Cardiomyocyte death induced by ischaemic/hypoxic stress is differentially affected by distinct purinergic P2 receptors

Abstract Blood levels of extracellular nucleotides (e.g. ATP) are greatly increased during heart ischaemia, but, despite the presence of their specific receptors on cardiomyocytes (both P2X and P2Y subtypes), their effects on the subsequent myocardial damage are still unknown. In this study, we aimed at investigating the role of ATP and specific P2 receptors in the appearance of cell injury in a cardiac model of ischaemic/hypoxic stress. Cells were maintained in a modular incubator chamber in a controlled humidified atmosphere of 95% N2 for 16 hrs in a glucose-free medium. In this condition, we detected an early increase in the release of ATP in the culture medium, which was followed by a massive increase in the release of cytoplasmic histone-associated-DNA-fragments, a marker of apoptosis. Addition of either apyrase, which degrades extracellular ATP, or various inhibitors of ATP release via connexin hemichannels fully abolished ischaemic/hypoxic stress-associated apoptosis. To dissect the role of specific P2 receptor subtypes, we used a combined approach: (i) non-selective and, when available, subtype-selective P2 antagonists, were added to cardiomyocytes before ischaemic/hypoxic stress; (ii) selected P2 receptors genes were silenced via specific small interfering RNAs. Both approaches indicated that the P2Y2 and P2χ7 receptor subtypes are directly involved in the induction of cell death during ischaemic/hypoxic stress, whereas the P2Y4 receptor has a protective effect. Overall, these findings indicate a role for ATP and its receptors in modulating cardiomyocyte damage during ischaemic/hypoxic stress.


Introduction
It is now widely accepted that, in addition to functioning as an intracellular energy source, extracellular adenine and uracil nucleotides initiate a wide range of intracellular signalling cascades through the activation of a plethora of purinergic receptors (the P2 receptors). The purines and pyrimidines ATP, adenosine diphosphate (ADP), uridine triphosphate (UTP), uridine diphosphate (UDP), and, as more recently recognized, sugar nucleotides such as UDP-glucose and UDP-galactose [1][2][3][4] indeed act on various P2X homo-and heteromultimeric ionotropic and P2Y metabotropic receptor subtypes [3]. In the cardiovascular system, evidence is accumulating to indicate that P2 receptors mediate important physiologic effects, including vasoconstriction and vasodilatation, growth of vascular smooth muscle and endothelial cells, angiogenesis, vascular remodelling, coagulation, platelet aggregation, inflammation and several aspects of cardiac function. The purinergic system and its receptors are also involved in development of myocardial infarction (MI), heart failure and xenograft rejection [5,6].
However, although the heart has been the first organ to be tested for the physiological actions of nucleotides in early work using adenylic acid [7], the effects mediated by specific myocardial P2 receptors are just now beginning to be elucidated. P2 receptors are abundantly expressed in the foetal human heart [8] as well as in the adult human heart [9][10][11]. In cardiomyocytes, ATP acting through P2Y and P2X receptors stimulates a pronounced positive inotropic effect and may also act in synergy with ␤-adrenergic agonists to augment myocyte contractility (Vassort,summarized in 12). Besides adenine nucleotides, uracil nucleotides # These authors contributed equally to this work. *Correspondence to: Maria P. ABBRACCHIO, Dipartimento di Scienze Farmacologiche, Università degli Studi di Milano, Via Balzaretti 9, 20133 Milano, Italy. Tel.: ϩ39-02-50318304/355 Fax: ϩ39-02-50318284 E-mail: mariapia.abbracchio@unimi.it have been also recently reported to act as positive inotropic agents in rat and mouse cardiomyocytes [10]. In addition, potent effects may be exerted by nucleotides on cardiomyocytes viability, under both physiological conditions and after hypoxia, when cardiomyocytes themselves release massive amounts of ATP. Given its importance as an energy substrate, ATP release is tightly controlled and occurs by several mechanisms such as secretory vesicles, ABC transporter cassette proteins, connexin hemichannels [13], pannexin channels [14] as well as through other channels with large unitary conductances [15]. In line with this hypothesis, short-term activation of UTP receptors protects cardiomyocytes against hypoxic damage and significantly reduces the cell death caused by hypoxia via activation of the P2Y2 and P2Y4 receptors [16]. In a previous study, we have investigated the effects of both adenine and uracil nucleotides on the viability of HL-1 cardiomyocytes, the only available cell line that spontaneously contracts in vitro and maintains a differentiated cardiac phenotype [17]. We showed that murine HL-1 cardiomyocytes express a wide panel of P2X and P2Y receptors known to either exclusively respond to adenine nucleotides (P2X receptors), to both adenine and uracil nucleotides (P2Y2, P2Y4, P2Y6) or to sugar nucleotides (P2Y14 receptor) [18]. Such a large heterogeneity of P2 receptor expression is consistent with previous studies [8,11], and suggests involvement of these receptors in multiple functional effects. We further demonstrated that the exposure of cardiomyocytes to high concentrations of adenine nucleotides (ATP, ADP or BzATP) induces cardiomyocyte cell death through a mechanism involving both P2Y and P2X receptors [18]. Thus, besides influencing cardiac contractility, P2 receptors may also directly regulate the viability of myocardial cells.
In this study, we set up and characterized an hypoxia/ ischaemia protocol in HL-1 cardiomyocytes to evaluate (a) whether ATP is endogenously released by these cells and possibly plays a role in induction of cell death under these conditions; (b) whether ischaemia-associated cardiomyocyte death is influenced by pharmacological agents known to act on either ATP release/availability or on P2 receptors, with the final aim to; (c) identify the specific P2 receptor subtypes involved in regulation of cardiomyocyte viability. Because apoptosis has a central role in MI, we focused our attention on this type of cell death. Results may have important therapeutic implications and set the basis for the development of novel cardioprotective agents that targetspecific P2 receptor subtypes.

