Mitochondrial volume homeostasis is a housekeeping cellular function essential for maintaining the structural integrity of the organelle. Changes in mitochondrial volume have been associated with a wide range of important biological functions and pathologies. Mitochondrial matrix volume is controlled by osmotic balance between cytosol and mitochondria. Any dysbalance in the fluxes of the main intracellular ion, potassium, will thus affect the osmotic balance between cytosol and the matrix and promote the water movement between these two compartments. It has been hypothesized that activity of potassium efflux pathways exceeds the potassium influx in functioning mitochondria and that potassium concentration in matrix could be actually lower than in cytoplasm. This hypothesis provides a clear-cut explanation for the mitochondrial swelling observed after mitochondrial depolarization, mitochondrial calcium overload, or opening of permeability transition pore. It should also be noted that the rate of water flux into or out of the mitochondrion is determined not only by the osmotic gradient that acts as the driving force for water transport but also by the water permeability of the inner membrane. Recent data suggest that the mitochondrial inner membrane has also specific water channels, aquaporins, which facilitate water movement between cytoplasm and matrix. This review discusses different phases of mitochondrial swelling and summarizes the potential effects of mitochondrial swelling on cell function.
- potassium homeostasis
- mitochondrial swelling
for decades, changes in mitochondrial volume have been associated with a wide range of important biological functions and pathologies. Loss of mitochondrial volume homeostasis and accompanying mitochondrial swelling are also the earliest and most striking signs of (ischemic) cell injury. Mitochondrial swelling could be result of variety of reasons, from modulations in ion channel and exchanger functions to uncontrolled flux of solutes when the mitochondrial permeability transition pore (PTP) opens. The relationship between the mitochondrial membrane potential and swelling has remained, however, controversial: mitochondrial swelling is often associated with the loss of mitochondrial membrane potential in intact cells but not in isolated mitochondria. Nevertheless, growing body of evidence suggests that mitochondrial swelling is not simply a manifestation of cell injury representing the final stage of mitochondrial dysfunction, but play a crucial role in cell injury. As an example, mitochondrial swelling is one likely mechanism by which cytochrome c and apoptosis inducing factor (AIF) are released (49). Despite evidence suggesting that mitochondrial volume affects several processes critical to the cell, the mechanisms of the mitochondrial volume regulation are not well understood. The present review is thus aimed to discuss the mechanism and relevance of the mitochondrial matrix volume regulation. In this study, we will not analyze the role of fission and fusion, which are largely discussed elsewhere (5, 8, 48).
MEASUREMENT OF MITOCHONDRIAL VOLUME
Since early studies, linking light transmission with mitochondrial swelling (57), the light scattering of the mitochondrial suspension has become the technique of choice to detect mitochondrial size change. This technique, however, has certain limitations. First, this technique cannot provide information on the situation in situ, in cytoplasmic milieu. Second, robust isolation of mitochondria in potassium-free sucrose medium damages probably the mitochondrial membranes, their contact sites and leads to dramatic matrix contraction (22). Third, the changes in mitochondrial light scattering may not always reflect changes in matrix volume. A recent study demonstrated that light scattering can give positive results when no change is seen by the isotope technique, and suggested that light scattering might be sensitive to conformational changes of the mitochondrial adenine nucleotide carrier (11). Moreover, the results of several fluorescent microscopy studies in situ do not support fully those obtained earlier by light scattering on mitochondrial suspension (21, 36, 42, 48, 50, 51).
In turn, the microscopy techniques have their disadvantages. Limited spatial resolution of fluorescence microscopy, and also confocal microscopy, precludes 3D imaging of submicron scale cellular compartments. Diffraction blur artificially enlarges objects with sizes at or below the diffraction limit (∼250 nm for green light), so that rod-shaped mitochondria of diameter ∼200 nm appear considerably thicker (400–500 μm). This may be the main reason why studies using confocal microscopy frequently overestimate mitochondrial dimensions. By contrast, swollen mitochondria are spherical and have a diameter above the diffraction limit (300–500 nm), so that diffraction blur affects their images to a relatively lesser extent. It is therefore difficult to determine the extent and even the direction of volume changes by conventional microscopy. These limitations may explain the inconsistency of previous studies, in which mitochondrial shortening has been associated with both, decreased (51) and increased mitochondrial volume (36).
