Mitochondrial permeability transition in stem cells, development, and disease

Abstract

The mitochondrial permeability transition (mPT) is a process that permits rapid exchange of small molecules across the inner mitochondrial membrane (IMM) and thus plays a vital role in mitochondrial function and cellular signaling. Formation of the pore that mediates this flux is well-documented in injury and disease but its regulation has also emerged as critical to the fate of stem cells during embryonic development. The precise molecular composition of the mPTP has been enigmatic, with far more genetic studies eliminating molecular candidates than confirming them. Rigorous studies in the recent decade have implicated central involvement of the F₁F_o ATP synthase, or complex V of the electron transport chain, and continue to confirm a regulatory role for Cyclophilin D (CypD), encoded by Ppif, in modulating the sensitivity of the pore to opening. A host of endogenous molecules have been shown to trigger flux characteristic of mPT, including positive regulators such as calcium ions, reactive oxygen species, inorganic phosphate, and fatty acids. Conductance of the pore has been described as low or high, and reversibility of pore opening appears to correspond with the relative abundance of negative regulators of mPT such as adenine nucleotides, hydrogen ion, and divalent cations that compete for calcium binding sites in the mPTP. Current models suggest that distinct pores could be responsible for differing reversibility and conductance dependent upon cellular context. Indeed, irreversible propagation of mPT inevitably leads to collapse of transmembrane potential, arrest of ATP synthesis, mitochondrial swelling, and cell death. Future studies should clarify ambiguities in mPTP structure and reveal new roles for mPT in dictating specialized cellular functions beyond cell survival that are tied to mitochondrial fitness including stem cell self-renewal and fate. The focus of this review is to describe contemporary models of the mPTP and highlight how pore activity impacts stem cells and development.

1. Introduction

The mitochondrial permeability transition (mPT) is a process wherein strict barrier function of the inner mitochondrial membrane (IMM) is abruptly lost. Under homeostasis, the IMM is impermeable, comprised of a specialized system of transporters and exchangers that restrict and tightly regulate the movement of molecules, ions, and metabolites between matrix and intermembrane space (Zorov et al., 2009). Mitochondrial function depends upon the electric insulation provided by the IMM to maintain a proton gradient sufficient for generation of mitochondrial membrane potential (ΔΨ_m) and ATP production (Mitchell, 1966). Although precise molecular composition remains debated, an entity referred to as the mitochondrial permeability transition pore (mPTP) is implicated in regulating the mPT and is sensitive to endogenous molecules, such as inorganic phosphate, adenine nucleotides, ROS, fatty acids, nitric oxide, H⁺, Mg²⁺, and Ca²⁺ (Antoniel et al., 2018; Chernyak and Bernardi, 1996; Costantini et al., 1996; Furuno et al., 2001; Halestrap and Pasdois, 2009; Halestrap et al., 2004; Haworth and Hunter, 1979; Hunter and Haworth, 1979b, 1979a; Kowaltowski et al., 1998, 1996; Wiȩckowski et al., 2000). Opening of the pore permits passage of any small molecular weight solutes < 1.5 kDa between the matrix and intermembrane space (IMS), resulting in depolarization of the IMM. Dissipation of ΔΨ_m caused by pore opening uncouples electron transport from the phosphorylation required for ATP synthesis and also disrupts other ΔΨ_m-dependent activities such as mitochondrial protein import (Bonora et al., 2013; Halestrap et al., 2002). Various mitochondrial matrix metabolites, such as Ca²⁺, NAD⁺, glutathione, and ROS are released into the intermembrane space. This breakdown of barrier function can lead to mitochondrial swelling and Bax/Bak-dependent OMM herniation that permits extrusion of matrix contents into the cytosol (Karch et al., 2013; McArthur et al., 2018). Release of matrix proteins and other metabolites into the cytoplasm can destabilize cellular homeostasis and amplify oxidative damage to proteins, nuclear DNA, ion channels, transporters, and membrane phospholipids (Zorov et al., 2014). Unrestrained, mPT can initiate cell death, but mPT is reversible and can produce transient flickering of IMM permeability when balance within the cellular environment is tipped toward negative regulation of mPT (Boyman et al., 2019).

Modulation of the mPTP is documented in embryonic development, normal physiology, disease, and injury. In the context of disease and injury, pathological opening of the mPTP can be triggered by an excess of endogenous molecules that cause oxidative stress and mitochondrial Ca²⁺ overload (Biasutto et al., 2016; Zoratti and Szabò, 1995). If conditions favorable to mPT are sustained, imbalance in matrix-IMS composition leads to mitochondrial swelling, breakdown of the outer mitochondrial membrane (OMM), and release of mitochondrial solutes that induce cell death and other pathologies (Karch and Molkentin, 2014; Suh et al., 2013). Substantial evidence supports that opening of the mPTP worsens outcomes following ischemia-reperfusion injury of cardiac and neural tissues (Bonora et al., 2020; Halestrap and Richardson, 2015; Hausenloy et al., 2020). Indeed, a large number of clinical trials have been designed around pharmacological inhibition of the mPTP and/or modulation of mitochondrial functions; yet, despite success in preclinical models, the vast majority have failed to produce improvement in clinical outcomes (Bonora et al., 2022; Carrer et al., 2021; Singh et al., 2021). Pharmacological inhibition of mPT does appear to provide protection in ischemia-reperfusion injuries (Leger et al., 2011; Matsumoto et al., 2018; Rekuviene et al., 2017) and surgery (Chiari et al., 2014). But, true benefit has yet to be definitively demonstrated in clinical trials for many ischemia-reperfusion injuries (Bøtker et al., 2020; Upadhaya et al., 2017). Beyond injury, opening of the pore is implicated in exacerbating metabolic diseases, including diabetes, thus targeting the mPTP could be considered for these disorders as well (Taddeo et al., 2014). Prolonged mPT activity is most frequently associated with pathophysiology, yet transient mPT could also play roles in cardiac development and recovery from injury (Elrod et al., 2010; Hausenloy et al., 2004; Korge et al., 2011), synaptic plasticity and efficacy (Mnatsakanyan et al., 2017), and in homeostasis of physiological Ca²⁺ levels (Bernardi and von Stockum, 2012). A growing body of evidence in stem cells indicate that mPTP closure is central to self-renewal, fate commitment, cell survival, metabolic reprogramming, and regenerative response to stress (Pérez and Quintanilla, 2017). Several excellent recent reviews cover in great depth the known endogenous and pharmacological inducers and inhibitors that could be leveraged for mPTP drug design, as well as mPTP regulation by post-translational modifications and signaling networks (Alves-Figueiredo et al., 2021; Bonora et al., 2022; Carrer et al., 2021; Morciano et al., 2021). Unclear is how targeting of the mPTP in stem cells could be leveraged for therapeutic use. In this review, we describe current knowledge regarding regulation of this enigmatic pore and focus on its role in stem cells and development.

