Mitochondrial permeability transition pore desensitization by a novel dispiranic derivative prevents cardiac reperfusion injury

TL;DR

A new compound prevents heart damage during reperfusion by inhibiting mitochondrial pore opening without harming cell energy production.

Contribution

A novel dispiranic compound (11d) is identified as a selective PTP desensitizer targeting ATP synthase c subunit.

Findings

Compound 11d inhibits PTP opening by ~70% in human cardiomyocytes at low micromolar concentrations.

11d selectively binds to ATP synthase c-ring and reduces cell death and ROS in hypoxia/reoxygenation models.

Ex vivo heart models show 11d improves cardiac recovery and reduces infarct cell death during reperfusion.

Abstract

Mitochondrial permeability transition pore (PTP) opening is a major determinant of cardiac ischemia/reperfusion (I/R) injury, contributing to cardiomyocyte death and impaired cardiac function following revascularization. Despite extensive research, effective pharmacological strategies targeting PTP remain limited. Here, we report the identification and characterization of a novel class of dispiranic small molecules designed to inhibit PTP opening by targeting the c subunit of FO-ATP synthase. Using a multidisciplinary approach combining chemical synthesis, cellular and mitochondrial assays, ex vivo cardiac models and in silico analyses, we identified compound 11d as a potent and selective PTP desensitizer. In living human cardiomyocytes, 11d inhibited PTP opening by approximately 70% at low micromolar concentrations without impairing mitochondrial bioenergetics under basal conditions. Mechanistically, 11d selectively inhibited Ca2+-activated F1FO-ATP synthase activity, consistent with PTP modulation, while sparing Mg2+-dependent ATP synthesis and succinate oxidase activity. Molecular docking and molecular dynamics simulations revealed stable binding of 11d within the canonical oligomycin-binding pocket of the ATP synthase c-ring, a finding corroborated by cellular thermal shift assays and subcellular fractionation showing preferential mitochondrial accumulation. Functionally, 11d significantly reduced cell death, mitochondrial reactive oxygen species (ROS) production and PTP opening in cardiomyocytes and endothelial cells subjected to hypoxia/reoxygenation (H/R). Importantly, in an ex vivo Langendorff rat heart model, 11d administration at reperfusion markedly improved cardiac functional recovery, reduced infarct-associated cell death and preserved myocardial architecture. Collectively, these results identify 11d as a promising lead compound for cardioprotection and support targeting the ATP synthase c subunit as a viable therapeutic strategy against reperfusion injury.

Figures

Validation of dispiranic structures as inhibitors of PTP. (A) Calcein−cobalt assay in living cardiomyocytes (AC16) pretreated with vehicle (DMSO) or the PTP inhibitors to be screened. PTP opening was stimulated by ionomycin administration and representative kinetics data are reported on the right. The raw values of slopes were expressed as a percentage of the vehicle. Each value is the mean of at least 30 cells from 3 biological and 3 technical replicates. (B) Mean percentages of PTP inhibition ± SEM for each compound (1 μM) have been reported. (C) Dose-responses graph for the 11c, 11d and 12c

Effect of dyspiropyrrolidine derivatives on mitochondrial bioenergetics. (A) Dose-response curve of F_1_F_O_-ATPase activated by Mg^2+^ (●) or by Ca^2+^ (○). (B) Kinetics inhibition of the mitochondrial Ca^2+^-activated F_1_F_O_-ATPase by 11d. (C) Mutual exclusion analysis between 11d and oligomycin (olig) or DCCD inhibitors of F_1_F_O_-ATPase. (D) NADH-O_2_ oxidase activity and Succinate-O_2_ oxidase activity without and with increasing concentrations of 11d, 11c and 12c (0.5, 1, 10 and 100 μM). (E) Representative curves of the calcium retention capacity (CRC) curves for assessing PTP opening

Biological characterization of dyspiropyrrolidines-derivatives at normoxic conditions. (A) OCR kinetics reporting mean values with SEM of all experiments done on the left; on the right are reported the statistics for the calculation of basal respiration rate, ATP-linked respiration and maximal respiratory capacity. Each value is the mean of at least 3 biological and 3 technical replicates. (B) Measurements of mitochondrial calcium uptake in AC16 cells using a mitochondrially targeted aequorin probe (mtAeq); [Ca^2+^] is detected after addition of 100 μM histamine (His) and 100 μM Bradykinin (Bk

Cytoprotective effects following H/R. (A) Calcein−cobalt assay in AC16 cells. The kinetics of PTP opening were evaluated at the moment of the reoxygenation in the presence of vehicle, 11d or 12c compounds. The raw values were expressed as a percentage of the vehicle. Each value is the mean of at least 10 cells from 3 biological and 3 technical replicates. (B) Quantification of AC16 cells positive to Propidium Iodide (PI^+^) staining following H/R. Each value is the mean of 3 biological and 4 technical replicates. (C) Calcein-cobalt assay in differentiated HCM cells. Data has been collected at

Evaluation of cell metabolism and cell viability in the presence of dyspiropyrrolidine compounds in pAECs. (A) Evaluation of the real-time ATP production rate by mitochondrial OXPHOS (MitoATP) or by glycolysis (GlycoATP) in pAECs treated with the compounds. The ATP rate index is shown on the y-axis (logarithmic scale) in pAECs treated with the dyspiropyrrolidine derivatives. (B) The mitochondrial respiration profile was obtained from the OCR without (Vehicle) and with compounds under basal respiration conditions and after the addition of 1.5 μM oligomycin (OLIG), 1.0 μM FCCP, and a mixture of

Cardioprotective effects of dyspiropyrrolidine compounds in a model of cardiac reperfusion injury. (A) Left ventricular peak developed pressure recording of rat hearts during the experimental protocol composed by the stabilization, ischemia and reperfusion phases. The data were detected by using the Langendorff *ex vivo* protocol. (B) End-diastolic pressure recording of a rat heart in the stabilization, ischemia, and reperfusion phases. Histograms of statistical and representative kinetics data are reported. Kinetics and statistics are representative of 5 different animals per experimental con

Docking results of compound 11d (blue) and compound 12c (green). (A) Overall view of the human ATP synthase C-ring (grey ribbons) displaying the top-ranked docking poses for both compounds. (B) Zoomed-in view of the binding site, showing the ligands and surrounding amino acids within 5 Å. (C) RMSD values of 11d plotted against simulation time (ns). (D) RMSD values of 12c during the time (ns). (E) Detailed view of the dominant binding pose of compound 11d (shown in blue), derived from the most populated cluster during MD simulations. Amino acids within 5 Å radius of the ligand are shown in oran

Full text

Introduction

1

Multiple pathological features link dysfunctional mitochondria to cardiovascular diseases (CVDs) onset and include quality control mechanisms impairments [1,2], defects in mitochondrial morphology [3,4], shift in the metabolism [5,6] and mitochondrial-mediated cell death [7]. In the last years, evidence of a novel mitochondrial target has become substantial [8], with challenging uses as therapeutic strategy in the cure of the residual damage following percutaneous coronary interventions (PCI) [9] in myocardial infarction (MI). This is the permeability transition pore (PTP) [10,11], a channel across mitochondrial membranes composed of multiple proteins with both structural and modulatory properties [12]. Among them, the subunit c of F_1_F_O_ -ATP synthase (Csub) constitutes the pore located in the inner mitochondrial membrane (IMM) [13], a pore whose opening is triggered by increased intracellular calcium concentrations ([Ca^2+^]), oxidative stress and changes in the amount of H^+^ ions in the mitochondrial matrix [14]. If the opening of the PTP occurs in the high conductance mode and in most mitochondria composing the cardiomyocyte, this generates an irreversible status of permeability transition of the IMM causing drastic osmotic dysregulation, mitochondrial swelling and mitochondria-dependent cell death [15].

At reperfusion time (i.e., following PCI), all the above-mentioned biochemical events occur; for this reason, PTP opening is considered to be an important determinant of cell death, tissue function loss and reperfusion damage [16]. Although the exact composition of this complex is still a matter of debate, current technology is able to finely measure the channel opening in living cells [17] and to evaluate the related damage; as consequence, also the cardioprotective impact of PTP inhibition can be carefully analyzed for translational purposes [18,19].

Moreover, recently, the importance of the interconnection between PTP activity and reperfusion damage has been elucidated also in studies involving humans, by reporting a significant intersubject variability of PTP opening among patients [20,21], how this variability is related to a different grade of reperfusion damage and how it can be also dependent on genetic determinants [20].

Clearly, this pathological role of the PTP in humans has attracted increasing attention in terms of drug discovery. Overall, Csub and Cyclophilin D (CypD) have been defined critical components of the PTP; however, CypD is considered only a regulator of the pore [22], thus its targeting cannot fully cover the clinical needs [[23], [24], [25]]. On the other hand, Csub is a protein of the F_O_ portion of ATP synthase, thus involved in proton translocation in the process of ATP production. Therefore, its inhibition may impair mitochondrial function and cell viability. However, considering the important role of the Csub in the process of PTP opening [26], many research efforts are aimed at developing functional Csub inhibitors with minimal or no toxicity [18,27,28].

To date, different scaffolds have been proven as interesting inhibitors of PTP opening. In particular, ureidic [28] and monospiropiperidinic scaffolds [18,29] were found to be quite useful in obtaining information about the validity of the assumptions described above, as well as demonstrating inhibitory activity on mPTP opening.

These studies have shown how specific classes of small molecules are active against ischemia/reperfusion (I/R) damage. We interestingly observed that elements such as carbonyl functionality, aromaticity, a precise steric hindrance, and heteroatoms distributed in particular positions (such as ureas, amides and tertiary and cyclic amines) seem to be particularly necessary in order to exert biological activity in this field.

The main and innovative idea of this work is to harmonize all these elements within a single scaffold, aiming to synthetize novel and more potent cardioprotective agents, fighting the challenge of I/R injury. This intention led us to the conception and design of the scaffolds discussed in this paper, exploiting simple and smart synthetic strategies, to achieve rather complex molecules in an atom-economical way and with good yields. Starting from previous data obtained from our group and mentioned above, we combined all the information in a single molecule, increasing structural rigidity with another spirocenter, designing a molecule that, although structurally complex, could be able to better interact with the target, because it is structurally rigid.

Initially, we maintained the piperidinic core, taking inspiration from the results previously obtained, so producing the piperidone scaffold-based dispiranic compounds described below. In a second step, we further modified the structure, by drawing inspiration from a core that is very common in pharmacologically active molecules, namely isatin [30]. In this way, we have obtained the class of pirrolidine scaffold-based dispiranic compounds. We validated them in a larger biological scenario including isolated mitochondria, living cardiomyocytes, endothelial cells and ex vivo models of rat and porcine origin. Furthermore, we proposed a potential target through which these compounds inhibit PTP activity.

Results

2

Identification of dispiranic structures as novel inhibitors of PTP

2.1

• ●Piperidone scaffold-based dispiranic compounds.

The first class of dispiranic compounds considered is that of the piperidine scaffold compounds. It seemed coherent to us to use the same main core analyzed deeply in other studies, as it has led to some remarkable results in monospiranic compounds [18]. Here we wanted to expand the structure to achieve a new library of dispiranic compounds, combining this scaffold with tetrahydrothiophene, taking inspiration from nature, since it is present in many natural and synthetic compounds with biological activity [31]. The library of dispiro-piperidone - tetrahydrothiophene heterocyclic compounds was synthesized with two steps. The first reaction is the alkylation of N-benzyl piperidone in position 3 and 5 with different aromatic aldehydes, through a double Knoevenagel-like reaction. N-benzyl piperidone in its enolic form, which works as nucleophile, attacks the aromatic aldehyde. Intermediates 3a-f are achieved after the dehydration of the aldol (Scheme 1).Scheme 1. Synthesis of intermediates 3a-f.Reaction conditions. a) NaOH 10%, EtOH abs, 2h, r.t.Scheme 1

Once the intermediates are obtained, with yields between 32% and 73%, the cycles are closed to form the final dispiranic compound. The reagent used for this purpose is 1,4-dithiane,2,5-diol, which releases two molecules of 2-mercaptoacetaldehyde: once deprotonated by triethylamine (TEA), it acts as a nucleophile, attacking the carbon in beta of the intermediate molecule, through a Michael addition. At this point, a monospiranic product is thus obtained. Through another Michael addition, the final dispiranic products 5a-f are then achieved (Scheme 2), with yields between 30% and 92%.Scheme 2. Procedure to obtain final dispiropiperidonic compounds 5a-f.Reaction conditions. a) TEA, EtOH abs, 3h, 0° to 90 °C.Scheme 2

It is interesting to know that the second annulation involving the reaction of another anion of 2-mercaptoacetaldehyde to the monospiranic intermediate again occurs on the same side as the initial Micheal addition presumably in a bid to obviate its steric interaction with the initially generated benzylic carbon [32].

• ●Pyrrolidine scaffold-based dispiranic compounds.

In a second phase, a new core has been selected and two series of compounds have been synthesized from isatin. This has been reacted with sarcosine and various cyclic ketones leading to the synthesis of several similarly structured dyspiropyrrolidinic compounds (Scheme 3). The reactions were conducted using the one pot reactions technique: the final product, although molecularly rather complex, was obtained in a single step. Isatin has been chosen because of its interesting activity in the biological field, as it is present as a precursor in many natural compounds and it can be found in several biologically active molecules [30,33,34]. In addition, as mentioned before, we selected this particular scaffold because it clearly presents some of the functionalities we have identified as indispensable to make the molecule active towards our goals (such as lactam and the aromatic moiety). Furthermore, in fluorine-isatin derivatives, we wanted to introduce the fluorine atom as it appears to have interesting pharmacodynamic properties concerning metabolic stability of the molecule [35]. Initially, monocyclic ketones were used; subsequently, α-tetralone and acenaphthenone were chosen, in an attempt to insert additional aromatic elements into the molecule, with the aim of increasing the hydrophobic interactions of the molecule with the receptor (see Scheme 4).Scheme 3. General procedure for the synthesis of the dispiropyrrolidinic compounds.Reaction conditions. a) MeOH, 2-10 h, 90 °C. n = 1 carbon atom (d); 2 carbon atoms (a-c).Scheme 3. Scheme 4General procedure for the synthesis of compounds 13 and 14.Reactions conditions: benzyl bromide, NaH, anhydrous DMF, r.t., 30 min.Scheme 4

In two cases, we decided also to further increase the steric hindrance of the molecule, so we added a benzylic functionality. To do this, we have alkylated the amidic nitrogen of the isatin with benzyl bromide, after deprotonation with NaH, thus obtaining the products 13 and 14, bulkier and with a new hydrophobic moiety (that is the benzylic group).

These tandem reactions are of paramount importance in the context of green chemistry as they offer a convenient strategy for the rapid, elegant and convergent construction of complex organic molecules without isolating and purifying the intermediates resulting in substantial minimization of waste, labor, time and cost.

