Relationship between V̇o_2peak, cycle economy, and mitochondrial respiration in untrained/trained

Abstract

Aerobic capacity is negatively related to locomotion economy. The purpose of this paper is to determine what effect aerobic exercise training has on the relationship between net cycling oxygen uptake (inverse of economy) and aerobic capacity [peak oxygen uptake (V̇o_2peak)], as well as what role mitochondrial coupled and uncoupled respiration may play in whole body aerobic capacity and cycling economy. Cycling net oxygen uptake and V̇o_2peak were evaluated on 31 premenopausal women before exercise training (baseline) and after 8–16 wk of aerobic training. Muscle tissue was collected from 15 subjects at baseline and post-training. Mitochondrial respiration assays were performed using high-resolution respirometry. Pre- (r = 0.46, P < 0.01) and postexercise training (r = 0.62, P < 0.01) V̇o_2peak and cycling net oxygen uptake were related. In addition, uncoupled and coupled fat respiration were related both at baseline (r = 0.62, P < 0.01) and post-training (r = 0.89, P < 01). Post-training coupled (r = 0.74, P < 0.01) and uncoupled carbohydrate respiration (r = 0.52, P < 05) were related to cycle net oxygen uptake. In addition, correlations between V̇o_2peak and cycle net oxygen uptake persist both at baseline and after training, even after adjusting for submaximal cycle respiratory quotient (an index of fat oxidation). These results suggest that the negative relationship between locomotion economy and aerobic capacity is increased following exercise training. In addition, it is proposed that at least one of the primary factors influencing this relationship has its foundation within the mitochondria. Strong relationships between coupled and uncoupled respiration appear to be contributing factors for this relationship.

NEW & NOTEWORTHY The negative relationship between cycle economy and aerobic capacity is increased following exercise training. The strong relationship between coupled and uncoupled respiration, especially after training, appears to be contributing to this negative relationship between aerobic capacity and cycling economy, suggesting that mitochondrial economy is not increased following aerobic exercise training. These results are suggestive that training programs designed to improve locomotion economy should focus on changing biomechanics.

INTRODUCTION

Exercise economy plays a role in contributing to ease of locomotion, whereas strength and ease of locomotion relate to reduced weight gain and increased free-living activity-related energy expenditure (11, 16, 20, 34, 35). Although, both aerobic capacity [peak oxygen uptake (V̇o_2peak)] (22), and exercise economy (6, 13, 24) often increase following exercise training, several investigators have shown that aerobic capacity is negatively related to economy while walking (2, 10), running (25, 29), and cycling (21). We have shown that increased V̇o_2peak following weight loss is associated with decreased economy, i.e., increased net oxygen uptake (V̇o₂) while walking (walking V̇o₂ minus resting V̇o₂) (2). In addition, Shaw et al. (32) showed that an increase in V̇o_2peak following training in middle-distance runners was positively associated with an increase in the energy cost of running (inverse of economy). Lucia et al. (21) reported a negative relationship between cycling economy and aerobic capacity. Since aerobic training may be expected to induce improvements in aerobic capacity as well as exercise economy, it is important to know if this negative relationship exists after exercise training. To our knowledge, no one has reported this relationship for cycling in an untrained state and following aerobic exercise training.

Using ³¹phosphate magnetic resonance spectroscopy, we have shown that oxidative capacity of the muscle as well as whole body maximal oxygen uptake (V̇o_2max) and plantar flexion muscle economy are negatively related (10), suggesting that the negative relationship between oxidative capacity and exercise economy may occur at the muscle tissue level. It would be of interest to know if this relationship exists at the muscle mitochondrial level, i.e., maximal mitochondrial coupled and uncoupled respiration, and determine if exercise training affects these relationships.

High-resolution respirometry experiments can be used to separate mitochondrial coupled respiration (state 3) and uncoupled respiration (state 4) as well as the respiratory control ratio (RCR; state 3/state 4 respiration, a measure of relative efficiency of respiration) while using either carbohydrate or fatty acids as fuel. Understanding the interrelationships between whole body maximal oxygen uptake, exercise economy, and in situ coupled and uncoupled mitochondrial respiration for both carbohydrate and fatty acid substrates may give us more understanding of the relationship between whole body aerobic capacity and exercise economy.

