Non-Conventional Mitochondrial Permeability Transition: Its Regulation by Mitochondrial Dynamics

Abstract

Mitochondrial permeability transition (MPT) is a phenomenon that the inner mitochondrial membrane (IMM) loses its selective permeability, leading to mitochondrial dysfunction and cell injury. Electrophysiological evidence indicates the presence of a mega-channel commonly called permeability transition pore (PTP) whose opening is responsible for MPT. However, the molecular identity of the PTP is still under intensive investigations and debates, although cyclophilin D that is inhibited by cyclosporine A (CsA) is the established regulatory component of the PTP. PTP can also open transiently and functions as a rapid mitochondrial Ca²⁺ releasing mechanism. Mitochondrial fission and fusion, the main components of mitochondrial dynamics, control the number and size of mitochondria, and have been shown to play a role in regulating MPT directly or indirectly. Studies by us and others have indicated the potential existence of a form of transient MPT that is insensitive to CsA. This “non-conventional” MPT is regulated by mitochondrial dynamics and may serve a protective role possibly by decreasing the susceptibility for a frequent or sustained PTP opening; hence, it may have a therapeutic value in many disease conditions involving MPT.

Introduction

Mitochondria play a critical role in eukaryotic cells, most notably, providing cellular energy in the form of ATP via oxidative phosphorylation (OXPHOS). Complete oxidation of carbon-containing organic nutrients to CO₂ occurs through the Krebs cycle in mitochondria, during which reducing equivalents (NADH and FADH₂) containing high-energy electrons are generated. The electron transport chain (ETC) in the inner mitochondrial membrane (IMM) converts high-energy electrons into a proton motive force across the IMM that is used by the ATP synthase for ATP production. The essential requirement for the OXPHOS is the well “sealed” IMM to maintain the electrochemical gradient created by the ETC. Maintaining the proper mitochondrial membrane potential (MMP) is also necessary for in-and-out of ions and metabolites. Therefore, the IMM maintains relative impermeability, and permits selective passages of ions and metabolites only through the specified channels, carriers, and transporters in regulated manners.

Mitochondrial permeability transition (MPT) is a phenomenon that the IMM loses selective permeability. It is thought to occur via the opening of a high conductance mega-channel called permeability transition pore (PTP), which allows free passage of molecules of up to 1.5 kDa in size. The phenomenon of MPT was initially appreciated as mitochondrial swelling and a loss of membrane integrity in the presence of Ca²⁺ in isolated mitochondria [1–4]. However, it is relatively recent that MPT drew attention due to its involvement in cell death in pathology. MPT often results in necrotic cell death because of the cellular energy crisis caused by the inability of mitochondria to produce ATP [5]. MPT is also involved in apoptosis. Mitochondrial swelling from massive influx of water during MPT can cause rupture of the outer mitochondrial membrane (OMM), which releases apoptotic effectors residing in the intermembrane space to the cytosol [6].

Despite decades of studies, the molecular identity of PTP is still not fully understood. The original model involving voltage-dependent anion channel (VDAC) of the OMM, adenine nucleotide translocase (ANT) of the IMM, and cyclophilin D (CypD) of the matrix was widely accepted for a while until the studies using mouse gene knockout (KO) technology. Gene KO studies demonstrated that VDAC and ANT are dispensable for MPT, although it requires higher Ca²⁺ concentrations to induce MPT in ANT KO [7, 8]. Most notably, CypD deletion greatly decreases PTP opening [9–12], establishing that CypD, which is inhibited by cyclosporine A (CsA), is the major regulator of PTP, albeit not a core component of the pore. Hence, PTP is conventionally defined by CsA-inhibitable large-conductance non-selective channel.

More recently, additional models have been proposed for PTP [13–15]. However, debates still continue about which protein (or its subunit) is responsible for the formation of PTP or whether there are multiple PTPs. Considering the long history of PTP studies and a recent development in the renewed identity of the PTP, there are many reviews that the readers can refer to [16–19]. In this short article, we will briefly describe the current state of conventional PTP, and then focus on a less-appreciated CsA-insensitive form of MPT that is regulated by mitochondrial dynamics.

