM) per 100 ng total RNA

Because we have previously observed that A₃AR agonist-mediated cardioprotection is blocked in some model systems by glibenclamide [12, 18], we proceeded to assess coupling of the A₃AR to K_ATP channel opening. I_KATP was recorded in freshly isolated mouse left ventricular cardiomyocytes using the whole-cell configuration of the patch clamp technique. Basal whole-cell current was monitored for 20 minutes to allow for diffusional exchange of ATP between the pipette solution, present at a low concentration (0.2 mM), and the intracellular milieu. This protocol reduced inhibition of K_ATP channels by ATP and mimicked reductions in ATP that occur during metabolic compromise. At this time, the effect of applying CP-532,903 (1 μM) or the A₁AR agonist CPA (1 μM) on I_KATP was measured by 100-ms test pulses from −120 to +60 mV from a holding potential of −40 mV. To confirm specificity of the agonists, recordings were obtained from myocytes isolated from global Adora1^−/− or global Adora3^−/− mice. Myocytes that exhibited spontaneous activation of outward current immediately upon establishing whole-cell configuration prior to drug application were discarded.

As shown in Figure 3, application of CPA to wild-type myocytes increased I_KATP, defined by glibenclamide (1 μM)-sensitivity. This effect of CPA was inhibited by exposure to pertussis toxin (Figure 5) and was also apparent in myocytes obtained from Adora3^−/− mice, but not from Adora1^−/− mice (Figure 3). Currents evoked in the presence of CPA were subtracted from those in the absence of the agonist to yield the drug-induced contribution to the current. Current densities, normalized to cell capacitance, were determined to be 2.70 ± 0.65 and 2.51 ± 0.76 pA/pF at test potentials of 0 and +40 mV, respectively (Figure 3). As demonstrated in Figures 4 and 5, CP-532,903 also increased I_KATP via a pertussis toxin-sensitive mechanism in wild-type cardiomyocytes, although it was slightly less in magnitude (1.86 ± 0.31 and 1.95 ± 0.55 pA/pF at test potentials of 0 and +40 mV) compared to that elicited by CPA. Based on findings with gene-deleted mice, this effect of CP-532,903 was dependent on expression of the A₃AR (but not the A₁AR). These findings indicate that both A₁ and A₃ARs, coupled to activation of K_ATP channels, are expressed in mouse ventricular cardiomyocytes. As hypothesized previously [22, 23], these findings are consistent with the idea that stimulation of these receptors activate G_i proteins in cardiomyocytes, which reduces sensitivity of K_ATP channels to ATP thereby causing them to open.

Figure 3. Evidence for coupling of A₁ARs to K_ATP channel activation in adult mouse cardiomyocytes.

A and B. Representative whole-cell current traces recorded from an isolated cardiomyocyte from a wild-type (WT) C57B/6 mouse after application of CPA (1 μM) in the presence or absence of glibenclamide (1 μM). The glibenclamide-sensitive current shown in B was obtained by digitally acquiring the CPA-elicited current that was sensitive to inhibition by glibenclamide. C and D. Representative whole-cell currents recorded from an isolated cardiomyocyte from either an Adora1^−/− mouse (C) or and Adora3^−/− mouse (D) after application of CPA with corresponding current-voltage relationships of the CPA-elicited current obtained by digital subtraction. E) Summary (n=5–7; means ± SEM) of the effects of CPA on I_KATP showing dependence on expression of the A₁ but not the A₃AR. Current density was determined at the 0 and +40 mV test pulses from a −40 mV holding potential. Current density, determined by normalizing current amplitude to cell capacitance, is presented to account for differences in cell size. All recordings were obtained as described in METHODS in the presence of ZM241385 (100 nM) and PSB603 (100 nM) to block A_2A and A_2BARs. *P<0.05 vs WT.

Figure 5. CPA- and CP-532,903-elicted I_KATP are blocked by treatment with pertussis toxin.

Representative whole-cell currents recorded from pertussis toxin-treated cardiomyocytes (2 μg/ml for a minimum of 1 h) isolated from wild-type C57Bl/6 mice after application of CPA (1 μM; A) or CP-532,903 (1 μM; B). Both agonists failed to significantly elicit I_KATP after pertussis toxin treatment. Similar results were obtained in recordings from 4–6 additional cells per treatment group. All recordings were obtained as described in METHODS in the presence of ZM241385 (100 nM) and PSB603 (100 nM) to block A_2A and A_2BARs.

