Intracellular energetic units in healthy and diseased hearts

Abstract

BACKGROUND:

The present review examines the role of intra-cellular compartmentation of energy metabolism in vivo.

OBJECTIVE:

To compare the kinetics of the activation of mitochondrial respiration in skinned cardiac fibres by exogenous and endogenous adenine nucleotides in dependence of the modulation of cellular structure and contraction.

METHODS:

Saponin-permeabilized cardiac fibres or cells were analyzed using oxygraphy and confocal microscopy.

RESULTS:

Mitochondria respiration in fibres or cells was upregulated by cumulative additions of ADP to the medium with an apparent K_m of 200 μM to 300 μM. When respiration was stimulated by endogenous ADP produced by intracellular ATPases, a near maximum respiration rate was achieved at an ADP concentration of less than 20 μM in the medium. A powerful ADP-consuming system, consisting of pyruvate kinase and phosphoenolpyruvate, that totally suppressed the activation of respiration by exogenous ADP, failed to abolish the stimulation of respiration by endogenous ADP, but did inhibit respiration after the cells were treated with trypsin. The addition of up to 4 μM of free Ca²⁺ to the actively respiring fibres resulted in reversible hypercontraction associated with a decreased apparent K_m for exogenous ADP. These changes were fully abolished in fibres after the removal of myosin by KCl treatment.

CONCLUSIONS:

Mitochondria and ATPases, together with cytoskeletal proteins that establish the structural links between mitochondria and sarcomeres, form complexes – intracellular energetic units (ICEUs) – in cardiac cells. Within the ICEUs, the mitochondria and ATPases interact via specialized energy transfer systems, such as the creatine kinase- and adenylate kinase-phosphotransfer networks, and direct ATP channelling. Disintegration of the structure and function of ICEUs results in dyscompartmentation of adenine nucleotides and may represent a basis for cardiac diseases.

According to the classic respiratory control theory, increased cytosolic ADP concentrations, due to ATP splitting by ATPases, is an ultimate signal for cellular respiratory stimulation (1,2). The theory is based on the observation that when ADP is added to isolated mitochondria it stimulates respiration in accordance with Michaelis-Menten kinetics and with a high affinity for ADP (K_m of 10 μM to 20 μM) (1). However, when one attempts to apply the respiratory control theory to living muscle in vivo, several complications arise. First, the ADP concentration in muscle cells is enigmatic because it has never been measured directly, but instead calculated from the tissue concentrations of ATP, phosphocreatine (PCr) and creatine determined using ³¹P-nuclear magnetic resonance (NMR), with the assumption that creatine kinase (CK) reactions are in equilibrium in the cells (3–5). The validity of this approach is questionable because recent studies have revealed the nonequilibrium state of the intracellular CK reactions (5–8). Second, calculations of ADP concentrations have yielded different results in different muscle types. Glycolytic muscle exhibits increased ADP concentrations in response to increased workload, which conforms to the respiratory control theory (9). However, in heart muscle, which is characterized by a linear relationship between the respiration rate and work, no significant changes in concentrations of ADP, ATP, PCr and creatine have been detected despite large variations in contractile activity, a condition known as metabolic stability (10–12). How then can the mitochondria be activated in the heart cell?

Currently, there exist two concepts regarding this problem. One group of researchers believe that a transient increase in cytoplasmic Ca²⁺ simultaneously activates the contractile apparatus and mitochondria, thus matching the increased energy demand with enhanced ATP production (13–15). This view is based on the argument that Ca²⁺ activates both the mitochondrial enzymes (dehydrogenases and ATP synthase) and actomyosin complexes (13,16,17). If this ‘parallel activation’ hypothesis is true, the changes in the concentrations of intracellular Ca²⁺ transients should be sufficient to cause the 15- to 20-fold increases observed in the respiration rate registered under conditions of the Frank-Starling law and metabolic stability in vivo. In fact, several experimental facts disagree with this assumption. The monitoring of intracellular Ca²⁺ concentration reveals that a stepwise stretch of the myocardium produces a rapid potentiation of twitch force but not of the Ca²⁺ transient (16,17). When the effects of Ca²⁺ on respiration, F_oF₁-ATPase and membrane potential were studied in isolated heart mitochondria, it was found that changes in mitochondrial Ca²⁺, although being rapid enough to participate in the regulation of respiration, could only maximally increase the respiration rate up to two times, with an increase of the free cytoplasmic Ca²⁺ concentration up to 600 nM (18,19). It is known that the mean free cytoplasmic Ca²⁺ concentration may extend up to 1 μM to 3 μM in maximally activated cardiac muscle (20), which largely exceeds the levels necessary for saturation of the mitochondrial Ca²⁺-sensitive enzymes. These data suggest that under in vivo conditions, mitochondrial respiration always proceeds with a rate close to V_max and, therefore, can not be further accelerated by Ca²⁺. Hence, the theory of ‘parallel activation’ fails to explain the 15- to 20-fold increases in the respiration rate of cardiac cells.