2. Ca²⁺-induced mitochondrial permeability transition

Opening of the mPTP is highly dependent upon mitochondrial Ca²⁺. Shuttling of Ca²⁺ between the extracellular, cytosolic, and mitochondrial compartments is vital to mitochondrial function and can communicate life or death signals via the mPT (Giorgi et al., 2008). Specialized channels in the OMM and IMM control accumulation of Ca²⁺ within the mitochondrial matrix (Nicholls, 2005). Flux of Ca²⁺ between the cytosol and the IMS is chiefly mediated by the beta-barrel voltage-dependent anion channel (VDAC) super-family, a high abundance mitochondrial porin localized to the OMM (Zeth and Zachariae, 2018). These porins serve as gatekeeper to small ions and high molecular weight metabolites, facilitate the exit of millions of molecules of ATP per second from the mitochondria, and regulate cell death through association with pro- and anti-apoptotic proteins (Choudhary et al., 2014; Noskov et al., 2016). Their role in formation of the mPTP is thought to stem from transport of Ca²⁺ and ATP/ADP across the OMM rather than as a core component or accessory factor of the mPTP (McCommis and Baines, 2012). From the IMS, Ca²⁺ crosses the highly impermeant IMM into the matrix through the selective mitochondrial calcium uniporter complex, MCU_cx, and is extruded by Na⁺/Ca²⁺ and H⁺/Ca²⁺ exchangers (Filadi and Greotti, 2021; Garg et al., 2021; Palty et al., 2010; Villa et al., 1998). Within the matrix, Ca²⁺ modulates the activity of enzymes important for ATP synthesis, making Ca²⁺ a vital messenger between the cell’s bioenergetic demand and the mitochondria’s ability to supply “free energy”. Yet, pathological conditions that favor excess sequestration of Ca²⁺ in the matrix, resulting in Ca²⁺ overload, can initiate mPTP opening and trigger a cascade leading to cell death. The mPTP itself does not appear to be important for Ca²⁺ efflux under stress although evidence supports its role in homeostasis via transient pore opening (Bernardi and Petronilli, 1996; Lu et al., 2016; Marchi et al., 2014).

Fundamental understanding of the mPT was first established from experiments conducted with mitochondrial isolates or permeabilized cells where increasing concentrations of Ca²⁺ (and other modulators) could be loaded to trigger mPTP opening (Bernardi et al., 2006; Rasola and Bernardi, 2007). Yet, cells are equipped with sophisticated machinery designed to protect them from excess extracellular Ca²⁺ associated with physiological stresses (Giorgi et al., 2018; Nicholls, 2005); thus, experimentation in vivo has required use of compounds capable of modulating Ca²⁺ transport and the mPTP in intact cells (Carrer et al., 2021). Endogenous molecules, ATP and potassium chloride (KCl), elevate mitochondrial Ca²⁺ by stimulation of efflux of Ca²⁺ stores from the endoplasmic reticulum and influx of extracellular Ca²⁺ via voltage-gated Ca²⁺ channels on the plasma membrane, respectively (Barsukova et al., 2011). The Ca²⁺ ionophore ionomycin (A23187) is also a popular compound used to trigger the mPT because it increases cytoplasmic and mitochondrial Ca²⁺ by formation of lipid-soluble complexes with Ca²⁺ that enable its transport across endoplasmic reticulum and plasma membranes (Pressman, 1976). In contrast, phenothiazines, anesthetics, and divalent cations Mg²⁺, Ba²⁺, and Sr²⁺, generally delay or inhibit sensitivity of the mPTP to opening by competing with Ca²⁺ for binding sites or obstructing Ca²⁺ influx (Zoratti and Szabò, 1995). Many modifiers of mPT exert their activity via modulation of the sensitivity of mPTP core components and regulatory factors to Ca²⁺ (Bernardi et al., 2015); nevertheless, some cellular states favor mPTP activation independently of Ca²⁺ concentration. For example, because the mPTP is a voltage-gated channel, loss of mitochondrial membrane potential alone can trigger its opening (Bernardi, 1992). Thus, precise understanding of the mPT has required careful isolation of the effects of matrix Ca²⁺, membrane potential, and other mechanisms thought to regulate conformational change of mPTP components.

3. Reactive oxygen species and other inducers of mPT

A host of endogenous and pharmacological compounds increase the sensitivity of the mPTP to opening (Zoratti and Szabò, 1995). The most well studied inducers include Ca²⁺ (detailed above), inorganic phosphate (Pi), protonophores, oxidizing agents, and reactive oxygen species (ROS). Respiratory complexes and other mitochondrial enzymes are responsible for the vast majority of oxidative stress in the cell, though reactive nitrogen species produced by nitric oxide synthase are also important sources of oxidative stress that can trigger mPT (Briston et al., 2017; Kaludercic and Giorgio, 2016; Vieira et al., 2001; Vorobjeva et al., 2020; Zorov et al., 2014). Indeed, the mPTP is generally sensitized by oxidants and desensitized by antioxidants that directly scavenge free radicals or otherwise increase antioxidant defenses (Chernyak and Bernardi, 1996; Costantini et al., 1996; Kaludercic and Giorgio, 2016; Kowaltowski et al., 1998; Petronilli et al., 1994). Early studies found that exposure of mitochondria to oxidizing agents activates opening of the pore, as by oxidation of mitochondrial glutathione with the ROS donor tert-butyl hydroperoxide (TBHP) or cross-linking of dithiols with arsenite oxide to produce oxidized pyridine nucleotide pools (NAD⁺, NADP⁺) (Chernyak and Bernardi, 1996; Connern and Halestrap, 1994; Petronilli et al., 1994). In contrast, sensitivity of the mPTP could be reversed by the reducing agent dithiothreitol. More recently, antioxidants have been designed to be targeted to the IMM to improve ROS scavenging and removal. Some of these, including the cardiolipin-binding peptide elamipretide (Szeto-Schiller 31 (SS-31), MTP-131, or Bendavia) and the compounds astaxanthin and MCI-186, potently inhibit mPT and prevent mitochondrial swelling and depolarization in isolated mitochondria and cell cultures (Baburina et al., 2019; Rajesh et al., 2003; Szeto, 2006; Zhao et al., 2004).

