Inhibiting Succinate Release Worsens Cardiac Reperfusion Injury by Enhancing Mitochondrial Reactive Oxygen Species Generation

Abstract

Background

The metabolite succinate accumulates during cardiac ischemia. Within 5 minutes of reperfusion, succinate returns to baseline levels via both its release from cells and oxidation by mitochondrial complex II. The latter drives reactive oxygen species (ROS) generation and subsequent opening of the mitochondrial permeability transition (PT) pore, leading to cell death. Targeting succinate dynamics (accumulation/oxidation/release) may be therapeutically beneficial in cardiac ischemia–reperfusion (IR) injury. It has been proposed that blocking MCT1 (monocarboxylate transporter 1) may be beneficial in IR injury, by preventing succinate release and subsequent engagement of downstream inflammatory signaling pathways. In contrast, herein we hypothesized that blocking MCT1 would retain succinate in cells, exacerbating ROS generation and IR injury.

Methods and Results

Using the mitochondrial ROS probe mitoSOX and a custom‐built murine heart perfusion rig built into a spectrofluorometer, we measured ROS generation in situ during the first moments of reperfusion. We found that acute MCT1 inhibition enhanced mitochondrial ROS generation at reperfusion and worsened IR injury (recovery of function and infarct size). Both of these effects were abrogated by tandem inhibition of mitochondrial complex II, suggesting that succinate retention worsens IR because it drives more mitochondrial ROS generation. Furthermore, using the PT pore inhibitor cyclosporin A, along with monitoring of PT pore opening via the mitochondrial membrane potential indicator tetramethylrhodamine ethyl ester, we herein provide evidence that ROS generation during early reperfusion is upstream of the PT pore, not downstream as proposed by others. In addition, pore opening was exacerbated by MCT1 inhibition.

Conclusions

Together, these findings highlight the importance of succinate dynamics and mitochondrial ROS generation as key determinants of PT pore opening and IR injury outcomes.

Nonstandard Abbreviations and Acronyms

AA5

atpenin A5

AR

AR‐C155858

CsA

cyclosporin A

Cx

mitochondrial complex

DMM

dimethyl malonate

IR

ischemia–reperfusion

λ_EM

emission wave length

λ_EX

excitation wavelength

MCT1

monocarboxylate transporter 1

NADH

nicotinamide adenine dinucleotide reduced form

NAD(P)H

NADH and nicotinamide adenine dinucleotide phosphate reduced forms

PT

permeability transition

ROS

reactive oxygen species

S1QEL

suppressor of site I_Q electron leak

SUCNR1

succinate receptor 1

TMRE

tetramethylrhodamine ethyl ester

ΔΨ_m

mitochondrial membrane potential

Clinical Perspective.

What Is New?

• A new method has been developed for simultaneous measurement of both cardiac function and small‐molecule probe fluorescence (reactive oxygen species and mitochondrial membrane potential) in the perfused mouse heart.

• Acutely blocking succinate release via MCT1 (monocarboxylate transporter 1) worsens ischemia–reperfusion injury by exacerbating reactive oxygen species generation during the first minute of reperfusion in a manner that can be abrogated by tandem blockade of succinate oxidation by mitochondrial complex II.

• Characterization of the kinetics of the mitochondrial permeability transition pore relative to reactive oxygen species generation reveals that reactive oxygen species lie upstream of the pore, not downstream as others have proposed.

What Are the Clinical Implications?

• Although inhibition of MCT1 may confer cardioprotection in vivo by preventing extracellular succinate from engaging inflammatory signaling, our data suggest inhibiting MCT1 may worsen acute injury by exacerbating damage from intracellular succinate in early reperfusion.

• Because the detrimental effects of MCT1 inhibitors are abrogated by mitochondrial complex II inhibition, optimal therapy for ischemia–reperfusion injury should incorporate both complex II and MCT1 inhibitors to deal with the consequences of both intracellular and extracellular succinate, respectively.

• Despite recent clinical trial failures, these basic science data advocate continued interest in the therapeutic potential of the permeability transition pore inhibitor cyclosporin A for ischemia–reperfusion injury.

The metabolite succinate has a central role in tissue ischemia. Several mechanisms exist for ischemic succinate accumulation, ¹ , ² and this process is conserved across diverse tissues, species, and physiologic contexts. ³ , ⁴ However, rapid oxidation of accumulated succinate at the onset of tissue reperfusion is a key event in ischemia–reperfusion (IR) injury, the underlying pathology of myocardial infarction (heart attack). ¹ , ² This has led to intense interest in succinate as a potential therapeutic target. ⁵ , ⁶ , ⁷ , ⁸ , ⁹

Approximately one‐third of succinate accumulated during ischemia is rapidly oxidized by mitochondrial complex II (Cx‐II) of the mitochondrial respiratory chain. ² , ¹⁰ This leads to the generation of reactive oxygen species (ROS) via mechanisms that are thought to involve either reverse electron transport at Cx‐I ¹¹ , ¹² , ¹³ or forward electron transport at Cx‐III. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ Mitochondrial ROS generation potentiates opening of the mitochondrial permeability transition (PT) pore, ¹⁹ , ²⁰ , ²¹ , ²² which triggers necrotic cell death.

