A novel class of cardioprotective small-molecule PTP inhibitors

Abstract

Ischemia/reperfusion (I/R) injury is mediated in large part by opening of the mitochondrial permeability transition pore (PTP). Consequently, inhibitors of the PTP hold great promise for the treatment of a variety of cardiovascular disorders. At present, PTP inhibition is obtained only through the use of drugs (e.g. cyclosporine A, CsA) targeting cyclophilin D (CyPD) which is a key modulator, but not a structural component of the PTP. This limitation might explain controversial findings in clinical studies. Therefore, we investigated the protective effects against I/R injury of small-molecule inhibitors of the PTP (63 and TR002) that do not target CyPD. Both compounds exhibited a dose-dependent inhibition of PTP opening in isolated mitochondria and were more potent than CsA. Notably, PTP inhibition was observed also in mitochondria devoid of CyPD. Compounds 63 and TR002 prevented PTP opening and mitochondrial depolarization induced by Ca²⁺ overload and by reactive oxygen species in neonatal rat ventricular myocytes (NRVMs). Remarkably, both compounds prevented cell death, contractile dysfunction and sarcomeric derangement induced by anoxia/reoxygenation injury in NRVMs at sub-micromolar concentrations, and were more potent than CsA. Cardioprotection was observed also in adult mouse ventricular myocytes and human iPSc-derived cardiomyocytes, as well as ex vivo in perfused hearts. Thus, this study demonstrates that 63 and TR002 represent novel cardioprotective agents that inhibit PTP opening independent of CyPD targeting.

1. Introduction

Cardiovascular diseases remain the leading cause of death worldwide, mainly due to acute myocardial infarction (MI) [1]. No therapeutic strategy is yet available to both reduce infarct size following ischemia and reperfusion, and to mitigate the risk of developing heart failure. Indeed, ischemia/reperfusion (I/R) injury is thought to play a major role in cell death after MI. Therefore, targeting pathways that lead to I/R injury are potential clinical targets that could significantly curtail the size and severity of myocardial injury following acute MI.

One causative event of I/R injury is opening of the permeability transition pore (PTP). The PTP is a regulated, transmembrane pore of the inner mitochondrial membrane which allows permeation of solutes up to 1.5 kDa. Classically, PTP activation can be triggered by Ca²⁺ overload and reactive oxygen species (ROS), while PTP opening can be blocked by other divalent cations, ADP and low matrix pH [2, 3]. PTP opening leads to mitochondrial depolarization, swelling, and respiratory inhibition which initiates cell death in cardiac cells [4, 5]. Many studies have explored the PTP as a potential drug target in cardioprotection [6–9]. These studies focused on the use of cyclosporine A (CsA), a cyclophilin D (CyPD)-binding cyclic peptide [10], as an inhibitor of the PTP because CyPD is a known regulator of the PTP. Indeed, CsA reduces infarct size in multiple preclinical models of cardiovascular disease [6, 11]. However, phase III clinical trials demonstrated that CsA provided no benefit for patients presenting with acute MI [12, 13]. While it is not clear why CsA failed in phase III clinical trials for acute MI, it is worth noting that CsA is not a direct inhibitor of the PTP. Thus, there is a critical need for selective PTP inhibitors in order to determine the therapeutic potential of targeting the PTP in acute MI and to elucidate the mechanisms by which the PTP contributes to I/R injury.

Recently, we have used a combination of high throughput screening and follow-up medicinal chemistry to discover novel small-molecule inhibitors of the PTP [14–16]. The most potent inhibitors are based on an isoxazole or triazole scaffold. These compounds prevented PTP opening in mouse liver mitochondria in a CyPD-independent manner, suggesting that they act via a novel mechanism—perhaps by directly targeting the PTP. The triazole-based variants showed improved plasma stability compared to the isoxazole-based inhibitors, and were particularly efficacious in a zebrafish model of collagen VI muscular dystrophy, which is characterized by PTP deregulation and chronic mitochondrial dysfunction [16]. Here, we tested the potential cardioprotective effects of two compounds in this series, the isoxazole 63 [14] and the structurally-related triazole analogue TR002 [16]. Our results indicate that targeting the PTP with small-molecule inhibitors independent of CyPD can protect cardiac mitochondria, cells, and tissue from I/R-induced cardiac injury.

2. Results

2.1. 63 and TR002 prevent PTP opening in isolated cardiac mitochondria

Small-molecule isoxazole (63) and triazole (TR002) compounds (Fig. 1A) are novel, potent inhibitors of the PTP that were identified and optimized following the screen of NIH MLPCN collection of 362,000 compounds [14] using mouse liver mitochondria. We assessed the ability of these inhibitors to desensitize the PTP in isolated heart mitochondria using a Ca²⁺ retention capacity (CRC) assay. CRC measures the Ca²⁺ load needed for PTP opening, which is marked by a sudden release of the previously accumulated cation. Mitochondrial suspensions were incubated with 1 μM 63, TR002 or CsA and subsequently challenged with a train of 10 μM Ca²⁺ pulses. Compared to CsA, we found that 63 and TR002 more effectively increased the threshold Ca²⁺ required for PTP opening (Fig. 1B). The CRC ratio was increased in a concentration-dependent manner for all three compounds, however, 63 was the most potent, followed by TR002 (Fig. 1C). At 10 μM TR002, we observed a decrease in the CRC ratio, suggesting that at this concentration TR002 may exhibit adverse effects (Fig. 1C); therefore, in all subsequent experiments we used concentrations of 1 μM or less for TR002. Taken together, these results demonstrate that 63 and TR002 were more effective inhibitors of the PTP than CsA in isolated heart mitochondria.

