Skeletal muscle ATP synthesis and cellular H(+) handling measured by localized (31)P-MRS during exercise and recovery

Abstract

(31)P magnetic resonance spectroscopy (MRS) is widely used for non-invasive investigation of muscle metabolism dynamics. This study aims to extend knowledge on parameters derived from these measurements in detail and comprehensiveness: proton (H(+)) efflux, buffer capacity and the contributions of glycolytic (L) and oxidative (Q) rates to ATP synthesis were calculated from the evolutions of phosphocreatine (PCr) and pH. Data are reported for two muscles in the human calf, for each subject and over a wide range of exercise intensities. 22 subjects performed plantar flexions in a 7T MR-scanner, leading to PCr changes ranging from barely noticeable to almost complete depletion, depending on exercise protocol and muscle studied by localized MRS. Cytosolic buffer capacity was quantified for the first time non-invasively and individually, as was proton efflux evolution in early recovery. Acidification started once PCr depletion reached 60-75%. Initial and end-exercise L correlated with end-exercise levels of PCr and approximately linear with pH. Q calculated directly from PCr and pH derivatives was plausible, requiring fewer assumptions than the commonly used ADP-model. In conclusion, the evolution of parameters describing cellular energy metabolism was measured over a wide range of exercise intensities, revealing a relatively complete picture of muscle metabolism.

Figures

Figure 1

Schematic overview of the common understanding of the temporal evolution of muscle PCr concentration, cytosolic pH and ATP synthesis rates during rest, exercise and recovery (as e.g. in ref. 21). Grey PCr and pH lines depict what one might call ‘pure oxidative’ exercise, i.e. exercise at relatively low intensity, below the lactate threshold, in which PCr breakdown buffers the temporary mismatch between the step increase in ATP demand and the slower (usually exponential) increase in oxidative ATP synthesis, with negligible contribution by glycolytic ATP synthesis; black PCr and pH lines depict behavior during ‘mixed’ exercise, i.e. exercise at higher intensity, above the lactate threshold, in which glycolytic ATP synthesis is also appreciable. The rates of ATP generation by PCr breakdown, glycolysis (L) and oxidative phosphorylation (Q) must equal ATP demand (U), which is shown here as constant for constant-intensity exercise below ‘critical power’: the notional pattern of L and Q illustrated here shows an intermediate case where glycolytic ATP synthesis is not negligible, but is transient, so that by the end of exercise ATP is supplied only by oxidative ATP synthesis. During recovery from exercise at any intensity, PCr is replenished at the expense purely of oxidative ATP synthesis, the kinetics being exponential at low exercise intensity, following more complicated kinetics after exercise where pH has significantly fallen. The changes in pH in exercise result from the balance between the alkalinizing effect of PCr breakdown (dominant at the start), the progressive acidifying effect of glycolytic ATP synthesis, and the alkalinizing (or acidification-moderating) effect of net acid efflux; during recovery pH is restored by net acid efflux despite the H+-generation which accompanies PCr resynthesis (whose early post-exercise dominance can cause a transient initial-recovery acidification).

Figure 2

Evolution of PCr (a,b) and pH (c,d) during exercise and recovery. Data from GM and SOL which were acquired at the same ergometer force, i.e. applying protocol A, are shown for two representative subjects (a,c). Time courses representing similar end-exercise PCr depletion in both muscles (two pairs with high or low depletion in each) were selected from data acquired in four different subjects, (b,d). This was achieved by applying different ergometer forces, i.e. using protocol B. Vertical lines indicate start and end of exercise.

Figure 3

Qmax of all subjects from protocol A (40% MVC ergometer force) and of four subjects from protocol B (different ergometer forces), selected by having reached similar PCr depletion in both muscles: Qmax vs. end-exercise values of PCr depletion (a) and pH (b). Qmax in GM and SOL in protocol A (c) and protocol B (d). Evolution of Q with [ADP] during recovery in protocol A (e) and the selected subjects from protocol B (f). Dashed lines depict the theoretical relation according to the ADP-model, at Qmax of 0.3, 0.5, 0.7 and 0.9 mM/s.

