0021-972X/05/$15.00/0
Printed in U.S.A.
The Journal of Clinical Endocrinology & Metabolism 90(11):6218 – 6225
Copyright © 2005 by The Endocrine Society
doi: 10.1210/jc.2005-0926
17␤-Estradiol Supplementation Decreases Glucose Rate
of Appearance and Disappearance with No Effect on
Glycogen Utilization during Moderate Intensity Exercise
in Men
Michaela C. Devries, Mazen J. Hamadeh, Terry E. Graham, and Mark A. Tarnopolsky
Departments of Pediatrics and Medicine (M.C.D., M.J.H., M.A.T.), McMaster University, Hamilton, Ontario, Canada L8Z
3N5; and Department of Human Biology and Nutritional Sciences (T.E.G.), University of Guelph, Guelph, Ontario, Canada
N1G 2W1
Context and Objective: Women use less carbohydrate during endurance exercise, as compared with men. In rodents, 17␤-estradiol
(E2) supplementation robustly increases lipid use and lowers muscle
and liver glycogen use during exercise. E2 supplementation has been
found to influence substrate selection by decreasing glucose rate of
appearance (Ra), disappearance (Rd), and metabolic clearance rate
during exercise in humans; however, neither a change in total carbohydrate use nor a sparing of muscle glycogen was demonstrated.
Subjects and Methods: We investigated the effect of 8 d of E2 (2
mg/d) supplementation on glucose turnover and net muscle glycogen
use in 11 men using a randomized, double-blind, placebo-controlled,
crossover design. Subjects underwent primed constant infusion of
[6,6-2H]glucose, and muscle biopsies were taken before and after 90
min of cycling at 65% maximal oxygen uptake.
M
ANY (1–7), BUT NOT all (8, 9), studies have reported
that women oxidize less carbohydrate (CHO) and
more lipid as compared with men during exercise, as evidenced by a lower respiratory exchange ratio (RER). Differences in body composition and training status between men
and women may account for the observed differences in fuel
selection during exercise; however, when men and women
are matched based on these criteria, gender differences remain (1–7, 10). It has been suggested that the sex hormone
17␤-estradiol (E2) may play a role in the observed gender
differences in substrate selection during exercise (11, 12).
Results from studies in rats have shown that E2 supplementation robustly influences substrate selection during endurance exercise (12–16). Administration of E2 to male or
oophorectomized female rats resulted in improved exercise
performance (12, 15), sparing of muscle and liver glycogen
during exercise (12, 15, 16), increased free fatty acid (FFA)
availability for oxidation (12, 14, 16), elevated intramyocelFirst Published Online August 23, 2005
Abbreviations: CHO, Carbohydrate; E2, 17␤-estradiol; FFA, free fatty
acid; FP, follicular phase; GC, gas chromatography; Gtot, total glycogen;
IMCL, intramyocellular lipid; LP, luteal phase; MCR, metabolic clearance rate; MG, macroglycogen; MS, mass spectrometry; PG, proglycogen; Ra, rate of appearance; Rd, rate of disappearance; RER, respiratory
exchange ratio; VO2max, maximal oxygen uptake.
JCEM is published monthly by The Endocrine Society (http://www.
endo-society.org), the foremost professional society serving the endocrine community.
Results: E2 supplementation decreased the respiratory exchange
ratio (P ⫽ 0.03) and glucose Ra and Rd (both P ⫽ 0.04) during exercise,
as compared with placebo. E2 supplementation lowered proglycogen
(P ⬍ 0.05) and total glycogen (P ⫽ 0.04) concentration, as compared
with placebo; however, there was no effect of E2 on net muscle glycogen use during exercise.
Conclusions: These findings show that E2 supplementation alters
fuel selection in exercising men by increasing lipid use and reducing
carbohydrate use, glucose Ra (primarily liver glucose production), and
Rd (primarily muscle glucose uptake). Furthermore, E2 reduces the
basal level of total muscle glycogen, particularly the proglycogen
form. (J Clin Endocrinol Metab 90: 6218 – 6225, 2005)
lular lipid (IMCL) concentration (13), and decreased lactate
concentration (12). Furthermore, E2 administration was
found to increase skeletal muscle lipoprotein lipase activity
(13, 18 –20). These effects were found to be dose dependent
but inducible with physiological doses of E2 (15). The results
from the aforementioned studies suggest that administration
of E2 to rats induces a shift in substrate selection resulting in
a decreased dependence on carbohydrate stores and an increased reliance on lipid stores during exercise.
Human trials with E2 administration have shown mixed
results. Ruby et al. (21) found that E2 administration to amenorrheic women resulted in higher FFA concentrations at
rest, lower glucose rate of appearance (Ra) and disappearance (Rd) with no difference in CHO or lipid use during
exercise. We found that E2 administration to young men
decreased glucose Ra and Rd, glucose metabolic clearance
rate (MCR) and heart rate, with an increased plasma glucose
concentration (22). Neither study found a change in wholebody substrate oxidation (21, 22), implying that there must
be a paradoxical (at least compared with the data in animals)
increased reliance on muscle glycogen stores to fuel exercise.
However, our laboratory has also shown that E2 administration did not result in any measurable change in muscle
glycogen use (23). In addition, we found no change in plasma
triglycerides, glycerol, glucose, or lactate concentrations,
with a trend toward a shift in fuel use (lower RER) after E2
administration (23). A type II error due to small sample size
6218
Devries et al. • Effects of 17-␤-Estradiol on Carbohydrate Metabolism
or low dose could be a factor in the aforementioned mixed
findings in the human studies (22, 23).
In previous studies involving E2 administration, either a
transdermal E2 patch (21, 23) or an oral E2 tablet (22) was
used to increase plasma E2 levels. In the studies that used the
transdermal patch, plasma E2 levels doubled (21, 23), mimicking E2 levels seen in the follicular phase (FP) of the menstrual cycle. However, these levels were well below those
observed in the animal studies. Consequently, Carter et al.
(22) used an oral tablet to achieve higher plasma E2 levels,
similar to those seen in the luteal phase (LP) of the menstrual
cycle. However, because this was the first study using the E2
tablet, it was not known what dosage to administer to obtain
such levels, and the plasma E2 levels that were reached were
well above what would normally be observed in the LP of the
menstrual cycle (22).