Validation of dispiranic structures as inhibitors of PTP

2.2

To investigate the effectiveness of these compounds against PTP opening, we used two different methods. The first one applied in human living cells derived from ventricular cardiomyocytes (AC16 cells) named as calcein-cobalt quenching assay (Fig. 1A) which included all small molecules derivatives of the study at a concentration of 1 μM as first screening; here, 11c, 11d and 12c significantly inhibited PTP opening as compared to the vehicle. Moreover, 11d and 12c are the most potent of the series as they further decrease the channel activity as compared to 11c (Fig. 1A). The mean percentages of PTP inhibition for the small molecules of this assay are reported in Fig. 1B; of note, 11d is able to inhibit channel activity by about 70%.Fig. 1. Validation of dispiranic structures as inhibitors of PTP. (A) Calcein−cobalt assay in living cardiomyocytes (AC16) pretreated with vehicle (DMSO) or the PTP inhibitors to be screened. PTP opening was stimulated by ionomycin administration and representative kinetics data are reported on the right. The raw values of slopes were expressed as a percentage of the vehicle. Each value is the mean of at least 30 cells from 3 biological and 3 technical replicates. (B) Mean percentages of PTP inhibition ± SEM for each compound (1 μM) have been reported. (C) Dose-responses graph for the 11c, 11d and 12c compounds compared with no inhibitor in the range 10 nM - 100 μM. Each value is the mean of at least 25 cells from 3 biological and 3 technical replicates. (D) Mitochondrial swelling assay in isolated heart mitochondria pretreated with DMSO or 11c, 11d and 12c compounds (1 μM) or CsA (1.6 μM) or Ruthenium Red (RR, 5 μM). Data were obtained by recording changes in the absorbance at 540 nm. On the right, immunoblot detection of Actin (cytosolic marker) and ATP5A (mitochondrial marker) proteins in the mitochondrial fraction obtained from the crude subcellular fractionation of the most promising compounds-related experiments. Each value is the mean of 3 biological (mice) and 3 technical replicates. One-way ANOVA was applied for statistical analysis for all graphs reported in the figure; (∗∗∗∗) p value < 0.0001; (∗∗∗) p value < 0.001; (∗∗) p value < 0.01; (∗) p value < 0.05.Fig. 1

Once the most promising compounds were identified, we performed additional dose–response assays within each compound library, to determine the lowest concentration that produced the maximal inhibitory effect on PTP opening (Fig. 1C). Overall, 1 μM was the lowest concentration that achieved the maximal effect in counteracting channel activity. Based on these results, treatments at 1 μM (as in panel A) were used to perform a secondary assay in isolated mitochondria from murine hearts (Fig. 1D), in order to confirm the efficacy of the most promising compounds identified in the initial analysis. To this purpose, freshly isolated mouse heart mitochondria were prepared by following already established protocols [18,36] and were then stimulated with 500 μM Ca^2+^ through the mitochondrial swelling assay; changes in the absorbance at 540 nm was monitored for 10 min at which a decrease was indicative of swelling (Fig. 1D). A couple of controls have been added to the experiment: 1.6 μM CsA as a positive control to check PTP inhibition and 5 μM Ruthenium Red (RR) as mitochondrial calcium uniporter (MCU) inhibitor to check the response of mitochondrial preparations in respect to Ca^2+^ addition. Taken together, these findings validate the significant ability of 11c, 11d and 12c to inhibit the channel.

In addition, if we compare this data with those obtained from known PTP inhibitors (DCCD, oligomycin A, c.10) already investigated in our previous publications [18,20,27,28] and applied to the same experimental settings, 11d results to have greater effect at a lower dose.

Characterization of F1FO-ATPase related effects upon 11c, 11d and 12c treatment

2.3

Due to the modulatory similarities between F_1_F_O_-ATPase and PTP, which share some subunits, we evaluated the dose-response effects of the compounds on F_1_F_O_-ATPase in isolated swine heart mitochondria, using either magnesium or calcium as cofactors. (Fig. 2A).Fig. 2. Effect of dyspiropyrrolidine derivatives on mitochondrial bioenergetics. (A) Dose-response curve of F_1_F_O_-ATPase activated by Mg^2+^ (●) or by Ca^2+^ (○). (B) Kinetics inhibition of the mitochondrial Ca^2+^-activated F_1_F_O_-ATPase by 11d. (C) Mutual exclusion analysis between 11d and oligomycin (olig) or DCCD inhibitors of F_1_F_O_-ATPase. (D) NADH-O_2_ oxidase activity and Succinate-O_2_ oxidase activity without and with increasing concentrations of 11d, 11c and 12c (0.5, 1, 10 and 100 μM). (E) Representative curves of the calcium retention capacity (CRC) curves for assessing PTP opening with 100 μM 11d, 11c and 12c. ΔF (%) calculated as (Fmax − F_0_) × 100 relative to the Fura-FF probe signal intensity for each experimental condition calculated on the representative CRC curves. All data represent the mean ± SD of 3 technical replicates performed on 3 biological replicates. One-way ANOVA was used for statistical analysis using the Students-Newman-Keuls test. (∗∗∗∗) p-value <0.0001; (∗∗∗) p-value <0.001; (∗∗) p-value <0.01; (∗) p-value <0.05.Fig. 2

The enzyme activity of Mg^2+^-activated F_1_F_O_-ATPase was insensitive to the compounds tested. Conversely, Compound 11d achieved 50% enzyme inhibition at the highest concentration tested (100 μM) with calcium-activated F_1_Fo-ATPase, the activity of which is linked to modulation of PTP [37]. Kinetic experiments based on the building of Dixon and Cornish-Bowden plots were carried out in order to understand the 11d inhibition mechanism leading to establishing the features of the enzyme–inhibitor interaction (Fig. 2B). 11d caused an uncompetitive mechanism of inhibition (K_i_' = 140.0 ± 2.0 μM) by binding to enzyme–substrate complex and forming the Enzyme-Substrate-Inhibitor (ESI) ternary complex. The inhibition cannot be counteracted by raising the concentration of substrate ATP. As a result, 11d did not affect the ATP binding site placed on the F_1_ domain, the hydrophilic portion of the enzyme. Using a mutual exclusion analysis, we aimed to verify a possible interaction of 11d with the F_O_ domain, the intramembrane portion of the F_1_F_O_-ATPase complex. On considering Oligomycin A and dicyclohexylcarbodiimide (DCCD) as known F_O_ inhibitors, binary mixtures of 11d plus Oligomycin A or 11d plus DCCD were employed to draw a graphical representation of mutual interaction between 11d and Oligomycin A as well as between 11d and DCCD with the F_O_ domain of enzyme complex (Fig. 2C). 11d exibited a higher affinity for the enzyme in the presence of Oligomycin A, as the dissociation constant for the complex ESI1I2 (K_x_') with Oligomycin A was lower than with DCCD.

NADH- and succinate-energized mitochondria were used to evaluate the effect of 11d, 11c, and 12c compounds on mitochondrial respiration. The NADH-O_2_ oxidoreductase activity was reduced at high-concentration tested (100 μM) with 11d (68% inhibition) and 12c (27% inhibition), whereas the succinate-O_2_ oxidoreductase activity was insensitive to the compounds tested (Fig. 2D). Moreover, also the calcium retention capacity (CRC) of mitochondria has been in principle evaluated by the use of Fura-FF probe. The increase in Fura-FF ratio intensity detected by the CRC decrease revealed the PTP opening. The derivative compounds tested did not affect the threshold value of Ca^2+^ concentration triggering PTP formation. However, 11d attained a lower Fura-FF ratio when compared to the vehicle, 11c and 12c treatments at the PTP opening (Fig. 2E). A lower Fura-FF ratio value might match the amount of Ca^2+^ released with the desensitization effect of PTP by 11d seen elsewhere and linked to inhibitory effect on F_1_F_O_-ATPase Ca^2+^-dependent [38].

Biological characterization of dyspiropyrrolidines – derivatives at normoxic conditions

2.4

To investigate the biological features of these small molecules, several in vitro assays have been performed by analyzing main cellular and mitochondrial parameters of human living cardiomyocytes (Fig. 3). Given the crucial role of the F_O_ portion of ATP synthase in cellular energy production, being recognized as an important constituent of the PTP complex and the putative site of action of these derivatives [18], we determined the state of the mitochondrial respiration of cardiomyocytes alone or in the presence of 11c, 11d or 12c compounds. By using Seahorse XF analyzer, we monitored in 11c treated cells a poor decrease in mitochondrial Oxygen Consumption Rate (OCR) and a lower maximal respiration capacity (MRC), data which becomes significant in correspondence of ATP production when compared to the untreated condition (Fig. 3A). On the contrary, 11d and 12c derivatives did not alter the untreated phenotype, appearing safer than 11c in regarding mitochondrial function (Fig. 3A). Then, we evaluated other important mitochondrial parameters such as mitochondrial membrane potential and mitochondrial calcium (Ca^2+^m) uptake. They are critical factors for the maintenance of physiological functions such as metabolism and a wide range of intracellular signaling pathways [[39], [40], [41], [42]]. Both of them were unchanged following 11d and 12c (Fig. 3B and C) with a slight, but significant decrease in correspondence of 11c treatment for mitochondrial membrane potential measurements. Moreover, mitochondrial morphology which plays an important role in cardiac performance, was unaltered. Indeed, the treatment with 11c, 11d or 12c did not affect the average number of mitochondria and the total mitochondrial network (Fig. 3D and E), as well as the number of branches and their average length reflecting the extent of mitochondrial elongation, and the number of junctions, points where branches meet, reflecting the degree of mitochondrial network complexity (Fig. 3F–H). The positive control PTP-inhibitor CsA did not alter physiological conditions when used in normoxia, assuming a behavior similar to the other channel inhibitors. Taken together, these data show the absence of side effects on mitochondrial population of treated cells, with the exception of 11c which shows some slight toxic features; however, this led to unchanged cell viability and proliferation under acute (from minutes to hours) and chronic (days) treatments (Fig. 3J and K).Fig. 3. Biological characterization of dyspiropyrrolidines-derivatives at normoxic conditions. (A) OCR kinetics reporting mean values with SEM of all experiments done on the left; on the right are reported the statistics for the calculation of basal respiration rate, ATP-linked respiration and maximal respiratory capacity. Each value is the mean of at least 3 biological and 3 technical replicates. (B) Measurements of mitochondrial calcium uptake in AC16 cells using a mitochondrially targeted aequorin probe (mtAeq); [Ca^2+^] is detected after addition of 100 μM histamine (His) and 100 μM Bradykinin (Bk) and previous 15min pretreatment with the tested compounds (1 μM) and CsA (1.6 μM). Each value is the mean of 3 biological and 4 technical replicates. (C) Mitochondrial membrane potential performed in AC16 cells using the TMRM probe reported as the slope of the kinetics following FCCP 10 μM administration. Each value is the mean of at least 30 cells from 3 biological and 3 technical replicates. (D-I) Evaluation of mitochondrial morphology parameters using a mtGFP probe. The histograms represent the total mitochondrial network (D), the average number of mitochondria per z stack (E), number of branches (F), the network complexity in terms of number of junctions (G) and the average branch length (H) with representative images (I) in AC16 cells. Scale bar (−) 5 μm. Each value is the mean of at least 15 cells from 3 biological and 3 technical replicates. (J) MTS assay in living AC16 cells to evaluate the toxicity of compounds after 1, 2, 3, and 4 h of treatment. Data was obtained by recording changes in absorbance at 490 nm. Each value is the mean of at least 3 biological and 2 technical replicates. (K) Quantification of cell viability with crystal violet assay in living AC16 cells treated for 24h, 48h, 72h. Each value is the mean of at least 3 biological and 3 technical replicates. Compounds concentration: 1 μM, CsA concentration: 1.6 μM. One-way ANOVA was applied for statistical analysis for all graphs reported in the figure; (∗∗∗∗) p value < 0.0001; (∗∗∗) p value < 0.001; (∗∗) p value < 0.01; (∗) p value < 0.05.Fig. 3

Cytoprotective effects of 11d and 12c in cardiomyocytes following hypoxia/reoxygenation

2.5

Considering the lower therapeutic potential of 11c and to validate the hypothesis that 11d and 12c compounds are effective against reperfusion damage, we took advantage of a hypoxic chamber, through which we first evaluated the kinetics of the PTP during hypoxia/reoxygenation (H/R) in human cardiomyocytes. H/R induction resulted in an increased PTP opening at reoxygenation time, which has been desensitized when cells are treated with 11d and 12c by 55% and 29%, respectively (Fig. 4A). A similar result has been obtained for differentiated HCM cells (Fig. 4C). In the same conditions, PTP inhibition prevented excessive cell death as measured with two independent techniques and evaluating both apoptotic and necrotic markers (Fig. 4B–D), increasing the rate of survival following H/R (Fig. 4E). Being known contributors of PTP opening after ischemic episodes, the effect of both compounds on mitochondrial ROS and Ca^2+^m uptake have been evaluated following H/R. As well established [43], H/R led to a significant increase in mitochondrial ROS production; also, mitochondrial Ca^2+^ is dysregulated following ischemia [44]. To investigate whether the new compounds are able to interfere with these pathways, cardiomyocytes were treated with 11d or 12c and ROS and Ca^2+^m have been analyzed. Surprisingly, 11d acquires additional features compared to 12c by strongly reducing ROS levels with a mild, but significant effect also on Ca^2+^m following H/R (Fig. 4F and G).Fig. 4. Cytoprotective effects following H/R. (A) Calcein−cobalt assay in AC16 cells. The kinetics of PTP opening were evaluated at the moment of the reoxygenation in the presence of vehicle, 11d or 12c compounds. The raw values were expressed as a percentage of the vehicle. Each value is the mean of at least 10 cells from 3 biological and 3 technical replicates. (B) Quantification of AC16 cells positive to Propidium Iodide (PI^+^) staining following H/R. Each value is the mean of 3 biological and 4 technical replicates. (C) Calcein-cobalt assay in differentiated HCM cells. Data has been collected at reperfusion time in the presence of vehicle or 11d compound. (D) Immunoblot detection of the main markers of apoptosis and necrosis such as Cleaved PARP, Cleaved Caspase 3 and Cleaved RIP. This is representative of 3 biological replicates. (E) Quantification of cell viability according to viable cells upon H/R staining with crystal violet. This is representative of at least 3 biological and 3 technical replicates. (F) Quantification of the MitoSox intensity of cells after H/R under the same experimental conditions. Each value is the mean of at least 15 cells from 3 technical and 3 biological replicates. (G) Mitochondrial calcium uptake in AC16 cells by using a mitochondrially targeted aequorin probe (mtAeq); [Ca^2+^] is detected at the time of reoxygenation and after addition of 100 μM His and 100 μM Bk. Each value is the mean of at least of 3 biological and 3 technical replicates. Histograms of statistical and representative kinetics data are reported. One-way ANOVA was applied for statistical analysis for all graphs reported in the figure; (∗∗∗∗) p value < 0.0001; (∗∗∗) p value < 0.001; (∗∗) p value < 0.01; (∗) p value < 0.05.Fig. 4

Cytoprotective effects of dyspiropyrrolidine derivatives in porcine aortic endothelial cells following hypoxia/reoxygenation

