Distinct effects of intracellular vs. extracellular acidic pH on the cardiac metabolome during ischemia and reperfusion

Abstract

Tissue ischemia results in intracellular pH (pH_IN) acidification, and while metabolism is a known driver of acidic pH_IN, less is known about how acidic pH_IN regulates metabolism. Furthermore, acidic extracellular (pH_EX) during early reperfusion confers cardioprotection, but how this impacts metabolism is unclear. Herein we employed LCMS based targeted metabolomics to analyze perfused mouse hearts exposed to: (i) control perfusion, (ii) hypoxia, (iii) ischemia, (iv) enforced acidic pH_IN, (v) control reperfusion, and (vi) acidic pH_EX (6.8) reperfusion. Surprisingly little overlap was seen between metabolic changes induced by hypoxia, ischemia, and acidic pH_IN. Acidic pH_IN elevated metabolites in the top half of glycolysis, and enhanced glutathione redox state. Meanwhile, acidic pH_EX reperfusion induced substantial metabolic changes in addition to those seen in control reperfusion. This included elevated metabolites in the top half of glycolysis, prevention of purine nucleotide loss, and an enhancement in glutathione redox state. These data led to hypotheses regarding potential roles for methylglyoxal inhibiting the mitochondrial permeability transition pore, and for acidic inhibition of ecto-5′-nucleotidase, as potential mediators of cardioprotection by acidic pH_EX reperfusion. However, neither hypothesis was supported by subsequent experiments. In contrast, analysis of cardiac effluents revealed complex effects of pH_EX on metabolite transport, suggesting that mildly acidic pH_EX may enhance succinate release during reperfusion. Overall, each intervention had distinct and overlapping metabolic effects, suggesting acidic pH is an independent metabolic regulator regardless which side of the cell membrane it is imposed.

Graphical Abstract

1. INTRODUCTION

pH is arguably one of the most fundamental environmental conditions that govern cell fate (along with factors such as temperature, oxygen and substrate availability). Along with enzymes such as carbonic anhydrase [1], cells express a variety of transporters and exchangers at the plasma membrane to regulate intracellular pH (pH_IN). This includes Na⁺/H⁺ exchangers (NHEs) [2], Na⁺/HCO₃⁻ cotransporters (NBCs) [3, 4], and monocarboxylate transporters (MCTs) [5, 6] (Figure 1A, see also [7] for review).

Figure 1. Cardiac pH_IN Manipulation.

(A): Endogenous regulation of cardiomyocyte intracellular pH (pH_IN). Glycolysis generates pyruvate, which can either be shuttled into mitochondria for TCA cycle activity (generating CO₂) or remain in the cytosol to generate lactate and a proton (H⁺). Cardiomyocytes regulate intracellular pH via monocarboxylate transporters (MCTs), Na⁺/H⁺ exchangers (NHEs), and Na⁺/HCO³⁻ cotransporters (NBCs). Pharmacologic inhibitors for these exchangers and transporters used herein are shown in red: AR-C155858 (MCT inhibitor), cariporide (NHE1 inhibitor), and S0859 (NBC inhibitor). (B): Representative images of isolated cardiomyocytes subject to control and enforced acid pH_IN (cariporide + S0859 + pH_EX 7.1). (C): Quantitation of pH_IN. Data are means ± SEM, N=10 (control) and 4 (acid pH_IN). Statistical significance was analyzed by one-way ANOVA with post-hoc Tukey’s HSD test. Other conditions tested are shown in Figure S2. (D): pH_IN measured in intact hearts using BCECF reflectance fluorescence [34]. Hearts were loaded with BCECF followed by washout then imposition of acid pH_IN conditions as indicated by the arrows. Gray traces are individual hearts, with average shown in yellow. (E): Graph showing heart rate x left ventricular developed pressure (i.e., rate pressure product, RPP) for control and enforced acid pH_IN perfused mouse hearts. Data are means ± SEM, N=8 hearts. Data for individual components (heart rate and developed pressure) are in Figure S3. Cardiac functional data for other conditions (hypoxia and ischemia) are in Figure S3.

Despite classical knowledge that nearly all metabolic enzymes are pH sensitive, only recently has acidic pH been recognized as a metabolic remodeling signal. Specific examples include an acid-induced neomorphic enzyme activity in lactate and malate dehydrogenases, to generate the metabolic side-product L-2-hydroxyglutarate [8, 9], and an acid-induced reversal of isocitrate dehydrogenase to reductively carboxylate α-ketoglutarate to citrate [8, 10]. Both these metabolic events are thought to be characteristic of and advantageous to cancer cells [11]. The expulsion of H⁺ from cells also serves important signaling roles such as maintaining the tumor microenvironment [12], or in the case of stem cells maintaining pluripotency/stemness [13]. Ligand binding to cell surface receptors [14] and metabolite dynamics (i.e. import and export) [15] are also known to be pH sensitive.

Cardiac ischemia (i.e., the cessation of coronary blood flow) or hypoxia (i.e., oxygen limitation with sustained substrate delivery and blood flow) are perhaps the most significant forms of metabolic disruption experienced by the heart, and are accompanied by large declines in cellular pH. Although the role of pH in determining outcomes in cardiac ischemia-reperfusion (IR) injury has been extensively studied, the acute effects of pH on cardiac metabolism are not completely understood, and have not been studied using agnostic tools such as metabolomics.

The primary metabolic sources of protons in normoxic cells are ATP hydrolysis and CO₂ generated from the TCA cycle [16, 17]. During cardiac ischemia or hypoxia, cessation of mitochondrial oxidative phosphorylation (Ox-Phos) leads to enhanced reliance on anaerobic glycolysis as a source of ATP [18]. While glycolysis results in both the generation of lactate and a decline of pH_IN, the term “lactic acidosis” often incorrectly implies that protons originate from lactate. In fact, lactate is ionized at physiologic pH (pK_a 3.9) and the source of protons in glycolysis is ATP hydrolysis at hexokinase and phosphofructokinase [19, 20]. In ischemia or hypoxia, the absence of buffering of these protons by Ox-Phos precipitates acidosis [21].