Figure 1. Effects of compounds 63 and TR002 on PTP opening in isolated cardiac mitochondria.

A) Chemical structures of the compounds 63 and TR002. B) PTP opening in isolated rat heart mitochondria treated with DMSO (Black line), 1 μM 63 (Green line), 1 μM TR002 (Blue line) or 1 μM CsA (Red line) as monitored by CRC assay. C) Data were quantified at different concentrations of the compounds (0.01 – 10 μM). *p < 0.05, **p < 0.01, ***p < 0.001 vs DMSO. CRC ratio is compound-to-DMSO treated mitochondria CRC ratios. Isolated heart mitochondria from D, E) wild type or F, G) CyPD-null mice were analyzed for CRC (D and F) or swelling (E and G) assays in presence of DMSO as control (Black line), 1 μM 63 (Green line), 1 μM TR002 (Blue line) or 1 μM CsA (Red line). One hundred μg of mitochondria were challenged with a train of 5 μM CaCl₂ pulses for CRC experiments or with a single 50 μM CaCl₂ pulse (arrow) for swelling.

Data are expressed as mean ± SEM. n=6.

We previously showed that isoxazole and triazole-based PTP inhibitors structurally related to 63 and TR002 inhibited PTP opening in liver mitochondria and HEK293 cells in a CyPD-independent manner [16]. To confirm that 63 and TR002 function independently of CyPD in heart mitochondria, we conducted CRC and mitochondrial swelling assays on isolated heart mitochondria from wild type and Ppif^−/− mice (Fig. 1D-G). We observed that 63, TR002, and CsA protected against PTP opening induced with a train of 5 μM Ca²⁺ pulses in wild type heart mitochondria (Fig. 1D-E). By contrast, 63 and TR002 provided additional protection against PTP opening in Ppif ^−/− heart mitochondria (Fig. 1F-G). These results were confirmed in liver mitochondria challenged with large boluses of Ca²⁺ (i.e. 100 μM) (Supplementary Fig. 1A-D). These results demonstrate that 63 and TR002 inhibit PTP independently of CyPD.

2.2. Compounds 63 and TR002 prevent PTP opening and mitochondrial membrane depolarization induced by calcium and ROS

We next investigated the effects of 63 and TR002 in intact cardiomyocytes. In these experiments, opening of the PTP was assessed in neonatal rat ventricular myocytes (NRVMs) by monitoring the decrease in calcein fluorescence [17]. NRVMs were pre-treated with DMSO, CsA, 63 or TR002, and PTP opening was induced by the Ca²⁺ ionophore calcimycin. Treatment with either 0.5 μM 63 or 0.5 μM TR002 delayed PTP opening (Fig. 2A). Notably, after 180 s of calcimycin treatment the decrease in calcein fluorescence was significantly lower compared to the DMSO control (DMSO: 71.9 ± 6.5 % vs 63: 25.5 ± 6.9 % or TR002: 33.9 ± 11.4 %, Fig. 2B).

Figure 2. Effects of compounds 63 and TR002 on PTP opening and mitochondrial membrane potential.

A) PTP opening monitored by decrease of calcein fluorescence in isolated NRVMs pretreated for 30 min with DMSO (Black line), 0.5 μM 63 (Red line), 0.5 μM TR002 (Orange line) or 1 μM CsA (Yellow line). a: 5 μM calcimycin. B) Data were quantified at different time points after treatment with calcimycin. *p < 0.05, §p < 0.01 vs DMSO 120s; #p < 0.05, çp < 0.01 vs DMSO 180s. C-D) Mitochondrial membrane potential (ΔΨm) monitored by TMRM fluorescence in isolated NRVMs following incubation for 1 h with DMSO as control (C) or 1 μM MitoPQ (D). Cells were pretreated for 30 min with DMSO (Black line), 0.5 μM 63 (Red line), 0.5 μM TR002 (Orange line) or 1 μM CsA (Yellow line). a: 5 μM calcimycin, b: 4 μM FCCP. E) Data were quantified after 10 minutes from the beginning of 129369MC 129369MC 129369MC 129369MC experiment and before FCCP injection. **p < 0.01 vs DMSO. Approximately 30 cells were analyzed per condition in each experiment and all the experiments were performed at least three times using three different animal preparations.

Data are expressed as mean ± SEM.

PTP inhibition was reported to prevent mitochondrial depolarization and cardiomyocyte derangements induced by I/R [18, 19]. To investigate whether 63 and TR002 could prevent such mitochondrial dysfunction, we monitored ΔΨm using tetramethylrhodamine (TMRM) fluorescence. NRVMs were pre-treated with 63, TR002, or CsA. Since CsA is also an inhibitor of the multidrug resistance P-glycoprotein (MDR), which may alter TMRM distribution, cells were also pre-treated with cyclosporine H which inhibits the MDR but does not affect PTP opening [20]. NRVMs treated with calcimycin displayed a decrease in ΔΨm (20%) that was prevented by treatment with both 63 and TR002 (Fig. 2C-E). Then, to investigate the increased propensity to PTP opening caused by oxidative stress, we used MitoParaquat (MitoPQ), which induces a primary increase in mitochondrial ROS levels [21]. NRVMs were pre-treated with 1 μM MitoPQ and then exposed to Ca²⁺ overload by means of calcimycin in presence or absence of 63, TR002 or CsA. NRVMs treated with calcimycin and MitoPQ displayed a dramatic drop in ΔΨm (50%) that was completely prevented by both 63 and TR002 (Fig. 2D-E). Taken together, these results demonstrate that compounds 63 and TR002 prevent PTP opening and the associated mitochondrial dysfunction induced by either Ca²⁺ or by the combination of Ca²⁺ overload and a selective increase in mitochondrial ROS levels.