2.6

The effect of derivative compounds on the OCR and extracellular acidification rate (ECAR) values under basal metabolic conditions has been detected to evaluate the rate of ATP production by oxidative phosphorylation (mitoATP) or glycolysis (glycoATP). The compounds tested did not affect the function of mitochondrial oxidative metabolism that was independent of dyspiropyrrolidine derivatives treatment of pAECs, model used for H/R studies [45]. The energetics profile was confirmed by ratio values between mitoATP production rate and glycoATP production rate (ATP Rate Index) that was always >1 (Fig. 5A). By considering the mitochondrial role in pAECs metabolism, the effect of different compounds was evaluated on bioenergetic parameters. However, mitochondria were unresponsive and the mitochondrial respiration in the presence or in the absence of derivatives did not differ from the control (Fig. 5B).Fig. 5. Evaluation of cell metabolism and cell viability in the presence of dyspiropyrrolidine compounds in pAECs. (A) Evaluation of the real-time ATP production rate by mitochondrial OXPHOS (MitoATP) or by glycolysis (GlycoATP) in pAECs treated with the compounds. The ATP rate index is shown on the y-axis (logarithmic scale) in pAECs treated with the dyspiropyrrolidine derivatives. (B) The mitochondrial respiration profile was obtained from the OCR without (Vehicle) and with compounds under basal respiration conditions and after the addition of 1.5 μM oligomycin (OLIG), 1.0 μM FCCP, and a mixture of 0.5 μM rotenone plus antimycin A (ROT + AA). Inhibitor injections are shown as dotted lines. On the right are shown mitochondrial parameters (basal respiration, ATP production, proton leak, maximal respiration, spare respiratory capacity, and ATP production) in the absence (Vehicle) or in the presence of the compounds. (C) Effect of 11d and 12c molecules on H/R injured pAEC's morphology and viability. Scale bar (−) 100 μm. (D) H/R effect on mitochondrial respiration profile and on mitochondrial parameters without and with 1 μM of 11d. One-way ANOVA was used for statistical analysis by Students–Newman–Keuls’ test. Data expressed as column charts or points represent the mean ± SD (vertical bars) from 3 technical replicates carried out on 3 biological replicates. (∗∗∗∗) p value < 0.0001; (∗∗∗) p value < 0.001; (∗∗) p value < 0.01; (∗) p value < 0.05.Fig. 5

pAECs’ viability was investigated after the in vitro H/R injury. After H/R treatment cells appeared detached losing their typical phenotype, the presence of the molecules during H/R injury at the highest dose, induced the restoration of adherent monolayer in order to evaluate the protective effect of compounds. H/R injury caused a significative decrease of pAEC viability. At high dose tested all the derivatives compounds significantly restored cell viability. Only 1 μM 11d improved the parameter above the control group (Fig. 5C). As a consequence, we have characterized the effect of 11d by considering the mitochondrial bioenergetic metabolism. Notably, H/R-injured pAECs restored a respiratory profile similar to that of control in the presence of 1 μM 11d, in fact, the proton leak, maximal respiration and ATP production, decreased upon H/R compared to control, were restored to control values (Fig. 5D).

Cardioprotective effects of 11d and 12c in a model of cardiac reperfusion injury

2.7

Considering the ability of these molecules to inhibit PTP opening and to escape from cell death following H/R and improving cell viability (Fig. 4, Fig. 5), we investigated the effects of compounds 11d and 12c in an animal model of cardiac reperfusion injury. We isolated beating rat hearts and placed them in a Langendorff system, which was continuously perfused with Krebs-Henseleit buffer (KHB) bubbled with oxygen at 37 °C. The ex vivo protocol included stabilization of the heart for 20 min, and then retrograde perfusion was progressively stopped to induce 30 min of global ischemia followed by 1 h of reperfusion. After stabilization, the mean left ventricular developed pressure (LVDP) was 84.9 mmHg in the ischemia/reperfusion (I/R) untreated group and no differences were identified among the other experimental groups (Fig. 6A). Following reperfusion, the LVDP decreased to 34.3 mmHg, with a mean reduction of 50.6 mmHg, indicating successful induction of ischemia. Compounds were administered in the reperfusion phase during the first 10 min of reflow. The dose of 10 μM of each compound was selected based on previous experiments to identify the highest dose that could be perfused in the heart without toxicity. In isolated hearts, perfusion with a constant volume of 11d or 12c compounds, induced a significant increase in LVDP, by +44.7 and + 35 mmHg, respectively (Fig. 6A) indicating reduced diastolic stiffness, vasoconstriction, and deterioration of myocardial performance. Also, we recorded a strong decrease in end diastolic pressure (EDP) compared to I/R, by −22.7 and −25.3 mmHg, respectively (Fig. 6B).Fig. 6. Cardioprotective effects of dyspiropyrrolidine compounds in a model of cardiac reperfusion injury. (A) Left ventricular peak developed pressure recording of rat hearts during the experimental protocol composed by the stabilization, ischemia and reperfusion phases. The data were detected by using the Langendorff ex vivo protocol. (B) End-diastolic pressure recording of a rat heart in the stabilization, ischemia, and reperfusion phases. Histograms of statistical and representative kinetics data are reported. Kinetics and statistics are representative of 5 different animals per experimental conditions. (C) TUNEL assay for apoptosis evaluation under experimental conditions and representative images: (green) apoptotic nuclei detected by TUNEL enzyme. The analyzed heart tissue slices coming from rat hearts used during Langendorff ex vivo protocol. Scale bar (−) 50 μm. This is representative of the acquisitions of 4 different sections from 5 hearts. (D) Analysis of the tissue architecture at reperfusion time in the same conditions of (C). This data was calculated by quantifying Eosin staining as marker of the cytoplasm of muscle fibers. The same hearts processed in A and B were then processed in slices for the analysis of C and D. Scale bar (−) 100 μm. This is representative of the acquisitions of 4 different sections from 5 hearts. One-way ANOVA was applied for statistical analysis for all graphs reported in the figure; (∗∗∗∗) p value < 0.0001; (∗∗∗) p value < 0.001; (∗∗) p value < 0.01; (∗) p value < 0.05.Fig. 6

At the end of the procedure, cell death was analyzed in the hearts by terminal deoxynucleotidyl transferase-mediated deoxyuridine 5-triphosphate nick-end labeling (TUNEL) assay. In the I/R untreated group, 77.7% of the cardiomyocytes were TUNEL positive; however, this number was significantly reduced in the presence of 11d and 12c to 20.6% and 21% respectively (Fig. 6C). Macroscopic view of heart tissue slices colored with Hematoxylin and Eosin staining, revealed deep changes in the structure of the muscle following I/R with evident loss of tissue (Fig. 6D). Treatment of ischemic hearts with 11d and 12c compounds preserved the tissue architecture at reperfusion time. This data was calculated by quantifying Eosin staining as marker of cytoplasm.

These findings confirmed the ability of the small molecules inhibitors to protect against reperfusion damage in a cardiac preclinical model. In addition, if we compare this data with the protective effects of known PTP inhibitors (cyclosporine A, DCCD, oligomycin A, c.10) already investigated in our previous publications [18,46] and applied to the same experimental setting, 11d has higher beneficial effects.

11d and 12c interact with c subunit to exert PTP inhibition

2.8

Following previous hypothesis and considered the rationale of compounds synthesis, the human ATP synthase C-ring was selected as the target system to investigate the interaction of novel dispiranic derivatives. The C-ring is a membrane-embedded homomeric complex composed of 8–14 identical subunits; in humans, each subunit contains 136 amino acids, with 76 residues forming the transmembrane region arranged in a helix–loop–helix topology [[47], [48], [49]]. As the starting point for the in silico analysis, the cryo-electron microscopy structure of the human C-ring (PDB ID: 8H9J) was used. To explore potential ligand binding modes, we began by analyzing the well-characterized inhibitor oligomycin A. Structural alignment was carried out in the Maestro Schrödinger Software Suite, superimposing the human C-ring onto the X-ray crystal structure of the yeast c_10_ ring in complex with oligomycin A (PDB ID: 4F4S). This procedure enabled accurate placement of oligomycin A into the human C-ring model. To assess the stability of the resulting complex and its interaction profile, three independent 100 ns molecular dynamics (MD) simulations were performed. Root mean square deviation (RMSD) values for the Cα atoms were calculated for each replica, using the initial structure as a reference, to monitor the structural stability of the oligomycin-bound C-ring over time (Fig. S1). The result in panel S1A indicated that all complexes achieved stable conformations after approximately 20 ns with most showing RMSD values below 2.0 Å, reflecting good structural stability. RMSD analyses of oligomycin A (Fig. S1B in Supplementary information) were computed to monitor the structural stability of the C-ring–oligomycin complex over time. In the first replica, RMSD was low, fluctuating between ∼1 and 2 Å, and reached a stable plateau after approximately 20 ns, indicating minimal deviation from the initial structure. The second replica showed a higher RMSD, stabilizing around 4 Å after 20 ns, suggesting moderate conformational rearrangement. In contrast, the third replica displayed a more gradual increase in RMSD, reaching a plateau only after ∼60 ns at < 5 Å. This behavior indicates greater conformational flexibility, particularly in the early phase of the simulation, even if the system ultimately stabilized and all three replicas converged to stable trajectories.

Cluster and interaction analysis of oligomycin A binding

2.8.1

To characterize the binding interactions of oligomycin A within the human ATP synthase C-ring, cluster and interaction analyses were performed on the merged MD trajectories (Fig. S2–S3 in Supplementary information). The cluster analysis identified a single representative structure for the complex (Fig. S2 in Supplementary information). A close-up of the binding site is shown in Fig. S3A in Supplementary information, highlighting residues within 5 Å of the ligand. The corresponding interaction profile is presented in Fig. S3B in Supplementary information, with residue numbering based on the PDB: 8H9J sequence. Throughout the simulations, oligomycin A remained stably positioned within the canonical C-ring binding pocket [48,50]. Due to the use of the cryo-EM structure the residue indices in our model are shifted by approximately +60 residues relative to full-length annotations (i.e. Glu119 in our model corresponds to Glu59 in literature [48]). In the most representative structure extracted from the most populated cluster and supported by the interaction histogram across the full trajectory, hydrogen bond (H-bond) interactions were observed with Glu119 and Leu123, while hydrophobic interactions involved Phe124, Leu117, and Ala120. Additional hydrophobic contacts with Ala116 and Leu113 further stabilized the ligand within the inter-subunit cavity formed by adjacent C-ring helices. These interaction patterns are consistent with previously reported binding modes of oligomycin, where Glu59, Leu63, and Phe64 (based on full-length sequence numbering) are key contributors to ligand recognition [29,48,50]. Overall, the binding mode remains highly conserved, the preserved interaction network confirms the structural integrity of the oligomycin binding site and supports the use of this system as a reference model for evaluating the binding modes of novel ATP synthase inhibitors.

Novel dispiranic derivatives 11d and 12c molecular docking

2.8.2

Following the protocol described in the Materials and Methods section, molecular docking of the two lead compounds, 11d and 12c, was carried out against the human ATP synthase C-ring. This analysis aimed to determine whether the compounds bind within the canonical oligomycin A-binding pocket or adopt alternative binding modes. For compound 11d, four refined poses were obtained, while three poses were retained for compound 12c. The top-ranked pose of compound 11d, highlighted in blue in Fig. 7A, achieved a docking score of −2.64 kcal/mol. In comparison, the best pose for compound 12c, shown in green in the same Figure, shows a score of −1.33 kcal/mol. These top-ranked poses, further illustrated in detail in Fig. 7B, were selected for subsequent analysis based on their favorable interactions with key residues within the ATP-synthase binding site. Furthermore, this initial analysis revealed that both compounds occupy the same binding site as oligomycin within the ATP synthase C-ring.Fig. 7. Docking results of compound 11d (blue) and compound 12c (green). (A) Overall view of the human ATP synthase C-ring (grey ribbons) displaying the top-ranked docking poses for both compounds. (B) Zoomed-in view of the binding site, showing the ligands and surrounding amino acids within 5 Å. (C) RMSD values of 11d plotted against simulation time (ns). (D) RMSD values of 12c during the time (ns). (E) Detailed view of the dominant binding pose of compound 11d (shown in blue), derived from the most populated cluster during MD simulations. Amino acids within 5 Å radius of the ligand are shown in orange. (F) Interactions formed by compound 11d throughout the 250 ns MD simulations (combined from three replicates), including hydrophobic contacts, polar interactions, H-bond, and salt bridges with residues in the binding pocket. (G) Close-up view of the most representative pose of 12c (shown in green), extracted from the most populated cluster during MD simulations. Amino acid residues within 5 Å of the ligand are highlighted in orange. (H) Interaction profile of compound 12c derived from the merged 250 ns of MD simulations, showing the frequency of hydrophobic contacts, polar interactions, H-bond, and salt bridges with surrounding residues. (I) Superposition of clusters of oligomycin (pink), compounds 11d (blue) and 12c (green within the ATP synthase C-ring (in grey ribbons). Residues within 5 Å of the ligand are highlighted in orange.Fig. 7

Binding stability of 11d and 12c compound

2.8.3

To investigate the stability and binding of the selected ATP synthase inhibitors, MD simulations were performed for 250 ns in triplicate for each derivates–protein complex. Structural deviations were quantified using RMSD analysis, for monitoring conformational changes during simulations. The RMSD of the Cα atoms was calculated, using the initial structure as a reference, to evaluate the stability of the protein across all replicates. As shown in Fig. S4 in Supplementary information, all systems exhibited convergence after approximately 50 ns, with backbone RMSD values remaining below 2.0 Å, indicating high structural stability of the ATP synthase C-ring after ligand binding. In addition, derivate-protein RMSD values were computed to assess the stability of each compound within the binding site. As shown in Fig. 7, compound 11d displayed stable conformation across all replicas, with RMSD values ranging between 1 and 3 Å after equilibration (Fig. 7C). In contrast, compound 12c showed greater positional variability, particularly in the first replica, where RMSD values fluctuated between 3 Å and 7 Å (Fig. 7D). Despite this fluctuation after 50 ns the system reaches a plateau. To further investigate the structural evolution of the derivate-protein complexes during the simulation, cluster analysis was performed on the MD trajectories (represented in Fig. S5 in Supplementary information). This allowed identification of the most representative binding conformations adopted by compounds 11d and 12c during the simulations. For compound 11d, cluster analysis revealed a predominant binding pose, reported in Fig. 7E, that was consistently maintained across all simulation replicates. This pose closely mirrors the interaction pattern of oligomycin A. Notably, it includes a stable hydrogen bond (H-bond) and salt bridge with Glu119, as well as hydrophobic contacts with Leu117 and Ala120. Additionally, π–π stacking interactions were observed with aromatic residues Phe115 and Phe124, contributing to the stabilization of the ligand within the binding pocket (Fig. 7F). Similarly, compound 12c exhibited a predominant binding pose (Fig. 7G); different from the initial docking pose as reported in Fig. S6 in Supplementary information. Interaction analysis of this pose revealed the conservation of key contacts with Glu119 and Phe124, as observed for oligomycin and derivate 11d. Additionally, new interactions were established, including a π–π stacking interaction with Phe130, also implicated in oligomycin binding (Fig. S3), and a stabilizing π–cation interaction with Arg99 (Fig. 7H). These findings suggest that although compound 12c undergoes a positional shift, it ultimately adopts a stable and potentially favorable binding conformation within the binding site.