Mitochondrial Permeability Transition Pore

Original model

Electrophysiological studies demonstrated that PTP opens in multiple sub-conductance states [20, 21], suggesting that the PTP has a multi-subunit structure that assembles into different oligomerization states. The initial clue for the molecular identity of the PTP was the sensitivity of MPT to atractyloside and bongkrekic acid that bind to the ANT [22]. ANT was shown to form complex with VDAC at the contact site along with the peripheral benzodiazepine receptor (now known as translocator protein of 18 kDa (TSPO)) [23–25]. VDAC as a PTP component was also suggested from the patch-clamp observation that the maximal conductance of PTP was twice of the VDAC conductance and that the most abundant sub-conductance state of the PTP happened to be the half-maximal state [26, 27]. Inhibition of MPT by CsA was known early on [28–30], and the CsA target protein CypD in the matrix was considered a regulatory component of the PTP. This series of information formulated an idea that VDAC and ANT form a core pore complex that is regulated by TSPO in the OMM and CypD in the matrix and possibly further associated with additional proteins found at the contact site including mitochondrial creatine kinase, hexokinase, and Bcl-2-related proteins [31, 32]. As indicated above, studies of genetic deletions of core pore components ANT and VDAC, as well as TSPO disproved this hypothetical model [7, 8, 33], but CypD remains as a main regulator of the PTP [9–12]. Indirect evidence also showed that the VDAC is a dispensable part of the PTP [34]. Inorganic phosphate sensitizes Ca²⁺-induced PTP opening [29]. Phosphate carrier (PiC) in the IMM binds to CypD in a CsA-sensitive manner and was also proposed to be a part of the PTP complex [35]. However, genetic manipulation of PiC did not affect PTP [36, 37]. Another report showed that PiC KO does not block PTP opening, but increases resistance to MPT, indicating that it is not the pore-forming component [38]. Since then, renewed efforts to identify PTP components resulted in new models involving the ATP synthase and spastic paraplegia 7 (SPG7).

Newly proposed models of PTP

It was found that CypD binds to oligomeric forms of the ATP synthase in a CsA-sensitive manner [39], which resembles the regulation of PTP by CypD. Follow-up studies showed that purified dimers of ATP synthase reconstituted in the lipid bilayer exhibit the electrophysiological characteristics of the PTP, and proposed that ATP synthase dimers form the PTP [13]. Oligomycin sensitivity conferring protein (OSCP) of the peripheral stalk was shown to be the site of CypD binding. In this model, the channel is formed at the interface of the two ATP synthases that is formed by a collection of small subunits of the Fo complex linked to the peripheral stalk [40, 41]. However, genetic deletions of the OSCP, the b subunit of the peripheral stalk, or other components of the ATP synthase dimerizing domain did not prevent PTP opening, disputing this model [42–46].

The ATP synthase as a PTP was also proposed by another group. In this case, however, the c-ring of the Fo complex was proposed to be the PTP [14]. An earlier study demonstrated that depletion of ATPase c subunit blocks MPT to the similar extent to CsA, and decreases cell death and neuronal excitotoxicity, suggesting a critical role of the ATPase c subunit in MPT [47]. In this new study, reconstituted c-subunit ring was shown to form a voltage-sensitive channel that has a current similar to that of PTP. Similarly, silencing c subunit in cells prevented ionomycin-induced MPT, mitochondrial depolarization, and cell death in a CsA-sensitive manner [14]. It was proposed that the c-ring becomes the PTP when the F1 complex dissociates from the Fo complex, which occurs with sustained high matrix Ca²⁺ concentration [14]. However, this model has also been disputed by the study employing gene KO of the three genes encoding c-subunits. The cells completely lacking c-subunits were shown to maintain PTP [46].

An shRNA-based loss-of function screening for PTP yielded several candidates for PTP components [15]. Among those, shRNAs for SPG7, VDAC1, and CypD were shown to decrease ROS-sensitized, Ca²⁺-induced PTP opening. This study showed that SPG7 interacts with VDAC1 and CypD and proposed that this tripartite complex forms the PTP at the contact sites [15]. However, SPG7 depletion did not completely block PTP opening, but decreased it, similar to CypD KO, indicating that SPG7 may be a regulatory component instead of a pore component of the PTP. SPG7 is a subunit of m-AAA protease of the IMM. In another study, however, it was shown that the loss of m-AAA induces an accumulation of the constitutively active mitochondrial calcium uptake channel MCU complex, resulting in Ca²⁺ overload and PTP opening [48]. This study further showed that depletion of SPG7 has no effect on Ca²⁺-induced PTP opening [48]. Therefore, SPG7 as a putative PTP is still far from conclusive.

