Essential role of the Na+-Ca2+ exchanger (NCX) in glutamate-enhanced cell survival in cardiac cells exposed to hypoxia/reoxygenation

Myocardial ischemia culminates in ATP production impairment, ionic derangement and cell death. The provision of metabolic substrates during reperfusion significantly increases heart tolerance to ischemia by improving mitochondrial performance. Under normoxia, glutamate contributes to myocardial energy balance as substrate for anaplerotic reactions, and we demonstrated that the Na+/Ca2+ exchanger1 (NCX1) provides functional support for both glutamate uptake and use for ATP synthesis. Here we investigated the role of NCX1 in the potential of glutamate to improve energy metabolism and survival of cardiac cells subjected to hypoxia/reoxygenation (H/R). Specifically, in H9c2-NCX1 myoblasts, ATP levels, mitochondrial activities and cell survival were significantly compromised after H/R challenge. Glutamate supplementation at the onset of the reoxygenation phase significantly promoted viability, improved mitochondrial functions and normalized the H/R-induced increase of NCX1 reverse-mode activity. The benefits of glutamate were strikingly lost in H9c2-WT (lacking NCX1 expression), or in H9c2-NCX1 and rat cardiomyocytes treated with either NCX or Excitatory Amino Acid Transporters (EAATs) blockers, suggesting that a functional interplay between these transporters is critically required for glutamate-induced protection. Collectively, these results revealed for the first time the key role of NCX1 for the beneficial effects of glutamate against H/R-induced cell injury.


Effect of glutamate on H/R injury: involvement of EAATs.
It is widely accepted that plasma membrane Excitatory Amino Acid Transporters (EAATs) are primarily responsible for glutamate entry into the cells 25 . We have recently demonstrated that, in particular physiological conditions, glutamate entry into the cells through EAATs relies upon NCX activity 17 . Thus, once we have demonstrated that a functional NCX1 was required for glutamate-induced protection against H/R injury, we investigated EAATs involvement in this phenomenon. To this aim, we used the non-transportable EAATs blocker DL-threo-β-Benzyloxyaspartic acid (DL-TBOA) 26, 27 at the concentration of 300 µM. Exposure to glutamate for the entire reoxygenation phase failed to prevent H/R-induced cell death in the presence of DL-TBOA (Fig. 3a). Same results were obtained in primary culture of rat adult cardiomyocytes (Fig. 3b). These findings support the substantial role of EAATs in the observed glutamate-induced protective response. DL-TBOA per se does not affect cell viability neither in normoxia (data not shown) nor after H/R (Fig. 3).
Effect of glutamate exposure on ATP production. Thereafter, we explored the potential mechanism underlying the protective action exerted by glutamate and the role of NCX1. We recently reported that, in physiological conditions, NCX activity supports glutamate-enhanced ATP synthesis in several cell models 17,18 . Therefore, we investigated whether this effect could be involved in the recovery of cardiac metabolism from hypoxic state. We tested this hypothesis in our experimental model, first in normoxia and then during H/R. As shown in Fig. 4a, when cells were exposed to different glutamate concentration (0.5 and 1 mM) for 1 h, a remarkable increase in ATP synthesis occurred in H9c2-NCX1 (nmol/mg protein: 15.9 ± 0.61 and 22.6 ± 0.97 versus 10.8 ± 0.26, for glutamate 0.5 and 1 mM, respectively) but not in H9c2-WT cells (nmol/mg protein: 10.5 ± 0.34 and 11.8 ± 0.58 versus 9.2 ± 0.28, for glutamate 0.5 and 1 mM, respectively), in line with our previous results 17 . Given that H9c2-WT cells were refractory to glutamate stimulation ( Fig. 2b and Fig. 4a), we further analyzed H9c2-NCX1 cells. We tested the ability of glutamate to fuel ATP recovery during reoxygenation after hypoxia, and explored whether a NCX/EAAT functional coupling could play any role. For this set of experiments, we used glutamate at 1 mM according to the protocols described in Fig. 1. As shown in Fig. 4b, in cells underwent to H/R injury, the ATP content was significantly reduced compared to the control (nmol/mg protein: 5.8 ± 0.3 versus 10.0 ± 0.3). Glutamate administration during the first hour of reoxygenation evoked a raise in ATP production up to the levels observed under normoxic conditions (nmol/mg protein: 11.0 ± 1.1 versus 10.0 ± 0.3). Interestingly, the ability of glutamate to restore ATP levels was abolished by SN-6 (nmol/mg protein: 11.0 ± 1.1 versus 6.84 ± 0.66), suggesting that NCX1 activity is critical for the metabolism recovery promoted by glutamate during the reoxygenation. Since we observed that the protective effect of glutamate relied both on NCX1 and EAATs (Figs 2 and 3), which functionally interact to allow glutamate entry into the cytosol, improving the energetic balance of the cells 17,18 , we also tested whether the ATP response to glutamate during the reoxygenation was prevented by DL-TBOA (300 µM). The analysis of the cellular ATP content revealed that the ability of glutamate to restore ATP levels after hypoxia was abolished by DL-TBOA addition (Fig. 4c) confirming the involvement of EAATs in such glutamate-induced protective response. Both SN-6 and DL-TBOA do not affect ATP levels neither in normoxia 17 nor after H/R protocol (Fig. 4b,c).