Not coincidentally, Ca²⁺ is a major contributor to ROS generation (Kanno et al., 2004). Mitochondrial Ca²⁺ alters activity of several dehydrogenases in the tricarboxylic acid cycle (TCA) and activity/conductance of complexes I, III, IV, and V in OXPHOS (Glancy et al., 2013; Territo et al., 2000). Thus, by perturbation of TCA and electron transport chain (ETC) enzyme activity, excess mitochondrial Ca²⁺ can lead to build up of reduced NADH and greater probability of electrons moving from ETC complexes to O₂ to form superoxide radicals (Bertero and Maack, 2018). Several ETC complexes, including complexes I, II, and III, and oxidoreductases of the TCA cycle have been directly implicated in ROS production associated with mPT (Batandier et al., 2004; Bonke et al., 2016; Korge et al., 2017b, 2017a). It is not surprising then that several inducers that trigger mPTP via ROS generation do so in conjunction with elevation of matrix Ca²⁺ and/or alteration of OXPHOS (Baumgartner et al., 2009; Davidson et al., 2012; Gu et al., 2015; Hansson et al., 2008; Hou et al., 2013; Krestinin et al., 2020; Lindsay et al., 2015). Yet, some studies also suggest that ROS generation can occur independently of Ca²⁺ dysregulation, with potential consequences for mPT (see discussion below).

Several studies have also demonstrated feedback between ROS production and mPTP opening, though much debate surrounded whether ROS acted independently of membrane potential to trigger mPTP opening (Zorov et al., 2014). In a seminal study of ROS and mPT, Zorov and colleagues observed a phenomenon they described as ROS-induced ROS release that could be inhibited by blocking mPT (Zorov et al., 2000). Specifically, photoexcitation of TMRM-loaded mitochondria induced production of superoxide anion radical (·O₂−) and hydroxyl radical (·OH), which further promoted a burst of ROS production. This secondary burst of ROS was accompanied by a collapse in membrane potential. Importantly, compound-based inhibition of mPTP or complex I of the ETC reduced magnitude of the ROS burst, strongly supporting that ROS was derived from the ETC and depended upon mPT. Likewise, scavenging of ROS or use of superoxide dismutase (SOD) mimetics inhibited mPT, highly suggestive of positive feedback between ROS signaling and opening of the mPTP. In this study, ROS appeared to be wholly responsible for opening of the mPTP, as chelation of Ca²⁺ or inhibition of the MCU_cx had no effect on ROS-induced mPT. Subsequent studies corroborated these findings, confirming that opening of the mPTP elevates ROS production (Batandier et al., 2004; Maciel et al., 2001). Interest in this area continues with more recent work showing that enforced opening of the mPTP contributes to elevated ROS production in cultured myofibers, thereby activating caspase-3 and subsequent cell death (Burke et al., 2021). In summary, oxidative stress and Ca²⁺ work in concert with other inducers to sensitize the mPTP to opening.

4. Negative regulators of mPT

Negative regulators of the mPT have attracted great attention for their potential therapeutic value. Endogenous molecules that desensitize the mPTP and thus generally increase the levels of Ca²⁺ needed to stimulate mPTP opening include divalent cations (Mg²⁺, Mn²⁺, Ba²⁺, Sr²⁺), K⁺, Na⁺, protons (H⁺)/acidic pH, adenine nucleotides (ADP, AMP, ATP), creatine, and reduced pyrimidines (NADH) (Bernardi et al., 1992; Dolder et al., 2003; Haworth et al., 1980; Hunter and Haworth, 1979b; Qian et al., 1997; Szabo et al., 1992). Elevated membrane potential, corresponding with energization of mitochondria, also prevents mPT (Hunter et al., 1976). A number of compounds have been found to inhibit pore opening, some without known mechanism of action (ER-000444793, triazoles TR001 and TR002) and others thought to sequester CypD (CsA and analogs (NIM811, JW47), Debio025, sanglifehrin A, 7I6, and cyclophilin inhibitor 1), inhibit ANT (bongkrekic acid and cinnamic anilides GNX-4728 and GNX-4975), bind to TSPO (TRO40303), block ETC activity (oligomycin and derivatives), interrupt fatty acid availability (nupercaine and tetracaine), or scavenge ROS (elamipretide, astaxanthin, and MCI-186). Recent reviews describe these pharmacological compounds in detail, along with disease applications and known targets/mechanism of action (Bonora et al., 2022; Carrer et al., 2021; Morciano et al., 2021). Identification of negative regulators of the mPTP remains an area of intense interest as development of inhibitory molecules of mPT could offer new therapeutic options for treatment of ischemic-reperfusion injuries and will no doubt be aided by better understanding of the proteins that comprise and regulate the channel.