Low conductance transient PTP (tPTP)

MPT by PTP opening is considered a terminal process that ends up with cell death as mentioned at the beginning. However, PTP was shown to open in multiple sub-conductance states by electrophysiological studies [20, 21]. In addition to the high conductance maximal state, PTP was observed to operate reversibly, opened and closed at a half-maximal state in a very high frequency [27, 49]. The bursting “flickering” of the half-conductance activity of the PTP was thought to be due to a cooperative nature of the mega-channel opening and/or a potential binary structure of the PTP comprised of two pore units [26, 27]. This transient low-conductance mode of PTP opening (tPTP) has been shown to function as a Ca²⁺-releasing channel subsequent to mitochondrial Ca²⁺ uptake [50–53]. In isolated mitochondria, tPTP induces transient depolarization of mitochondria, causing Ca²⁺ release [50]. The known inhibitors of PTP opening, CsA and ADP, completely blocked transient depolarization and Ca²⁺ release, indicating that the PTP opening is responsible for this event [50, 53]. In cells, tPTP was responsible for Ca²⁺ efflux following mitochondrial uptake of Ca²⁺ released from the ER/SR, suggesting that tPTP plays a role in the amplification of intracellular Ca²⁺ signals [50, 53].

Mitochondrial dynamics and MPT

Mitochondrial dynamics is the cellular process describing dynamic changes in mitochondrial shape and location. Mitochondrial fission and fusion are the main processes that regulate mitochondrial size and number, and are mediated by mitochondria-associated dynamin-related large GTPases. Dynamin-like/related protein 1 (DLP1/Drp1) is a cytosolic protein that is recruited to the mitochondrial outer surface where it constricts and divides mitochondria for fission. Mitochondrial fusion requires two separate machineries for fusion of OMM and IMM. Two isoforms of mitofusin, Mfn1 and Mfn2, in the OMM and the optic atrophy 1 (OPA1) protein in the IMM mediate fusion of respective membranes. The mechanisms and additional factors necessary for mitochondrial fission and fusion can be found in recent review articles [54–59].

Mitochondrial fission and fusion are implicated in pathological conditions, as they are involved in regulating mitochondrial function and cell death. As mentioned earlier, MPT can lead to necrosis and apoptosis. It was first observed that mitochondria in cells undergoing apoptosis were fragmented and swollen, which was prevented by inhibition of mitochondrial fission, potentially implicating mitochondrial fission in MPT [60]. Additional studies showed that ischemia-reperfusion (I-R) injury and heart failure (HF) as well as neuronal excitotoxicity where the MPT is a major cell death mechanism were reduced by inhibiting fission (Drp1 siRNA / Drp1-K38A dominant negative / Drp1 chemical inhibitor mdivi-1) or increasing fusion (Mfn / OPA1 overexpression) [61–65]. Therefore, general notion is that mitochondrial fragmentation by fission sensitizes to cell death [66]. Although preventing mitochondrial fragmentation by fission/fusion manipulations is protective in pathological conditions, only limited studies directly assessed MPT in altered mitochondrial fission or fusion. It was found that expressing the dominant-negative fission mutant Drp1-K38A or the fusion protein Mfn1 or Mfn2 in cardiac cells decreased MPT and cell death upon simulated I-R injury [61]. A similar observation was made in HF model [65]. However, this study used a loss of MMP as a surrogate marker of MPT, which is not a direct measure of the MPT. Quenching of mitochondrial calcein fluorescence by cobalt upon MPT is a well-established method to test the MPT [67, 68]. Using this assay, it was shown that inhibition of mitochondrial fission by Drp1-K38A decreases MPT and cell death in hyperglycemia [69].

