Effects of hyper- and hypoosmolality on whole body
protein and glucose kinetics in humans
KASPAR BERNEIS,1 RONALD NINNIS,1 DIETER HÄUSSINGER,2 AND ULRICH KELLER1
of Research and of Internal Medicine, University Hospital Basel,
4031 Basel, Switzerland; and 2Department of Internal Medicine,
University Hopital Düsseldorf, 40225 Düsseldorf, Germany
1Departments
dehydration; nonesterified fatty acids; protein turnover; lipolysis; stable isotopes
CELL VOLUME ALTERATIONS induced by changes of extracellular osmolality have been reported to regulate
intracellular metabolic pathways (16, 17). In vitro
studies demonstrated that administration of hormones
such as insulin and glucagon not only affect liver
protein and glycogen metabolism but also liver cell
volume (4, 16). Hypoosmotic cell swelling counteracts
proteolysis in liver, whereas hyperosmotic cell shrinkage promotes protein breakdown (16). It has also been
suggested that the antiproteolytic effect of several
amino acids is linked to amino acid-induced cell swelling (15, 16, 37). Amino acid-induced stimulation of
glycogen synthesis (4, 29) or inhibition of proteolysis
(37) can be mimicked by hypoosmotic swelling of the
cells (2, 15, 16, 24). Low et al. (33) demonstrated that
glycogen synthesis in skeletal muscle in association
The costs of publication of this article were defrayed in part by the
payment of page charges. The article must therefore be hereby
marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.
E188
with changes in cell volume requires integrin-extracellular matrix interactions and presumably an intact
cytoskeleton.
Clinical investigations in physiological states in which
cell swelling and cell shrinking may be expected are
rare. Electrolyte disorders resulting in changes of
plasma osmolality are often observed in hospitalized
patients, and it has been suggested that the loss of
protein and potassium from body stores observed during major trauma or sepsis is accompanied by progressive cellular dehydration (9). Recently, it also has been
demonstrated that dehydration in decompensated diabetes mellitus is associated with protein catabolism
and insulin resistance of glucose metabolism (24).
However, the metabolic effects of acute alterations of
extracellular osmolality have not been assessed previously in humans. Effects on lipid metabolism have not
been assessed in vitro and in vivo. The present study
examined, therefore, the effects of acute hyperosmolality and hyposmolality on whole body protein and glucose kinetics and on lipid metabolites of human subjects; they were studied after an overnight fast and
during intravenous administration of insulin, glucose,
and amino acids to create a protein-anabolic milieu (3).
METHODS
Subjects. Written informed consent was obtained from 10
healthy young male volunteers aged 25 6 1 yr with a body
mass index of 23.0 6 0.8 kg/m2. Their medical history,
physical examination, and routine laboratory tests before the
studies provided no evidence of cardiopulmonary, renal, hepatic, or metabolic diseases. The subjects were on no medication and did not perform vigorous exercise during the study
period. The study protocol was reviewed and approved by the
ethical committee of the Basel University Hospital.
Procedures. Each subject underwent three sets of studies
(hyperosmolality, hypoosmolality, and isoosmolality) in randomized order and with intervals of $1 wk in between (Fig.
1). Each subject remained fasting after 11 AM and was
admitted at 4:30 PM to the metabolic study unit of the
hospital. Thereafter, baseline kinetics were measured by
placing a Teflon cannula into the right antecubital vein for
infusions; a superficial dorsal vein of the right hand was
cannulated in a retrograde manner with a 21-gauge butterfly
needle for blood sampling. The hand was kept in a thermostatcontrolled warming chamber at ,55°C, allowing arterialization of hand venous blood (35). Blood and breath samples
were subsequently obtained to determine background isotopic enrichments of plasma [2H2]glucose, [1-13C]leucine,
a-[1-12C]ketoisocaproate (a-KIC), and breath 12CO2, followed
by primed-continuous infusions of [1-12C]leucine (3 µmol/kg
bolus; 0.06 µmol · kg21 · min21 infusion; 99% enriched; sterile
and pyrogen free; Mass Trace, Somerville, MA) and of [6,62H ]glucose (3 mg/kg bolus; 0.04 mg · kg21 · min21 infusion;
2
0193-1849/99 $5.00 Copyright r 1999 the American Physiological Society
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.2 on June 15, 2017
Berneis, Kaspar, Ronald Ninnis, Dieter Häussinger,
and Ulrich Keller. Effects of hyper- and hypoosmolality on
whole body protein and glucose kinetics in humans. Am. J.