Therefore, the purposes of this paper are to examine the interrelationships between state 3 respiration, state 4 respiration, RCR, cycling net V̇o₂, and V̇o_2peak, both in an untrained state (baseline) and following 8–16 wk of aerobic training (trained).

METHODS

Study participants.

This is a secondary analysis of a study designed to evaluate insulin sensitivity, resting energy expenditure, and blood pressure following a bout of moderate-intensity or high-intensity exercise as compared with no exercise (no exercise for 72 h before evaluation) (3, 17). Since the 3 postexercise training conditions (no exercise within 72 h, following moderate-intensity exercise, and following high-intensity exercise) were randomly assigned to avoid an order effect in analysis, the post-training evaluation following 72 h of no exercise used in this study varied from 8 wk to 16 wk. No differences in V̇o_2peak or submaximal V̇o₂ were observed between 8-, 12-, or 16-wk evaluations. Only the post-training evaluation corresponding to the no acute exercise session is reported. V̇o_2peak and state 3 and 4 fat respiration were previously reported (16). Women (n = 52) between 20 and 40 yr of age participated in this study. A subset of 25 subjects volunteered for the muscle biopsy. Participants reported normal menstrual cycles and were not taking oral contraceptives or any medications known to influence glucose and/or lipid metabolism. Additional inclusion criteria included: 1) normotensive; 2) nonsmoker; 3) sedentary as defined by participating in any exercise-related activities less than once per week; and 4) normoglycemic as evaluated by postprandial glucose response to a 75-g oral glucose tolerance test. All participants provided written informed consent. Study procedures were approved by the Institutional Review Board at the University of Alabama at Birmingham and conformed to the guidelines set forth by the Declaration of Helsinki.

Procedures.

After initial screening and fitness assessments, all participants were evaluated two times during the follicular phase of two different menstrual cycles. Participants stayed in a room calorimeter for the 23 h before testing. The first evaluation was considered baseline. Post-training evaluations took place after 8, 12, or 16 wk of exercise training. Training duration varied because of the experimental design of the parent study in which there was a random assignment of test order over 8–16 wk of exercise training. Only the results of the baseline tests (before initiation of training) and the post-training condition in which the subjects participated in no exercise for 72 h before evaluation are included. The other two conditions, either a bout of moderate-intensity or a bout of high-intensity cycle exercise 22 h before evaluation are not included in this paper. There were no differences between the 3 post-training conditions for V̇o_2peak, suggesting that aerobic fitness was similar for the 8-, 12-, and 16-wk evaluations.

Food intake.

Food was provided the day before room calorimeter visits and during the days spent in the room calorimeter. Diets were prepared by the Clinical Research Unit kitchen staff and consisted of ≈60% of energy as carbohydrates, ≈25% as fat, and ≈15% as protein. Macronutrient content of the diet was held constant.

Energy balance clamp.

One goal for having provided food was to try to achieve energy balance (energy intake matching energy expenditure) during the stay in the room calorimeter. Caloric intake during the room calorimetry visit was based on estimates generated from 330 double-labeled water estimates of free-living energy expenditure of sedentary premenopausal women collected in our laboratory. The equation was as follows: equation 1 = 750 kcal + [(31.47 × fat free mass (FFM)) − (0.31 × fat mass) − (155 × race (race coded 1 for African Americans and 0 for European Americans)]. An equation for estimating the room calorimeter energy intake was developed from over 200 room calorimeter visits of premenopausal women: equation 2 = 465 kcal + [(27.8 × fat free mass (FFM)) − (2.4 × fat mass) − (188 × race) (race coded 1 for African Americans and 0 for European Americans)]. However, we recognized that the estimates may result in overfeeding or underfeeding individual subjects. Therefore, we developed a correction equation for the room calorimeter visit that was: based on energy expenditure during the room calorimeter stay up to 5:30 PM. This equation was equation 3 = (390 kcal + average energy expenditure in kcal/min between 8:00 AM and 5:30 PM) × 925 kcal) − equation 3 estimate of energy expenditure. We then adjusted the food intake of the evening meal to match the results of equation 3. The energy balance clamp was deemed effective since there were no significant differences in energy balance between the 2 conditions (mean values varying between −19.8 ± 207 and −21.7 ± 182) or significant difference from 0 for the 2 conditions.