These studies demonstrated that mitochondrial dynamics plays a role in MPT, apparently acting upstream of PTP opening. Reactive oxygen species (ROS) is a major permissive factor for PTP opening [70–73]. Our studies showed that increased mitochondrial ROS levels in hyperglycemic conditions increase MPT and cell death, and that inhibiting mitochondrial fission blocks the ROS increase and MPT [69, 74]. Similar experiments in prediabetic heart demonstrate an increase in ROS production, which trigger mitochondrial swelling together with changes in fission/fusion protein expression, leading to cell death [75, 76]. As I-R injury and excitotoxicity also involve mitochondrial ROS increase, the reduced MPT by fission inhibition may be due to indirectly decreasing ROS production, not directly acting on PTP components. In another study, it is shown that overexpression of Drp1 decreases mitochondrial calcium retention capacity (mCRC) in permeabilized cells, indicating that increasing Drp1 levels promotes MPT [77]. This finding raises the possibility that Drp1 may directly regulate PTP opening; however, it is also plausible that the effect is secondary to Drp1 influencing on the mitochondrial Ca²⁺ influx mechanism. A recent study suggested that Drp1 promotes hypoxia-induced PTP opening by binding to potential PTP constituents, Bax and PiC, and releasing hexokinase II from the PTP [78]. Bax, PiC, and hexokinase have been suggested to be regulatory proteins for PTP; therefore, it is possible that Drp1 directly interacts with components in the PTP supercomplexes, and thereby changes pore opening probability. Further studies in identifying the interacting proteins in the PTP complexes (interactome) may help to elucidate and distinguish the direct and indirect effects of Drp1 in MPT (Fig. 1).

Figure 1. Mitochondrial dynamics and MPT.

Mitochondria undergo fission-fusion cycle. Various insults including hyperglycemia (HG), ischemia-reperfusion (I-R), and neuronal excitotoxicity drive fission and cause mitochondrial ROS overproduction and Ca²⁺ overload, leading to PTP opening and MPT for cell injury. During hypoxia, the fission protein Drp1 may interact with Bax and phosphate carrier (PiC), and lead to a release of hexokinase II (HKII) from mitochondria, which induces PTP opening. Bax may prevent fusion, allowing mitochondria to remain fragmented.

Transient loss and recovery of MMP has been observed in cells and isolated mitochondria [73, 79–85]. This ‘flickering’ of trans-membrane potential is reminiscent of aforementioned tPTP. However, there are mixed conclusions regarding its dependence on ROS, Ca²⁺ or MPT. We reported earlier that mitochondrial fission plays a profound role in regulating mitochondrial function by controlling this transient mitochondrial depolarization [86, 87]. We will describe a possibility of a new form of tPTP regulated by mitochondrial dynamics, and discuss potential mechanisms of how it controls mitochondrial function and thus its application for therapeutic strategy.

Transient depolarization of mitochondria: MMP flickering

Transient depolarization of mitochondria has been observed since late 1990s using the mitochondrial potentiometric probe tetramethylrhodamine ethyl ester (TMRE) or its methyl version (TMRM). Some of these observations attributed the transient depolarization to an artifact of ROS generated from excited dye molecules interacting with oxygen under high energy light [73, 82, 83, 88, 89]. In these cases, tPTP facilitated by ROS is responsible for transient depolarization and thus sensitive to CsA. In other studies, however, transient mitochondrial depolarization was shown to be spontaneous, and independent of ROS, Ca²⁺, or MPT [79, 84, 85]. In addition, spontaneous mitochondrial pH flashes occur concurrently with transient mitochondrial depolarization [90]. This matrix pH flash reflects a compensatory increase in proton pumping immediately following the spontaneous loss of MMP, so that the proton motive force can be maintained. The transient mitochondrial depolarization coupled to pH flash was shown to be independent of ROS, Ca²⁺, and MPT [90]. These observations indicate that there likely exist different forms of transient mitochondrial depolarization, and that, aside from the light-induced artifactual phenomenon, spontaneous flickering of MMP is a normal cellular process. Then, why do mitochondria do this? What is the role of mitochondrial dynamics in this?

MMP flickering: why and how?