Electron microscopy provides the necessary resolution, but does not enable the study of functioning mitochondria; besides, the fixation procedure itself affects mitochondrial volume. The latter could be the reason why neuronal mitochondria appear to be spherical in electron micrographs and not thread-like, as in confocal images.
Recently, modifications of traditional confocal microscopy have been made it possible to improve the resolution below the diffraction limit. As an example, beam-scanning multifocal multiphoton 4Pi confocal microscopy enables 3D visualization of yeast mitochondria at an equilateral resolution of 100 nm (17). In several studies, traditional confocal microscopy has been combined with deconvolution microscopy improving resolution and image quality beyond what is generally attainable with either technique alone (9). By combining confocal microscope with deconvolution, we succeeded in measuring the volume of single functioning neuronal mitochondrion in situ (52). Properly calibrated 3D deconvolution analysis permitted the removal of diffraction-induced blur, so that the mitochondrion can be seen at its real size (Fig. 1). Absolute values of mitochondrial volume measured by deconvolution confocal microscopy, 0.04–0.08 μm3, correspond well to those estimated by electron microscopy (0.01–0.1 μm3; Refs. 4 and 33). Moreover, we were able to quantify the volume and other morphological parameters of the same mitochondrion before and after different treatments. Figure 2 shows an example of the mitochondrial remodeling in response to valinomycin. Thus these new approaches make possible the accurate estimation of the mitochondrial volume in the cellular environment.
MITOCHONDRIAL MATRIX VOLUME IS CONTROLLED BY K+ FLUXES
Mitochondria function in a cytosolic milieu containing Na+, K+, Ca2+ as well as other cations and anions. The inner mitochondrial membrane, however, is impermeable to these ions and their flux and concentrations in the mitochondrial matrix are controlled by specific channels and exchangers. The mitochondrial potassium balance is controlled by ATP-dependent (KATP) and Ca2+-dependent (KCa) K+ channels responsible for influx and by K+/H+ exchanger responsible for removal of excess of matrix potassium. Sodium balance is governed by Na+/Ca2+ (influx) and Na+/H+ (efflux) exchangers and calcium balance by Ca2+ channel (influx) and Na+/Ca2+ exchanger (efflux) (summarized in Fig. 3; for reviews, see Refs. 3 and 47). The balance of the main cytoplasmic anions, phosphate and chloride is regulated by numerous carriers and channels.
Any dysbalance in the flux of these ions could affect the osmotic balance between cytosol and the matrix and promote the water movement between these two compartments. However, because intracellular potassium concentration is considerably higher (∼150 mM) than for other ions one may suggest that osmotic balance between the cytoplasm and mitochondrial matrix is controlled mainly by the potassium fluxes. Altered potassium fluxes could mediate also the mitochondrial matrix volume changes observed with depolarization, calcium overload and opening of PTP. Potassium homeostasis could be considered therefore as the main regulator of the mitochondrial matrix volume.
ROLE OF MITOCHONDRIAL MEMBRANE POTENTIAL
Disruption of the electrochemical potential seems to affect both influx and efflux of potassium. The negative internal mitochondrial membrane potential drives K+ as well as other cationic molecules electrophoretically into the mitochondrial matrix. Potassium extrusion via K+/H+ antiporter, although electroneutral, requires the proton gradient (for review, see Ref. 3). A priori, it is not therefore possible to predict the net K+ displacement and concomitant volume change when the potential drops. Hence, there are currently two opposing views how this will happen if the membrane potential drops. Garlid and Paucek (22) propose that potassium concentration in the matrix exceeds the potassium concentration in cytoplasm and that K+/H+ antiporter removes only the excess of matrix potassium. On the other hand, in a recent paper, Safiulina et al. (52) hypothesized that activity of K+/H+ antiporter could exceed the potassium influx and that potassium concentration in matrix could be actually lower than in cytoplasm. The latter hypothesis is supported by earlier findings: Wainio et al. (58) showed a 3.4-fold lower amount of K+ in the mitochondrial matrix compared with the cytosol (25–30 mM) and Zoeteweij et al. (56) showed that resting potassium concentration in matrix is only ∼15 mM [measured by K+-sensitive fluorescence indicator benzoturan isophthalate (PBFI)].
