Subjects.

Forty-eight subjects (20 healthy untrained subjects and 28 cyclists ranging from recreationally trained to professional) volunteered to participate in this study (Table 1). Cyclists competed at the Olympic or (inter)national level, except for four amateur cyclists. Previously published data from our laboratory of 14 chronic heart failure (CHF) patients and 6 healthy subjects (4) were reexamined (see details below). Therefore, 68 subjects were included in our analysis. Before participation, experimental procedures and risks of the study were explained, and all subjects provided written informed consent. The study was conducted according to the principles of the Declaration of Helsinki and was approved by the medical ethics committee of the VU Medical Center, Amsterdam, the Netherlands (NL43423.029.13 and NL49060.029.14).

Whole body V̇o_{2 max} during cycling.

Subjects performed a maximal incremental exercise test on a cycle ergometer to voluntary exhaustion, despite verbal encouragement (respiratory exchange ratio of 1.20 ± 0.07). V̇o_{2 peak} was calculated as the highest 30-s value achieved during the maximum-effort incremental test and is considered a valid index of V̇o_{2 max} in subjects exercising to their limit of exercise tolerance (15). Respiratory data were analyzed breath-by-breath using open circuit spirometry and expressed at STPD (Cosmed Quark CPET; Cosmed, Rome, Italy). Before every test, the gas analyzer and volume transducer were calibrated according to the manufacturer's instructions. Subjects were instructed to avoid strenuous exercise and alcohol consumption within 24 h before the test and caffeinated beverages and meals within 3 h before the test.

Skeletal muscle biopsy.

Biopsy samples were obtained from the vastus lateralis using a modified Bergström needle technique. The vastus lateralis muscle (part of the quadriceps femoris muscle) acts as knee extensor and for cycling exercise is predominantly involved in the pushing phase and power-producing phase of the pedal cycle (67). Biopsy sites were locally anesthetized with a 2% lidocaine solution, and an incision of <1 cm was made through the skin and fascia latae. The biopsy needle was inserted ∼15 cm above the patella to a depth of ∼4 cm. Biopsy samples were carefully removed and aligned according to their muscle fiber arrangement using a magnifying glass. Subsequently, samples were frozen in liquid nitrogen. After being frozen, biopsy samples were placed in a cryostat and cut in 10-μm-thick sections at −20°C. Sections were collected on polylysine-coated slides.

SDH histochemistry.

SDH activity was determined using quantitative histochemistry, which has been described in detail elsewhere (4, 51). In short, biopsy sections were incubated at 37°C in a medium consisting of 0.4 mM tetranitroblue tetrazolium (Sigma, St. Louis, MO), 75 mM sodium succinate, 5 mM sodium azide, and 37.5 mM sodium phosphate buffer, pH 7.6 (51). Biopsy sections were incubated for 20 min. Images were made with ×10 or ×20 objectives, and absorbance was measured by microdensitometry (46) with an interference filter at 660 nm (51) using National Institutes of Health ImageJ https://imagej.nih.gov/ij/). Weighed average SDH activity was determined from spatially averaged SDH activity of >40 randomly selected individual cells, including the subsarcolemma mitochondria. The randomly selected cells mirrored the distribution of high oxidative (i.e., type I) and low oxidative (i.e., type II) fibers within the biopsy section (i.e., SDH activity of individual fibers can be used to discriminate between fiber types similar to myofibrillar ATPase activity; see Ref. 5). Reproducibility of quantitative SDH histochemistry was assessed by comparison of absorbance values of two subsequent measurements of the same muscle tissue.

Temperature affects SDH activity in rats, mice, and fish (13, 51), but the effect of temperature on SDH activity has not been assessed in humans. In the present study, we assessed the relationship between incubation temperature and SDH activity in human muscle tissue sections that were incubated at 32, 37, and 42°C (Fig. 3). Subsequently, the temperature quotient Q₁₀ (the increase in SDH activity with a 10°C increase in temperature) was determined.