Cell culture
HL-1 cells, a cardiac muscle cell line derived from the AT-1 mouse atrial myocyte tumour lineage, were a gift from William C. Claycomb, and maintained according to described protocols [17,19]. In separate experimental groups, cells received no intervention (normoxia control, 95% air and 5% CO2) or were exposed to ischaemic/hypoxic stress. Hypoxia was produced by exposure to 5% CO2 and 95% N2 in a modular incubator chamber for 16 hrs in the presence of serum-and glucose-free DMEM medium. Control cells received vehicle or the indicated compounds.

RNA interference and cell transfection
Validated high-performance purity grade small interfering RNAs (siRNA) against mouse P2Y2, P2Y4, P2Y6, and P2ϫ7 were synthesized by Qiagen. Control siRNA, with a non-silencing oligonucleotide sequence that does not recognize any known homology to mammalian genes, was also generated as a negative control, as described in Ref. [20].

Nuclear staining of adhering cells
After cell treatment, cells were fixed in 4% paraformaldehyde in phosphate buffered saline and nuclear chromatin stained by using the fluorescent dye Hoechst 33258 (Sigma-Aldrich), as described in Ref. [21].

Ischaemic/hypoxic stress induces ATP release in HL-1 cardiomyocytes
Released cellular ATP can serve as a paracrine signalling molecule for intercellular communication, or can act in an autocrine manner to regulate cellular functions, including those of cardiomyocytes. Cultured HL-1 cardiomyocytes exhibited a basal release of ATP (approximately equal to 1.25 nmol/l) that was significantly increased at 30 min. after the induction of an ischaemic/hypoxic stress (approximately 2.28 nmol/l; Fig. 1A). To ensure that, at this time point, membrane damage did not contribute to ATP release, we analysed the conditioned media from ischaemic/hypoxic cells for the presence of the cytosolic enzyme lactate dehydrogenase (LDH), a marker of cellular damage. We consistently found that ischaemic/hypoxic-induced ATP release occurred in the absence of LDH release, hence with no significant plasmalemmal damage compared with normoxic controls (data not shown), suggesting that, at this early stage of hypoxia, ATP does not derive from cell necrosis but from release via transmembrane channels/pumps.