During cardiac ischemia/hypoxia, acidic pH is thought to confer protection against cell death via a number of mechanisms including lowering ATP demand by inhibition of contractile machinery [22, 23], maintaining the mitochondrial permeability transition (PT) pore in a closed state [24, 25], and the inhibition of proteases and phospholipases [26, 27]. Although both ischemia and hypoxia have several overlapping features including the acidification of pH_IN, the degree to which acid pH_IN can act as an independent factor that acutely remodels metabolism is unclear. In the first part of this study, we applied targeted metabolomics to compare the effects of hypoxia, ischemia or forced acid pH_IN on cardiac metabolism. Our goal was to ask what portion of the acute metabolic changes seen in hypoxia or ischemia are actually due to acidosis?

Following a cardiac ischemic event, paradoxically much of the subsequent pathology is triggered by tissue reperfusion (re-establishment of coronary blood flow), leading to ischemia-reperfusion (IR) injury. Upon reperfusion, mitochondrial Ca²⁺ overload and a burst of reactive oxygen species (ROS) generation combine to trigger opening of the mitochondrial PT pore, leading to necrosis [28–32]. This process is exacerbated by the re-alkalinization of myocardial pH, relieving inhibition of the pore [24, 25]. The major source of ROS at reperfusion is the rapid mitochondrial oxidation of the TCA cycle metabolite succinate, which accumulates during ischemia [33–35]. However, around ⅔ of accumulated succinate washes out from the heart upon reperfusion, in a pH-sensitive manner. An acidic extracellular pH (pH_EX 6.0) can promote retention of succinate inside cells [15], which suggests that the local pH environment in the heart at reperfusion (i.e., extracellular as well as intracellular pH) may influence succinate-driven ROS and IR injury [36].

Further highlighting the importance of pH during early reperfusion, it has been widely demonstrated that extending ischemic acidosis into reperfusion via several independent methods confers cardioprotection. This includes: (i) perfusing the heart during early reperfusion with mildly acidic extracellular media (pH 6.4–6.9) [37–50] (see Table S1), (ii) pharmacologic inhibition of carbonic anhydrase [49, 51], (iii) acute stimulation of glycolytic lactate production via bolus delivery of the NAD⁺ precursor NMN [52], and (iv) ischemic post-conditioning (aka, staccato reperfusion) [53]. The method of acidification used to confer protection does not appear important, with similar results reported for the use of pH buffers such as HEPES or hypercapnia to achieve an altered carbonic acid equilibrium (Table S1).

In addition, although the pharmacologic inhibition of NHEs confers cardioprotection [54, 55], this effect has been mostly attributed to preventing myocardial Na⁺ overload, which subsequently is thought to prevent Ca²⁺ overload from reverse mode operation of the Na⁺/Ca²⁺ exchanger (see [56] for review). Notably, the potential therapeutic effects of NHE inhibition on cardiac pH_IN within the setting of IR injury have been largely overlooked, and it is unknown whether extension of acidosis into the reperfusion period accounts for any of the cardioprotective benefits of NHE inhibiting drugs.

As mentioned above, acidic pH_IN is a known inhibitor of mitochondrial PT pore opening [24, 25]. However, the potential downstream mediators of protection by acidic pH_EX reperfusion are not fully understood. Furthermore, an apparent paradox exists, in which acidic pH_EX at reperfusion is cardioprotective (Table S1), but it should also drive succinate retention in cells [15], leading to more ROS generation and damage [34]. With a goal of understanding the acute effects of acid pH_EX reperfusion on metabolism, the second part of this study applied targeted metabolomics analysis to examine cardiac metabolism following 2 minutes of reperfusion with normal (pH 7.4) or acidic (pH 6.8) buffer. Downstream metabolic mechanisms were also explored as potential mediators of acidic pH_EX induced cardioprotection. Overall, in addition to providing reference data sets for the effects of pH on cardiac metabolism, our results indicate that acidic pH_IN and pH_EX are powerful metabolic regulators with both overlapping and independent effects.

2. MATERIALS AND METHODS

2.1. Animals and Reagents

Animal experiments complied with the NIH Guide for Care and Use of Laboratory Animals (8^th edition, 2011) and were approved by the University of Rochester Committee on Animal Resources. Male and female C57BL/6J mice (8–16 weeks old) were housed on a 12 hr. light/dark cycle with food and water ad libitum. Terminal anesthesia was achieved via intra-peritoneal 2,2,2-tribromoethanol (Avertin) at 250 mg/kg. BCECF-AM was from Molecular Probes (Eugene, OR, USA) and collagenase was from Roche (Indianapolis, IN, USA). All other reagents were from Sigma (St. Louis, MO, USA) or MedChemExpress (Monmouth Jct., NJ, USA).

2.2. Adult Mouse Cardiomyocyte Isolation

Following anesthesia, the aorta was rapidly cannulated, then the heart excised and perfused for 3 min. with Isolation Buffer (IB), comprising (in mM): NaCl (120), KCl (15), Na₂HPO₄ (0.6), KH₂PO₄ (0.6), MgSO₄ (1.2), HEPES (10), NaHCO₃ (4.6), taurine (30), glucose (5.5), butanedione monoxime (10), pH 7.4 at 37 °C. The heart was then perfused for 10 min. with Digestion Buffer (DB), comprising IB plus 12.5 μM CaCl₂, 0.025 % (wt/vol) trypsin, 6.525 U collagenase A, and 15.375 U collagenase D. Ventricular tissue was teased apart, resuspended in Stop Buffer (SB), consisting of IB plus 12.5 μM CaCl₂ and 10 % (vol/vol) FBS, and filtered through 200 μm mesh. Cells were settled by gravity for 10 min., then sequentially suspended/settled in SB with [Ca²⁺] increased stepwise to 1.8 mM. Finally, cells were suspended in DMEM with (in mM) L-glutamine (4), Na-pyruvate (0.1), glucose (5), L-carnitine (0.5), palmitate conjugated to BSA (0.1), pH 7.4 at 37 °C. This protocol [57] yielded ~8 × 10⁵ cells per heart, with ~80% rod-shaped cells excluding Trypan blue dye.

