Am J Physiol Endocrinol Metab 307: E539–E552, 2014.

First published August 5, 2014; doi:10.1152/ajpendo.00276.2014.

Metabolic and hormonal responses to isoenergetic high-intensity interval

exercise and continuous moderate-intensity exercise
Jonathan M. Peake,1,2 Sok Joo Tan,3 James F. Markworth,4 James A. Broadbent,1 Tina L. Skinner,3
and David Cameron-Smith4
1
School of Biomedical Sciences and Institute of Health and Biomedical Innovation, Queensland University of Technology,
Brisbane, Australia; 2Centre of Excellence for Applied Sport Science Research, Queensland Academy of Sport, Brisbane,
Australia; 3School of Human Movement Studies, The University of Queensland, Brisbane, Australia; and 4Liggins Institute,
University of Auckland, Auckland, New Zealand
Submitted 16 June 2014; accepted in final form 30 July 2014

Peake JM, Tan SJ, Markworth JF, Broadbent JA, Skinner TL, out” effort or at intensities close to maximum oxygen con-

Downloaded from http://ajpendo.physiology.org/ by 10.220.33.6 on June 20, 2017

Cameron-Smith D. Metabolic and hormonal responses to isoener- sumption (V̇O2 max) (16). The body of literature supporting the
getic high-intensity interval exercise and continuous moderate-inten- health and fitness benefits of high-intensity interval training
sity exercise. Am J Physiol Endocrinol Metab 307: E539 –E552, 2014. (HIIT) is expanding. So too is our knowledge of the molecular
First published August 5, 2014; doi:10.1152/ajpendo.00276.2014.—
mechanisms that accompany adaptations to this form of train-
This study investigated the effects of high-intensity interval training
(HIIT) vs. work-matched moderate-intensity continuous exercise ing (19).
(MOD) on metabolism and counterregulatory stress hormones. In a Improved metabolism is integral to the benefits of HIIT.
randomized and counterbalanced order, 10 well-trained male cyclists Carbohydrate (CHO) oxidation increases with exercise inten-
and triathletes completed a HIIT session [81.6 ⫾ 3.7% maximum sity, whereas fat oxidation increases during exercise up to
oxygen consumption (V̇O2 max); 72.0 ⫾ 3.2% peak power output; 792 ⫾ 95 65–75% V̇O2 max and decreases at higher workloads (52, 60).
kJ] and a MOD session (66.7 ⫾ 3.5% V̇O2 max; 48.5 ⫾ 3.1% peak Exercise also stimulates amino acid metabolism in skeletal
power output; 797 ⫾ 95 kJ). Blood samples were collected before, muscle and plasma (3, 4), but the effects of exercise intensity
immediately after, and 1 and 2 h postexercise. Carbohydrate oxidation are not well established (55). During recovery from exercise,
was higher (P ⫽ 0.037; 20%), whereas fat oxidation was lower (P ⫽ circulating fatty acids and CHO and fat oxidation remain
0.037; ⫺47%) during HIIT vs. MOD. Immediately after exercise,
elevated for several hours (22). However, the effects of prior
plasma glucose (P ⫽ 0.024; 20%) and lactate (P ⬍ 0.01; 5.4⫻) were
higher in HIIT vs. MOD, whereas total serum free fatty acid concen- exercise intensity and modality (i.e., intermittent vs. continu-
tration was not significantly different (P ⫽ 0.33). Targeted gas ous) on substrate metabolism after exercise are currently un-
chromatography-mass spectromtery metabolomics analysis identified clear (30, 31, 36, 63).
and quantified 49 metabolites in plasma, among which 11 changed As an extension to this early research on intensity-dependent
after both HIIT and MOD, 13 changed only after HIIT, and 5 changed metabolic changes, more recent research has compared molec-
only after MOD. Notable changes included substantial increases in ular adaptations to high-intensity interval exercise and contin-
tricarboxylic acid intermediates and monounsaturated fatty acids after uous, moderate-intensity exercise. Most research in this area
HIIT and marked decreases in amino acids during recovery from both has focused on short intervals of 30 s; however, several other
trials. Plasma adrenocorticotrophic hormone (P ⫽ 0.019), cortisol studies have investigated the effects of longer intervals lasting
(P ⬍ 0.01), and growth hormone (P ⬍ 0.01) were all higher imme-
2– 4 min (2, 21, 57, 65). Markers of mitochondrial biogenesis
diately after HIIT. Plasma norepinephrine (P ⫽ 0.11) and interleu-
kin-6 (P ⫽ 0.20) immediately after exercise were not significantly increase to a similar extent after these forms of exercise, both
different between trials. Plasma insulin decreased during recovery acutely (2, 17) and chronically (7). Few studies have compared
from both HIIT and MOD (P ⬍ 0.01). These data indicate distinct differences in CHO, fat, and amino acid metabolism between
differences in specific metabolites and counterregulatory hormones high-intensity intervals and continuous, moderate-intensity ex-
following HIIT vs. MOD and highlight the value of targeted metabo- ercise (2). Furthermore, other studies comparing the effects of
lomic analysis to provide more detailed insights into the metabolic high-intensity intervals and continuous, moderate-intensity ex-
demands of exercise. ercise on long-term changes in blood lipids and body fat
exercise intensity; metabolites; stress hormones; amino acids; free percentage have reported variable findings (45, 56). Consider-
fatty acids; tricarboxylic acid intermediates ing the potential benefits of HIIT (19), further research to
compare differences in metabolism between this form of train-
ing and more traditional continuous, moderate-intensity exer-
INTEREST IN HIGH-INTENSITY exercise as an effective mode of cise is warranted.
exercise training has intensified over the past decade in recog- Metabolomics is a relatively new analytical platform within
nition of three main factors that are commonly cited as limi- the domain of exercise biochemistry and physiology. Coupled
tations to regular physical activity: lack of time, lack of with the greater availability and ease-of-use of mass-spectrom-
motivation, and chronic diseases that restrict work capacity etry and nuclear magnetic resonance equipment, we now have
during exercise (64). This type of training generally consists of better access to powerful data acquisition and advanced data-
relatively brief, intermittent exercise performed either at “all- processing techniques to identify large numbers of analytes
simultaneously in biological specimens (48). Metabolomics
Address for reprint requests and other correspondence: J. Peake, Institute of
and proteomics were first applied to exercise studies on ani-
Health and Biomedical Innovation, Kelvin Grove, QLD 4059, Australia mals (5) and then humans (51). These early studies helped to
(e-mail: [email protected]). refine these techniques for use in exercise research but did not
http://www.ajpendo.org 0193-1849/14 Copyright © 2014 the American Physiological Society E539
E540 METABOLIC AND HORMONAL RESPONSES TO EXERCISE

yield any detailed information on specific metabolic responses and adding this to the power output in the prior completed stage of the
to exercise (51). Subsequently, metabolomics and proteomics test. The test was repeated at least 48 h later, and the higher V̇O2 max
have been used to determine responses to acute exercise (34, and associated peak power output were used to determine power
42, 43, 48) and in response to exercise training (49, 67). output during the experimental trials.
Familiarization trial. To determine whether the athletes could
However, we are not aware of any studies that have system-
complete the HIIT session, they completed a familiarization trial at
atically employed metabolomics to compare metabolic re- least 48 h following the second V̇O2 max test. This required them to
sponses to different modes of exercise. complete 10 ⫻ 4 min intervals at a power output corresponding to
CHO and fat metabolism during exercise is regulated by ⬃80% V̇O2 max, with a 2-min recovery at 50 W (11.4 ⫾ 0.9% peak
various hormones, including insulin, glucagon, cortisol, adre- power output) between the intervals. The work completed during each
nocorticotrophic hormone (ACTH), growth hormone, epineph- interval was calculated by multiplying the power output for each
rine, and norepinephrine (37). The cytokine interleukin (IL)-6 athlete at 80% V̇O2 max by the duration of the interval (240 s). The
also influences metabolism during exercise in a hormone-like work completed during each recovery period was calculated by
manner (50). Combining metabolomics with a range of coun- multiplying the power output (50 W) by the duration of the recovery
terregulatory hormones provides a more complete perspective period (120 s). These two figures were added together and then
multiplied by 10 to calculate the total amount of work that each athlete
on cross talk between the endocrine system and metabolic