The experimental data on mitochondrial matrix volume changes, which follow modulations of the mitochondrial potential, are inconsistent. It has been suggested that a high membrane potential corresponds to the swollen state and that when the potential collapses the mitochondria contract (see review by Garlid and Paucek in Ref. 22). In contrast, several works based on fluorescence microscopy studies suggest that mitochondria swell when they lose their membrane potential (21, 42, 45), although others reported no volume change or small change (36, 51). Our recent findings demonstrate that loss of mitochondrial membrane potential leads to mitochondrial swelling (52). These results suggest that in situ and in vivo, the balance of potassium movement is shifted toward influx rather than to efflux, as suggested by Garlid and Paucek (22). Since mitochondrial depolarization could not increase the potassium flux to matrix via potassium channels this could be explained only by an inhibition of potassium extrusion from the mitochondrial matrix and/or by induction of permeability transition. This means that at least in living cells mitochondrial depolarization leads to mitochondrial swelling rather than to the contraction.
ROLE OF MITOCHONDRIAL CALCIUM
The role of calcium in mitochondrial matrix volume regulation is a topic of particular interest during last decades. Halestrap et al. (28) demonstrated that an increase in intramitochondrial [Ca2+] can increase the electrogenic flux of K+ into mitochondria by an unknown mechanism and thereby cause swelling. Since then, several mechanisms have been proposed to explain these findings. First, elevated matrix Ca2+ could open recently discovered mitochondrial large conductance KCa channels (14, 53, 60). In their elegant paper, Xu et al. (60) demonstrated the presence of mitoKCa-mediated current at the resting cytosolic Ca2+ concentrations in myocytes and its enhancement at high Ca2+ concentrations. They suggested that the regulatory site for Ca2+ on the channel is likely to face the mitochondrial matrix and proposed that the channel is therefore regulated by intramitochondrial Ca2+. Second, some evidences indicate that also the activity of potassium efflux pathway may be regulated by bivalent cations as Ca2+. It has been shown that depletion of endogenous bivalent cations induces electroneutral K+/H+ exchange activity (15, 34). Gogvadze et al. (23) proposed that high mitochondrial Ca2+ might suppress the activity of K+/H+ exchange and uncompensated K+ entry could thus cause the net movement of water and stimulate mitochondrial swelling. Third, activation of calcium efflux pathways (Na+/Ca+ and Na+/H+ exchangers) during intramitochondrial Ca2+ overload could dissipate the mitochondrial membrane potential/proton gradient. Loss of proton gradient could suppress the activity of K+/H+ exchange and lead finally the mitochondrial swelling. Finally, if the Na+/Ca2+ exchanger becomes saturated, high intramitochondrial Ca2+ could rise to the level sufficient to facilitate or open directly the mitochondrial PTP (2, 13, 30), which leads to entry of potassium as well as other ions and solutes and inducing massive swelling of the matrix. Increase in matrix [Ca2+] up to micromolar level inhibits pyrophosphatase and leads to elevation in matrix pyrophosphate. Elevated matrix pyrophosphate, in turn, may transiently displace adenine nucleotides from adenine nucleotide translocase and convert the latter into a potassium channel (12, 26). Thus, whatever is the mechanism of Ca2+-induced mitochondrial swelling, it is likely to be caused by potassium influx.