Temperature-adjusted SDH activity.

For comparison of SDH activity to whole body V̇o_{2 max}, SDH incubation temperature should equal quadriceps muscle temperature at maximal exercise. Previously, similar data on SDH activity and V̇o_{2 max} during cycling have been obtained in 14 CHF patients and 6 healthy controls, but SDH activity was not adjusted to muscle temperature at maximal exercise. In the present study, SDH activity was adjusted to temperature using the Q₁₀ for human tissue and estimated quadriceps muscle temperature at maximal effort, being 36.5°C in CHF patients (65) and ∼41°C in healthy subjects and cyclists (60).

Skeletal muscle fiber V̇o_{2 max}.

Previously, it has been shown that paired determination of SDH activity by quantitative histochemistry and ex vivo maximum rate of oxygen consumption of intact single fibers under hyperoxic conditions are proportionally related in Xenopus (both determined at 20°C; see Ref. 75) and rat myocardial trabeculas (both determined at 38°C; see Ref. 16). Therefore, skeletal muscle fiber V̇o_{2 max} (in nmol·mm⁻³·s⁻¹) was calculated as 6,000 × the temperature-adjusted SDH staining rate [in change of absorbance at 660 nm (ΔA₆₆₀)·μm section thickness⁻¹·s incubation time⁻¹] as described previously (16, 75).

Mitochondrial oxidative capacity predicted from SDH activity.

Oxidative capacity of mitochondria was predicted from temperature-adjusted SDH activity, total skeletal muscle mass (see details below), and body mass (Eq. 1). This prediction involves the following assumptions: 1) that all oxygen at V̇o_{2 max} is consumed by skeletal muscle mitochondria, 2) that all mitochondria are active at their V̇o_{2 max} during maximal cycling exercise, and 3) that SDH activity of the vastus lateralis locomotor muscle represents SDH activity of whole body musculature. Because these assumptions can be debated (see details below), another prediction of mitochondrial oxidative capacity is made that also accounts for 1) differences in SDH activity between arm and leg muscles, 2) active skeletal muscle mass during maximal cycling exercise, and 3) oxygen consumption of other organs (see details below).

where mitochondrial oxidative capacity is calculated at STPD (in ml·kg⁻¹·min⁻¹), SDH is weighed SDH activity of vastus lateralis adjusted for temperature (in ΔA₆₆₀·μm⁻¹·s⁻¹), V_m is molar volume of oxygen at STPD (22.4 l/mol), δ is muscle density (1.04 kg/l), M_m is skeletal muscle mass (kg), and M_b is body mass (kg).

Skeletal muscle mass.

Total skeletal muscle mass (including respiratory muscles) was estimated from an anthropometric regression model that has been shown to predict total skeletal mass obtained from whole body magnetic resonance imaging in 244 nonobese subjects (47):

where Ht is height (m), sex = 1 for males or 0 for females, ethnicity = −1.2 for Asian (1 CHF patient) and 0 for Caucasian.

SDH activity of skeletal muscle.

In animals, mean mitochondrial volume density of whole body skeletal musculature is not significantly different from mitochondrial volume density of the hindlimb musculature (33). In humans, however, mitochondrial volume density of arm and leg musculature may differ due to bipedal locomotion. In untrained subjects SDH activity of the deltoid muscle was 80% of the SDH activity in vastus lateralis (26), whereas in leg-trained subjects this was 65% (26). Therefore, we assume that SDH activity of arm musculature is 80% of vastus lateralis SDH activity in healthy controls and CHF patients and 65% in cyclists. Here, arm muscle mass accounts for 20% of total skeletal muscle mass (i.e., ∼6 kg; see Ref. 10).

Active skeletal muscle during cycling exercise.