Downloaded from http://ajpendo.physiology.org/ by 10.220.33.6 on June 20, 2017

completed in this familiarization trial. This figure was then used to
organs such as skeletal muscle, adipose tissue, and the liver determine the duration of exercise required to complete the same
during exercise. amount of work while cycling continuously at the power output
The primary aim of this study was to compare the metabolic corresponding to ⬃65% V̇O2 max (moderate-intensity continuous ex-
responses after high-intensity interval and continuous, moder- ercise, MOD). The required duration was calculated by dividing the
ate-intensity cycling matched for total workload. To achieve total amount of work that each athlete completed in this familiariza-
this goal, we employed targeted gas chromatography-mass tion trial by the power output for each athlete at 65% V̇O2 max.
spectrometry (GC-MS)-based metabolomic analysis to assess Exercise trials. At least 1 wk after the familiarization trial the
changes in tricarboxylic acid (TCA) intermediates, fatty acids, athletes completed either the HIIT or MOD trial in a randomized and
and amino acids. A secondary aim was to investigate changes counterbalanced order. These two trials were separated by a minimum
of 7 days. On the morning of each trial, the athletes arrived at the
in counterregulatory hormones that regulate metabolism during
laboratory around 6:30 A.M. Upon arrival, a preexercise venous blood
exercise and are responsive to exercise intensity and duration sample (20 ml) was collected from an antecubital vein. The athletes
(e.g., ACTH hormone, cortisol, catecholamines, growth hor- then proceeded with a 10- to 15-min warm-up at a self-selected
mone, glucagon, IL-6, and insulin). We hypothesized that CHO intensity before each experimental trial. The intensity and duration of
and TCA metabolism would be greater, whereas fat metabo- warm-up were recorded for replication in the second trial. HR and
lism would be lower, after high-intensity interval cycling ratings of perceived exertion at the end of warm-up were also noted.
compared with continuous, moderate-intensity cycling. HR, ratings of perceived exertion, V̇O2, and V̇CO2 were recorded
continuously from 0- to 4-min periods at four different time points
METHODS during both trials (i.e., 0 – 4, 18 –22, 36 – 40, and 48 –52 min). The
data-recorded V̇O2 and V̇CO2 were averaged over 4 min, and these
Participants. Ten well-trained male cyclists and triathletes volun- averages were then used to calculate rates of substrate oxidation (see
teered to participate in the study. The characteristics of these athletes details below). Immediately after each period of gas analysis, ear
were as follows: mean ⫾ SD age 33.2 ⫾ 6.7 yr; height 1.83 ⫾ 0.06 prick blood samples were collected and analyzed to measure capillary
m; body mass 78.6 ⫾ 8.7 kg; body mass index 23.4 ⫾ 2.7 kg/m2; lactate concentration using a Lactate Pro kit (Arkray Factory, Shiga,
V̇O2 max 4.8 ⫾ 0.3 l/min; peak power output 443 ⫾ 37 W; and Japan). Within 5 min of the end of exercise, another venous blood
maximum heart rate (HR) 188 ⫾ 11 beats/min. Inclusion criteria for sample was collected. Additional venous blood samples were col-
this study required: 1) one year competitive cycling or triathlon lected 1 and 2 h after exercise while the athletes rested quietly and
experience, 2) high-intensity training in the 2 mo before participation, consumed water ad libitum.
and 3) good physical health as indicated by a medical screening Dietary and exercise training control. The athletes were requested
questionnaire. Written informed consent was obtained before study to avoid alcohol and caffeine and consume similar foods 24 h before
participation, and ethical clearance was obtained from the Human both experimental trials with the assistance of a food diary. In addition
Research Ethics Committee at The University of Queensland. they were provided with a standardized preexercise meal consisting of
Preliminary testing. The athletes were required to visit an exercise six Sanitarium Weet-bix and sufficient Sustagen Sport powder mixed
laboratory in the School of Human Movement Studies at The Uni- with milk to provide each athlete with 1.5 g CHO/kg body mass. The
versity of Queensland on three occasions before the experimental athletes consumed this meal 1.5 h before arrival at the laboratory for
trials. On the first two occasions, the athletes were tested to determine both experimental trials. They also completed training diaries and
their V̇O2 max. On the third occasion, the athletes completed a famil- were requested to follow the same training schedule in the week
iarization trial for the HIIT. before each experimental trial. In the 24 h before each trial, partici-
V̇O2 max and peak power output. V̇O2 max was determined before the pants were reminded to keep physical activity to a minimum and
experimental trials using a graded exercise test to volitional fatigue on avoid intense training.
an electromagnetically braked cycle ergometer (Lode Excalibur; Substrate oxidation. Rates of total CHO and fat oxidation were
Quinton Instruments, Seattle, WA). Following 5–10 min warm-up at calculated using stoichiometric equations (25), with the assumption
a self-selected intensity, the test began at 100 W, and the power that protein oxidation was negligible:
increased by 15-W increments every 30 s until volitional exhaustion.
Athletes were required to maintain their cadence above 60 revolu- CHO oxidation 共g/min兲 ⫽ 4.210 · V̇CO2 ⫺ 2.962 · V̇O2
tions/min (rpm). HR was recorded continuously during the test using
a radiotelemetry HR monitor (Polar Electro; Öy, Kempele, Finland). Fat oxidation 共g/min兲 ⫽ 1.695 · V̇O2 ⫺ 1.701 · V̇CO2
Peak power output was established either as 1) the power output
associated with the final completed stage of the test or 2) by multi- The rate of energy expenditure was calculated using the following
plying the fraction of time spent in the final incomplete stage by 15 W formula, where V̇O2 and V̇CO2 are in l/min

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00276.2014 • www.ajpendo.org

 METABOLIC AND HORMONAL RESPONSES TO EXERCISE E541
upper aqueous layer was removed, and anhydrous sodium sulfate
Energy expenditure 共kJ/min兲 ⫽ 16.318 · V̇O2 ⫹ 4.602 · V̇CO2
(100 –150 mg) was added to remove any remaining water. Once dried,
Blood sampling. At each of the four time points (preexercise, the remaining chloroform solution was transferred to a GC-MS vial
postexercise, 1 and 2 h postexercise), blood samples were collected in assembled with a glass insert and airtight lid.
four 6-ml Vacutainers. Three Vacutainers contained anticoagulants GC-MS instrument parameters were set according to the protocol
(EDTA, lithium heparin, or fluoride/oxalate). EDTA, lithium heparin, described previously (53). Analysis was performed using an Agilent
and fluoride/oxalate samples were placed on ice, whereas serum 7890A gas chromatograph coupled to an MSD-5975C inert mass
sample collected in a serum separation tube was left to clot at room spectrometer (Agilent). A split/splitless inlet was used for the identi-
temperature for 20 –30 min. Once the serum sample had clotted, all fication of metabolites. With the use of a CTC PAL autosampler (CTC
samples were centrifuged at 2,500 rpm for 10 min at 4°C. Plasma and Analytics), 1 ␮l of sample was injected into a glass split/splitless
serum samples were separated into aliquots and stored in Eppendorf 4-mm-inner diameter (ID) straight inlet liner packed with deactivated
tubes at ⫺80°C until further processing. One aliquot of lithium glass wool (Supelco). The inlet was set to 290°C and 56.8 kPa. It was
heparin plasma was mixed with 11 ␮l 5.26 mmol/l sodium metabisul- then pulsed splitless at 180 kPa for 1 min with column flow of 1
fite before freezing to prevent oxidation of catecholamines. ml/min. Together, an average initial linear velocity of 35 cm/s was
Blood analysis. Enzymatic assay kits were used to measure glucose achieved. Approximately 1 min after the injection, the purge flow was
and lactate (bioMérieux) in fluoride/oxalate plasma and total nonest- set to 25 ml/min.
A ZB1701 (Zebron) capillary column [30 m ⫻ 250 ␮m (ID) ⫻ 0.15

Downloaded from http://ajpendo.physiology.org/ by 10.220.33.6 on June 20, 2017

erified free fatty acids (Wako Diagnostics, Richmond, VA) in serum.
These assays were performed in duplicate on an automated biochem- ␮m film thickness] (Phenomenex) was used for all measurements.
istry analyzer (Cobas Mira; Roche Diagnostics). The intra-assay Carrier gas was ultra-high-purity grade helium (99.99%; BOC). GC
coefficient of variation was 2.0% for glucose, 2.2% for lactate, and oven temperature programming started isothermally at 45°C for 2
2.6% for nonesterified free fatty acids. ACTH, cortisol, growth hor- min. It was then increased at 9°C/min to 180°C and maintained for 5
min, increased by 40°C/min to 220°C and maintained for 5 min,
mone, and insulin concentrations were measured in duplicate using
increased by 40°C/min to 240°C and maintained for 11.5 min, and
EDTA plasma and immunoassay kits customized on an automated
finally increased by 40°C/min to 280°C and maintained for 2 min. The
analyzer (Cobas e411; Roche Diagnostics, Mannheim, Germany). The
transfer line to the mass selective detector was maintained at 250°C,
intra-assay coefficient of variation was 8.3% for ACTH, 2.2% for
the source at 230°C and the quadropole at 150°C. At 5.5 min into the
cortisol, 2.3% for growth hormone, and 4.6% for insulin. Cortisol-
run, the detector was turned on and operated in positive-ion, electron-
binding globulin was measured in duplicate by ELISA (Cusabio,
impact ionization mode at 70 eV electron energy with the electron
Hubei, China). Free cortisol was then calculated and reported as the
multiplier set with no additional voltage relative to the autotune value.
ratio between cortisol and cortisol-binding globulin (33). IL-6 and
To monitor instrumental carryover, N-hexane blanks (Merck) were
glucagon were also measured in duplicate by ELISA (R&D Systems run after every 10 samples. Identification of species was carried out
Quantikine). The intra-assay coefficient of variation was 6.0% for using mass spectra acquired in scan mode from 38 to 550 amu, with
cortisol-binding globulin, 6.4% for IL-6, and 7.8% for glucagon. detection threshold of 100 ion counts. The raw data obtained from
Epinephrine and norepinephrine were measured in single by HPLC. GC-MS was analyzed using the R package Metab and Automated
The intra-assay coefficient of variation was 6.5% for norepinephrine Mass Spectral Deconvolution and Identification System. The obtained
and 5.3% for epinephrine. data were corrected for internal standard (D4-alanine) and volume.
Metabolomics analysis. Lithium heparin plasma samples were Data for metabolomic species are reported as internal standard-
thawed, and 300 ␮l were separated into aliquots with the addition of normalized peak heights.
20 ␮l of 10 mmol/l DL-alanine-2,3,3,3-d4 (D4-alanine; Sigma-Al- Statistical analysis. Data were analyzed using SPSS version 18.0.0
drich) as an internal standard. All samples were vortexed for 1 min (Chicago, IL). Normality of the distribution for outcome measures
and frozen at ⫺80°C. The samples were then lyophilized overnight was tested using the Shapiro-Wilk test. When necessary, raw data
using an SC250 Express SpeedVac Concentrator and RVT4104 Re- were log-transformed to obtain normality. Normally distributed data
frigerated Vapor Trap (Thermo Scientific Savant). Dried pellets were were analyzed using 2 (trial) ⫻ 4 (time) factor repeated-measures
resuspended in 500 ␮l of 50% methanol and vortexed vigorously for analysis of variance. If this analysis revealed any significant time and
1 min. The suspensions were centrifuged for 5 min at 4°C at 3,500 time ⫻ trial interaction effects (P ⬍ 0.05), paired t-tests were used to
rpm. The supernatants were collected, transferred into a 15-ml falcon compare changes over time within trials, and differences between
tube, and kept on dry ice. The pellets were resuspended in 500 ␮l of trials. Data that were not normally distributed after log transformation
80% methanol, and the suspensions were recentrifuged under the were analyzed using the nonparametric Friedman’s test. If the result of
same conditions mentioned above. After the supernatants were pooled this test was significant (P ⬍ 0.05), Wilcoxon’s signed-rank tests were
from the two centrifuge steps, the volume of supernatant was in- used to compare changes within trials, and differences between trials.
creased to 5– 6 ml using milli-Q water, vortexed, and frozen overnight The false discovery rate was used to adjust P values for multiple
at ⫺80°C. Extracted frozen supernatants were further lyophilized comparisons (8). Respective P values at or below critical values of
overnight as described previously. All tubes were then stored at 0.05, 0.033, and 0.016 for changes within trials and P values at or
⫺80°C until the derivatization step. below critical values of 0.05, 0.038, 0.025, and 0.013 for difference
Methyl chloroformate derivatization was based on an optimized between trials were accepted as statistically significant. Data that did
protocol reported previously (53). Lyophilized samples were resus- not require log transformation are reported as means ⫾ SD, whereas
pended in 1 mol/l sodium hydroxide (200 ␮l) and vigorously vor- log-transformed data are reported as geometric means ⫾ 95% confi-
texed. After the resuspended samples were transferred to silanized dence interval. Data that were not normally distributed are reported as
tubes, methanol (334 ␮l) and pyridine (67 ␮l) were added. To initiate median, 25th, and 75th percentiles.
the derivatization process, methylchloroformate (40 ␮l) was added Data processing. The metabolomic compound names were entered
followed by vigorous mixing for 30 s. After this step was repeated, into the Human Metabolome Database version 3.5 (http://www.hmdb.
chloroform (400 ␮l) was added, and the mixture was vortexed for 10 ca/) to retrieve KEGG compound identifiers. This list of identifiers
s to separate the methylchloroformate derivatives from the reaction was then used to search the KEGG BRITE compound classification
mixture. Sodium bicarbonate (400 ␮l, 50 mmol/l) was then added, and hierarchy (http://www.genome.jp/kegg/tool/map_brite1.html). This
the mixture was vortexed for 10 s. Thereafter, samples were centri- hierarchical information was used to create a binary network of
fuged for 5 min at 5°C at 2,000 rpm to separate the aqueous layer from classification terms and species to import into Cytoscape (http://
the organic layer. With the use of a glass Pasteur pipette and bung, the www.cytoscape.org/) to generate a network map. Statistical analysis