Exercise training.

After baseline testing, all participants aerobically exercise trained 3 times/wk on a cycle ergometer for the duration of the 16-wk study. Post-training evaluation in which subjects did no exercise for 72 h before evaluation varied from 8 to 16 wk. Participants trained initially for 20 min at 65% of maximum heart rate and progressively increased their training until they were training continuously for 40 min at 80% of maximum heart rate by week 4. Heart rate was monitored throughout each session by a Polar Vantage XL heart rate monitor (Polar Beat, Port Washington, NY). All sessions were under the supervision of an exercise physiologist in a training facility dedicated to research.

Peak aerobic capacity and submaximal steady-state cycling V̇o₂.

Two to four days before each room calorimeter visit, peak aerobic capacity and submaximal, steady-state cycling V̇o₂ were measured. After an initial warm-up, V̇o₂, respiratory exchange ratio (RER), and heart rate were evaluated during a 3-min submaximal, steady-state cycling task at 50 W. No significant differences were observed between the last 20 s and second-to-last 20 s of the third minute for RER (0.89 ± 0.08 vs. 0.89 ± 0.09). Since subjects demonstrated steady-state during the last minute of the submaximal cycling task, RER is considered respiratory quotient (RQ) under steady-state conditions and gives an index of fat oxidation during the submaximal task. Resting V̇o₂ was subtracted from submaximal cycling V̇o₂ to obtain submaximal net V̇o₂ (see Room calorimeter section for resting oxygen uptake methods). Following the submaximal task, participants completed a graded cycle ergometer test to measure peak oxygen uptake (V̇o_2peak) as determined by the highest oxygen uptake reached in the final stage of exercise. Starting at 70 W, every 2 min, power was raised 20 W until participants reached volitional exhaustion. Sixty-revolutions-per-minute cycle cadence was maintained throughout the test. Oxygen uptake, ventilation, and respiratory exchange ratio were determined by indirect calorimetry using a MAX-II metabolic cart (Physio-Dyne Instrument Company, Quogue, NY). Heart rate was continuously monitored by Polar Vantage XL heart rate monitors (Polar Beat, Port Washington, NY). Although we do not claim a true maximum oxygen uptake since tests were done on a cycle ergometer rather than a treadmill, criteria for achieving a true maximum were heart rate within 10 beats of estimated maximum, RER of at least 1.1, and plateauing of V̇o₂. All subjects reached at least one criterion, and all but three subjects reached at least two criteria at each of the four test time points.

Room calorimeter.

Participants spent 23 h in a whole-room respiration calorimeter (3.38 m long, 2.11 m wide, and 2.58 m high) for measurement of total energy expenditure and REE before and after the exercise training period. The design characteristics and calibration of the calorimeter were described previously (33). Oxygen consumption and carbon dioxide production were continuously measured with the use of a magnetopneumatic differential oxygen analyzer (Magnos206; ABB, Frankfurt, Germany) and a nondispersive infrared industrial photometer differential carbon dioxide analyzer (Uras26, ABB). The calorimeter was calibrated before each participant entered the chamber. Prior to each test, calibration was carried out on the oxygen and carbon dioxide analyzers using standard gases. The full scale was set for 0%–1% for the carbon dioxide analyzer and 0%–2% for the oxygen analyzer. Each participant entered the calorimeter at 8:00 AM. Metabolic data were collected throughout the 23-h stay. Each participant was awakened at 6:30 AM the next morning in the calorimeter. Resting oxygen uptake was then measured for 30 min before the subject left the calorimeter at 7:00 AM. Energy expenditure was calculated by the Weir equation (5).

Body composition.