5. Molecular composition of the mPTP

The precise molecular composition of the mPTP has been enigmatic, with far more genetic studies eliminating molecular candidates than confirming them. Rigorous studies in the recent decade continue to confirm a chief regulatory role for Cyclophilin D (CypD) in modulating the sensitivity of the pore to opening and have implicated central involvement of the F₁F_o ATP synthase in classical mPTP high-conductance behavior. Recently suggested is that members of the adenine nucleotide translocator (ANT) family serve as a pore forming component in what may be an independent low-conductance channel, raising the possibility that two distinct mPTP channels comprised of different structural components mediate response to Ca²⁺ and other inducers (Bround et al., 2020; Karch et al., 2019) (Fig. 1). Full conductance channels are thought to produce long-lasting mPTP permeability resulting in mitochondrial swelling and cell death; whereas, the low conductance channels transiently release Ca²⁺ and rapidly restore IMM barrier function. Several other proteins have been tested for their contributions to mPT but have largely been excluded as core components. These include the PiC in the IMM, mitochondrial creatine kinase (mtCK) in the IMS, and hexokinase (HK), translocator protein of 18 kDa (TSPO), and VDAC in the OMM. Current models incorporate them as regulators or accessories to transport of mPTP modulators (Baines et al., 2007; Dolder et al., 2003; Gutiérrez-Aguilar et al., 2014; Krauskopf et al., 2006; Kwong et al., 2014; Šileikyte et al., 2014; Vyssokikh and Brdiczka, 2003; Whittington et al., 2018). Recent expert reviews of contemporary models of the mPTP are available that provide comprehensive discussion of the critical roles that upstream signaling pathways and post-translational regulation of mPTP core components play in modulation of pore activity (Alves-Figueiredo et al., 2021; Bonora et al., 2022; Morciano et al., 2021). Indeed, the dynamic conformational change of proteins required to control opening and closing of the pore appears to depend upon post-translational modifications, such as phosphorylation, acetylation, S-nitrosylation, S-glutathionylation, sulfenylation, nitration, deamination, succinylation, and ubiquitination (Alves-Figueiredo et al., 2021). Thus, composition of key components will be only briefly summarized here.

Figure 1. Models of mPTP structural composition and conformational change.

Current working models of the mPTP include a high-conductance channel formed by (A) ATP synthase dimers or (B) the c-ring component of monomeric ATP synthase. (C) In addition, literature suggest that an independent low conductance channel is formed by ANT.

CypD is a mitochondrial matrix protein encoded by Ppif that serves as a chief regulatory component of the mPTP (Elrod and Molkentin, 2013). CypD is poised to orchestrate a host of activities as scaffold to a large number of signaling and structural proteins and through its effects on ETC complex activity and regulation of synthasome assembly; yet, it is most widely appreciated for its regulation of the mPTP (Beutner et al., 2017; Porter and Beutner, 2018). CypD is not a core component of the channel nor does it fundamentally alter pore properties but it does sensitize the pore to opening (Baines et al., 2005; De Marchi et al., 2006). CypD was the first protein identified that firmly established mPTP as a bona fide molecular entity and not simply an artifact of IMM breakdown due to lipid phase, as part of early work identifying CypD as the target of an immunosuppressant capable of suppressing Ca²⁺ and ROS-induced mPT, cyclosporin A (CsA) (Broekemeier et al., 1989; Crompton et al., 1988; Halestrap and Davidson, 1990; Halestrap et al., 1997; Tanveer et al., 1996). Although CypD’s mechanism of action is still largely unknown, it has been shown that CypD binding partners include components of the mPTP, including the F₁F_o ATPase oligomycin sensitive conferring protein subunit (OSCP) (Giorgio et al., 2009). Also proposed is that CypD normally occupies and thus masks an inhibitory phosphate-binding site within the mPTP (Basso et al., 2008). This site is exposed when CypD is displaced by CsA treatment or by Ppif knockout, thereby decreasing mPTP sensitivity to inducers such as Ca²⁺ and ROS. Indeed, a large body of research has leveraged the Ppif knockout mouse and CsA’s ability to modify mPTP sensitivity to understand mPTP composition, modulators, and participation in Ca²⁺ handling and cell death. In fact, the first studies of Ppif knockout mice provided some of the most compelling support for the contribution of CypD to regulation of mPT (Baines et al., 2005; Nakagawa et al., 2005). Ablation of CypD conferred mitochondria with resistance to depolarization and protection from cell death following Ca²⁺ overload or H₂O₂-induced oxidative stress. Further, research from the Molkentin group found that cardiac-specific transgenic mice overexpressing CypD were more susceptible to mitochondrial swelling and cell death, two hallmarks of mPT (Baines et al., 2005). These studies also demonstrated that CsA-sensitive mPT and CypD were responsible for regulating necrotic cell death induced by ROS and Ca²⁺ overload but that apoptotic cell death remained intact in Ppif null cells. Additionally, some of the first in vivo studies to suggest a physiologic role for the mPTP in maintenance of Ca²⁺ homeostasis came from investigation of the Ppif null mouse. With sustained exercise, Ppif knockout resulted in heart failure, which was attributed to impaired Ca²⁺ efflux from the mitochondrial matrix and associated activation of Ca²⁺-dependent dehydrogenases (Elrod et al., 2010). This increased dehydrogenase activity caused by elevated matrix Ca²⁺ impaired the heart’s ability to utilize fatty acids as a mitochondrial fuel source, thus limiting metabolic flexibility necessary to adapt to increased bioenergetic demand under exercise-induced stress. These observations were significant to the mPTP field and pointed to a possible role for mPTP in permitting efflux of Ca²⁺ out of the mitochondria to prevent Ca²⁺ overload via the “flickering” or transient opening of the mPTP, which was first described a decade earlier (Petronilli et al., 1999). Interestingly, these data are also consistent with early reports of CsA’s effect on respiration, noting that CsA caused a large accumulation of Ca²⁺ in the mitochondrial matrix (Fournier et al., 1987; Jung and Pergande, 1985). Importantly, CypD is implicated as a regulator in most models of the mPTP, and its role in mPTP sensitization has served as a unifying litmus test for interrogating mPTP activity since the earliest observations of mPT.