Effect of glutamate on mitochondrial function following H/R challenge. The set of experiments
presented here were performed in order to further (a) characterize the metabolic response to glutamate in the post-hypoxic phase, (b) assess whether glutamate-induced protection could be related to an improvement of oxidative metabolism, (c) explore the role of NCX1 in mitochondrial responses to glutamate. We first probed the metabolic phenotype of H9c2-WT and H9c2-NCX1 cells by analyzing mitochondrial respiration, assessed as oxygen consumption rate (OCR) and glycolytic activity, measured as extracellular acidification rate (ECAR) Figure 1. Timeline of the experimental protocols (H/R). Schematic diagram showing the H/R timeline protocol in H9c2 cells (a) and in isolated rat adult cardiomyocytes (b). Control groups were incubated under normoxic conditions at 37 °C for the entire protocol. Glutamate (1 mM)-alone or in combination with 1 µM SN-6, 300 µM DL-TBOA or 3 μg/ml oligomycin (for ATP experiments conducted in H9c2-NCX1 cells)-was administered during the reoxygenation phase. Cell viability (assessed by extracellular LDH measurement) and ROS were evaluated at the end of the reoxygenation phase in both experimental protocols. ATP content, OCR and ECAR were assessed after the first hour of reoxygenation in H9c2-NCX1 cells. CTL = control; H/R = hypoxia/ reoxygenation; G = glutamate.
of the surrounding media (which predominately reflects the excretion of lactic acid converted from pyruvate). Experimental operative protocols are summarized in Fig. 5 (see "Methods" for further details). In normoxic conditions, ECAR at baseline was not different between H9c2-WT and H9c2-NCX1 cells, whereas maximal respiratory capacity was slightly but significantly smaller in H9c2-NCX1 cells (Fig. 5a). When H/R stress was applied, both cell lines showed a marked decrease in mitochondrial oxygen consumption that was better compensated in H9c2-WT by an increase in glycolysis (Fig. 5b). Next, we measured OCR and ECAR parameters in hypoxic H9c2-NCX1 cells reoxygenated with glutamate. As shown in Fig. 6a, after glutamate treatment, OCR profiles Effect of NCX inhibition on glutamate-induced protection against H/R injury. Extracellular LDH activity measured 5 h after the hypoxic insult (3 h) in H9c2 cells (a,b) and 2 h after the hypoxic insult (1.5 h) in rat adult cardiomyocytes (d) in different experimental conditions. 1 mM glutamate, alone or in combination with 1 μM SN-6, was added during the reoxygenation phase. Differences among means were assessed by one-way ANOVA followed by Dunnet's post hoc test. Each column represents the mean ± S.E.M. of almost 5 independent experiments performed in duplicate. (a) *p < 0.001 versus CTL, CTL + G and H/R + G; # p < 0.001 versus CTL, p < 0.01 versus CTL + G and p < 0.05 versus H/R + G; § p < 0.001 versus CTL and CTL + G, p < 0.01 versus H/R + G. (b) *p < 0.01 versus CTL and CTL + G; # p < 0.001 versus CTL and p < 0.01 versus CTL + G. (d) LDH levels were normalized to the control (normoxia-exposed) group and expressed as percentage. Each column represents the mean ± S.E.M. of almost 4 independent experiments performed in duplicate. *p < 0.001 versus CTL and CTL + G and p < 0.05 versus H/R + G; # p < 0.001 versus CTL + G, p < 0.01 versus CTL and p < 0.05 versus H/R + G; § p < 0.001 versus CTL and CTL + G, and p < 0.05 versus H/R and H/R + G. (c) Analysis of H9c2 cell survival by FDA/PI staining. Images are representative of 3 independent experiments. CTL = control; H/R = hypoxia/reoxygenation; G = glutamate. greatly improved toward normoxic values. In particular, in H9c2-NCX1 cells reoxygenated in the presence of glutamate we observed a significant recovery of maximal respiratory capacity, which indicates increased activity of the electron transport chain (ETC), as well as a significant improvement in spare respiratory capacity, which estimates cell's ability to cope with large increases in energy demand and reflects the amount of extra ATP that can be produced by oxidative phosphorylation. Notably, the recovery of