Figure 4. Evidence for coupling of A₃ARs to K_ATP channel activation in adult mouse cardiomyocytes.

Representative whole-cell currents recorded from an isolated cardiomyocyte from either a wild-type (WT) C57B/6 (A), Adora1^−/− (B), or Adora3^−/− (C) mouse after application of CP-532,903 (1 μM) with corresponding current-voltage relationships of the CP-532,903-elicited current obtained by digital subtraction. D. Summary (n=5–7; means ± SEM) of the effects of CP-532,903 on I_KATP showing dependence on expression of the A₃ but not the A₁AR. Current density was determined at the 0 and +40 mV test pulses from a −40 mV holding potential. All recordings were obtained as described in METHODS in the presence of ZM241385 (100 nM) and PSB603 (100 nM) to block A_2A and A_2BARs. *P<0.05 vs WT.

3.2. Production of Adora3^{LoxP/LoxP;Myh6-Cre} Mice

Mice with the experimental genotype Adora3^{LoxP/LoxP;Myh6-Cre} and the control genotype Adora3^LoxP/LoxP were prepared (Figure 1). The LoxP sites located in the 5’ non-coding region and in the first intron were designed to undergo recombination by Cre recombinase, the Myh6-Cre driver for which becomes strongly expressed in ventricular cardiomyocytes near the time of birth [30]. This resulted in ablation of the first exon, along with ~2 kb of non-coding DNA upstream of exon 1 that contained the Adora3 promoter [32]. As shown in Figure 6, PCR amplification of the Adora3 allele in Adora3^{LoxP/LoxP;Myh6-Cre} mice revealed that the LoxP-targeted allele was disrupted in a tissue-specific manner, as revealed by the 190 bp product indicative of recombination in the heart, which was not detected in samples from seven other organs or from skeletal muscle. Placement of LoxP sites into the Adora3 allele did not interfere with expression or function of the A₃AR, demonstrated by retained ability of the A₃AR agonist Cl-IB-MECA to increase plasma histamine levels in Adora3^LoxP/LoxP mice resultant from A₃AR-induced mast cell degranulation (Figure 6). Cl-IB-MECA also increased plasma histamine levels in Adora3^{LoxP/LoxP;Myh6-Cre} mice to an equivalent extend as compared to both Adora3^LoxP/LoxP and C57BL/6 wild-type mice, indicating specificity of the Myh6-Cre transgene (Figure 6).

Figure 6. Myh6-Cre disrupts the LoxP-flanked A₃AR gene (Adora3) specifically in myocardium.

Panel A: Depiction of Adora3^LoxP allele before and after Myh6-Cre-mediated recombination. Recombination was assessed by PCR amplification of genomic DNA obtained from a panel of organs of a 12 week-old adult male Adora3^{LoxP/LoxP;Myh6-Cre} mouse. Each lane contains a PCR reaction of genomic DNA obtained from the indicated tissues. The positive control lane used DNA isolated from cultured ear fibroblasts from an Adora3^LoxP/LoxP mouse infected with an adenovirus that expresses Cre recombinase. The negative control lane contained a PCR reaction using genomic DNA obtained from heart tissue of a 10 week-old male Adora3^LoxP/LoxP mouse. An amplicon was only generated with heart DNA from the Adora3^{LoxP/LoxP;Myh6-Cre} mouse, indicating cardiomyocyte-selective recombination. Panel B: Administration of Cl-IB-MECA (100 μg/kg i.p.) to Adora3^LoxP/LoxP or Adora3^{LoxP/LoxP;Myh6-Cre} mice (n=3–7) increased plasma histamine levels resultant from A₃AR-mediated mast cell degranulation to an equivalent extent as compared to wild-type C56Bl/6 mice.

3.3. Isolated Heart Studies

Hearts were perfused by the Langendorff procedure, as described in METHODS. The experimental designed involved comparing the effects of treatment with CP-532,903 on hearts from mice with the experimental genotype Adora3^{LoxP/LoxP;Myh6-Cre} or control mice having the Adora3^LoxP/LoxP genotype while pacing at a rate of 420 beats/min. CP-532,903 was infused into the coronary circulation of the hearts for 10 min immediately before 20 min of global ischemia. Recovery of left ventricular contractile function (developed pressure, maximal rate of contraction, and maximal rate of relaxation) was used as an index of the extent of ischemia/reperfusion injury.