2.3. Measurement of cardiomyocyte pH_IN

Cardiomyocytes were seeded on laminin-coated 35 mm coverslips (20 μg/ml) for 1 hr., incubated with the pH-sensitive dye 2’−7’-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein, acetoxymethyl ester (BCECF-AM) at 3.84 μM for 30 min., then subject to various combinations of the following treatments for 1 hr.: (i) Control pH 7.4, (ii) pH_EX 7.1, (iii) 100 nM of the protonophore carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP), (iv) 30 μM of the NBC inhibitor S0859, (v) 20 μM of the NHE1 inhibitor cariporide, (vi) 1 mM of the NAD⁺ precursor nicotinamide mononucleotide (NMN), (vii) 1 μM of the MCT-1 inhibitor AR-C155858. Ratiometric fluorescence was measured at λ_EX /λ_EM 440/535 and 440/490 nm using an Eclipse TE2000-S microscope (Nikon, Avon MA) and data were analyzed using a TILL Photonics System as previously described [8]. pH_IN was calculated from a determined Boltzmann factor calibration curve using the ionophore nigericin based on fluorescence ratios. For each coverslip, four images were taken, capturing 5–16 individual cardiomyocytes (technical replicates) to determine pH_IN values. Groups of 3–4 coverslips on a given day were considered as biological replicates (N).

2.4. Langendorff-perfused mouse hearts

Following anesthesia, the aorta was rapidly cannulated, the heart excised and perfused at 4 ml/min. with Krebs-Henseleit buffer (KHB) comprising (in mM): NaCl (118), KCl (4.7), MgSO₄ (1.2), NaHCO₃ (25), KH₂PO₄ (1.2), CaCl₂ (2.5), glucose (5), pyruvate (0.2), lactate (1.2), palmitate conjugated 6:1 with fat-free BSA (0.1). KHB was gassed with 95 % O₂ and 5 % CO₂ at 37 °C to maintain pH 7.4. A water-filled balloon connected to a pressure transducer was inserted into the left ventricle and expanded to provide a diastolic pressure of 6–8 mmHg. Cardiac function was recorded digitally at 1 kHz (Dataq, Akron OH).

To determine the effect of hypoxia, ischemia, or acidic pH_IN on cardiac metabolism, hearts were equilibrated for 15 min., and then further subjected to 20 min. of either: (i) control perfusion, (ii) hypoxia (buffer gassed with 95% N₂, 5% CO₂), (iii) global no-flow ischemia, or (iv) acidic pH_IN (Figure 2A). Optimal acidic pH_IN conditions (20 μM cariporide + 10 μM S0859, at pH_EX 7.1) were determined from cardiomyocyte experiments, see Figure S2. A pH_EX of 7.1 was obtained using modified KHB containing 10 mM NaHCO₃, gassed with 95 % O₂, 5 % CO₂.

Figure 2. Metabolomics study in control, hypoxia, ischemia and acidosis.

(A): Schematic showing design of the 4 experimental perfusion conditions. Hearts were snap frozen for steady-state metabolomics analysis by LC-MS indicated by the arrow. Colors and terms used to denote each condition are used throughout. Each condition used of 8 hearts. (B): Sparse partial least-squares discriminant analysis (sPLSDA) plot, prepared using MetaboAnalyst 5.0, incorporating 59 metabolites from 32 samples. Key at upper right denotes symbols for each group. The top 5 metabolite weightings contributing to each principal component (Component 1: x-axis, Component 2: y-axis) are shown alongside each axis.

To examine the impact of acidic pH_EX reperfusion on cardiac metabolism, following a 15 min. equilibration hearts were either (i) sampled immediately, or (ii) after being further subjected to 25 min. global no-flow ischemia, or (iii) 25 min. global no-flow ischemia plus 2 min. control reperfusion (pH_EX 7.4), or (iv) 25 min. global no-flow ischemia plus 2 min. acidic reperfusion (pH_EX 6.8) (Figure 4A).

Figure 4. Metabolomics study of acidic pH_EX reperfusion.

(A): Schematic showing design of the 4 experimental perfusion conditions. Hearts were snap frozen for steady-state metabolomics analysis by LC-MS as indicated by the arrows. Colors and terms used to denote each condition are used throughout. Number of hearts studied for each group: control 15, ischemia 19, control reperfusion 12, acid reperfusion 15. (B): Sparse partial least-squares discriminant analysis (sPLSDA) plot, prepared using MetaboAnalyst 5.0, incorporating 79 metabolites from 61 samples. Key at upper right denotes symbols for each group. The top 5 metabolite weightings contributing to each principal component (Component 1: x-axis, Component 2: y-axis) are shown alongside each axis.

Acidic pH_EX of 6.8 was obtained using modified KHB containing 3.8 mM NaHCO₃, gassed with 95 % O₂, 5 % CO₂. During reperfusion, cardiac effluents were collected at 1 min. intervals for 2 min. Samples were immediately treated with 10 % perchloric acid, followed by addition of 100 nmols butyrate as internal standard, and stored at −80 °C for later analysis.

To examine the effect of inhibiting ecto-5′-nucleotidase (5NT), hearts were subject to 25 min. global no-flow ischemia followed by 60 min. reperfusion, with optional infusion of 10 μM adenosine 5′-(α,β-methylene) diphosphate (meth-ADP), for 10 min. pre-ischemia and 2 min. post-ischemia into reperfusion.

For metabolomics, sampling comprised clamping the heart with pre-cooled Wollenberger tongs, plunging into liquid N₂ and grinding to powder, followed by storage under N₂ at −80 °C until analysis. Post-reperfusion hearts not used for metabolomics were transverse sliced (2 mm thick), stained with 1 % tetrazolium chloride, formalin fixed, and imaged for infarct measurement. Live (red) versus infarcted (white) tissue was quantified by planimetry using a custom MATLAB script.