An alternative concept considers the high degree of structural and functional organization in cardiac cells (21). It states that the metabolic channelling by means of intracellular phosphotransfer systems constitutes a major mechanism linking the mitochondria and ATPases within specific structures –intracellular energetic units (ICEUs) (21–23). To date, three different systems (CK-phosphotransfer, adenylate kinase [AK]-phosphotransfer and direct transfer of adenine nucleotides) have been described that operate within the ICEUs. The important feature of these systems is that they ensure effective stimulation of oxidative phosphorylation without significant changes in cytosolic adenine nucleotide and PCr contents, thus conferring a metabolic stability on the cardiac cells in conditions of increasing workloads.

The present review summarizes recent data in support of the existence of ICEUs in heart cells and of a pathogenic role of the altered structural organization of energy metabolism in cardiac diseases.

DIFFERENT KINETICS OF REGULATION OF MITOCHONDRIAL RESPIRATION BY ENDOGENOUS VERSUS EXOGENOUS ADP

The saponin-permeabilized (skinned) muscle fibre technique, in conjunction with oxygraphy, confocal microscopy, enzyme kinetics analysis, proteomics and mathematical modelling, has important advantages compared with isolated mitochondrial fractions in studies of energy metabolism. It allows the avoidance of artifacts due to the isolation of mitochondria, it enables the analysis of mitochondria in small specimens from human muscle biopsies and, given the preservation of interactions of mitochondria with other cellular structures, it provides a reliable means for studying their function in vivo (24,25). Already, initial experiments show that skinned fibres from glycolytic fast-twitch muscles (eg, musculus gastrocnemius) display a high apparent affinity to ADP in the regulation of respiration (K_m of 10 μM to 20 μM), comparable with that in isolated mitochondria; however, oxidative slow-twitch muscles, such as heart and musculus soleus, show a much lower affinity to ADP (K_m of 200 μM to 400 μM) (26–32). This contrast between the two muscle types is not related to diversities in diffusion distances between the medium and core of the fibres, or in the tissue contents of mitochondria and total ATPase activities (25), but instead results from distinct mechanisms of the regulation of O₂ consumption (9,31). Another interesting property of skinned cardiac cells – the dependency of mitochondrial respiration on the source of activator ADP – was also revealed. When ATP was added to the skinned fibres, it activated respiration to the same extent and with similar kinetics as exogenous ADP. However, high respiratory rates were observed with ATP at much lower ADP concentrations in the medium than with exogenous ADP (22). These studies led to the conclusion that the endogenous ADP produced by ATPases does not easily equilibrate with ADP in the bulk-water phase. In permeabilized cells, the bulk phase is a medium, but in cells in vivo, it is probably the cytoplasm. Thus, diffusion of ADP out of the cytoplasm was shown to be restricted.

When added to the medium, creatine and AMP also activate respiration in skinned fibres, provided that small amounts of ATP or ADP are available. Figure 1 shows that creatine activates respiration by increasing the apparent affinity of respiration to exogenous ADP. Compared with creatine, activation of respiration by AMP requires much less ATP. In fact, AMP stimulates respiration even without added ATP (Figure 2); hence, the trace amounts of ATP bound to intracellular structures are sufficient to trigger the respiration-linked AK reaction (33).

Figure 1).

Effect of creatine (Cr) (20 mM) on O₂ consumption (V_O2) versus ADP concentration in human skinned atrial fibres (presentation of data in Lineweaver-Burk plots). Cr increases the apparent affinity (K_m) of mitochondria to ADP. Here and elsewhere, ADP was added with MgCl₂ (0.6 mol/mol ADP). ww Wet weight

Figure 2).

Respirometric investigation of the coupling of mitochondrial adenylate kinase to oxidative phosphorylation in human atrial fibres by applying increasing concentrations of AMP in the absence of added ATP. Atr Atractyloside; V_O2 O₂ consumption; ww Wet weight. Adapted with permission from reference 33

To understand the mechanisms of respiratory stimulation by ATP, creatine and AMP, it is important to define the source and localization of the activator ADP. The activation of kinases is associated with the accumulation of large amounts of ADP (50 μM to 90 μM) in the medium surrounding the skinned fibres (34), which exceed the K_m values for ADP in the regulation of respiration in isolated mitochondria. One may therefore reason that it is the cytoplasmic (exogenous) ADP produced by the ATPases, muscle type (MM)-CK or AK₁ outside of the mitochondria (in myofibrils and the sarcoplasmic reticulum [SR]) that activates respiration after being diffused from the cytoplasm through the outer mitochondrial membrane to the adenine nucleotide translocator (ANT). Alternatively, respiration could be upregulated by endogenous (intramitochondrial) ADP produced near the ANT due to its functional coupling to mitochondrial CK (mi-CK) or AK₂ (22,23,35,36). Analysis of the respiration of the skinned fibres in the presence of pyruvate kinase (PK) and phosphoenolpyruvate (PEP) at concentrations (20 U/mL and 5 mM, respectively) sufficient to eliminate the ADP by its phosphorylation in the surrounding medium (ie, cytoplasm) allowed the discrimination between these two alternatives (22,23,34). Figure 3 shows that 2 mM ATP added to skinned fibres markedly activated respiration via ADP produced in ATPase reactions. The inclusion of PK (20 U/mL) negligibly inhibited respiration in the presence of ATP, and did not abolish the activation of respiration caused either by creatine (Figure 3A) or AMP (Figure 3B). Thus, most of the ADP produced by ATPases or kinases remained inaccessible to exogenous PK in skinned fibres and, therefore, represented the endogenous activator ADP.