Previous studies attempting to determine the effect of E2
supplementation on glycogen metabolism have investigated
E2 effect only on either glucose turnover or muscle glycogen
use during exercise, not both. To date, no study has been
conducted using both stable isotope tracers and muscle samples to investigate the effect of E2 supplementation on wholebody and muscle glucose/glycogen metabolism. The purpose of the present study was to determine the effect of E2
administration on whole-body glucose turnover and skeletal
muscle glycogen use during moderate intensity endurance
exercise in young men. Furthermore, we used a dosage that
more closely mimics E2 levels observed in the LP, which is
higher than that used previously (23), and increased the
sample size as compared with our previous research (22) to
determine whether previously observed trends were artifacts of type II error. In addition, we investigated the effect
of E2 administration on the proglycogen (PG) and macroglycogen (MG) fractions of glycogen as the effects of E2 on
muscle glycogen use and storage may be more apparent in
the PG form because it is more dynamic (24, 25), and failure
to measure both forms may miss a shift in glycogen use
and/or storage.
Subjects and Methods
Participants
Eleven healthy, recreationally active male volunteers participated in
this study [age 23 ⫾ 2 yr (mean ⫾ sd); weight 80 ⫾ 3 kg; height 178 ⫾
1 cm; body mass index 25 ⫾ 1 kg/m2; fat-free mass 64 ⫾ 2 kg (80 ⫾ 1%
of body weight); maximal oxygen uptake (VO2max) 44 ⫾ 2 ml O2 per
kilogram body weight per minute]. Informed consent was obtained
before commencing the study following a description of the study and
advisement of the risks and benefits of participation. This study was
conducted under approval of the Research Ethics Committee of McMaster University.
Protocol
At least 1 wk before the first trial date, subjects performed a progressive exercise test on an electronically braked bike (Excalibur Sport,
Lode, Groningen, The Netherlands) to determine their VO2max, as previously described (26). The VO2max was used to determine the work
intensity needed to elicit 65% of the subject’s VO2max for subsequent
testing and was confirmed by measuring oxygen uptake at the calculated
workload intensity for each subject approximately 30 min after the test.
Subjects were randomly assigned to receive either placebo (400 mg/d
glucose polymer, Polycose; Abbott Laboratories, Ross Division, St. Laurent, Québec, Canada) or E2 [1 mg/d for 2 d and 2 mg/d for 6 d (Estrace;
J Clin Endocrinol Metab, November 2005, 90(11):6218 – 6225
6219
Shire BioChem, Inc., St. Laurent, Québec, Canada) ⫹ ⬃300 mg Polycose]
for 8 d in a randomized, double-blind, crossover manner. On the morning of the ninth day, subjects reported to the laboratory and performed
90 min of exercise on a cycle ergometer at 65% VO2max. At least a 10-d
washout period was allowed between trials for the hormone levels to
return to baseline. After the washout period, the subjects repeated the
9-d protocol on the alternate treatment. The tablets were placed in
gelatin-filled capsules and filled with a glucose polymer and taken once
per day. Subjects were instructed to take the pills at the same time each
day and return any unused pills. All subjects reported 100% compliance.
During the 8-d dosage period, subjects were asked to maintain and
record their normal activity level. Food intake was recorded in all subjects for at least 3 d of the dosage period (minimum of 2 weekdays and
1 weekend day) and was returned to the subjects during the second arm
of the study to allow them to keep their diet similar between trials. Three
subjects completed diet records on both arms of the study in order to
ensure that habitual diet did not change between trials. Diet records
were analyzed using commercially available analysis software (Nutritionist Pro, version 2.2; First DataBank, Inc., San Bruno, CA). Subjects
were not tested on or around major holidays or during times when their
diets deviated from normal.
Subjects consumed the same meal on the evening before both test
days. On the morning of the test day, subjects reported to the laboratory
10 –12 h postabsorptive. Body weight was recorded and body composition was determined via bioelectric impedance analysis (RJL Systems
BIA-101A, Mt. Clemens, MI). A 20-gauge plastic catheter (Becton Dickinson, Lincoln Park, NJ) was placed into the antecubital vein of the right
arm to allow for blood sampling for the duration of the testing. To
arterialize the blood, the arm was placed in a heating pad (65 ⫾ 5 C) 15
min before the withdrawal of each blood sample when the subjects were
not exercising. Another plastic catheter was placed into the antecubital
vein of the left arm to allow for infusion of stable isotopes with a constant
infusion pump (model 74900; Cole-Parmer Instrument Co., Vernon
Hills, IL). Subjects then underwent primed constant infusion of [6,62
H]glucose (99% enriched; Cambridge Isotope Laboratories, Inc., Cambridge, MA) for 2.5 h. Glucose was mixed with 0.9% saline and filtered
through a 0.2-␮m filter (sterile, nonpyrogenic, Acrodisc; Pall Gelman
Corp., Ann Arbor, MI) into single-use vials at the McMaster University
Medical Center pharmacy. The solutions were cultured for 5 d at room
temperature and 35 C to ensure there was no bacterial contamination.
Before initiating the infusion protocol, a baseline blood sample was
taken to allow for determination of the natural background enrichment
of [6,6-2H]glucose. A priming dose of [6,6-2H]glucose (17 ␮mol/kg) was
given followed immediately by a constant infusion at a rate of 0.22
␮mol/kg䡠min for 60 min to reach steady state. At the onset of exercise,
the infusion rate was increased in a stepwise fashion at t ⫽ 0, 5, and 10
min of exercise to 0.33 ␮mol/kg䡠min, 0.44 ␮mol/kg䡠min, and 0.55 ␮mol/
kg䡠min. The infusion rate remained at 0.55 ␮mol/kg䡠min for the remainder of the 90-min exercise session.