Total and regional body composition (i.e., fat mass and lean mass) was determined by dual-energy X-ray absorptiometry (iDXA, GE-Lunar, Madison, WI) both before exercise training and post-training. Participants were evaluated 3–4 days before the room calorimeter visit following an overnight fast. Participants wore light clothing and remained supine with arms at their side but not touching the body in compliance with manufacturer-recommended testing procedures. Scans were analyzed by the same investigator with ADULT software, LUNAR-DPX-L version 1.33 (GE Medical Systems Lunar).

Tissue biopsy and preparation of permeabilized muscle fibers.

Following no exercise for 72 h, both at baseline and following 8–16-wk exercise training, a subset of 25 women underwent muscle biopsies (70–140 mg) of the vastus lateralis (only 15 also had muscle biopsies post-training). Muscle biopsy samples were obtained from the lateral side of the vastus lateralis under local subcutaneous anesthesia (1% lidocaine) by percutaneous needle biopsy using a 5-mm Bergstrom needle under suction, as previously described (1). The second biopsy was performed on the thigh contralateral to the prior biopsy A portion of the biopsy sample was immediately placed and transported in an ice-cold relaxing and preservation solution BIOPS, containing (in mM) 2.77 Ca-EGTA buffer, 0.0001 free calcium, 50 K-MES, 7.23 K₂EGTA, 20 imidazole, 0.5 DTT, 20 taurine, 5.7 ATP, 14.3 PCr, and 6.56 MgCl₂-6 H2O (pH 7.1, 290 mOsm) (18, 30) and was used to prepare permeabilized muscle fiber bundles (PmFB). Briefly, small pieces of skeletal muscle (~20–25 mg) were placed immediately in fresh ice-cold BIOPS, trimmed of fat and connective tissue on ice, and separated into 4 small muscle bundles (~2–6 mg wet wt). The PmFBs were mechanically separated by gentle blunt dissection with a pair of needle-tipped, antimagnetic forceps under magnification (Zeiss, Stemi S2000-C Stereo Microscope, Diagnostic Instruments). They were then treated with 30 μg/mL saponin and gently rocked (Rocker II, model 260350, Boekel Scientific) at 4°C for 30 min in BIOPS. PmFBs were then rinsed twice by gentle rocking to wash out saponin and ATP at 4°C for at least 15 min and <30 min in MiR05 containing (in mM) 105 K-MES, 30 KCl, 1 EGTA, 10 K₂HPO₄, and 5 MgCl₂-6 H2O, with 0.5 mg/mL BSA (pH 7.1, 290 mOsm). The PmFBs were then transferred to a fresh MiR05/creatine solution (500 μL) and blebbistatin (BLEB, 25 μM).

High-resolution mitochondrial respirometry in permeabilized fibers.

Mitochondrial respiration assays were performed using high-resolution respirometry by measuring oxygen consumption in 2 mL of MiR05/creatine/blebbistatin buffer in a 2-channel respirometer (Oroboros Oxygraph-2k with DatLab software; Oroboros Instruments Corp., Innsbruck, Austria) with constant stirring at 750 revolutions/min (19) and following a modified substrate-uncoupler-inhibitor titration protocol to evaluate respiratory control in a sequence of coupling and inhibitory states induced by multiple titrations in each assay (31). Ethanol (70%) was run in both chambers for a minimum of 30 min, rinsed 3 times with Milli-Q ultrapure ddH₂O, and the chambers calibrated after a stable air-saturated signal was obtained before every experiment. Reactions were conducted with PmFB (2–8 mg wet wt) at 37°C with hyperoxygenation to maintain oxygen concentrations above air saturation (~500–200 µM) (9) and prevent oxygen diffusion restrictions, which have been shown to limit oxygen supply to the core of the fiber bundle (7). All experiments were completed before the oxygraph chamber [O₂] reached 150 µM. Respiration rates were measured using 2 substrate protocols (4 mM malate, 9 mM pyruvate, and 2.5 mM succinate) to drive convergent electron input to complexes I and II of the Electron Transport System (ETS) or malate (2 mM; anapleurotic intermediate) and palmitoylcarnitine (40 μM) to determine mitochondrial β-oxidation per se with all electron transport chain complexes, independent of the step catalyzed by carnitine palmitoyltransferase I, which may differ across individuals and races (23). Polarographic oxygen measurements are expressed as pmol·s⁻¹·mg⁻¹ wet wt. Determination of state 2, 3, and 4 respiration rates were determined in the presence of substrate alone (state 2; LEAK state; low ATP), after the addition of ADP (2 mM; state 3; OxPhos state), and after inhibition of ATP synthase (Complex V) with oligomycin (state 4; LEAK state; high ATP). For quality control and to ensure outer mitochondrial membrane integrity, cytochrome c (10 μM) was added to the assay after activation by ADP and only preparations with <10% increase after addition were included. Respiratory control ratios (RCRs) were determined as the ratio of state 3/state 4 respiration rates.