Spontaneous transient depolarization (MMP flickering) is an intrinsic property of mitochondria [79, 84–87, 90, 91]. The MMP generated by electrogenic proton efflux by the ETC is the major component of the proton motive force that drives ATP synthesis [92]. Therefore, maintaining proper MMP is critical for mitochondrial and cellular health, as it supports not only ATP synthesis, but also in-and-out of ions and metabolites across the IMM, and mitochondrial protein import. Barring inhibitory conditions, the ETC is a self-regulating system, which strives to maintain the MMP within an adequate range, by accelerating the electron transport and proton extrusion when MMP is low whereas decelerating when it is high [93, 94]. An example of this property, in their extremities, is the experimental evaluation of the ETC activity by oxygen consumption measurement, in which the maximum capacity of electron transport is measured in the presence of an uncoupler (e.g. with FCCP), and the proton leak by blocking the proton channel activity of the ATP synthase (e.g. with oligomycin). Mitochondria do not function properly when their MMPs remain outside of the adequate range for a prolonged period. While sustained low MMP is unable to support mitochondrial function, excessively high MMP (hyperpolarization) is also harmful, because hyperpolarization increases the effective concentration of free radical semiquinones (QH•) at CoQ, FMN, and FAD of the ETC that can directly react with oxygen for ROS production [95, 96]. Mitochondrial hyperpolarization also increases the risk of Ca²⁺ overload for MPT. Therefore, one can envision that spontaneous transient depolarization of mitochondria is the cellular protective mechanism to prevent harmful ROS overproduction and mitochondrial Ca²⁺ overload caused by mitochondrial hyperpolarization.

MMP flickering is stochastic, occurring in random frequency and intervals [84, 86, 87]. Furthermore, there is no correlation between MMP flickering and the extent of MMP, as we observed that some mitochondria with low MMP flickers more robustly than those with high MMP and that some mitochondria with high MMP were quiescent [86, 87]. These observations suggest that individual mitochondria in a cell are in different energetic states. Indeed, in cardiac and skeletal muscle cells, MMP differs among the different subpopulations of mitochondria as well as their energetic characteristics and its Ca²⁺ and ROS regulation [97–99]. How this spatial difference in MMP is achieved is unclear. However, it is likely that the intracellular compartmentalization of mitochondria and thus MMP is dependent on spatial relationships with other organelles and localized metabolic processes under different energy demands, as well as on the cytoskeletal system. For example, Ca²⁺ is an important activator of mitochondrial metabolism [100, 101]; therefore, the energetic state of mitochondria in the vicinity of the ER and plasma membrane where locally high concentration of Ca²⁺ is available would be different from that in mitochondria away from them. The same would be true for mitochondria in the metabolic carrier-enriched areas. Furthermore, cytosolic Ca²⁺ regulates mitochondrial motility, and high Ca²⁺ concentrations stop mitochondrial movement [102], indicating that mitochondria may move to the high Ca²⁺ areas where they remain and provide energy locally [103]. Thus, dynamic changes in localized mitochondrial energetic states would contribute to the intracellular heterogeneity of the MMP.

As to how the MMP flickering occurs, we can speculate that the MMP above a ‘set ceiling’ for the maximum tolerable MMP evokes spontaneous transient depolarization. Indeed, we showed that transient mitochondrial depolarization occurs with a buildup of MMP through active electron transport [87]. Furthermore, oligomycin treatment that increases the MMP activates MMP flickering, whereas the complex III inhibitor antimycin A decreases it [87, 104], indicating that MMP flickering is likely caused by elevation of MMP. As just described, mitochondrial energetic states differ locally inside a cell. Therefore, it is likely that individual mitochondria have their own ‘set ceilings’ for the MMP, and a buildup of MMP above the set MMP ceiling would evoke transient depolarization (Fig. 2A). Individual mitochondria in a cell experience fluctuations of metabolic fluxes because of spatial and temporal differences of metabolic inputs [99]. Accordingly, MMPs of individual mitochondria also fluctuate. The randomness of the MMP flickering is likely from the differences in MMP ceilings as well as metabolic states and their fluctuations in individual mitochondria.

Figure 2. MMP flickering and the impact of mitochondrial connectivity on mitochondrial energetics.