Figure 3).

Respirometric investigations of the coupling of kinases to oxidative phosphorylation in atrial fibres from a patient surgically operated on for aortic valve replacement. A Additions: fibres, 2 mM ATP, 20 U/mL pyruvate kinase (PK), 20 mM creatine (Cr), 2 mM ATP and 0.1 mM atractyloside (Atr). B The same as in (A), except that 2 mM AMP was added instead of Cr to stimulate respiration. Note the effective control over oxidative phosphorylation at the level of the adenine nucleotide translocator (inner mitochondrial membrane) revealed by Atr. V_O2 O₂ consumption; ww Wet weight. Adapted with permission from reference 33

The PK/PEP system also allowed direct monitoring of the fluxes of ADP between the endogenous and cytoplasmic compartments (Figure 4). The rate of decrease measured spectrophotometrically at 340 nm after the addition of Mg²⁺-ATP to skinned fibres, equals the total ATPase activity, which comprises both Mg²⁺- and Ca²⁺-activated enzymes (the reaction medium contained 0.1 μM Ca²⁺) (Figure 4B). The addition of glutamate and malate, the respiratory substrates for mitochondria, significantly decreased that flux, whereas atractyloside, a blocker of the ANT, exerted the opposite effect. These observations suggest that in conditions of oxidative phosphorylation, the ADP formed by ATPases is directly channelled (37) to mitochondria to be rephosphorylated into ATP, and the latter is directed back toward the ATPases to maintain continuous cycling of the endogenous nucleotides. The magnitude of direct channelling, quantitated as the difference between the fluxes before and after the addition of atractyloside, is approximately 50% of the maximal Mg²⁺-Ca²⁺-ATPase activity (35). The same experiment (adding 20 mM PCr into the medium after atractyloside) was used to estimate the function of MM-CK in the ATPase end of the CK-phosphotransfer network. As a result, the ADP flux available to the PK/PEP system was markedly decreased (by 66%) compared with that available without PCr (Figure 4B). This change is explained by the functional coupling between MM-CK and ATPase reactions, which enables the rephosphorylation of ADP produced by ATPases at the expense of PCr (Figure 4A) (38), thereby preventing ADP from being captured by the PK/PEP system.

Figure 4).

A The principle of competition between mitochondria (Mit) and the cytoplasmic pyruvate kinase (PK)/phosphoenolpyruvate (PEP) system or muscle type creatine kinase (MM-CK) for phosphorylation of ADP produced by the ATPases. B Original trace of registration of the ADP flux through the PK reaction in the presence of 5 mM PEP and the coupled lactate dehydrogenase (LDH) system in human atrial fibres. Optical density (OD) at 340 nm is equivalent to NADH concentrations in the medium. Atr Atractyloside; Cr Creatine; Glut/Mal Glutamate/Malate; PCr Phosphocreatine; Pi Inorganic phosphate. Adapted with permission from reference 33

It is known that the apparent K_m for exogenous ADP in mitochondrial respiration depends on the structural organization and intactness of membranous structures in cardiac cells. First, the hypo-osmotic treatment of skinned cardiac fibres changes the kinetics of mitochondrial respiration from monophasic to biphasic, which is attributable to two populations of mitochondria with apparent K_m values for exogenous ADP of approximately 300 μM and 30 μM for mitochondria with intact and destructed outer membranes, respectively (29,39). Second, the skinned cardiac fibres of desmin-deficient mice (40) also display populations of mitochondria with high and low K_m values for ADP, in cells with normal and disintegrated (due to lack of desmin) structures, respectively. Third, mild treatment of the skinned cardiac fibres with trypsin markedly increases the affinity of mitochondrial respiration to exogenously added ADP. In this process, the fibres displayed a very high sensitivity to trypsin (50 nM), as a shift in K_m from 300 μM to 350 μM in untreated fibres to 100 μM was induced (23). Along with this effect, trypsin drastically potentiated the inhibitory influence of the exogenous PK/PEP system on ATP-stimulated respiration (41). Both effects of trypsinization were associated with significant disorganization of the mitochondrial arrangement in the cardiac cells (41). Thus, it became clear that the accessibility of exogenous ADP to mitochondria and diffusion of endogenous ADP to the cytoplasm are restricted due to the existence of proteinous barriers in skinned cardiac cells (23,41,42). Only after selective destruction of these barriers with trypsin can exogenous ADP diffuse rapidly to the mitochondria (thus explaining the decreased K_m in activation of respiration) and endogenous ADP can feasibly equilibrate within the cytoplasmic phase, thereby becoming available for the PK/PEP system, which has an inhibititory effect on respiration.