ROLE OF PTP
Permeability transition is described as an abrupt increase of inner membrane permeability to solutes with molecular masses of <1.5 kDa. These events are caused by the opening of the PTP. The pore consists of the several inner and outer membrane proteins (voltage-dependent anion channel, adenine nucleotide translocase, cyclophilin D, and others) and its opening is triggered by multiple effectors that include elevated mitochondrial calcium and mitochondrial depolarization (see Refs. 24 and 27 for reviews). The only primary consequence of PTP opening is mitochondrial depolarization due to the equilibration of matrix and cytosolic ion gradients generated by the electron transport chain. Most likely the total net flux of ions is directed toward the mitochondrial matrix and this ion movement, accompanied by osmotically obliged water, leads finally to mitochondrial swelling. It has been also suggested that high concentrations of proteins in matrix exert an oncotic pressure, driving water into the mitochondria after opening of PTP (for review, see Ref. 29). However, the opening of PTP could not increase the oncotic pressure, as the mitochondrial inner membrane remains impermeable to proteins and relatively constant oncotic pressure should be already well balanced with water since the inner membrane is permeable to water. As a matter of fact, it has been not yet specified which cytosolic ions and metabolites will enter via PTP to matrix and could be responsible for this osmotic imbalance. Keeping in mind that the mitochondrial membrane potential (ΔΨ) in depolarized mitochondria cannot be a driving force for potassium influx, it is tempting to suggest that swelling associated with opening of PTP is induced by potassium entry down to the concentration gradient, which exists in energized mitochondria.
POSSIBLE ROLE OF AQUAPORINS
The rate of water flux into or out of the mitochondrion is determined not only by the osmotic gradient that acts as the driving force for water transport but also by the water permeability of the inner membrane. It has been agreed by most of the scientists that water could pass through the inner membrane by simple diffusion through the lipid bilayer or accompany the ions through the channels. However, recent data suggest that water flux across the inner membranes occur not only through the lipid bilayer or ion channels, but also through specific water channels, aquaporins. Recent reports demonstrate the presence of a member of the aquaporin family, AQP8 in the liver (6, 19), kidney (39), and central nervous system stem cells (38). A shorter isoform, AQP9, generated by alternative splicing has been detected in brain mitochondria (in astrocytes and subset of neurons; see Ref. 1). These water pores are completely impermeable to charged species, such as protons that is critical for the conservation of mitochondrial membrane potential. Water permeation through aquaporins is a passive process that follows the direction of osmotic pressure across the membrane (37). In their recent review, Lee and Thevenod (40) suggest that mitochondrial aquaporins may mediate water transport associated with physiological volume changes and participate in osmotic swelling induced by apoptotic stimuli. This hypothesis is supported by their earlier finding that AgNO3, a potent inhibitor of aquaporins, inhibits Cd2+-induced mitochondrial swelling (39). However, recent work by Yang et al. (61), which provides evidence against functionally significant expression of aquaporins in mitochondria, questions this hypothesis.
PHASES OF MITOCHONDRIAL SWELLING
Mitochondrial swelling is defined as an “increase in the volume of mitochondria due to an influx of fluid”.
Early phase of mitochondrial swelling involves movement of water from the intercristal spaces into the matrix. Matrix swelling thus results in collapse of intermembrane space in cristae without increase in surface area of inner boundary membrane and change in external shape of mitochondria. It is proposed that this phase involves fission and tubularization of cristae to allow additional matrix expansion (43).
The tubular connections of cristae to the inner boundary membrane are not rigid structures and probably do not prevent the recruitment of crista membrane regions to the inner membrane (44). As an example, tBid, well-known inducer of apoptosis, causes opening of junctions between the cristae and the intermembrane space (54). Further matrix expansion could thus increase the surface area of inner boundary membrane, which would exceed that of outer membrane and the later should change its shape to accommodate the extra volume (21). In several cell types the mitochondria are already spherical or oval in steady-state conditions and have achieved already the maximal volume-to-surface area ratio. As an example the relatively spherical cardiomyocyte mitochondria could increase their matrix volume only 20–30% without loosing their outer membrane integrity (11). However, in cell types where mitochondria are elongated, the volume increase could be more substantial. Elongated mitochondrion from cerebellar granule neuron depicted in Fig. 2 could increase its volume up to twofold without loosing the outer membrane integrity (52).
However, when water influx and increased mitochondrial matrix volume will not balance the osmotic pressure in the matrix and in the cytoplasm, and water influx continues, then the pressure applied to the outer membrane from the matrix side increases. This could lead to the opening of PTP balancing osmotic pressures or rupture of the mitochondrial outer membrane allowing further expansion of matrix. It could be also hypothesized that this promotes fusion of adjacent mitochondria and leads to formation of megamitochondria (59).