During maximal leg cycling exercise, mitochondria in the arm musculature are likely not active at their V̇o_{2 max}. In healthy subjects, oxygen uptake of the arms during leg pedaling at maximal effort (26.1 ml·kg muscle⁻¹·min⁻¹) yielded ∼35% of V̇o_{2 max} of the arms attained during arm cranking (77.8 ml·kg muscle⁻¹·min⁻¹; see Ref. 12). Therefore, we assume that during maximal leg cycling exercise mitochondria in the arm musculature are active at 35% of their V̇o_{2 max}.

Oxygen consumption of other organs.

At maximal effort, oxygen consumption by the brain was assumed to be similar to resting conditions, ∼20% of measured oxygen consumption in rest (∼1 ml·kg⁻¹·min⁻¹; cf. Ref. 17). Maximal myocardial oxygen consumption has been estimated to be 0.45 mM/s (18, 57), that is, 670 ml·kg heart muscle⁻¹·min⁻¹ at core temperature during maximal exercise and ∼3 ml·kg body mass⁻¹·min⁻¹ using an average heart mass of 380 g in males and 330 g in females (22). Note that oxygen consumption of other organs such as the liver, kidneys, and digestive system is neglected but is assumed to be less than in resting conditions (<2 ml·kg⁻¹·min⁻¹; cf. Ref. 17).

where V̇o_2rest is oxygen consumption at rest, V̇o_{2max, heart} is 670 ml·kg heart muscle⁻¹·min⁻¹, and M_h is heart mass (330 g in females, 380 g in males). SDH = 1/5 × 0.35 × (SDH_arm/SDH_leg) × SDH_VL + 4/5 × SDH_VL, that is, SDH activity of arm musculature is not fully active during maximal cycling exercise (35%) and is lower than leg SDH activity (65% in cyclists, 80% in controls and CHF patients) and accounts for ∼20% of whole body skeletal muscle mass, with SDH_VL the SDH activity determined from biopsy samples of the vastus lateralis.

Statistics.

All data are presented as individual values or as means ± SD, unless otherwise indicated. Differences between groups were assessed by one-way ANOVAs (between-factor group: CHF patients, controls, cyclists). Differences in SDH activity with incubation temperature were assessed by repeated-measures ANOVA (within-factor temperature: 32, 37, and 42°C). Post hoc comparisons with Bonferroni correction were performed to detect differences between groups and temperatures. The relationships between SDH activity and whole body V̇o_{2 max}, SDH activity and V̇o_{2 max} normalized for skeletal muscle mass, and between mitochondrial oxidative capacity and whole body V̇o_{2 max} were assessed by linear regressions. To check whether regression lines differed from the line of identity or regression line from literature, differences in slope and intercept were tested by confidence intervals of the regression coefficients. Differences were considered to be significant if P < 0.05.

Histochemical method for SDH activity.

Figure 1 shows muscle fiber cross sections of vastus lateralis that were incubated for SDH using quantitative enzyme histochemistry. SDH activity was significantly different between CHF patients, controls, and cyclists.

Reproducibility was determined from two measurements of mean SDH absorbance without incubation and in high oxidative (i.e., type I) and low oxidative (i.e., type II) muscle fibers (5) after 20 min of incubation (Fig. 2). It is concluded that average SDH absorbance values were similar for both measurements.

The relationship between SDH absorbance and incubation temperature in human muscle tissue is shown in Fig. 3. Mean absorbance values of high and low oxidative fibers were significantly different between 32, 37, and 42°C (P < 0.001). At 42°C, SDH absorbance is 1.76 times higher compared with that at 32°C. Hence, SDH enzyme activity of all subjects was adjusted for muscle temperature using a Q₁₀ of 1.76.

Relationship between SDH activity and whole body V̇o_{2 max}.