OCR profiles induced by glutamate was significantly inhibited when NCX1 was blocked with SN-6 during the reoxygenation phase (Fig. 6a). Glutamate ability to promote mitochondrial ATP generation (and thereby survival) in H/R H9c2-NCX1 cells was also supported by the capacity of oligomycin 18 (an inhibitor of the ATP synthase, the final enzyme in the oxidative phosphorylation pathway) to fully prevent the ATP response to glutamate during reoxygenation phase ( Supplementary Fig. 2). Finally, as shown in Fig. 6b, glycolytic activity was significantly reduced by glutamate during H/R in H9c2-NCX1 cells, and also this metabolic response was sensitive to NCX1 blockade. Collectively, these data lend further support to the hypothesis that NCX1 is critical for the glutamate-dependent metabolic boost in myocytes recovering from hypoxic insult, and that the ATP production stimulated by glutamate during the reoxygenation phase essentially relies on mitochondrial oxidative phosphorylation.

Analysis of NCX 1 and EAATs expression following H/R challenge.
In our previous studies, we demonstrated that both in the heart and in H9c2-NCX1 cells NCX1 protein expression is upregulated under stressful conditions, including hypertrophy and ischemic injury 19,20 . Since we observed that the protective effect of glutamate against H/R damage relied both on NCX1 and EAAT activities (Figs 2 and 3), we explored whether EAATs expression could also be modified by H/R challenge. As shown in Fig. 7, protein expression analysis revealed that, in H9c2-NCX1 cells, NCX1 levels were increased after H/R (Fig. 7a), in line with our previous report 19 . Again, in H9c2-WT cells NCX1 protein expression was undetectable both in normoxia and after H/R challenge (Fig. 7a). The expression of the three main EAATs expressed in H9c2 cells 17 , namely EAAC1, GLAST and GLT-1 was unmodified after H/R challenge, in both H9c2-WT and H9c2-NCX1 cells (Fig. 7b-c).

Effect of glutamate on NCX1 activity alteration following H/R. Cardiac ischemia typically increases
expression of NCX1 19 , and the accompanying alterations of exchange activity during I/R injury fuel a vicious circle of further damage by promoting intracellular Ca 2+ overload 28,29 . We therefore verified that NCX1 function was altered in our H/R model and explored the possibility that the normalization of its activity accompanies the protection induced by glutamate. Exchanger activity was monitored in Fluo-4 loaded H9c2-NCX1 cells subjected to isotonic extracellular Na + removal at the end of the experimental protocol reported in Fig. 1 (see LDH levels were normalized to the control (normoxia-exposed) group and expressed as percentage. Each column represents the mean ± S.E.M. of 4 independent experiments performed in duplicate. *p < 0.001 versus control groups and p < 0.05 versus H/R + G; # p < 0.001 versus control groups and H/R + G; § p < 0.01 versus CTL and p < 0.05 versus H/R + G. CTL = control; H/R = hypoxia/reoxygenation; G = glutamate; TBOA = DL-TBOA.
"Methods" for further details). No change in fluorescence baseline was observed in H9c2-WT cells subjected to the same Na + removal protocol (data not shown) 17 , confirming that the Ca 2+ responses observed in H9c2-NCX1 are mediated by NCX1 reverse mode. As shown in Fig. 8, when NCX1 reverse mode was activated by superfusing a Na + -free extracellular solution, in control cells a rise in intracellular Ca 2+ concentration ([Ca 2+ ] i ) of about 80% occurred, as revealed by the increase in fluorescence signal. When H/R group was analyzed, we observed that the NCX1-mediated increase in [Ca 2+ ] i was enhanced (about 50%) compared to what observed in control cells. When glutamate was added during the entire reoxygenation phase, NCX1 reverse activity was normalized to normoxic values and was significantly reduced compared to H/R group. Glutamate failed to prevent the H/R-induced increase of NCX1 reverse activity in the presence of both DL-TBOA and SN-6. As shown in Fig. 8b,d, the inhibitors do not affect per se Ca 2+ responses, neither in normoxia nor in H/R conditions.