Statistics.

Sample size for this study was based on detecting differences for insulin sensitivity and resting energy expenditure following a bout of high-intensity exercise. Fifty-two subjects participated in the study, but only twenty-five subjects volunteered for the muscle biopsies. In addition, there were missing data at different time points because of experimenter error or equipment malfunction. Finally, one data point for the baseline state 4 carbohydrate respiration that was over 4 standard deviations above the mean and over 2 standard deviations above the next highest value was removed as an outlier. Therefore, sample size for various variables and variable pairs for correlations are different. Correlations for the baseline time point were calculated using only the subjects for which post-training values were obtained, so the same subjects were used in both the pre- and post-training correlation analysis. Because no study has shown a decrease in state 3 respiration (coupled respiration) following exercise training and most studies have shown an increase in state 3 respiration, we used a one-tailed test of significance for state 3 respiration. Paired t tests between baseline and post-training were run for all variables of interest. Simple Pearson product correlations were run between variables of interest at baseline and post-training. Finally, linear regression was used to determine the independent relationships between post-training submaximal cycling net V̇o₂ and V̇o_2peak and RQ.

RESULTS

Table 1 shows the results of the pre- to post-training changes in study variables. No significant changes were found for any study variable except for the increase in V̇o_2peak. Table 2 shows the pretraining correlations between whole body V̇o_2peak, net submaximal cycling V̇o₂, and the mitochondrial respiration variables. Net submaximal cycling V̇o₂ was significantly related to submaximal cycle RQ (0.46) and V̇o_2peak (r = 0.52). State 3 carbohydrate respiration was significantly related to carbohydrate RCR (measure of mitochondrial carbohydrate coupling efficiency, r = 0.69). Carbohydrate RCR and state 4 respiration (LEAK) were not related, whereas state 3 fat respiration was significantly related to state 4 carbohydrate respiration (uncoupled carbohydrate respiration, r = 0.60) and the RCR for fat respiration (fat oxidation mitochondrial coupling efficiency, r = 0.56). No other significant correlations were observed at baseline.

Table 1.

Pre-post–training variables of interest

	Baseline	Post-Training	P
Body weight, kg (n = 31)	74.1 ± 14.6	74.2 ± 14.7	0.73
Percent fat (n = 31)	37.9 ± 6.8	37.5 ± 6.7	0.18
Cycle submaximal V̇o2, mL O2·kg−1·min−1 (n = 31)	11.9 ± 2.3	12.0 ± 2.7	0.66
Cycle submaximal net V̇o2, mL·kg−1·min−1 (n = 31)	9.1 ± 2.2	9.2 ± 2.5	0.79
Cycle submaximal RQ (n = 31)	0.88 ± 0.08	0.87 ± 0.07	0.28
V̇o2peak, mL O2·kg−1·min−1 (n = 31)	25.3 ± 6.1	27.0 ± 6.0	<0.01
State 3 carbohydrate respiration, pmol·s−1·mg wet wt−1 (n = 15)	19.0 ± 6.4	21.5 ± 12.5	0.30
State 4 carbohydrate respiration, pmol·s−1·mg wet wt−1 (n = 15)	8.5 ± 3.3	10.1 ± 4.9	0.34
RCR carbohydrate (n = 15)	2.4 ± 1.0	2.1 ± 0.6	0.37
State 3 fat respiration, pmol·s−1·mg wet wt−1 (n = 15)	8.8 ± 4.6	11.3 ± 4.9	0.04
State 4 fat respiration, pmol·s−1·mg wet wt−1 (n = 14)	3.9 ± 2.7	4.5 ± 2.4	0.24
RCR fat (n = 15)	2.7 ± 0.6	2.6 ± 0.5	0.62

Dara are presented as means ± SD. n = no. of subjects. RCR, respiratory control ratio; RQ, respiratory quotient; V̇o₂, oxygen uptake.