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00276.2014 • www.ajpendo.org

E542 METABOLIC AND HORMONAL RESPONSES TO EXERCISE

Table 1. Exercise intensity and substrate oxidation during during exercise was not significantly different between the
exercise HIIT and MOD trials (P ⫽ 0.71) (Table 1). HR, V̇O2,
%V̇O2 max, power output, %peak power output, and the rate of
HIIT MOD
energy expenditure were all higher during HIIT compared with
Heart rate, beats/min 179.9 ⫾ 9.6* 154.3 ⫾ 10.7 the corresponding periods during MOD (P ⬍ 0.05) (Table 1).
V̇O2, l/min 3.7 ⫾ 0.3* 2.9 ⫾ 0.2
V̇O2max, % 81.6 ⫾ 3.7* 66.7 ⫾ 3.5
The rates of CHO oxidation (Table 1) and blood lactate
Power output, W 319 ⫾ 26* 216 ⫾ 26 concentration (Fig. 1) were also higher during HIIT compared
Peak power output, % 72.0 ⫾ 3.2 48.5 ⫾ 3.1 with MOD (P ⬍ 0.01). Conversely, the rate of fat oxidation
CHO oxidation, g/min 4.20 ⫾ 1.04* 2.75 ⫾ 0.64 tended to be lower (P ⫽ 0.037), whereas respiratory exchange
Fat oxidation, g/min 0.41 ⫾ 0.26 0.47 ⫾ 0.25
RER 0.95 ⫾ 0.04 0.91 ⫾ 0.04
ratio tended to be higher (P ⫽ 0.063), during HIIT compared
Lactate, mmol/l 7.5 ⫾ 1.2* 1.5 ⫾ 1.1 with MOD (Table 1).
Rate of EE, kJ/min 19.4 ⫾ 1.6* 15.7 ⫾ 1.5 Glucose, lactate, and free fatty acids. Blood lactate was
Total work, kJ 792 ⫾ 95 797 ⫾ 95 higher throughout HIIT compared with MOD (P ⬍ 0.01;
Data are means ⫾ SD. HIIT, high-intensity interval training; MOD, mod- Fig. 1A). Compared with preexercise values, plasma glucose
concentration (P ⬍ 0.01; Fig. 1B) and total serum free fatty

Downloaded from http://ajpendo.physiology.org/ by 10.220.33.6 on June 20, 2017

erate-intensity continuous exercise; CHO, carbohydrate; RER, respiratory
exchange ratio; EE, energy expenditure. Data are for 4-min blocks at minutes acid concentration (P ⬍ 0.001; Fig. 1C) increased after both
0 – 4, 18 –22, 36 – 40, and 48 –52, corresponding to the first, fourth, seventh,
and ninth intervals in the high-intensity interval exercise. *P ⬍ 0.05 for HIIT HIIT and MOD. Compared with MOD, both plasma glucose
vs. MOD. and lactate (Fig. 1D) were higher immediately after HIIT
(P ⬍ 0.05). Plasma lactate concentration increased after
HIIT (P ⬍ 0.01) but not MOD (P ⫽ 0.29) (Fig. 1). Total
including time and interaction effects and fold changes for each
metabolomic compound were then imported as a separate network
serum free fatty acid concentration was not significantly
attributes file and used to specify visual properties on the network. In different between the trials (P ⫽ 0.33). During the postex-
addition to quantitative statistical analysis, a heat map depiction of the ercise recovery period, plasma glucose concentration re-
percentage change in each of the plasma metabolites from preexercise turned to preexercise values within 1 h after both trials.
levels was generated for the entire metabolomics dataset using the Plasma lactate concentration remained higher 1 h after HIIT
gplots package within the R statistical software (heat map.2 function) compared with MOD (P ⬍ 0.05). Lactate was below pre-
to display relative changes over time in families of metabolites in exercise values 2 h after HIIT (P ⬍ 0.05). Serum free fatty
response to exercise trials.
acid concentration remained higher at 1 and 2 h after both
RESULTS HIIT and MOD (P ⬍ 0.01).
Metabolomic compounds. Targeted metabolomic analysis
Exercise intensity and substrate oxidation. The duration of identified 49 compounds in the plasma samples. According to
exercise was exactly 60 min for the HIIT trial and 61 min ⫾ 14 KEGG classification, these species were categorized into var-
s for the MOD trial. The total amount of work completed ious types of metabolites, including fatty acids (saturated,

A 15 HIIT MOD C 1.2

* *
Free fatty acids (mmol/l)

#
Blood lactate (mmol/l)

0.9
10

0.6

5
0.3

Fig. 1. Free fatty acid, glucose, and lactic 0 0.0

acid concentrations following high-intensity 0-4 18-22 36-40 48-52
interval training (HIIT) and moderate-inten- Time (min)
sity, continuous exercise (MOD). Data rep-
resent means ⫾ SD for blood lactate (during B 10 *
D 12

exercise) and free fatty acids and geometric #

*
Plasma lactate (mmol/l)

10
mean ⫾ 95% confidence interval for plasma
Glucose (mmol/l)

glucose and plasma lactate. *P ⬍ 0.05 vs. 8

8
preexercise. #P ⬍ 0.05 for HIIT vs. MOD. *
6
6
4
#
4 2 *
0 0
E

ST

1H

2H

ST

1H

2H

ST

1H

2H

ST

1H

2H
PR

PR

PR

PR
PO

PO

PO

PO

H IIT MOD H IIT MOD

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00276.2014 • www.ajpendo.org

 METABOLIC AND HORMONAL RESPONSES TO EXERCISE E543
monounsaturated, polyunsaturated), amino acids (branched compared with MOD (P ⬍ 0.01). Other carboxylic acids such
chain, essential, nonessential), carboxylic acids, fatty acyl as 4-methyl-2-oxopentanoic acid and 3-methyl-2-oxopentanoic
carnitines, and unclustered species. The network diagram in acid also increased after HIIT (P ⬍ 0.01) (data not shown).
Fig. 2A summarizes the distribution of these metabolites, and Only succinic acid increased significantly after MOD (P ⬍
significant changes observed over time. Among the 49 metab- 0.01).
olites that were identified, 29 changed significantly after exer- Most amino acids did not change during HIIT or MOD, with
cise (P ⬍ 0.05); among these 29 metabolites, 11 changed after the exception of the nonessential amino acids alanine (Fig. 5A),
both HIIT and MOD, 13 changed only after HIIT, and 5 glutamate (Fig. 5B), and tyrosine (data not shown), which
changed only after MOD (Fig. 2, B and C). The heat map in increased following HIIT (P ⬍ 0.05). Alanine was also higher
Fig. 3 summarizes the relative magnitude of changes in each of immediately after HIIT compared with MOD (P ⬍ 0.01).
the detected metabolites from preexercise levels in response to During recovery from exercise, the branched-chain amino
HIIT and MOD. acids leucine (Fig. 5C), valine (Fig. 5D), and isoleucine (data
The TCA cycle intermediates citric acid (Fig. 4A), succinic not shown); the essential amino acid methionine (Fig. 5E); and
acid (Fig. 4B), aconitic acid (Fig. 4C), and malonic acid (Fig. the nonessential amino acids alanine (Fig. 5A) and proline (Fig.
4D) all increased after HIIT (P ⬍ 0.01). Furthermore, citric 5F) all decreased below preexercise values, particularly after

Downloaded from http://ajpendo.physiology.org/ by 10.220.33.6 on June 20, 2017

acid, succinic acid, and aconitic acid were all higher after HIIT MOD (P ⬍ 0.01).