Physiol. 276 (Endocrinol. Metab. 39): E188–E195, 1999.—To
investigate the effect of acute changes of extracellular osmolality on whole body protein and glucose metabolism, we
studied 10 male subjects during three conditions: hyperosmolality was induced by fluid restriction and intravenous infusion of hypertonic NaCl [2–5%; (wt/vol)] during 17 h; hypoosmolality was produced by intravenous administration of
desmopressin, liberal water drinking, and infusion of hypotonic saline (0.4%); and the isoosmolality study consisted of
ad libitum oral water intake by the subjects. Leucine flux
([1-13C]leucine infusion technique), a parameter of whole
body protein breakdown, decreased during the hypoosmolality study (P , 0.02 vs. isoosmolality). The leucine oxidation
rate decreased during the hypoosmolality study (P , 0.005
vs. isoosmolality). Metabolic clearance rate of glucose during
hyperinsulinemic-euglycemic clamping increased less during
the hypoosmolality study than during the isoosmolality study
(P , 0.04). Plasma insulin decreased, and plasma nonesterified fatty acids, glycerol, and ketone body concentrations and
lipid oxidation increased during the hypoosmolality study. It
is concluded that acute alterations of plasma osmolality
influence whole body protein, glucose, and lipid metabolism;
hypoosmolality results in protein sparing associated with
increased lipolysis and lipid oxidation and impaired insulin
sensitivity.
EFFECTS OF HYPOOSMOLALITY ON PROTEIN AND GLUCOSE METABOLISM
E189
Fig. 1. Diagram of study protocol. p.os,
Per os (by mouth); Minirin, desmopressin; Vamina, standard mixed amino
acid solution.
Vamina (in g/l) was 2.5 aspartate, 4.2 glutamate, 5.9 glycine,
12.0 alanine, 8.4 arginine, 0.42 cysteine, 5.1 histidine, 4.2
isoleucine, 5.9 leucine, 6.8 lysine, 4.2 methionine, 5.9 phenylalanine, 5.1 proline, 3.2 serine, 4.2 threonine, 1.4 tryptophane, 0.17 tyrosine, and 5.5 valine. The amino acid solution
was enriched with 0.295 g [1-13C]leucine per 1,000 ml solution
to ,5% tracer-to-tracee ratio (TTR) to maintain plasma
a-KIC TTR during infusion unchanged. Plasma was rapidly
obtained by refrigerated centrifugation (4°C) and was stored
at 270°C until later assay. Expired air was collected into
gastight 20-ml glass tubes (Vacutainer, Becton-Dickinson,
Meylan, France) for later 13CO2 analysis.
Analytical methods. All tracer infusates were ultrafiltrated
(0.1 µm) and analyzed by gas chromatography-mass spectrometry (GC-MS model 5890 /5790, Hewlett-Packard, Palo Alto,
CA) for tracer concentration, isotopic enrichment, and chemical purity. Plasma TTR of [1-13C]leucine and of a-[1-13C]KIC
was measured by GC-MS selected ion monitoring (39). Plasma
concentrations of leucine and a-KIC were determined by the
same methodology with [2H7]leucine and a-[2H3]KIC as internal standards, respectively. Isotopic enrichment of 12CO2 in
expired air was measured by isotope ratio mass spectrometry
(SIRA series II, VG Isotech, Cheshire, UK). CO2 production
rate (V̇CO2) was determined by indirect calorimetry with a
ventilated hood metabolic monitor (Deltatrac II MBM-200,
Datex, Helsinki, Finland). Plasma arginine vasopressin was
measured by RIA as described by Liard et al. (32). Plasma
glucose concentrations were measured with glucose oxidase
and a hydrogen peroxide sensor (glucose analyzer 2300 STAT
Plus, YSI, Yellow Springs, OH). Plasma sodium and potassium concentrations were measured by indirect potentiometry (Du Pont Dimension AR, Dade, Düdingen, Switzerland),
and plasma osmolality was measured by cryoscopic technique
(micro osmometer 3 MO, Advanced Instruments, Norwood,
MA)
Plasma concentrations of C-peptide (23), glucagon, and
insulin were measured with RIAs (CIS BIO International,
F91192 Gif-Sur-Yvette Cedex, France; Diagnostic Products,
Los Angeles, CA; and Insik-5 P2796 kit, Sorin Biomedica,
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.2 on June 15, 2017
98% enriched; sterile and pyrogen free; Mass Trace) administered during 3 h. A single injection of sodium [1-13C]bicarbonate (1.76 µmol/kg; 99% enriched; sterile and pyrogen free;
Mass Trace) was used to accelerate 13C labeling of the
bicarbonate pool (1). After 150 min of tracer equilibration,
blood and breath samples were obtained in 15-min intervals
during 30 min (baseline kinetics). At 8 PM, the subjects were
served a standard meal (600 kcal) and remained fasting
thereafter until the end of the study on the following day.