2.3. 63 and TR002 prevent Ca²⁺ dysregulation and sarcomeric impairment in NRVMs following anoxia/reoxygenation

We next investigated if inhibition of PTP opening by 63 and TR002 could prevent alterations of intracellular [Ca²⁺] homeostasis induced by anoxia and reoxygenation (A/R) in NRVMs. NRVMs were subjected to either 3h or 6h of anoxia followed by reoxygenation, and intracellular [Ca²⁺] transients were monitored by means of Fluo-4 fluorescence during the reoxygenation phase. Upon reoxygenation after 3h of anoxia the number of beating cells was significantly higher in the cells treated with 63 or TR002 compared to the control (DMSO: 22.8 ± 11.3 % vs 63: 83.3 ± 12.9 % or TR002: 80.0 ± 25.8 %, Fig. 3F). However, if one compares the beating cells, NRVMs treated with either 63 or TR002 did not display significant differences in Ca²⁺ transients compared to the anoxic control. Indeed, although NRVMs treated with the compounds exhibit an oscillatory pattern similar to the normoxic cells (Fig. 3A, Supplementary Fig. 2A), there were no differences in transient amplitude, in transient frequency and in response to caffeine (Fig. 3C-D-E). Interestingly, the ability of NRVMs to decrease the level of Ca²⁺ in the cytosolic compartment following caffeine addition, which releases Ca²⁺ from sarcoplasmic reticulum (SR), was lower in DMSO-treated cells than in cardiomyocytes treated with either 63 or TR002 (Supplementary Fig. 2B), especially after 6h of anoxia. Moreover, following 6h of anoxia, 63 and TR002 preserved Ca²⁺ transients in NRVMs that display a higher number of spontaneous beating cells compared to the control (DMSO: 0 % vs 63: 44.0 ± 15.9 % or TR002: 37.5 ± 19.8 %, Fig. 3F). Indeed, cells treated with either compound did not show significant differences in transient amplitude, frequency and response to caffeine than normoxic control (Fig. 3C-D-E), while cells treated with DMSO were unexcitable (Fig. 3B). Moreover, NRVMs treated with both 63 and TR002 displayed a preserved response to caffeine (DMSO: 35.2 ± 12.7 % vs 63: 100.0 % or TR002: 93.3 ± 9.4 %, Fig. 3F).

Figure 3. Effects of compounds 63 and TR002 on cytosolic [Ca²⁺] homeostasis in NRVMs exposed to A/R.

A-E) Cytosolic [Ca²⁺] homeostasis monitored by Fluo-4 AM in isolated NRVMs treated for 30 min with DMSO, 0.5 μM 63 or 0.5 μM TR002. Cells were exposed either to 3h and 6h of anoxia and the Ca²⁺ transients were evaluated during reoxygenation. A-B) Comparison between cytosolic Ca²⁺ oscillatory patterns in reoxygenation after (A) 3h or (B) 6h of anoxia. C) Transients amplitude average. D) Transients frequency average. E) Average of caffeine peak amplitude. F) Comparison between beating cells and response to caffeine in NRVMs treated with DMSO, 0.5 μM 63 or 0.5 μM TR002 after either 3h and 6h of anoxia. C-E) #p < 0.05, ###p < 0.001 vs Anoxia 6h DMSO. F) *p < 0.05 vs DMSO 3h Anoxia; ## p < 0.01 vs DMSO 6h Anoxia. Approximately 20/30 cells were analyzed per condition in each experiment and all the experiments were performed at least three times using three different animal preparations.

Data are expressed as mean ± SEM.

Subsequently, we investigated whether treatment with 63 or TR002 prevents A/R-induced derangements of sarcomeric organization in cardiomyocytes. NRVMs were exposed to 6h of anoxia followed by 30 min of reoxygenation in the presence or absence of CsA, 63 and TR002, and the distribution of the α-sarcomeric actinin was evaluated by immunocytochemistry. After 6h of anoxia, NRVMs displayed disrupted sarcomeric structure. In contrast, NRVMs pre-treated with all three compounds displayed a well aligned sarcomeric pattern similar to that observed in normoxic NRVMs (Fig. 4A, Supplementary Fig. 2C). Moreover, the impairment of the mitochondrial morphology observed in DMSO-treated cells was ameliorated in NRVMs treated with the three compounds upon 6h of anoxic injury (Fig. 4B, Supplementary Fig. 2D).

Figure 4. Effects of compounds 63 and TR002 on sarcomeres arrangement and mitochondrial shape.

Sarcomeres arrangement and B) mitochondrial morphology monitored by confocal microscopy in NRVMs exposed to 6h of anoxia and pretreated for 30 min with DMSO, 0.5 μM 63, 0.5 μM TR002 or 1 μM CsA. Confocal images were processed for double immunofluorescent labeling of α-sarcomeric actinin (green) and TOM20 (red). A) The fluorescence intensity of the α-sarcomeric actinin was plotted and reported on the graph to show the effect of the treatments on the distribution of the sarcomeres within the cell. Both representative images (A and B upper row) and regions of interests (A and B lower row) are provided.

Approximately 20/30 cells were analyzed per condition in each experiment and all the experiments were performed at least three times using three different animal preparations.