Fig. 2. Network map (A) and KEGG classification of metabolomic species (B) identified and quantified in plasma by gas chromatography-mass spectrometry.
Numbered nodes represent measured metabolites. Nodes with colored letters represent a significant change with time for specific species after HIIT (H) and MOD
(M) exercise. Numbered nodes with red boarders represent those metabolites that had a significantly different response between high-intensity intervals and
continuous, moderate-intensity exercise. The Venn diagram (C) illustrates the distribution of common and trial-specific changes in metabolomic species after
exercise.

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00276.2014 • www.ajpendo.org

E544 METABOLIC AND HORMONAL RESPONSES TO EXERCISE

Downloaded from http://ajpendo.physiology.org/ by 10.220.33.6 on June 20, 2017

Fig. 3. Heat map showing fold changes in metabolomic compounds after high-intensity intervals and continuous, moderate-intensity exercise. PUFA,
polyunsaturated fatty acids; SFA, saturated fatty acids; MUFA, monounsaturated fatty acids. Numbers on the right correspond to compound numbers in Fig. 2
and KEGG compound identifiers.

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00276.2014 • www.ajpendo.org

 METABOLIC AND HORMONAL RESPONSES TO EXERCISE E545

A 0.25 C 0.015
#
# *
0.20 *
0.010
Citric acid

Aconitic acid
0.15

0.10
0.005

0.05

0.00 0.000 Fig. 4. Plasma tricarboxylic acid intermedi-

ates following HIIT and MOD. Data are in
arbitrary units and represent means ⫾ SD for
B 0.08 D 0.003
aconitic and succinic acid and median ⫾
*
# *
interquartile range for citric and malonic
0.06
acid. *P ⬍ 0.05 vs. preexercise. #P ⬍ 0.05
for HIIT vs. MOD.

Downloaded from http://ajpendo.physiology.org/ by 10.220.33.6 on June 20, 2017

*
Succinic acid

Malonic acid 0.002

0.04

0.001
0.02

0.00 0.000
H

H
ST
E
E

E
ST

ST

ST
1H

1H

2H

1H

2H
R
2

2
PR

PR

PR
PO
P
PO

O
P

P
H IIT MOD H IIT MOD

The saturated fatty acids myristic acid (Fig. 6A), dodecanoic plasma concentrations of ACTH, free cortisol, and growth
acid (Fig. 6B), and decanoic acid (Fig. 6C) all increased after hormone were all higher immediately after HIIT (P ⬍ 0.05).
both trials (P ⬍ 0.01). The monounsaturated fatty acids palmi- During the postexercise recovery period, growth hormone
toleic acid (Fig. 6D) and heptadecenoic acid (data not shown) concentration remained elevated up to 1 h after MOD and up
increased after both trials (P ⬍ 0.01), whereas myristoleic acid to 2 h after HIIT (P ⬍ 0.05), whereas IL-6 remained elevated
(Fig. 6E) and oleic acid (Fig. 6F) increased only after HIIT up to 2 h after both HIIT and MOD (P ⬍ 0.01). Norepinephrine
(P ⬍ 0.05) (Fig. 6). The polyunsaturated fatty acid ␥-linolenic was also higher 1 h after MOD (P ⬍ 0.01), whereas ACTH was
acid increased after MOD (P ⫽ 0.005), and there was a trend lower than preexercise values 2 h after HIIT (P ⬍ 0.01).
toward an increase after HIIT (P ⫽ 0.037) (data not shown). Plasma insulin concentration did not change significantly fol-
The increase in some of these fatty acids (e.g., oleic acid, lowing HIIT (P ⫽ 0.14) or MOD (P ⫽ 0.055) (Fig. 7G).
myristoleic acid) after exercise was relatively brief, whereas However, during the postexercise recovery period, insulin was
other fatty acids (e.g., myristic acid, palmitoleic acid) remained lower than preexercise values at 1 and 2 h after both HIIT and
elevated during the 2-h recovery period. MOD (P ⬍ 0.01). Plasma glucagon also did not change
The percentage molar distribution of all nonesterified fatty significantly during HIIT (P ⫽ 0.12) or MOD (P ⫽ 0.32) (Fig.
acid (NEFA) detected in plasma was calculated using normal- 7H), whereas it was lower than preexercise values at 2 h after
ized peak heights as previously described (39) and is presented both HIIT and MOD (P ⬍ 0.05).
in Table 2. Among the fatty acids that increased after exercise,
oleic acid was the most abundant fatty acid, whereas myristo- DISCUSSION
leic acid was the least abundant. The molar distribution of
dodecanoic acid, myristic acid, and palmitoleic acid increased In this study, we systematically compared the metabolic and
after both HIIT and MOD (P ⬍ 0.05), the molar distribution of hormonal responses to matched-work high-intensity interval
octanoic acid decreased after MOD (P ⬍ 0.05), and the molar and continuous moderate-intensity exercise. CHO oxidation,
distribution of linoleic acid decreased after HIIT (P ⬍ 0.05). plasma glucose, and lactate were markedly higher after high-
Other metabolomic species that were present in plasma intensity interval exercise compared with continuous moder-
(including glycine, phenylalanine, cysteine, threonine, orni- ate-intensity exercise, whereas fat oxidation and total serum
thine, arachidonic acid, adrenic acid, docosahexaenoic acid, eico- NEFA were not significantly different between the two exer-
sapentaenoic acid, bis-homo-␥-linolenic acid, azelaic acid, oc- cise trials. Targeted metabolomic analysis of plasma samples
tanoylcarnitine, palmitic acid, stearic acid, arachidic acid, quinic revealed detailed insights into the exercise intensity-dependent
acid, heptadecanoic acid, phenylacetic acid) did not change changes in TCA cycle intermediates, specific NEFA, and
significantly after exercise (P ⬎ 0.05). amino acids. Although there was no significant difference in
Hormones. The plasma concentrations of growth hormone total serum NEFA between the two trials after exercise, we did
(Fig. 7A), norepinephrine (Fig. 7B), and IL-6 (Fig. 7C) in- notice some differences in the abundance of individual fatty
creased after both HIIT and MOD, whereas ACTH (Fig. 7D), acids. This finding highlights the importance of considering
free cortisol (Fig. 7E), and epinephrine (Fig. 7F) increased more than changes in total serum NEFA when evaluating the
only after HIIT (P ⬍ 0.05) (Fig. 7). Compared with MOD, the metabolic demands of exercise. The significant increase in

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00276.2014 • www.ajpendo.org

E546 METABOLIC AND HORMONAL RESPONSES TO EXERCISE

A 2.0
#
D 0.9

* * *
0.8
1.5 #
0.7

Alanine

Valine
1.0
0.6

0.5
0.5

0.0 0.4

B 0.05 E 0.05
* * *
0.04 0.04

Downloaded from http://ajpendo.physiology.org/ by 10.220.33.6 on June 20, 2017

Methionine
0.03 0.03
Glutamate

Fig. 5. Plasma amino acids following HIIT

and MOD. Data are in arbitrary units and
0.02 0.02
represent means ⫾ SD. *P ⬍ 0.05 vs. pre-
exercise. #P ⬍ 0.05 for HIIT vs. MOD.
0.01 0.01

0.00 0.00

C 0.5
* * F 1.5
*
0.4

1.0
0.3
Leucine

Proline
0.2
0.5

0.1

0.0 0.0
E

ST

1H

2H

ST

1H

2H

ST

1H

2H

ST

1H

2H
PR

PR

PR

PR
PO

PO

PO

PO
H IIT MOD H IIT MOD

TCA cycle intermediates suggests that during exercise these (48), urinary succinate (12), plasma/serum succinate, malate,
metabolic compounds spill over from muscle into the circula- fumarate, citrate-isocitrate, aconitic acid, and ␣-ketoglutarate
tion and/or are released systemically from other metabolic (34, 42) after exercise. Lewis et al. (34) observed that succi-
organs (e.g., the liver). The individual NEFA that were most nate, malate, and fumarate (but not citrate-isocitrate, aconitic
responsive to exercise were not necessarily the most abundant acid, or ␣-ketoglutarate) were higher in plasma from fast
NEFA in plasma. This finding raises some questions about the marathon runners compared with slow marathon runners. This
factors that determine the preference(s) for mobilization of finding also provides some indirect evidence of the intensity-
NEFA during exercise. Furthermore, the changes in the per- dependent changes in these TCA intermediates after exercise
centage molar distribution of NEFA contrast with previous that we observed. The biological role (if any) of TCA inter-
findings after exercise and challenge the previously held notion mediates in urine and plasma after exercise is unclear because
that changes in plasma NEFA composition reflect the compo- they are primarily active in skeletal muscle (14, 15) and the
sition of adipose tissue. These metabolic changes were accom- liver (24) during exercise. It is likely that the presence of TCA
panied by changes in counterregulatory hormones, some of intermediates in urine and plasma after exercise likely reflects
which increased only after HIIT. Collectively, these findings “spillover” from skeletal muscle or the liver.
highlight the value of targeted metabolomics to obtain more The intensity-dependent increase in plasma TCA intermedi-
detailed and specific information on metabolic responses to ates in the present study agrees with other reports that the
exercise beyond basic measures of substrate oxidation, glu- concentration of TCA intermediates in skeletal muscle also
cose, and free fatty acids. increases with exercise intensity (15). The increase in the
TCA intermediates, including citric acid (1.8-fold), succinic concentration of TCA intermediates in skeletal muscle, and
acid (3.0-fold), aconitic acid (2.0-fold), and malonic acid possibly plasma too, during high-intensity exercise may result
(1.6-fold), all increased in plasma during high-intensity inter- from epinephrine-induced accumulation of pyruvate (54). Gi-
val exercise in the present study. Of the TCA intermediates, bala et al. (15) found that of all of the TCA intermediates in
only succinic acid increased in response to moderate-intensity skeletal muscle, malate increased to the greatest extent (7.5⫻)
exercise, and this response was considerably lower than the during exercise and contributed the most to the expansion of
response following high-intensity interval exercise. Others the pool of TCA intermediates at the end of exercise. By
have reported an increase in urinary citrate and ␣-ketoglutarate contrast, succinic acid showed the greatest fold change in