Figure 4 shows that SDH activity adjusted for muscle temperature is closely related to whole body V̇o_{2 max} normalized to body mass (r² = 0.81, P < 0.001) and whole body V̇o_{2 max} normalized to skeletal muscle mass (r² = 0.83, P < 0.001). Without adjustment for muscle temperature, correlation coefficients are slightly lower for V̇o_{2 max} normalized to body mass (r² = 0.79, P < 0.001) and V̇o_{2 max} normalized to skeletal muscle mass (r² = 0.80, P < 0.001). These results indicate that SDH activity is proportional to the V̇o_{2 max} in humans, irrespective of training status (V̇o_{2 max} ranging from 9.8 to 79.0 ml·kg⁻¹·min⁻¹). Normalization of V̇o_{2 max} to skeletal muscle mass allows direct comparison with ex vivo measurements from intact animal myocytes in hyperoxic Tyrode solution (represented by the solid line; see methods). In humans, the relationship between temperature-adjusted SDH activity and V̇o_{2 max} differed significantly from the predicted relationship from ex vivo measurements [slope <7.6 × 10⁶, confidence interval (CI) = (5.3 × 10⁶, 6.7 × 10⁶), P < 0.05; intercept = 0, P > 0.05; see Refs. 16 and 75]. At V̇o_{2 max}, the oxidative enzyme activity is not fully exploited. Note that V̇o_{2 max} normalized to skeletal muscle mass may provide an overestimation of the true value because it also includes oxygen uptake by tissues other than skeletal muscles (Fig. 4B).

Relationship between mitochondrial oxidative capacity and whole body V̇o_{2 max}.

The relationship between body-mass-specific V̇o_{2 max} during leg cycling and mitochondrial oxidative capacity from temperature-adjusted SDH activity and total skeletal muscle mass is shown in Fig. 5A. In CHF patients, controls, and cyclists, mitochondrial oxidative capacity is proportionally related to V̇o_{2 max} during cycling (r² = 0.88–0.89, P < 0.001). The relationship between V̇o_{2 max} and mitochondrial capacity based on equation 1 significantly differed from the line of identity [slope <1, CI = (0.72 0.86), P < 0.05; intercept = 0, P > 0.05] and indicates that subjects do not fully use (∼85%) the oxidative enzyme capacity at V̇o_{2 max} during cycling exercise. Note that in Fig. 5A it is assumed that all oxygen at V̇o_{2 max} is consumed by skeletal muscle mitochondria, that all mitochondria are active at their V̇o_{2 max} during maximal cycling exercise, and that SDH activity of vastus lateralis represents SDH activity of whole body musculature. Figure 5B displays the relationship between V̇o_{2 max} and mitochondrial oxidative capacity based on equation 3. This prediction suggests that V̇o_{2 max} during cycling is 90 ± 14% of the mitochondrial oxidative capacity. Even though the relationship in Fig. 5B tended to differ from the line of identity, the slope was not significantly different from 1 [slope = 1, CI = (0.84 1.00), P > 0.05; intercept = 0, P > 0.05]. Note that there is still considerable individual variation between mitochondrial oxidative capacity (Eq. 3) and measured V̇o_{2 max} during cycling (coefficient of variation = 15.2%), which is also illustrated by lower correlation coefficients in separate subgroups (CHF patients r = 0.47, P = 0.107; healthy controls r = 0.61, P < 0.001, cyclists r = 0.67, P < 0.001). Mitochondrial oxidative overcapacity did not differ between CHF patients, controls, and cyclists (P = 0.108).

Symmorphosis in humans: Mitochondrial oxidative capacity is proportionally related to V̇o_{2 max}.