Discussion
In this report we provide evidence that glutamate supplementation from the start of the reoxygenation phase counteracted the H/R-induced injury in cardiac cells. In particular, we demonstrated that such glutamate protective action was related to its ability to sustain oxidative metabolism, leading to an increase in ATP cellular content. Noteworthy, we showed that this protection disappeared in the absence of a functional NCX1, disclosing a key role of this transporter in sustaining cell viability.
NCX1 is central to many pathophysiological functions of the heart 19,20,30,31 , and in particular its role in cardiac ischemia has been investigated in different in vitro and in vivo models 30 . On one hand, a detrimental role of NCX1 during myocardial I/R emerges when the unbalanced exchange activity contributes to myocyte electrical instability and promotes Ca 2+ overload 28,29 , so those strategies that ultimately normalize NCX1-mediated Na + and Ca 2+ ionic fluxes have therapeutic potential 32 . We have previously found in different cardiac models that pharmacological blockade of NCX1throughout the entire H/R protocol is protective 19 . On the other hand, NCX1 activity is strategic for cardioprotection against I/R evoked by conditioning programs 19,33 . Intriguingly, we disclosed here a new beneficial and essential role of NCX1 in glutamate-induced cell survival in cardiac models of H/R. This conclusion is supported by the following evidence: (1) in H9c2 cells, glutamate productively sustained mitochondrial ATP synthesis thereby promoting survival in H9c2-NCX1 (expressing NCX1) but not in H9c2-WT (not expressing NCX1) 17 (Figs 2 and 4); (2) in H9c2-NCX1 cells, pharmacological blockade of NCX1 during the reoxygenation phase completely prevented the beneficial effects of glutamate in terms of recovery of ATP synthesis (Fig. 4) and mitochondrial respiration (Fig. 6), cell protection (Fig. 2) and normalization of Na + /Ca 2+ exchanger activity (Fig. 8); (3) in rat adult cardiomyocytes, glutamate failed to protect cells against H/R injury when NCX1 was inhibited at the reoxygenation (Fig. 2). Interestingly, we found that SN-6 applied at 1 μM only during the reoxygenation phase failed to protect H9c2-NCX1 cells or rat adult cardiomyocytes against H/R injury, in line with previous data obtained in different cell lines expressing NCX1 34 . In the heart, glutamate is one of the main constituent of the free intracellular amino acid pool 7 . Beyond its role as building block during anabolic macromolecular synthesis, glutamate acts as a key metabolite of myocardial energy metabolism. Indeed, its activity in coupling cytosolic and intra-mitochondrial energetic states through the malate-aspartate shuttle is well known 9 . Some experimental interventions aimed to improve cardiac metabolism against I/R challenge are based on the attempt to favor glutamate or glutamine utilization, through the classical ischemic preconditioning, the increase in glutamate transporter activity or the perfusion of glutamine supplement [35][36][37] . The administration of high doses of glutamate during post-ischemic reperfusion has been shown to improve left ventricular function 38 , and, in line with this observation, we found that glutamate supplementation from the start of the reoxygenation phase dramatically improved cell viability in our cardiac models of H/R. Specifically, protection was observed in H9c2-NCX1 cells, but not in H9c2-WT exposed to glutamate. Moreover, in the H9c2-NCX1 cells, the ability of glutamate to ameliorate viability was abolished by the NCX1 inhibitor SN-6. Taken together, these findings provide the first clear evidence for the involvement of NCX1 activity in promoting glutamate-induced cell survival.
Since an important step in the recovery of cardiac cells upon reperfusion is the resumption of the oxidative phosphorylation, we explored the effect of glutamate supplementation on ATP production and the involvement of NCX1 in such metabolic response. As expected, in H9c2-NCX1 cells subjected to H/R challenge, ATP levels were dramatically reduced after already 1 h of reoxygenation. Interestingly, this drop in ATP content was fully counteracted by glutamate administration in the first hour of reoxygenation. In connection with these results, a previous work by Kristiansen and coworkers showed that, in a rat isolated perfused heart model, the administration of exogenous glutamate from the beginning of the reperfusion reduces infarct size to the same extent as its administration during both ischemia and reperfusion, indicating that glutamate main effect is linked to the latter phase 5 . Additionally, we found that the NCX1 inhibitor SN-6 abolished the ability of glutamate to ameliorate ATP production. Our data demonstrated that NCX1 may play a critical role in the glutamate-induced ATP synthesis under both pathological and physiological conditions 17,18 .