Since changes in mitochondrial dynamics could be upstream of PTP opening [22, 23], expression and distribution of proteins involved in mitochondrial fusion and fission were evaluated in normoxic NRVMs. To this aim, cardiomyocytes were treated in normoxia for 6 h with 1 μM 63 or 1 μM TR002, and their effect were compared with CsA that is reported to increase DRP1 phosphorylation by inhibiting calcineurin [24]. Both compounds did not modify DRP1 phosphorylation, that was markedly increased by CsA (Supplementary Fig. 3A). On the contrary, MFN2 expression levels was not altered by any treatment (Supplementary Fig. 3B). In addition, 63 and TR002 did not modify mitochondrial morphology as evaluated by immunocytochemistry (Supplementary Fig. 3C).

Taken together, our data demonstrate that treatment with either 63 or TR002, preserves mitochondrial function, intracellular [Ca²⁺] transients and cardiomyocyte contractility in NRVMs following A/R. This protection from A/R injury is associated with the preservation of the mitochondrial morphology and the sarcomeric array in cardiomyocytes. Moreover, the cardioprotective efficacy of both compounds is not contributed by changes in mitochondrial dynamics as might be the case with CsA.

2.4. Compounds 63 and TR002 reduce cardiomyocyte death following anoxia/reoxygenation and protect hearts from ischemic damage

Since PTP opening is one of the main causes of cell death occurring during post-ischemic reperfusion (or post-anoxic reoxygenation) [25], we next investigated whether 63 and TR002 could decrease the extent of A/R-induced cardiomyocyte death.

Cardiomyocytes (i.e. NRVMs, adult mouse ventricular myocytes (AMVMs), human iPSc-derived cardiomyocytes (hiPSc-CMs)) were pre-treated with increasing concentrations of 63 or TR002 (Supplementary Fig. 4A). Similar to CsA, 63 and TR002 at concentrations as low as 0.1 μM protected cardiomyocytes against A/R-induced loss of viability as shown by a decrease in lactate dehydrogenase (LDH) release (Fig. 5). In NRVMs, 63 was more protective than TR002 (Fig. 5A). In AMVMs, higher concentrations (i.e. ≥ 0.5 μM) of 63 and TR002 were required for decreasing cell death (Fig. 5B). Similar to results obtained in NRVMs, all concentrations of 63 and TR002 decreased susceptibility to cell death in hiPSc-CMs (Fig. 5C).

Figure 5. Effects of compounds 63 and TR002 on cardiomyocytes viability.

Cell death measured by LDH release from A) isolated NRVMs, B) isolated AMVMs and C) hiPSc-CMs treated for 30 minutes with different concentrations of either 63 or TR002 or with 1 μM CsA. *p < 0.05, **p < 0.01, *** p< 0.001 vs DMSO Reoxygenation. D) Cell death measured by LDH release in ex vivo hearts perfused in a Langendorff model of I/R, with or without 63, TR002 or CsA. *p < 0.05, **p < 0.01, ***p < 0.001 vs DMSO.

For in vitro experiments (A, B, C), approximately 3/4 wells were analyzed per condition in each experiment and all the experiments were performed at least three times using three different animal preparations or cell differentiations. For ex vivo experiments (D), n = 5 to 13.

Data are expressed as mean ± SEM.

We further examined cardioprotection by 63 or TR002 in isolated mouse hearts perfused at constant flow and subjected to a protocol of I/R. Hearts of 5 month old mice were perfused with different concentrations of 63 or TR002 for 10 minutes and then subjected to 40 minutes of ischemia followed by 15 minutes of reperfusion (Supplementary Fig. 4B). Both 63 and TR002 decreased the release of LDH in a dose-dependent manner up to 0.5 μM (Fig. 5D).

Protective effects were also observed in a different perfusion setup with constant pressure and a latex balloon in the left ventricle for hemodynamic monitoring. The hearts were subjected to 20 minutes of ischemia followed by 90 minutes of reperfusion. Hearts were perfused with 0.5 μM of either 63 or TR002 for 10 minutes prior to ischemia and for the first 10 minutes of reperfusion (Supplementary Fig. 4B). Treatment with 63 resulted in significantly smaller infarct sizes (DMSO: 65.55% vs 63: 44.90%, p < 0.001, Fig. 6A, Supplementary Fig. 4C) and significantly better post-ischemic recovery compared to vehicle-perfused hearts (DMSO: 37.52% vs 63: 54.35%, p<0.001, Fig. 6B). Treatment with TR002 trended toward cardioprotection as measured by a decrease in infarct size (DMSO: 41.45% vs TR002: 33.83%, p=0.120, Fig. 6C) and an increase in rate pressure product (RPP) recovery (DMSO: 45.38% vs TR002: 54.60%, p=0.33, Fig. 6D) but did not reach statistical significance.

Figure 6. Effects of compounds 63 and TR002 on functional recovery and cell death by TTC staining.

A-C) Cell death measured by TTC staining at 90 minutes of reperfusion in ex vivo murine hearts perfused in a Langendorff model with a standard I/R procedure after treatment with A) 63, and C) TR002 compared to DMSO control. B-D) Functional recovery calculated by left ventricle developed pressure (LVDP) x heart rate as assessed by intraventricular balloon after treatment with B) 63, and D) TR002 as compared to DMSO control. *** p < 0.001 vs DMSO. n > 6.

Data are expressed as mean ± SEM.

Taken together, these results indicate that 63 and TR002 reduce the susceptibility to both anoxic and ischemic injury, both in vitro and ex vivo by decreasing the opening of the PTP in cardiomyocytes derived from different species. In general, 63 was more cardioprotective than TR002.