Our results confirmed that V̇o_{2 max} scales proportionally with SDH activity (r² = 0.81) and mitochondrial oxidative capacity (r² = 0.89) across heart failure patients, healthy controls, and cyclists. These results are in line with previous cross-sectional studies that have shown that in heterogeneous healthy populations the body-mass-specific V̇o_{2 max} is related to SDH activity (r = 0.79; see Ref. 14) and mitochondrial volume density (r = 0.82; see Ref. 34). Note that in more homogeneous populations correlations between SDH activity and V̇o_{2 max} were found to be substantially lower (r = 0.23; see Ref. 21). Our subgroups also showed lower correlations, which in combination with our coefficient of variation indicated considerable individual differences in mitochondrial oxidative overcapacity (see Unexplained variance of mitochondrial oxidative capacity and V̇o_{2 max}). However, mitochondrial oxidative overcapacity did not differ between CHF patients, controls, and cyclists, which supports existence of symmorphosis in humans. CHF impairs V̇o_{2 max} and mitochondrial oxidative capacity due to the limited cardiac output (52), as a result of reduced O₂ flux to the mitochondria from impaired red blood cell (RBC) flux and velocity, reduced percentage of flowing capillaries supporting RBC flux during rest and exercise, impaired Q̇o₂/V̇o₂ matching, low microvascular Po₂, and high fractional oxygen extraction (19, 29, 52, 55). Our results indicate that these impairments in both V̇o_{2 max} and mitochondrial oxidative capacity are proportional in CHF patients and therefore support existence of symmorphosis in humans. Also trained cyclists did conform to the proportional relationship between V̇o_{2 max} and mitochondrial oxidative capacity. In humans, however, individual increases in V̇o_{2 max} induced by endurance training are generally not proportional to changes in oxidative enzyme activity of the skeletal muscle involved in endurance training (cf. Refs. 30 and 61). This discrepancy may be explained by the time course of (acute) training adaptations. For instance, V̇o_{2 max} tends to increase by 15–30% during the first 2–3 mo of endurance training, whereas 40–50% increases in V̇o_{2 max} may occur over 9–24 mo of training (61). However, concomitant changes in the tricarboxylic acid cycle and respiratory chain enzymes are much faster, displaying half-lives in the order of 1–3 wk (30, 61). Therefore, increases in V̇o_{2 max} and mitochondrial capacity are likely disproportional during acute training adaptations (<9 mo) but may become proportional with chronic training adaptations (>9 mo), such as in our trained cyclists with an extensive training history (120 mo on average). Also note that trainability of V̇o_{2 max} may markedly vary between subjects and is strongly related to heritability (63). Thus, mitochondrial oxidative capacity scales with V̇o_{2 max} and largely accounts for differences in body-mass-specific V̇o_{2 max} in this heterogeneous population.

Animals conform to the concept of symmorphosis assuming that all parts of the O₂ pathway, contributing to V̇o_{2 max}, are designed proportionally (70). Across sedentary and athletic animal species, it was shown that mitochondrial volume density is proportional to body-mass-specific V̇o_{2 max} explaining 97% of its variance (35). Athletic animals presented a 2.5-fold greater body-mass-specific V̇o_{2 max}, which was matched by a 2.5-fold larger mitochondrial volume density in their muscles (35). Therefore, the maximal rate of O₂ consumption by skeletal muscle has been considered to be invariant across animals (4–5 ml O₂·ml mitochondria⁻¹·min⁻¹; see Ref. 5). Our cross-sectional observation suggests that also in a heterogeneous population of human subjects V̇o_{2 max} was matched by mitochondrial oxidative capacity: both V̇o_{2 max} and SDH or predicted oxidative capacity were ∼1.5-fold greater in cyclists compared with controls, ∼3-fold greater in cyclists compared with CHF patients, and ∼2-fold greater in controls compared with CHF patients. Our results contradict previous findings (24), which challenged the existence of symmorphosis in humans and suggested that the relationship between V̇o_{2 max} during cycling and mitochondrial oxidative capacity (from permeabilized fibers) varied with exercise training status. In this previous study (24), measurements in untrained subjects did conform to the concept of symmorphosis, whereas trained subjects did not. However, this conclusion is based on data of only four trained subjects and on a smaller observed range of V̇o_{2 max} (31–66 ml·kg⁻¹·min⁻¹). In the present study mitochondrial oxidative capacity is proportional to V̇o_{2 max} explaining 81–89% of its variance when observing a larger range of V̇o_{2 max} (9.8–79.0 ml·kg⁻¹·min⁻¹). Hence, similar to animals, a higher oxygen uptake demand is met by increasing mitochondrial enzyme activity while oxidative capacity per mitochondria remains invariant across human individuals as well. Note that the observed marginal mitochondrial oxidative overcapacity does not challenge the concept of symmorphosis, since small excess in mitochondrial capacity at maximal exercise may serve as a safety factor (70). Thus, our data support the concept of symmorphosis in a heterogeneous population of human subjects indicating that mitochondrial oxidative capacity is proportional to V̇o_{2 max}.