Ischaemic/hypoxic stress promotes apoptosis of HL-1 cardiomyocytes
Ischaemic/hypoxic stress induced significant increases in the appearance of HL-1 cardiomyocytes apoptosis, as assessed by two complementary assays including annexin V and nucleosomal immunoassay. An increased number of annexin-positive cardiomyocytes was registered after 16 hrs of exposure to ischaemic/ hypoxic stress (Fig. 1B). We found that, at this time after ischaemic/ hypoxic stress induction, the number of PI-positive cells was also increased compared to control value (ϩ72.2% Ϯ 9.9, P Ͻ 0.05, data not shown). To confirm the membrane-related changes associated with apoptosis, we performed nucleosomal ELISA assay, which detect later stages of apoptosis. ELISA analysis showed a strong increase in the apoptotic index, up to 320.9% Ϯ 15.8 of control value at 16 hrs (Fig. 1C). Because one important sign of apoptosis is the appearance of specific morphological changes, including nuclear condensation and chromatin fragmentation, we used Hoechst 33258, a fluorescent DNA-binding dye, to stain nuclear chromatin in HL-1 cardiomyocytes, and examined cell morphology by phase-contrast microscopy. Figure 1D and E shows the phase contrast morphology of HL-1 cardiomyocytes exposed to ischaemic/hypoxic stress or to normoxia. Control cultures have normal cell bodies and intact features with nuclei homogeneous in size and shape and very lightly stained with Hoechst 33258 (Fig. 1D). However, when HL-1 cardiomyocytes were exposed to ischaemia/hypoxia for 16 hrs, we noticed morphologic alterations typical of cells undergoing apoptosis, which became rounded and condensed and lost adhesion from the monolayer. Furthermore, HL-1 cells displayed condensed chromatin that was brightly and uniformly stained by Hoechst 33258 and ranged in shape from a single uniform sphere to a collection of multiple chromatin dots typical of fragmented apoptotic nuclei (Fig. 1E).

Role of ATP in the ischaemic/hypoxic stress-induced apoptosis in cardiomyocytes
Endogenously released ATP plays an important role in mediating HL-1 cardiomyocytes apoptosis, as treatment with apyrase, a nucleotidase that degrades nucleotide triphosphates into nucleotide monophosphates and thus completely abolishes the amount of ATP released in the medium during ischaemia/hypoxia stress ( Fig. 2A), significantly prevented the induction of apoptosis (Fig. 2B). Because there is, as expected, a constitutive basal release of ATP even in the absence of hypoxia, apyrase also significantly reduced basal ATP levels during normoxia ( Fig. 2A). However, this was not associated to any change in cell viability, to confirm that, under control conditions, ATP does not mediate cell death. To investigate the role of connexin hemichannels in the induction of apoptosis following ischaemic/hypoxic stress experiments were carried out in the presence of established inhibitors of these channels. Gap26, a highly specific inhibitor of connexin hemichannels, significantly prevented the induction of apoptosis induced by ischaemic/hypoxic stress (Fig. 2C). A further less specific inhibitor of connexin channels, 18aGA, was also protective (Fig. 2D). The maxi-anion channel blocker GdCl3, which has been previously involved in ATP release [15], also inhibited ischaemic/ hypoxic stress-associated apoptosis (data not shown). To further confirm the role of released ATP in the induction of apoptosis by ischaemic/hypoxic stress in HL-1 cells, we found that stimulation of HL-1 cardiomyocytes by ATP itself for 24 hrs at 250-500 mol/l increased apoptosis up to 254% Ϯ 37 of control values set to 100% (Fig. 2E).

P2 receptors antagonists prevent the appearance of HL-1 cardiomyocyte apoptosis induced by ischaemia/hypoxia
We have previously shown that HL-1 cardiomyocytes express all known P2X receptors (except for P2ϫ2), as well as the P2Y2,4,6,14 subtypes [18]. To verify if the apoptotic effect of the ischaemic/ hypoxic stress was actually due to activation of P2 receptors by ATP or other nucleotides released during hypoxia, HL-1 cardiomyocytes were exposed to ischaemic/hypoxic stress in the presence of non-selective P2 receptor antagonists, such as PPADS and suramin [2,23]. These treatments almost completely prevented the appearance of apoptosis (Ϫ82.7% Ϯ 7.7 after PPADS; Ϫ68% Ϯ 12 after suramin; Fig. 2F).