In the hyperosmolality study, the subjects were instructed
not to drink after 8 PM after the first day of the study; during
the night, they received 1 ml · kg21 · h21 of a saline [2% NaCl
(wt/vol)] infusion, and at 8 AM, they received an infusion of
200 ml/h of 5% NaCl. In the hypoosmolality study, the
subjects received 4 µg desmopressin (Minirin) intravenously
at 8 PM on the first study day and at 8 AM on the second study
day, respectively, and they were instructed to drink 2–2.5
liters of tap water during the night. A 0.4% NaCl (200 ml/h)
infusion was started in the morning until the end of the study.
The isoosmolality study consisted of an identical protocol in
which osmolality was maintained constant throughout the
study by access to oral water ad libitum. In all studies,
plasma Na1 concentrations and osmolality were measured
before the treatment was started and hourly after 8 AM on
the second day. The subjects were supervised throughout the
study by a physician.
Leucine and glucose kinetics were measured a second time
on study day 2 in an identical fashion, starting at 8 AM and
continuing during 5 h. After the total experimental period of
180 min, insulin (Actrapid, Norvonordisc, Küsnacht) was
infused continuously at 60 mU · m22 · min21 during 3 min, and
then at 15 mU · m22 · min21 during 117 min (7). In addition,
20% glucose (wt/vol) was infused at variable rates and
adjusted every 5–10 min according to rapidly measured
plasma glucose concentrations to maintain euglycemia; a
standard mixed amino acid solution (Vamina 10%, Pharmacia, Stockholm, Sweden) was administered intravenously at a
rate of 0.0144 ml · kg21 · min21, corresponding to a leucine
infusion rate of 0.65 µmol · kg21 · min21. The composition of
E190
EFFECTS OF HYPOOSMOLALITY ON PROTEIN AND GLUCOSE METABOLISM
mmol/l (P , 0.0001) between 8 PM (end of baseline
period) and 1 PM (end of the study) (Fig. 2). Osmolality
(mmol/kgH2O) decreased slightly during the isoosmolality study from 286 6 1 to 283 6 1, increased during the
hyperosmolality study from 283.4 6 0.5 to 296.4 6 0.7
(P , 0.0001), and decreased during the hypoosmolality
study from 286 6 1 to 265 6 1 (P , 0.0001) between 8
PM (end of baseline period) and the end of the study at
1 PM on the next day. Plasma potassium concentrations (mmol/l) during the baseline period were 3.7 6
0.04, 3.9 6 0.06, and 3.8 6 0.06 during the isoosmolality, hyperosmolality, and hypoosmolality studies, respectively. They remained unchanged until the end of the
study in the hyperosmolality (3.9 6 0.04) and isoosmolality (3.8 6 0.04) studies and decreased in the hypoosmolality study to 3.6 6 0.06 (P , 0.02 vs. baseline).
RESULTS
Plasma sodium, osmolality, potassium, and water
balance. Plasma sodium concentrations remained unchanged in the isoosmolality study (baseline values,
142 6 0.4; end of study, 140 6 0.6 mmol/l); they
increased in the hyperosmolality study from 142 6 0.2
to 149 6 0.4 mmol/l (P , 0.0001) and decreased in the
hypoosmolality study from 142 6 0.4 to 131 6 0.5
Fig. 2. Sodium plasma concentrations (mmol/l) and plasma osmolality (mmol/kgH2O) in isoosmolality (G, middle curve), hyperosmolality
(E, top curve), and hypoosmolality (C, bottom curve) studies. Data are
means 6 SE; n 5 10 subjects/group.