(A) Fluctuation of metabolic input into mitochondria results in temporal changes in MMP. When MMP goes above a hypothetical MMP upper limit (green dashed line), mitochondria become depolarized to prevent Ca²⁺ overload and ROS production. (B) TMRE images of primary hepatocytes with and without Drp1-K38A expression. Ball-shape mitochondria that are mostly discrete are prevalent in control cells, and they undergo small, isolated flickering (arrowhead). Inhibiting fission by expressing Drp1-K38A induces mitochondrial interconnection and enlargement. Large-scale MMP flickering occurs in fission-deficient cells, as all connected mitochondria undergo simultaneous loss and recovery of MMP (bottom panels). Partly reproduced from reference #86. (C) Drawing illustration of (B). Spontaneous MMP flickering is a normal cellular process. Isolated MMP flickering in discrete mitochondria (grey mitochondrion) represents normal level of proton leak. When mitochondria become interconnected by fusion, all connected mitochondria are simultaneously depolarized during MMP flickering, which is reflected as increased proton leak in respiration measurement of a cell population.

The MMP flickering may possibly be the manifestation of the aforementioned tPTP that induces mitochondrial depolarization and functions as a mitochondrial Ca²⁺ releasing mechanism. Unlike tPTP, however, the spontaneous MMP flickering is insensitive to CsA [84, 86, 87, 90, 99], suggesting a distinct cellular process and possibly a novel form of tPTP. Our studies demonstrated that mitochondrial connectivity regulated by mitochondrial dynamics has a profound impact on MMP flickering and thus mitochondrial function.

The role of mitochondrial dynamics in MMP flickering and cellular energetics

While we observed rare but consistent occurrence of MMP flickering in normal conditions, the inhibition of mitochondrial fission, thereby increasing mitochondrial interconnection drastically augments it [86, 87, 104]. It is known that a single mitochondrion is electrically connected regardless of its length and size [105, 106]. In normal cells, spontaneous depolarization occurs in discrete mitochondria as localized events (Fig. 2B control cell). In contrast, with fission inhibition, many of mitochondria become interconnected, and thus, local depolarization spreads to all connected mitochondria (Fig. 2B Drp1-K38A cell), causing large-scale MMP flickering [86, 87, 104]. Therefore, mitochondrial interconnection amplifies small local flickering into large-scale MMP flickering (Fig. 2B and C).

In mitochondria with well-sealed IMM, translocation of protons through the ATP synthase to the matrix powers the phosphorylation of ADP, representing a “coupled” respiration. Protons can “leak” to the matrix away from the ATP synthase, decreasing the OXPHOS coupling (“uncoupling”). Proton leak is measured by leak respiration, which is the respiration (oxygen consumption) in the presence of oligomycin that blocks the proton translocation through the ATP synthase. As protons leak to the matrix, proton gradient and MMP become lower, and mitochondria are depolarized. Oligomycin does not block MMP flickering, indicating that the transient depolarization in MMP flickering occurs by proton leak. In normal cells, at any given instance, the population of mitochondria undergoing transient depolarization is small, having little impact on overall cellular energetics, which represents the normal cellular activity of the ETC. Upon increasing mitochondrial interconnection, however, the spatial amplification of MMP flickering in individual cells is collectively manifested as an increase of proton leak (i.e. respiration uncoupling) in respiration measurement of a cell population (Fig. 2C) [86, 87]. An increase of proton leak accelerates electron transport, and thus, decreases the chance of ectopic electron-oxygen reaction, preventing superoxide production in the ETC. This is likely the reason why inhibition of mitochondrial fission (or enhancing fusion) decreases ROS and is protective in hyperglycemic conditions, cardiac I-R, neuronal excitotoxicity and other pathologies [61–65]. Therefore, mitochondrial connectivity that is determined by mitochondrial dynamics is an important factor that regulates the OXPHOS activity.

Does MMP flickering represent a novel form of tPTP?

Genetic studies have thus far disputed all models of PTP including the recently proposed ones – ATP synthase dimer, c-ring of Fo-ATP synthase, and SPG7 [42, 43, 46, 48]. In addition, the presence of an “unregulated” PTP (versus conventional regulated PTP) that is CypD-independent and is not regulated by Ca²⁺ has been reported [107]. It was also shown that CsA-insensitive PTP or channels can be formed by experimental addition of palmitic acid, prooxidants, or mitochondrial targeting sequences to isolated mitochondria [108–110]. Possibly, one of the reasons for the ongoing unresolved identity of PTP might be the existence of multiple forms of pores [13–15, 27, 35, 107–111]. As mentioned above, MMP flickering is not inhibited by CsA [84, 86, 87, 90], indicating that it is not mediated by conventional tPTP opening. Then, the question is whether MMP flickering represents a novel form of tPTP.