The remaining two‐thirds of succinate accumulated during ischemia is released upon reperfusion, in a pH‐dependent manner via MCT1 (monocarboxylate transporter 1). ¹⁰ , ²³ The physiologic function of this released succinate is unclear, with suggestions that it may serve as a metabolic signal of hypoxia. ²⁴ Succinate is a ligand for the widely expressed SUCNR1 (succinate receptor 1, formerly orphan G protein coupled receptor 91, GPR91), which elicits a range of physiologic responses. ²⁵ Most relevant to IR injury, SUCNR1 signaling promotes inflammation via macrophage activation, which may contribute to the pathology of IR injury. ¹⁰ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ As such, it has been postulated that inhibiting MCT1 would be beneficial in IR injury, via blunting of extracellular succinate ➔ SUCNR1 signaling. ¹⁰ , ³¹

In contrast, given the key role of intracellular succinate for ROS generation during reperfusion, we hypothesized herein that acute blockade of succinate release via MCT1 would worsen IR injury, and that simultaneous blockade of Cx‐II would abrogate this effect. This hypothesis was tested by measuring mitochondrial ROS generation in situ in perfused mouse hearts using a custom‐built perfusion rig within the chamber of a benchtop spectrofluorometer. In addition, because it has been proposed that PT pore opening itself may lie upstream of the burst of ROS seen upon reperfusion, ²³ , ³² we used this apparatus to interrogate the temporal relationship between ROS generation and PT pore opening during early reperfusion.

Methods

All original data used to prepare the figures are available in spreadsheet form on the data‐sharing website FigShare (https://doi.org/10.6084/m9.figshare.19319627). All methods used are as detailed below, with further information and materials or other resources available from the authors upon request.

Animals and Reagents

Animal and experimental procedures complied with the National Institutes of Health Guide for Care and Use of Laboratory Animals (8th Edition, 2011) and were approved by the University of Rochester Committee on Animal Resources (protocol number 2007–087). Male and female C57BL/6J adult mice (8–20 weeks old) were housed in a pathogen‐free vivarium with 12‐hour light–dark cycles and food and water ad libitum. Mice were administered terminal anesthesia via intraperitoneal 2,2,2‐tribromoethanol (Avertin) ≈250 mg/kg. Avertin was prepared in amber glass vials and stored at 4 °C for not >1 month. This agent was chosen for anesthesia because it does not impact cardioprotection as reported for volatile anesthetics or opioids ³³ , ³⁴ , ³⁵ and does not have mitochondrial depressant effects as reported for barbiturates. ³⁶ Euthanizing occurred via cardiac extirpation (see below).

MitoSOX red was from Thermo (NJ) and was stored aliquoted under argon before use. Fully oxidized mitoSOX was prepared by reaction with Fremy salt as described elsewhere. ³⁷ AR‐C155858 (AR) was from MedChemExpress (Monmouth Junction, NJ). Unless otherwise stated, all other reagents were from MilliporeSigma (St. Louis, MO).

Perfused Mouse Hearts

Following establishment of anesthetic plane (toe‐pinch response), beating mouse hearts were rapidly cannulated, excised, and retrograde perfused at a constant flow (4 mL/min) with Krebs‐Henseleit buffer consisting of (in millimoles per liter): NaCl (118), KCl (4.7), MgSO₄ (1.2), NaHCO₃ (25), KH₂PO₄ (1.2), CaCl₂ (2.5), glucose (5), pyruvate (0.2), lactate (1.2), and palmitate (0.1, conjugated 6:1 to bovine serum albumin). Krebs‐Henseleit buffer was gassed with 95% O₂ and 5% CO₂ at 37 °C. A water‐filled balloon connected to a pressure transducer was inserted into the left ventricle and expanded to provide a diastolic pressure of 6 to 8 mm Hg. Cardiac function was recorded digitally at 1 kHz (Dataq, Akron, OH) for the duration of the protocol. Following equilibration (10–20 minutes) IR injury comprised 25 minutes. Global no‐flow ischemia plus 60 minutes. Reperfusion. Hearts were then sliced and stained with triphenyltetrazolium chloride for infarct quantitation by planimetry (red=live tissue, white=infarct). Infarction analysis was blinded to experimentalists.