The most relevant conclusion from all of these experiments is that in heart muscle cells, mitochondria and adjacent ATPases form structural and functional complexes (Figure 5), for convenience, called ICEUs (22,23). Within the ICEUs, energy is transferred from mitochondria to ATPases in the form of energy-rich phosphoryl groups via three specialized energy transfer systems. One of them, the CK-phosphotransfer network, comprising different CK isoenzymes, functions as follows: ATP generated in the matrix and exported to the intermembrane space by the ANT is converted into PCr by mi-CK due to its functional coupling to the ANT. PCr, in turn, is used to locally replenish ATP for ATPases by means of coupling with MM-CK (7,43). The sites of mitochondrial PCr production and its use are connected through the cytosolic MM-CK (44). The second system, the AK-phosphotransfer network is based on the coupling of the AK₂ isoenzyme to the ANT in mitochondria and of the AK₁ isoenzyme to ATPases (44). It has been shown that ADP locally produced by mitochondrial kinases is more effective in stimulating mitochondrial respiration than is cytoplasmic ADP. The third mechanism is direct channelling of adenine nucleotides between mitochondria and ATPases (22,36). Altogether, these systems exactly match the increased energy demand with enhanced ATP energy production when the workload increases, without significant fluctuations of cytosolic adenine nucleotide, PCr or creatine concentrations, thus maintaining the state of metabolic stability and a high ATP to ADP ratio near the ATPases (45). In normal conditions, the CK-phosphotransfer system represents the predominating mechanism of energy transduction (21,44).

Figure 5).

Schematic presentation of intracellular energetic units (ICEUs) in the cardiomyocyte. Through interaction with cytoskeletal elements, the mitochondria and sarcoplasmic reticulum (SR) are precisely fixed with respect to the structure of the sarcomeres of myofibrils between two Z-lines and correspondingly between two transverse (T)-tubules. Adenine nucleotides within the ICEU do not equilibrate rapidly with adenine nucleotides in the bulk-water phase (cytoplasmic phase). The mitochondria and ATPases of the SR and myofibrils are interconnected by metabolic channelling of reaction intermediates and energy transfer via creatine kinase (CK)- and adenylate kinase (AK)-phosphotransfer networks and direct ATP/ADP exchange. The protein factors (still unknown and marked as ‘X’), most likely connected to the cytoskeleton, fix the position of mitochondria and likely also control the permeability of the voltage-dependent anion channels for ADP and ATP. Adenine nucleotides within the ICEU and bulk-water phase may be connected by some more rapidly diffusing metabolites, such as creatine (Cr) and phosphocreatine (PCr). Synchronization of functioning of ICEUs within the cell may occur by the same metabolites (for example, inorganic phosphate [Pi] or PCr) and synchronized release of Ca²⁺ by SR (not shown) during the excitation-contraction coupling process (for details, see reference 23). ANT Adenine nucleotide translocator; mb Mitochondrial membrane

It is likely that in ICEUs, there exist localized diffusion barriers that isolate part of the cellular adenine nucleotides from their cytoplasmic bulk-phase pool, resulting in ATP compartmentation. This hypothesis is confirmed by recent studies (41,46) which analyzed the diffusion of adenine nucleotides in muscle cells by comparing the results of experimental kinetic measurements with the solutions of mathematical models and assuming that two different types of diffusion restriction for adenine nucleotides take place. One type of restriction was considered uniformly distributed, arising from the molecular crowding effects of proteins and being proportional to the diffusion distance (ie, one-half of the cardiac cell diameter). Another type of restriction was expected to result from local diffusion barriers at the borders or within the ICEUs (localized or nonuniform diffusion restriction). To discriminate between these two restriction types, the experimental results were compared with data obtained from mathematical models for the following processes: inhibition of endogenous ADP-stimulated respiration by the PK/PEP system competing with mitochondria for ADP; the kinetics of O₂ consumption stabilization after the addition of 2 mM ATP or ADP; the ADP concentration buildup in the medium after the addition of ATP; and the ATPase activity with inhibited mitochondrial respiration. It was revealed that when uniformly distributed intracellular diffusion was assumed, only the measurements of the respiration rate as a function of exogenously added ADP or ATP and PK/PEP system inhibition of respiration were mathematically reproducible (43). In contrast, when localized diffusion restriction near mitochondria and surrounding ATPases was postulated, the model reproduced all of the measurements (46). Thus, most likely, the organization of mitochondria into ICEUs results in the heterogeneity of the intracellular diffusion of ADP or ATP (41,46), which agrees with ³¹P-NMR studies showing the heterogeneity of phosphorus metabolites in skeletal muscles of fish (47) and rats (48).