As shown in Table 3, baseline functional parameters and coronary flow rates were not significantly different between Adora3^{LoxP/LoxP;Myh6-Cre} and Adora3^LoxP/LoxP hearts. Neither coronary flow rate nor contractile properties were changed during infusion of CP-532,902 into hearts of either genotype (Table 3). These observations concur with our previous study, which demonstrated that infusion of CP-532,903 into spontaneously beating isolated mouse hearts up to a concentration of 100 nM did not alter left ventricular function, coronary flow rate, or heart rate, whereas infusion of an A₁ (CCPA) or an A_2AAR (CGS 21680) agonist produced prominent changes in all of these parameters [18].

Table 3.

Functional parameters of isolated hearts at baseline and after infusion of CP-532,903 or vehicle while pacing at 420 beats/min.

Genotype	N	Treatment	Devel. Pressure (mmHg)	+dP/dt (mmHg/s)	−dP/dt (mmHg/s)	Coronary Flow (ml/min/g)
Adora3LoxP/LoxP	Male = 9	baseline	93.3 ± 2.0	3,704 ± 263	−2,006 ± 406	22.4 ± 1.8
	Female = 4	vehicle	94.9 ± 1.8	3,765 ± 253	−1,985 ± 401	22.1 ± 2.0
	Total = 13
	Male = 7	baseline	86.5 ± 3.5	3,098 ± 467	2,254 ± 118	23.1 ± 1.1
	Female = 3	CP-532,903	86.2 ± 3.2	3,238 ± 476	2,273 ± 116	22.3 ± 1.0
	Total = 10
Adora3LoxP/LoxP;Myh6-Cre	Male = 7	baseline	79.2 ± 7.0	3,248 ± 408	−2,108 ± 226	21.9 ± 1.1
	Female = 6	vehicle	80.5 ± 6.7	3,299 ± 390	−2,127 ± 229	21.0 ± 0.9
	Total = 13
	Male = 7	baseline	88.6 ± 1.8	3,553 ± 197	−2,249 ± 111	25.2 ± 1.8
	Female = 6	CP-532,903	88.8 ± 1.9	3,571 ± 192	−2,255 ± 112	24.6 ± 2.1
	Total = 13

With all hearts, contractile function diminished upon the onset of ischemia with complete arrest within 5 min. With the onset of reperfusion, the hearts resumed spontaneous contraction within ~2 min, and then after a short period of decline contractile function progressively improved for the remainder of the reperfusion period. Left ventricular developed pressure data during reperfusion of control Adora3^LoxP/LoxP hearts treated with either vehicle or CP-532,903 is shown in Figure 7 (panel a). As is common practice with the isolated mouse heart model [15, 18, 19, 33], the data in Figure 7 are expressed as a percentage of pre-ischemic values to control for interindividual variability within the experimental groups. As reported previously [18], control Adora3^LoxP/LoxP hearts treated with CP-532,903 exhibited better recovery of left ventricular developed pressure compared to those treated with vehicle at nearly all time-points assessed during reperfusion. Final recovery of developed pressure was 70.1 ± 2.7% of baseline in CP-532,903-treated hearts versus 58.1 ± 3.3% in vehicle-treated hearts (P<0.05). Other contractile parameters including maximal +dP/dt and −dPdt were also significantly improved by CP-532,903 by ~15–20% (P<0.05), whereas the rate of coronary flow was unaffected (Figure 7a). Although group sizes were too small to show significant differences, when analyzed individually there was a clear trend for CP-532,903 to provide improvement in hearts from both male and female Adora3^LoxP/LoxP mice (not shown).

Figure 7. CP-532,903 does not protect hearts from Adora3^{LoxP/LoxP;Myh6-Cre} mice in which Adora3 is deleted specifically in cardiomyocytes form ischemia/reperfusion injury.