Thus mitochondrial swelling is a multiphasic process involving several phases that are summarized in Fig. 4.
CHANGES IN MITOCHONDRIAL MORPHOLOGY AFFECT CELL FUNCTION
Mitochondrial volume homeostasis is a housekeeping function that is essential for maintaining the structural integrity of the organelle. In addition to this function, mitochondrial volume seems to affect several other processes in the cell.
Increase in matrix volume could activate respiratory chain, thereby increasing ATP production. Halestrap with colleagues (25, 26, 41) demonstrated that increase in matrix volume correlates well with increase in respiration rates. Thus, it could be hypothesized that during energetic stress, when the need for ATP is high and ΔΨ decreases, an increase in matrix volume could further activate the respiratory chain. It is interesting to note that matrix volume is increased in ischemic mitochondria as well as following reperfusion (41).
Another hypothetical role of the mitochondrial matrix volume homeostasis is related to the reactive oxygen species (ROS) formation. Indeed, drugs are thought to induce mitochondrial swelling like mitoKATP channel can interfere with ROS production although data describing these effects are rather disparate. Ferranti et al. (18) have shown that in isolated heart, liver, and brain mitochondria, diazoxide reduces ROS generation. In contrast, cells treated with diazoxide showed an increase in ROS production (7, 20); however, Dzeja et al. (16) have found that in perfused rat hearts, diazoxide (mitoKATP opener) diminishes the generation of malondialdehyde, a marker of oxidative stress. These discrepancies probably reflect the fact that increased K+ influx and/or matrix swelling are associated with other cellular processes, which have an impact on the ROS generation, like matrix alkalinisation (22), mild uncoupling (32), or enhancement of fatty acid oxidation (25).
Mitochondrial swelling is also one of the key players in cytochrome c release associated with apoptotic cell death. Cytochrome c can be released from the mitochondria via both PTP-dependent and PTP-independent mechanisms (46). Recently, Gogvadze et al. (23) suggested that release of cytochrome c might occur via moderate modulation of mitochondrial volume, irrespective of the mechanism leading to the mitochondrial swelling. For example, stimulation of K+ influx by low (nanomolar) concentrations of the K+-selective ionophore valinomycin induced swelling, which was accompanied by cyclosporine A independent cytochrome c release although the amount of cytochrome c released from mitochondria was markedly lower than that during permeability transition (10, 23). Holmuhamedov et al. (31) demonstrated that potassium channel openers increase matrix volume and release cytochrome c and adenylate kinase 2. Furthermore, Garlid and co-workers (10) clearly demonstrated that increasing K+ influx and stimulation of K+ cycling makes the outer mitochondrial membrane permeable to cytochrome c. Our observation that depolarization leads to mitochondrial swelling suggests that also depolarization could be one of the factors facilitating cytochrome c release.
Recent evidence suggests that mitochondrial swelling could have also other, rather unexpected effects on cell function. Our previous study suggests that in cardiomyocytes increases in mitochondrial volume can impose mechanical constraints inside the cell, leading to an increase in force developed by myofibrils (35). Such a mechanical signaling could exist also in neurons. Mitochondrial traffic in axonal shafts could be very sensitive to mitochondrial volume modulations because the diameter of these shafts precisely matches mitochondrial diameter under normal conditions (55). We hypothesized that mitochondrial swelling (induced either by cation influx or disruption of mitochondrial potential) could be one of the primary factors regulating mitochondrial traffic in neurites: increase in mitochondrial diameter could mechanically hamper the passage of mitochondria in narrow neurites (52).
Altogether, these data show that that mitochondrial volume seems to affect mitochondrial electron transport, ROS production, cytochrome c release in the process of apoptosis and as well participate in mechanical signaling pathways.
This research was supported Estonian Science Foundation Grant 6227, European Union FP6 contract MTKD-CT-2004–517176, Estonian and French governments short-term travel program Parrot, and by Institut National de la Santé et de la Recherche Médicale.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society