THE CYTOSKELETAL NETWORK CONTROLS THE ARRANGEMENT OF MITOCHONDRIA AND ATPases WITHIN THE ICEUs IN CARDIAC CELLS

At present, the nature of the proteins locally forming barriers for the intracellular displacement of ADP is unclear, as is their role in organizing the mitochondria within the cell structure. These proteins are most likely related to the cytoskeleton because high K_m values for exogenous ADP in the regulation of respiration are preserved in ‘ghost’ fibres, from which myosin has been extracted and which contain mostly mitochondria, SR and cytoskeletal structures (42). The cytoskeleton consists of different fibres, classified according to their diameter as actin microfilaments (6 nm), intermediate filaments (10 nm), such as desmin and vimentin, and microtubules (25 nm) (49). Several of these proteins are known to maintain mitochondria in normal spatial arrangement. First, Appaix et al (42) recently showed that, in permeabilized cells, the mitochondria are wrapped into the microtubular network. After selective treatment with trypsin, the regular network progressively disappeared along with disarrangement of mitochondria and a decrease in K_m for ADP in the regulation of respiration. These data suggest an important role for microtubules in regulation of mitochondrial function. The mechanism may involve binding of the microtubules to the mitochondrial outer membrane via microtubule-associated proteins (50). Second, desmin filaments connect the mitochondria with the Z-disks of the sarcomere, thereby localizing these organelles in a series exactly near the A-band of the adjacent sarcomere (51). It has been suggested that desmin influences mitochondrial function by altering the proportion and content of the respiratory chain complexes of subsarcolemmal and interfibrillar mitochondria. Desmin may also regulate the mitochondrial shape, by contracting and stretching the mitochondrial membrane, in association with the contraction-relaxation cycle of the sarcomere. Its effect on mitochondrial affinity toward ADP can be mediated by the formation or stabilization of the mitochondrial contact sites between the inner and outer mitochondrial membranes or by binding to voltage-dependent anion channels directly, or via microtubule-associated protein 2 or plectin (52,53). The important physiological role of desmin has been convincingly demonstrated in desmin-deficient mice, whose myocardium exhibits disintegration of myofibrils and rearrangement of mitochondria (seen as their accumulation between the myofibrils and in subsarcolemmal clumps) (40). Thus, loss of desmin disintegrates the connections of mitochondria with other cellular structures. This results in severe functional consequences, such as increased apparent affinity to exogenous ADP in the regulation of respiration, decreased maximal rate of respiration and impaired functional coupling of mi-CK and the ANT, all observed in skinned fibres (40,54). Mitochondria in desmin-deficient cardiomyocytes are also more prone to opening of the mitochondrial permeability transition pore, which simulates apoptotic cell death (54). Trypsinization of skinned fibres does not affect the desmin molecules, whereas plectin is cleaved, resulting in the disappearance of its immunostain (42). Thus, increased mitochondrial affinity to exogenous ADP, observed after treatment of skinned cardiac fibres with trypsin, may result from impaired plectin-mediated connections of mitochondria to cytoskeletal proteins.

Recently, Kadaja et al (55) applied circulatory autoantibodies of patients with liver diseases to localize the cytoskeletal proteins potentially involved in organization of the ICEUs. This work recognized that cytoskeletal proteins (such as actin, myosin, desmin, tropomyosin, alpha-actinin, filamin and vimentin) are the autoantigenic targets in primary biliary cirrhosis and chronic hepatitis. Therefore, the effects of the immunoglobulin G (IgG) fraction purified from the sera of healthy persons and patients with these diseases on ADP-dependent respiration were studied in skinned fibres of oxidative (heart and musculus soleus) and glycolytic muscle (musculus gastrocnemius). It was found that IgGs from healthy persons and patients with primary biliary cirrhosis or chronic hepatitis had no effect in glycolytic muscles, but markedly inhibited the respiration of mitochondria in oxidative muscles, with a stronger inhibitory effect of IgGs from patients than from healthy persons. Laser confocal microscopy indicated that IgG was binding to the sarcomeric structures projecting at the Z-disk and M-line areas. Thus, it appears that the proteins of these structures may limit the access of exogenous adenine nucleotides to mitochondria, thereby decreasing the rate of ADP-dependent respiration.

CONTRACTION OF CARDIAC MUSCLE IS ASSOCIATED WITH ALTERED KINETICS OF REGULATION OF MITOCHONDRIAL FUNCTION

As mentioned previously, the mitochondria are very precisely arranged between the myofilaments, with each mitochondrion being positioned near the adjacent sarcomere (51). Similar observations were recently made with the analysis of confocal micrographs that permitted the quantification of the distances between mitochondrial centres across the whole muscle (56). It was demonstrated that these centres are arranged very regularly, in a crystal-like pattern throughout the whole muscle, due to the fixed juxtaposition of mitochondria with sarcomeres (56). The proteins controlling such an exact interaction between sarcomeres and mitochondria have yet to be defined; the desmin filaments connecting mitochondria predominantly to the Z-disks of sarcomeres are likely of major importance (51). Nozaki et al (51) demonstrated a linear relationship between changes in sarcomere length and mitochondrial length; although caused by different mechanisms. More recently, Kaasik et al (57) reported that controlled swelling of mitochondria is associated with increased developed tension due to changes in physical forces in skinned cardiac fibres. Thus, structural junctions between mitochondria and sarcomeres appear to be strong enough to transmit forces from one to another.