3. Discussion

This study shows that our recently-described small-molecule inhibitors of the PTP, isoxazole 63 and triazole TR002, protect against cardiomyocyte injury caused by A/R in vitro in isolated cardiomyocytes, and by I/R ex vivo in perfused hearts. Notably, the inhibition of PTP opening by 63 and TR002 is independent of CyPD since PTP opening was equally inhibited in mitochondria isolated from WT and Ppif^−/− mice. Besides reducing the extent of I/R-induced loss of viability, 63 and TR002 maintained mitochondrial function, intracellular [Ca²⁺] homeostasis, cardiomyocyte contractility and sarcomeric organization.

Previous studies have documented the protective effect of indirect PTP inhibition via targeting of its regulator CyPD with CsA in animal models [6–9]; however, this protection failed to be translated in a human clinical trial [12, 13]. Recently, new CyPD-independent, small-molecule inhibitors of the PTP were discovered [14–16]. Here, we show that two of these compounds, 63 and TR002, are more potent than CsA in inhibiting PTP opening in both isolated cardiac mitochondria and cardiomyocytes. Protection against A/R-induced cell death by 63 and TR002 was observed in mouse, rat and human cardiac cells, as well as in ex vivo mouse hearts. Besides demonstrating species-independent protective effects, the present findings show the occurrence of protection due to PTP inhibition in human cardiomyocytes. This result represents a relevant pre-requisite for translating our results to the clinics.

The present findings strongly support the critical role of PTP opening in cardiac injury induced by post-ischemic reperfusion. Besides showing improved viability as documented previously [26–28], here we find that inhibition of PTP opening by 63 and TR002 not only maintains mitochondrial function, but also prevents morphological and functional perturbations occurring at the cytosolic level in cardiomyocytes during reperfusion. Therefore, the disruption of mitochondrial bioenergetics induced by PTP opening is likely to be upstream of the loss of cardiomyocyte function and viability caused by I/R, reinforcing the importance of targeting PTP to obtain a robust cardioprotection.

Previous efforts at PTP inhibition as a therapeutic strategy for cardioprotection focused on the indirect regulator CyPD. However, this approach was not successful in clinical studies [29]. The reason why CsA was not cardioprotective is unclear, but could be due to the fact that CyPD is an indirect regulator of the PTP, and that pore opening can occur also in its absence [30]; alternatively, poor pharmacokinetics of CsA could have contributed to the lack of clinical success [29, 30]. In addition, CyPD inhibition has been described to result in detrimental effects, especially exacerbating adverse remodeling of hypertrophied hearts [26, 31]. This could be due to an increase in matrix [Ca²⁺] as a consequence of a decreased occurrence of PTP opening. If this is the case, however, the present compounds could mimic the negative effects of CyPD targeting. A recent study demonstrates that the negative consequence of matrix Ca²⁺ accumulation resulting from the deletion of the mitochondrial Na⁺/Ca²⁺ exchanger is prevented by CyPD deletion [32]. Therefore, rather than causing cardiac injury by increased matrix [Ca²⁺] PTP inhibition appears to prevent cardiac derangements resulting from intra-mitochondrial [Ca²⁺] overload. The negative outcome of pharmacological or genetic targeting of CyPD could be related to the loss of its activity and/or interaction with other proteins. Since CyPD is a prolyl isomerase required for the correct folding of intra-mitochondrial proteins, its absence or inhibition could hamper mitochondrial proteostasis [33]. This negative effect is not likely to be generated by CyPD-independent PTP inhibitors.

In any case, additional studies are needed to investigate whether the prolonged administration of the novel PTP inhibitors is devoid of negative effects.

4. Methods

4.1. Cell culture and treatments

Neonatal rat ventricular myocytes (NRVMs) were isolated from 1–3 day old Wistar rats as described previously [34]. Cardiomyocytes were plated in 0.1% porcine gelatin (Sigma) coated plates at variable density (at least 3 × 10⁵ cells/ml) in MEM supplemented with 10% FBS (Thermo Fisher Scientific), 1% penicillin/streptomycin (Thermo Fisher Scientific), 1% non-essential amino acids (Thermo Fisher Scientific), 1 mM 5-Bromo-2-Deoxyuridine (Sigma). Cells were maintained at 37°C in the presence of 5% CO₂. The medium was changed to MEM supplemented with 1% FBS, 1% penicillin/streptomycin and 1% non-essential amino acids after 24h of plating.

Adult mouse ventricular myocytes (AMVMs) were isolated from the hearts of 12-week old C57Bl6/J mice, as previously described [35]. Collagenase-digested isolated myocytes were incubated in buffer with increasing concentrations of Ca²⁺, achieving a final concentration of 1.2 mM Ca²⁺ as in DMEM plating media. Cells were seeded at 3 × 10⁴ rod-shaped myocytes/well in DMEM plating media supplemented with 10 mM 2,3-Butanedione monoxime (BDM, SIGMA) in 24-well plates coated with laminin. After 1h incubation at 37°C, 5% CO₂, plating media was replaced to remove unattached cells.