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00276.2014 • www.ajpendo.org

 METABOLIC AND HORMONAL RESPONSES TO EXERCISE E547

A 0.25 E 0.06
* * * *
0.20

0.04

Palmitoleic acid
Myristic acid

0.15

0.10
0.02

0.05

0.00 0.00

B 0.10 C 0.0015
* * * *
0.08

Downloaded from http://ajpendo.physiology.org/ by 10.220.33.6 on June 20, 2017

0.0010
Dodecanoic acid

Myristoleic acid

0.06

0.04
Fig. 6. Plasma fatty acids following HIIT and
0.0005 MOD. Data are in arbitrary units and repre-
sent means ⫾ SD. *P ⬍ 0.05 vs. preexercise.
0.02

0.00 0.0000

D 0.04 F 0.5
* *
0.4
*
0.03
Decanoic acid

0.3
Oleic acid

0.02
0.2

0.01
0.1

0.00 0.0
E

E
ST

2H

ST
ST

ST

1H
E

E
1H

2H

1H

2H

PR

2H
PR
1
PR

PR

PO
PO

PO
PO

H IIT MOD H IIT MOD

plasma after exercise in the present study. One explanation for hetadecenoic acid (C17:1), and palmitoleic acid (C16:1), after
the rise in succinic acid is that carbon skeletons may enter the both exercise trials. Myristoleic acid (C14:1) and oleic acid
cycle at the level of 2-oxoglutarate through the alanine ami- (C18:1n-9) increased only after HIIT, whereas ␥-linolenic acid
notransferase reaction, thereby causing accumulation of inter- (C18:3) increased only after continuous, moderate-intensity
mediates in the second span of the TCA cycle (i.e., malate, exercise.
fumarate, succinate) (15). Typically, TCA intermediates accu- Among the NEFA that increased after exercise in the present
mulate in skeletal muscle early during exercise and then study, dodecanoic acid increased to the greatest extent (3.2-
progressively decline thereafter (18). In the present study, the fold), palmitoleic acid and myristic acid increased moderately
plasma concentrations of TCA intermediates may therefore (2.5-fold), and oleic acid increased only slightly (1.3-fold). The
have been higher during exercise compared with the end of percentage molar distribution of these NEFA also increased
exercise. after exercise (Table 2). Interestingly, with the exception of
Most exercise studies have focused on changes in plasma/ oleic acid, these NEFA generally comprised ⬍4% of the sum
serum NEFA; comparatively little research has investigated of all NEFA that we measured in plasma. Palmitoleic acid and
exercise-induced changed in other lipid classes (e.g., phospho- myristic acid also make up only a small proportion of fatty
lipids, mono-, di- or triacylglcerols, diacylglycerols, choles- acids in adipose tissue (23, 29). Changes in plasma fatty acids
teryl esters) (44). NEFA represent a relatively small proportion during exercise represent the balance between fatty acid release
of fatty acids in tissues. Nevertheless, they are important in the from adipose tissue and fatty acid uptake into skeletal muscle
context of exercise from both a physiological standpoint (i.e., and the liver (41). There are several possible explanations as to
alterations in the composition of fatty acids delivered to tis- why dodecanoic acid, palmitoleic acid, and myristic acid were
sues) and a practical perspective (i.e., increasing the likelihood preferentially mobilized during exercise despite their low rel-
of detecting significant changes) (44). In addition to an in- ative abundance in plasma and adipose tissue. One possibility
crease in total serum NEFA, we detected increases in the is that transport proteins for these particular fatty acids are
plasma concentration of specific NEFA, including myristic more abundant in skeletal muscle and adipose tissue, and/or
acid (C14:0), decanoic acid (C10:0), dodecanoic acid (C12:0), they are more sensitive to muscle contractions and hormones.

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00276.2014 • www.ajpendo.org

E548 METABOLIC AND HORMONAL RESPONSES TO EXERCISE

Table 2. Percentage molar distribution of nonesterified fatty acids

HIIT MOD

Pre Post 1h 2h Pre Post 1h 2h

Octanoic acid (C8:0) 0.4 (0.3) 0.3 (0.3) 0.3 (0.1) 0.2 (0.1) 0.3 (0.1) 0.3 (0.1) 0.2 (0.1)* 0.2 (0.1)*
Decanoic acid (C10:0) 0.3 (0.2) 0.5 (0.3) 0.5 (0.3) 0.4 (0.3) 0.4 (0.2) 0.6 (0.2) 0.4 (0.2) 0.5 (0.3)
Dodecanoic acid (C12:0) 0.6 (0.2) 1.5 (0.6)* 1.0 (0.8) 1.3 (0.9)* 0.7 (0.4) 1.8 (0.5)* 1.1 (0.6)* 1.3 (0.8)*
Myristic acid (C14:0) 1.8 (0.5) 3.4 (0.9)* 2.7 (1.0)* 3.4 (1.6)* 2.0 (0.7) 4.4 (1.2)* 3.1 (0.9)* 3.3 (1.0)*
Palmitic acid (C16:0) 30 (0.8) 29 (0.8) 30 (0.8) 30 (0.8) 32 (5.8) 29 (1.3) 30 (0.7) 30 (1.0)
Margaric acid (C17:0) 1.1 (0.2) 1.3 (0.3) 1.2 (0.2) 1.2 (0.3) 1.0 (1.0) 1.2 (0.3) 1.2 (0.2) 1.3 (0.2)
Stearic acid (C18:0) 24 (1.2) 23 (1.2) 23 (2.3) 23 (3.2) 22 (3.5) 21 (2.2) 23 (1.6) 22 (1.0)
Arachidic acid (C20:0) 0.2 (0.34) 0.2 (0.19) 0.1 (0.09) 0.1 (0.02) 0.2 (0.17) 0.1 (0.03) 0.1 (0.02) 0.1 (0.01)
Myristoleic acid (C14:1) 0.02 (0.004) 0.03 (0.005) 0.02 (0.005) 0.02 (0.003) 0.06 (0.127) 0.02 (0.005) 0.02 (0.003) 0.03 (0.004)
Palmitoleic acid (C16:1) 0.3 (0.1) 0.6 (0.2)* 0.5 (0.2)* 0.7 (0.3)* 0.4 (0.1) 0.9 (0.4)* 0.6 (0.2)* 0.7 (0.2)*
Vaccenic acid (C18:1n-7) 9.1 (0.1) 8.7 (0.7) 9.0 (0.6) 8.7 (0.6) 8.8 (1.6) 8.5 (0.9) 8.7 (0.6) 8.7 (0.5)
Oleic acid (C18:1n-9) 7.5 (1.0) 8.1 (1.0) 7.9 (0.9)* 8.7 (1.1) 8.1 (0.8) 8.8 (0.7) 8.2 (0.7) 8.3 (0.7)
Linoleic acid (C18:2) 16 (0.9) 15 (1.1)* 16 (1.3) 15 (0.9)* 16 (1.6) 15 (1.8) 15 (1.1) 15 (0.8)

Downloaded from http://ajpendo.physiology.org/ by 10.220.33.6 on June 20, 2017

␥-Linolenic acid (C18:3) 0.7 (0.3) 0.8 (0.2) 0.7 (0.2) 0.7 (0.3) 0.7 (0.5) 0.9 (0.3) 0.7 (0.2) 0.7 (0.2)
Arachidonic acid (C20:4) 5.4 (0.8) 4.8 (0.8) 5.1 (0.9) 4.7 (0.6) 4.8 (1.3) 4.6 (0.7) 4.9 (0.7) 4.9 (0.7)
Eicosapentaenoic acid (C20:5) 0.9 (0.3) 0.8 (0.3) 0.8 (0.3) 0.8 (0.2) 1.0 (0.3) 0.9 (0.4) 0.9 (0.4) 1.0 (0.3)
Docosahexaenoic acid (C22:6) 1.9 (0.5) 1.9 (0.6) 1.8 (0.4) 1.8 (0.3) 1.8 (0.5) 1.7 (0.6) 1.8 (0.5) 1.9 (0.4)
Adrenic acid (C22:4) 0.15 (0.07) 0.14 (0.04) 0.14 (0.03) 0.13 (0.02) 0.24 (0.35) 0.13 (0.03) 0.13 (0.02) 0.14 (0.04)
Unsaturated-to-saturated ratio 0.73 (0.04) 0.70 (0.03) 0.72 (0.04) 0.71 (0.04) 0.72 (0.04) 0.69 (0.03) 0.71 (0.04) 0.70 (0.04)
*Significant difference vs. preexercise, P ⬍ 0.05.