To monitor whether the mechanical interactions between mitochondria and sarcomeres are associated with functional changes, we studied the kinetics of the regulation of respiration in skinned cardiac fibres and cells in conditions of sarcomere shortening caused by increasing concentrations of cytosolic free Ca²⁺ (58). In the presence of ATP, increasing free Ca²⁺ concentrations up to 3 μM to 4 μM induced a strong contraction of permeabilized cardiomyocytes, with significant alterations in mitochondrial arrangement (Figure 6). In parallel, the apparent K_m for exogenous ADP or ATP in the regulation of respiration drastically decreased (Figure 7B), with only a modest increase in the V_max of respiration (by 40%), and with an optimum Ca²⁺ concentration of 0.4 μM and a subsequent decrease at higher Ca²⁺ levels (Figure 7A). These changes were not related to mitochondrial Ca²⁺ overload, and they were fully reversible after decreasing the free Ca²⁺ to diastolic levels (results not shown).

Figure 6).

The effect of increasing the free Ca²⁺ concentration from zero (A and C) to 1 μM (B and D) in permeabilized cardiomyocytes (A and B) and ‘ghost’ cells (from which myosin was extracted by KCl treatment) (C and D) in the presence of ATP (1 mM). Note that Ca²⁺ caused contraction in the permeabilized cells but not in the ‘ghost’ cells. In these experiments, the mitochondrial localization was examined using confocal microscopy and the autofluorescence of mitochondrial flavoproteins (A and B) and from the Rhod-2 signal (5 μM) in the mitochondrial matrix (C and D). Adapted with permission from reference 58

Figure 7).

Effect of free Ca²⁺ concentrations on the regulation of respiration in skinned cardiac fibres (SF) in the presence of 2 mM or 1.5 mM exogenously added ADP or ATP, respectively. A The dynamics of the maximal rate of respiration. B The dynamics of the apparent K_m in the regulation of respiration. Means and SD of three to six experiments are shown. V_O2 O₂ consumption; ww Wet weight. Adapted with permission from reference 58

An interesting observation was that excess Ca²⁺ exerted no effect on contractile and mitochondrial functions in ‘ghost’ fibres, from which myosin ATPase was extracted (Figures 6C and 6D). This is similar to data by Khuchua et al (59), who analyzed the effects of caffeine and Ca²⁺ on mitochondrial oxidative phosphorylation and actomyosin ATPase in saponin-skinned skeletal muscles. In contrast to intact skinned muscle fibres, in which caffeine stimulated both respiration and actomyosin ATPase activity due to Ca²⁺ efflux from SR, caffeine exerted no effect on respiration in fibres without myosin. It was concluded that the effects of caffeine on the velocity of mitochondrial oxidative muscles fibres can be explained exclusively by the activation of actomyosin ATPase. Again, these results do not support the concept of ‘parallel activation’ of mitochondria and ATPases. Instead, they suggest that the most probable primary role for Ca²⁺ in muscle cells is to activate the contractile system, which secondarily stimulates respiration via the augmented production of ADP. Notably, this view does not exclude the direct activation of mitochondrial enzymes by Ca²⁺, but leaves the mitochonrial enzymes a role for tuning the mitochondrial systems into their maximally activated state, thus being ready to handle the metabolic signals (eg, ADP – a major factor that stimulates respiration).

Clearly, the effectiveness of metabolic signalling by ADP strongly depends on structure-functional relationships of the interaction between mitochondria and sarcomeres. This is exactly what the concept of ICEUs assumes: if mitochondria and the contractile apparatus form strongly related complexes then changes in sarcomere length should modify mitochondrial function. Figure 8 demonstrates that hypercontraction significantly decreased the mitochondrial role of providing ATP for contraction and increased the proportion of ADP that became available for phosphorylation by the PK/PEP system. At the same time, a decreased K_m for ADP in regulation of respiration (Figure 7B) indicates that exogenous ADP became more readily available to the mitochondria. Thus, the ICEU concept offers novel mechanisms for controlling the interaction between ATPases and mitochondria (via contraction-related alterations in localized restrictions of the intracellular diffusion of adenine nucleotides). It is possible that the mitochondrial outer membrane represents a barrier and, in this case, sarcomere contraction causes the opening of the voltage-dependent anion channels to adenine nucleotides. Another possibility is that strong sarcomere shortening disintegrates the border structures of the ICEUs so that the diffusion restriction for adenine nucleotides through its barriers decreases, thus facilitating the diffusion of cytosolic ADP into the ICEUs.

Figure 8).

The effect of free Ca²⁺ concentration ([A] 0 μM and [B] 2 μM) on the kinetics of ADP production in dependence of ATP concentration in the presence (dashed line) or absence (continuous line) of oxidative phosphorylation in skinned cardiac fibres. The measurements were performed spectrophotometrically in solution B supplemented with 5 mM phosphoenolpyruvate, 0.24 mM NADH, a large excess of pyruvate kinase (20 U/mL) and lactate dehydrogenase (20 U/mL) at 340 nm. Effect of respiratory substrates 10 mM glutamate (Glut) and 2 mM malate (Mal) (Triangle). Effect of atractyloside (Atr) (98 μM) (Square). Note that in the absence of free Ca²⁺ (A), the mitochondria are capable of rephosphorylating approximately 50% of the total ADP flux produced by ATPases (see also Figure 4). In conditions of hypercontraction due to high free Ca²⁺ concentrations (B), the proportion of ADP directly channelled to mitochondria markedly decreases compared with that in (A). Means and SEM of three to six measurements are shown. *P<0.05 compared with the parameter value in the absence of oxidative phosphorylation. ww Wet weight. Adapted with permission from reference 58

ARE THE ICEUs AFFECTED IN DISEASE?