Oligomycin A and novel dispiranic derivatives

2.8.4

To assess the extent to which the compounds diverged from their initial docking conformations over time, we analyzed the most representative structures derived from MD simulations. As shown in Fig. 7I, both 11d and 12c remained within the canonical oligomycin-binding pocket of the C-ring of human ATP synthase. However, a marked difference in their binding behavior was observed. Compound 11d, despite the lower value of the docking score, maintained a binding orientation closely aligned with both its original docking position and the established oligomycin binding mode. In contrast, compound 12c underwent a shift in its binding position, diverging from both its initial docking configuration and the typical oligomycin orientation, while maintaining some interactions with key residues.

11d accumulates at mitochondria and interacts with c subunit

2.8.5

To investigate the subcellular localization of 11d, we performed subcellular fractionation to isolate whole cytoplasmic fraction and mitochondria from treated cells. The purity of the mitochondrial fraction was confirmed by Western blot analysis (Fig. 8A). Subsequent UPLC-MS/MS analysis was used to determine the relative abundance of 11d in this compartment compared to the cytosol. As shown in Fig. 8B, 11d was found to be predominantly localized in the mitochondria with an abundance 3.4-fold higher than in the cytosol. To directly observe binding within cells, we carried out a cellular thermal shift assay (CETSA [51,52]) following a protocol that measures ligand-induced thermal stabilization of the target protein, the c subunit of F_O_-ATP synthase. Since thermal shifts at elevated compound concentrations typically correspond to IC_50_ values and binding affinities, we used higher doses of compound 11d to confirm its interaction with the c subunit of F_O_-ATP synthase (Fig. 8C and D), observed between 59 °C and 74 °C. All together these findings demonstrate how 11d preferentially accumulates in mitochondria of treated cells and it thermally stabilizes the Csub protein (Fig. 8C, red line), an effect not seen in other mitochondrial proteins such as ATP5A (black line).Fig. 8(A) Evaluation of the purity of subcellular fractions by Western blot analysis evaluated through B-tubulin, VDAC and C-subunit primary antibodies incubation. H: homogenate; C: whole cytosol; M: mitochondria. (B) UPLC-MS/MS analysis on the amount of 11d in mitochondria and whole cytosol after 15min and 60min incubation. (C) Cellular Thermal Shift Assay (CETSA) in AC16 cells to evaluate the thermal stabilization between 53 °C and 80 °C of two mitochondrial proteins (ATP5A and the Csub) in the presence or absence of the 11d compound. Cells have been treated with either 10 μM 11d or vehicle for 15 min before performing the assay. Kinetics and statistics are reported from the analysis of 3 different experiments; Y axis measures the band intensity (in percentage) quantified from the Western blot analysis for which a representative result can be found in (D).Fig. 8

Discussion and conclusions

3

A lasting PTP opening is a complicated and dynamic phenomenon whose effect is attributable to extensive cell death in a plethora of pathologies. Current literature considers the PTP opening a maladaptive rearrangement of the ATP synthase enzyme under pathological conditions, a key complex in cell life already used as crucial target for several therapies in the past (i.e., the selectively removal of strains of microorganisms by targeting the Csub and mycobacterial subunit ε with antimicrobials [53]). Moreover, PTP retains some ATP synthase proteins in its structure and as many modulate it to exert deadly functions [13,46,[54], [55], [56], [57]]. Among them, and in addition to the best known CypD [22], in the last decade Csub binding and inhibition has been proposed as new target in limiting cell death following IRI [18] and as landmark for further drug discovery [12,58,59].

On the other side of this coin, it must be said that there is no a single consensus on the exact composition of this complex [60]; this does not make easier the identification of specific candidates to be pharmacologically targeted. However, it remains ample and shared the evidence on the contribution of the PTP in the development of myocardial damage at reperfusion time in the preclinical setting [10,14]; this issue has been further validated through recent investigations carried out directly in human patients [20,21]. The cure for the residual damage following revascularization remains a hot topic in the clinic, and is our opinion that the advancements in the translational research must propose new compounds to be refined and tested in future clinical trials before the target is finally abandoned.

Taking into account all these considerations and taking advantage from current biotechnologies which are able to finely measure the channel opening and the related damage, we focused the study to looking for new small molecule able to powerfully inhibit the whole channel opening by minimizing the resulting damage as much as possible, without performing specific assays to unveil the target.

This method allowed us to identify for the first time a novel small molecule inhibitor able to counteract the PTP opening by about 70% with an almost total recovery of cell viability parameters in cardiomyocytes and endothelial cells subjected to in vitro H/R setting and the cardiac performance in preclinical models of cardiac I/R. Indeed, especially 11d, but also 12c administration at reperfusion time, fully recovered the LVDP to values achieved before ischemia induction and also preserved the EDP; this reflected significant reductions in diastolic stiffness, vasoconstriction, and deterioration of myocardial performance. This phenotype was accompanied by the significant reduction of cell death and the preservation of the amount of muscle fibers and their architecture.

The mitochondrial ATP synthase can change from an energy-producing to an energy-dissipating molecular mechanism [61]. The dual function is related to the cation cofactor that supports the enzyme catalysis. Indeed, the natural cofactor Mg^2+^ bound to the catalytic sites, performs the dual function of ATP production or hydrolysis [62] coupled to H^+^ translocation. Otherwise, when the concentration of Ca^2+^ in the mitochondria abruptly increases, the substitution of Mg^2+^ in the catalytic sites with Ca^2+^ establishes the monofunctional feature of ATP hydrolysis that powers the H^+^ pump [63]. Moreover, Ca^2+^-activated ATP synthase is most likely the molecular component that supports the features of PTP formation [56,64,65]. The dyspiropyrrolidine compounds did not affect the enzyme activated with Mg^2+^, whereas selectively inhibited the hydrolytic activity of Ca^2+^-activated ATP synthase. In particular, 11d showed a higher uncompetitive inhibitory efficiency than 11c and 12c and the binding site of 11d did not overlap the region of interaction of Oligomycin A and DCCD on Csub. The ability to modulate in a different way ATP synthase catalysis depending on cation cofactors was corroborated by the decrease in the intensity of CRC on PTP opening.

The effect of these compounds on AC16 and pAECs’ cellular metabolism results highlighted by the absence of toxic effects on ATP production by oxidative phosphorylation and glycolysis. In particular, the compounds did not affect the profile and key bioenergetic parameters of cellular respiration of all cellular models tested. Importantly, by considering dyspiropyrrolidine compounds on cell viability investigated after the in vitro H/R injury, we can assert that the molecules have a protective effect. Of all the compounds tested, 11d proved to be the best performing in recovering the cell viability after H/R. Since the negative effect of H/R on cell viability may be attributed to an impaired bioenergetics of cell metabolism, we have confirmed the mitochondria as biological target of protective effect of compound 11d. Indeed, mitochondrial respiration reduced by H/R was restored at values similar to control in the presence of 11d, improving both maximum respiration and ATP production, the key parameters supporting the capability of the cells to respond to changes in energetic demand and indicates the fitness of the cells. The compound did not alter mitochondrial and cellular parameters in resting conditions; although not significant, we reported only a slight decrease in the total mitochondrial volume of cells treated with 11d; this parameter will be better investigated in the future, especially in 11d derivatives, as the decrease of mitochondrial network may correspond to mitophagy induction, another cardioprotective pathway in charge of mitochondria [66]. Besides 11d, also 12c has been taken into consideration for functional studies due to the significant result in PTP inhibition when compared to its derivative, 11c. However, 12c did not confirmed the same pattern once H/R is induced, producing values of PTP inhibition, cell viability and cell death more similar to 11c than 11d. Nevertheless, when used at increased concentrations in isolated rat hearts, 12c triggered cardioprotection in a manner similar to 11d. Overall, the better biological profile of 11d may be addressed to some additional peculiarities observed following H/R, like a strong action against ROS and a mild and transient decrease in Ca^2+^ accumulation in mitochondria which are other determinants of PTP opening. On the other hand, although preserving a modest cytoprotective role, 11c use in future studies should be avoided due to some negative effects on ATP production and mitochondrial membrane potential.

In this study, we also employed a molecular docking and MD simulation approach to investigate the binding mode of two novel derivatives, compounds 11d and 12c, targeting the ATP synthase C-ring, a well-established site of action for the inhibitor oligomycin A. This data has been confirmed also with an in vitro approach. Starting from the human ATP synthase structure (PDB ID: 8H9J), we used the known binding mode of oligomycin A, confirmed via superposition with its co-crystallized yeast homolog (PDB ID: 4F4S) and MD simulations, to guide the setup and analysis of ligand interactions within the canonical inter-subunit binding cavity. Initial docking studies positioned both compounds 11d and 12c within the oligomycin A binding pocket. Subsequent 250 ns MD simulations performed in triplicate allow the evaluation of the dynamic stability of these ligand–protein complexes. RMSD analysis revealed that derivate 11d maintained a stable binding conformation across all replicates, consistent with the low Cα and ligand RMSD values observed. Moreover, cluster and interaction analyses showed that 11d preserved key contacts characteristic of oligomycin, including a persistent H-bond and salt bridge with Glu119 (i.e. Glu59 in full-length numbering), and hydrophobic interactions with Leu117, Ala120, and Phe124, underscoring its compatibility with the known binding mode. Conversely, compound 12c exhibited higher ligand RMSD values and adopted a shifted binding orientation after approximately 50 ns of simulation in all replicates. This alternative pose, although still located within the oligomycin binding site, displayed altered interaction patterns, including a new π-cation contact with Arg99 and additional aromatic stacking with Phe130. These findings suggest a more flexible binding mode for 12c, potentially affecting its retention and affinity. Nevertheless, both compounds engaged Glu119 and Phe124, indicating partial overlap with the conserved binding features of oligomycin. Overall, compound 11d emerges as a promising candidate due to its stable binding profile and conservation of key interactions, while the distinct binding adaptation observed for 12c highlights the potential for alternative binding modes from those of oligomycin. The results from the molecular docking were extended in in vitro experiments through which we confirmed the preferential accumulation of 11d in mitochondria, the putative site of action, and its ability to interact with Csub to exert the biological function.

Summary

4

The identification of compound 11d as a potent desensitizer of the mitochondrial PTP highlights its therapeutic potential in a broad spectrum of pathophysiological conditions characterized by mitochondrial Ca^2+^ overload, oxidative stress and bioenergetic failure. Beyond its robust cardioprotective effects in models of hypoxia and I/R injury, the ability of 11d to preserve mitochondrial function, limit ROS production and prevent maladaptive PTP opening suggests a possible application in chronic cardiovascular disorders such as hypertension [67], where sustained mitochondrial stress and endothelial dysfunction contribute to disease progression. By selectively modulating Ca^2+^-dependent ATP synthase activity without impairing physiological ATP production, 11d represents a promising lead compound for the development of mitochondria-targeted therapies aimed at reducing acute and chronic tissue damage associated with pathological PTP activation.

Experimental section

5

Cell lines and culture

5.1

The AC16 human cardiomyocyte cell line was grown in DMEM/F12 containing 2 mM l-glutamine, 12.5% FBS and 1x PS solution. All experiments were performed using AC16 between passages 5–10, seeded and routinely cultured in Petri dishes in a 5% CO_2_ atmosphere and 37 °C. Differentiated Human Cardiac Myocyte (HCM) was grown in Cardiac Myocyte Medium in similar conditions. Primary cell cultures of porcine Aortic Endothelial Cells (pAECs) were isolated and maintained as previously described [68]. All experiments were performed using pAECs between passages 3–10. The cells were seeded and routinely cultured in T25 or T75 primary culture flasks (2 × 10^4^ cells/cm^2^) in a human Endothelial Serum-Free Medium (hESFM) with 5% FBS and 1 × antibiotic/antimycotic solution in a 5% CO_2_ atmosphere and 38.5 °C.

Chemical details

5.2

Reaction progress and product mixtures were monitored by thin-layer chromatography (TLC) on silica gel (precoated F254 Macherey-Nagel plates) and visualized with a UV lamp (254 nm light source). The organic solutions from extractions were dried over anhydrous sodium sulphate. Chromatography was performed on Merck 230−400 mesh silica gel or using Isolera One (Biotage Sweden). ^1^H, ^13^C, DEPT NMR spectra were recorded on a VARIAN Mercury Plus 400 MHz spectrometer. Chemical shifts (δ) are reported in parts per million (ppm) using the peak of deuterated solvents as an internal standard and coupling constants (J) are reported in Hertz. Splitting patterns are designed as follows: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; and b, broad. Melting points for purified products were determined in a glass capillary on a Stuart Scientific electrothermal apparatus SMP3 and are uncorrected. Mass spectra were recorded by an ESI single-quadrupole mass spectrometer Waters ZQ 2000 (Waters Instruments UK). For analytical controls, Beckmann System Gold 168 HPLC was used with an LC column Kinetex 5-μm EVO C18 100 Å (250 Å∼ 4.6 mm) and a variable-wavelength UV detector fixed to 220 nm. The analysis was conducted using two solutions, A and B, containing 100:0.1H_2_O:TFA and 40:60:0.1H_2_O:CH_3_CN:TFA, respectively, with a gradient elution of 0−50% solution B over 30 min. The purity of all compounds was determined by HPLC and was greater than 95%. Mass spectrometric detection was carried out using an electrospray interface (ESI) operated in positive ionization mode. Nitrogen was used as a desolvation gas at a 300 L/h flow rate with the desolvation temperature set at 150 °C and the source temperature set at 130 °C.

General procedure for the synthesis of intermediates 3a-f

5.3

To a solution of the appropriate aldehyde (2 eq) in EtOH (10 ml), N-benzyl piperidone (1 eq) was added, followed by the NaOH solution (1 eq) dropwise. The reaction was stirred for 1-3 h at room temperature and monitored by mass analysis. Once the reaction was completed, a yellowish precipitate was clearly visible: after filtration under reduced pressure, the solid was isolated and analyzed as the desired product. These compounds were used without need of further purifications.

(3E,5E)-1-benzyl-3,5-bis(4-fluorobenzylidene)piperidin-4-one (3a)

5.3.1

MS (ESI): [M+H]^+^ = 402,16; Yield: 62%;^1^H NMR (400 MHz, Chloroform-d) δ 7.79 (s, 1H), 7.39 – 7.22 (m, 5H), 7.18 – 7.02 (m, 2H), 3.85 (d, J = 1.9 Hz, 2H), 3.73 (s, 1H). ^13^C NMR (101 MHz, Chloroform-d) δ 187.55, 164.26, 161.76, 137.22, 135.66, 132.95, 132.44, 132.35, 131.43, 129.11, 128.53, 127.63, 115.97, 115.75, 61.50, 54.32. ^19^F NMR (376 MHz, Chloroform-d) δ −110.89.