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.2 on June 15, 2017
Italy, respectively). Plasma concentrations of nonesterified
fatty acids (NEFA) (36), glycerol (5), acetoacetate, and b-hydroxybutyrate (43, 44) were determined with enzymatic
methods.
Calculations. Estimates of whole body leucine and glucose
kinetics were made at steady-state conditions during 30 min
on the evening before the study day (baseline period), during
30 min on the study day (experimental period), and during
the end of the clamping period (270–300 min) on the study
day. Total leucine flux was calculated by dividing the infusion
rate of [1-13C]leucine by the a-[1-13C]KIC TTR according to
the reciprocal pool model (6, 25). Leucine oxidation rate
(representing irreversible leucine catabolism) was calculated
by dividing the product of 13CO2 atom percent excess and
V̇CO2 in expired air by plasma a-[1-13C]KIC TTR. A CO2
retention factor of 0.81 was used (1).
Infusion of glucose during glucose clamping with a different natural 13CO2 content may lead to an error of the
calculated leucine oxidation rate by influencing the background 13CO2-to-12CO2 ratio. To address this question, Laager
et al. (30) recently determined background 13C enrichment
while infusing high doses of insulin and glucose but no 13C
leucine tracer during 8 h. They found a 6% overestimation of
the leucine oxidation rate (30) during glucose infusion, a
number which would be lower in the present study due to the
lower glucose infusion rate (,40% of that studied by Laager
et al.). Thus there was only a minimal effect of glucose
infusions on calculated leucine oxidation rates during clamping in all protocols of the present study.
Nonoxidative leucine disappearance (representing whole
body protein synthesis) was calculated by subtracting the
rate of leucine oxidation from total leucine flux. Endogenous
leucine flux (a parameter of protein breakdown) was calculated by subtracting the infusion rate of unlabeled leucine
from total leucine flux. Net balance of leucine metabolism was
calculated as the difference between the rate of nonoxidative
disappearance and endogenous flux of leucine. Endogenous
glucose rate of appearance (Ra ) was calculated by dividing the
[6,6-2H2]glucose infusion rate by plasma glucose TTR. During
glucose clamping, total plasma glucose Ra was the sum of the
glucose infusion rate and endogenous glucose Ra. Glucose Ra
divided by the corresponding plasma glucose concentration
yielded the glucose metabolic clearance rate (MCR).
Respiratory quotients were calculated by dividing V̇CO2 (ml/
min) by V̇O2 (ml/min), where V̇O2 is oxygen consumption.
Resting energy expenditures (kcal/24 h) and utilization of fat
and carbohydrates as percentage of nonprotein energy expenditure were determined by indirect calorimetry (28).
Statistical analysis. Repeated-measures ANOVA of Statview and Student’s paired t-tests (Abacus Concepts, Berkeley,
CA) on a Power Macintosh 7100/80 were used to detect
differences between and within the three protocols. BonferroniDunn and Scheffé’s F procedures were performed for correction of double comparisons. Statistical tests were only performed to assess differences between the hypoosmolality and
isoosmolality studies and the hyperosmolality and isoosmolality studies and within all groups. Results are means 6 SE.
EFFECTS OF HYPOOSMOLALITY ON PROTEIN AND GLUCOSE METABOLISM
E191
Fig. 4. Changes of leucine oxidation rate and endogenous leucine flux
from baseline to experimental period. Data are means 6 SE; n 5 10.
ISO, isoosmolality; HYPER, hyperosmolality; HYPO, hypoosmolality.
* P , 0.05 and **P , 0.005 vs. ISO (repeated-measures ANOVA).
Fig. 3. Tracer-to-tracee ratio (TTR) of a-ketoisocaproate (a-KIC) and
13CO atom percent excess (APE) rate in the isoosmolality (G),
2
hyperosmolality (E), and hypoosmolality (C) studies during 3 successive periods of observation. Results of consecutive measurements
obtained in 10-min intervals of baseline, experimental, and clamping
periods are shown. Data are means 6 SE; n 5 10 subjects/group.
ing the hypoosmolality study than during the
isoosmolality study (P , 0.02 vs. isoosmolality, paired
t-test). Leucine oxidation increased to 0.42 6 0.02,
0.43 6 0.01, and 0.41 6 0.02 µmol · kg21 · min21 in the
isoosmolality, hyperosmolality, and hypoosmolality studies, respectively, during the clamping period (P ,
0.0001 vs. experimental period for all groups). Nonoxidative leucine flux remained unchanged between the
baseline and the experimental period of all studies. It
increased during clamping in all studies (P , 0.0001 vs.
experimental period), without difference between them.