The cobalt-quenching calcein assay has been used for testing MPT [67, 68]. This assay uses the property of cobalt ion quenching calcein fluorescence. IMM is normally impermeable to cobalt ions, but once PTP opens, cobalt ions as well as calcein molecules (MW 622 Da) freely go in and out of mitochondria and thus calcein fluorescence is quenched. Our experiments with adult cardiomyocytes indicate that depolarizing mitochondria concomitantly lose calcein fluorescence (Fig. 3). These observations support the idea that transient depolarization occurs through a non-specific pore, which we like to refer to as non-conventional tPTP (nc-tPTP). Isolated mitochondria were also used to test the presence of nc-tPTP. We loaded calcein to isolated heart mitochondria by using the membrane-permeable calcein-AM. Mitochondrial calcein fluorescence was stable without cobalt. The addition of cobalt caused a minimal loss of calcein fluorescence. However, upon addition of the complex I-linked substrates glutamate/malate to establish the electrochemical gradient along with cobalt in calcium-free buffer, we found that a significant number of mitochondria lost or decreased calcein fluorescence, indicating that those mitochondria underwent MPT (Fig. 4A). This loss of calcein fluorescence in energized mitochondria was mostly insensitive to CsA, indicating that the observed MPT was likely nc-tPTP (Fig. 4A). Fluorometric analyses also detected Ca²⁺- and CsA-independent nc-tPTP. Whereas Ca²⁺ caused a rapid loss of calcein fluorescence by CsA sensitive PTP opening, simply energizing mitochondria with glutamate/malate also induced a significant decrease of calcein fluorescence in the absence of Ca²⁺ (Fig. 4B), indicating spontaneous pore opening independent of Ca²⁺. CsA has no effect on this energy-dependent spontaneous pore opening. Of note, cobalt alone in non-energized mitochondria causes a small steady decrease of calcein fluorescence, likely reflecting background cobalt permeability and/or MPT of deteriorating mitochondria. Additionally, we observed that oligomycin does not inhibit nc-tPTP, but FCCP completely blocks it (Fig. 4C). The requirement of ETC substrates and MMP for nc-tPTP opening, and its insensitivity to CsA and independence of Ca²⁺ are the properties observed for MMP flickering. It can be speculated that MMP flickering occurs by nc-tPTP.

Figure 3. Mitochondrial depolarization during MMP flickering occurs by MPT.

Mitochondria of adult cardiomyocytes were co-labeled with TMRE and calcein. Time sequence imaging shows a loss of mitochondrial calcein fluorescence coinciding with MMP loss, demonstrating mitochondrial depolarization by PTP opening.

Figure 4. The nc-tPTP in isolated mitochondria.

(A) Isolated mouse heart mitochondria loaded with calcein (5μM calcein-AM) were imaged in the presence of 5mM glutamate/2.5mM malate plus CoCl₂ (1mM) with and without CsA (1μM). A significant number of mitochondria show decreases or losses of calcein fluorescence regardless of CsA addition. Scale bar: 5μm. (B) Calcein fluorometry of isolated mouse heart mitochondria. The addition of CoCl₂ along with glutamate/malate (arrow) in Ca²⁺-free buffer significantly decreases calcein fluorescence by nc-tPTP. Including CsA does not affect the nc-tPTP. Arrowheads indicate the addition of Triton X-100 to solubilize mitochondria. (C) FCCP addition completely inhibits nc-tPTP opening whereas oligomycin has little impact on it. Co: cobalt, GM: glutamate/malate, Ca: calcium, Olm: oligomycin

The identity and mechanism of nc-tPTP are currently unknown. It was found that MMP flickering requires the mitochondrial dynamics protein OPA1 [87, 90, 104]. OPA1 is a membrane remodeling dynamin-related protein that mediates IMM fusion and cristae maintenance [112–114]. Initially, the MMP flickering was proposed to be the process of MMP equilibration between fusing mitochondria through the fusion pore formed by OPA1 [90]. However, MMP flickering occurs in already fused interconnected mitochondrial networks of fission-deficient cells [87, 104]. Furthermore, fusion between IMMs would not cause MPT, and thus, it is unlikely that OPA1-mediated IMM fusion forms nc-tPTP. As mentioned earlier, mitochondrial fission and fusion play a critical role for the extent of MMP flickering by determining mitochondrial connectivity. While how OPA1 participates in MMP flickering and nc-tPTP is undefined, OPA1 may play an additional role possibly in membrane leak for nc-tPTP through its novel IMM remodeling activity [87]. Indeed, it has been shown that OPA1 along with components of the mitochondrial contact site and cristae organizing system (MICOS complex) can establish tight cristae junctions that cause heterogeneity in localized MMP and pH [115], which may enhance the susceptibility for nc-tPTP opening.

