The Reduction in the Mitochondrial Membrane Potential in Aging: The Role of the Mitochondrial Permeability Transition Pore

It is, therefore, quite likely that enhancement of mPTP by the excess mROS generated by the gas-1 mutant increased the fraction of depolarized mitochondria. This may also explain the reduction in the mitochondrial density since depolarized mitochondria are expected to be cleared by mitophagy [16], and cells with a high number of depolarized mitochondria may proceed to apoptosis [32]. More recently Kwon et al. [68] found that the reduction in TMRE fluorescence in the gas-1 mutant could be accounted for entirely by the reduction in mitochondrial density.

Lemire et al. [73] also studied ∆Ψm in C. elegans mutants that affected the C. elegans lifespan. They used the lipophilic cation diSC3(3) at high concentration (4 μM) to measure ∆Ψm in several strains that had extended C. elegans lifespan (i.e., daf-2, age-1, clk-1, isp-1 and eat-2). At high concentration diSC3(3), aggregates in the mitochondria and the fluorescence maximum shifted to the red. Lemire et al. determined the fluorescence maximum of diS-C3(3) in whole worms at the larval stage (L4) of the various life-extending strains and found that at the larval stage, ∆Ψm, as estimated from the fluorescence maximum shift, was lower in all the life-extending strains compared to that of the wild type (N2). These results appear to contradict the results of Brys et al. [66], but it should be emphasized that these measurements were undertaken with the whole larva and not with mitochondria isolated from the adult worms; there are significant differences between the properties of adult and larval mitochondria [43]. Moreover, disruption of mitochondrial function during development was shown to increase lifespan [74].

Cho et al. [75] studied the effect of flavonoids on neurodegeneration in aging C. elegans. They found that flavonoids mitigate neurodegeneration in aging C. elegans and that this effect was associated with an early reduction in ∆Ψm. They measured ∆Ψm with TMRE at 100 nM in whole worms at the larval stage (L4). The effect of flavonoids on ∆Ψm was comparable to the weak uncoupler DNP, suggesting that this uncoupling effect enhanced mitophagy, which protected the worms against neurodegeneration. Uncouplers are known activators of the mPTP [76], and mPTP activation enhances mitophagy [16]. Thus, activating mPTP at an early age could lead to activation of protection mechanisms, such as mitophagy or the mitochondria unfolded protein response, UPRmt (see below). Cho et al. [75] also reported that ∆Ψm was much lower in the adult worms than in the larval state regardless of the flavonoids but did not present the evidence.

Dilberger et al. [69] studied the effects of paraquat on C. elegans longevity, mROS production and ∆Ψm. Paraquat is known to induce excess mROS production, as in aging, and can, therefore, be considered a pharmacological model of aging in C. elegans. Indeed, Dilberger et al. [69] showed that paraquat, depending on the concentration, greatly reduced C. elegans lifespan. This was associated with greatly increased mROS production, mitochondrial fragmentation and greatly reduced ∆Ψm. Dilberger et al. [69] measured ∆Ψm in isolated mitochondria from the fluorescence quenching of R123; paraquat (5 mM) reduced ∆Ψm by ~50%. In addition, they showed that paraquat increased mitochondrial fragmentation. Paraquat is known to induce the activation of mPTP [77], which is known to increase mitochondrial fragmentation in C. elegans [70]; therefore, the reduction in ∆Ψm by paraquat and the resulting reduction in lifespan can be attributed to the activation of mPTP.

Since paraquat induces the mitochondrial unfolded protein response, UPRmt, in C. elegans [24,78], the results of Dilberger et al. [69] suggest that the activation of mPTP is sufficient to activate UPRmt. Berry et al. [24] also showed that reduction in ∆Ψm by the uncoupler FCCP was sufficient to activate UPRmt in C. elegans. However, FCCP is known to activate mPTP [75], and Angeli et al. [79] showed that genetic or pharmacological inhibition of mPTP inhibited UPRmt. Moreover, it was recently shown that two signals from the mitochondria to the cytosol induced UPRmt, excess mROS and the accumulation of mitochondrial protein precursors [80]. Excess mROS was released by the activation of mPTP, and the depolarization of mitochondria resulting from the activation of mPTP inhibited the uptake of mitochondrial protein precursors, providing further support for the conclusion that activation of UPRmt depends on the activation of mPTP.