It is known that glutamate can get access to the mitochondrial matrix via the aspartate/glutamate carriers, a required component of the malate/aspartate shuttle 39,40 . We have recently proposed an alternative and innovative pathway, whereby EAAC1 -a member of the EAATs family 25,41,42 -and NCX1 cooperate in order to favor glutamate entry into the cytoplasm and then into the mitochondria, stimulating ATP synthesis 17,18 . Therefore, once confirmed the requirement of a functional NCX1 for glutamate to improve cell survival by enhancing the ATP response under hypoxic conditions, an involvement of EAATs was also tested. The ability of glutamate to stimulate ATP recovery and restore cell viability was fully abolished in the presence of the non-transportable EAATs inhibitor DL-TBOA (300 µM), confirming that glutamate entrance into the cells was mediated by EAATs.
We have previously demonstrated that when H9c2-NCX1 (but not H9c2-WT) are acutely exposed to glutamate an increase of the reverse mode of NCX1 (i.e. Ca 2+ influx/Na + efflux exchange cycle) is observed, and this increase is selectively inhibited by NCX or EAATs blockers 17 . Such increase of the reverse mode of NCX1 develops within seconds and relies on the presence of extracellular glutamate that, being cotransported with Na + into the cell via EAATs, influences both Na + gradient across plasma membrane and membrane potential 17 . Overall, results from our and other groups 43,44 lend support to the existence of a functional coupling between EAAT and NCX transporters in different cell types, whereby the EAAT-induced NCX reverse mode maintains the Na + -driving force for an effective glutamate uptake. After being picked up by cells, glutamate can be used as metabolic fuel for mitochondrial ATP synthesis, and in this process, Ca 2+ signals originated by the reversed NCX activity can also play a role 17 .
The glutamate-induced recovery in ATP-linked respiration and other OCR parameters indicates an improved activity of ETC, which otherwise would be unable to support the increased ATP demand (and thereby viability) during reoxygenation phase. Overall, our data indicate that H/R H9c2-NCX1 cells treated with glutamate are better equipped to function in conditions of increased energy needs. The improvement of oxygen consumption by glutamate in H/R was significantly suppressed when NCX1 was blocked with SN-6, further highlighting the key role of this exchanger in glutamate-dependent protection of ischemic myocytes.
The mechanisms underlying the glutamate-induced decrease of ECAR in H/R cells were no further investigated in the present study. Albeit speculative, several mechanisms may come into play. It has been shown that in ischemic cardiomyocytes glutamate decreases lactate levels by shunting pyruvate to alanine [45][46][47][48] , and promotes diversion of glucose into glycogen rather than undergoing glycolytic oxidation 38 . In any case, the decrease of glycolytic rate induced by glutamate in H/R H9c2-NCX1 cells may also contribute to the observed protection. This is because glycolysis produces significant level of acidic by-products (i.e. lactate), and proton accumulation accounts for a substantial proportion of dysfunctions of myocytes in ischemic settings 1 .
During reoxygenation the molecular oxygen reintroduced to hypoxic myocytes can be converted to oxygen free radicals above viable levels 49 , and glutamate can improve the reducing power of myocytes 4 . Therefore, it is possible that the restoration of energetic metabolism and the preservation of free radical scavengers 6 acted as synergistic components of glutamate-induced protection. Indeed, we found that H/R injury increased reactive oxygen species (ROS) in H9c2-NCX1 cells and that glutamate partially, but significantly, reduced such increase ( Supplementary Fig. 3).
Our use of glutamate at 1 mM, which is somehow higher than basal plasma levels (between 0.05-0.2 mM, depending on mammalian species 6,38 ), was based on the following considerations. First, at the concentration tested in this study, glutamate has no evident toxic effect on H9c2 cells and rat adult cardiomyocytes. Second, protection against ischemic injury typically requires glutamate supplementation in the millimolar range 5,6,38,50 . Third, although glutamate has very large muscle/plasma ratios at the baseline, during myocardial ischemia such large concentration gradient dissipates 4 and high exogenous concentrations of glutamate are required to compensate this loss and enable adequate intracellular glutamate loading 6 . Fourth, ATP synthesis in H9c2-NCX1 cells is significantly stimulated when glutamate is used at 1 mM concentration 17 (Fig. 4a).