Table 2.

Intercorrelations at baseline

	Submaximal Cycle, Net V̇o2	Submaximal Cycle, RQ	V̇o2peak	State 3 Carbohydrate Respiration	State 4 Carbohydrate Respiration	RCR Carbohydrate	State 3 Fat Respiration	State 4 Fat Respiration	RCR Fat
Submaximal cycle, net V̇o2		0.46†	0.52†	−0.08	−0.48	0.32	−0.20	0.06	−0.09
Submaximal cycle, RQ			−0.23	0.30	−0.21	0.45	−0.15	−0.16	−0.31
V̇o2peak				−0.17	−0.19	−0.13	−0.17	0.28	−0.16
State 3 carbohydrate respiration					0.36	0.69†	0.26	0.62*	−0.40
State 4 carbohydrate respiration						−0.39	0.60*	0.30	0.05
RCR carbohydrate							−0.13	0.33	−0.28
State 3 fat respiration								0.39	0.56*
State 4 fat respiration									−0.26

n = 31 Subjects for correlations between V̇o_2peak and submaximal cycle net V̇o₂ and n = 15 subjects for all other correlations, except submaximal cycle net V̇o₂ and state 4 carbohydrate respiration for which n = 14 subjects. RCR, respiratory control ratio; RQ, respiratory quotient; V̇o₂, oxygen uptake; V̇o_2peak, peak oxygen uptake.

^*Significance at P < 0.05;

^†significance at P < 0.01.

Table 3 shows the post-training intercorrelations. Net submaximal cycling V̇o₂ was significantly related to V̇o_2peak (r = 0.62), state 3 carbohydrate respiration (r = 0.74), and state 4 carbohydrate respiration (r = 0.52). State 3 carbohydrate respiration was significantly related to state 4 carbohydrate respiration (r = 0.79) and V̇o_2peak (r = 0.50), whereas state 3 fat respiration was significantly related to state 4 fat respiration (r = 0.89). No other significant correlations were observed post-training.

Table 3.

Intercorrelations post-training

	Submaximal Cycle, net V̇o2	Submaximal Cycle, RQ	V̇o2peak	State 3 Carbohydrate Respiration	State 4 Carbohydrate Respiration	RCR Carbohydrate	State 3 Fat Respiration	State 4 Fat Respiration	RCR Fat
Submaximal cycle, net V̇o2		0.32	0.62†	0.74†	0.52*	0.44	0.37	0.19	0.21
Submaximal cycle, RQ			−0.05	−0.11	0.00	0.00	−0.15	−0.13	−0.04
V̇o2peak				0.50*	0.42	0.10	−0.16	−0.25	0.18
State 3 carbohydrate respiration					0.79†	0.50	0.25	0.16	−0.07
State 4 carbohydrate respiration						−0.06	0.06	0.09	−0.29
RCR carbohydrate							0.22	0.06	0.27
State 3 fat respiration								0.89†	−0.04
State 4 fat respiration									−0.45

n = 31 Subjects for correlation between V̇o_2peeak and submaximal cycle net V̇o₂ and n = 15 subjects for all other parameters, except those that involved state 4 carbohydrate respiration for which n = 14 subjects. RCR, respiratory control ratio; RQ, respiratory quotient; V̇o₂, oxygen uptake; V̇o_2peak, peak oxygen uptake.

^*Significance at P < 0.05;

^†significance at P < 0.01.

Correlations were run between the whole body measures (V̇o_2peak, submaximal cycling V̇o₂, and submaximal cycling RQ) with only the subjects who had muscle biopsies. No significant correlations were obtained; however, the correlation between V̇o_2peak and submaximal cycling V̇o₂ approached significance (r = 0.45 baseline and r = 0.44 post-training, both P < 0.10) and were of similar magnitude as the full sample size correlations, suggesting that the reason for lack of significance was due to the reduced sample size.