Human induced pluripotent stem cells (hiPSc) were differentiated into cardiomyocytes (CMs) by means of the STEMdiff™ Cardiomyocyte Differentiation Kit (Stemcell Technologies) following manufacturer’s instructions. Briefly, hiPSc were plated at the density of 3.5 × 10⁵ cells/well in a 12-well plate. After reaching 90% confluence, cells were incubated with the induction medium A for the mesodermal commitment. Two days after, cardiac mesoderm commitment was achieved by the addition of medium B. After four days, the medium was replaced with medium C for the cardiac induction. After eight days from the beginning of the differentiation protocol, the induction medium C was replaced with maintenance medium and cells were re-fed every 48h. Cells differentiated with this protocol display spontaneous beating usually on Day 7 and robust contracting syncytium by Day 10. On Day 20, beating cardiomyocytes have been selected as described before [36]. Briefly, cells were cultured in Glucose-free DMEM (no glucose, no pyruvate, Life Technologies) supplemented with 4 mM Na-DL-Lactate (Sigma) for 48h. Medium was renewing every day and it was replaced after 2 days with maintenance medium. Cardiomyocytes were used for the experiments after 4 days from the selection protocol.

To evaluate the effect of the compounds, cells were pre-treated with different concentrations of either 63 and TR002, from 0.1 to 1 μM, for 30 min unless specified in section 3. Moreover, to prevent PTP opening cells were pre-treated with 1 μM CsA (Sigma) [4] for 30 min as control.

4.2. Tissue mitochondria isolation.

Cardiac mitochondria were isolated from Sprague Dawley rats (2–3 months old), control mice (C57BL/6 wild type mice) and mice lacking the Ppif gene which encodes for CyPD (Ppif^−/− mice) which have been described previously [37]. Mouse heart mitochondria were isolated in mouse heart mitochondria (MHM) buffer (in mM: 225 mannitol, 75 sucrose, 10 HEPES-Tris and 1 EGTA, pH 7.4) as described before [4]. Briefly, hearts were rapidly excised, rinsed with cold MHM to remove blood and cut into small pieces. These were re-suspended into 1 ml of MHM supplemented with 0.1 mg/ml Nagarse (Sigma) and homogenized with a pre-chilled glass-Teflon homogenizer for 3 min. Then 6 ml of MHM supplemented with 0.2% BSA were added to stop Nagarse action and homogenate was spun at 500 g for 10 min at 4°C. Supernatant was filtered through a 150 μm mesh, transferred into a new tube and spun at 10000 g for 10 min at 4°C. Pellet was resuspended in MHM buffer, spun at 7500 g for 6 min at 4°C and washed again with MHM. After another centrifuge at 7500 g for 6 min pellet was re-suspended in 100 μl MHM and stored on ice. Protein quantification was determined with BCA method.

Mouse liver mitochondria were isolated in MS-EGTA buffer (in mM: 225 mannitol, 75 sucrose, 5 HEPES, and 1 EGTA, pH 7.4, Sigma). Tissues were minced into 2 mm X 2 mm pieces and then homogenized using a glass and teflon tissue homogenizer (8 strokes) on ice and in MS-EGTA buffer. Tissue homogenates were then subjected to an 800 g centrifugation for 5 min (1X) and the supernatants were then centrifuged at 10000 g for 10 min. Isolated mitochondria were then suspended in KCl buffer (in mM: 125 KCl, 20 HEPES, 2 MgCl₂, 2 KH₂PO₄, and 0.04 EGTA, pH 7.2, Sigma) for subsequent assays.

All experimental procedures with animals conducted at the Cincinnati Children’s Medical Center were approved by the Institutional Animal Care and Use Committee of Cincinnati Children’s Medical Center, protocols IACUC 2018–0047. We have complied with the relevant ethical considerations for animal usage overseen by this committee. The number of mice used in this study reflects the minimum number needed to achieve statistical significance based on experience and previous power analysis. Both sexes of mice were used in equal ratios.

4.3. Assessment of calcium retention capacity (CRC) and mitochondrial swelling assays.

CRC was determined using Calcium Green-5N (Invitrogen), a low affinity membrane-impermeant probe that increases its fluorescence emission upon Ca²⁺ binding. Just prior to initiation of the assay, mitochondrial suspension was spun at 9500 g for 5 min at 4°C to remove EGTA, supernatant was then discarded and the pellet suspended at 0.25 mg/ml in a buffer containing 250 mM sucrose, 10 mM MOPS-Tris, 1 mM Pi-Tris, 10 μM EGTA-Tris, 5 mM glutamate-Tris, 2.5 mM malate-Tris and 0.5 μM Calcium Green-5N, pH 7.4. Mitochondrial suspension was then dispensed to wells of black 96-well plate (final volume 0.2 ml) and subjected to trains of 5 μM (mouse mitochondria) or 10 μM (rat mitochondria) Ca²⁺ pulses followed by 3s linear shaking and fluorescence (ex/em = 485/535) read for 4 min with Tecan Infinite F200 plate reader. Experiments were performed at 25°C.

The CRC assay and the mitochondrial swelling assay in Ppif^−/− liver mitochondria were performed simultaneously using a dual-detector (one to measure fluorescence and the other to measure absorbance) single-cuvette based fluorimeteric system (Horiba Scientific). Isolated mitochondria (2 mg) were loaded into the cuvette along with 0.25 μM Calcium Green-5N (Invitrogen), 7 mM pyruvate (Sigma), and 1 mM malate (Sigma) and brought up to 1 ml using KCl buffer (as described above). Mitochondria were then pulsed with sequential additions of 100 μM CaCl₂ until PTP opening occurred or until the mitochondria reached a CaCl₂ saturation point and could no longer take up more Ca²⁺.

4.4. Live Imaging

Experiments using NRVMs were carried out in HBSS at pH 7.4 (adjusted with NaOH) and at 37°C.