Mitochondrial oxidative overcapacity estimated from SDH activity.

Our subjects used on average 90% of their mitochondrial capacity during maximal cycling exercise, suggesting that mitochondrial oxidative capacity exceeded O₂ use by ∼11%. Based on the relationship between mitochondrial respiration and cellular Po₂, one could calculate at what percentage of their maximal capacity the mitochondria operate during dynamic exercise. Assuming first-order Michaelis Menten kinetics for mitochondrial oxygen consumption (58), the maximum obtainable V̇o₂ relative to the theoretical V̇o_{2 max} (under hyperoxic conditions) is:

where K_m is the Michaelis constant for oxygen of the mitochondria (0.5 Torr; see Ref. 58) and Po₂ is the mitochondrial oxygen tension (3.1 Torr during maximal cycling exercise estimated from myoglobin saturation and the P₅₀ of myoglobin for oxygen; see Ref. 54). Following equation 4, V̇o₂/V̇o_{2 max} = 0.86. Therefore, during maximal cycling exercise, the mitochondria should theoretically operate at ∼86% of their maximal oxidative capacity, which agrees well with the present findings. Further mitochondrial inhibition (e.g., by nitric oxide) is therefore not necessary to explain the reported oxidative overcapacity. It should be noted that equation 4 assumes homogeneous oxygen tension inside the muscle cells, which is physiologically impossible (for further details, see Refs. 20 and 23). Note that the use of mitochondrial oxidative capacity (90 ± 14%) is similar to the exploited mitochondrial oxidative capacity (90 ± 15%) when based on the assumption that skeletal muscle mass accounts for ∼90% of the total oxygen consumption at maximal exercise (35) (i.e., instead of estimating oxygen consumption by the heart and brain). Moreover, considering that V̇o_{2 max} is 5–10% higher during running vs. leg cycling (1, 6) and with arm musculature being fully active, the exploited mitochondrial oxidative capacity during running is estimated to be comparable to leg cycling (between 86 ± 13 and 90 ± 14%, respectively). Therefore, mitochondrial oxidative overcapacity of the present study appears to be robust and agrees well with mitochondrial overcapacity calculated from cellular oxygen tension.

Mitochondrial oxidative overcapacity with enhanced oxygen supply.

Our results indicate that circumventing limitations to V̇o_{2 max} can increase V̇o_{2 max} on average by ∼11% at most. What happens to V̇o_{2 max} when limitations of O₂ supply are reduced? Because the O₂ pathway from atmosphere to mitochondria is an in-series system, it is evident that every step significantly impacts V̇o_{2 max} (72). Acute interventions that enhance O₂ supply have shown to increase V̇o_{2 max}. Hyperoxia, for instance, by increasing inspired O₂ fraction (Fi_{O2) increased whole body V̇o2 max with 2–5% (74) and leg V̇o2 max with 8% (40) by enhancing oxygen diffusion at the lungs and from microvessels to mitochondria (40, 74). Blood reinfusion of 900-1,350 ml elevated oxygen-carrying capacity of the blood and thereby increased V̇o2 max by 4–9% (68). Moreover, V̇o2 max of one leg derived from blood flow and arteriovenous O2 difference measurements was 5–14% higher in one-legged cycling compared with two-legged cycling exercise because of the higher leg blood flow in one-legged exercise (38). The lower leg blood flow during maximal exercise with two legs compared with one leg is likely due to vasoconstriction of active muscles that is required to avoid hypotension and increase transit time at the muscle capillaries (38, 56). The theoretical model of Wagner, linking lungs, circulation, and muscles to V̇o2 max, predicts that an isolated 20% increase in ventilation, FiO2, lung diffusion capacity, muscle O2 diffusion conductance, hemoglobin concentration, or blood flow will increase V̇o2 max by only 1.3–5% (59, 71). Enhancement of oxygen supply can also be studied together with noninvasive assessment of postexercise PCr recovery kinetics with ³¹P-MRS (27, 28, 45), since PCr recovery kinetics have been used to estimate mitochondrial oxidative capacity (43). Previous studies have shown that enhanced oxygen supply with hyperoxia (27) and reactive hyperemia induced by cuff occlusion (45) improved both PCr recovery kinetics and estimated mitochondrial respiration rate; however, the quantitative interpretation of muscle mitochondrial capacity remains difficult (for review, see Ref. 37). Therefore, evidence indicates that increases in whole body V̇o2 max due to enhanced O2 supply are generally within a 10% range and agree well with our findings of excess mitochondrial capacity.}