Interestingly, in their study, Berry et al. [24] measured ∆Ψm from TMRE fluorescence from the whole body of the worms after incubating the worms for 24 h with 100 nM TMRE. It appears, however, that at this high TMRE concentration, the dye in the mitochondrial matrix was largely aggregated and largely quenched. This can be deduced from the parallel titrations of the effect of FCCP on TMRE fluorescence and the activation of UPRmt. In the very wide range of FCCP concentrations from 10 pM to 10 μM, UPRmt was activated only at a narrow range around 10 nM, presumably resulting in a small reduction in ∆Ψm. However, the titration of TMRE fluorescence over the entire range of FCCP concentration did not show the expected linear titration curve. Already, at the very low concentration of 10 pM, there was a ~30% reduction in TMRE fluorescence but increasing concentrations even up to 3 uM did not reduce the fluorescence further. Only at 25 μM, presumably when the mitochondria became completely depolarized, was the fluorescence significantly reduced (~80%). Thus, the TMRE fluorescence values at 10 pM FCCP and 1 nM FCCP, which were insufficient to activate UPRmt, and the TMRE fluorescence value at 10 nM FCCP, that was sufficient to activate UPRmt, were all the same, demonstrating that, at this concentration of TMRE, the fluorescence magnitude over nearly the entire range of ∆Ψm values was not sensitive to the magnitude of ∆Ψm, and only when the mitochondria were depolarized was the TMRE fluorescence reduced as expected.

Berry et al. [71] targeted a fungal light-driven proton pump to the mitochondria of C. elegans to selectively increase the mitochondrial protonmotive force. They called this optogenetic tool mtON. They used TMRE at a high quenching concentration (300 nM) to measure ∆Ψm in isolated mitochondria from the fluorescence quenching of the suspending medium. They showed that in isolated non-respiring mitochondria, applying light and all trans retinal (ATR), the catalytic center of the light-driven proton pump, resulted in the generation of ∆Ψm that was collapsed by the uncoupler CCCP. Using the pH indicator BCECF, they also showed that mtON generated a pH gradient as expected. The ∆Ψm that was generated by the substrate succinate (in the presence of rotenone, “state 2”) was not significantly different from that generated by mtON at maximal illumination. However, applying both succinate and mtON did not increase ∆Ψm above the value that was generated by either succinate or mtON alone, suggesting that the magnitude of ∆Ψm in state 2 was limited by the membrane proton permeability. Berry et al. [71] also showed that mtON alone (without succinate) was able to generate ATP, resembling ATP generated by succinate oxidation. In “state 3” (i.e., succinate + ADP), mtON inhibited respiration by 50%, presumably because mtON was able to increase ∆Ψm in state 3 (however, Berry et al. [71] did not measure ∆Ψm in “state 3”). Electron transport inhibitors (i.e., rotenone, antimycin A and azide) killed most C. elegans worms within hours, presumably because they collapse ∆Ψm, and mtON was able to significantly increase worm survival in the presence of electron transport inhibitors. Finally, Berry et al. [71] showed that turning on mtON in the gas-1 mutant significantly increased motility that was otherwise greatly reduced in the mutant compared to wt.

Since the reported reduction in ∆Ψm in the gas-1 mutant [68] undoubtedly resulted from the enhancement of the activation of mPTP by the excess mROS produced by the gas-1 mutation (see above), the restoration of motility to the gas-1 mutant by mtON could be attributed to inhibition of the voltage-gated mPTP by the ∆Ψm that was generated by mtON.

mtON should be an effective inhibitor of mPTP not only because high ∆Ψm decreases the probability of opening of the voltage-gated mPTP, but also, and perhaps more importantly, because it enables the channel to close quickly after releasing the excess calcium and mROS that activate the channel opening. Normally, electron transport proton pumps cannot easily restore ∆Ψm to close the channel because respiratory substrates are also released by the channel and electron transport is inhibited. However, the mtON-generated ∆Ψm is not dependent on electron transport and can, therefore, close the channel quickly after a relatively short opening.