Inhibition of P2؋ 7 partially prevents the ischaemia/hypoxia-induced apoptosis
Our previous data are in support of a role of the P2X receptors in the induction of apoptosis mediated by ATP in HL-1 cardiomy-ocytes [18]. Indeed, a 24 hr-exposure of HL-1 cardiomyocytes to the selective P2ϫ7 agonist BzATP resulted in a dramatic reduction of the number of cells adhering to the culture substrate [18]. In this study, pre-treatment with TNP-ATP, an antagonist for P2ϫ1-4 receptors [26], failed to significantly prevent the appearance of apoptosis induced by ischaemic/hypoxic stress (Fig. 3A), thus ruling out a role for these P2X receptor subtypes in our model. By contrast, when the specific P2ϫ7 antagonist KN-62 was tested, a significant reduction of apoptosis in the ischaemic/hypoxic condition was monitored (Fig. 3A). To validate the effect of the P2ϫ7 antagonist, we silenced the expression of P2ϫ7 receptor using small interfering (si) RNAs. Treatment of HL-1 cells with a P2ϫ7 receptor siRNA construct reduced the P2ϫ7 receptor mRNA levels by 54% Ϯ 3.4 (Fig. 3B). The decrease in P2ϫ7 receptor expression resulted in a partial, albeit statistically significant, inhibition of apoptosis in response to the ischaemic/hypoxic stress (Ϫ34% Ϯ 2.7; Fig. 3C). Taken together, these results point to a possible contribution of the P2ϫ7 receptor in the apoptosis in HL-1 cardiomyocytes induced by ischaemic/hypoxic stress.

Inhibition of P2Y 2 prevents ischaemia/ hypoxia-induced apoptosis
To evaluate the role of P2Y receptor subtypes in the induction of apoptosis by the ischaemic/hypoxic stress, we used selective P2Y2 and P2Y6 antagonists. AR-C118925, the only known selective, heterocyclic antagonist of the P2Y2 receptor [27], significantly prevented the appearance of apoptosis by Ϫ50.8% Ϯ 2 (Fig. 4). By contrast, the potent antagonist of P2Y6 nucleotide receptor diisothiocyanate derivative MRS2578 [28] did not affect hypoxia-induced apoptosis in HL-1 cardiomyocytes (Fig. 4). This pharmacological approach could not be adopted for the P2Y4 receptor, because no selective antagonists are currently available for this receptor subtype. Globally, these pharmacological data support a role for P2Y2, but not P2Y6, in ischaemic/hypoxic stress-induced apoptosis. To validate these results, we adopted a second approach based on the use of specific siRNAs against the various P2Y receptors. Analysis of P2Y2, and P2Y6 mRNAs after transfection of HL-1 cardiomyocytes with specific siRNAs showed a reduction of the corresponding receptor mRNAs of 70% Ϯ 6.8 and 89% Ϯ 3.6, for the P2Y2 and P2Y6 receptors, respectively ( Fig. 5A and B). Semi-quantitative real-time analysis could not be performed for the P2Y4 receptor in silenced cells, due to the lack of efficacious oligonucleotide primers for this type of analysis [29]. However, standard RT-PCR in cardiomyocytes transfected with P2Y4 siRNA showed a dramatic reduction of P2Y4 amplification product at the expected 447 bp (Fig. 5C) with respect to cells transfected with corresponding negative siRNAs.
In line with the pharmacological data of Figure 5, silencing of P2Y2 resulted in a marked prevention of apoptosis following ischaemic/hypoxic stress (Fig. 5D). No effect was instead detected after silencing of P2Y6 (Fig. 5E). Of interest, gene silencing of P2Y4 resulted in a significant increase of cell apoptosis during ischaemic/hypoxic stress, thus suggesting a potential protective role for this receptor (Fig. 5F).