Change in V̇o_2peak or change in net submaximal cycling V̇o₂ were not related to each other or to changes to state 3 or state 4 carbohydrate or fat respiration. However, change in state 3 and change in state 4 carbohydrate respiration (r = 0.79, P < 0.01) as well as change in state 3 and change in state 4 fat respiration (r = 0.65, P < 0.02) were significantly related.

Since fat oxidation requires more oxygen to produce ATP than carbohydrate oxidation, it is possible the correlation between submaximal cycling net V̇o₂ and V̇o_2peak was caused by larger fat oxidation for those individuals with low submaximal net V̇o₂. Therefore, a multiple regression analysis was run with post-training submaximal cycling RQ (an index of relative fat metabolism) and V̇o_2peak as independent variables and net submaximal cycling V̇o₂ as the dependent variable (Table 4). Both submaximal cycling RQ (partial R = 0.45) and V̇o_2peak (partial R = 0.67) were independently related to submaximal cycling net V̇o₂, showing that the relationship between cycling net V̇o₂ and V̇o_2max are not mediated by increased fat oxidation in the post-training evaluation. A similar result was obtained at baseline when observing the relationship between cycling net V̇o₂ and V̇o_2max after adjusting for RQ (partial R between cycling net V̇o₂ and V̇o_2max = 0.58, P < 0.01).

Table 4.

Multiple regression for submaximal bike net V̇o₂ after adjusting for either V̇o_2peak or submaximal bike RQ

R2		Intercept/Slope	Partial R	P
0.51	Model 1: submaximal cycle, net V̇o2 (n = 35)	−9.5		<0.01
	V̇o2peak	0.57	0.67	<0.01
	Submaximal cycle RQ	12.1	0.45	<0.05

n = No. of subjects. RQ, respiratory quotient; V̇o₂, oxygen uptake; V̇o_2peak, peak oxygen uptake.

DISCUSSION

The purpose of this paper was twofold: 1) determine what effects cycle ergometer exercise training has on the positive relationship between V̇o_2peak and net submaximal cycling V̇o₂ (inverse of bike economy) and 2) determine what role coupled and uncoupled mitochondrial respiration might contribute to this relationship. V̇o_2peak was significantly related to submaximal cycling net V̇o₂ both before training and in the trained state (Tables 2 and 3 and Fig. 1). In fact, the relationship did not become weaker with training, with a correlation 0.52 at baseline and 0.62 following training. In addition, no respiration variable was related to submaximal cycling net V̇o₂ before exercise training. However, following training, state 3 (Fig. 2) and state 4 carbohydrate respiration (Table 3) were significantly related to submaximal cycling net V̇o₂, with carbohydrate RCR showing a strong trend (r = 0.44) toward a significant relationship. Taken together, these results suggest that exercise training may have strengthened the relationship between submaximal cycling net V̇o₂ and aerobic capacity, and this relationship may be caused by factors occurring in the mitochondria, i.e., a strong relationship between uncoupled and coupled carbohydrate respiration. Repeated bouts of exercise training induce higher rates of, and demand for, oxidative phosphorylation, which increased reactive oxygen species (ROS) formation, including superoxide formation (26). With the potential for increased ROS formation, greater control of the coupling and uncoupling of oxidative phosphorylation would facilitate and allow increased uncoupling as a protective mechanism to counter increased mitochondrial ROS. Uncoupling shunts protons across the inner mitochondrial membrane, thus decreasing membrane potential and ROS. This comes with an increased oxygen consumption rate (8). Supporting this is the doubling of the relationship between coupled and uncoupled carbohydrate respiration (state 3 and state 4, respectively) from baseline (r = 0.38) to the trained state (r = 0.80, Table 3). These data are supportive of the hypothesis that high carbohydrate uncoupling in the mitochondria leads to less economical work on the cycle ergometer, whereas high uncoupling is associated with greater potential for mitochondrial oxidative capacity and possible improved control of ROS formation.

Fig. 1.