Unexplained variance of mitochondrial oxidative capacity and V̇o_{2 max}.

Even though V̇o_{2 max} and mitochondrial oxidative capacity are related, there are still considerable individual differences in mitochondrial oxidative overcapacity. These individual differences may arise from methodological errors in estimating mitochondrial oxidative capacity. For instance, the location of biopsy sampling affects fiber type distribution by 2–3%, and since SDH activity is closely related to fiber type (5) this could also affect weighed SDH activity. Nonetheless, we showed that SDH activity from quantitative histochemistry is highly reproducible (Fig. 2). We could not account for individual variation in muscle temperature, which in healthy controls is likely to be 1–1.5% (60, 62) and ∼3.5% in CHF patients (65). Even though skeletal muscle mass was estimated from an anthropometric model (47), the estimated skeletal mass as percentage of body mass (34.7% in CHF patients, 37.5% in controls, and 42.3% in cyclists) was in agreement with skeletal muscle mass derived from MRI measurements in subgroups of 468 subjects with similar age (33.8, 39.1, and 42.3%, respectively; see Ref. 36). Errors in oxygen consumption by the brain and heart (e.g., due to differences in heart mass and mitochondrial volume density between CHF patients, controls, and cyclists) may have attributed to the unexplained variance between measured V̇o_{2 max} and mitochondrial oxidative capacity. However, note that maximal oxygen consumption of the heart (∼3 ml·kg⁻¹·min⁻¹) and oxygen consumption of the brain (∼1 ml·kg⁻¹·min⁻¹) only marginally contributed to the mitochondrial oxidative capacity in control subjects and cyclists. Moreover, we could not account for individual variation in SDH activity differences between arm and leg musculature and submaximally active mitochondria in arm musculature during leg cycling exercise. V̇o_{2 max} has shown to be obtained reproducibly within our laboratory (intraclass correlation coefficient = 0.98, P < 0.001; unpublished observations), and all subjects terminated exercise due to voluntary exhaustion. Although V̇o_{2 peak} attained during the maximum-effort incremental test to voluntary exhaustion is considered a valid index of V̇o_{2 max} (15) that is not different from V̇o_{2 peak} attained during supramaximal exercise (2), small differences in maximal effort may have contributed to individual differences in mitochondrial oxidative overcapacity because underestimates of V̇o_{2 max} may have occurred in some subjects. Taken together, unexplained variance between mitochondrial oxidative capacity and V̇o_{2 max} is only 11% (r² = 0.89) in the present study, and therefore it is likely that possible effects of methodological errors are partially cancelled out in our group of human subjects.