Plasma leucine concentrations were 153 6 10, 157 6
11, and 152 6 5 µmol/l during the baseline period of the
isoosmolality, hyperosmolality, and hypoosmolality studies, respectively. They remained unchanged during the
isoosmolality (145 6 5) and hyperosmolality studies
(147 6 5) but decreased slightly during the hypoosmolality study to 143 6 2 µmol/l (P , 0.05 vs. baseline
values). During insulin-glucose-amino acid clamping
plasma leucine concentrations were 149 6 6, 149 6 4,
and 150 6 3 µmol/l during the isoosmolality, hyperosmolality, and hypoosmolality studies, respectively (nonsignificant vs. experimental period).
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.2 on June 15, 2017
Water balance in the isoosmolality study and the
hyperosmolality study was zero; fluid balance increased in the hypoosmolality study by 2.01 6 0.07 kg,
as calculated from urinary output and water administration.
Leucine kinetics. TTR of a-KIC and of CO2 reached
nearly steady-state conditions during all three measurement periods in all three studies (Fig. 3). Figure 4
demonstrates that endogenous leucine flux
(µmol · kg21 · min21 ) decreased in the hypoosmolality
study from the baseline to the experimental period of
the next morning from 1.9 6 0.05 to 1.79 6 0.06
compared with the isoosmolality study (P , 0.02 vs.
hypoosmolality, repeated-measures ANOVA) and remained unchanged in the isoosmolality and hyperosmolality studies (from 1.79 6 0.09 to 1.88 6 0.1 and from
1.8 6 0.06 to 1.71 6 0.06, respectively). During clamping endogenous leucine flux decreased to 1.4 6 0.06,
1.31 6 0.03, and 1.4 6 0.04 in the isoosmolality study,
and hypoosmolality studies, respectively (P , 0.0001
vs. experimental period for all groups), with no significant difference between them. Leucine oxidation decreased from baseline values to the experimental period from 0.34 6 0.03 to 0.27 6 0.01 µmol · kg21 · min21
(P , 0.03 vs. baseline, P , 0.005 vs. isoosmolality,
repeated-measures ANOVA) during the hypoosmolality
study and remained unchanged in the isoosmolality
study (0.31 6 0.02 during the baseline period, 0.32 6
0.02 during the experimental period) and in the hyperosmolality study (0.31 6 0.02 and 0.31 6 0.01, respectively). Leucine oxidation was significantly lower dur-
E192
EFFECTS OF HYPOOSMOLALITY ON PROTEIN AND GLUCOSE METABOLISM
Table 1. Glucose plasma concentrations and kinetics
Isoosmolality study
Glucose plasma concentrations, mmol/l
Glucose MCR,
ml · kg21 · min21
Endogenous glucose
Ra , µmol · kg21
· min21
Hyperosmolality study
Glucose plasma concentrations, mmol/l
Glucose MCR,
ml · kg21 · min21
Endogenous glucose
Ra , µmol · kg21
· min21
Hypoosmolality study
Glucose plasma concentrations, mmol/l
Glucose MCR,
ml · kg21 · min21
Endogenous glucose
Ra , µmol · kg21
· min21
Baseline
Experimental
Clamping
4.8 6 0.2
4.9 6 0.1
4.9 6 0.1
2.8 6 0.1
2.3 6 0.1c
4.4 6 0.4d
13 6 0.7
11.2 6 0.4c
5.3 6 0.7
4.9 6 0.05
5.1 6 0.05a
4.9 6 0.1
2.8 6 0.1
2.5 6 0.1c
4.2 6 0.4d
13.5 6 0.6
12.4 6 0.4a
6 6 0.6
4.8 6 0.1
4.7 6 0.05a
4.8 6 0.1
2.7 6 0.2
2.4 6 0.1
3.3 6 0.2b,d
12.8 6 0.1
11.1 6 0.3
4.6 6 0.6
Data are means 6 SE; n 5 10 subjects/group. MCR, metabolic
clearance rate; Ra , rate of appearance. a P , 0.05 vs. isoosmolality
(paired t-tests); b P , 0.05 vs. isoosmolality (repeated-measures
ANOVA); c P , 0.05 vs. baseline (paired t-tests); d P , 0.005 vs.
experimental (paired t-tests).