The following conditions (Figure 1F) were examined: (1) Control: DMSO vehicle infusion 5 minutes before ischemia and 5 minutes into reperfusion. (2) S1QEL: Suppressor of site I_Q electron leak 1.1: 1.6 μmol/L, delivered 5 minutes before ischemia and 5 minutes into reperfusion. (3) DMM5: Dimethyl malonate (DMM), 5 mmol/L at the onset of reperfusion and 5 minutes into reperfusion. (4) AR: MCT1 inhibitor AR‐C155858 (AR), 10 μmol/L infusion 5 minutes before ischemia and 5 minutes into reperfusion. (5) AR+DMM5: 10 μmol/L AR infusion 5 minutes before ischemia and 5 minutes into reperfusion, plus 5 mmol/L DMM at the onset of reperfusion and 5 minutes into reperfusion. (6) AR+DMM10: 10 μmol/L AR infusion 5 minutes before ischemia and 5 minutes into reperfusion plus 10 mmol/L DMM at the onset of reperfusion and 5 minutes into reperfusion. (7) AR+AA5: 10 μmol/L AR infusion 5 minutes before ischemia and 5 minutes into reperfusion plus 100 nmol/L atpenin A5 (AA5) at the onset of reperfusion and 5 minutes into reperfusion. (8) CsA: 0.8 μmol/L cyclosporin A (CsA) infusion ³⁸ 5 minutes before ischemia and 5 minutes into reperfusion. Simultaneously with these conditions, hearts were delivered either 1.5 μmol/L mitoSOX for 5 minutes or 500 nmol/L tetramethylrhodamine ethyl ester (TMRE) for 20 minutes before ischemia.

Figure 1. Fluorescence cardiac perfusion system.

A, Schematic showing arrangement of the perfused mouse heart relative to light source and photomultiplier tube (PMT). B, Photograph of heart in situ before mounting inside spectrofluorometer. C, Photograph of cardiac perfusion apparatus from above mounted inside spectrofluorometer. Components of the system are labeled, including a custom 3‐dimensionally printed water‐jacketed cuvette holder, heated umbilical lines to deliver Krebs‐Henseleit buffer at 37 °C to the heart, and vacuum line to remove perfusate/effluent. D, Nicotinamide adenine dinucleotide phosphate (NAD[P]H) autofluorescence (340/460 nm) during ischemia and reperfusion. Traces in gray are individual hearts, with the average (N=7) shown in black. E, Flavoprotein autofluorescence (462/520 nm) during ischemia and reperfusion. Traces in gray are individual hearts, with the average (N=4) shown in black. Fluorescence data are normalized to preischemic levels. F, Schematic showing design of the 7 experimental perfusion conditions examined herein. Where indicated, suppressor of site I_Q electron leak 1.1 (S1QEL1.1) (mitochondrial complex [Cx]‐I reactive oxygen species inhibitor), dimethylmalonate (DMM; Cx‐II inhibitor), AR‐C155858 (AR; MCT1 [monocarboxylate transporter 1] inhibitor), or atpenin A5 (AA5; Cx‐II inhibitor) were administered at the listed concentrations. Colors and terms used to denote each condition are used throughout the remainder of the figures.

Perfusion Apparatus Within Spectrofluorometer

A cardiac perfusion apparatus was custom built with an umbilicus to position the heart within the light‐proof enclosure of a Varian/Cary Eclipse benchtop spectrofluorometer (Agilent, Santa Clara, CA), as shown in Figure 1A through 1C. Hearts were placed against the wall of a water‐jacketed cuvette maintained at 37 °C in which the excitation light source strikes the left ventricle of the heart at a 45° angle relative to the photomultiplier tube window. The cuvette holder was 3‐dimensionally printed (the .stl file is available upon request). Data were collected using Cary WinUV Kinetics software, which permitted real‐time fluorescence monitoring (1‐second reads/15‐second cycle, photomultiplier tube voltage 600, Ex/Em slit width=5 nm) for the duration of the perfusion protocol. At the heart rates observed (mean±SD, 438±54 bpm; N=54) a 1 second fluorescent read was averaged across approximately 7 heart beats, so gating for motion artifacts was unnecessary. Initial validation was performed by monitoring endogenous NAD(P)H fluorescence (λ_EX 340 nm, λ_EM 460 nm) and flavoprotein fluorescence (λ_EX 460 nm, λ_EM 520 nm), as shown in Figure 1D and 1E.

Measuring In Situ Mitochondrial ROS

Mitochondrial ROS generation was measured using the probe mitoSOX ³⁹ , ⁴⁰ (triphenylphosphonium‐conjugated dihydroethidium) at a concentration of 1.5 μmol/L, because the triphenylphosphonium moiety is known to uncouple mitochondria at concentrations ≥2.5 μmol/L. ⁴¹ , ⁴² , ⁴³ Hearts subject to conditions 1 to 7 described above were equilibrated for 10 minutes, loaded with 1.5 μmol/L mitoSOX for 5 minutes, and then immediately subjected to 25 minutes ischemia. Fluorescence was monitored at λ_EX 510 nm and λ_EM 580 nm.