ATP synthase, or mitochondrial complex V of the electron transport chain (EC 7.1.2.2), is a conserved enzyme of the IMM, and its primary function is synthesis of ATP from adenosine diphosphate (ADP). Beyond its role in conversion of electrochemical potential of the mitochondrial membrane to ATP, ATP synthase is now also believed to serve as a key structural component of the high-conductance mPTP (Morciano et al., 2015). ATP synthase consists of the water-soluble F₁ domain that extends into the mitochondrial matrix and the highly polar F_o domain integral to the IMM (subunits c, e, f, and g) (Spikes et al., 2020). The F₁ domain is comprised of a catalytic portion of alpha-beta trimers and a regulative OSCP. Though several models for assembly of an ATP synthase pore exist, two of these warrant comment here. The first model hypothesizes that channel formation at the interface between two ATP synthase monomers serves as the mPTP conducting core (Fig. 1A). The first study to suggest this model demonstrated that ATP synthase dimers, but not monomers, conducted current activated by Ca²⁺ and oxidizing agents when introduced into artificial planar lipid bilayers (Giorgio et al., 2013). mPTP activity dependent upon this dimer model has since been corroborated by other studies, including extension of the concept to pinpoint a critical interface between subunit g and subunit e of two interacting monomers (Carraro et al., 2018; Giorgio et al., 2017). Yet, other studies have argued that monomeric ATP synthase is sufficient to produce mPTP activity (Mnatsakanyan et al., 2019). Evidence support another model in which subunit c of F_o ATP synthase is the core (Alavian et al., 2014; Azarashvili et al., 2014; Bonora et al., 2013) (Fig. 1B). In this model, mPTP forms at sites of ATP synthase dimers but requires dissociation of the dimers to open (Bonora et al., 2017). Detachment of F₁ domains is thought to trigger a conformational change within ATPase. A possible mechanism is that subunit e on the IMS aspect of the IMM interacts with a polar ring-like pore formed across the IMM by subunit c (Pinke et al., 2020). Deformation of the e subunit is proposed to cause the release of a lipid plug that moves out of the barrel-like c-ring, thereby enabling passage of solutes from the matrix into the IMS. This model is still debated, however, as genetic disruption of the c subunit has produced conflicting results. For example, depletion of the three genes that encode the c subunit, ATP5MC1, ATP5MC2, and ATP5MC3 (previously ATP5G1, ATP5G2, and ATP5G3), was shown to reduce OMM rupture, loss of membrane potential, and cell death by mPTP inducers like ionomycin and H₂O₂. (Bonora et al., 2013). Consistent with a role for the c subunit, transient overexpression of ATP5MC1 amplified mPT response. An independent group also deleted ATP5MC1, ATP5MC2, and ATP5MC3, but observed intact mPTP activity, arriving at the conclusion that the c subunit (and other F_o subunits involved in proton translocation, A6L and a) were dispensable (He et al., 2017). Moreover, simulations of c-ring structure from two species, S. cerevisiae and B. pseudofirmus, predicted that the lumen of the c ring is highly hydrophobic, rendering it unlikely to be capable of permitting passage of cations typical of the mPT (Zhou et al., 2017) though more recent cryo-EM may address some of these concerns as the ring appears to distort and widen upon ATPase exposure to Ca²⁺ (Pinke et al., 2020). Neither model is invulnerable to criticism, and some incongruency could be due to complexities of independent mPTP channels. For instance, the observation that loss of the c subunit produces CsA-sensitive channel activity distinct from classical mPT lends support to the notion that CypD-regulated pores with differing conductance exist (Neginskaya et al., 2019).

The adenine nucleotide transporter (ANT) was one of the first components proposed to serve a structural role in the mPTP and has since emerged as a key candidate for the core of an independent low-conductance channel (Fig. 1C). Four ANT isoforms have been identified in human (ANT1–4) and three in mice (Ant1, Ant2, and Ant4) (Ellison et al., 1996; Lim et al., 2011; Rolland et al., 2011; Schiebel et al., 1993). ANT isoforms are encoded by distinct loci in the genome, and their unique expression profiles depend upon tissue type, developmental stage, and cell cycle status (Brower et al., 2007; Cozens et al., 1989; Karch et al., 2019; Ku et al., 1990; Li et al., 1989; Lunardi et al., 1992; Schiebel et al., 1993; Stepien et al., 1992). ANT2 (SLC25A5) and ANT3 (SLC25A6) are ubiquitously expressed in many cell types (Karch et al., 2019; Stepien et al., 1992; Torroni et al., 1990). ANT1 (SLC25A4) is expressed in heart, muscle, and brain; whereas, ANT4 (SLC25A31) is restricted to embryonic stem cells, germ cells, embryonic ovaries, and testis (Karch et al., 2019; Li et al., 1989; Lim et al., 2011; Rolland et al., 2011). Early studies implicated ANT in mediating the mPT and indicated that ANT regulates contact sites between the IMM and OMM, positioning ANT to mediate flux across mitochondrial membranes (Bücheler et al., 1991; Halestrap and Davidson, 1990; LêQuôc and LêQuôc, 1988). ANT later appeared in protein fractions that could generate complexes in artificial proteoliposomes containing mPTP activity (Beutner et al., 1996, 1998). Importantly, pull-down assays designed to identify CypD binding proteins revealed that ANT and VDAC bound to CypD and, when these fractionated complexes were reconstituted into liposomes, they recapitulated properties of the mPTP in a CsA-dependent manner (Crompton et al., 1998). Yet, the first study of ANT knockout mice showed that Ant1;Ant2-doubly deficient cells retained functional mPTP, leading to exclusion of ANT as an essential unit of the pore (Kokoszka et al., 2004). Important work by Molkentin’s group has shown that loss of one isoform results in upregulation of the others, demonstrating that ANT isoforms can compensate for one another and express in multiple tissue types even when atypical for the isoform, with the exception of Ant4 (Karch et al., 2019). Their recent triple knockout of all mouse ANT isoforms, Ant1, Ant2, and Ant4, has since invited speculation that ANT contributes to low-conductance mPTP activity (Karch et al., 2019). This study demonstrated through comparisons of membrane depolarization, Ca²⁺ retention, and mitochondrial swelling that mitochondria lacking all isoforms of ANT were desensitized to Ca²⁺-induced mPTP opening but that mPTP activity was still present at high levels of Ca²⁺. Triple-null mitochondria treated with CsA showed no mPTP activity, as measured by mitochondrial swelling, suggesting that remaining pore activity was dependent upon CypD. Indeed, quadruple deficiency of Ant1, Ant2, Ant4, and CypD (Ppif^−/−) confirmed that all mPTP activity could be accounted for by the ANT family members and CypD. Patch clamping currents of the IMM further revealed that the conductance properties of Ant-triply deficient mitoplasts (mitochondria stripped of the OMM), contained fewer pores that were relatively unresponsive. These data were the first indication that two separate pores might comprise total mPTP activity, one of which was composed of ANT family members and another highly dependent upon CypD (likely representing a complex with ATP synthase depicted in Fig. 1).