Table 2. Plasma concentrations of insulin,
C-peptide, glucagon, NEFAs, glycerol, acetoacetate,
and b-hydroxybutyrate
Isoosmolality study
Insulin, pmol/l
C-peptide, nmol/l
Glucagon, ng/l
NEFA, µmol/l
Glycerol, µmol/l
Acetoacetate, µmol/l
b-Hydroxybutyrate, µmol/l
Hyperosmolality study
Insulin, pmol/l
C-peptide, nmol/l
Glucagon, ng/l
NEFA, µmol/l
Glycerol, µmol/l
Acetoacetate, µmol/l
b-Hydroxybutyrate, µmol/l
Hypoosmolality study
Insulin, pmol/l
C-peptide, nmol/l
Glucagon, ng/l
NEFA, µmol/l
Glycerol, µmol/l
Acetoacetate, µmol/l
b-Hydroxybutyrate, µmol/l
Experimental
Clamping Period
40.8 6 4.8
0.378 6 0.025
74 6 3
414 6 17
43 6 0.003
42 6 5
66 6 13
154.8 6 4.8c
0.322 6 0.027b
83 6 5c
241 6 23c
18 6 0.2c
30 6 2c
13 6 5
37.8 6 3.6
0.379 6 0.022
81 6 5
392 6 31
45 6 0.5
54 6 7
86 6 26
139.8 6 6.6
0.307 6 0.038b
99 6 6c
200 6 14c
16 6 0.2c
36 6 2c
12 6 4
29.4 6 3a
0.359 6 0.027
75 6 5
435 6 20
49 6 0.3a
74 6 10a
144 6 37a
148.2 6 9c
0.277 6 0.033a,c
92 6 5c
217 6 20c
17 6 0.1c
40 6 3a,c
11 6 4
Data are means 6 SE; n 5 10/group. NEFA, nonesterified fatty
acids. a P , 0.05 vs. isoosmolality (paired t-tests); b P , 0.05, c P ,
0.005 vs. experimental period (paired t-tests).
tions increased during clamping, and C-peptide decreased similarly in all studies (P , 0.05 or less).
C-peptide concentrations during clamping were lower
during the hypoosmolality study than during the isoosmolality study (P , 0.05). Glucagon plasma concentrations increased similarly during clamping in all studies
(P , 0.005 or less). NEFA concentrations were slightly
higher during the hypoosmolality study compared with
the isoosmolality study (nonsignificant), and they decreased similarly during clamping in all studies (P ,
0.005 or less). Plasma glycerol concentrations during
the experimental period were higher during the hypoosmolality study than during the isoosmolality study (P ,
0.05). Plasma glycerol concentrations were similarly
decreased during clamping in all studies. Plasma acetoacetate concentrations were higher during the hypoosmolality study than during the isoosmolality study
(P , 0.05 or less). Plasma acetoacetate concentrations
were similarly decreased in all studies during clamping
but decreased more in the hypoosmolality study compared with isoosmolality (P , 0.05, paired t-test).
b-Hydroxybutyrate concentrations were higher during
the hypoosmolality study than during the isoosmolality
study (P , 0.05). b-Hydroxybutyrate plasma concentrations decreased similarly in all studies during clamping. Plasma concentrations of antidiuretic hormone
were 1.1 6 01 during the isoosmolality study and 1.7 6
0.2 pg/ml during the hyperosmolality study (P , 0.001
vs. isoosmolality). At the end of the studies, antidiuretic
hormone plasma concentrations were 0.7 6 0.1 (isoosmolality study) and 3.6 6 0.2 pg/ml (hyperosmolality
study), respectively (P , 0.0001 vs. isoosmolality).