A portion of unexplained variance between mitochondrial capacity and V̇o_{2 max} may also be attributed to physiological differences between individuals. These differences may arise from limitations to the convective O₂ supply, for instance, arterial and venous saturation, systemic hematocrit, hemoglobin content, cardiac output, and end-diastolic heart volume (3, 48, 72). Also, limitations to diffusive O₂ supply may explain individual differences in mitochondrial oxidative overcapacity at maximal exercise. Oxygen diffusion at the muscle level is determined by oxygen transport from sarcolemma to mitochondria by myoglobin and by blood-myocyte O₂ flux set by the oxygen tension gradient between the capillaries and sarcolemma, which depends on (changes in) RBC velocity and capillary hematocrit (i.e., RBC flux), on capillary density, and the proportion of flowing capillaries (3, 39, 52, 72). Moreover, differences in mitochondrial oxidative overcapacity may be investigated in light of individual fiber type distribution, since vasomotor control may be modulated as a function of fiber type and/or oxidative capacity of the muscle fibers (29, 41), displaying better Q̇o₂/V̇o₂ matching and thus higher microvascular Po₂ in high vs. low oxidative muscle or muscle parts (29). Furthermore, individual differences may arise from the effectiveness in energy distribution within the muscle cells, which depends on metabolite-facilitated diffusion (i.e., oxygen transport by the oxy-deoxy myoglobin shuttle and ATP diffusion by the creatine kinase shuttle) and on membrane potential conduction via the mitochondrial reticulum (25). Because SDH activity using quantitative enzyme histochemistry is determined in individual muscle fibers, it can be related to other fiber-specific variables (e.g., myosin heavy chain type and muscle fiber size, capillary density, diffusion distance, or myoglobin concentration). Future studies should investigate whether differences in determinants of convective and diffusive O₂ supply or energy distribution may also account for part of the unexplained variance in the relationship between mitochondrial oxidative capacity and V̇o_{2 max}, i.e., explain differences between individual subjects with similar V̇o_{2 max} but different SDH activity.

Use of mitochondrial oxidative overcapacity.

Even though oxygen use during maximal cycling exercise is on average ∼90% of the oxidative capacity, a greater proportion of the mitochondrial oxidative capacity may be used when oxygen supply is not limiting. For instance, during exercise with smaller muscle groups (without cardiac output limitation), blood flow heterogeneity may facilitate matching of oxygen supply to oxygen demand in the active muscle groups, although this does not come at the expense of Q̇o₂/V̇o₂ matching in another region (29, 41). Also at submaximal exercise intensities, a higher mitochondrial oxidative capacity may enable mitochondria to operate at a lower percentage of their maximal oxidative rate for a given oxygen consumption. Consequently, following Michealis Menten kinetics of mitochondria, V̇o₂ kinetics will be faster and result in glycogen sparing and faster adaptations of steady state during submaximal exercise. In addition, lower substrate concentrations are required, and mitochondria are less likely to become hypoxic, thereby reducing damage from oxidative stress induced by hypoxia. Moreover, hypoxia in the muscle fibers will likely occur at higher exercise intensity and therefore could increase intensity at the lactate or ventilatory threshold (i.e., intensity at which aerobic ATP resynthesis can no longer match ATP use in the working muscles), which is an important determinant of endurance performance in subjects with similar V̇o_{2 max} (3). Therefore, with higher mitochondrial oxidative capacity a higher rate of ATP production may be sustained during endurance performance (32). Because mitochondria conform Michaelis Menten kinetics, it is highly unlikely that mitochondria operate at 100% of their maximal oxidative capacity, since this requires very high intracellular Po₂ (that would likely require an increase in the P₅₀ of the blood and enhanced diffusion capacity in the lungs) and extremely high substrate concentrations (that potentially could have an osmotic effect similar to lactate accumulation and thereby increase blood viscosity).

Conclusions.

Human V̇o_{2 max} attained during cycling exercise is related to mitochondrial oxidative capacity predicted from skeletal muscle SDH activity. Measured whole body V̇o_{2 max} is ∼90% of the mitochondrial oxidative capacity, which can be explained by limited oxygen supply to muscle mitochondria.