Because the λ_EX and λ_EM of mitoSOX overlap with the absorbance spectra of endogenous chromophores in cardiomyocytes (eg, myoglobin and cytochromes), changes in chromophore absorbance during the course of IR could impact the fluorescent signal. Furthermore, the distribution of the mitoSOX probe between cytosolic and mitochondrial matrix compartments is determined by the mitochondrial membrane potential (ΔΨ_m), which would also be expected to change during IR. To correct for these potential cofounding effects, a series of hearts were loaded with fully oxidized mitoSOX ³⁷ that fluoresces in a manner that is independent of ROS generation, but is still subject to the effects of chromophore absorbance and probe distribution. Fluorescent data from these hearts were used to correct data obtained with naïve mitoSOX (conditions 1–7), yielding a net signal that originated only from in situ probe oxidation, without contribution from other factors. The data process for this correction is illustrated in Figure S1, which also shows a minimal change in signal during IR when no mitoSOX was present. In addition, absorbance spectra of compounds tested in this study (eg, AR, DMM, S1QEL) were also measured to ensure these chemicals did not absorb significant amounts of light at the λ_EX and λ_EM of mitoSOX (Figure S2).

Measuring In Situ ΔΨ_m

To assess the timing of mitochondrial PT pore opening, mitochondrial membrane potential was measured using TMRE at 500 nmol/L delivered for 20 minutes before ischemia. The PT pore inhibitor CsA (0.8 μmol/L) was optionally delivered for 5 minutes before ischemia. To determine changes upon reperfusion, TMRE data were normalized to the fluorescent signal during the last 5 minutes of ischemia. Two conditions were examined: control IR and IR with MCT1 inhibition (same as condition 7 above). Attempts to determine the effect of CsA on PT pore opening in the presence of AR were confounded by interactions between these molecules and TMRE, resulting in precipitation of components in Krebs‐Henseleit buffer.

Effluent Analysis by High‐Performance Liquid Chromatography

Effluents from control and AR‐treated hearts were collected in 1‐minute intervals for the first 3 minutes of reperfusion and immediately treated with 10% perchloric acid. Following addition of 100 nmols butyrate as an internal standard, samples were frozen in liquid N₂ and stored at −80 °C until analysis. Samples were centrifuged at 20 000g to remove insoluble materials. Metabolites were resolved on high‐performance liquid chromatography (Shimadzu [Kyoto, Japan] Prominence 20 system) using two 300×7.8 mm Aminex HPX‐87H columns (BioRad, Carlsbad, CA) in a series with 10 mmol/L H₂SO₄ mobile phase (flow rate: 0.7 mL/min) and a 100‐μL sample injected on the column. Succinate and lactate were detected using a photodiode array measuring absorbance at 210 nm as previously described. ² A standard curve was constructed for calibration. Lactate data were corrected for 1.2 mmol/L lactate contained in the Krebs‐Henseleit buffer.

Statistical Analysis

Comparisons between the 2 groups were made using unpaired Student t tests. For multiple group comparisons, ANOVA was used with post hoc Tukey honest significant difference tests. Data are shown as mean±SEM. Numbers of biological replicates are noted in the figures. Significance was set at α=0.05.

Results

Cardiac Fluorescence

Although similar spectrofluorometric cardiac perfusion apparatus has previously been constructed, ³² , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ prior efforts have used larger animal hearts (rats, guinea pigs) or have not simultaneously measured cardiac function. To the best of our knowledge, this is the first study of mouse hearts with simultaneous fluorescence and functional assessment.

The spectrofluorimetric perfusion system was first validated by monitoring NAD(P)H and flavoprotein autofluorescence during ischemia and reperfusion (Figure 1D and 1E). As expected, NADPH (λ_EX 340 nm, λ_EM 460 nm) autofluorescence immediately rose upon ischemia, in agreement with previous reports. ³² , ⁴⁴ , ⁴⁹ , ⁵⁰ Likewise, flavoprotein fluorescence (λ_Ex 460 nm, λ_Em 520 nm) decreased upon ischemia. ³² , ⁵⁰ , ⁵¹ Both parameters returned to baseline levels immediately upon reperfusion.

NAD(P)H autofluorescence in the heart is thought to mainly represent mitochondrial NADH, the substrate for Cx‐I, which accumulates during ischemia because of a highly reduced respiratory chain. ⁵⁰ The mitochondrial ATP synthase is thought to operate in reverse during ischemia to maintain ΔΨ_m, ⁵² , ⁵³ and this also contributes to feedback inhibition of Cx‐I, resulting in high NADH. ¹¹ Notably, a decrease in NAD(P)H signal was observed approximately 12 minutes into ischemia, concurrent with the onset of ischemic hypercontracture (not shown). The latter is thought to indicate the onset of an energetic crisis (no ATP to relax contractile machinery). ⁵⁴

mitoSOX Fluorescence

Having validated the cardiac fluorescence system, we next sought to use it for measurement of ROS generation during IR. Figure 2A shows cardiac functional measurements throughout IR (heart rate×pressure product), whereas Figure 2B shows the corrected mitoSOX fluorescent readout over the same period.