6. mPTP in stem cells and development

A growing body of literature has reported critical importance of metabolism in stem cell biology (Shyh-Chang and Ng, 2017). The mPT is implicated in modulating metabolism during cellular differentiation of various stem cells, including pluripotent stem cells and those of the cardiac, neural, hematopoietic, endothelial, and hepatic lineages. Diversity in metabolic requirements of different tissue lineages and cell types makes universal conclusions about how the mPTP dictates stemness and fate choice unlikely, but generally an open state correlates with greater multi-/pluripotency (Fig. 2). Mechanistically, several studies demonstrate that the mPTP shapes stem cell bioenergetics by altering mitochondrial network maturation, inner membrane complexity, OXPHOS, matrix Ca²⁺, and superoxide flashes. Reports detailed below suggest that closing of the mPTP promotes the mitochondrial maturation necessary to meet elevated OXPHOS demands typical in a variety of differentiating stem cells. Still other studies suggest that regulatory components of the mPTP promote glycolysis as the dominant bioenergetic pathway. For example, when Ppif was deleted to desensitize the mPTP, matrix Ca²⁺ was elevated and TCA activity was amplified, favoring glycolytic glucose metabolism over mitochondrial OXPHOS (Tavecchio et al., 2015). CypD is known to exert direct effects on metabolism via transcription of mtDNA-encoded dehydrogenases and ETC enzymes, modulation of oxygen consumption, and regulation of metabolic pathways; thus, some differing conclusions regarding the effects of mPTP inactivation on metabolism could be due to mPTP-independent mechanisms and should be considered with caution (Menazza et al., 2013; Radhakrishnan et al., 2015; Tubbs et al., 2014). The vast majority of studies reveal roles for closing of the mPTP in cellular differentiation, many of which are detailed below. Importantly, one study has shown the reverse, that reprogramming of somatic cells to a pluripotent state can be enhanced by transient enforced mPTP opening (Ying et al., 2018, 2021). Opening of the mPTP triggers a mitochondrial ROS/miR-101c pathway that leads to epigenetic alteration (demethylation) at promoters of pluripotency genes, leading to greater chromatin accessibility. Thus, in addition to playing roles in stem cell differentiation, self-renewal, and survival, the mPTP appears to impact mechanisms of chromatin remodeling that dictate stemness.

Figure 2. Stem cell fate is modulated by the state of the mPTP.

Common themes among the most multi-/pluripotent stem cells is that they generally prefer glycolysis for energy production, have less active mitochondria, exhibit an open-state mPTP, and have greater chromatin accessibility especially at genetic loci important for multipotency. For adult stem cells, this bioenergetic metabolism protects their quiescence and limits oxidative stress. As stem cells proliferate and differentiate, they typically undergo metabolic reprogramming associated with greater oxygen consumption. Closing of the mPTP promotes maturation of mitochondrial machinery and increased utilization of mitochondrial fuels, which differentiating stem cells need to meet elevated demands for ATP generation and other mitochondrial metabolites. Epigenetic modifications reinforce lineage-specific gene expression programs.

Many compelling reports of the critical role that mPT plays in stem biology come from studies of the developing heart. Adaptations in bioenergetics and mitochondria occur during differentiation of cardiac stem cells and reflect normal developmental changes within the maturing embryonic heart (Beutner et al., 2014; Chung et al., 2007, 2008; Porter et al., 2011). Immature cardiac stem cells have fragmented, perinuclear mitochondria that are relatively inactive by comparison to mature cardiomyocytes (Beutner et al., 2014; Porter et al., 2011). Undifferentiated myocytes rely instead upon anaerobic glycolysis. Several reports from George Porter’s group show that the mPTP is open at early stages of cardiomyocyte differentiation (Hom et al., 2011; Lingan and Porter, 2016; Lingan et al., 2017). Age-associated remodeling in differentiating cardiomyocytes of the embryo and neonate produces longer mitochondria with tightly packed cristae, a more-polarized mitochondrial membrane, elevated OXPHOS, and greater energy efficiency marked by lower output of ROS (Hom et al., 2011; Porter et al., 2011). These same adaptations could be accelerated in utero and in vitro by inhibition of mPT with CsA or NIM811, thereby improving cardiomyocyte differentiation and function (Hom et al., 2011; Lingan and Porter, 2016; Lingan et al., 2017). Independent studies have corroborated the role of mPTP closure in cardiomyocyte differentiation and the importance of mitochondrial maturation in fate determination from mouse and human pluripotent stem cells (Cho et al., 2014; Fujiwara et al., 2011; Yan et al., 2009). Inhibition of the mPTP by CsA or NIM811 increased expression of genes required for mitochondrial activity and improved several metrics of mitochondrial function, including OXPHOS activity, membrane potential, and ATP generation (Cho et al., 2014). Interestingly, authors observed synergistic benefit of quenching ROS in differentiation cultures such that concurrent treatment with CsA and antioxidants Trolox or NAC enhanced the fraction of cells committing to the cardiomyocyte fate (Cho et al., 2014). Enhanced myogenesis was highly dependent upon closure of the mPTP, as antioxidants alone had no effect on frequency of cells expressing markers of the cardiomyocyte lineage.