Figure 4

pH vs. PCr depletion during exercise (left) and recovery (right). Data from GM and SOL at the same ergometer force of two representative subjects (a,b), and GM of all 15 datasets measured with protocol A (c,d). Group averages are depicted as underlying wide dark-grey trajectory (datasets without acidification excluded). Grey solid lines represent neutral pH and ADP isolines, the red circle indicates the PCr depletion range at the starting acidification (in line with the ‘lactate-threshold’ model). Exemplary data from protocol B with ~20% and ~50% end-exercise PCr depletion from each muscle (e,f), same datasets as shown in Fig. 2b,d. Arrows indicate direction of temporal evolution, dashed grey lines in recovery plots depict prior evolution during exercise. Spectra of two consecutive time points were averaged, successive data points are equidistant in time (every 12 s).

Figure 5

Glycolytic ATP synthesis rate (L), as calculated from the time-derivatives of PCr and pH evolution, is very low at exercises resulting in low PCr depletion. Above about 65% a higher end-exercise PCr depletion is correlated with much higher L, both for initial (a) and end-exercise (c) values of L, which is in line with the ‘lactate-threshold’ model. A linear regression analysis shows that L is correlated linearly with end-exercise pH, again both initial (b) and end-exercise L (d). Initial values of L are quantified at 45 s of exercise (cf. grey vertical line in Fig. 6) ensuring absence of E and all its associated uncertainties (cf. Eq. 10). Protocol A: 5 min exercise, Protocol B: 3 min exercise. Two spectra averaged per data point before processing.

Figure 6

ATP synthesis rates during and after exercise, at strong (a) and medium (b) activation measured in GM using protocol A, and at low activation measured in GM (c) and SOL (d) using protocol B. L (red) and Q (green) were calculated from the derivatives of PCr (blue) and pH time courses, and corrected for H+ efflux which was assumed to be linear with pH (Eqs 13 and 14). QADP (dashed) was quantified using the ADP time courses (Eqs 4 and 5) and is shown for comparison. The black vertical lines mark the end of exercise, the grey lines at 45 s indicate the moment defining initial L used in Fig. 5. Two averages per time point (12 s temporal resolution).

Figure 7

Individual cytosolic buffer capacity β was non-invasively quantified from the change in pH and PCr within the first 6 seconds of exercise, where contribution by L is considered to be negligible. This was successful in most measurements (see text), and yielded realistic results.

Figure 8

In early recovery (a) the individually quantified proton efflux rates E showed a strong increase at acidifications below ΔpH ≈ −0.4, whereas at higher acidifications E was decreasing (the arrows indicate the direction of temporal evolution). In later recovery (b,c), after lowering acidification below about ΔpH ≈ −0.2, the E showed approximately linear correlation with acidification (shown for comparative purposes as grey dashed line), supporting the applicability of the linear E estimation during exercise (cf. Eq. 13). E was calculated from derivatives of pH and PCr evolutions (cf. Eq. 11) using 2x averaged data in initial recovery (a) and 5x averaged data in later recovery (b,c) where SNR nearly vanishes, especially in measurements at higher acidification (c). All data from GM, protocol A (same color-code as in Fig. 4c,d), 3 measurements not included for lack of acidification. Successive data points are equidistant in time with 12 s in (a) and 30 s in (b,c). The fat grey trajectory depicts E calculated from group averages of PCr and pH (a).

Figure 9

Typical position of the voxels in soleus and gastrocnemius medialis muscle, respectively, for localized 31P -MRS at 7 Tesla (a). Unaveraged sample spectra of the respective muscles at rest and at the end of exercise (b) show the prominent PCr peak and the smaller Pi peak.