1-benzyl-3,5-bis((E)-4-nitrobenzylidene)piperidin-4-one (3b)

5.3.2

MS (ESI): [M+H]^+^ = 455,17; Yield: 47%; ^1^H NMR (400 MHz, Chloroform-d) δ 8.10 – 8.02 (m, 4H), 7.83 – 7.75 (m, 4H), 7.38 – 7.22 (m, 7H), 3.83 (s, 4H), 3.70 (d, J = 0.8 Hz, 2H).

(3E,5E)-1-benzyl-3,5-bis(4-chlorobenzylidene)piperidin-4-one (3c)

5.3.3

MS (ESI): [M+H]^+^ = 434,16; Yield: 77%; ^1^H NMR (400 MHz, Chloroform-d) δ 7.76 (s, 1H), 7.40 – 7.32 (m, 2H), 7.28 (d, J = 0.5 Hz, 1H), 3.87 – 3.80 (m, 2H), 3.73 (s, 1H). ^13^C NMR (101 MHz, Chloroform-d) δ 135.50, 135.20, 133.66, 131.65, 129.09, 128.98, 128.57, 127.69, 61.53, 54.34.

(3E,5E)-1-benzyl-3,5-bis(4-cyanobenzylidene)piperidin-4-one (3d)

5.3.4

MS (ESI): [M+H]^+^ = 416,16; Yield: 32%;^1^H NMR (400 MHz, Chloroform-d) δ 7.77 (s, 1H), 7.70 – 7.64 (m, 2H), 7.46 – 7.37 (m, 2H), 7.33 – 7.20 (m, 3H), 3.82 (d, J = 1.8 Hz, 2H), 3.72 (s, 1H). ^13^C NMR (101 MHz, Chloroform-d) δ 187.00, 139.52, 136.82, 135.61, 134.77, 132.38, 130.60, 129.01, 128.78, 128.62, 127.84, 118.53, 112.46, 61.50, 54.16.

(3E,5E)-1-benzyl-3,5-bis(benzylidene)piperidin-4-one (3e)

5.3.5

MS (ESI): [M+H]^+^ = 366,18; Yield: 60%; ^1^H NMR (400 MHz, Chloroform-d) δ 7.84 (s, 1H), 7.45 – 7.32 (m, 6H), 7.32 – 7.16 (m, 3H), 3.89 (s, 2H), 3.73 (s, 1H). ^13^C NMR (101 MHz, Chloroform-d) δ 136.77, 135.32, 130.48, 130.44, 129.14, 129.10, 128.79, 128.66, 128.54, 128.47, 127.52, 61.47, 54.47.

(3E,5E)-1-benzyl-3,5-bis(naphtylidene)piperidin-4-one (3f)

5.3.6

MS (ESI): [M+H]^+^ = 466,21; Yield: 32%; ^1^H NMR (400 MHz, Chloroform-d) δ 8.00 (t, J = 1.6 Hz, 1H), 7.91 – 7.80 (m, 4H), 7.61 – 7.42 (m, 3H), 7.35 – 7.11 (m, 3H), 4.01 (d, J = 1.8 Hz, 2H), 3.75 (s, 1H). ^13^C NMR (101 MHz, Chloroform-d) δ 187.80, 137.34, 136.89, 133.68, 133.31, 133.19, 132.94, 130.62, 129.17, 128.63, 128.49, 128.28, 127.79, 127.55, 127.21, 126.67, 61.60, 54.66.

Synthesis of compounds 5a-f

5.4

(1R,4R,5R,7S,8S,11S)-13-benzyl-1,8-bis(4-fluorophenyl)-4,11-dihydroxy-2,9-dithia-13-azadispiro[4.1.47.35]tetradecan-6-one (5a)

5.4.1

Compound 3a (1 eq) and 1,4-dithiane-2,5-diol (2 eq) were solubilized in absolute ethanol (5 ml). The reaction mixture was maintained at 0 °C using an ice bath. Once the temperature was reached, TEA (2 eq) was added dropwise and the reaction was stirred at room temperature for 10 min. Then, it was set at 90 °C for 3 h and monitored by TLC (A1P4). Once the reaction was completed, the solvent was evaporated and the crude residue was purified by flash silica gel chromatography with an eluent mixture of ethyl acetate and petroleum ether (A1P4). After evaporation, a white solid was obtained. Exact mass (m/z): 554.16278; Yield: 32%; m.p.: 130-132 °C; HPLC Rt: 24,92; ^1^H NMR (400 MHz, DMSO‑d6) δ 7.49 – 7.21 (m, 8H), 7.15 – 6.99 (m, 4H), 5.53 (s, 2H), 5.22 (d, J = 4.6 Hz, 2H), 3.44 (s, 4H), 3.14 (dd, J = 12.2, 4.8 Hz, 2H), 2.73 (dd, J = 14.8, 12.0 Hz, 2H), 1.56 (d, J = 12.1 Hz, 2H). ^13^C NMR (101 MHz, DMSO‑d6) δ 203.59, 162.45, 160.02, 138.45, 131.82, 131.56, 131.48, 128.80, 128.35, 127.33, 114.61, 114.40, 79.14, 68.18, 60.95, 54.53, 50.70, 36.28.

(1R,4R,5R,7S,8S,11S)-13-benzyl-4,11-dihydroxy-1,8-bis(4-nitrophenyl)-2,9-dithia-13-azadispiro[4.1.47.35]tetradecan-6-one (5b)

5.4.2

Compound 3b (1 eq) and 1,4-dithiane-2,5-diol (2 eq) were solubilized in absolute ethanol (5 ml). The reaction mixture was maintained at 0 °C using an ice bath. Once the temperature was reached, TEA (2 eq) was added dropwise and the reaction was stirred at room temperature for 10 min. Then, it was set at 90 °C for 3 h and monitored by TLC (A1P4). Once the reaction was completed, the solvent was evaporated and the crude residue was purified by flash silica gel chromatography with an eluent mixture of ethyl acetate and petroleum ether (A1P4). After evaporation, a with pale yellow solid was obtained.

Exact mass (m/z): 586.10413; Yield: 90%; m.p.:160-162 °C; HPLC Rt: 24,50; ^1^H NMR (400 MHz, DMSO‑d6) δ 7.49 – 7.21 (m, 8H), 7.15 – 6.99 (m, 4H), 5.53 (s, 2H), 5.22 (d, J = 4.6 Hz, 2H), 3.44 (s,4H), 3.14 (dd, J = 12.2, 4.8 Hz, 2H), 2.73 (dd, J = 14.8, 12.0 Hz, 2H), 1.56 (d, J = 12.1 Hz, 2H). ^13^C NMR (101 MHz, DMSO‑d6) δ 203.59, 162.45, 160.02, 138.45, 131.82, 131.56, 131.48, 128.80, 128.35, 127.33, 114.61, 114.40, 79.14, 68.18, 60.95, 54.53, 50.70, 36.28.

13-benzyl-1,8-bis(4-chlorophenyl)-4,11-dihydroxy-2,9-dithia-13-azadispiro[4.1.47.35]tetradecan-6-one (5c)

5.4.3

Compound 3c (1 eq) and 1,4-dithiane-2,5-diol (2 eq) were solubilized in absolute ethanol (5 ml). The reaction mixture was maintained at 0 °C using an ice bath. Once the temperature was reached, TEA (2 eq) was added dropwise and the reaction was stirred at room temperature for 10 min. Then, it was set at 90 °C for 3 h and monitored by TLC (A1P4). Once the reaction was completed, the solvent was evaporated and the crude residue was purified by flash silica gel chromatography with an eluent mixture of ethyl acetate and petroleum ether (A1P4). After evaporation, a white solid was obtained. MS (ESI): [M+H]^+^ = 585,09; Yield:83%; ^1^H NMR (400 MHz, DMSO‑d6) δ 7.45 – 7.19 (m, 1H), 5.51 (s, 2H), 5.20 (t, J = 4.4 Hz, 2H), 5.09 (d, J = 4.1 Hz, 2H), 3.43 (s, 2H), 3.34 (s, 2H), 3.14 (dd, J = 12.1, 4.7 Hz, 2H), 2.72 (dd, J = 25.8, 12.1 Hz, 2H), 1.58 (d, J = 12.1 Hz, 2H). ^13^C NMR (101 MHz, dmso) δ 203.58, 138.47, 134.81, 131.94, 131.38, 128.79, 128.35, 127.77, 127.30, 79.21, 68.26, 60.85, 54.49, 50.85, 36.66.

4,4'-((1R,4R,5R,7S,8S,11S)-13-benzyl-4,11-dihydroxy-6-oxo-2,9-dithia-13 azadispiro[4.1.47.35]tetradecane-1,8-diyl)dibenzonitrile (5d)

5.4.4

Compound 3d (1 eq) and 1,4-dithiane-2,5-diol (2 eq) were solubilized in absolute ethanol (5 ml). The reaction mixture was maintained at 0 °C using an ice bath. Once the temperature was reached, TEA (2 eq) was added dropwise and the reaction was stirred at room temperature for 10 min. Then, it was set at 90 °C for 3 h and monitored by TLC (A1P4). Once the reaction was completed, the solvent was evaporated and the crude residue was purified by flash silica gel chromatography with an eluent mixture of ethyl acetate and petroleum ether (A1P4). After evaporation, a white solid was obtained. Exact mass (m/z): 568.17224; Yield: 79%; m.p.: 181-183 °C; HPLC Rt: 21,117; ^1^H NMR (400 MHz, DMSO‑d6) δ 7.45 – 7.19 (m, 8H), 5.51 (s, 4H), 5.20 (t, J = 4.4 Hz, 2H), 5.09 (d, J = 4.1 Hz, 2H), 3.43 (s, 1H), 3.34 (s, 4H), 3.14 (dd, J = 12.1, 4.7 Hz, 2H), 2.72 (dd, J = 25.8, 12.1 Hz, 2H), 1.58 (d, J = 12.1 Hz, 2H). ^13^C NMR (101 MHz, DMSO‑d6) δ 203.58, 138.47, 134.81, 131.94, 131.38, 128.79, 128.35, 127.77, 127.30, 79.21, 68.26, 60.85, 54.49, 50.85, 36.46.

(1R,4R,5R,7S,8S,11S)-13-benzyl-4,11-dihydroxy-1,8-diphenyl-2,9-dithia-13-azadispiro[4.1.47.35]tetradecan-6-one (5e)

5.4.5

Compound 3e (1 eq) and 1,4-dithiane-2,5-diol (2 eq) were solubilized in absolute ethanol (5 ml). The reaction mixture was maintained at 0 °C using an ice bath. Once the temperature was reached, TEA (2 eq) was added dropwise and the reaction was stirred at room temperature for 10 min. Then, it was set at 90 °C for 3 h and monitored by TLC (A1P4). Once the reaction was completed, the solvent was evaporated and the crude residue was purified by flash silica gel chromatography with an eluent mixture of ethyl acetate and petroleum ether (A1P4). After evaporation, a white solid was obtained. Exact mass (m/z): 518.18256; Yield: 92%; m.p.: decomp 202-204 °C; HPLC Rt: 24,85; ^1^H NMR (400 MHz, DMSO‑d6) δ 7.75 – 7.59 (m, 8H), 7.47 – 7.39 (m, 4H), 7.40 – 7.26 (m, 2H), 7.26 – 7.14 (m, 2H), 5.57 (s, 2H), 5.22 (t, J = 4.4 Hz, 2H), 5.16 (d, J = 4.1 Hz, 2H), 3.40 (s, 1H), 3.34 (s, 4H), 3.18 (dd, J = 12.1, 4.5 Hz, 2H), 2.79 (d, J = 12.1 Hz, 2H), 2.70 (d, J = 12.1 Hz, 2H), 1.56 (d, J = 12.2 Hz, 2H). ^13^C NMR (101 MHz, DMSO‑d6) δ 203.34, 141.97, 131.69, 130.66, 128.76, 128.29, 118.51, 110.10, 79.25, 68.59, 60.73, 54.60, 51.32, 36.66.

(1R,4R,5R,7S,8S,11S)-13-benzyl-4,11-dihydroxy-1,8-di(naphthalen-2-yl)-2,9-dithia-13-azadispiro[4.1.47.35]tetradecan-6-one (5f)

5.4.6

Compound 3f (1 eq) and 1,4-dithiane-2,5-diol (2 eq) were solubilized in absolute ethanol (5 ml). The reaction mixture was maintained at 0 °C using an ice bath. Once the temperature was reached, TEA (2 eq) was added dropwise and the reaction was stirred at room temperature for 10 min. Then, it was set at 90 °C for 3 h and monitored by TLC (A1P4). Once the reaction was completed, the solvent was evaporated and the crude residue was purified by flash silica gel chromatography with an eluent mixture of ethyl acetate and petroleum ether (A1P4). After evaporation, a with white solid was obtained. Exact mass (m/z): 618.2124; Yield: 57%; m.p.: decomp. 199.201 °C; ^1^H NMR (400 MHz, DMSO‑d6) δ 7.47 – 6.99 (m, 8H), 5.56 (s, 4H), 5.25 (t, J = 4.4 Hz, 2H), 5.13 (d, J = 4.0 Hz, 2H), 3.42 (s, 4H), 3.34 (s, 2H), 3.14 (dd, J = 12.1, 4.7 Hz, 2H), 2.83 – 2.67 (m, 2H), 1.61 (s, 2H). ^13^C NMR (101 MHz, DMSO‑d6) δ 203.57, 138.56, 135.83, 129.61, 128.79, 128.37, 127.73, 127.26, 79.40, 68.28, 61.01, 54.57, 51.45, 36.24.

Synthesis of the dispiropyrrolidinic compounds 11a-d and 12b-d

5.5

1′-methyldispiro[cyclohexane-1,3′-pyrrolidine-2′,3″-indoline]-2,2″-dione (11a)

5.5.1

To a solution of isatin (1 eq, 3.4 mmol) in MeOH (15 ml), sarcosine (1.5 eq, 5.1 mmol) and cyclohexan-1-one (1 eq, 3.4 mmol) were added. The reaction was stirred at 90 °C under reflux for 10 h. The reaction was monitored through MS analysis and TLC (A1P1). Upon completion of the reaction, this was quenched with NaHCO_3_ sat and deionized water. The organic solvent was evaporated and the aqueous phase was extracted three times with ethyl acetate (40 ml). The organic phases, taken together, after treatment with anhydrous sodium sulphate, were evaporated and an amorphous orange solid was obtained. The crude product was then purified by flash silica gel chromatography with eluent mixture of ethyl acetate and petroleum ether (A1P1). After evaporation of the solvent, a pale pink solid was obtained. Exact mass (m/z): 285.1593; Yield: 51%; mp: 144-146 °C; HPLC Rt: 11,55; ^1^H NMR (400 MHz, Chloroform-d) δ 8.86 (s, 1H), 7.32 – 7.09 (m, 1H), 7.09 – 6.93 (m, 2H), 6.90 (d, J = 7.8 Hz, 1H), 3.26 (h, J = 3.9, 3.4 Hz, 2H), 2.68 (ddd, J = 12.7, 7.5, 5.3 Hz, 1H), 2.44 (dt, J = 13.5, 3.9 Hz, 1H), 2.22 (dd, J = 14.6, 4.4 Hz, 1H), 2.08 (s, 3H), 2.03 – 1.91 (m, 1H), 1.77 (ddd, J = 14.9, 11.5, 7.0 Hz, 1H), 1.67 (ddd, J = 14.7, 10.9, 3.7 Hz, 2H), 1.59 – 1.45 (m, 2H), 1.43 – 1.32 (m, 1H). ^13^C NMR (101 MHz, Chloroform-d) δ 212.01, 179.02, 141.71, 135.09, 129.55, 128.33, 127.09, 123.44, 122.75, 115.29, 110.23, 62.83, 52.14, 41.30, 35.85, 35.64, 34.10, 27.70, 25.44, 21.69.