2.5. Monitoring pH_IN in intact hearts

To examine cardiac pH_IN, hearts were perfused on a custom-built Langendorff system within a spectrofluorometer [34]. Hearts were equilibrated for 5 min., loaded with BCECF-AM (2.5 μM) for 20 min., followed by a 20 min. washout before imposition of acidic pH_IN conditions as described above. Fluorescence of BCECF was monitored (0.5 s read every 5 s.) at λ_EX 440 nm/λ_EM 530 nm and at λ_EX 490 nm/λ_EM 530 nm, with a 5 nm slit width, and expressed as ratio of fluorescence intensities (490/440 nm). pH_IN was calculated from a Boltzmann factor calibration curve, created using the protonophore FCCP (1 μM) and KHB at a range of pH values.

2.6. Metabolite Analysis

Metabolites were serially extracted in 80 % methanol, evaporated under N₂ then resuspended in 50 % methanol. Liquid Chromatography Tandem Mass Spectrometry (LC-MS/MS) analysis was performed as previously described [58], with reverse-phase LC on a Synergi Fusion-RP column (Phenomenex, Torrence CA) at 35 °C. Samples were analyzed by single reaction monitoring on a Thermo Quantum triple-quadrupole mass spectrometer. HPLC effluent was subject to electrospray ionization in negative ion mode using a HESI-II ion source. Metabolites were identified against a library of validated standards based on retention time, intact mass, collision energy, and fragment masses. Data were collected and analyzed using Thermo XCaliber 4.0 software. Post-hoc data analysis utilized Metaboanalyst 5.0 [59] for the generation of sPLSDA (sparse partial least-squares discriminant analysis) plots.

Each daily batch of 12 samples was bookmarked with a pooled sample to monitor instrument sensitivity. Following batch normalization, peaks were normalized to the sum of all peaks within each sample. Outliers were identified and discarded, as those values falling outside the 99.99 % confidence intervals for each group.

For the control/hypoxia/ischemia/acid pH_IN study, N for each group was 8. No samples were excluded, and across the 32 samples a total of 3.0 % of data were missing or discarded as outliers. A Jarque-Bera test for normality revealed 89 % of metabolites were normally distributed, permitting parametric statistical analysis. For the acidic pH_EX reperfusion study, the initial N for each group was: control 20, ischemia 20, normal reperfusion 14, acidic pH_EX reperfusion 17. Of these 71 original samples, 10 were excluded due to excessive noise (high numbers of outliers), and across the remaining 61 samples, a total of 3.3 % of data were missing or discarded as outliers. A Jarque-Bera test for normality revealed 84 % of metabolites were normally distributed.

Across the data sets for both studies, all missing values were imputed as medians of remaining values for each group [60]. Both studies contained control and ischemic groups, and although they were performed almost 3 years apart, and the latter study offered greater metabolome coverage (79 vs. 59 metabolites), comparison of the changes induced by ischemia in both studies revealed a significant correlation (r² 0.83, Figure S1), suggesting that comparison between them is valid.

2.7. Effluent Analysis by High-Performance Liquid Chromatography and UV-Vis Spectrophotometry

Effluents from control and acidic pH_EX reperfused hearts were analyzed by HPLC (Shimadzu Prominence 20A system) using two 300 × 7.8 mm Aminex HPX-87H columns (BioRad, Carlsbad CA, USA) in series, with 10 mM H₂SO₄ mobile phase (flow 0.7 ml/min) at 35 °C. 100 μl sample was injected. Carboxylic acids (succinate, fumarate, and lactate) were quantified via photodiode array (A₂₁₀) as previously described [61], with normalization to internal butyrate standard, and a standard curve for each metabolite. Effluents from the 1^st minute of reperfusion were also analyzed by spectrophotometry in the range 225–300 nm using a Beckman DU800 spectrophotometer (Carlsbad CA, USA), to approximate the content of nucleotides (A₂₅₀) and nucleosides (A₂₆₀).

2.8. Mitochondrial PT pore opening

Mouse liver mitochondria were isolated by differential centrifugation, and PT pore opening measured spectrophotometrically as osmotic swelling-induced decrease in light scatter/absorbance at 540 nm, both as previously described [58]. Liver mitochondria were used so that liver tissue could be procured from the same animals used for heart perfusion experiments, thus limiting animal usage. Furthermore, liver mitochondria are more mechanically fragile and better suited to swelling-based PT pore assays than heart mitochondria. MGO or the pore inhibitor cyclosporin A (CsA) were added at concentrations indicated in figure legends, prior to triggering of PT pore opening by 100 μM Ca²⁺.

2.9. Western blotting

Heart or cardiomyocyte samples were separated by SDS-PAGE and immunoblotted to nitrocellulose, then probed with a mouse monoclonal anti-MGO adduct antibody (AbCam ab243074). Blot detection used secondary antibodies, ECL reagents, and a KwikQuant imaging system, all from Kindle Bioscience.

2.10. Statistical Analysis

Student’s t-test (paired and unpaired) and ANOVA with post-hoc Tukey’s HSD test were applied where appropriate. Significance was defined as α=0.05. For metabolomics data from the control/hypoxia/ischemia/acid pH_IN study, the Storey modification of the Benjamini-Hochberg correction for false discovery rate (FDR) was applied [62], with q-values (FDR-corrected p-values) displayed in results. For the acidic pH_EX reperfusion study, application of FDR correction revealed all metabolites exhibiting a significant p-value (Tukey’s HSD test) remained significant after FDR correction, so p-values are displayed in results.

3. RESULTS

The complete original data set is available on the data sharing site FigShare (DOI: 10.6084/m9.figshare.16602281 – DOI reserved, unembargoed upon publication).