Very recently, Berry et al. [25] showed that the optogenetic generation of ∆Ψm, (mtON) in C. elegans during the entire adult lifespan increased C. elegans lifespan. They also showed that both 4-day-old and 10-day-old worms increased thrashing in liquid (but not in solids) when mtON was activated, demonstrating that muscle activity, that normally deteriorates in old worms, can be partially rescued by ∆Ψm generated by mtON. In this study, they estimated ∆Ψm not in isolated mitochondria, but from the whole worm incubated with TMRE (100 nM). The TMRE fluorescence was collected from the head or the pharyngal bulbs that have a high concentration of mitochondria. They showed that, compared to TMRE fluorescence from the worms at day 1, TMRE fluorescence was greatly reduced with aging with a ~75% reduction in day 4 and a ~50% reduction in day 10. Berry et al. measured mitochondrial density at day 4 but not at day 1, so the reduction in TMRE fluorescence was most likely due, partially at least, to decreased mitochondrial density and partially due to decreased ∆Ψm (cf. [68]). In a separate experiment, with 4-day-old worms, Berry et al. [25] showed that turning on mtON greatly increased median TMRE fluorescence. In this experiment, they separately measured mitochondrial density with mitotracker and normalized the TMRE fluorescence to the mitochondrial density.

As discussed above, at 100 nM TMRE, the probe was aggregated and quenched in the mitochondrial matrix, so it was not sensitive to small changes in ∆Ψm, and the large reduction in fluorescence in aging worms observed was most likely due to an increased fraction of depolarized mitochondria and/or mitophagy-cleared depolarized mitochondria. If the average ∆Ψm in the mitochondria of aged worms was 50–75% lower than that of young worms, as claimed by Berry et al. [25], it would not be expected that an uncoupler like FCCP would stimulate respiration in old worms since the enhancement of respiration by an uncoupler results from the collapse of the protonmotive force. However, the fact that Berry et al. [25] observed no difference in the effect of FCCP on respiration between young and old worms suggests that there was no difference in the ∆Ψm of respiring mitochondria in old and young cells, and the lower TMRE fluorescence resulted from lower mitochondria density and/or a larger fraction of depolarized mitochondria that did not respire. The explanation that the low ∆Ψm that was observed in old worms resulted from enhanced activation of mPTP, and that mtON increased the lifespan and healthspan by inhibiting mPTP, is compatible with all the observations reported by Berry et al. [25].

In a more recent report, Berry et al. [26] investigated the relationship between dietary restriction (DR) that extended lifespan in C. elegans and ∆Ψm. Here, they showed again that aging from day 1 to day 4 was associated with a ~50% reduction in TMRE fluorescence under normal dietary conditions, while, when on DR, the reduction in TMRE fluorescence was much smaller, ~20%. In this study, they also measured TMRE fluorescence from the pharyngal bulb of whole worms, but in this case, they incubated the worms with 1 μM TMRE compared to 100 nM in their previous study. They did not explain this change of protocol, but it is clear that, under these new conditions, TMRE fluorescence from the mitochondrial matrix is in full quenching mode and the fluorescence cannot be proportional to the magnitude of ∆Ψm (see above). Therefore, in this study as well, the reduction in TMRE fluorescence in aged cells could indicate either a reduction in the density of mitochondria or an increase in the number of depolarized mitochondria, most likely both. Berry et al. [26] also found that several genetic and pharmacological manipulations modulated the effect of DR on longevity and TMRE fluorescence. The uncoupler FCCP inhibited the effect of DR on both longevity and TMRE fluorescence as expected. The mutant eat −2, a genetic model of DR, increased TMRE fluorescence at day 4 and extended lifespan. This effect was also inhibited by FCCP. a deletion of the uncoupling protein, unc-p, a deletion of the adenine nucleotide translocases ant-1,2, and the deletion of the IF1 protein (an ATP synthase inhibitor) mal-2. All raised the TMRE fluorescence at day 1, but only ant-1,2 and mal-2 inhibited the reduction in TMRE fluorescence at day 4 and extended lifespan.

While the authors sought to explain these effects as resulting from modulations of the bioenergetic functions of mitochondria, all these effects can be attributed directly to the activation or inhibition of mPTP. It was shown that life extension by the eat-2 mutant depended on the inhibition of mPTP [67], and that FCCP was an activator of mPTP [76]. ANT 1 and ANT 2 are components of mPTP [18] and their deletion inhibited mPTP [70]. The protein IF1 interacts with ATP synthase, another component of mPTP [18], and was shown to activate mPTP [81]; therefore, the deletion of mal-2 is expected to inhibit mPTP. Thus, all the genetic and pharmacological effects on ∆Ψm that are described by Berry et al. [26] can be attributed to activation or inhibition of mPTP. Finally, cyclosporin A, an inhibitor of mPTP was shown to extend lifespan in C. elegans [82], and it was also shown that mPTP was inhibited in DR in mammals [83,84,85].