Images were acquired using an inverted fluorescence microscope (Leica DMI6000B equipped with DFC365FX camera) with PL APO 40x/1.25 oil objective. Fluorescence intensity was quantified using the Fiji distribution of the Java-based image processing program ImageJ [38], and background signal was subtracted from all analyzed regions of interest. For Ca²⁺ imaging, traces were analyzed using the “Peak Analyzer” tool of Origin Pro 9.1.

To monitor mitochondrial membrane potential (ΔΨm), cells were incubated with 25 nM tetramethylrhodamine (Thermo Fisher Scientific) in presence of 1.6 μM cyclosporin H for 30 min at 37°C in a humidified incubator. Tetramethylrhodamine fluorescence intensity was monitored following addition of 4 μM oligomycin (Sigma) and images were acquired before and after the addition of 4 μM carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (Sigma) [34]. PTP opening was induced by 5 μM of the Ca²⁺ ionophore calcimycin, in presence or absence of 1 μM MitoParaquat that evokes a primary increase in mitochondrial ROS [21].

To monitor PTP opening, cells were incubated with 1 μM calcein acetoxymethyl (AM) ester (Thermo Fisher Scientific) in presence of 2 mM Cobalt Chloride (CoCl₂) for 15 min at 37°C in a humidified incubator followed by 20 min of de-esterification as previously described [17].

To monitor [Ca²⁺]_I homeostasis, cells were incubated with 5 μM Fluo-4 acetoxymethyl (AM) ester (Thermo Fisher Scientific), 0.01% w/v pluronic F-127 (Sigma) and 250 μM sulfinpyrazone (Sigma), for 20 min at 37°C in a humidified incubator followed by 20 min of de-esterification. The experiments were performed in normoxia and after both 3h and 6h of anoxia followed by 30 min of reoxygenation.

4.5. Immunocytochemistry

NRVMs were prepared for staining by means of Cardiomyocyte Immunocytochemistry Kit (Thermo Fisher Scientific) following manufacturer’s instruction. Briefly, NRVMs were fixed with Fixative Solution for 30 min, permeabilized with Permeabilization Solution S for 15 min at room temperature and then blocked with Blocking Solution for 30 min. Mitochondria were stained using anti-TOM20 antibody (Santa Cruz; 1:500, rabbit) and sarcomeres were stained using anti-α-sarcomeric actinin antibody (Sigma; 1:500, mouse) over-night at 4°C. The day after, samples were incubated with Alexa Fluor 594 conjugated anti-rabbit (Life Technologies, 1:250) and Alexa Fluor 488 conjugated anti-mouse (Life Technologies, 1:250) for 1h at room temperature. Coverslips were mounted using NucBlue™ Fixed Cell ReadyProbes™ Reagent with DAPI to stain nuclei (Invitrogen). Images were collected at Zeiss LSM 700 confocal system equipped with a PlanApo 40x/1.2 oil objective, and an Argon laser used to excite the fluorophores at the appropriate wavelength. Images were collected at 2048×2048 pixels per image with a 100 Hz acquisition rate. Images were analyzed using the Fiji distribution of the Java-based image processing program ImageJ [38]. The experiments were performed both in normoxia and after 6h of anoxia followed by 30 min of reoxygenation.

4.6. Western blot analysis

NRVMs were plated in 12-well plates at density of 4 × 10⁵ cells/well. Five days after, cells were incubated for 6 h in HBSS supplemented with DMSO, 1 μM 63, 1 μM TR002 or 1 μM CsA. NRVMs were homogenized in RIPA Lysis Buffer (1X, EMD Millipore) containing protease (1X, cOmplete mini protease inhibitor cocktail, Roche) and phosphatase (1X, PhosStop, Roche) inhibitors. Protein concentration was determined using BCA protein assay (BCA Kit, Euroclone). Proteins were separated using 4–12 % Bis-Tris gels (Life Technologies) in MES/SDS running buffer 1X (i.e. 1 M MES, 1 M Tris Base, 69.3 mM SDS, 20.5 mM EDTA, pH 7.3). Proteins were transferred to nitrocellulose membrane (Bio-Rad) by western blot in transfer buffer (i.e. 250 mM Tris Base, 1.92 M glycine, 0.1 % SDS) over-night (O/N) at 4°C. Membranes were blocked for 1 h with 5 % BSA in TBS-Tween, and incubated O/N with the following primary antibodies diluted in 5 % BSA:

• Anti p-DRP1 (Ser637) → 1:1000 (Cell Signaling, #4867, rabbit)

• Anti DRP1 → 1:1000 (BD Biosciences, #611738, mouse)

• Anti MFN2 → 1:1000 (Abnova, #H00009927-M03, mouse)

• Anti ATP Synthase subunit β → 1:1000 (Abcam, #14730, mouse)

Following incubation with primary antibodies, membranes were washed three times in TBS-Tween, 5 min each, and incubated for 1 h with the following secondary antibodies diluted in 5% BSA:

• Anti mouse (m-IgGκ BP-HRP) → 1:3000 (Santa Cruz Biotechnology, #sc-516102)

• Anti rabbit (mouse anti-rabbit IgG-HRP) → 1:3000 (Santa Cruz Biotechnology, #sc-2357)

Following incubation with primary and secondary HRP-conjugated antibodies, bands were detected by the PierceTM ECL Western Blotting Substrate using the UVITEC Cambridge, mini HD instrument. The results were expressed as densitometry analysis performed using the Fiji distribution of the Java-based image processing program ImageJ [38].