3.1. Cardiomyocyte pH_IN

Before proceeding to heart perfusions, we first screened a range of pharmacologic and physiologic treatments to induce pH_IN acidification in isolated primary adult mouse cardiomyocytes. Although cardiomyocytes are only 20 % of cells in the heart by number, they occupy 80 % of cardiac volume, suggesting treatment regimens validated in isolated cardiomyocytes could be applied to achieve acidification of pH_IN in the intact heart. As shown in Figures 1B/C and Figure S2, the combination of cariporide (NHE1 inhibitor), S0859 (NBC inhibitor) and pH_EX 7.1 yielded a drop in pH_IN of ~0.3 units, to pH 6.87. Although a similar degree of acidification was observed with MCT-1 inhibition (AR-C155858), we chose not to pursue this strategy due to its known effects on key metabolic pathways (lactate/succinate dynamics). As shown in Figure 1D, using a recently developed in-situ reflectance-fluorescence system [34] and the pH sensitive probe BCECF, we also found that the cariporide/S0859/pH_EX 7.1 combination lowered cardiac pH_IN by the same degree in intact perfused hearts.

3.2. Acidic pH_IN lowers but does not stop cardiac function

Figure 1E shows that hearts treated with the combination of cariporide/S0859/pH_EX 7.1, exhibited a gradual decline in contractile function (rate x pressure product), stabilizing at around 55% of that seen in controls. The decline was driven by a lowering of both heart rate (chronotropy) and left-ventricular developed pressure (inotropy) (Figure S3A/B). A fall in coronary root pressure was also seen (data not shown), consistent with the well-known vasodilatory effects of acid. The impact of other treatments (hypoxia and ischemia) on cardiac function is shown in Figure S3C/D.

3.3. Control, hypoxia, ischemia, acid pH_IN effects on cardiac metabolism

A total of 59 metabolites were reliably measured across all 32 samples examined (N=8 per condition). Application of sparse partial least squares discriminant analysis (sPLSDA, a dimensionality reduction tool, Figure 2) revealed that the 4 treatment conditions distributed into independent clusters along 2 axes. Differences between control and ischemia moved primarily along component 1 (x-axis), whereas differences induced by hypoxia or acidic pH_IN moved primarily along component 2 (y-axis). This suggests the fundamental character of metabolic changes brought about by ischemia is different than the other 2 conditions. This is borne out in volcano plots (Figures 3A–C), illustrating the magnitude and statistical significance of metabolic changes brought about by the 3 perturbations. Ischemia induced a far greater number of changes than did hypoxia or acidic pH_IN. A potential confounding factor in these divergent effects is the absence of coronary flow in ischemia, versus persistent flow in the hypoxia and acid pH_IN conditions.

Figure 3. Metabolomics reveals differential effects of hypoxia, ischemia, acidosis.

Volcano plots showing relative metabolite abundance in (A) hypoxia, (B) ischemia, and (C) acid pH_IN, vs. control perfusion. Log₁₀ fold changes are on x-axes and Log₁₀ FDR-corrected p-values (q-values) on y-axes. Metabolites in gray shaded areas are significantly (q-value < 0.05) up or down (>1.5 fold). Data are means, N=8 hearts per condition. (D/E): Venn diagrams showing commonalities and differences between metabolites either decreased (D) or increased (E) by each of the 3 perturbations.

A convenient tool to visualize the different metabolic changes induced by the 3 perturbations is Venn diagrams, as shown in Figures 3D/E. Surprisingly, only 1 metabolite change was shared between all 3 perturbations – namely a decrease in the glycogen synthetic precursor UDP-glucose. This suggests that acidic pH_IN may be an important signal in stimulating glycogen synthesis during ischemia or hypoxia. It has been reported previously that acidosis directly stimulates hepatic glycogen synthesis [63]. In addition, although glycogen is an important source of glucose during ischemia [35], it is known glycogen turnover is significant in the ischemic heart [64], suggesting that acidic pH_IN during ischemia may be an important signal for maintaining glycogen synthesis under conditions of enhanced glycogen utilization.

Further examining metabolites that were decreased, a lowering of fructose-1,6,-bis-phosphate (F-1,6-BP) was seen in both hypoxia and ischemia, but notably this metabolite was elevated by acidic pH_IN. This is somewhat counterintuitive, since acidic pH is known to inhibit PFK1, the enzyme that produces F-1,6-BP. Furthermore, several other metabolites in the “top-half” of glycolysis (i.e., glucose-6-P, glucose-1-P, and fructose-6-P) were all significantly increased by acidic pH_IN. This is consistent with an acid-induced inhibition at the level of GAPDH (the forward reaction of which generates a H⁺), and this mid-point of glycolysis is increasingly viewed as a key regulatory stage for the pathway [65, 66]. These differential effects of acid pH_IN alone, versus acidification that occurs in the context of ischemia or hypoxia, would appear to suggest that in the latter 2 conditions, other factors may over-ride the effect of acid on glycolysis. This could include limited substrate availability in the case of no-flow ischemia, or the need to maintain glycolytic flux to provide ATP in the absence of mitochondrial Ox-Phos under hypoxic conditions. Overall, it appears that perturbations to glycolytic activity seen in ischemia are unlikely driven purely by the effects of acidic pH_IN.

A number of other changes were observed in hearts subjected to acidic pH_IN. In particular, a 32% drop in oxidized glutathione (GSSG), coupled with a 10-fold rise in reduced glutathione (GSH, although this latter change did not reach statistical significance) together suggest that acidic pH_IN may favor a more reduced cellular environment, enhancing resistance to oxidative stress. This is consistent with the general notion that acidic pH is cardioprotective [37, 38, 49, 51–53]. The pK_a of the GSH thiol is around 9 [67, 68] such that it is 99% protonated at physiologic intracellular pH values. As such, we do not consider it likely that further acidification would directly impact the GSH redox couple by enhancing protonation of the GSH thiol group.