Figure 2. Cardiac function vs mitoSOX fluorescence during ischemia–reperfusion (IR).

A, Cardiac functional measurements. Graph shows heart rate×left ventricular developed pressure (ie, rate pressure product) for the period of 25 minutes. Ischemia is indicated. Data are mean±SEM, N=16. B, Corrected mitoSOX fluorescence (510/580 nm) during IR. Fluorescent data were processed as described in the Methods section and as shown in Figures S1 through S5, to correct for changes in absorbance of endogenous cytochromes at the fluorescence wavelengths. Data are mean±SEM, N=6. MitoSOX fluorescent measurements were not performed for the entire perfusion time for all hearts in (A) (eg, some hearts were used for controls or other measurements), hence the different N between panels.

To account for changes in the absorbance of endogenous chromophores during IR, or probe distribution, additional hearts were perfused with oxidized mitoSOX or with no probe at all. No significant change in the fluorescent signal was observed in hearts without mitoSOX (Figure S1A). However, delivery of oxidized mitoSOX resulted in a rapid increase in the fluorescent signal during dye loading (Figure S1B). At the onset of ischemia, an additional signal increase was seen. This is expected, because it has been shown that cardiac tissue absorbance at 510 and 580 nm decreases during ischemia. ⁵⁵ Upon reperfusion, after a small increase, the oxidized mitoSOX signal steadily declined for the remainder of the experiment. This is possibly because of loss of the dye from mitochondria as a result of changes in ΔΨ_m as mitochondrial integrity becomes compromised.

Raw mitoSOX fluorescence traces (Figure S1C) also increased slightly during loading, but not to the same extent as the fully oxidized probe. A sharp signal increase at the onset of ischemia may represent a burst of ROS generation as the terminal respiratory chain becomes inhibited. Upon reperfusion, a sustained signal increase was observed for ≈10 minutes, followed by a decline. Using the oxidized mitoSOX data to correct the raw mitoSOX data, Figure 2B (Figure S1D) shows the signal resulting from oxidation of the probe during IR. Upon reperfusion, a sustained increase in the redox‐dependent mitoSOX signal was seen, and this is consistent with the concept that a burst of mitochondrial ROS generation occurs during the first minutes of reperfusion. ¹ , ⁵⁶ , ⁵⁷ , ⁵⁸

It has been posited that reverse electron transport at Cx‐I is the primary source of ROS during reperfusion, and recently, a novel series of inhibitors that target ROS generation at the ubiquinone (Q) binding site of Cx‐I (termed S1QELs) were shown to elicit cardioprotection against IR injury. ⁵⁹ Succinate levels return to baseline within the first 5 minutes of reperfusion, ² and accordingly, examining mitoSOX fluorescence during this time period revealed that the immediate signal increase at the onset of reperfusion in control hearts was suppressed in hearts treated with a S1QEL (Figure 3A through 3C). These data suggest that the ROS signal detected by mitoSOX during the first minutes of reperfusion originates from Cx‐I reverse electron transport. However, by 2 minutes, the rate of signal increase in S1QEL hearts had returned to that seen in control hearts, potentially indicating a role for other sources of ROS. ¹⁷ Although no effect of S1QEL on flavoprotein fluorescence was observed (Figure 3E), S1QEL did cause a slight detriment in the elevation of NAD(P)H fluorescence at the start of ischemia (37±4% with S1QEL versus 53±5% in controls, P=0.042). However, it also blunted the NAD(P)H response to ischemic hypercontracture (Figure 3D). The combination of these effects was such that the drop in NAD(P)H signal at the onset of reperfusion was not significantly different between S1QEL versus control (30±4% versus 36±4%, respectively), suggesting that consumption of NADH by Cx‐I during early reperfusion was not impacted by the compound.

Figure 3. Impact of suppressor of site I_Q electron leak (S1QEL) on ischemia–reperfusion (IR) injury dynamics.

Hearts were treated with the mitochondrial complex I Q‐site reactive oxygen species generation inhibitor S1QEL1.1 for 5 minutes pre‐ and postischemia (Figure 1 schematic). A, Control mitoSOX fluorescent data during the first 5 minutes of reperfusion (from Figure 2B) for comparative purposes. B, mitoSOX fluorescent data from S1QEL‐treated hearts. Gray traces show individual data, with averages shown as a bold line. C, Calculated slopes from the first minute of reperfusion, for the data in (A and B). For each condition, individual data points are shown on the left, with mean±SEM on the right (N for each condition can be seen from the number of data points). P values (ANOVA followed by unpaired Student t test) for differences between groups are denoted. D, Nicotinamide adenine dinucleotide phosphate (NAD(P)H autofluorescence (340/460 nm) during IR. Traces in gray are individual hearts, with the average (N=5) shown in green. E, Flavoprotein autofluorescence (462/520 nm) during ischemia and reperfusion. Traces in gray are individual hearts, with the average (N=5) shown in green. D and E, Average data from control hearts (from Figure 1) are shown in black for comparative purposes. Fluorescence data are normalized to preischemic levels. Ctrl. indicates control.