Before and after exercise, a muscle biopsy was taken under local
anesthetic from the vastus lateralis muscle approximately 20 cm proximal to the knee joint using a modified Bergström needle (5 mm diameter) with suction modification, as previously described (23, 27). The
vastus lateralis muscle was chosen because it is one of the four quadriceps muscles, which is one of the primary muscles used during cycling
activity and has been used in similar studies (23, 27). Biopsies were taken
from the same leg before and after exercise and from the contralateral
leg on the subsequent test day. The site of the second biopsy was
approximately 3 cm above the first biopsy site to ensure homogeneity
for muscle sampling source. Approximately 30 mg was immediately
snap frozen and stored in liquid nitrogen until transferred to ⫺86 C for
subsequent PG/MG analyses.
Blood samples were taken at baseline and 15 min before commencing
exercise and at t ⫽ 60, 75, and 90 min of exercise. For glucose and lactate
analysis, blood samples were collected in heparinized tubes, placed on
ice, centrifuged at 1750 ⫻ g at 4 C for 10 min and stored at ⫺50 C for
subsequent analysis. For determination of hormone levels (E2, testosterone, progesterone), blood samples were collected and allowed to
stand for 30 min in untreated tubes, centrifuged at 1200 ⫻ g at 4 C for
30 min and the serum stored at ⫺50 C for subsequent analysis.
Respiratory measures (oxygen uptake, volume of carbon dioxide
expired, RER) were taken at t ⫽ 0, 5, 30, 60, 75, and 90 min during exercise
using a computerized open-circuit gas collection system (Moxus Mod-
6220
J Clin Endocrinol Metab, November 2005, 90(11):6218 – 6225
Devries et al. • Effects of 17-␤-Estradiol on Carbohydrate Metabolism
ulator VO2 system with O2 analyzer S-3A/I and CO2 analyzer CD-3A;
AEI Technologies, Inc., Pittsburgh, PA).
where C1 and C2 are plasma glucose concentrations at sampling times
1 and 2.
Biochemical glycogen determination
Blood samples
Muscle PG and MG content was analyzed as described by Adamo and
Graham (24). Briefly, snap-frozen muscle samples were freeze dried for
24 h, powdered, and dissected free of any blood and connective tissue
and weighed. Ice-cooled perchloric acid (200 ␮l, 1.5 m) was added to
1.5–2.5 mg tissue and pressed with a plastic inoculating loop for 20 min
on ice. Samples were then centrifuged for 15 min at 3000 rpm, and an
aliquot of 100 ␮l of the supernatant was removed for MG determination.
The remaining supernatant was aspirated off. HCl (1 ml, 1 m) was added
to each sample, and PG samples was briefly pressed and MG samples
were vortexed for several seconds. Samples were hydrolyzed for 2 h at
100 C and neutralized with 2 m Tris base, vortexed, centrifuged at 3000
rpm for 5 min, and stored at ⫺86 C until subsequent determination of
glucosyl units. Muscle PG and MG content was determined as described
previously (28). Values for PG and MG were then added to give total
muscle glycogen content. Muscle glycogen use during exercise was
calculated as follows:
Plasma was analyzed for lactate and glucose and serum for estrogen,
testosterone, and progesterone levels. Plasma lactate and glucose concentrations were analyzed with an automated lactate and glucose analyzer (2300 STAT plus; YSI, Farnborough, UK). Serum E2, testosterone,
and progesterone were analyzed using a single-incubation RIA (Coata-Count, kit TKE21, kit TKTE1 and kit TKIN5; Diagnostics Products
Corp., Los Angeles, CA).
glycogen use ⫽ [glycogen]pre ⫺ [glycogen]post
Isotopic enrichment, glucose Ra and Rd, and MCR
Isotopic enrichment of glucose was determined using gas chromatography (GC)-mass spectrometry (MS; GC model 6890; Hewlett-Packard, Palo Alto, CA; and MS model 5973; Agilent Technologies, Inc., Palo
Alto, CA) of the pentaacetate derivative, as described previously (22). To
isolate the pentaacetate derivative, the samples were first deproteinized
using barium hydroxide and zinc sulfate. The supernatant was passed
through an anion-cation (AG 1X8 – 400 and AG 50X8 – 400; Sigma Chemicals, St. Louis, MO) exchange column, and the eluted extract was evaporated using a rotary evaporator (SpeedVac Plus SC210A; ThermoSavant, Holbrook, NY). The derivative was prepared for GCMS analysis
by adding 100 ␮l of 2:1 acetic anhydride and pyridine solution to each
sample.
A 15-m fused silica capillary column with 0.2 mm diameter and 0.2
␮m film thickness (Supelco, Bellefonte, PA) was used in the GC oven.
The initial oven temperature was set at 105 C and increased 45 C/min
until a final temperature of 250 C was reached. The oven was held at this
temperature for 3 min and then increased 50 C/min until a final temperature of 300 C was reached. The carrier gas was helium and the
split-pulse injection flow rate was set at 25 ml/min. The source temperature was set at 230 C and the quadrupole temperature was set at 150
C. For each sample, 1 ␮l was injected into the GC oven. To monitor
selected ions, a mass-charge ratio of 200 and 202 atomic mass units was
used within the electron impact ionization mode to perform mass
analysis.
Glucose Ra and Rd were calculated using the non-steady-state Steele
equation (29) as modified by Romijn et al. (30) for use with stable isotopes
because the amount of tracer infused is no longer considered negligible.
The modified Steele equation is as follows:
Cm dE
F⫺Vd
1 ⫹ E dt
Ra ⫽
E
dE
dCm
共1 ⫹ E兲⫺Cm
Vd
dt
dt
Rd ⫽ Ra⫺
共1 ⫹ E兲2
where F is the infusion rate, Cm is the measured glucose concentration,
E is the enrichment, t is the time, and Vd is the volume of distribution,
assumed to be 100 ml/kg.
Glucose Ra and Rd were calculated at rest and at 60, 75, and 90 min
during exercise. The average of the exercise time points were combined
and reported as the average Ra and Rd over the exercise bout. MCR was
determined as follows:
MCR (milliliters per kilogram per minute) ⫽ Rd/[(C1 ⫹ C2)/2]
Statistical analysis
For RER and indirect calorimetry, rest samples were analyzed using
paired t tests. For all other analyses, including RER and indirect calorimetry exercise time points, two-way, repeated-measures ANOVA with
treatment and time being the experimental variables was used. Onetailed tests were used for RER, glucose oxidation, and substrate oxidation because we hypothesized a priori that E2 supplementation would
decrease RER and carbohydrate and glucose oxidation and increase lipid
oxidation. When significance was attained, Tukey’s honestly significant
difference post hoc test was used to determine the location of the difference. Analyses were performed using a computerized statistics program (STATISTICA for Windows, version 5.1; StatSoft, Tulsa, OK).