Another notable feature of the acidic pH_IN metabolome was a depression in the levels of glutamate. Previously, we showed that anaplerosis into the TCA cycle from glutamate is a significant feature of ischemia [35], although the impact of acidic pH_IN on this was not considered. The current data suggest glutamate anaplerosis to α-ketoglutarate may be sensitive to by pH_IN. Importantly this would not impact succinate accumulation, since the primary source for accumulation of succinate in ischemia is canonical TCA activity with TCA endogenous metabolites serving as the source [35].

3.4. Acidic pH_EX during the metabolic transition from ischemia to reperfusion

In addition to enforced pH_IN, we examined the metabolic impact of acidic pH_EX upon reperfusion, because this regimen has been shown to induce cardioprotection. First, we confirmed that delivering a pH_EX of 6.8 for the first 2 minutes of reperfusion was cardioprotective, observing a mild but non-significant improvement in functional recovery and a significant lowering of infarct size (Figure S4). This is consistent with a meta-analysis of 13 previous studies examining cardioprotection by acidic reperfusion, in which impacts of acid pH_EX reperfusion on functional recovery were highly variable, while infarct size reductions were more robust (Table S1).

Although application of pH_EX 6.8 to isolated cardiomyocytes did invoke a small drop in cellular pH to ~6.99 (Figure S2), attempts to measure the impact of 2 minutes of pH_EX 6.8 reperfusion on the pH_IN of reperfused hearts were unsuccessful, likely owing to the massive and rapid re-alkalinization of pH_IN already occurring during this time frame [38].

A total of 79 metabolites were reliably measured across 61 samples included in the acid pH_EX reperfusion study (N=12–19 per condition). Application of sPLSDA (Figure 4B) revealed that the 4 treatment conditions distributed into independent clusters along 2 component axes. Similar to the control/hypoxia/ischemia/acid pH_IN study, differences between control and ischemia moved primarily along component 1 (x-axis), with several shared highly-weighted contributors between both studies (compare to Figure 2B). Upon reperfusion, both control and acidic pH_EX groups moved back along component 1 toward the control state, but also introduced novel features along a different axis, component 2 (y-axis). Notably, acidic pH_EX reperfusion diverged further along this axis than control reperfusion. Together these data suggest that control and acidic pH_EX reperfusion are largely similar in character, with the latter inducing greater changes.

Figures 5A/B show volcano plots for changes in metabolite levels with either control or acidic pH_EX reperfusion, compared to ischemia. Figure 5C shows key differences between metabolites in control or acidic pH_EX reperfused hearts. Visualization of these data as Venn diagrams (Figures 5D/E) indicates that the vast majority of metabolic changes induced by control reperfusion were also present in acidic pH_EX reperfusion (i.e., there were very few metabolic changes that were unique to control reperfusion and lost in acidic pH_EX reperfusion). In contrast, acidic pH_EX reperfusion induced a large number of additional metabolic changes that were unique to this condition and were not seen in control reperfusion.

Figure 5. Metabolomics reveals differential effects of control vs. acid pH_EX reperfusion.

(A/B): Volcano plots showing relative metabolite abundance transitioning from ischemia into either (A) control reperfusion (pH_EX 7.4) or (B) acid reperfusion (pH_EX 6.6). (C): Relative metabolite abundance comparing control and acid pH_EX reperfusion. For volcano plots, Log₁₀ fold changes are on the x-axis and Log₁₀ p-values are on the y-axis. Metabolites in gray boxes are significantly (p-value < 0.05) up or down (>2.0 fold). Data are means. All metabolites shown as significant remained so after Storey false discovery rate (FDR) correction. (D/E): Venn diagrams showing commonalities and differences between metabolites either decreased (D) or increased (E) as a result of either control or acid pH_EX reperfusion. (F): Relative changes in abundance of reduced and oxidized glutathione (GSH and GSSG respectively) from baseline into ischemia and then into either control or acid pH_EX reperfusion. Data are means ± SEM, N = CTRL 15, ISC 19, CTRL REP 12, ACID pH_EX REP 15.

Metabolite changes upon reperfusion that were common to both reperfusion conditions, were largely as expected. Most metabolites that decreased during ischemia were then increased upon reperfusion. This included several TCA cycle metabolites (citrate, isocitrate), as well as high energy phosphates (ATP, CTP, PCr) and metabolites in the lower half of glycolysis. Likewise, most metabolites that were elevated in ischemia were then decreased upon reperfusion in a manner that was not impacted by pH_EX. This included succinate, lactate, and 3-OH-butyrate.

There were several metabolite differences between control and acidic pH_EX reperfusion hearts, including the redox state of glutathione. As Figure 5F shows, GSH was elevated and GSSG was lower in ischemia, and these changes were reversed upon control reperfusion. However, in acidic pH_EX reperfusion both GSH and GSSG were maintained at their ischemic levels into reperfusion. This suggests acidic pH_EX reperfusion may carry a lower burden of oxidative stress than control. Although we have shown that cytosolic and mitochondrial pH can impact ROS generation by the respiratory chain [36], whether the cardioprotective benefits of acidic pH_EX reperfusion are conferred by a lowering of ROS generation, is currently unclear.

3.5. Cardioprotective roles for metabolites uniquely elevated by acidic pH_EX reperfusion

While most metabolites in the lower half of glycolysis returned to normal levels regardless of the reperfusion pH_EX, this was not the case for the top half of the pathway. As shown in Figure 6, hexose phosphates were all significantly elevated with acidic pH_EX reperfusion. This is consistent with the elevation of these metabolites caused by acidic pH_IN (Figure 3), and suggests that glycolysis responds rapidly to both pH_IN and pH_EX.

Figure 6. Metabolomics of glycolysis.

The pathway is shown with each metabolite that was measured highlighted in gray and linked to its data. Key denotes the color scheme for each treatment group as used in previous figures. Data are means of relative metabolite abundance (arbitrary units) ± SEM, N = CTRL 15, ISC 19, CTRL REP 12, ACID pH_EX REP 15.