1′,5-dimethyldispiro[cyclohexane-1,3′-pyrrolidine-2′,3″-indoline]-2,2″-dione (11b)

5.5.2

To a solution of isatin (1 eq, 3.4 mmol) in MeOH (15 ml), sarcosine (1.5 eq, 5.1 mmol) and 4-methylcyclohexan-1-one (1 eq, 3.4 mmol) were added. The reaction was stirred at 90 °C under reflux for 10 h. The reaction was monitored through MS analysis and TLC (A1P1). Upon completion of the reaction, this was quenched with NaHCO_3_ sat and deionized water. The organic solvent was evaporated and the aqueous phase was extracted three times with ethyl acetate (40 ml). The organic phases, taken together, after treatment with anhydrous sodium sulphate, were evaporated and an amorphous orange solid was obtained. The crude product was then purified by flash silica gel chromatography with eluent mixture of ethyl acetate and petroleum ether (A1P1). After evaporation of the solvent, a pale pink solid was obtained. Exact mass (m/z): 299.1753; mp: 110-112 °C; Yield: 51%; HPLC Rt: 13,37; ^1^H NMR (400 MHz, Chloroform-d) δ 9.01 (s, 1H), 7.29 – 7.09 (m, 1H), 7.05 – 6.75 (m, 3H), 3.29 – 3.13 (m, 2H), 2.83 – 2.69 (m, 1H), 2.60 (dt, J = 14.6, 3.1 Hz, 1H), 2.26 – 2.11 (m, 1H), 2.03 (s, 3H), 1.90 – 1.62 (m, 2H), 1.40 – 1.06 (m, 3H), 0.79 (d, J = 6.4 Hz, 3H). ^13^C NMR (101 MHz, CDCl_3_) δ 211.40, 179.40, 141.51, 129.51, 127.51, 126.93, 122.64, 110.30, 76.34, 62.84, 52.07, 45.37, 41.70, 35.71, 34.66, 34.54, 28.66, 21.86.

1′-methyl-3″,4″-dihydro-1″H-dispiro[indoline-3,2′-pyrrolidine-3′,2″-naphthalene]-1″,2-dione (11c)

5.5.3

To a solution of isatin (1 eq, 3.4 mmol) in MeOH (15 ml), sarcosine (1.5 eq, 5.1 mmol) and 3,4-dihydronaphthalen-1(2H)-one (1 eq, 3.4 mmol) were added. The reaction was stirred at 90 °C under reflux for 10 h. The reaction was monitored through MS analysis and TLC (A1P1). Upon completion of the reaction, this was quenched with NaHCO_3_ sat and deionized water. The organic solvent was evaporated and the aqueous phase was extracted three times with ethyl acetate (40 ml). The organic phases, taken together, after treatment with anhydrous sodium sulphate, were evaporated and an amorphous orange solid was obtained. The crude product was then purified by flash silica gel chromatography with eluent mixture of ethyl acetate and petroleum ether (A1P1). After recrystallization with a mixture of ethyl ether/petroleum ether (3:7), a dark yellow solid was obtained. Exact mass (m/z): 333.1594; Yield: 52%; mp: 194-196 °C; HPLC Rt: 14,38; ^1^H NMR (400 MHz, Chloroform-d) δ 10.20 (s, 1H), 8.35 – 8.01 (m, 4H), 7.62 (ddd, J = 8.6, 7.3, 1.5 Hz, 1H), 7.41 – 6.43 (m, 2H), 4.39 (s, 3H), 3.49 (s, 4H), 2.79 – 2.02 (m, 9H). ^13^C NMR (101 MHz, CDCl_3_) δ 210.13, 162.80, 152.08, 143.47, 138.59, 135.13, 133.27, 131.81, 128.45, 127.94, 126.87, 125.20, 123.52, 118.73, 115.15, 114.57, 109.84, 57.72, 51.38, 27.74, 26.04.

1′-methyl-2H-dispiro[acenaphthylene-1,3′-pyrrolidine-2′,3″-indoline]-2,2″-dione (11d)

5.5.4

To a solution of isatin (1 eq, 3.4 mmol) in MeOH (15 ml), sarcosine (1.5 eq, 5.1 mmol) and acenaphtenone (1 eq, 3.4 mmol) were added. The reaction was stirred at 90 °C under reflux for 10 h. The reaction was monitored through MS analysis and TLC (A1P1). Upon completion of the reaction, this was quenched with NaHCO_3_ sat and deionized water. The organic solvent was evaporated and the aqueous phase was extracted three times with ethyl acetate (40 ml). The organic phases, taken together, after treatment with anhydrous sodium sulphate, were evaporated and an amorphous orange solid was obtained. The crude product was then purified by flash silica gel chromatography with eluent mixture of ethyl acetate and petroleum ether (A1P1). After recrystallization with a mixture of ethyl ether/petroleum ether (3:7), an orange solid was obtained. Exact mass (m/z): 355.1440; Yield: 15% to 30%; mp: 210-212 °C; HPLC Rt: 13,70; ^1^H NMR (400 MHz, Chloroform-d) δ 7.90 (d, J = 8.1 Hz, 1H), 7.85 (d, J = 7.0 Hz, 1H), 7.70 (dd, J = 7.7, 5.8 Hz, 2H), 7.56 (dd, J = 8.1, 7.0 Hz, 1H), 7.47 (dd, J = 8.4, 7.0 Hz, 1H), 6.91 (td, J = 7.7, 1.3 Hz, 1H), 6.76 (td, J = 7.6, 1.1 Hz, 1H), 6.46 (d, J = 7.7 Hz, 1H), 3.84 (s, 1H), 3.64 (td, J = 9.5, 8.6, 4.1 Hz, 1H), 3.06 (td, J = 11.9, 5.1 Hz, 1H), 2.46 (ddd, J = 12.8, 8.5, 4.2 Hz, 1H), 2.30 (s, 4H). ^13^C NMR (101 MHz, Chloroform-d) δ 204.87, 177.95, 142.05, 141.52, 132.66, 132.53, 131.68, 131.28, 130.34, 129.92, 129.44, 128.94, 128.68, 128.22, 128.12, 126.65, 124.89, 122.66, 122.53, 121.70, 121.58, 109.67, 109.60, 79.00, 63.48, 52.45, 35.98, 31.73.

5″-fluoro-1′-methyldispiro[cyclohexane-1,3′-pyrrolidine-2′,3″-indoline]-2,2″-dione (12b)

5.5.5

To a solution of 5-Fluoro-isatine (1 eq, 3.03 mmol) in MeOH (15 ml), sarcosine (1.5 eq, 4.54 mmol) and cyclohexanone (1 eq, 3.03 mmol) were added sequentially. The reaction was then stirred at 90 °C under reflux for 10 h. The reaction was monitored by mass analysis and TLC (A1P1). Upon the completion of the reaction, this was quenched with NaHCO_3_ sat and deionized water. The organic solvent was evaporated and the aqueous phase was extracted three times with ethyl acetate (40 ml). The organic phases, taken together, after treatment with anhydrous sodium sulphate, were evaporated and an amorphous orange solid was obtained. The crude product was then purified by flash silica gel chromatography with eluent mixture of ethyl acetate and petroleum ether (A1P1), obtaining a white solid. Exact mass (m/z): 303.1502; Yield: 72%; mp: 153-155 °C; HPLC Rt: 11,27; ^1^H NMR (400 MHz, Chloroform-d) δ 8.56 (s, 1H), 6.95 (td, J = 8.7, 2.6 Hz, 1H), 6.89 – 6.76 (m, 2H), 3.33 – 3.12 (m, 2H), 2.62 (ddd, J = 12.8, 7.4, 5.7 Hz, 1H), 2.43 – 2.22 (m, 2H), 2.07 (s, 3H), 1.99 (ddd, J = 12.6, 8.4, 6.8 Hz, 1H), 1.83 (ddd, J = 15.0, 11.1, 7.0 Hz, 1H), 1.75 – 1.55 (m, 2H), 1.45 – 1.37 (m, 1H). ^13^C NMR (101 MHz, Chloroform-d) δ 211.58, 179.03, 160.25, 157.86, 137.59, 116.13, 115.89, 115.21, 114.96, 110.65, 110.57, 62.95, 52.02, 41.19, 35.79, 35.36, 34.03, 25.34, 21.67. ^19^F NMR (376 MHz, Chloroform-d) δ −119.61.

5-fluoro-1′-methyl-3″,4″-dihydro-1″H-dispiro[indoline-3,2′-pyrrolidine-3′,2″-naphthalene]-1″,2-dione (12c)

5.5.6

To a solution of 5-Fluoro-isatine (1 eq, 3.03 mmol) in MeOH (15 ml), sarcosine (1.5 eq, 4.54 mmol) and 3,4-dihydronaphthalen-1(2H)-one (1 eq, 3.03 mmol) were added sequentially. The reaction was then stirred at 90 °C under reflux for 10 h. The reaction was monitored by mass analysis and TLC (A1P1). Upon the completion of the reaction, this was quenched with NaHCO3 sat and deionized water. The organic solvent was evaporated and the aqueous phase was extracted three times with ethyl acetate (40 ml). The organic phases, taken together, after treatment with anhydrous sodium sulphate, were evaporated and an amorphous orange solid was obtained. The crude product was then purified by flash silica gel chromatography with eluent mixture of ethyl acetate and petroleum ether (A1P1), obtaining an orange solid. Exact mass (m/z): 351.1498; Yield: 48%; HPLC Rt: 16,27; ^1^H NMR (400 MHz, Chloroform-d) δ 8.29 (s, 1H), 8.09 (dd, J = 7.7, 1.7 Hz, 1H), 7.37 – 7.16 (m, 3H), 6.94 – 6.81 (m, 1H), 6.76 – 6.62 (m, 2H), 6.62 – 6.44 (m, 1H), 3.53 (q, J = 8.4 Hz, 1H), 3.25 (ddd, J = 10.6, 8.8, 4.3 Hz, 1H), 2.58 (ddd, J = 12.7, 9.1, 4.2 Hz, 3H), 2.42 (dt, J = 14.5, 4.1 Hz, 1H), 2.34 – 2.19 (m, 2H), 2.10 (s, 3H). ^13^C NMR (101 MHz, Chloroform-d) δ 198.74, 159.86, 157.46, 143.25, 137.47, 137.45, 133.26, 133.03, 128.14, 127.85, 126.97, 115.82, 115.58, 110.06, 109.98, 57.97, 51.89, 35.57, 33.37, 31.96, 26.07. ^19^F NMR (376 MHz, Chloroform-d) δ −120.56.

5″-fluoro-1′-methyl-2H-dispiro[acenaphthylene-1,3′-pyrrolidine-2′,3″-indoline]-2,2″-dione (12d)

5.5.7

To a solution of 5-Fluoro-isatine (1 eq, 3.03 mmol) in MeOH (15 ml), sarcosine (1.5 eq, 4.54 mmol) and acenaphtenone (1 eq, 3.03 mmol) were added sequentially. The reaction was then stirred at 90 °C under reflux for 10 h. The reaction was monitored by mass analysis and TLC (A1P1). Upon the completion of the reaction, this was quenched with NaHCO_3_ sat and deionized water. The organic solvent was evaporated and the aqueous phase was extracted three times with ethyl acetate (40 ml). The organic phases, taken together, after treatment with anhydrous sodium sulphate, were evaporated and an amorphous orange solid was obtained. The crude product was then purified by flash silica gel chromatography with eluent mixture of ethyl acetate and petroleum ether (A1P1), obtaining an amorphous solid. Re-crystallization with a mixture of diethyl ether and petroleum ether (3:7) gave the desired product as an orange solid. Exact mass (m/z): 373.1345; Yield 40%; mp: 207-209 °C; HPLC Rt: 14,95; ^1^H NMR (400 MHz, Chloroform-d) δ 7.98 – 7.87 (m, 1H), 7.77 – 7.66 (m, 1H), 7.60 (ddd, J = 8.1, 7.1, 5.5 Hz, 1H), 7.50 (dd, J = 8.3, 7.1 Hz, 2H), 7.02 (dd, J = 8.6, 2.6 Hz, 1H), 6.66 – 6.57 (m, 1H), 6.39 (dd, J = 8.4, 4.2 Hz, 2H), 3.87 – 3.78 (m, 1H), 3.61 (ddd, J = 10.7, 8.8, 4.1 Hz, 3H), 3.05 (ddd, J = 13.0, 10.7, 5.4 Hz, 2H). ^13^C NMR (101 MHz, Chloroform-d) δ 204.60, 160.14, 157.74, 142.01, 132.47, 132.00, 131.43, 130.35, 128.70, 128.28, 128.06, 125.01, 122.59, 121.83, 121.77, 116.04, 115.81, 114.67, 114.41, 110.15, 110.07, 63.59, 52.43, 35.89, 31.78. ^19^F NMR (376 MHz, Chloroform-d) δ −120.60.

1-benzyl-1′-methyl-3″,4″-dihydro-1″H-dispiro[indoline-3,2′-pyrrolidine-3′,2″-naphthalene]-1″,2-dione (13)

5.5.8

In a two-necked flask with an anhydrous environment, the amide (1 eq) was dissolved in DMF (5 ml). Then, sequentially, NaH (1 eq) and benzyl bromide (1 eq) were added. A chromatographic change from yellow to purple to green was evident. The reaction was monitored by TLC (A1P1). After 30 min, the reaction was completed and the organic solvent was evaporated, followed by a treatment with diethyl ether (5 ml). The solid phase was then filtered on Gooch and the diethyl ether evaporated. Separation by chromatographic column with eluent mixture A1P1 was carried out. After evaporation of the organic solvent, a yellow solid was obtained. Exact mass (m/z): 423.2068; Yield: 65%; HPLC Rt: 18,02; mp:150-152 °C; ^1^H NMR (400 MHz, Chloroform-d) δ 8.07 (dd, J = 7.3, 2.1 Hz, 1H), 7.43 – 7.13 (m, 8H), 6.95 (t, J = 7.6 Hz, 1H), 6.86 – 6.72 (m, 2H), 6.62 (d, J = 7.8 Hz, 1H), 6.51 (t, J = 7.6 Hz, 1H), 4.99 (d, J = 15.3 Hz, 1H), 4.84 (d, J = 15.3 Hz, 1H), 3.58 (s, 1H), 3.40 – 3.20 (m, 1H), 2.57 (ddd, J = 12.6, 8.6, 4.2 Hz, 1H), 2.49 – 2.12 (m, 5H), 2.08 (s, 3H). ^13^C NMR (101 MHz, Chloroform-d) δ 198.48, 142.95, 142.73, 135.37, 132.57, 132.20, 128.53, 128.28, 127.41, 127.30, 127.16, 127.07, 126.21, 126.02, 121.47, 107.98, 57.18, 51.30, 43.13, 35.01, 32.83, 31.37, 25.34.