Statistical significance was set at P ⱕ 0.05. Data are presented as means ⫾
sem unless otherwise indicated. All data were checked for normality
using the Kolmogorov-Smirnov test.
Results
Diet and body weight
Subjects consumed 2,416 ⫾ 331 kcal (mean ⫾ sd, 10,080 ⫾
1,381 kJ), 298 ⫾ 69 g of CHO, 94 ⫾ 28 g of protein, and 87 ⫾
18 g of fat per day. CHO, protein, and fat made up 49 ⫾ 2,
16 ⫾ 1, and 33 ⫾ 2% of total dietary intake, respectively. The
caloric intake and macronutrient composition of the three
subjects who completed diet records on both arms of the
study did not change between trials. Body weight did not
change between trials.
Hormone concentrations
E2 supplementation for 8 d significantly increased serum
E2 concentrations (P ⬍ 0.001) and significantly decreased
serum testosterone (P ⫽ 0.001) and serum progesterone (P ⫽
0.03) (Table 1).
RER and lipid and carbohydrate metabolism
Data from one subject were not included in the statistical
analysis due to an error that occurred when taking breath
samples; hence, data presented are for 10 subjects. RER,
CHO, and lipid oxidation increased during exercise, as compared with rest (P ⬍ 0.001). E2 supplementation decreased
TABLE 1. Serum hormone concentrations for 11 men after 8 d of
placebo or E2 supplementation
Treatment
Placebo
E2
Total testosterone (nmol/liter)
E2 (pmol/liter)
Progesterone (nmol/liter)
21 ⫾ 1
129 ⫾ 13
3.0 ⫾ 0.3
16 ⫾ 1a
946 ⫾ 167a
2.3 ⫾ 0.2b
Data are means ⫾ SEM, n ⫽ 11. Repeated-measures ANOVA: E2 is
significantly different from placebo.
a
P ⬍ 0.002.
b
P ⫽ 0.03.
Devries et al. • Effects of 17-␤-Estradiol on Carbohydrate Metabolism
RER (P ⫽ 0.03) during exercise, as compared with placebo
(Fig. 1). E2 supplementation also decreased CHO oxidation
(P ⫽ 0.03) and increased lipid oxidation (P ⫽ 0.03) (Fig. 1).
There was no difference in energy expenditure between
trials.
Plasma glucose and lactate concentrations
Plasma glucose concentration was lower at t ⫽ 75 and 90
min exercise as compared with the resting levels (P ⫽ 0.01)
(Table 2). Plasma lactate concentration was higher during
exercise as compared with rest (P ⬍ 0.001). There was no
effect of E2 supplementation on plasma glucose or lactate
concentrations at rest or during exercise.
Glucose Ra and Rd and MCR
J Clin Endocrinol Metab, November 2005, 90(11):6218 – 6225
6221
TABLE 2. Plasma glucose and lactate concentrations in 11 men
at rest and during 90 min of cycling at 65% VO2max
Glucose (mmol/liter)
Estrogen
Placebo
Lactate (mmol/liter)
Estrogen
Placebo
Data are means ⫾
0 min
60 min
75 min
90 min
4.9 ⫾ 0.1
5.1 ⫾ 0.1
4.7 ⫾ 0.1
4.8 ⫾ 0.2
4.5 ⫾ 0.1
4.7 ⫾ 0.1
4.5 ⫾ 0.2
4.7 ⫾ 0.2
0.9 ⫾ 0.1
1.0 ⫾ 0.1
2.3 ⫾ 0.3
2.5 ⫾ 0.3
2.4 ⫾ 0.4
2.4 ⫾ 0.2
2.3 ⫾ 0.4
2.6 ⫾ 0.3
SEM,
n ⫽ 11. Repeated-measures ANOVA.
during exercise, as compared with rest (P ⬍ 0.001) (Fig. 2).
E2 supplementation decreased glucose Ra (P ⫽ 0.04) and Rd
(P ⫽ 0.04) during exercise with no effect on MCR, as compared with placebo (Table 3).
Data from a different subject were not included in the
statistical analysis due to technical problems with the infusion on one of the test days; hence, data are presented for 10
subjects. Glucose Ra and Rd were equivalent, indicating that
subjects were in steady-state glucose flux for the duration of
the exercise bout. Glucose Ra and Rd and MCR all increased
Muscle PG, MG, and total glycogen (Gtot) concentration
decreased significantly during 90 min of endurance exercise
at 65% VO2max (P ⬍ 0.001) (Fig. 3). E2 supplementation
FIG. 1. RER (A), glucose (B), and lipid oxidation (C) for 10 men at rest
and during 90 min of cycling at an intensity of 65% VO2max on placebo
(PL, f) or after 8 d of E2 supplementation (䡺). Data are means ⫾ SEM.
*, Rest lower than exercise, P ⬍ 0.001. †, E2 lower (A and B) and
higher (C) than PL during exercise, P ⫽ 0.03.
FIG. 2. Glucose Ra (A), Rd (B), and MCR (C) in 10 men at rest (pre)
and during 90 min of cycling (EX) at an intensity of 65% VO2max on
placebo (PL, f) or after 8 d of E2 supplementation (䡺). Data are
means ⫾ SEM. *, Rest lower than exercise, P ⬍ 0.001. †, E2 lower than
PL during exercise, P ⫽ 0.04.
Muscle glycogen concentration and use
6222
J Clin Endocrinol Metab, November 2005, 90(11):6218 – 6225
FIG. 3. Muscle PG (A), MG (B), and Gtot in 11 men at rest (pre) and
after 90 min of cycling at an intensity of 65% VO2max on placebo (PL,
f) or after 8 d of E2 supplementation (䡺). Data are means ⫾ SEM. *,
Rest higher than after exercise, P ⬍ 0.001. †, E2 lower than PL, P ⬍
0.05. DW, Dry weight.
significantly lowered PG by 18% (P ⬍ 0.05) and Gtot by 20%
(P ⫽ 0.04) with a strong trend toward a lowered MG (24%,
P ⫽ 0.059) at rest and after exercise, as compared with placebo (Table 3). E2 supplementation had no effect on PG
(172 ⫾ 17 vs. 191 ⫾ 8, E2 vs. placebo), MG (118 ⫾ 27 vs. 113 ⫾
8), or Gtot (290 ⫾ 42 vs. 304 ⫾ 8) use.