A key observation was that levels of glyceraldehyde-3-phosphate / dihydroxyacetone-phosphate (DHAP) were retained in an elevated state upon acidic pH_EX reperfusion, versus declining with control reperfusion (Figure 6). An important byproduct generated from DHAP is the reactive metabolite methylglyoxal (MGO), which can post-translationally modify proteins (i.e., advanced glycation end products, AGEs) [69, 70]. It has also been shown that a related metabolite phenylglyoxal can inhibit the mitochondrial permeability transition pore [71, 72]. Given that acidic pH is a known inhibitor of the PT pore [24, 25], we hypothesized that MGO generation may be a molecular link between acidic pH and PT pore inhibition. As shown in Figure S5, MGO was able to inhibit PT pore opening in isolated mitochondria. However, western blotting for MGO protein adducts did not reveal any difference between control and acidic pH_EX reperfused hearts. Similar results were seen in primary cardiomyocytes exposed to acidic pH_EX. Together these data suggest that MGO is unlikely responsible for the cardioprotection induced by acidic pH_EX.

3.6. Cardioprotective roles for metabolites uniquely decreased by acidic pH_EX reperfusion

A number of differences were observed in the levels of purine nucleotides and their metabolites, between control and acid pH_EX reperfused hearts (Figure 5E). As shown in Figure 7, acid pH_EX reperfused hearts consistently had lower levels of nucleosides (guanosine, adenosine, inosine, xanthine, xanthosine, hypoxanthine) and elevated ratios of nucleotide triphosphates vs. diphosphates (GTP/GDP and ATP/ADP). This led us to hypothesize that the activity of ecto-5′-nucleotidase (5′-NT) may be inhibited by acidic pH_EX, preventing purine nucleotide loss upon reperfusion. To determine whether this could play a role in the mechanism of cardioprotection elicited by acidic pH_EX reperfusion, we subjected hearts to IR in the presence of the 5′-NT inhibitor meth-ADP. However, as shown in Figure S6, this intervention failed to elicit cardioprotection. Thus, although acidic pH_EX reperfusion appears to modulate the purine salvage pathway, this may not be responsible for its concomitant cardioprotection.

Figure 7. Metabolomics of the purine nucleotide salvage pathway.

The pathway is shown with each metabolite that was measured highlighted in gray and linked to its data. Key denotes the color scheme for each treatment group as used in previous figures. Data are means of relative metabolite abundance (arbitrary units) ± SEM, N = CTRL 15, ISC 19, CTRL REP 12, ACID pH_EX REP 15.

3.7. Impact of acid pH_EX reperfusion on metabolite efflux.

Upon reperfusion, around ⅓ of accumulated succinate is oxidized to generate ROS, while ⅔ washes out from the heart [15, 35]. Although acidic pH_EX has been shown to inhibit succinate release from cells via MCT-1, this occurs at much lower pH_EX values (pH 6.0 in [15]) than employed herein (pH 6.8). Furthermore, ischemic succinate that accumulates in the extracellular space will presumably wash out in a manner unaffected by pH.

To examine the impact of acidic pH_EX reperfusion on metabolite release, we measured levels of carboxylates (i.e., succinate, fumarate and lactate) and purines (i.e., nucleotides and nucleotides) in effluents collected from hearts during the first minutes of control or acid pH_EX reperfusion. As shown in Figure S7, acid pH_EX reperfusion significantly enhanced succinate efflux during the first minute of reperfusion. This is contrary to the finding that pH_EX 6.0 can inhibit succinate release [15], and suggests that mild acidification may enhance succinate efflux. Highlighting the speed with which succinate is oxidized upon reperfusion [34], we observed that acid pH_EX reperfusion also enhanced the efflux of fumarate, for which the likely origin is succinate oxidation by mitochondrial complex II. It should be noted that post-reperfusion succinate returns to normal levels within the heart very quickly [35], such that no difference was seen in the cardiac levels of succinate between control and acid pH_EX reperfused hearts (Figure 5), as these cardiac samples were collected after 2 minutes of reperfusion.

No impact of pH_EX on the efflux of lactate was observed, suggesting the effect of pH_EX on succinate seen here is not a general impact on MCT-1 activity. There was also no impact of pH_EX during reperfusion on the efflux of purines (Figure S6), suggesting that the observed differences in cardiac levels of purines (Figure 7) do not originate from differential wash-out of these molecules into effluents.

4. DISCUSSION

Acidic pH is a hallmark of the enhanced glycolysis that occurs to compensate for impaired mitochondrial Ox-Phos during tissue hypoxia/ischemia. However, surprisingly little is known about the impact of this acidosis on metabolism within the context of hypoxia/ischemia. Furthermore, although extension of ischemic acidosis into early reperfusion is proven to significantly reduce myocardial infarction, the mechanism of this protection is not fully understood, including the role (if any) of altered metabolism.

It is well known that acidosis decreases contractility of cardiac muscle, impacting all stages of excitation-contraction coupling [22, 73, 74], including Ca²⁺ dynamics and myofilament function [23, 75, 76]. As such, when regarding the impact of forced acidic pH_IN on metabolism, it is possible that lowering of cardiac function (as shown in Figure 1E) may have contributed to the observed metabolite differences. Likewise, the extension of ischemic acidosis into reperfusion, as seen with acidic pH_EX reperfusion, may have a beneficial effect via lowering ATP demand thus allowing cardiomyocytes to re-establish ionic homeostasis.

Although NHE1 inhibitors have exhibited robust cardioprotection in pre-clinical animal models of ischemia-reperfusion injury, these agents faltered in human clinical trials, reportedly due to sub-optimal trial design (timing of delivery relative to infarct) and neurovascular side effects [54, 55, 77]. The mechanism of action for these agents is largely attributed to cardiac Na⁺ dynamics, but the role of acidosis as an additional protective mechanism downstream of NHE1 inhibition [78] has received little attention. Examining the impact of NHE inhibition alone on cardiac metabolism may be useful, to determine if acid-induced metabolic changes underlie some of the mechanism of action for these drugs. In this regard, it has been shown that only about 35% of the recovery of cardiomyocyte pH_IN during reperfusion can be accounted for by NHE activity [79]. Furthermore, our cell data (Figure S2) show that NHE inhibition alone did not lower pH_IN as much as acidic pH_EX did. As such, we might predict that the magnitude of metabolic changes that could be bought about by NHE inhibitors alone, would be smaller than those observed here with acid pH_EX. A detailed examination of this is beyond the scope of the current study.