Preventing Succinate Efflux Exacerbates Mitochondrial ROS at Reperfusion

Upon reperfusion of ischemic heart, one‐third of accumulated succinate is oxidized by Cx‐II, driving ROS generation. ² , ¹⁰ Consistent with this, the Cx‐II competitive inhibitor malonate is reported to be cardioprotective when delivered at reperfusion. ⁶ , ⁷ , ⁶⁰ The remaining two‐thirds of succinate accumulated during ischemia is released from tissue upon reperfusion, and the pathway for this release was recently elucidated as MCT1. ² , ¹⁰ To test the impact of MCT1 inhibition on ROS generation at reperfusion, hearts were infused peri‐ischemically with the MCT1 inhibitor AR, ⁶¹ which resulted in a significant decrease in succinate release into the postcardiac effluent during the first 3 minutes of reperfusion (Figure S3). As expected, lactate efflux was also significantly diminished, confirming MCT1 inhibition.

Figure 4C and 4F show that AR resulted in a significantly greater rate of ROS generation during the first minute of reperfusion compared with control. To investigate the requirement for Cx‐II in the additional AR‐induced mitoSOX signal, the competitive Cx‐II inhibitor DMM (5 mmol/L) was used. Surprisingly, DMM alone at this concentration did not significantly impact the mitoSOX signal, and it also did not significantly blunt the additional signal induced by AR (Figure 4B, 4D, and 4F). We hypothesized that because malonate is a competitive Cx‐II inhibitor, it may not be able to outcompete the additional succinate present in cells caused by MCT1 inhibition. Supporting this hypothesis, tandem administration of a higher dose of DMM (10 mmol/L) was capable of blocking the elevated mitoSOX signal elicited by AR, returning it to control levels (Figure 4E and 4F). Furthermore, the potent Cx‐II inhibitor AA5 (100 nmol/L) was also effective in blocking the additional mitoSOX signal induced by AR (Figure S4A and S4B). Overall, these data suggest that MCT1 inhibition enhances mitochondrial ROS generation in the first minute of reperfusion, in a manner that can be blocked by inhibitors of Cx‐II.

Figure 4. Impact of AR‐C155858 (AR) and mitochondrial complex II inhibitors on mitoSOX fluorescence at reperfusion.

A through E, Corrected mitoSOX traces 5 minutes pre‐ and postreperfusion for the labeled conditions. Gray traces are individual data, with averages shown as bold colored lines (see Figure 1 for color scheme). F, Calculated slopes from first minute of reperfusion for the data in (A through E). For each condition, individual data points are shown on the left, with mean±SEM on the right (N for each condition can be seen from the number of data points). P values (ANOVA followed by Tukey honest significant difference test) for differences between groups are denoted. Ctrl. indicates control; DMM, dimethyl malonate; and N.S., not significant.

Blocking Succinate Release Via MCT1 Worsens IR Injury

In agreement with the role of succinate‐derived ROS as a driver of post‐IR pathology, DMM alone improved and AR worsened IR injury (Figure 5A through 5C), although these effects were only significant in terms of infarct size while failing to reach significance (Tukey honest significant difference test) for functional recovery. In addition, tandem administration of low‐dose (5 mmol/L) or high‐dose (10 mmol/L) DMM reversed the impact of AR. Furthermore, administration of the potent Cx‐II inhibitor AA5 also blocked the impact of AR on functional recovery and infarct (Figure S4C through S4E) and was more protective than AA5 alone. ² These effects of MCT1 inhibition and Cx‐II inhibition were more significant for myocardial infarction (Figure 5C) than for functional recovery (Figure 5B), but all trended in the same direction, similar to mitoSOX data (Figure 4) (ie, interventions that increased mitoSOX signal at reperfusion worsened IR injury, and those that decreased it improved IR injury). Overall, a correlation was observed between the effect of interventions on mitoSOX at reperfusion and on infarct size and functional recovery, as illustrated in Figure 6.

Figure 5. Impact of AR‐C155858 (AR) and mitochondrial complex II inhibitors on outcomes of ischemia–reperfusion (IR) injury.

A, Cardiac functional measurements during IR. Graphs show heart rate×left ventricular developed pressure (ie, rate pressure product [RPP]) for each of the conditions examined for the period of 25 minutes ischemia is indicated. Data are mean±SEM. B, Quantitation of percent functional recovery (ie, cardiac function at 60 minutes of reperfusion as a percentage of that immediately before ischemia [−25 minutes time point]). For each condition, individual data points are shown on the left, with mean±SEM shown on the right (N for each condition can be seen from number of data points). No significant differences were observed between groups by ANOVA followed by Tukey honest significant difference (HSD) test, although P values from ANOVA followed by unpaired Student t test did denote differences between groups as shown above the data. C, Myocardial infarct size for each condition. Images above the graph show representative triphenyltetrazolium chloride–stained heart cross‐sections, with pseudocolored mask used for planimetry below. Graph shows data for each condition, with individual data points on the left and mean±SEM on the right (N for each condition can be seen from number of data points). P values (ANOVA followed by Tukey HSD test) for differences between groups are denoted above the data. Ctrl. indicates control; DMM, dimethyl malonate; and N.S., not significant.