Discussion
The main findings of this study were that 8 d of E2 supplementation lowered RER, glucose Ra and Rd, and muscle
PG and total glycogen concentration with no effect on glucose MCR or net muscle glycogen use. These findings suggest that E2 influences CHO use during endurance exercise
through an effect on glucose turnover (hepatic glucose production and peripheral uptake). E2 supplementation also
reduced basal muscle glycogen stores but not net muscle use
during exercise.
The E2 supplementation regimen used in the current study
was somewhat different from that used in studies conducted
Devries et al. • Effects of 17-␤-Estradiol on Carbohydrate Metabolism
previously. Two previous studies used a patch delivery
method to administer E2 to subjects (21, 23), and serum E2
concentrations varied dramatically between studies due in
part to differences in the duration of supplementation. In one
study, a 72- and 144-h E2 supplementation regimen to amenorrheic women significantly increased serum E2 concentrations (21), but the levels were still well below what is
typically observed in the FP of the menstrual cycle (31–33).
The second study used a patch delivery method during an
11-d protocol that resulted in serum E2 concentrations between what is typically observed in the late FP and early LP
of the menstrual cycle (23). Another investigation used a
novel delivery method that resulted in serum E2 concentrations five times the values typically observed in the LP of the
menstrual cycle (22). With the oral E2 delivery method used
in the current study (dose and duration), the serum E2 concentrations were at the upper limit of the range for the LP of
the menstrual cycle and less than the supraphysiological
levels seen in our previous study (22) and was enough to
induce metabolic changes in line with previous research
(21–23).
There were also obvious gender differences in the amenorrheic women recruited in the aforementioned studies for
Ruby et al. (21), whereas we used males in our previous
studies (22) and in the current one. Because it is likely that
any metabolic effects of E2 are to be receptor mediated, it is
important to note that E2 receptor-␤ mRNA and protein are
present in the skeletal muscle of men (34, 35). The only study
that we are aware of that has looked at E2 receptor mRNA
and protein content in men and women used a small number
of subjects (three men, three women), and thus no gender
comparisons were conducted (35), although, subjective examination of the presented data suggested that the men had
consistently lower E2 receptor mRNA content and percent E2
receptor-positive nuclei (35). Consequently, it is likely that
higher levels of serum E2 are required in studies using men
before a change in substrate oxidation is seen. This could
partially explain why we did not find changes in exercise
metabolism when we supplemented men with an E2 dose
that resulted in much lower serum E2 concentrations than
those found in the current study (23).
In previous animal studies, E2 was given to male (13, 15,
16) or oophorectomized female rats (12, 14). E2 supplementation decreased the reliance on CHO and increased reliance
on lipid during endurance exercise, regardless of whether
male or female mice were used (12–16). Essentially, oophorectomized rats are the animal equivalent to amenorrheic
women because both models lack ovarian hormones. A concern associated with amenorrheic women as subjects is that
their energy intake is lower, as compared with eumenorrheic
women (37). Lower energy intake results in lower body
stores and may alter substrate selection during exercise. For
example, lower caloric intake results in lower glycogen
stores, which in turn promotes lipid oxidation during exercise (38). As such, it may be more appropriate to use males
as subjects to determine the effects of E2 administration per
se on substrate use during exercise. The only potentially
confounding issue with using males in E2 supplementation
studies is the effect testosterone may have on substrate selection during exercise. In this and other studies (22, 23), E2
Devries et al. • Effects of 17-␤-Estradiol on Carbohydrate Metabolism
J Clin Endocrinol Metab, November 2005, 90(11):6218 – 6225
6223
TABLE 3. Glucose Ra, Rd, and MCR and muscle proglycogen, MG, and total glycogen values before and after 90 min of cycling at 65%
VO2max
E2
Glucose Ra (mg/kg䡠min)
Glucose Rd (mg/kg䡠min)
Glucose MCR (ml/min)
PG (mmol/kg DW)
MG (mmol/kg DW)
Total glycogen (mmol/kg DW)
Placebo
Pre
During/post
Pre
During/post
22.6 ⫾ 0.6a
22.6 ⫾ 0.7a
4.7 ⫾ 0.2a
279 ⫾ 22c
152 ⫾ 29
431 ⫾ 48c
56.6 ⫾ 3.1b
56.6 ⫾ 3.1b
12.8 ⫾ 0.9
107 ⫾ 17c,d
34 ⫾ 8d
141 ⫾ 23c,d
22.0 ⫾ 0.4a
21.9 ⫾ 0.4a
4.3 ⫾ 0.1a
330 ⫾ 24
179 ⫾ 29
510 ⫾ 45
62.1 ⫾ 3.2
62.1 ⫾ 3.2
13.4 ⫾ 0.6
140 ⫾ 24d
66 ⫾ 21d
206 ⫾ 43d
Data are means ⫾ SEM. DW, Dry weight.
a
Repeated-measures ANOVA, pre lower than exercise, P ⬍ 0.001.
b
Repeated-measures ANOVA, E2 lower than placebo, P ⫽ 0.04.
c
Repeated-measures ANOVA, E2 lower than placebo, P ⬍ 0.05.
d
Repeated-measures ANOVA, post exercise lower than pre, P ⬍ 0.001.
supplementation in men resulted in a significant lowering of
testosterone; however, the levels were still well above female
levels and within the normal range for adult males. However,
in the animal studies, both male and oophorectomized female rats were used, conditions characterized by high and
low levels of testosterone, and no differences in liver or
muscle glycogen use, performance, or any metabolic blood
parameters were observed (12, 14 –16). Most importantly, a
recent study involving human subjects found that pharmacologic manipulation of testosterone in men had no influence
on substrate metabolism (39). As such, it is most likely that
the increase in E2 is the predominant hormonal factor that
caused the observed shift in fuel selection during exercise in
the current study.