Maintaining redox homeostasis is also critical for cardiac function and survival, and pH is a known regulator of redox processes [80, 81]. Herein, both forced pH_IN and acidic pH_EX reperfusion (Figures 3 and 5) revealed that acidic pH caused a shift toward a more reduced intracellular glutathione redox state (higher GSH, lower GSSG). This would be favorable in detoxifying mitochondrial ROS generated at reperfusion [82], and may be a contributing factor that underlies the cardioprotection conferred by acidic pH_EX reperfusion. Although the biochemistry of glutathione is pH sensitive (i.e., the GSH <–> GS⁻ + H⁺ equilibrium), the pK_a of its thiol is around 9, such that at physiologic pH it will be largely protonated [67, 68]. As such, further acidification would be unlikely to alter the availability of the thiolate anion, and simple pH effects on GSH thiol protonation are unlikely to underlie the observed alterations in glutathione redox state. This does not preclude the possibility that pH may have a significant impact on other thiols in the cell with lower pK_a values, which may impact their redox state. A potential limitation to this aspect of the study was our use of the C57BL/6J mouse sub-strain, which carries a mutation in the mitochondrial nicotinamide nucleotide transhydrogenase that is important for maintaining mitochondrial antioxidant networks [83]. It would be interesting to see if acid pHEX reperfusion has a similar effect on glutathione redox state in the C57BL/6N sub-strain.

In considering other potential metabolic differences between control vs. acid pH_EX reperfusion that may underlie cardioprotection, the top half of glycolysis was elevated in acid pH_EX (Figure 6), but this did not appear to be linked to the reactive metabolite MGO (Figure S5B), suggesting MGO is not a protective mediator. Likewise, although purine nucleotide salvage appeared to be altered by acid pH_EX reperfusion (Figure 7), with a potential inhibition of 5′-NT activity, pharmacologic inhibition of 5′-NT was unable to recapitulate a protected phenotype. This is possibly because some products of 5′-NT such as adenosine and inosine are endogenously generated during conditions such as ischemic preconditioning, and are known to confer cardioprotection [84–87]. As such, 5′-NT inhibition may have a detrimental effect by removing endogenously produced adenosine or inosine from the heart. These findings do not preclude the possibility that some of the protective effect of acid pH_EX reperfusion may be mediated by prevention of purine loss.

A surprising finding was the impact of acidic pH_EX reperfusion on the release of metabolites from myocardium during the first minutes of reperfusion. Although no impact of pH_EX on the release of lactate was seen, there was a significantly larger release of succinate with acid pH_EX during the first minute of reperfusion. Previously it has been shown that extreme pH_EX acidosis (pH 6.0) can inhibit succinate export via MCT-1, but the current results suggest a complex relationship between pH and succinate release, with mild acidosis causing an increase in release, not a decrease. Such an enhanced release of succinate would be expected to confer a cardioprotective benefit, by lowering the amount of succinate available inside cardiomyocytes for ROS generation [34]. An important caveat is that although more succinate was released from the heart in acid pH_EX vs. control reperfusion (Figure S6), this did not result in less succinate in the heart itself 2 min. into reperfusion (Figure 5). This is likely because succinate is consumed rapidly, such that by the time of sampling, cardiac succinate levels have returned almost to baseline [35]. Sampling at earlier times of reperfusion would be required to track the fate of succinate (both intracellular and extracellular) on a more rapid timescale. However, such short sampling times bring increased experimental error.

In summary, the first study (control/hypoxia/ischemia/acid pH_IN) highlights that acid pH can independently rewire metabolism, although not nearly to the extent that hypoxia or ischemia can. In-fact, there were surprisingly few metabolic alterations unique to acid pH_IN that were not also seen with the other two interventions. As such, acid pH alone does not seem to underlie the majority of metabolic changes seen in hypoxia or ischemia. In contrast the second study, on acid pH_EX reperfusion, shows that this intervention brings about a large number of metabolic changes in addition those seen in control reperfusion. Some of these novel acid-induced metabolic alterations (enhanced succinate release and prevention of purine loss) may partly underlie the cardioprotective benefits of acid pH_EX reperfusion. Further studies are required to discern the impact of acid pH_EX reperfusion on downstream signaling pathways such as mitochondrial ROS generation and PT pore opening. Comparing the two studies, while there are commonalities between both forms of acidosis (acid pH_IN and acid pH_EX reperfusion), the latter was surprisingly more impactful, perhaps reflecting an enhanced vulnerability of metabolism during the first minutes of reperfusion.

HIGHLIGHTS.

• Hypoxia, ischemia and acidic pH_IN each induce unique cardiac metabolic profiles.

• Acidic pH_EX at reperfusion prevents purine loss and enhances succinate release.

• Surprisingly, acidic pH_EX reperfusion has a greater metabolic impact than forced acid pH_IN

ABBREVIATIONS

5NT

ecto-5′-nucleotidase

A-pH_EX

Acidic extracellular pH

ATP

adenosine triphosphate

BCECF-AM

2′−7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein, acetoxymethyl ester

CHIA

Control, Hypoxia, Ischemia, Acidic pH_IN

IR

ischemia-reperfusion

MCT

monocarboxylate transporter

meth-ADP

adenosine 5′-(α,β-methylene) diphosphate

NBC

Na⁺/HCO₃⁻ cotransporter

NHE

Na⁺/H⁺ exchanger

NMN

Nicotinamide mononucleotide

Ox-Phos

oxidative phosphorylation

pH_EX

extracellular pH

pH_IN

intracellular (cytosolic) pH

PT pore

permeability transition pore

ROS

reactive oxygen species

RPP

rate x pressure product

TCA

tricarboxylic acid