Effect of CP-532,903 on recovery of left ventricular developed pressure (expressed as a percentage of baseline) of isolated hearts from Adora3^LoxP/LoxP (Panel A) and Adora3^{LoxP/LoxP;Myh6-Cre} (Panel B) mice subjected to 20 min of global ischemia and 45 min of reperfusion. Hearts were treated with either CP-532,903 (100 nM) or equivalent vehicle for 10 min prior to induction of ischemia. Insets show final recovery of left ventricular maximal ±dP/dt and coronary flow rate data at the end of the 45-min reperfusion period. Values shown are means ± SEM. *P<0.05 vs. vehicle-treated group. n = 10–13.

The effect of CP-532,903 on hearts from Adora3^{LoxP/LoxP;Myh6-Cre} mice in which Adora3 is deleted specifically in cardiomyocytes is shown in Figure 7 (panel b). In these studies, CP-532,903 did not improve functional recovery. If anything, the rate of recovery of CP-532,903-treated hearts appeared to be delayed compared to vehicle-treated hearts.

4. Discussion

A₃AR agonists hold promise as therapeutic agents for the treatment of several diseases including myocardial ischemia/reperfusion injury and have the advantage of being effective while producing no adverse hemodynamic effects. In the current study, we provide definitive evidence for the presence of A₃ARs coupled to activation of K_ATP channels in adult mouse ventricular cardiomyocytes. We subsequently found that the selective A₃AR agonist CP-532,903 lost effectiveness to reduce ischemia/reperfusion injury in an isolated, buffer-perfused heart model when Adora3 was deleted in cardiomyocytes. These findings identify cardiomyocyte A₃ARs as an important target of A₃AR agonist-mediated cardioprotection. Taking into consideration results of prior published work [12, 18], our findings are consistent with the hypothesis that A₃AR agonists provide ischemic tolerance, at least in part, by facilitating opening of myocardial K_ATP channels.

The current study utilized an isolated, buffer-perfused heart model to exclude contribution of blood components and a complete inflammatory response to the injury process. It is well established that the A₃AR is expressed in several different leukocyte populations, in particular granulocytic cells including neutrophils, where A₃AR signaling modulates chemotaxis and oxidant production by the NADPH oxidase complex [6–10]. In a previous study, we found that the A₃AR agonist Cl-IB-MECA reduced infarct size in an in vivo mouse model of ischemia/reperfusion when given at the time of reperfusion, and that this was dependent on expression of A₃ARs in bone marrow-derived cells [14]. In a separate study, IB-MECA reduced contractile dysfunction caused by infusion of neutrophils into the coronary circulation of post-ischemic, buffer-perfused isolated rat hearts [16]. Thus, the cardioprotective mechanisms of A₃AR agonists are likely multifactorial. Current evidence suggests some contribution of inflammation suppression and reduced immune cell-mediated reperfusion injury. Because A₃AR agonists stimulate mast cell degranulation in some rodents including mice [6, 11], another contributing mechanism in these species may relate to potential benefit from depleting mast cells of their contents prior to the onset of ischemia [6, 17]. With the isolated heart model, it is important to recognize that post-ischemic contractile dysfunction results from a combination of both reversible injury processes (i.e., myocardial ‘stunning’) and cell death (i.e., infarction). Whether CP-532,903 protected against one or both of these forms of injury was not addressed in the present investigation, although we speculate a broad effect considering A₃AR agonists have been shown in a number of different model systems to reduce infarct size, to promote cardiomyocyte survival, and to attenuate myocardial stunning [5, 12–20].

Traditionally, the A₁AR has been viewed as the only AR subtype expressed in cardiac muscle. In ventricular muscle cells, A₁AR signaling reduces the positive inotropic effects of β adrenergic receptor stimulation, by coupling through G_i proteins to multiple effectors including adenylyl cyclase, potassium channels, and calcium channels [24]. As reported previously using neonatal (rat)/juvenile (guinea pig) cells [22, 23] and as demonstrated in this study using adult mouse cells, the A₁AR also signals to facilitate opening of K_ATP channels in ventricular cardiomyocytes. With the development of more selective pharmacological tools and genetically modified mice, however, it has become apparent that multiple ARs are expressed in cardiomyocytes [2]. For instance, it is now well accepted that A_2AARs are expressed in the myocardium that modulate contractile responses and potentially the release of pro-growth/pro-inflammatory factors [2]. There is evidence for cardiomyocyte expression of A_2BARs that may modulate metabolism and responses to hypoxic stress [34]. Our finding in the current study for presence of A₃ARs in mouse cardiomyocytes, despite low transcript levels, needs to be confirmed in other species. It will be important to determine whether the A₃AR may couple to additional signaling pathways in cardiomyocytes.