Therapeutic Perspectives of nc-tPTP

Sustained opening of the PTP is inherently pathologic as it causes mitochondrial dysfunction. I-R injury is caused by PTP opening, and therefore, it is predicted that targeting the PTP is beneficial in these pathological conditions by decreasing mitochondrial dysfunction and cell death [116–119]. However, except for CypD, the molecular identity of the PTP is still unresolved, which makes molecular targeting difficult. Surprisingly, in a clinical trial involving about 1,000 patients with acute myocardial infarction, cyclosporine A did not result in better clinical outcomes than those with placebo [120]. Although the molecular identity nc-tPTP is also unknown, it can be controlled by manipulating known mitochondrial dynamics proteins. In early reperfusion during I-R injury, increased metabolic influx along with limited ATP synthase activity causes mitochondrial hyperpolarization, Ca²⁺ overload, and ROS increase for PTP opening [121–123]. Under this condition, the nc-tPTP can serve as a relief valve for excess proton gradient and matrix Ca²⁺ during reperfusion, which would prevent PTP opening. Indeed, inhibiting mitochondrial fission, thereby increasing nc-tPTP, decreases PTP opening, ameliorating heart I-R injury in both cell and animal models [61, 64, 124]. In this context, the conventional tPTP functioning as Ca²⁺ releasing channel [50, 51, 125] would be a relief valve as well. However, during I-R injury, the opening of PTP is sustained not transient. Finally, PTP and tPTP are molecularly the same channel, which makes separate manipulations difficult.

The functional significance of mitochondrial dynamics has mainly been appreciated as to maintain healthy population of mitochondria, which involves the segregation of dysfunctional mitochondria by fission and mitochondrial content mixing by fusion [126–128]. However, as discussed in this review, mitochondrial dynamics can play a more direct and fundamental role in regulating mitochondrial energetics by modulating mitochondrial connectivity and mediating nc-tPTP. Unlike the conventional PTP, nc-tPTP is likely a cellular protective mechanism. As nc-tPTP affects mitochondrial proton leak, pathological conditions involving mitochondrial ROS overproduction benefit from manipulating nc-tPTP, including I-R and excitotoxic injuries and diabetes. In addition, many other pathological conditions are also accompanied by PTP opening and mitochondrial dysfunction at some stage of disease progression. This pathologic PTP opening causes mitochondrial dysfunction-induced mitochondrial fragmentation, decreasing the nc-tPTP effect, which predicts the exacerbation of injury by forming a vicious cycle. Therefore, we can postulate that shifts in balance between the extents of PTP (“push” to injury) and nc-tPTP (“pull” from injury) determine pathological outcomes (Fig. 5). An increase of nc-tPTP would put a brake on this vicious cycle and slow pathological progression. Manipulation of mitochondrial dynamics has been shown to benefit many pathological models. Possibly, those are the result of unknowingly targeting nc-tPTP. Identifying the true nature of nc-tPTP will further increase therapeutic opportunities for many diseases.

Figure 5. Opposite roles of PTP and nc-tPTP in pathology.

Sustained PTP opening results in mitochondrial dysfunction and tissue injury whereas nc-tPTP and tPTP may function as a relief valve for mitochondria. In PTP-prone conditions, shifting the balance to nc-tPTP may improve pathological outcomes.

Highlights.

• This review summarizes the studies on a form of mitochondrial permeability transition (MPT) that occurs stochastically and transiently and is insensitive to cyclosporine A and Ca²⁺

• We discuss the regulation of this non-conventional transient opening of permeability transition pore (nc-tPTP) by mitochondrial dynamics

• We propose the “push and pull” concept of PTP (“push” to injury) and nc-tPTP (“pull” from injury) in controlling life and death of cells and therapeutic perspectives of nc-tPTP