There may be a dose-response effect of E2 on substrate use
and storage. In previous studies, E2 had no effect on RER
(21–23). However, in two of these studies (21, 23), the E2 dose
that was administered was lower than what was used in the
current study and may explain the lack of change in RER.
However, no change in RER was observed in the study
conducted by Carter et al. (22), even though serum E2 concentrations were well above those observed in the current
study. One possible explanation could be the relatively small
sample size (type II error) used in that study (22). Another
possible explanation could be that E2 is known to increase
IMCL content and lipoprotein lipase activity in rats (17, 19,
20). Given that there is simultaneous esterification of FFA to
IMCL and IMCL hydrolysis during exercise (40) and that an
increase in IMCL synthesis would tend to elevate the RER,
it is possible that the RER could be falsely elevated with E2.
Consequently, it is possible that in the study conducted by
Carter et al. (22), the dosing regimen was so high that the
effects of E2 on lipid synthesis exceeded those on lipid mobilization leading to an inability to detect an effect of E2 on
lowering RER.
Our study is the first to measure muscle glycogen and use
stable isotope tracers in the same subjects to study the effects
of E2 supplementation on whole-body and muscle CHO
metabolism during exercise. We found that 8 d of E2 supplementation decreased RER during exercise, indicating a
decrease in CHO use and an increase in lipid use. There was
no effect of E2 supplementation on muscle glycogen use;
however, muscle glycogen levels were lowered after supplementation. This finding may suggest that to elicit a shift
in muscle glycogen use, there first needs to be a shift in fuel
storage. Support for this theory comes from a recent study
that showed that after a low CHO diet designed to reduce
muscle glycogen content, leg lipid oxidation during exercise
increased by 122% (41), suggesting that lowering muscle
glycogen content promotes an increase in lipid oxidation
during exercise. Second, it has been shown that a high muscle
glycogen concentration inhibits lipid oxidation during exercise (41). Hence, in our study the slight, yet significant, decrease in muscle glycogen stores after E2 supplementation
could have further enhanced lipid oxidation and promoted
muscle glycogen sparing had the supplementation regimen
continued. Concomitant to this lowering in muscle glycogen
storage, it is possible that an increase in IMCL content is also
required before a shift in substrate use becomes quantifiable.
Animal studies have shown that E2 supplementation initially
increased fat oxidation followed by a subsequent decrease in
glucose oxidation during exercise (14). Furthermore, Rooney
et al. (16) found that E2 supplementation in rats increased
resting IMCL concentration causing a subsequent downregulation of CHO use during exercise. These two studies
further support that an initial change in fat availability and
oxidation needs to occur before a shift in CHO use is observed after E2 supplementation. Future studies should investigate the effect of E2 supplementation on IMCL content
and whether an increase in lipid content is a requisite to
promote lipid use and whether a longer E2 dosing regimen
promotes muscle glycogen sparing during exercise.
Ours is the first study to investigate the effect of E2 supplementation on muscle PG, MG, and Gtot storage and use.
Previous studies have failed to find a reduction in muscle
glycogen use after E2 treatment (23); therefore, we decided
to investigate whether there was a fraction-specific decrease
in glycogen use that was not translated to whole-muscle
glycogen use. We anticipated a decrease in muscle PG use if
there was an effect of E2 on muscle glycogen fraction use
because the PG fraction is more dynamic than the MG fraction (24, 25). Interestingly, the only effect of E2 on muscle
glycogen was that 8 d of E2 supplementation decreased PG
and Gtot with a strong trend toward a decrease in MG stores.
However, it is likely that we lacked sufficient power due to
the high variability in MG concentrations to detect a significant decrease in MG stores because such a strong trend (24%,
P ⫽ 0.059) was observed; however, it still remains that PG is
the more dynamic form of glycogen because the shift in
storage is observed in this pool more readily.
6224
J Clin Endocrinol Metab, November 2005, 90(11):6218 – 6225
Our study confirms the previous observations that E2 supplementation caused a decrease in glucose Ra and Rd (21, 22).
The consistency of this finding reaffirms that E2 acts to promote whole-body CHO sparing during exercise by preferentially attenuating hepatic glucose production (Ra) and glucose uptake (Rd). Animal studies support a net sparing of
hepatic glycogen use during exercise in response to E2 supplementation (12, 15, 16). From the current data, we cannot
differentiate between an effect of E2 on reducing hepatic
gluconeogenesis and/or glycogenolysis; however, a previous study showed that women had a lower glucose recycling
rate than men (2), suggesting a lesser flux of CHO precursors
through gluconeogenesis. The main precursors for gluconeogenesis are lactate, glycerol, and alanine. No effect of E2
supplementation has been found on glycerol concentration
(21–23), and the current study and others (21–23) found no
effect of E2 supplementation on plasma lactate concentration, suggesting that E2 does not decrease gluconeogenesis.
Furthermore, amino acid oxidation in skeletal muscle during
exercise, the main source for alanine, is lower in women (4,
42). The possible reasons that E2 appears to have a predominant effect on net hepatic, but not skeletal muscle, glycogenolysis are unclear. To date, no study has investigated
differences in E2 receptor number or affinity between the two
tissues, which could account for the observed differences.
Alternatively, it could be due to differences in the rate of
glycogen turnover between liver and muscle; because liver
has a quicker protein turnover rate as compared with muscle
(44), it could be more sensitive to E2 supplementation and/or
show effects more readily in a short-term study.
Unlike previous studies (21, 22), we failed to find an effect
of E2 on MCR. It is possible that we did not have the statistical
power necessary to find an effect of E2 on MCR, therefore
making a type II error. However, the two studies that found
an effect of E2 on MCR were slightly different from the
current study. Our dose was lower than that used in one
study (22), which achieved supraphysiological E2 concentrations and the effect on MCR thus may be a dose-response
relation, and our dose may not have been high enough to
elicit this response. Although the dose in the other study that
found an effect of E2 on MCR was lower than that used in
the current study, the subjects were women, who, as previously discussed, may respond more readily to E2 treatment
because of gender differences in estrogen receptor number
and affinity. From the current data, we are unable to determine whether we have made a type II error or whether there
was no effect of our E2 dose on MCR.