Literature also strongly support a role for the mPTP in the nervous system. Proper neurological development relies upon tight regulation of metabolism in neural stem and progenitor cells. As newly formed neurons mature, they undergo dynamic changes in metabolic pathway utilization and fuel source (Mattson et al., 2008). Indeed, neural progenitors switch from anaerobic glycolysis to aerobic OXPHOS during differentiation (Candelario et al., 2013). These findings are consistent with increases in mitochondrial superoxide flashes observed during neuronal differentiation, which appear to be critical for promoting fate commitment and exit from the cell cycle (Hou et al., 2012). Hou and colleagues found that limiting mitoflash frequency with ROS scavengers or mPTP inhibitors elevated progenitor proliferation; whereas, increasing superoxide flashes promoted neural differentiation. These data demonstrate that transient flickering of the mPTP, producing brief superoxide flashes, negatively regulates self-renewal of neural progenitor cells in conjunction with activation of differentiation programs and signaling that inhibits proliferation. The mPTP can also cause death of neural stem cells in response to stressors, such as anesthetics or other toxic compounds. The neurotoxicity of anesthetics is poorly understood but, consistent with typical triggers of mPTP opening, is predicted to be linked to elevated Ca²⁺ signaling, ROS, and neuroinflammation (Bai et al., 2013; Orrenius et al., 2003). For example, neurons derived from human embryonic stem cells have been shown to undergo mPTP-dependent cell death upon response to supratherapeutic doses of anesthetic (Twaroski et al., 2015). Propofol at high doses induced activation of fission protein Drp1 (phospho-Ser616) and its respective regulator cyclin dependent kinase 1, leading to death of neurons (Twaroski et al., 2015). Treatment with an inhibitor of mitochondrial fission, mdivi-1, prevented cell death in conjunction with delay of mPTP opening and reduction of mitochondrial depolarization and fragmentation. Similarly, chronic activation of macrophages of the central nervous system, known as microglia, can cause death of healthy neurons and progressive degenerative neurological disorders through production of ROS, inflammatory cytokines, and chemokines. In a study highly suggestive of mPTP regulation in neural stem cells, overexpression of Hsp75 decreased apoptosis and preserved mitochondrial membrane potential. It was found that Hsp75 overexpression inhibited formation of CypD-dependent mPTP, thereby suppressing activation of a mitochondrial mediated cell death cascade initiated by microglia (Wang et al., 2015). Together, these studies suggest that, in addition to determining fate selection and proliferation, regulation of the mPTP is critical for cell survival of neural stem and progenitor cells.

Metabolic adaptation is also critical for differentiation of blood lineages and development of the immune system. A number of studies have found that mPT activity can modulate self-renewal of blood stem cells and subsequent maturation of their progeny. Throughout adulthood, blood cells are replenished by hematopoietic stem and progenitor cells that reside in the bone marrow. Any perturbation in the balance of self-renewal, proliferation, or maturation of these cells can result in hematologic malignancies, cytopenias, anemia, and infections. Hematopoietic cell transplantation remains the most common and curative stem cell therapy in the clinic; yet, collection and culture of donor hematopoietic stem cells can be problematic due to the detrimental effects of oxidative stress on stem cell self-renewal. Studies led by Hal Broxmeyer’s group found that exposure of hematopoietic stem and progenitor cells to normoxia outside the body produced superoxide flashes that impaired regenerative function. Reports describe loss of regenerative potential of adult hematopoietic stem cells exposed to normoxia in a process termed extra physiologic oxygen shock/stress (Broxmeyer et al., 2015). These findings are complemented by studies showing that the damaging effects of ambient air could be forestalled by processing of adult mouse bone marrow or human cord blood with CsA. Donor cells processed with CsA exhibited enhanced numbers of phenotypically identified hematopoietic stem cells and functional competitive repopulating activity (Broxmeyer et al., 2015; Mantel et al., 2011, 2015). The mechanism of enhanced engraftment with CsA treatment is believed to derive from temporary protection from ROS and/or genotoxic stress associated with hours of handling in normoxia. Interestingly, even earlier studies pointed to a role for the mPTP in the blood lineage. The receptor tyrosine kinase c-Kit is a prototypical surface marker of hematopoietic stem cells, but is also expressed on stem cells of the gut, germline, nervous system, and melanocytes. Hallmarks of mPT, including depolarization of membrane potential and ROS production, could be induced in an erythroleukemia cell line engineered to activate p53-induced apoptosis (Lee, 1998). Apoptosis, loss of membrane potential, and ROS generation could all be suppressed in these cells by stimulation of c-Kit with its cognate ligand SCF, suggesting that signaling downstream of c-Kit is important for protecting hematopoietic cells from mPT-dependent cell death. CsA has been used in a vast number of studies of the immune system and as a therapy to suppress graft-versus-host disease after transplantation, largely due to its inhibition of T-cell receptor signaling through calcineurin (Flores et al., 2019; Shevach, 1985). Conclusions from studies using CsA for its immunosuppressive activity must be made with caution, but it is worth noting that collective evidence supports a role for mPT in B cell lymphopoiesis. During B cell development, CsA interrupts differentiation of B-1 lymphocytes, a highly specialized subset of B cells, but promotes production of conventional B-2 lymphocytes (Arnold et al., 2000). Similar enhancement by CsA has been observed in NK cell differentiation, though mechanism is not fully understood (Flanagan et al., 1999; Kosugi and Shearer, 1991). Genetic models have also been used in isolation, but can be suggestive of a role for mPT. One likely component of the low-conductance mPTP, ANT, appears to be required for development of B lymphocytes and red blood cells in embryogenesis. Curiously, hypomorphic deletion of Slc25a5 which encodes Ant2 in these lineages was accompanied by opening of the mPTP, reduction in respiration capacity and ATP production, elevated ROS, and increase in death of B cells and erythroid cells (Cho et al., 2015). Data supporting a role for mPTP in maturation of developing red blood cells have been corroborated by other independent genetic studies in mouse and human. Deletion of the mitochondrial chaperonin Hsp60, encoded by Hspd1, in hematoendothelial precursors of the mid-gestation mouse embryo causes anemia and other vascular defects (Duan et al., 2019). Authors found reduction in mitochondrial membrane potential and decreased expression of VDAC in erythrocytes emerging from the yolk sac. CsA was able to significantly decrease mPTP-dependent cell apoptosis and partially restored VDAC expression. In another study, it was shown that human hematopoietic stem cells that overexpress the fission factor FIS1 undergo arrest in maturation toward the red blood cell lineage, resulting from fragmentation of the mitochondrial network, impairment in mitochondrial membrane potential, decreased ETC complex abundance, and elevated ROS production (Gonzalez-Ibanez et al., 2020). Treatment of these cells with CsA rescued mitochondrial morphology and restored the cells’ ability to properly differentiate and synthesize hemoglobin, lending additional support to the notion that the mPTP could regulate metabolic reprogramming required for maturation of blood lineages.