Both G proteins and protein kinase C are involved in the cardiomyocyte apoptosis induced by hypoxia and ATP
To shed some light on the post-receptor mechanisms involved in ischaemic/hypoxic stress and ATP-induced apoptosis of HL1 cardiomyocytes, we also performed experiments with either the selective Gi protein inhibitor PTX, the pan-G protein inhibitor GDP-␤-S or the protein kinase C inhibitor GF. Both G-protein inhibitors partially but significantly inhibited hypoxia-associated apoptosis, whereas the protein kinase C inhibitor showed a non-statistically significant trend to a reduction (Fig. 6A). When apoptosis was induced by exposure to high ATP concentrations, a highly significant reduction of cell death was observed in the presence of either PTX or GF (Fig. 6B). No significant effects were induced by these agents on normoxic cells.

Discussion
Here we show that ATP and P2 receptors play a key role in ischaemic/hypoxic stress-induced apoptosis of HL-1 cardiomy-ocytes, that express various P2X and P2Y receptors (Ref. 18 and this issue). Specifically, we demonstrate that (i) ischaemic/ hypoxic stress induces rapid ATP release from cultured cardiomyocytes, and that (ii) distinct P2 receptors regulate cardiomyocyte death, with both pro-and anti-apoptotic effects. These actions are likely mediated by direct stimulation of P2 receptors by ATP, although we cannot exclude that other concomitantly released nucleotides (e.g. uracil nucleotides, see also below) are also involved. Moreover, by using pharmacological inhibitors and gene silencing of selected P2 receptors, we provide evidence for a specific role of P2ϫ7 and P2Y2 receptors in ischaemic/hypoxic stress-induced cardiomyocyte apoptosis. Our data also suggest a protective role for P2Y4 receptors; conversely, no effects are apparently induced by co-expressed P2Y6 receptors. Finally, we provide initial hints involving G-proteins and PKC in the postreceptor mechanisms at the basis of hypoxia-and ATP-induced cardiomyocyte apoptosis.
These data are relevant to the pathogenic mechanisms of myocardial damage during hypoxia/ischaemia in patients. Under physiological conditions, in both rat and human heart and in human plasma, extracellular ATP levels are around 20-40 nmol/l [30]. Upon hypoxia/ischaemia, due to both increased exocytosis from cells' vesicular pool and to cytoplasmic release, these levels are markedly increased to micromolar concentrations [31]. More recently, during MI, human venous plasma levels of UTP have been also demonstrated to increase to about 10% of the ATP concentrations [10]. It has been postulated that, under these conditions, nucleotides release is originally meant at augmenting heart function via a positive inotropic effect mainly exerted through myocardial P2Y2 receptors (although P2Y4 and P2Y6 receptors may also contribute) (ibidem). However, if heart reperfusion is impaired or prevented and nucleotides are not removed, their local concentrations (especially those of ATP) increase to toxic levels and contribute to cardiomyocyte death. This hypothesis is consistent with our data demonstrating that ATP is released from cardiomyocytes very early after hypoxia/ischaemia, and that this release is later associated to apoptosis. It has to be underlined that, in our experiments, after ischaemic/hypoxic stress, the ATP concentration was found to double to 2.28 nM, a value that may seem still low to activate P2Y receptors (which are generally responsive to microMolar nucleotide concentrations, e.g. http://www.iuphardb.org/DATABASE/ObjectDisplayForward? objectIdϭ324&familyIdϭ52).
However, it may well be that the local ATP concentration at membrane receptors is much higher than that measured in the culture medium. It has been indeed previously recognized that, after being released, the ATP concentration in the vicinity of the cell surface is often different from the bulk concentration, being influenced by both delayed diffusion (due to unstirred layer effects), and by immediate degradation by membrane bound ectoATPAses [32].
Regarding the threshold exogenous ATP concentration needed to induce apoptosis, the micromolar values found in our experimental model are in line with previous literature data reporting apoptosis induction between 10 and 50 M to mM concentrations, depending on the cell type and specific conditions [33][34][35][36].