1-benzyl-5-fluoro-1′-methyl-3″,4″-dihydro-1″H-dispiro[indoline-3,2′-pyrrolidine-3′,2″-naphthalene]-1″,2-dione (14)

5.5.9

In a two-necked flask with an anhydrous environment, the amide (1 eq) was dissolved in DMF (5 ml). Then, sequentially, NaH (1 eq) and benzyl bromide (1 eq) were added. A chromatographic change from yellow to purple to green was evident. The reaction was monitored by TLC (A1P1). After 30 min, the reaction was completed and the organic solvent was evaporated, followed by a treatment with diethyl ether (5 ml). The solid phase was then filtered on Gooch and the diethyl ether evaporated. Separation by chromatographic column with eluent mixture A1P1 was carried out. After evaporation of the organic solvent, a yellow solid was obtained.

Exact mass (m/z): 441.1594 Yield: 98%; mp: 152-154 °C; HPLC Rt: 21,58; ^1^H NMR (400 MHz, Chloroform-d) δ 8.09 (dd, J = 7.5, 1.9 Hz, 1H), 7.42 – 7.18 (m, 7H), 6.89 – 6.76 (m, 1H), 6.65 (dd, J = 8.6, 2.6 Hz, 1H), 6.51 (dd, J = 8.5, 4.2 Hz, 1H), 5.01 – 4.76 (m, 2H), 3.56 (s, 1H), 3.31 (td, J = 9.7, 8.6, 4.2 Hz, 1H), 2.58 (ddd, J = 12.7, 8.6, 4.2 Hz, 1H), 2.52 – 2.15 (m, 4H), 2.08 (s, 3H). ^13^C NMR (101 MHz, Chloroform-d) δ 198.75, 159.99, 157.58, 143.11, 139.51, 135.70, 133.15, 129.02, 128.10, 127.84, 127.78, 126.96, 115.56, 115.32, 114.97, 109.16, 58.15, 52.00, 43.94, 35.67, 33.47, 31.91, 26.02. ^19^F NMR (376 MHz, Chloroform-d) δ −125.90.

Calcein−Cobalt assay

5.6

AC16 cells were pretreated with DMSO (vehicle) and different compounds for 15 min. For the initial screening experiments, all the compounds were used 1 μM. For the dose-response experiments, incremental concentration was used: 100 nM, 500 nM, 1 μM and 5 μM, then the best concentrations in inhibiting PTP opening were chosen for the consequent experiments.

Cells were loaded with calcein acetoxymethyl ester and Co^2+^ as previously described [17]. Staining solution was added to the cells for 15 min at 37 °C in a 5% CO_2_ atmosphere. Image acquisitions were performed with a motorized Nikon AX R confocal microscope with a 40X/0.6 PlanApo objective and laser LU-N4S 405/488/561/640. Ionomycin (1 μM) was administered 30 s after the beginning of the experiment to induce PTP opening.

Mitochondrial isolation and swelling assay

5.7

The study was approved by the Local Ethics Committee (prot. CBCC2.N.L1R) and all procedures are conformed to the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals. In agreement with this approval, euthanasia of mice was induced by cervical dislocation and mitochondria were isolated by conventional procedures involving differential centrifugation. Freshly excised SV129 mouse hearts were washed and then homogenized in medium containing 50 mmol/L Tris-HCl, 25.67 g/L sucrose, and 40.98 g/L D-Mannitol (pH 7.4) supplemented with 0.5 mmol/L EGTA and 5 g/L bovine serum albumin (BSA). The homogenate was then transferred to microcentrifuge tubes and centrifuged at 0.8 rcf for 5 min for at least 2 cycles; the supernatants were collected, and the pellets were discarded. Subsequently, the sample was centrifuged at 10.0 rcf for 10 min to separate the mitochondrial fraction; the pellet was resuspended and ground in a loose-fitting glass Potter Elvehjem homogenizer for a fixed number of times. Samples were then centrifuged at 10.0 rcf for 10 min, and the pellet (mitochondria) was resuspended in 1 mL of Respiration Buffer (pH 7.4, Tris-HCl 50 mmol/L; 25.67 g/L sucrose; 40.98 g/L D-Mannitol) supplemented with 5 mmol/L succinate. Mitochondria were quantified and diluted to a final concentration of 1 mg/mL for each mitochondrial swelling assay to monitor the changes in absorbance at 540 nm as previously described [36]. Incubations with small molecules were carried out at 25 °C, and PTP opening was induced by the addition of 500 μM Ca^2+^. Data are shown as percentage compared to initial absorbance (equal to 100%).

Cell viability

5.8

For crystal violet assay, the cells seeded in 12-well plates were treated with vehicle or compounds for 24 h, 48 h and 72 h. At the end of the treatment, the cells were washed with PBS, fixed in 4% paraformaldehyde, and stained with 0.1% crystal violet. Crystal violet was dissolved in 1 mol/L acetic acid, and the absorbance at 595 nm was measured with a spectrophotometer.

For MTS assay, equal numbers of cells were seeded onto each well of a 96-well plate. The cells were incubated in normal growth medium supplemented with vehicle or compounds for 24 h and the MTS test (Abcam, ab197010) was performed according to manufacturer's instructions. The resulting absorbance was read at 490 nm.

In vitro hypoxia/reoxygenation (H/R)

5.9

The protective effect of 11d, 11c, 12c was assayed by a model of in vitro H/R on pAECs, previously described [45]. Briefly, pAECs were seeded in 96-well plates at a density of 1 × 10^4^ cells/well. The day after cell culture medium was replaced by the acid buffer (137 mM NaCl, 12 mM KCl, 0.9 mM CaCl_2_, 0.49 mM MgCl_2_, 4 mM HEPES and 20 mM sodium dl-lactate at pH 6.2) and cells were placed in a modular incubator chamber (Billups-Rothenberg USA) containing a gas mixture (1% O_2_, 5% CO_2_ and 94% N_2_). After 13 h, the normoxic conditions and the complete culture medium were restored for 24 h (reoxygenation) before cell viability assay and cell metabolism assay, as previously described [45]. Control group (CTR) was represented by pAECs cultured in normoxic conditions and in cell culture medium with the vehicle (0.01% DMSO).

For AC16H/R experiments, cells were rinsed with PBS, and the culture medium was changed according to the oxygen-glucose deprivation protocol described in Ref. [69] with slight variation. Cells were exposed to 1% O_2_ at 37 °C for 2 h to simulate ischemia. At the time of reperfusion, cells were cultured with complete fresh medium and measured 48 h later, for treated cells, compounds were added for throughout the entire reperfusion period (5 μM).

Mitochondrial Ca2+ measurements

5.10

AC16 cells were transfected with mtAEQwt. 48 h later, the coverslips were incubated with 5 mM coelenterazine for 1.5 h in Krebs-Ringer modified buffer (KRB) supplemented with 1 mM CaCl_2_ (KRB: 125 mM NaCl, 5 mM KCl, 1 mM Na_3_PO_4_, 1 mM MgSO_4_, 5.5 mM glucose, and 20 mM HEPES, pH 7.4, at 37 °C). Aequorin signals were measured in KRB supplemented with 1 mM CaCl_2_ using a purpose-built luminometer. The agonist (a mix of 500 mM bradykinin and histamine for AC16) was added to the same medium. The experiments were terminated by lysing the cells with Triton X-100 in a hypotonic Ca^2+^-rich solution (10 mM CaCl_2_ in H_2_O), thus discharging the remaining aequorin pool. The light signals were collected and calibrated with [Ca^2+^] values. Further experimental details have been previously described in Ref. [70].

Preparation of the mitochondrial fractions

5.11

Swine hearts (Sus scrofa domesticus) were collected at a local abattoir and transported to the lab within 2 h in ice buckets at 0–4 °C. After removal of fat and blood clots as much as possible, approximately 30–40 g of heart tissue was rinsed in ice-cold washing Tris-HCl buffer (medium A) consisting of 0.25 M sucrose, 10 mM Tris(hydroxymethyl)-aminomethane (Tris), pH 7.4 and finely chopped into fine pieces with scissors. Each preparation was made from one heart. Once rinsed, tissues were gently dried on blotting paper and weighed. Then tissues were homogenized in medium B consisting of 0.25 M sucrose, 10 mM Tris, 1 mM EDTA (free acid), 0.5 mg/ml BSA fatty acid free, pH 7.4 with HCl at a ratio of 10 ml medium B per 1 g of fresh tissue. After a preliminary gentle break-up by Ultraturrax T25, the tissue was carefully homogenized by a motor-driven Teflon pestle homogenizer (Braun Melsungen Type 853202) at 650 rpm with 3 up-and-down strokes. The mitochondrial fraction was then obtained by stepwise centrifugation (Sorvall RC2-B, rotor SS34). Briefly, the homogenate was centrifuged at 1,000×g for 5 min, thus yielding a supernatant and a pellet. The pellet was re-homogenized under the same conditions of the first homogenization and re-centrifuged at 1,000×g for 5 min. The gathered supernatants from these two centrifugations, filtered through four cotton gauze layers, were centrifuged at 10,500×g for 10 min to yield the raw mitochondrial pellet. The raw pellet was resuspended in medium A and further centrifuged at 10,500×g for 10 min to obtain the final mitochondrial pellet. The latter was resuspended by gentle stirring using a Teflon Potter Elvejehm homogenizer in a small volume of medium A, thus obtaining a protein concentration of 30 mg/ml [71]. All steps were carried out at 0–4 °C. The protein concentration was determined according to the colourimetric method of Bradford by Bio-Rad Protein Assay kit II with BSA as standard. The mitochondrial preparations were then stored in liquid nitrogen until the evaluation of F_1_F_O_-ATPase activities.

Kinetic analyses

5.12

The inhibition mechanism of 11d on the Ca^2+^-activated F_1_F_O_-ATPases was explored by the graphical methods of Dixon and Cornish-Bowden plots, which complement one another [72]. To this aim, the 1/v (reciprocal of the enzyme activity v) in Dixon plot or the S/v ratio in Cornish-Bowden plot were plotted as a function of 11d concentration. In all plots the enzyme specific activity was taken as the expression of v. To build these plots, different experimental sets were designed in which the F-ATPase activity was evaluated in the presence of increasing 11d concentrations at two ATP concentrations, keeping the divalent cofactor (Ca^2+^) concentration constant. The values of K_i_', which represent the dissociation constant of the ternary ESI complex, were calculated as the abscissa (changed to positive) of the intercept of the straight lines obtained in the Cornish-Bowden plots.

Kinetic studies have been conducted on the mutual exclusion of different inhibitors on the same F-ATPase activity. These analyses aimed to shed light on the possible interaction on the F_O_ domain between compounds 11d and oligomycin (olig) of N,N-dicyclohexylcarbodiimide (DCCD). The value of -K_X_', which represents the dissociation constant of the quaternary complex ESI1I2, was calculated from the abscissa (changed to positive) of the intersection point of the two lines obtained in the presence and absence of inhibitor of F_O_ domain [73].

Mitochondrial respiration assays

5.13

Immediately after thawing, the mitochondrial fractions were used to evaluate mitochondrial respiration. To detect mitochondrial respiratory activities, the oxygen consumption rates were polarographically evaluated by Clark-type electrode using a thermostated Oxytherm System (Hansatech Instruments) equipped with a 1 mL polarographic chamber. The reaction mixture (120 mM KCl, 10 mM Tris-HCl buffer pH 7.2), maintained under Peltier thermostatation at 37 °C and continuous stirring, contained 0.25 mg mitochondrial protein. To evaluate the NADH-O_2_ oxidoreductase activity, the mitochondrial oxidation was run under saturating substrate conditions (75 μM NADH) after 2 min of stabilization of the oxygen signal. Preliminary tests assessed that under these conditions O_2_ consumption was suppressed by 2.5 μM rotenone, known inhibitor of complex I. The succinate-O_2_ oxidoreductase activity by complex II was evaluated by detecting the succinate oxidation in the presence of 2.5 μM rotenone. The reaction was started by the addition of 10 mM succinate after 2 min of stabilization of oxygen signal. Also in this case preliminary tests assessed that, under the conditions applied, succinate oxidation was suppressed by of 1 μg/mL antimycin A, selective inhibitor of CIII [74].

Evaluation of calcium retention capacity (CRC)

5.14

Immediately after the preparation of swine heart mitochondrial fractions, fresh mitochondrial suspensions (1 mg/mL) were energized in the assay buffer (130 mM KCl, 1 mM KH_2_PO_4_, 20 mM HEPES, pH 7.2 with TRIS), incubated at 25 °C with 1 μg/mL rotenone and 5 mM succinate as respiratory substrate. To evaluate NaHS effect, selected NaHS concentrations were added to the mitochondrial suspensions before PTP evaluation. PTP opening was induced by the addition of low concentrations of Ca^2+^ (10 μM) as CaCl_2_ aqueous solution at fixed time intervals (1 min). The calcium retention capacity, whose lowering indicates PTP opening, was spectrofluorophotometrically evaluated in the presence of 0.8 μM Fura-FF. The probe has different spectral properties in the absence and in the presence of Ca^2+^, namely it displays excitation/emission spectra of 365/514 nm in the absence of Ca^2+^ (Fura-FF low Ca^2+^) and shifts to 339/507 nm in the presence of high Ca^2+^ concentrations (Fura-FF high Ca^2+^). PTP opening, was evaluated by the increase in the fluorescence intensity ratio (Fura-FF high Ca^2+^Fura-FF low Ca^2+^), which indicates a decrease in CRC [71].

Basal mitochondrial membrane potential

5.15

Cells were loaded with 20 nM tetramethylrhodamine methyl ester (TMRM) for 30 min at 37 °C. To obtain and analyze basal levels, cells were stimulated with 10 nM carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP), a strong uncoupler of oxidative phosphorylation. Image acquisitions were performed with a motorized Nikon AX R confocal microscope with a 40X/0.6 PlanApo objective and laser LU-N4S 405/488/561/640.

Evaluation of mitochondrial morphology parameters by microscopic analysis

5.16

Mitochondrial morphology was assessed by transfecting mtGFP and staining with Hoechst 33342 to mark the nucleus. Image acquisitions were performed with a motorized Nikon AX R confocal microscope with a 60X/0.6 PlanApo objective and laser LU-N4S 405/488/561/640. Image analysis involved the use of Fiji software and two different plugins: 3D object counter for the detection of total mitochondrial network and number of mitochondria; 2D/3D skeleton for the calculation of branches and junctions.