Up to five different fatty acid transport proteins are expressed exercise trials. Consistent with other studies, there were no
in skeletal muscle and adipose tissue (20). It is yet to be significant differences in plasma fatty acid concentrations dur-
determined whether these individual proteins transport specific ing recovery from the high-intensity exercise and the continu-
fatty acids. However, the expression or function of specific ous moderate-intensity exercise (22, 41). In support of the
substrate transporters may be adaptable to alter preferences for present study in relation to postexercise lipid metabolism,
certain cellular substrates (20). Another possibility is that the several studies have found no difference in fat oxidation or
chemical structure of these fatty acids favors a greater ATP respiratory exchange ratio after high-intensity exercise com-
yield (relative to O2 demand) during exercise compared with pared with moderate-intensity exercise (30, 31). By contrast,
other fatty acids. Further research is needed to examine in more others have reported that fat oxidation was higher while respi-
detail why certain fatty acids are mobilized into the circulation ratory exchange ratio was lower after high-intensity exercise
to a greater or lesser extent during exercise. compared with moderate-intensity exercise (36, 63). These
We observed that the ratio of unsaturated/saturated NEFA in discrepancies may reflect differences between these studies in
plasma did not change after exercise, which contrasts with exercise protocols and the training status of participants. It also
other research reporting an increase in this ratio. Mougios et al. remains unclear whether differences in intensity and substrate
(39) reported that the molar distribution of palmitic acid and metabolism during exercise translate to differences in energy
stearic acid increased, whereas the molar distribution of oleic expenditure during recovery from exercise (31, 38).
acid and linoleic acid decreased, after exercise. It is difficult to In the present study, changes in the concentration of NEFA
compare these studies with our results because we measured a and other specific fatty acids (e.g., dodecanoic acid, myristic
greater number of plasma NEFA. Consequently, the ratio of acid, palmitoleic acid) remained high for 2 h after exercise.
unsaturated/saturated fatty acids was lower in the present Mulla et al. (41) reported similar changes in NEFA and
study. Nevertheless, we observed that the molar distribution of glycerol output from adipose tissue during and after exercise.
palmitic acid and stearic acid did not change, oleic acid Together, these findings reflect a shift in substrate oxidation in
increased slightly, and linoleic acid decreased slightly (Table favor of lipid metabolism (28, 30). This response may occur to
2). The changes in these specific NEFA after exercise are continue supplying energy while facilitating restoration of
generally consistent with other studies (6, 39). In addition to glucose homeostasis and glycogen restoration (30).
differences in the number of NEFA that we detected, the This is the first study to compare the effects of exercise
variation between our findings and others (39) may be related intensity on changes in a broad range of amino acids in plasma.
to the intensity and duration of exercise and the sex and The plasma concentrations of alanine, glutamate, and tyrosine
training status of the participants. increased significantly only during high-intensity interval ex-
A strength of the present study is that we monitored meta- ercise. Furthermore, alanine was significantly higher after
bolic changes not only immediately following completion of high-intensity interval exercise compared with continuous
exercise but additionally during the early hours of recovery moderate-intensity exercise. Other studies have reported an
from exercise. We found that total serum NEFA remained increase in alanine concentration in both venous plasma (3) and
elevated 2 h after both high-intensity interval exercise and arterial plasma in response to exercise (40, 59). This rise in
continuous, moderate-intensity exercise. Among the individual alanine within the circulation reflects its synthesis and release
fatty acids that we measured in plasma by metabolomics, from skeletal muscle (58). Previous research indicates that
heptadecenoic acid, dodecanoic acid, myristic acid, oleic acid, alanine release from skeletal muscle is more sustained during
and palmitoleic acid were also still elevated 2 h after both exercise with low vs. normal muscle glycogen (59). We did not

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00276.2014 • www.ajpendo.org

 METABOLIC AND HORMONAL RESPONSES TO EXERCISE E549

A * # E
#
16384
*
150 *
8192
4096
log 2 GH (pg/ml)

2048 100
*

Freecortisol
1024
*
512
256 50
128
64
32 0

B 40 F 1.5
*
*

Downloaded from http://ajpendo.physiology.org/ by 10.220.33.6 on June 20, 2017

30
Noradrenaline (nmol/l)

Adrenaline (nmol/l)

1.0

20
*
0.5
10 * Fig. 7. Plasma hormones following HIIT and
MOD. Data represent means ⫾ SD for gluca-
0 0.0 gon; geometric mean ⫾ 95% confidence in-
terval for ACTH, insulin, and growth hor-
mone; and median ⫾ interquartile range for
C 5 * G 200 norepinephrine, epinephrine, cortisol, and in-
terleukin (IL)-6. *P ⬍ 0.05 vs. preexercise.
4 #P ⬍ 0.05 for HIIT vs. MOD. NB, free
150 cortisol represents the ratio between total cor-
* tisol and cortisol-binding globulin.
Insulin (pmol/l)

3
IL-6 (pg/ml)

100
2
* *
50
1

0 0

D 40 H 300
#
*
30
* *
200
Glucagon (pg/ml)
ACTH (pmol/l)

20

100
10
*
0 0
E

E
1H

2H

1H

2H

1H

2H

1H

2H
ST

ST

ST

ST
PR

PR

PR

PR
PO

PO

PO

PO

H IIT MOD H IIT MOD

measure muscle glycogen in the present study, but the greater We can only speculate, but the increase in venous plasma
reliance on CHO metabolism could possibly account for the glutamate concentration that we observed may have served
higher plasma concentration of alanine after high-intensity to generate more TCA intermediates during high-intensity
interval exercise. In contrast with our findings, other studies interval exercise. We detected increases in the plasma con-
have reported a decrease (40) or no change (59) in the con- centrations of 4-methyl-2-oxopentanoic acid and 3-methyl-
centration of glutamate in arterial plasma. In skeletal muscle, 2-oxopentanoic acid after high-intensity intermittent exer-
glutamate consumption increases markedly during the first few cise. These carboxylic acids may represent breakdown prod-
minutes of exercise, particularly when muscle glycogen is low ucts of leucine (46).
(59). The consumption of glutamate in muscle may shift the The other amino acids that we measured in plasma (i.e.,
alanine aminotransferase reaction in favor of the formation of threonine, phenylalanine, glycine, cysteine, proline, and orni-
alanine and TCA intermediates such as ␣-oxoglutarate (58). thine) did not change significantly during either exercise trial.

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00276.2014 • www.ajpendo.org

E550 METABOLIC AND HORMONAL RESPONSES TO EXERCISE

Some of these amino acids, such as threonine, phenylalanine, role for these hormones in regulating fat metabolism during
and glycine, are not metabolized in skeletal muscle (58), which exercise. At rest, IL-6 induces lipolysis in skeletal muscle (but
could explain why they did not change after exercise in this not adipose tissue) (66), yet its role in regulating lipid metab-
study. It is possible that phenylalanine is converted to tyrosine, olism during exercise remains unresolved (62). Cortisol medi-
which might partially account for the rise in plasma tyrosine ates protein metabolism during exercise (11) and could possi-
concentration that occurred during high-intensity interval ex- bly account for the higher plasma alanine concentration after
ercise. high-intensity intermittent exercise in the present study.
During recovery from exercise, branched-chain amino acids, There were several limitations to the present study. First, we
methionine, alanine, and proline were all significantly lower could not determine the precise intensity for high-intensity
compared with preexercise, particularly after continuous mod- interval exercise and continuous moderate-intensity exercise
erate-intensity exercise. This sustained decline in these amino that would elicit the greatest difference in CHO and fat me-
acids during recovery from exercise reflects their role in tabolism. Second, the short 30-s stages that we employed
gluconeogenesis. Wahren et al. (61) provided early evidence during the V̇O2 max test may not have provided stable readings
that amino acids (particularly alanine), pyruvate and lactate to determine relative workloads in an accurate manner. Last,
assist in restoring hepatic glucose metabolism after exercise. we did not assess the metabolic and hormonal response to