In our model, early release of ATP does not involve cell lysis; moreover, subsequent ATP-mediated apoptosis is prevented by the blockage of connexin hemichannels through which ATP is released. To confirm a causal relationship between ATP release and cardiomyocyte death, both the connexin hemichannels blockers (Gap26 and 18aGa), apyrase and P2 blockers inhibited cardiomyocyte apoptosis (the present data; Ref. 37). In line with these findings, apyrase expression was increased in ischaemic heart samples of patients [5,38]. Moreover, it has been recently shown that connexin hemichannels inhibitors abolished the release of ATP induced by simulated ischaemia also in rat neonatal cardiomyocytes [39]. As highlighted earlier, our data point to differential roles of various P2 receptors in cardiomyocyte viability during ischaemic/hypoxic stress. The pro-apoptotic role of P2Y2 receptors under hypoxic conditions is in line with previous data demonstrating this as the most expressed P2 receptor undergoing up-regulation in the human failing heart [40]. Moreover, three single nucleotide polymorphisms of the human P2Y2 receptor gene have been associated with MI and proposed as genetic markers for MI in Japanese men [41]. P2Y2 receptors also promote a profibrotic response in isolated myocardial fibrob-lasts; in mice lacking this receptor, extracellular nucleotides are no longer able to increase the expression of pro-fibrotic plasminogen activator inhibitor type 1 and ␣-smooth muscle actin [42]. Regarding P2ϫ7 receptor, despite abundant information on its role in induction of apoptosis in macrophages, monocytes, microglia and other cell types [43], very little is known on its presence and role in MI. Although no direct evidence for P2ϫ7 presence in rodent heart is available, in dog, this receptor has been originally cloned from heart cDNA [44]. Significant levels of P2ϫ7 have been found in human myocardial samples from normal subjects and patients with chronic heart failure [11]. Moreover, a loss-of-function P2ϫ7 polymorphism has been reported in heart failure patients [45], but the relevance of this finding is unknown. More recently, pannexin-1/P2ϫ7 channels have been shown to mediate the release of cardioprotectants upon ischaemic pre-and post-conditioning [46]. This is the first time that, based on both pharmacological and gene-silencing data, P2ϫ7 is directly implicated in ischaemia-associated cardiomyocyte death.
Our data also highlight a protective role for P2Y4. In cultured hypoxic cardiomyocytes, activation of some P2Y receptors by UTP reduced LDH release [16], prevented intracellular calcium elevation, maintained ATP intracellular levels [47] and preserved mitochondrial activity [48]. In line with these findings, in vivo experiments showed that UTP administration reduced infarct size and improved myocardial function [49]. However, the exact P2 receptor involved was not univocally identified in these studies. UTP can activate both P2Y2 and P2Y4 receptors [2]; this data suggest, for the first time, that it is the latter that mediates UTPinduced protection. Regarding the P2Y6 receptor that has been proposed to mediate positive inotropic effects [10], recent papers have involved this receptor in pressure overload-induced cardiac fibrosis [5,50]. However, our data suggest that, despite playing a role in heart remodelling during chronic heart failure, P2Y6 is not directly involved in regulation of cardiomyocyte viability during acute hypoxic stress. Although our study focused on the role of P2 receptors it is known that, upon addition to cells, ATP is quickly degraded to adenosine, which can then activate the P1 receptors (subdivided into the A1, A2A, A2B and A3 subtypes [51]). However in our experimental model, ATPinduced cytotoxicity was not counteracted by either 8cyclopentyl-1,3-dipropylxanthine (CPX, which primarily antagonizes the A1 and A2B receptors [52][53][54]), or by the selective A2A antagonist SCH58261, or by the selective A3 antagonist MRS1220 [51] (data not shown), thus ruling out a role for these receptors in the detected effects. Finally, this data shed some light on the post-receptor mechanisms involved in P2 receptor regulation of cardiomyocyte viability. Inhibition of both G-proteins and PKC resulted in apoptosis reduction, in line with our proposal that the G-protein coupled P2Y2 receptor participates to cardiomyocyte death. Globally, our results suggest that the pharmacological modulation of P2Y2 and P2ϫ7 receptors by mean of selective antagonists may represent a new therapeutic strategy in ischaemic heart disease.