It is necessary to point out in our studies that post-ischemic recovery of contractile function of vehicle-treated Adora3^{LoxP/LoxP;Myh6-Cre} hearts with Adora3 deleted in cardiomyocytes tended to be better compared to vehicle-treated Adora3^LoxP/LoxP hearts. Final recovery of developed pressure was 58.1 ± 3.3% of baseline in vehicle-treated Adora3^LoxP/LoxP hearts vs. 70.0 ± 2.2% in vehicle-treated Adora3^{LoxP/LoxP;Myh6-Cre} hearts. Recovery of other contractile parameters was also better (Figure 7). It is tempting to conclude from this comparison that cardiomyocyte A₃ARs play a damaging role during the pathogenesis of ischemia/reperfusion injury even though these same receptors are the molecular site of action of A₃AR agonist-mediated cardioprotection. Perhaps both A₃AR agonism and antagonism, depending on timing, may attenuate ischemia/reperfusion injury, potentially explained by the unique biology of the A₃AR such as exceptionally rapid receptor desensitization [35, 36] or susceptibility to biased agonism [37, 38]. However, we believe that it is not prudent to make direct comparisons between these genotypes in our studies due to the presence of the Myh6-Cre transgene in Adora3^{LoxP/LoxP;Myh6-Cre} hearts. It is also possible that unanticipated compensatory changes may arise from cardiomyocyte-specific deletion of Adora3 [15, 39]. For example, it is theoretically possible for expression or cellular localization of other members of the AR family to be impacted by loss of A₃AR signaling [40]. This observation requires further investigation and must be aided by experimental data with A₃AR antagonists as they become available. Interestingly, mice with global deletion of Adora3 have also been reported to be more resistant to ischemia/reperfusion injury, but whether this is attributed to loss of A₃AR signaling, non-specific compensatory changes, or genetic background effects remains uncertain [15, 33, 39, 41]. Our targeting strategy for depletion of the A₃AR in the present investigation was designed to avoid disruption of the transmembrane and immunoglobulin domain containing 3 splice variant (TMIGD3i3 or Adora3i3) that shares exon 2 of Adora3 with the A₃AR [42].

In conclusion, we achieved three important goals in this study. First, we provide evidence for functional expression of A₃ARs in the mouse myocardium, which are coupled to K_ATP channel activation. Second, and perhaps most importantly, we successfully prepared a genetically modified line of mice that allows for conditional depletion of the A₃AR in mice by means of the LoxP-Cre recombinase strategy. This new tool will be a valuable resource for future investigation of the complex biology of the A₃AR in mice. Finally, utilizing this tool we provide new information regarding the underlying causes of A₃AR agonist-mediated cardioprotection. Additional studies are needed to assess additional A₃AR agonists with different signaling profiles [37, 38], to examine the effect of deletion of Adora3 in other cell types, and to expand the analyses to in vivo models. Our findings further support the idea that A₃AR agonism may be a useful approach for limiting ischemia/reperfusion injury of the heart.

Abbreviations:

AR

(adenosine receptor)

IB-MECA

[N⁶-(3-iodobenzyl)adenosine-5’-N-methyluronamide]

Cl-IB-MECA

[2-chloro-N⁶-(3-iodobenzyl)adenosine-5’-N-methyluronamide]

K_ATP

(ATP-dependent potassium channel)

CP-532,903

(N⁶-(2,5-dichlorobenzyl)-3’-aminoadenosine-5’-N-methylcarboxamide]

CCPA

(2-chloro-N⁶-cyclopentyladenosine)

CPA

(N⁶-cyclopentyladenosine)

Adora3

(A3 adenosine receptor gene)

ZM241385

[4-[2-[7-(2-furyl)[1,2,4]triazolo-[2,3-a][1,3,5]triazine-5-ylamino]ethyl]phenol]

PSB603

([8–4-[4-[4-[chlorophenzyl)piperazide-1-sulfonyl)phenyl]]-1-propylxanthine]