4.7. Assessment of cell death in vitro

For anoxia/reoxygenation experiments, NRVMs and hiPSc-CMs were seeded in 24-well plates at density of 1.5 × 10⁵ cells/well and incubated in (in mM) 118 NaCl, 5 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 2 CaCl₂, 25 MOPS at pH 6.4 during anoxia or pH 7.4 during reoxygenation, while AMVMs were seeded in 24-wells plates at density of 3 × 10⁴ cells/well and incubated in (in mM) 118 NaCl, 5 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 2 CaCl₂, 10 HEPES, 4 NaHCO₃, supplemented with 25 μM Blebbistatin (Sigma) at pH 6.4 during anoxia or pH 7.4 during reoxygenation [39]. Anoxia was induced adding 10 mM 2-deoxy-D-glucose and incubating in a BD GasPak™ EZ Anaerobe Gas-generating Pouch System with an indicator (BD Biosciences) at 37°C. NRVMs were exposed to anoxia for 12h, hiPSc-CMs were exposed to anoxia for 10h, and AMVMs were exposed to anoxia for 8h [40]. To induce reoxygenation, plates were removed from the GasPak™ pouch, 2-deoxy-D-glucose was replaced with 10 mM D-glucose, the pH was restored at 7.4. The plates were then incubated in a humidified incubator at 37°C (i.e. 4h NRMVs, 4h hiPsc-CMs, 10h AMVMs). A schematic description of the protocol applied in vitro is reported in Supplementary Fig. 4A.

The release of lactate dehydrogenase (LDH) from cardiomyocytes was measured to evaluate cell death occurring in normoxia, anoxia and reoxygenation as described before [4, 41]. Supernatant aliquots were collected after both anoxia and reoxygenation. At the end of every experiment, intact cells were lysed by incubating with 1% Triton X-100 (Sigma) for 30 min and supernatants were collected to evaluate the total amount of LDH. LDH enzymatic activity was measured spectrophotometrically by the absorbance of nicotinamide adenine dinucleotide (Roche) at 340 nm, indicative of the reduction of pyruvate to lactate.

4.8. Assessment of cell death ex vivo

In one series of studies, hearts were subjected to ischemia/reperfusion (I/R) using a non-recirculating Langendorff model at constant flow [42, 43] with a bicarbonate buffer (in mM: 118.5 NaCl, 3.1 KCl, 1.18 KH₂PO₄, 25 NaHCO₃, 1.2 MgCl₂, 1.4 CaCl₂, and 11 D-glucose, pH 7.4) insufflated with 95% O₂ – 5% CO₂ at 37°C, supplemented with different concentrations of DMSO, CsA, 63, or TR002 (Supplementary Fig. 4B). For LDH measurements, the hearts were rapidly excised from cervical-dislocated 5 months’ sex-matched C57BL/6J mice and perfused retrogradely for 20 min at a constant flux of 5 ml/min. The hearts were subjected to 40 min of global no-flow ischemia, and 15 minutes of reperfusion. Coronary effluent during the reperfusion period was collected every minute to measure LDH. At the end of reperfusion, the hearts were quickly homogenized in 0.5% Triton X-100 for total LDH measurement. Since all values were normalized to heart weight, the amount of LDH release was expressed as % of total LDH (i.e. effluent + homogenate) to rule out possible changes due to variations in heart size [4].

All experiments conducted at the National Institutes of Health received humane treatment in accordance with National Institutes of Health guidelines and the “Guiding Principles for Research Involving Animals and Human Beings”. This portion of the study was reviewed and approved by the Institutional Animal Care and Use Committee of the National Heart, Lung, and Blood Institute. In experiments testing hemodynamic recovery and infarct size, the hearts were excised from 12–16-week-old C57BL/6N (Taconic Farms, USA) mice under pentobarbital anaesthesia and perfused retrogradely at 100 cm of water at 37°C (Supplementary Fig. 4B) as previously described [44, 45]. Both heart rate (HR) and left ventricular developed pressure (LVDP) were measured using a latex balloon connected to a pressure transducer and inserted into the left ventricle. Data was recorded and digitized using a PowerLab system (ADInstruments, Colorado Springs, CO). The rate-pressure product (RPP = LVDP X heart rate) was used as an index of cardiac contractile function. RPP recovery is presented as a percent recovery at the end of reperfusion as compared to baseline perfusion. For measurement of myocardial infarct size, at the end of the 90 min of reperfusion, hearts were perfused with 1% (wt/vol) of 2,3,5-triphenyltetrazolium chloride (TTC) and incubated in TTC at 37°C for 30 min, followed by fixation in 10% (wt/vol) formaldehyde. Infarct size was expressed as the percentage of the total cross-sectional area of the heart.

4.9. Data analysis

All values are expressed as mean ± S.E.M. Comparison between groups was performed by one-way ANOVA, followed by a Tukey’s post hoc multiple comparison where data were normally distributed. Data that did not follow the normal distribution were statistically analysed by Kolgomorov-Smirnov’s test. Comparisons between two groups were performed using a two-tailed Student’s t-test. A value of p < 0.05 was considered significant.

Abbreviations

A/R

anoxia/reoxygenation

AMVMs

adult mouse ventricular myocytes

CRC

calcium retention capacity

CsA

cyclosporine A

CyPD

cyclophilin D

hiPSc-CMs

human iPSc-derived cardiomyocytes

I/R

ischemia/reperfusion

LDH

lactate dehydrogenase

MI

myocardial infarction

MitoPQ

MitoParaquat

NRVMs

neonatal rat ventricular myocytes

PTP

permeability transition pore

ROS

reactive oxygen species

TMRM

tetramethylrhodamine

ΔΨm

mitochondrial membrane potential