Downloaded from http://ajpendo.physiology.org/ by 10.220.33.6 on June 20, 2017

More recent work by Mourtzakis et al. (40) revealed that efflux individual intervals. The metabolic and hormonal responses
of branched-chain amino acid, glutamine, and alanine from over the short period of these intervals may be different
skeletal muscle stimulates a significant rise in arterial blood compared with that of 40 min of continuous exercise. Because
glucose concentration during recovery from exercise. Despite a we only reported pre- and postexercise data for both trials, the
difference in plasma glucose concentration after exercise, the results of the two exercise trials are not directly comparable.
plasma concentrations of amino acids during recovery were not In summary, the present study provides detailed insights into
significantly between the two exercise trials in the present the metabolic and hormonal responses to high-intensity inter-
study. Conceivably, amino acids may play a more significant val exercise, which is rapidly gaining popularity as a time-
role in supporting gluconeogenesis during recovery from ex- effective form of training. Our findings highlight the benefits of
ercise that depletes rather than raises blood glucose concentra- targeted metabolomic profiling to characterize changes in me-
tion (40). tabolism through the TCA cycle and lipid and protein path-
We detected an increase in several other metabolic species in ways. The athletes in the present study consumed breakfast
plasma, including 2-hydroxybutyric acid, itaconic acid, and before exercise; the metabolomic response to exercise may be
citraconic acid, most notably after high-intensity intermittent different under fasting conditions. Future research could use a
exercise. These metabolites do not fall into any of the classical targeted metabolomic approach to investigate differences in
metabolic clusters, and relatively little is known about their metabolism relative to individual thresholds for maximal fat
biological functions, at least in humans. However, 2-hydroxy- oxidation. We only measured a small subset of the array of
butyric acid is elevated in urine from patients with lactic possible lipids. Future studies employing targeted metabolo-
acidosis and ketoacidosis (32), which suggests that it plays mics analysis could investigate changes in other lipids, includ-
some sort of role under conditions of metabolic stress. Itaconic ing triglycerides, phospholipids, and cholesterol esters. Further
acid may be derived from aconitic acid (which is a TCA research is also needed to examine metabolomics adaptations
intermediate), whereas citraconic acid may be a by-product of to exercise training.
glutamine metabolism (47).
We attempted to link intensity-dependent changes in metab- GRANTS
olism with specific hormonal responses. ACTH, cortisol, and This research was supported by a student scholarship from the School of
growth hormone were significantly higher after high-intensity Human Movement Studies at The University of Queensland and a grant from
intermittent exercise compared with continuous moderate-in- the Centre of Excellence for Applied Sports Science at the Queensland
Academy of Sport, Brisbane Australia.
tensity exercise. Cortisol and growth hormone play an impor-
tant role in regulating blood glucose concentration at rest (9, DISCLOSURES
10), but less so during exercise (11, 26). It seems unlikely that
these hormones were solely responsible for the higher plasma No conflicts of interest, financial or otherwise, are declared by the authors.
glucose concentration after high-intensity intermittent exercise.
AUTHOR CONTRIBUTIONS
Catecholamines and IL-6 may also have induced a rise in
plasma glucose during exercise (27, 35). Exercise suppressed Author contributions: J.M.P. and D.C.-S. conception and design of research;
J.M.P., S.J.T., J.F.M., and J.A.B. analyzed data; J.M.P., S.J.T., J.F.M., J.A.B., and
insulin release, whereas plasma glucagon concentration did not D.C.-S. interpreted results of experiments; J.M.P., J.F.M., and J.A.B. prepared
change significantly after exercise. It is possible that ⬃1 h figures; J.M.P. drafted manuscript; J.M.P., S.J.T., J.F.M., J.A.B., T.L.S.,
exercise was not long enough, and/or the catecholamine re- and D.C.-S. edited and revised manuscript; J.M.P., S.J.T., J.F.M., J.A.B., T.L.S.,
sponse to exercise was not sufficient to induce the release of and D.C.-S. approved final version of manuscript; S.J.T., J.F.M., and T.L.S.
glucagon (13). Either way, it would appear that glucagon performed experiments.
played a minor role (if any) in regulating glucose metabolism
REFERENCES
during exercise in the present study.
Cortisol, growth hormone, and catecholamines stimulate 1. Arner P, Kriegholm E, Engfeldt P, Bolinder J. Adrenergic regulation of
lipolysis during exercise (1, 11, 26). However, because the lipolysis in situ at rest and during exercise. J Clin Invest 85: 893–898,
1990.
plasma concentrations of NEFA were not significantly differ- 2. Bartlett JD, Hwa Joo C, Jeong TS, Louhelainen J, Cochran AJ,
ent after high-intensity intermittent exercise vs. continuous Gibala MJ, Gregson W, Close GL, Drust B, Morton JP. Matched work
moderate-intensity exercise, it is difficult to assign a specific high-intensity interval and continuous running induce similar increases in

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00276.2014 • www.ajpendo.org

 METABOLIC AND HORMONAL RESPONSES TO EXERCISE E551
PGC-1alpha mRNA, AMPK, p38, and p53 phosphorylation in human 25. Jeukendrup AE, Wallis GA. Measurement of substrate oxidation during
skeletal muscle. J Appl Physiol 112: 1135–1143, 2012. exercise by means of gas exchange measurements. Int J Sports Med 26,
3. Bergstrom J, Furst P, Hultman E. Free amino acids in muscle tissue and Suppl 1: S28 –S37, 2005.
plasma during exercise in man. Clin Physiol 5: 155–160, 1985. 26. Kanaley JA, Dall R, Moller N, Nielsen SC, Christiansen JS, Jensen
4. Blomstrand E, Saltin B. Effect of muscle glycogen on glucose, lactate MD, Jorgensen JO. Acute exposure to GH during exercise stimulates the
and amino acid metabolism during exercise and recovery in human turnover of free fatty acids in GH-deficient men. J Appl Physiol 96:
subjects. J Physiol 514: 293–302, 1999. 747–753, 2004.
5. Boluyt MO, Brevick JL, Rogers DS, Randall MJ, Scalia AF, Li ZB. 27. Keller C, Steensberg A, Pilegaard H, Osada T, Saltin B, Pedersen BK,
Changes in the rat heart proteome induced by exercise training: Increased Neufer PD. Transcriptional activation of the IL-6 gene in human con-
abundance of heat shock protein hsp20. Proteomics 6: 3154 –3169, 2006. tracting skeletal muscle: influence of muscle glycogen content. FASEB J
6. Borsheim E, Knardahl S, Hostmark AT. Short-term effects of exercise 15: 2748 –2750, 2001.
on plasma very low density lipoproteins (VLDL) and fatty acids. Med Sci 28. Kiens B, Richter EA. Utilization of skeletal muscle triacylglycerol during
Sports Exerc 31: 522–530, 1999. postexercise recovery in humans. Am J Physiol Endocrinol Metab 275:
7. Burgomaster KA, Howarth KR, Phillips SM, Rakobowchuk M, Mac- E332–E337, 1998.
donald MJ, McGee SL, Gibala MJ. Similar metabolic adaptations 29. Kunesova M, Hlavaty P, Tvrzicka E, Stankova B, Kalouskova P,
during exercise after low volume sprint interval and traditional endurance Viguerie N, Larsen TM, van Baak MA, Jebb SA, Martinez JA,
training in humans. J Physiol 586: 151–160, 2008.
Pfeiffer AF, Kafatos A, Handjieva-Darlenska T, Hill M, Langin D,
8. Curran-Everett D. Multiple comparisons: philosophies and illustrations.

Downloaded from http://ajpendo.physiology.org/ by 10.220.33.6 on June 20, 2017

Zak A, Astrup A, Saris WH. Fatty acid composition of adipose tissue
Am J Physiol Regul Integr Comp Physiol 279: R1–R8, 2000.
triglycerides after weight loss and weight maintenance: the DIOGENES
9. De Feo P, Perriello G, Torlone E, Ventura MM, Fanelli C, Santeusa-
study. Physiol Res 61: 597–607, 2012.
nio F, Brunetti P, Gerich JE, Bolli GB. Contribution of cortisol to
30. Kuo CC, Fattor JA, Henderson GC, Brooks GA. Lipid oxidation in fit
glucose counterregulation in humans. Am J Physiol Endocrinol Metab
257: E35–E42, 1989. young adults during postexercise recovery. J Appl Physiol 99: 349 –356,
10. De Feo P, Perriello G, Torlone E, Ventura MM, Santeusanio F, 2005.
Brunetti P, Gerich JE, Bolli GB. Demonstration of a role for growth 31. Laforgia J, Withers RT, Shipp NJ, Gore CJ. Comparison of energy
hormone in glucose counterregulation. Am J Physiol Endocrinol Metab expenditure elevations after submaximal and supramaximal running. J
256: E835–E843, 1989. Appl Physiol 82: 661–666, 1997.
11. Del Corral P, Howley ET, Hartsell M, Ashraf M, Younger MS. 32. Landaas S, Pettersen JE. Clinical conditions associated with urinary
Metabolic effects of low cortisol during exercise in humans. J Appl excretion of 2-hydroxybutyric acid. Scand J Clin Lab Invest 35: 259 –266,
Physiol 84: 939 –947, 1998. 1975.
12. Enea C, Seguin F, Petitpas-Mulliez J, Boildieu N, Boisseau N, Delpech 33. le Roux C, Sivakumaran S, Alaghband-Zadeh J, Dhillo W, Kong W,
N, Diaz V, Eugene M, Dugue B. 1H NMR-based metabolomics approach Wheeler M. Free cortisol index as a surrogate marker for serum free
for exploring urinary metabolome modifications after acute and chronic cortisol. Ann Clin Biochem 39: 406 –408, 2002.
physical exercise. Anal Bioanal Chem 396: 1167–1176, 2010. 34. Lewis GD, Farrell L, Wood MJ, Martinovic M, Arany Z, Rowe GC,
13. Galbo H, Holst JJ, Christensen NJ. Glucagon and plasma catecholamine Souza A, Cheng S, McCabe EL, Yang E, Shi X, Deo R, Roth FP,
responses to graded and prolonged exercise in man. J Appl Physiol 38: Asnani A, Rhee EP, Systrom DM, Semigran MJ, Vasan RS, Carr SA,
70 –76, 1975. Wang TJ, Sabatine MS, Clish CB, Gerszten RE. Metabolic signatures
14. Gibala MJ, MacLean DA, Graham TE, Saltin B. Anaplerotic processes of exercise in human plasma. Sci Transl Med 2: 33–37, 2010.
in human skeletal muscle during brief dynamic exercise. J Physiol 502: 35. Marker JC, Hirsch IB, Smith LJ, Parvin CA, Holloszy JO, Cryer PE.
703–713, 1997. Catecholamines in prevention of hypoglycemia during exercise in humans.
15. Gibala MJ, MacLean DA, Graham TE, Saltin B. Tricarboxylic acid Am J Physiol Endocrinol Metab 260: E705–E712, 1991.
cycle intermediate pool size and estimated cycle flux in human muscle 36. McGarvey W, Jones R, Petersen S. Excess post-exercise oxygen con-
during exercise. Am J Physiol Endocrinol Metab 275: E235–E242, 1998. sumption following continuous and interval cycling exercise. Int J Sport
16. Gibala MJ, McGee SL. Metabolic adaptations to short-term high-inten- Nutr Exerc Metab 15: 28 –37, 2005.
sity interval training: a little pain for a lot of gain? Exerc Sport Sci Rev 36: 37. McMurray RG, Hackney AC. Interactions of metabolic hormones,
58 –63, 2008. adipose tissue and exercise. Sports Med 35: 393–412, 2005.
17. Gibala MJ, McGee SL, Garnham AP, Howlett KF, Snow RJ, Har- 38. Melanson EL, Sharp TA, Seagle HM, Horton TJ, Donahoo WT,
greaves M. Brief intense interval exercise activates AMPK and p38 Grunwald GK, Hamilton JT, Hill JO. Effect of exercise intensity on
MAPK signaling and increases the expression of PGC-1alpha in human 24-h energy expenditure and nutrient oxidation. J Appl Physiol 92:
skeletal muscle. J Appl Physiol 106: 929 –934, 2009. 1045–1052, 2002.
18. Gibala MJ, Tarnopolsky MA, Graham TE. Tricarboxylic acid cycle 39. Mougios V, Kouidi E, Kyparos A, Deligiannis A. Effect of exercise on
intermediates in human muscle at rest and during prolonged cycling. Am the proportion of unsaturated fatty acids in serum of untrained middle aged
J Physiol Endocrinol Metab 272: E239 –E244, 1997. individuals. Br J Sports Med 32: 58 –62, 1998.
19. Gillen JB, Gibala MJ. Is high-intensity interval training a time-efficient
40. Mourtzakis M, Saltin B, Graham T, Pilegaard H. Carbohydrate me-
exercise strategy to improve health and fitness? Appl Physiol Nutr Metab
tabolism during prolonged exercise and recovery: interactions between
39: 409 –412, 2014.
pyruvate dehydrogenase, fatty acids, and amino acids. J Appl Physiol 100:
20. Glatz JF, Luiken JJ, Bonen A. Membrane fatty acid transporters as
1822–1830, 2006.
regulators of lipid metabolism: implications for metabolic disease. Physiol
Rev 90: 367–417, 2010. 41. Mulla NA, Simonsen L, Bulow J. Post-exercise adipose tissue and
21. Helgerud J, Hoydal K, Wang E, Karlsen T, Berg P, Bjerkaas M, skeletal muscle lipid metabolism in humans: the effects of exercise
Simonsen T, Helgesen C, Hjorth N, Bach R, Hoff J. Aerobic high- intensity. J Physiol 524: 919 –928, 2000.
intensity intervals improve VO2max more than moderate training. Med Sci 42. Nieman DC, Gillitt ND, Henson DA, Sha W, Shanely RA, Knab AM,
Sports Exerc 39: 665–671, 2007. Cialdella-Kam L, Jin F. Bananas as an energy source during exercise: a
22. Henderson GC, Fattor JA, Horning MA, Faghihnia N, Johnson ML, metabolomics approach. PLoS One 7: e37479, 2012.
Mau TL, Luke-Zeitoun M, Brooks GA. Lipolysis and fatty acid metab- 43. Nieman DC, Shanely RA, Luo B, Meaney MP, Dew DA, Pappan KL.
olism in men and women during the postexercise recovery period. J Metabolomics approach to assessing plasma 13- and 9-hydroxy-octadec-
Physiol 584: 963–981, 2007. adienoic acid and linoleic acid metabolite responses to 75-km cycling. Am
23. Hodson L, Skeaff CM, Fielding BA. Fatty acid composition of adipose J Physiol Regul Integr Comp Physiol 307: R68 –R74, 2014.
tissue and blood in humans and its use as a biomarker of dietary intake. 44. Nikolaidis MG, Mougios V. Effects of exercise on the fatty-acid com-
Prog Lipid Res 47: 348 –380, 2008. position of blood and tissue lipids. Sports Med 34: 1051–1076, 2004.
24. Huang CC, Lin WT, Hsu FL, Tsai PW, Hou CC. Metabolomics 45. Nybo L, Sundstrup E, Jakobsen MD, Mohr M, Hornstrup T, Simon-
investigation of exercise-modulated changes in metabolism in rat liver sen L, Bulow J, Randers MB, Nielsen JJ, Aagaard P, Krustrup P.
after exhaustive and endurance exercises. Eur J Appl Physiol 108: 557– High-intensity training versus traditional exercise interventions for pro-
566, 2010. moting health. Med Sci Sports Exerc 42: 1951–1958, 2010.