The concept of ICEUs predicts that any disturbance in the diffusion barriers permits ADP to accumulate in the ICEUs. If this process is accompanied with increased permeability of the mitochondrial outer membrane, high concentrations of ADP are transferred into the intermembrane space where they directly stimulate respiration and, thereby, nullify respiratory control via coupled reactions of mi-CK and the ANT. Thus, decreases in the apparent K_m for the regulation of respiration by ADP and reduced efficacy of creatine to attenuate this parameter may be criteria for the impaired structure of the ICEUs.

In support of this assumption, numerous studies (40,60–62) have shown that during ischemia/reperfusion injury or cardiac failure, the K_m for ADP decreases in association with significant ultrastructural alterations, such as disruption of the mitochondrial outer membrane and separation of mitochondria from the sarcomeres. Likewise, the coupling between mi-CK and the ANT is impaired (62–65). On the other hand, different protocols of experimental cardioprotection via ischemic or pharmacological preconditioning have resulted in the preservation of a high K_m for ADP and mi-CK/ANT interactions, and improved contractility (64,65). Considerable evidence exists that relates these kinetic changes to an altered cytoskeletal network. Nozaki et al (51), while studying the effect of hypoxia on the relationship between the changes in sarcomere and mitochondrial length, found the relationship to be less steep in hypoxic cardiac muscles than in normoxic specimens. In contrast to normoxic muscles, in which mitochondria were similar in length to the adjacent sarcomeres, the mitochondria were significantly shorter than the sarcomeres in the hypoxic muscle group, especially at sarcomere lengths above 2.0 μM (51). Undoubtedly, the connections between mitochondria and sarcomeres were disturbed by hypoxia. According to Ganote and Armstrong (66), hypoxia or ischemia harms many cytoskeletal components, thereby disintegrating the cell structure and causing cell injury and death. These processes involve the disruption of desmin, which impairs the linking of mitochondria to Z-disks (53,66–68) and alterations in other proteins. For example, microtubule alterations have been considered to be an early cellular reaction to ischemic challenge because microtubule disassembly precedes lactate dehydrogenase release, a sign of irreversible cellular injury (69,70).

The cytoskeletal system is largely perturbed in the failing heart as well. A progressive increase in desmin expression was observed during the transition from hypertrophy to failure in guinea pig hearts and in explanted failing human myocardium (71). However, desmin expression later becomes downregulated, so that in end-stage heart failure, the number of desmin-positive myocytes is decreased in association with reduced cardiac function (72). Also, a missense mutation of desmin (methionine for isoleucine at codon 451) has been detected and is considered to be responsible for idiopathic dilated cardiomyopathy (73). There is agreement that in failing heart, tubulin expression is increased, occurring in parallel to increased left ventricular end-diastolic pressure in humans (74). Reduction of microtubule hyperpolymerization by colchicine reverses myocyte stiffness and normalizes the contractile parameters, whereas taxol treatment of normal myocytes potentiates the accumulation of microtubules and functional disturbances (71). Therefore, it has been concluded that cytoskeletal abnormalities, rather than changes in contractile apparatus, cause the contractile dysfunction observed in compensated hypertrophy and in failing heart (71).

Along with the role of desmin and microtubules, the participation of dystrophin in the pathogenesis of cardiac disease also deserves attention. In contrast with skeletal muscles, where the transverse (T)-tubules lack dystrophin, this protein is found in cardiac T-tubules, which suggests that dystrophin not only serves in the transmission of contractile force, but controls other cellular processes as well (75). Yoshida et al (76) demonstrated that experimental chronic heart failure induced by coronary artery ligation in rats results in reduced myocardial tissue content of sarcoglycans and dystrophin. According to Meng et al (77), in cardiac cells, approximately 40% of dystrophin is tightly bound to the contractile apparatus, and the loss of that dystrophin fraction results in cardiac insufficiency. On the other hand, genetic mutation in dystrophin (X-linked trait) is the most common cause of muscular dystrophy, accounting for both the Duchenne and Becker phenotypes of the disease (78,79). In these diseases, progressive muscular weakness and atrophy are associated with impaired CK function (80,81). In skeletal muscle of dystrophin null mice (mdx), a widely used model of Duchenne muscular dystrophy, the mitochondria isolated from skeletal muscle exhibit impaired oxidative phosphorylation (80,82,83). Because these observations point to a potential role of dystrophin-mediated control over systems integrating mitochondria with ATPases, we directly addressed this issue in skinned muscle fibres of mdx mice (35). We found that both slow-twitch oxidative muscles (heart and musculus soleus) of mdx mice exhibited attenuated control of respiration by creatine, expressed as an attenuated decrease in the K_m for ADP in the regulation of respiration compared with that in wild-type muscle. This defect in the effect of creatine could not be ascribed to differences in mi-CK activity because the total CK activity of mdx ventricles was similar to that of wild-type ventricles. It is known that the muscles of mdx mice exhibit abnormally high rates of reactive oxygen species production (84), which entail peroxidation of cardiolipin (85), a component of the mitochondrial inner membrane that binds mi-CK (86). Therefore, impaired coupling of mi-CK to the ANT may result from altered properties of cardiolipin. The same study (35) revealed that compared with normal musculus soleus, the K_m for ADP in the regulation of respiration was decreased in mdx counterparts. A decreased K_m for ADP has been observed in oxidative skeletal muscle after genetic or chemical modification of the CK system (30,87). As suggested previously, such a change can be explained by decreased barrier function of the ICEUs, exposing mitochondria to control by cytosolic ADP. In line with this assumption, the cytosolic ADP is abnormally high in the skeletal muscle of patients with Duchenne muscular dystrophy (80,81,88).