Figure 6. Correlations between mitoSOX fluorescence during reperfusion and ischemia–reperfusion outcomes.

Graphs show correlation between mitoSOX slope data and (A) infarct size or (B) functional recovery. Each data point is the mean±SEM for a condition (see color scheme in Figure 1). mitoSOX data are from Figure 3F and Figure S5B. Infarct data are from Figure 4C and Figure S4E. Functional data are from Figure 4B and Figure S4C. Linear curve fits are shown, with correlation coefficient (r ²) listed alongside. AR indicates AR‐C155858; Ctrl., control; and DMM, dimethyl malonate.

Timing of Mitochondrial ROS at Reperfusion Versus PT Pore Opening

Although it is largely accepted that ROS can trigger PT pore opening, it has also been proposed that during reperfusion injury, the PT pore itself may drive ROS generation. ²³ , ³² To test this, we measured the effect of the PT pore inhibitor CsA on mitoSOX fluorescence during reperfusion. CsA was chosen at 0.8 μmol/L, because this was previously shown to elicit protection in mouse hearts. ³⁸ As shown in Figure 7B and 7C, CsA had no impact on mitoSOX fluorescence during reperfusion. To confirm that PT pore opening did occur, hearts were loaded with the ΔΨ_m indicator TMRE. Upon reperfusion, an immediate rise in TMRE fluorescence was observed. Subsequently, in control hearts the TMRE signal declined from approximately 5 minutes into reperfusion, whereas in CsA‐treated hearts, the signal was sustained (Figure 7D and 7E). We thus infer that PT pore opening occurs no sooner than 5 minutes into reperfusion. Despite a small blip in the mitoSOX signal in control hearts at approximately 6.5 minutes, no substantial difference was seen between control and CsA hearts at this time point, concurrent with divergence of the TMRE traces, thus suggesting no secondary ROS burst because of PT pore opening.

Figure 7. Temporal relationship between mitoSOX and permeability transition (PT) pore opening during ischemia–reperfusion.

A, Schematic showing perfusion conditions. Where indicated, mitoSOX (reactive oxygen species indicator), tetramethylrhodamine ethyl ester (TMRE) (mitochondrial membrane potential indicator) and/or cyclosporin A (CsA) (PT pore inhibitor) were administered at the listed concentrations. B and C, mitoSOX signal during early reperfusion under control (Ctrl.) or CsA condition. The last 5 minutes of the ischemic period is indicated. Gray traces show individual data, with averages shown as bold lines. Inset to (C) shows averages for control and CsA superimposed. D and E, Normalized TMRE fluorescence during early reperfusion under control or CsA condition. The last 5 minutes of the ischemic period is indicated. Gray traces show individual data, with averages shown as bold lines. Inset to (C) shows averages for control and CsA superimposed. Red arrow indicates the point at which traces diverge 5 minutes into reperfusion.

Finally, as expected from its impact on mitoSOX fluorescence during early reperfusion (Figure 4), inhibition of MCT1 led to an accelerated loss of the TMRE signal during reperfusion, indicating faster opening of the PT pore (Figure S5).

Discussion

Cardiac IR is a complex pathology, with numerous links between early and late events in the development of myocardial infarction. Shortly after reperfusion, a burst of ROS generation and mitochondrial Ca²⁺ overload both trigger opening of the mitochondrial PT pore, a key event in necrotic cell death of cardiomyocytes. Following this, inflammatory cells (macrophages, neutrophils) are recruited to the heart, where they mediate responses that lead to cardiac remodeling, fibrosis, and eventual development of hypertrophy and heart failure. ⁶² , ⁶³ , ⁶⁴

Accumulation of succinate in hypoxia/ischemia is highly conserved, ³ , ⁴ but this succinate drives ROS generation upon tissue reperfusion, and this has led to a consensus that intracellular succinate plays a detrimental role in reperfusion injury. ¹ , ² , ¹⁶ , ⁶⁵ , ⁶⁶ In contrast, succinate release from cells via MCT1, ² , ¹⁰ , ²³ along with the recent identification of a succinate receptor (GPR91/SUCNR1 ²⁵ ), have led to the notion that extracellular succinate may also play a role in IR. ⁶ , ¹⁰ In this regard, findings from an in vivo model of cardiac IR injury demonstrated that blocking succinate release via MCT1 was cardioprotective, likely because of inhibiting immune system activation. ⁶ , ¹⁰ However, herein our data reveal that acutely blocking MCT1 in a perfused heart system (where there are no inflammatory cells) leads to greater ROS generation and the worsening of IR injury. Reconciling these findings, it is possible that MCT1 inhibition is detrimental in the acute setting at the level of cardiomyocytes, but this is balanced in vivo by a longer‐term effect of MCT1 inhibition, possibly involving receptor‐mediated succinate effects on other cell types, including inflammatory cells. Together, these studies highlight that the therapeutic targeting of MCT1 may require careful timing and titration to balance detrimental versus beneficial effects of blocking succinate release.