Mitochondrial dysfunction is a well-established “hallmark of aging”, which strongly interacts with most other “hallmarks of aging”. However, there are various manifestations of mitochondrial dysfunction in aging, which raises two questions: what are the causes of the mitochondrial dysfunctions that are associated with aging, and what are the consequences for health and longevity for each of the aging-associated mitochondrial dysfunctions? As discussed in detail in this review, one of the most frequently reported mitochondrial dysfunctions in aging is the reduction in ∆Ψm that is observed in many types of aging cells, but the cause of this effect is still not established. It was suggested previously that this may be the result of oxidative damage to mitochondrial phospholipids, such as cardiolipin [48]. Such damage could increase proton permeability and, thus, reduce ∆Ψm [15]. However, there is no evidence that the mitochondria of old cells are uncoupled. Alternatively, it was suggested that the lower ∆Ψm in aging is a result of an aging-specific metabolic state that impairs bioenergetic functions, and that this leads to disease and a shortened lifespan [23,24,25,26,27]. However, it was not specified what exactly this metabolic state is and how aging leads to this state.

Here we offer a new interpretation: the lower ∆Ψm that is observed in aging cells results from the enhanced activation of the mPTP, which increases the fraction of depolarized mitochondria. This interpretation is compatible with all the evidence reviewed here and is strongly supported by many of the reviewed studies. Several studies demonstrated that inhibition of mPTP restored ∆Ψm in aging cells (Table 1). Most of the studies that reported aging-associated reduction in ∆Ψm were conducted with whole cells, tissues, or even intact animals (e.g., worms). However, most of these studies employed high concentrations of fluorescent ∆Ψm indicators in which the fluorescence was in the quenching mode and the fluorescence was not proportional to the magnitude of ∆Ψm. Therefore, the fluorescence was only proportional to the number of mitochondria with significant ∆Ψm. Any reduction in probe fluorescence in aging cells under these assay conditions simply indicates loss of mitochondria with ∆Ψm, either due to depolarization or elimination of depolarized mitochondria by mitophagy, both of which could be the result of mPTP activation. Only a few studies measured mitochondria density with an appropriate probe, and these suggest that aging is associated with both a reduction in mitochondrial density and an increased fraction of depolarized mitochondria. Many of the studies that reported lower ∆Ψm in aged cells also reported enhanced activation of mPTP (Table 1), which supports this interpretation of the results. Of all the studies of the reduction in ∆Ψm in whole aging cells, only the study of Morris et al.

[60] that reported a 20% lower average ∆Ψm in aging intestinal stem cells from Drosophila, used a valid protocol for ∆Ψm measurement in whole cells, with a low non-quenching TMRE concentration (20 nM), and where the fluorescence was collected directly from within the labeled mitochondria. However, Morris et al. showed that the old intestinal stem cells in Drosophila were in a very distinct metabolic state with a largely inhibited mitochondrial calcium transporter, MCU, which could explain the lower ∆Ψm. This state was not reported in any other aging cells study and does not appear to represent a general feature of mitochondrial dysfunction in aging cells.

Several experiments with isolated mitochondria from tissues of aged animals showed a small reduction (~10%) in the average ∆Ψm. While the protocols of these experiments are more reliable, these results are also compatible with the suggestion that in aging cells there is a larger fraction of depolarized mitochondria. Whether this is the case or not can only be decided by measuring ∆Ψm of individual mitochondria (cf. [43]). If there was really a significant reduction in the average ∆Ψm in aging cells, it would be expected that there would be a reduction in mROS production and increased uncoupling, but this is not observed in aging cells—it is widely reported that mROS production is increased in aged cells and that the coupling is not impaired.

The conclusion that the reduction in ∆Ψm in aging cells results from mitochondrial depolarization by the enhanced activation of mPTP in aging cells provides a straightforward explanation of all the reported effects of ∆Ψm on aging. Enhanced activation of mPTP leads to excess release of mROS and destruction of NAD+ and this leads to many of the reported effects of aging (Figure 1), to age-dependent degenerative diseases, and, eventually, to cell death by necrosis or apoptosis. On the other hand, limited activation of mPTP, particularly at an early age, initiates protective mechanisms, such as UPRmt, mitophagy, and mitochondria biogenesis. Inhibition of mPTP activation by dietary restriction, drugs or artificially produced ∆Ψm, retards the process of aging, inhibits degenerative diseases, and increases longevity.