Concerning the molecular basis, it is probable that the dystrophin molecules associated with T-tubules (75), extending from the cell surface to each Z-disk of the sarcomere, make up the localized barriers for diffusion of adenine nucleotides, thus participating in formation of the ICEUs. A lack of dystrophin in this region in mdx mice would then increase the permeability of that barrier, thus easing the access of cytoplasmic ADP to mitochondria. On the other hand, Tkatchenko et al (89) have demonstrated attenuated or increased expression of titin, troponin I, alpha-tubulin and Rac1 in skeletal muscle of mdx mice. Given the important role of these proteins in the structural organization of the muscle cell, their alterations may also facilitate the intracellular diffusion of ADP.

METABOLIC CONSEQUENCES OF THE DISINTEGRATION OF THE ICEUs

Impairment or loss of the structural interactions between mitochondria and ATPases may cause serious alterations of intra-cellular energy metabolism. Accumulation of ADP within the ICEUs is detrimental for several reasons. First, it decreases the ATP to ADP ratio and, thereby, reduces the free energy from ATP hydrolysis (45). Second, high ADP levels directly inactivate the ATPases and, as a result, the Ca²⁺ uptake by the SR becomes attenuated, leading to excessive Ca²⁺ accumulation in the cytoplasm, activation of Ca²⁺-dependent proteases and phospholipases and, eventually, degradation of cellular structures. At the level of the sarcomere, cross-bridge cycling becomes inhibited, giving rise to increased rigour tension, systolic and diastolic dysfunction, and a depressed Frank-Starling relationship, all typical for ischemic or failed heart. Third, an excess of ADP in the mitochondrial intermembrane space inhibits the synthesis of PCr in coupled reactions of mi-CK and the ANT. Decreased PCr synthesis, in turn, reduces the capacity of MM-CK coupled to ATPases to rephosphorylate the produced ADP and, thus, further promotes its accumulation. The feasibility of such a scenario in vivo is supported by ³¹P-NMR studies (90–93) that demonstrate a reduced intracellular PCr to ATP signal ratio, together with an increased ADP signal. In a failing heart, several other mechanisms are known to contribute to an impaired CK-phosphotransfer system. Depending on the duration of heart disease and the level of volume or pressure overload, the myocardial content of free creatine markedly decreases (94–99) and, thus, limits PCr synthesis. Down-regulated expression of mi-CK (63,93,99) and dyscoupling between mi-CK and the ANT(due to oxidation of mi-CK by oxygen free radicals and/or nitrosylation of its SH- groups [100]) could likely be additional factors in decreasing PCr levels. Dysfunction of the CK-phosphotransfer system, in turn, may promote cellular death via apoptosis because impaired interaction between mi-CK and the ANT favours the opening of the mitochondrial permeability transition pore (101).

Of note, there exist several means by which the heart adapts to the impaired function of the CK-phosphotransfer system. First, along with decreased expression of mi-CK and MM-CK isoenzymes, the expression of brain-type CK isoforms significantly increases (94); this increase is considered to improve the rephosphorylation of ADP near ATPases because brain-type CK has a higher affinity toward ADP than does MM-CK (102). Second, in conditions of metabolic inhibition of CK, hypoxia and heart failure, the role of AK in intracellular energy transfer may increase because intracellular accumulation of ADP favours the use of its beta-phosphoryls by AK for ATP formation (103). However, activation of the AK-phospho-transfer system cannot fully compensate for weakened energy transfer via the CK system because the sum of the phosphoryl transfer by these two systems lags behind the rate of ATP turnover in failed heart (103). This means that exhange by other means (eg, simple diffusion or direct channelling of ADP) must increase. This was experimentally proven by Braun et al (35), who demonstrated that direct ADP channelling between ATPases and the mitochondria significantly increased in the skinned cardiac fibres of mdx mice. Finally, accumulation of ADP in the cytoplasm of cardiac cells inevitably stimulates the degradation of adenine nucleotides, which is also a characteristic feature of diseased heart. One of the metabolites, AMP, may exert additional influences via the activation of AMP-activated protein kinases (AMPKs) (eg, inhibition of protein synthesis known to result from energy depletion [104]). The observation that 5-aminoimidazole-4-carboxyamide 1-beta-D-ribofuranoside, an activator of AMPK, also causes increases in eukaryotic elongation factor 2 phosphorylation, points to a key role for AMPK in the effects associated with energy depletion (104).