Inhibiting MCT1 in the heart lowered succinate release into the effluent by approximately 30%. Assuming this succinate was retained in cells and available for oxidation by Cx‐II, it is therefore not surprising that MCT1 inhibition also led to enhanced ROS generation. Furthermore, although 5 mmol/L of the competitive Cx‐II inhibitor malonate was ineffective at blocking this additional ROS, doubling its concentration to 10 mmol/L effectively overrode the impact of AR. Unfortunately, because of the physical nature of dimethylmalonate (liquid) and our drug infusion system, it was not possible to test additional even higher doses of DMM. Instead, we tested the potent Cx‐II inhibitor AA5 ( IC₅₀10 nmol/L ⁶⁷ ), and found that it was able to completely abrogate the additional ROS induced by AR, and this resulted in cardioprotection to a level greater than baseline control IR injury (Figure S4). The combination of AR and AA5 may be considered optimal from a therapeutic perspective, keeping succinate inside cells to prevent its signaling effects and preventing its oxidation to generate ROS.

Although there is a general consensus that mitochondrial ROS generation during early reperfusion is an upstream event that triggers PT pore opening, it has also been suggested that pore opening itself is the main driver of ROS generation in early reperfusion. ²³ , ³² Our results (Figure 7) suggest that the burst of ROS at reperfusion is independent of the PT pore, because evidence for pore opening (ie, a CsA‐sensitive loss of ΔΨ_m) was not observed until at least 5 minutes into reperfusion, by which time a large amount of ROS generation had already occurred.

It should not go unmentioned that Ca²⁺ is also considered to be a major trigger for PT pore opening, ⁶⁸ , ⁶⁹ with both Ca²⁺ and ROS thought to potentiate each other's effects at promoting pore formation. ⁷⁰ Thus, future experiments should be directed at using fluorescent Ca²⁺ probes in this or similar perfusion systems to understand in situ Ca²⁺ kinetics and how they relate to ROS and pore‐opening kinetics in early reperfusion.

Several caveats on the use of mitoSOX as a probe for mitochondrial ROS should be addressed. First, it is known the probe can be oxidized by reactants other than ROS, although notably, 1‐electron oxidations result in nonfluorescent products that would be undetectable in our measurement system. ⁷¹ Furthermore, the mitoSOX signal increase in early reperfusion was inhibited by S1QEL, a specific inhibitor of ROS generation at Cx‐I, ⁵⁹ so we consider any contribution from other poorly characterized oxidation sources to be minimal. Second, the mitoSOX signal is impacted both by its distribution between extracellular, cytosolic, and mitochondrial compartments, ³⁷ and by the primary and secondary filter effects from endogenous chromophores ⁴⁷ , ⁵⁵ (see Methods section). However, our use of fully oxidized mitoSOX as a control (Figure S1) ensures that any changes in the mitoSOX signal we observed originated only from in situ probe oxidation, and not from redistribution of the probe or changes in the absorbance of myoglobin or cytochromes.

Although mitoSOX does confer limitations on the assignment of the signal to a particular ROS, similar issues of probe specificity are also broadly applicable to genetically encoded biosensors. ⁷² , ⁷³ In addition, more precise analytical methods, such as liquid‐chromatography mass spectrometry (LC–MS) separation of mitoSOX oxidation products, ³⁷ , ⁷⁴ require time‐consuming isolation steps that may release components that further oxidize the probe during isolation, and therefore may not be readily compatible with the rapid kinetics of the events observed herein. The development of more precise genetically encoded and non–ΔΨ_m‐dependent mitochondrial ROS probes may therefore be useful in future studies.

Overall, the findings herein demonstrate that blocking MCT1 in cardiac IR leads to inhibition of succinate release, which worsens IR injury because of enhanced mitochondrial ROS generation. Concurrent inhibition of Cx‐II abrogates these effects, highlighting the importance of succinate oxidation at Cx‐II in the pathology of IR injury. Furthermore, these events appear to lie temporally upstream of PT pore opening. Future experiments could explore the role of local succinate signaling in the heart to determine if SUCNR1 may modulate responses to IR in a manner independent of inflammatory cells, providing further insight on the delicate balance of succinate dynamics in IR injury.

Sources of Funding

This work was supported by a grant from the National Institutes of Health (R01‐HL071158). A. S. Milliken is funded by an American Heart Association Predoctoral Fellowship (number 21PRE829767) and formerly National Institutes of Health T32‐GM068411.

Disclosures

None.

Supporting information

Figures S1–S5