Post-training relationship between submaximal bike net oxygen uptake (V̇o₂) and peak oxygen uptake (V̇o_2peak) (r = 0.62, P < 0.01).

Fig. 2.

Post-training relationship between submaximal bike net oxygen uptake (V̇o₂) and state 3 carbohydrate respiration (r = 0.74, P < 0.01).

No fat respiration variables were significantly related to submaximal cycling net V̇o₂ either at baseline or following training, and thus these data suggest that fat respiration plays little role in the relationship between V̇o_2peak and submaximal cycling net V̇o₂. Although, it should be pointed out that state 4 fat respiration was strongly related to state 3 fat respiration, both at baseline and in the trained state and state 3 fat respiration, and was significantly related to fat RCR at baseline. Interestingly, state 3 fat respiration lost its significant relationship with coupling efficiency (RCR) in the trained state, suggesting that after training individuals who have high capacity for generating ATP from fat oxidation also have high uncoupling when using fat as a substrate.

We have previously shown that high exercise economy is related to increased endurance, which in turn is related to reduced subsequent weight gain (11, 16, 20, 35). Those individuals that move more economically also experience more ease in locomotion and are more likely to be more physically active (4, 12, 35) and less prone to weight gain (35). It would therefore be important to understand factors that may influence exercise economy. Muscular strength associates positively with walking economy (14), whereas strength training induced increases in strength associated with reduced muscle activation (15), increased economy (13, 28), increased free-living activity-related energy expenditure (14, 16, 27), and reduced weight gain (35). This is in contrast to the negative relationship between exercise economy and aerobic capacity shown in this study as well as by others (2, 10, 21, 25, 29, 32).

We expected to observe a significant increase in state 3 respiration following the exercise training. However, no significant pre-post increase for state 3 carbohydrate was observed, although mean values for all respiration state variables increased over 13% and state 3 respiration rates with fat substrates were significantly increased (P = 0.04). We obtained the post-training muscle samples 72 h following the last exercise bout, so it is possible that the mitochondrial respiration values may have been elevated more with a shorter time interval from the last training session.

Based on the results of the present study, it is proposed that the basis of the inverse relationship between pedaling economy and aerobic capacity has its foundation at the mitochondrial level and that increased capacity for generating ATP from oxidative processes is accompanied with an equal or greater increase in uncoupled respiration. On the other hand, increases in strength and muscle size are associated with a reduced dependence on inefficient fast-twitch muscle fibers, allowing the more efficient slow-twitch muscle fibers to do a larger percentage of the total work (12), thus improving overall exercise economy.

Because of the design limitations of the parent study for this secondary analysis, the post-training evaluations were conducted between 8 and 16 wk following commencement of training. No differences were observed between either V̇o_2peak or submaximal cycle V̇o₂ between any of the post-training evaluations (8, 12, or 16 wk), suggesting that aerobic fitness changed little after the first 8 wk of training. We therefore feel confident that the relationships between aerobic capacity, mitochondrial respiration, and cycle economy were not affected by the variation in training time.

In conclusion, these results suggest that the inverse relationship between pedaling economy and aerobic capacity remains following exercise training. In addition, it is proposed that at least one of the primary factors influencing this relationship has its foundation within the mitochondria and may be a protective mechanism against ROS formation. Strong relationships between coupled and uncoupled respiration appear to be contributing factors for explaining the relationship between whole body aerobic capacity and pedaling economy.

GRANTS

This work was supported by NIH Grants R01-AG-027084-01, R01-AG-27084-S, P30-DK-56336, P60-DK-079626, UL-1RR-025777, and R01-DK-49779.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

G.R.H. conceived and designed research; G.R.H., D.R.M., S.T.W., M.M.B., and G.F. performed experiments; G.R.H., D.R.M., M.M.B., and G.F. analyzed data; G.R.H., D.R.M., S.L.M., M.M.B., and G.F. interpreted results of experiments; G.F. prepared figures; G.R.H. and G.F. drafted manuscript; G.R.H., D.R.M., S.T.W., S.L.M., M.M.B., and G.F. edited and revised manuscript; G.R.H., D.R.M., S.T.W., S.L.M., M.M.B., and G.F. approved final version of manuscript.