Oxygen Consumption Rate (OCR)

5.17

The rate of oxygen consumption was assayed with an XF96 Extracellular Flux Analyzer (Seahorse Biosciences—Agilent Technologies, Santa Clara, CA USA). Briefly, mitochondrial respiration was evaluated by measuring the OCR under basal conditions and after the addition of oligomycin (1 μM for AC16, 1.5 μM for pAECs), carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP) (2 μM for AC16, 1 μM for pAECs), rotenone (1 μM for AC16, 0.5 μM for pAECs) and antimycin A (1 μM for AC16, 0.5 μM for pAECs). OCR values were normalized to cell density using the crystal violet assay.

Reactive Oxygen Species (ROS) measurements

5.18

ROS measurements were performed according to Ref. [75]. Briefly, measurement of mitochondrial superoxide (mtO_2_^●−^) production was performed on cells incubated for 30 min at 37 °C in the presence of 5 μM mitoSOX red. The fluorescence was evaluated on Nikon AX R confocal microscope with a PlanApo 60X/1.4 objective and laser LU-N4S 405/488/561/640.

Propidium Iodide (PI) assay

5.19

After H/R, AC16 cells were stained according to the manufacturer's protocols with PI (Thermo Fisher Scientific, P3566) in binding buffer (10 mM HEPES, 5 mM KCl, 150 mm NaCl, 1.8 mM CaCl_2_, 1 mM MgCl_2_, pH 7.4) and left in the dark at room temperature for 15 min. PI-positive cells were quantified by flow cytometry (Attune NxT Flow cytometer, Thermo Fisher Scientific), and data were analyzed with Attune NxT Software (Thermo Fisher Scientific).

Ex vivo model

5.20

H/R was studied ex vivo using the Langendorff model with minor modifications [46]. The study was approved by the Local Ethics Committee (n° 373/2021-PR, prot. CBCC2.48) and all procedures are conformed to the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals. In agreement with this approval, euthanasia of rats was induced by decapitation by using a guillotine with a sharp blade. In brief, upon euthanasia, the hearts of Wistar rats weighing 270−280 g at inclusion into the study were rapidly excised, immediately arrested in ice-cold KHB (pH 7.4; 4 °C), cannulated, and retrograde perfused at a fixed-flow rate (11 mL/min) through the aorta with warm KHB (37 °C) bubbled with 95% O_2_ and 5% CO_2_. Upon removal of the left atrial appendage, a latex fluid-filled balloon was inserted into the left ventricular chamber through the atrium to obtain an isovolumetrically beating preparation and connected to a pressure transducer (APT300, Hugo-Sachs, Grünstrasse, Germany) by a fluid-filled polyethylene catheter to monitor performance. An additional transducer above the aortic cannula monitored the CPP. At the start of each experiment, the fluid in the balloon was increased incrementally to achieve a constant EDP of 4 ± 1 mmHg. The left ventricular developed pressure (LVDP) was then measured. The LVDP, end diastolic pressure (EDP), and coronary perfusion pressure (CPP) were continuously recorded using a programmable acquisition system (HSE Isoheart Software for Isolated Heart, Hugo-Sachs, Grünstraße, Germany).

TUNEL assay

5.21

Hearts from ex vivo experiments were cut into 10-μm-thick sections and processed for a terminal deoxynucleotidyl transferase-mediated deoxyuridine 5-triphosphate nick-end labeling (TUNEL) assay (DeadEnd™ Fluorometric TUNEL System, G3250, Promega) to identify fragmented nuclei in the tissue. Before immunochemistry staining the samples were subjected to an antigen retrieval step in a solution of EDTA 1 mmol/L (pH 8.0) at 100 °C. Samples were processed according to the manufacturer's instructions. TUNEL intensity was evaluated on Nikon AX R confocal microscope with a PlanApo 60X/1.4 objective and laser LU-N4S 405/488/561/640. TUNEL-positive cells were counted in 10 sections and analyzed with Fiji software.

Hematoxylin and eosin histological staining

5.22

Heart slices were stained with hematoxylin for 1 min, rinsed in running tap water, stained with eosin for 30 s and rinsed, dried, and mounted. Images were taken with Nikon Upright Eclipse Ni–U equipped with digital High-definition Color Camera Head DS-Fi2 5-megapixel CCD and PlanFluor 20X/0.5 objective. Images were analyzed using Colour Deconvolution plugin of ImageJ.

Immunoblot analysis

5.23

For immunoblotting, cells were lysed in RIPA buffer and then quantified by the Lowry method, and 10 μg of protein was loaded on a 4%–20% precast gel. After electrophoretic separation, proteins were transferred onto nitrocellulose membranes that were incubated overnight with the following primary antibodies: cleaved PARP (Cell signaling, 9541, 1:1000), Caspase 3 (Cell signaling, 9662, 1:1000), RIP (Cell signaling, 3493, 1:1000), β-Actin (Merck, A1978, 1:5000), ATP5A (Abcam, ab14748, 1:5000). The membranes were then treated with specific HRP-labeled secondary antibodies, followed by chemiluminescence detection using a ChemiDoc Touch Gel Imaging System. Western blots shown in figures are representative of at least three different independent experiments.

Human ATP synthase in complex with oligomycin

5.24

The cryo-electron microscopy structure of the human C-ring was retrieved from the Protein Data Bank (PDB ID: 8H9J) [47] and was used as the starting model for all subsequent computational analyses. In this study, residue numbers are based on the PDB structure, where the sequence starts at residue 62. The human C-ring system was prepared using the Protein Preparation Wizard within the Schrödinger Software Suite (version 2020-3) [[76], [77], [78]]. To accurately position oligomycin within the binding site, the structure was superimposed onto the crystallographic structure of the yeast F1Fo ATPase c10 ring (PDB ID: 4F4S) [48].

Molecular docking 11d and 12c

5.25

Molecular docking of compounds 11d and 12c was performed using the Maestro Schrödinger Software Suite (version 2020-3) [[76], [77], [78]]. To fully explore potential binding modes of these two derivates, the docking grid was generated along the α-helical regions of the C-ring, corresponding to the known oligomycin binding site. Although the C-ring is composed of eight identical subunits arranged symmetrically, the docking grid was defined to cover only two adjacent chains. This approach takes advantage of the ring's symmetry, as the binding environment is repeated across the subunits, allowing for a representative sampling of potential ligand interactions. The grid indeed, was centered at coordinates (26.39, 9.67, 57.09), with an inner box size of 25 Å and an outer box of 30 Å. The receptor structure used for docking was extracted from cluster analysis of the merged MD trajectories of the oligomycin-bound system. The docking protocol involved extra precision (XP) docking using Glide [[76], [77], [78]] and refinement of the protein–ligand complexes using Prime [79]. Ligands were prepared using the LigPrep tool in Maestro, applying the OPLS3e force field [80]. The van der Waals radii were scaled by a factor of 0.50 for both the receptor and ligands. Multiple binding poses were generated, and the top-ranked pose, based on docking scores was selected for subsequent MD simulations.

Molecular dynamics simulations

5.26

All Molecular Dynamics (MD) simulations were performed using GROMACS 2021.4 [81] with the CHARMM36 m force field [81], including the WYF parameters [82] to accurately model π–cation interactions. Solvation was carried out using the TIP3P water model [83]. System setup was conducted via the CHARMM-GUI server [84], employing the Membrane Builder module to embed the C-ring structures within a 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) lipid bilayer, following protocols established in previous studies [29,85,86]. Potassium (K^+^) and chloride (Cl^−^) ions were added to neutralize the system and to reproduce a physiological ionic strength of 0.15 M. All simulations were performed in the NPT ensemble at 303.15 K with a 2 fs integration time step. Oligomycin-bound systems were simulated in triplicate for 100 ns each, while the two selected compounds (11d and 12c) were each simulated in triplicate for 250 ns, resulting in a total of nine independent MD simulations.

MD simulation analysis

5.27

Root mean square deviation (RMSD) values were calculated over time for both Cα atoms and ligands to assess the structural stability of each system. To identify representative conformational states and explore distinct ligand binding modes, cluster analysis was performed on concatenated trajectories for each complex (ATP synthase–oligomycin, ATP synthase–11d and ATP synthase–12c), incorporating the full simulation time from all replicates. Clustering was conducted using the GROMOS algorithm as implemented in GROMACS, based on the RMSD of Cα atoms and ligand positions, with a cutoff of 0.15 nm. Trajectories were sampled at every frame (step size = 1). In addition, protein–ligand contact analysis was performed on the combined trajectories using a custom Python script incorporating the MDAnalysis [87] and PLIP [88] libraries, following the approach described in a previous work [29].

Cellular thermal shift assay

5.28

CETSA was performed as described in previous publications with some important modifications. Briefly, AC16 cells were incubated with vehicle or 11d dissolved in normal growth medium at 37 °C for 15 min. Then, the cells were detached and pelleted by centrifugation at 1300 rpm for 5 min. The cells were resuspended in EBC buffer and divided into equal amounts. Subsequently, the samples were exposed to increasing temperatures ranging from 53 °C to 80 °C, stabilized at room temperature and lysed by 3 cycles of incubation in liquid nitrogen. Next, the samples were pelleted at 13,200 rpm for 15 min. Supernatants were analyzed by immunoblots using antibodies against the Csub (Abcam, ab181243, 1:1000) and ATP5A (Abcam, ab14748, 1:1000) proteins after SDS-PAGE. CETSA curves were obtained by quantifying immunoblot bands and converting the data into percentages.

Subcellular fractionation

5.29

AC16 cells were processed according to Qproteome Mitochondria Isolation Kit (37612, Qiagen) and cytosolic and mitochondrial fractions were isolated following to the manufacturer's instructions starting from 1 x 10^7^ cells. Briefly, cells were harvested, lysed on ice, homogenized, and centrifuged sequentially to obtain a cytosolic supernatant and a mitochondrial pellet, which was resuspended in the provided storage buffer. Protein concentration was determined by Lowry assay, and equal protein amounts from each fraction were analyzed by SDS–PAGE and Western blot. Fraction purity was assessed using β-tubulin (Cell signaling, 2128S, 1:5000) as a cytosolic marker and VDAC (Abcam, ab15895, 1:1000) as a mitochondrial marker, with enrichment of each marker in the appropriate fraction confirming successful subcellular separation.

UPLC-MS/MS analysis

5.30

Cytosolic and mitochondrial fractions were processed using a protein precipitation method. A 50 μL aliquot of each fraction was spiked with 50 μL of melatonin-d_4_ internal standard solution (final concentration: 100 nM). Proteins were precipitated by adding 150 μL of ice-cold acetonitrile/methanol/water (8:1:1, v/v/v) solution. The mixture was vortexed for 2 min, incubated on ice for 10 min, and then centrifuged at 14,000×g for 15 min at 4 °C. The resulting supernatant was collected for UPLC-MS/MS analysis. The UPLC-MS/MS analysis was performed using an ExionLC 2.0 system (AB Sciex LLC, USA) coupled to a QTRAP 6500+ mass spectrometer (AB Sciex LLC, USA). Separation was achieved on an ACE Excel C18 column (100 mm × 2.1 mm, 2 μm; Avantor-ACE, UK) maintained at 40 °C. The mobile phase consisted of 0.25% formic acid in water (solvent A) and 0.25% formic acid in acetonitrile/methanol (95:5, v/v) (solvent B). A gradient elution was applied at a flow rate of 0.2 mL/min with an injection volume of 5 μL. The gradient profile was as follows: 90% A was held for 3 min, linearly decreased to 10% A between 3 and 10 min, maintained at 10% An until 15 min, and then returned to 90% A at 16.5 min. The column was re-equilibrated under initial conditions for the remaining 3.5 min, resulting in a total run time of 20 min. Mass spectrometric detection was conducted using an electrospray ionization (ESI) source operating in positive ion mode. Data acquisition was performed in multiple reaction monitoring (MRM) mode with a dwell time of 50 ms for each transition. The specific MRM parameters were as follows: for the internal standard (melatonin-d_4_), the transition m/z 237.1 → 178.1 (DP: 80 V, CE: 20 V); for the analyte 11d, the quantifier transition m/z 355 → 175 (DP: 80 V, CE: 30 V) and the qualifier transitions m/z 355 → 284 (DP: 80 V, CE: 40 V) and m/z 355 → 195 (DP: 80 V, CE: 35 V). Nitrogen was used as the collision gas. The ion source parameters were set as follows: ion-spray voltage, 5500 V; source temperature, 350 °C; curtain gas, 35; ion source gas 1 (GS1), 45; and ion source gas 2 (GS2), 35 (all in arbitrary manufacturer units). Under these conditions, the retention times were approximately 8.5 min for 11d and 7.8 min for the internal standard. For the comparative analysis of 11d levels between subcellular fractions, the analyte-to-internal standard peak area ratio was used as a direct measure of relative abundance. This ratio was normalized to the total protein content of each fraction, determined by a Lowry assay. The final data are expressed as normalized peak area ratios (arbitrary units, AU) per mg of total protein to allow for direct comparison of 11d′s relative abundance between cytosolic and mitochondrial fractions. All data processing was performed using the SCIEX OS software.

Statistical analysis

5.31

The statistical method used is one-way ANOVA with multiple comparisons performed by GraphPad Prism 9 (Prism, La Jolla, CA, USA). A P-value <0.05 was considered significant. P-values, replicates, SEM/SD are reported in the figure legends.

Declaration of interest

The authors declare that they have no competing interests.

The small-molecules presented in this study are patented (PCT/IB2024/052951).

Funding

The Signal Transduction Laboratory is supported by the Italian Ministry of Health grant GR-2018-12367114 and GR-2019-12369862 (to G.M.); the Italian Association for Cancer Research grants IG-23670 (to P.P.) and IG-19803 (to C.G.), A-ROSE (Associazione Ricerca Oncologica Sperimentale Estense); European Research Council grant 853057-InflaPML (to C.G.); and local funds from the University of Ferrara (to P.P. and C.G.). This research was funded by the Carisbo Foundation, grant number 2019.0534 (to S.N.).

CRediT authorship contribution statement

Giampaolo Morciano: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Validation, Writing – original draft, Writing – review & editing. Gaia Pedriali: Data curation, Formal analysis. Giulia Turrin: Methodology. Cristina Algieri: Data curation, Formal analysis. Esther Densu Agyapong: Data curation, Formal analysis. Debora La Mantia: Data curation, Formal analysis. Chiara Bernardini: Data curation, Formal analysis, Investigation, Supervision, Validation. Giorgia Canini: Data curation, Formal analysis, Software. Anna Fantinati: Data curation, Formal analysis, Methodology. Elena Nicoletta Colarusso: Data curation, Formal analysis. Fabio Mollica: Data curation, Formal analysis, Software. Daniela Ramaccini: Data curation, Formal analysis. Alessandra Pagliarani: Supervision, Validation. Carlotta Giorgi: Validation. Elena Tremoli: Supervision, Validation. Alessandro Arcovito: Software, Supervision, Validation, Visualization. Salvatore Nesci: Data curation, Investigation, Supervision, Validation. Claudio Trapella: Conceptualization, Investigation, Supervision, Validation. Paolo Pinton: Conceptualization, Supervision, Validation.

Declaration of competing interest

All authors have nothing to disclose.