Considering that ionic disturbances occurring during H/R are leading causes of cell death 2 , we investigate whether the ability of NCX1 to control intracellular Ca 2+ levels might be involved in the glutamate-induced cardioprotection. Experiments performed by using the fluorescent Ca 2+ indicator Fluo-4 showed an increased NCX1 reverse-mode activity in cells that were subjected to H/R, rather than in controls. It is interesting to note that the increase in NCX1 activity tended toward normalization when cells were reoxygenated in the presence of 1 mM glutamate. We hypothesize that ATP produced from glutamate by H/R cells can support Na + /K + -pump and Ca 2+ -ATPase to restore intracellular Na + and Ca 2+ levels, so that the NCX-driven Ca 2+ overload is limited and cell survival promoted. An alternative (but not exclusive) explanation could rely on the finding that increased ATP levels, in the presence of Ca 2+ transients, may evoke a rapid massive endocytosis, which can involve NCX1 51 . It is possible to speculate that glutamate-enhanced ATP cellular content may serve as a trigger for rapid massive endocytosis, which in turn may remove NCX1 from the cell surface, thereby limiting its activity. Since both the hypothesis depends on the glutamate entry into the cells, the involvement of the EAATs in this metabolic pathway was also confirmed by testing NCX1 activity in the presence of DL-TBOA. As expected, we found that glutamate failed to attenuate the Ca 2+ increase evoked by H/R insult when cells were incubated with the inhibitor DL-TBOA.
Collectively, our data provide clear evidence that glutamate supplementation from the beginning of the reoxygenation phase can positively affect cell viability by sustaining the oxidative metabolism and increasing ATP content, with NCX1 and EAATs playing a critical role. In particular, as for normoxic conditions, we propose an alternative and regulated mechanism whereby EAATs activity would stimulate NCX1 reverse mode of operation, leading to an increase in mitochondrial Ca 2+ concentration, to a higher physiological steady-state level likely stimulating Ca 2+ -sensitive dehydrogenase activity and the rate of ATP synthesis. Indeed, Ca 2+ may play a dual role within the cells: on one hand this ion can be essential to stimulate ATP synthesis, on the other hand it can be harmful, by triggering cell death pathways 52 . There must be a critical point representing the boundary between cytoprotective and cytotoxic effects related to the increase in [Ca 2+ ] i , and our results demonstrate that this point might also critically depend upon NCX1 activity.

Isolation of rat adult ventricular cardiomyocytes.
In vitro hypoxia/reoxygenation challenge. The day before the H/R experiment, cells were plated in 6 multiwell plates (120,000 cells/well for H9C2 cells or 10,000 cells/cm 2 for cardiomyocytes). Hypoxia was induced in an airtight chamber in which O 2 was replaced with N 2 in a glucose-free Tyrode's solution containing (in mM): NaCl 137, KCl 2.7, MgCl 2 1, CaCl 2 1.8, NaH 2 PO 4 0.2 and NaHCO 3 10, pH 7.4. After closing all sealable connectors, the chamber was transferred to an incubator and the cells were subjected to hypoxia (as described in Fig. 1) at 37 °C. Reoxygenation was initiated by opening the chamber and then replacing the glucose-free Tyrode's solution with fresh Tyrode's solution containing 5.5 mM glucose 53 . The cells were then maintained in the incubator under an atmosphere of 5% CO 2 , at 37 °C, as described in Fig. 1. Glutamate supplementation does not significantly modifies medium osmolarity.
Evaluation of cell viability. H/R-induced cell injury was quantified by measurement of lactate dehydrogenase (LDH) activity released from the cytosol of damaged cells in the experimental media 53 , and by the method of double staining with FDA/PI 21 . At the end of the H/R experiment, 100 μl of cell culture medium were removed and added to a 96 well plate. Then, 100 μl of the reaction mixture (Diaphorase/NAD + mixture premixed with iodotetrazolium chloride/sodium lactate) were added to each well and the plate was incubated for 30 min at room temperature, protected from light. LDH activity was assessed by reading the absorbance of the sample medium at 490 nm in a Victor Multilabe Counter plate reader (Perkin Elmer, Waltham, MA, USA). For FDA/PI staining, cells were plated on glass coverslips and subjected to H/R. Afterwards, cells were treated with 36 μM FDA (Sigma) and 7 µM PI (Calbiochem., San Diego, CA, U.S.A.) for 10 min at 37 °C in PBS. Stained cells were examined immediately with an inverted Zeiss Axiovert 200 microscope (Carl Zeiss, Milan, Italy) and then analyzed. When FDA crosses the cell membrane it is hydrolyzed by intracellular esterases producing a green-yellow fluorescence. Cell damage curtails FDA staining and allows cell permeation by PI that, interacting with nuclear DNA, yields a bright red fluorescence 21 .