In the current study, we used 11 men to investigate the effect
of E2 on glucose kinetics and muscle glycogen use. In this
randomized, double-blind, crossover, placebo-controlled design, the subjects acted as their own control, hence reducing
errors due to small sample size and the high between-subject
variability. The statistical analyses that we used complemented
the strong study design to reduce both type I and type II errors.
First, we used a higher number of subjects than in the previously published articles addressing similar issues. Second, we
observed significant differences due to estrogen supplementation in RER, glucose Ra and Rd, and total glycogen and PG
concentrations, indicating that, at least with respect to these
outcome variables, we achieved sufficient power. Furthermore,
Devries et al. • Effects of 17-␤-Estradiol on Carbohydrate Metabolism
alpha (also known as P value, which indicates the probability
of a type I error) would be the same, even with 10 times the
number of subjects because alpha would remain at a level of
0.05. This confirms and extends the findings by other researchers. Indeed, this manuscript complements our observations of
the effect of estrogen on whole-body substrate oxidation and
leucine kinetics (26). Alternatively, a low sample size will alter
power (1-␤), and this may have resulted in the lack of significant
findings in glycogen use. However, we calculated the number
of individuals necessary to elicit an effect of E2 on glycogen use
and found it to be 300. The differences in glycogen use between
the estrogen and placebo groups were 4.2% for macroglycogen
and 4.6% for total glycogen, which is unlikely to be statistically
significant or, with a sample size of 300 subjects, physiologically
significant. Furthermore, it is unlikely that the lack of difference
in glycogen use was due to a type I error because our results fall
in line with results observed in previous estrogen supplementation studies.
In summary, we have demonstrated that E2 administration decreases RER, glucose Ra and Rd, and muscle glycogen
content without altering muscle glycogen use during exercise. Future studies should concentrate on longer duration
supplementation regimens to determine whether a longer
dosage period would decrease muscle glycogenolysis. Also,
future research should focus on determining whether the
decreased glucose Ra after E2 supplementation is due to a
decrease in glycogenolysis or gluconeogenesis. Lastly, research needs to be conducted to investigate the effect of E2
supplementation on IMCL stores and whether E2 controls
lipid availability and oxidation primarily, resulting in a subsequent shift in CHO use.
Acknowledgments
We thank Christine Rodriguez, Brian Timmons, Laura Phillips, and
Christopher Westbrook for assisting with the subjects and Jennifer Day
for analyzing the diet records. We also thank Shire BioChem, Inc. (St.
Laurent, Québec, Canada), for supplying the E2 and Abbott Laboratories, Ross Division (St. Laurent, Québec, Canada), for supplying the
Polycose.
Received April 28, 2005. Accepted August 16, 2005.
Address all correspondence and requests for reprints to: Mark A. Tarnopolsky, M.D., Ph.D., Department of Pediatrics, McMaster University
Medical Center, Room 4U4, McMaster University, 1200 Main Street West,
Hamilton, Ontario, Canada L8Z 3N5. E-mail: [email protected].
This work was supported by National Sciences and Engineering
Research Council (NSERC) Canada and Hamilton Health Sciences Foundation. M.C.D. was the recipient of the NSERC of Canada Industrial
Scholarship with Nestlé Canada 2003–2005. M.J.H. was the recipient of
the National Institute of Nutrition 2002–2004 Postdoctoral Fellowship.
References
1. Carter SL, Rennie C, Tarnopolsky MA 2001 Substrate utilization during
endurance exercise in men and women after endurance training. Am J Physiol
280:E898 –E907
2. Friedlander AL, Casazza GA, Horning MA, Huie MJ, Piacentini MF, Trimmer JK, Brooks GA 1998 Training-induced alterations of carbohydrate metabolism in women: women respond differently from men. J Appl Physiol
85:1175–1186
3. Horton TJ, Pagliassotti MJ, Hobbs K, Hill JO 1998 Fuel metabolism in men
and women during and after long-duration exercise. J Appl Physiol 85:1823–
1832
4. Phillips S, Atkinson SA, Tarnopolsky MA, MacDougall JD 1993 Gender
differences in leucine kinetics and nitrogen balance in endurance athletes.
J Appl Physiol 75:2134 –2141
Devries et al. • Effects of 17-␤-Estradiol on Carbohydrate Metabolism
5. Romijn JA, Coyle EF, Sidossis LS, Rosenblatt J, Wolfe RR 2000 Substrate
metabolism during different exercise intensities in endurance-trained women.