Endothelial cells are key constituents of blood vessels, and circulating endothelial progenitor cells have attracted great attention for their ability to populate and repair damage to the vascular system (Yoder, 2012). From a regenerative medicine perspective, differentiation of mature endothelium from human induced pluripotent stem cells is attractive, yet remains challenging due to incomplete development of functional features critical to flow sensing in the vasculature. A recent study showed that one of these serious maturation defects – failure to synthesize a glycocalyx – was caused by incomplete mitochondrial maturation due to a constitutively open mPTP (Tiemeier et al., 2019). Generation of this “hairy” polysaccharide surface of the apical surface of the endothelium could be stimulated by enforced closure of the mPTP with CsA. mPTP closure improved mitochondrial function and enabled production of functional glycocalyx capable of flow sensing typical of mature vascular endothelial cells from the body. Other studies generally corroborate an important role for cyclophilins and possibly mPTP opening in endothelial cells through examination of the effects of CsA and its analogs on vascular and endothelial cell development. Whether the mPTP must be open or closed for proper regulation of these processes is unclear, but a body of evidence suggests that some mPTP activity during the course of cell maturation could be important. For example, an early study of embryonic stem cell differentiation showed that CsA shifted fate commitment away from the endothelial cell (CD31⁺) lineage in favor of cardiomyocyte progenitors (Yan et al., 2009). Additionally, angiogenic sprouting and proliferation of endothelial cells and their progenitors are impaired by CsA and an analog that does not inhibit calcineurin/NFAT signaling, N-Methyl-valyl-4-cyclosporin A (Davies et al., 2005; Nacev et al., 2011). Further, CsA treatment in utero causes collapse of the cardiovascular system in developing embryos (Pandey et al., 2015). CsA treatment results in progressive reduction in blood flow and disappearance of luminal structures in the vasculature. In this study, changes were attributed to CsA’s blockade of Jagged 1-induced Notch activation. Important to note is that CsA and its analogs inhibit multiple members of the cyclophilin family and could interfere with mPTP-independent prolyl isomerase functions such as protein folding. Thus, important limitations of many of these endothelial cell studies are that CypD and mPTP were not directly evaluated. Future work should be aimed at careful examination of mPT in endothelial progenitors during fate commitment and maturation.

Lastly, evidence points to roles for the mPTP in progenitor cell survival and regeneration in the liver. In a study of drug toxicity, induced pluripotent stem cells were derived to model a neurometabolic disorder caused by mutations in mitochondrial DNA polymerase gamma (POLG), Alpers-Huttenlocher Syndrome (AHS). Patients with AHS commonly receive valproic acid for control of epilepsy, but acute liver failure caused by genotoxic stress can lead to serious morbidity. Authors found that mPTP and superoxide flashes occurred more frequently in hepatocyte-like cells derived from AHS patients. Further, AHS cells undergo cell death in response to treatment with valproic acid, which could be blocked by CsA, highlighting the mPTP as a potential target for preventing hepatotoxicity of this therapy for AHS patients (Li et al., 2015). Stress response and regulation of metabolism in the liver is highly dependent upon management of cytosolic and mitochondrial Ca²⁺. Deficiency of MICU1, the Ca²⁺-sensing regulator of the MCU_cx, in hepatocytes sensitizes livers to Ca²⁺ overload, impairs bioenergetics and cell functions, increases mPTP opening, and subsequently results in liver failure (Antony et al., 2016; Pan et al., 2013). Importantly, mPTP inhibition by NIM811 is sufficient to accelerate hepatocyte proliferation and rescue liver regeneration in mice challenged by partial hepatectomy (Antony et al., 2016). In summary, a growing body of literature centered on stem cell biology implicates the mPTP as an important component of the stem cell’s toolkit for adapting to evolving metabolic demands associated with fate commitment and regeneration.

7. Concluding Remarks

The mPTP is renowned for inducing programmed cell death and necrosis in response to ischemia-reperfusion injuries. Yet, sustained or transient opening of the mPTP is a crucial parameter in deciding cell fate. Long-term opening leads to bioenergetic collapse, release of cytotoxic matrix molecules, oxidative stress, and cell death; whereas, short-term flickering results in variable conductance of the channel and is purported to support protective intracellular signaling by release of matrix Ca²⁺ and ROS. Emerging evidence supports a central role for mPTP modulation in metabolic programming and survival signaling required for stem cell self-renewal and tissue repair. Although stem cell studies represent a small fraction of all mPT research, new clarity regarding the molecular identity and structural regulation of the mPTP will undoubtedly accelerate this budding area of research and enable new capabilities in stem cell engineering and therapeutics for regenerative medicine.

Abbreviations

AHS

Alpers-Huttenlocher Syndrome

ANT

Adenine nucleotide translocator

CsA

Cyclosporin A

CypD

Cyclophilin D

ETC

Electron transport chain

IMM

Inner mitochondrial membrane

IMS

Intermembrane space

MCU_cx

Mitochondrial calcium uniporter complex

mPT

Mitochondrial permeability transition

mPTP

Mitochondrial permeability transition pore

NIM811

(Melle-4)cyclosporin

OMM

Outer mitochondrial membrane

OSCP

Oligomycin sensitive conferring protein subunit

OXPHOS

Oxidative phosphorylation

Pi

Inorganic phosphate

ROS

Reactive oxygen species

SOD

Superoxide dismutase

TCA

Tricarboxylic acid cycle

VDAC

Voltage-dependent anion channel