Analysis of ATP production. ATP synthesis was evaluated using a commercially available luciferase-luciferin system (ATPlite, Perkin Elmer, Waltham, MA). The day before the experiment, cells were plated (5,000 cells/well) in 96 multiwell plates. The day after, cells were first washed with Tyrode's solution containing (in mM): NaCl 137, KCl 2.7, MgCl 2 1, CaCl 2 1.8, NaH 2 PO 4 0.2, NaHCO 3 10, glucose 5.5 mM, pH 7.4 and then exposed to different glutamate concentrations (0.5 and 1 mM) in the same Tyrode's solution for 1 h at 37 °C 17 . When ATP content was evaluated after H/R, glutamate and the specific pharmacological tools were added at the beginning of the reoxygenation phase and maintained for 1 h. After the incubation period, ATP levels were analyzed with a luminescence counter (Victor Multilabel Counter, Perkin Elmer) and normalized to the respective protein content 17,18 . Bioenergetic analysis. Seahorse XF24 Extracellular Flux Analyzer (Seahorse Bioscience, North Billerica, MA, USA) was used to detect oxygen consumption rate (OCR) and extracellular acidification rate (ECAR), representing oxidative phosphorylation and glycolysis, respectively, as previously described 54,55 . The general scheme of the mitochondrial stress test is shown in Fig. 5a. Oligomycin (1.5 μM), FCCP (2 μM), rotenone/antimycin A (0.5 μM) were sequentially introduced to measure basal respiration, ATP production, proton leak, maximal respiration, spare respiratory capacity, and non-mitochondrial respiration. Maximal respiratory capacity was estimated by inducing maximal OCR via chemical dissipation of the mitochondrial membrane potential with the protonophore FCCP on the background of oligomycin (used to prevent the ATP-consuming reverse activity of ATP synthase, which may lead to cellular metabolic dysfunction and death). Maximal respiratory capacity is a measure of the maximal ability of the ETC to produce energy. Spare respiratory capacity is derived from the difference between maximal OCR and basal respiration. A cell with a larger spare respiratory capacity can produce more ATP to maintain adequate levels of energetic molecules and overcome more stress.
The general scheme of glycolysis stress test is shown in Fig. 5b. Sequential injections of 3 μM rotenone (to block complex I, thereby eliminating mitochondrial respiration and force cells to rely on glycolysis), 10 mM glucose, and 100 mM 2-deoxyglucose (2-DG; glucose analog and inhibitor of glycolytic ATP production) were used to measure glycolysis, glycolytic capacity and allow estimation of glycolytic reserve and non-glycolytic acidification.
H9c2 cells (40,000 cells/well) were seeded on the XFp cell culture mini plates (Seahorse Bioscience, Billerica MA, USA) and subjected to the H/R challenge (Fig. 1). At the end of the first hour of reoxygenation, the Tyrode's solution was replaced with 500 μl/well of XF24 running media. The plates were pre-incubated at 37 °C for 20 min in the XF Prep Station incubator (Seahorse Bioscience, Billerica MA, USA) in the absence of CO 2 and then run on the XF24 analyzer to obtain OCR and ECAR.
OCR and ECAR were recorded during specified programmed time periods (three readings each) as the average numbers between the injections of inhibitors mentioned above. The final data calculation was performed after the readings had been normalized for total protein/well. Western blotting. Protein extraction and western blotting analysis were performed as previously described 19 . Immunoblots were probed overnight at 4 °C with the appropriate primary antibody: NCX1 17,19 (R3F1, Swant, Bellinzona, Switzerland), dilution 1:500; mouse anti-EAAC1 17 (Chemicon International, CA, USA), dilution 1:1,000; rabbit anti-GLAST and rabbit anti-GLT1 17 (Alpha Diagnostic International) both used at