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00276.2014 • www.ajpendo.org

E552 METABOLIC AND HORMONAL RESPONSES TO EXERCISE

46. Oddy VH, Lindsay DB. Determination of rates of protein synthesis, gain 57. Tjonna AE, Lee SJ, Rognmo O, Stolen TO, Bye A, Haram PM,
and degradation in intact hind-limb muscle of lambs. Biochem J 233: Loennechen JP, Al-Share QY, Skogvoll E, Slordahl SA, Kemi OJ,
417–425, 1986. Najjar SM, Wisloff U. Aerobic interval training versus continuous
47. Palazoglu M, Fiehn O. Metabolite Identification in Blood Plasma Using moderate exercise as a treatment for the metabolic syndrome: a pilot study.
GC/MS and the Agilent Fiehn GC/MS Metabolomics RTL Library. Agilent Circulation 118: 346 –354, 2008.
Technologies, Inc. Accessed June 2014 from https://www.chem.agilent. 58. van Hall G, Saltin B, Wagenmakers AJ. Muscle protein degradation and
com/Library/. . ./5990 –3638en_lo%20CMS.pdf. amino acid metabolism during prolonged knee-extensor exercise in hu-
48. Pechlivanis A, Kostidis S, Saraslanidis P, Petridou A, Tsalis G, mans. Clin Sci (Lond) 97: 557–567, 1999.
Mougios V, Gika HG, Mikros E, Theodoridis GA. 1H NMR-based 59. van Hall G, van der Vusse GJ, Soderlund K, Wagenmakers AJ.
metabonomic investigation of the effect of two different exercise sessions
Deamination of amino acids as a source for ammonia production in human
on the metabolic fingerprint of human urine. J Proteome Res 9: 6405–
skeletal muscle during prolonged exercise. J Physiol 489: 251–261, 1995.
6416, 2010.
49. Pechlivanis A, Kostidis S, Saraslanidis P, Petridou A, Tsalis G, 60. van Loon LJ, Greenhaff PL, Constantin-Teodosiu D, Saris WH,
Veselkov K, Mikros E, Mougios V, Theodoridis GA. 1H NMR study on Wagenmakers AJ. The effects of increasing exercise intensity on muscle
the short- and long-term impact of two training programs of sprint running fuel utilisation in humans. J Physiol 536: 295–304, 2001.
on the metabolic fingerprint of human serum. J Proteome Res 12: 470 – 61. Wahren J, Felig P, Hendler R, Ahlborg G. Glucose and amino acid
480, 2013. metabolism during recovery after exercise. J Appl Physiol 34: 838 –845,
50. Pedersen BK. IL-6 signalling in exercise and disease. Biochem Soc Trans 1973.

Downloaded from http://ajpendo.physiology.org/ by 10.220.33.6 on June 20, 2017

35: 1295–1297, 2007. 62. Wan Z, Ritchie I, Beaudoin MS, Castellani L, Chan CB, Wright DC.
51. Pohjanen E, Thysell E, Jonsson P, Eklund C, Silfver A, Carlsson IB, IL-6 indirectly modulates the induction of glyceroneogenic enzymes in
Lundgren K, Moritz T, Svensson MB, Antti H. A multivariate screening adipose tissue during exercise. PLoS One 7: e41719, 2012.
strategy for investigating metabolic effects of strenuous physical exercise 63. Warren A, Howden EJ, Williams AD, Fell JW, Johnson NA. Postex-
in human serum. J Proteome Res 6: 2113–2120, 2007. ercise fat oxidation: effect of exercise duration, intensity, and modality. Int
52. Romijn JA, Coyle EF, Sidossis LS, Gastaldelli A, Horowitz JF, Endert J Sport Nutr Exerc Metab 19: 607–623, 2009.
E, Wolfe RR. Regulation of endogenous fat and carbohydrate metabolism 64. Weston KS, Wisloff U, Coombes JS. High-intensity interval training in
in relation to exercise intensity and duration. Am J Physiol Endocrinol patients with lifestyle-induced cardiometabolic disease: a systematic re-
Metab 265: E380 –E391, 1993. view and meta-analysis. Br J Sports Med 48: 1227–1234, 2014.
53. Smart KF, Aggio RB, Van Houtte JR, Villas-Boas SG. Analytical 65. Wisloff U, Stoylen A, Loennechen JP, Bruvold M, Rognmo O, Haram
platform for metabolome analysis of microbial cells using methyl chloro- PM, Tjonna AE, Helgerud J, Slordahl SA, Lee SJ, Videm V, Bye A,
formate derivatization followed by gas chromatography-mass spectrome-
Smith GL, Najjar SM, Ellingsen O, Skjaerpe T. Superior cardiovascu-
try. Nat Protoc 5: 1709 –1729, 2010.
lar effect of aerobic interval training versus moderate continuous training
54. Spencer MK, Katz A, Raz I. Epinephrine increases tricarboxylic acid
cycle intermediates in human skeletal muscle. Am J Physiol Endocrinol in heart failure patients: a randomized study. Circulation 115: 3086 –3094,
Metab 260: E436 –E439, 1991. 2007.
55. Struder HK, Hollmann W, Platen P, Wostmann R, Ferrauti A, Weber 66. Wolsk E, Mygind H, Grondahl TS, Pedersen BK, van Hall G. IL-6
K. Effect of exercise intensity on free tryptophan to branched-chain amino selectively stimulates fat metabolism in human skeletal muscle. Am J
acids ratio and plasma prolactin during endurance exercise. Can J Appl Physiol Endocrinol Metab 299: E832–E840, 2010.
Physiol 22: 280 –291, 1997. 67. Yan B, AJ, Wang G, Lu H, Huang X, Liu Y, Zha W, Hao H, Zhang
56. Thomas TR, Adeniran SB, Etheridge GL. Effects of different running Y, Liu L, Gu S, Huang Q, Zheng Y, Sun J. Metabolomic investigation
programs on VO2 max, percent fat, and plasma lipids. Can J Appl Sport Sci into variation of endogenous metabolites in professional athletes subject to
9: 55–62, 1984. strength-endurance training. J Appl Physiol 106: 531–538, 2009.

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00276.2014 • www.ajpendo.org