J Appl Physiol 88:1701–1714
6. Tarnopolsky LJ, MacDougall JD, Atkinson SA, Tarnopolsky MA, Sutton JR
1990 Gender differences in substrate for endurance exercise. J Appl Physiol
68:302–308
7. Tarnopolsky MA, Atkinson SA, Phillips SA, MacDougall JD 1995 Carbohydrate loading and metabolism during exercise in men and women. J Appl
Physiol 78:1360 –1368
8. Costill D, Fink W, Getchell L, Ivy J 1979 Lipid metabolism in skeletal muscle
of endurance-trained males and females. J Appl Physiol 47:787–791
9. Roepstorff C, Steffensen CH, Madsen M, Stallknect B, Kanstrup I, Richter
EA, Kiens B 2002 Gender differences in substrate utilization during submaximal exercise in endurance-trained subjects. Am J Physiol 282:E435–E447
10. Venables MC, Achten J, Jeukendrup AE 2005 Determinants of fat oxidation
during exercise in healthy men and women: a cross-sectional study. J Appl
Physiol 98:160 –167
11. Bunt J 1990 Metabolic actions of estradiol: significance for acute and chronic
exercise responses. Med Sci Sports Exerc 22:286 –290
12. Kendrick ZV, Steffen CA, Rumsey WL, Goldberg DI 1987 Effect of estradiol
on tissue glycogen metabolism in exercise oophorectomized rats. J Appl
Physiol 63:492– 496
13. Ellis GS, Lanza-Jacoby S, Gow A, Kendrick ZV 1994 Effects of estradiol on
lipoprotein lipase activity and lipid availability in exercise male rats. J Appl
Physiol 77:209 –215
14. Hatta H, Atomi Y, Shinohara S, Yamamoto Y, Yamada S 1988 The effects of
ovarian hormones on glucose and fatty acid oxidation during exercise in
female ovariectomized rats. Horm Metab Res 20:609 – 611
15. Kendrick ZV, Ellis GS 1991 Effect of estradiol on tissue glycogen metabolism
and lipid availability in exercised male rats. J Appl Physiol 71:1694 –1699
16. Rooney TP, Kendrick ZV, Carlson J, Ellis GS, Matakevich B, Lorusso SM,
McCall JA 1993 Effect of estradiol on the temporal pattern of exercise-induced
tissue glycogen depletion in male rats. J Appl Physiol 75:1502–1506
17. Klover P, Mooney R 2004 Hepatocytes: critical for glucose homeostasis. Int
J Biochem Cell Biol 36:753–758
18. Hamosh M, Hamosh P 1975 The effect of estrogen on the lipoprotein lipase
activity of rat adipose tissue. J Clin Invest 55:1132–1135
19. Ramirez I 1981 Estradiol-induced changes in lipoprotein lipase, eating, and
body weight in rats. Am J Physiol 240:E533–E538
20. Wilson D, Flowers C, Carlile S, Udall K 1976 Estrogen treatment and gonadal
function in the regulation of lipoprotein lipase. Atherosclerosis 24:491– 499
21. Ruby BC, Robergs RA, Waters DL, Burge M, Mermier C, Stolarczyk L 1997
Effects of estradiol on substrate turnover during exercise in amenorrheic females. Med Sci Sports Exerc 29:1160 –1169
22. Carter SL, McKenzie S, Mourtzakis M, Mahoney DJ, Tarnopolsky MA 2001
Short-term 17-␤-estradiol decreases glucose Ra but not whole body metabolism during endurance exercise. J Appl Physiol 90:139 –146
23. Tarnopolsky MA, Roy BD, MacDougall JD, McKenzie S, Martin J, Ettinger
S 2001 Short-term 17-␤-estradiol administration does not affect metabolism in
young males. Int J Sports Med 22:175–180
24. Adamo KB, Graham TE 1998 Comparison of traditional measurements with
J Clin Endocrinol Metab, November 2005, 90(11):6218 – 6225
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
6225
macroglycogen and proglycogen analysis of muscle glycogen. J Appl Physiol
84:908 –913
Lomanko J, Lomanko WM, Whelan WJ 1993 Proglycogen: a low molecularweight form of muscle glycogen. FEBS Lett 279:223–228
Hamadeh MJ, Devries MC, Tarnopolsky MA 2005 Estrogen supplementation
reduces whole body leucine and carbohydrate oxidation and increases lipid
oxidation in men during endurance exercise. J Clin Endocrinol Metab 90:3592–
3599
Phillips S, Green H, Tarnopolsky M, Heigenhauser G, Grant S 1996 Progressive effect of endurance training on metabolic adaptations in working
skeletal muscle. Am J Physiol 270:E265–E272
Harris R, Hultman E, Nordesjo L 1974 Glycogen, glycolytic intermediates and
high-energy phosphates determined in biopsy samples of musculus quadriceps femoris of man at rest. Methods and variance of values. Scand J Clin Lab
Invest 33:109 –120
Steele R 1959 Influences of glucose loading and of injected insulin on hepatic
glucose output. NY Acad Sci 82:420 – 430
Romijn JA, Coyle EF, Sidossis LS, Gastaldelli A, Horowitz JF, Endert E,
Wolfe RR 1993 Regulation of endogenous fat and carbohydrate metabolism
in relation to exercise intensity and duration. Am J Physiol 256:E380 –E391
Bonen A, Haynes W, Graham TE 1991 Substrate and hormonal responses to
exercise in women using oral contraceptives. J Appl Physiol 70:1917–1927
Campbell SE, Angus DJ, Febbraio MA 2001 Glucose kinetics and exercise
performance during phases of the menstrual cycle: effect of glucose ingestion.
Am J Physiol Endocrinol Metab 281:E817–E825
Nicklas BJ, Hackney AC, Sharp RL 1989 The menstrual cycle and exercise:
performance, muscle glycogen, and substrate responses. Int J Sports Med
10:264 –269
Lemoine S, Granier P, Tiffoche C, Rannou-Bekono F, Thieulant M, Delamarche P 2003 Estrogen receptor ␣ mRNA in human skeletal muscle. Med Sci
Sports Exerc 35:439 – 443
Wiik A, Glenmark B, Ekman M, Esbjornsson-Liljedahl M, Johansson O,
Bodin K, Enmark E, Jansson E 2003 Oestrogen receptor ␤ is expressed in adult
human skeletal muscle both at the mRNA and protein level. Acta Physiol
Scand 179:381–387
Lamont L, McCullough A, Kalhan S 2001 Gender differences in leucine, but
not lysine, kinetics. J Appl Physiol 91:357–362
Koop-Woodroffe S, Manore M, Dueck C, Skinner J, Matt K 1999 Energy and
nutrient status of amenorrheic athletes participating in a diet and exercise
training intervention program. Int J Sports Med 9:70 – 88
Roepstorff C, Halberg N, Hillig T, Saha A, Ruderman N, Wojtaszewski J,
Richter E, Kiens B 2005 Malonyl-CoA and carnitine in regulation of fat oxidation in human skeletal muscle during exercise. Am J Physiol Endocrinol
Metab 288:E133–E142
Gerson L, Hagobian T, Grow D, Chipkin SR, Braun B 2004 Role of testosterone in substrate use during exercise. Physiologist 47:31
Guo Z, Burguera B, Jensen MD 2000 Kinetics of intramuscular triglyceride
fatty acids in exercising humans. J Appl Physiol 89:2057–2064
Roepstorff C, Halberg N, Hillig T, Saha A, Ruderman N, Wojtaszew J,
Richter EA, Kiens B 2005 Malonyl-CoA and carnitine in regulation of fat
oxidation in human skeletal muscle during exercise. Am J Physiol Endocrinol
Metab 288:E133–E142
JCEM is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the
endocrine community.