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.2 on June 15, 2017
Glucose kinetics. Plasma glucose concentrations were
higher during the hyperosmolality study (P , 0.03,
paired t-test) and lower during the hypoosmolality
study (P , 0.03 vs. isoosmolality study) compared with
the isoosmolality study (Table 1). Ra decreased from
baseline to the experimental period during the isoosmolality study (P , 0.02 paired t-test) and remained
unchanged during the hyperosmolality and hypoosmolality studies. Glucose Ra during the experimental
period was higher in the hyperosmolality study compared with the isoosmolality study (P , 0.03 paired
t-test). Glucose MCR decreased from baseline to the
experimental period during the isoosmolality (P ,
0.005 paired t-test) and hyperosmolality studies (P ,
0.05) but missed statistical significance in the hypoosmolality study (P , 0.1). Glucose MCR increased
during clamping in all studies (P , 0.005 or more
significant vs. experimental period); the increase in
MCR in the hypoosmolality study was blunted compared with the isoosmolality study (P , 0.02, repeatedmeasures ANOVA). The rate of glucose infusion to
maintain euglycemia was 14.98 6 1.1 during the
isoosmolality study, 14.43 6 1.7 during the hyperosmolality study, but only 10.5 6 0.55 µmol · kg21 · min21
during the hypoosmolality study (P , 0.01 vs. isoosmolality).
Plasma concentrations of insulin, C-peptide, glucagon, glycerol, NEFA, acetoacetate, and b-hydroxybutyrate and of antidiuretic hormone. Plasma insulin concentrations were lower during the experimental period of
the hypoosmolality study compared with the isoosmolality study (P , 0.05 ) (Table 2). Plasma insulin concentra-
EFFECTS OF HYPOOSMOLALITY ON PROTEIN AND GLUCOSE METABOLISM
DISCUSSION
The present study examined for the first time the
effect of acute hyperosmolality and hypoosmolality on
whole body protein and glucose metabolism in humans.
Administration of a vasopressin analog and liberal
water drinking, on the one hand, or infusion of a
hypertonic saline solution and restriction from drinking, on the other hand, were used to produce changes of
Table 3. Resting energy expenditure, respiratory gas
exchange respiratory quotient, and utilization of
carbohydrates and fat as percentage of nonprotein
energy expenditure
Baseline
Experimental
Clamping
Period
Isoosmolality study
REE, kcal/24 h
1,724 6 73
1,590 6 65c
1,651 6 79e
V̇O2 , ml/min
247 6 12
228 6 9c
235 6 11
V̇CO2 , ml/min
221 6 9
207 6 8c
219 6 11e
RQ
0.91 6 0.02 0.90 6 0.02
0.93 6 0.01
C%
64 6 5
73 6 5
78 6 5
F%
36 6 5
27 6 5
22 6 5
Hyperosmolality study
REE, kcal/24 h
1,729 6 71
1,668 6 90
1,776 6 77
V̇O2 , ml/min
251 6 12
242 6 14
248 6 13
V̇CO2 , ml/min
214 6 7
207 6 10
224 6 8f
RQ
0.86 6 0.03 0.86 6 0.02
0.91 6 0.03
C%
64 6 8
50 6 6
67 6 9e
F%
36 6 8
50 6 6
33 6 9e
Hypoosmolality study
REE, kcal/24 h
1,813 6 99
1,583 6 83d
1,691 6 92e
V̇O2 , ml/min
263 6 15
232 6 13d
245 6 12
V̇CO2 , ml/min
223 6 10
189 6 8a,d
212 6 11f
RQ
0.85 6 0.02 0.82 6 0.03a 0.87 6 0.02b
C%
53 6 8
43 6 9a
51 6 6b
F%
47 6 8
57 6 9a
49 6 6b
Data are means 6 SE; n 5 10 subjects/group. REE, resting energy
expenditure; RQ, respiratory gas exchange respiratory quotient; C%
and F%, utilization of carbohydrates and fat as percentage of
nonprotein energy expenditure, respectively; V̇O2 and V̇CO2 , oxygen
consumption and carbon dioxide production rates, respectively. a P ,
0.05 vs. isoosmolality (repeated-measures ANOVA); b P , 0.05 vs.
isoosmolality (paired t-tests); c P , 0.05, d P , 0.005 vs. Baseline
(paired t-tests); e P , 0.05, f P , 0.005 vs. Experimental (paired
t-tests).
extracellular osmolality, resulting presumably in a
modest state of cell swelling or cell shrinkage (12).
Leucine release from endogenous proteins (representing protein breakdown) and leucine oxidation (indicating irreversible catabolism) were diminished during
hypoosmolality compared with isoosmolality. Because
protein synthesis (nonoxidative leucine disappearance)
during hypoosmolar conditions was not altered, this
indicates that net protein balance was improved during
the hypoosmolal state. In contrast, there was no significant change of whole body protein synthesis or protein
breakdown during hyperosmolal conditions. The measurements during the baseline periods were performed
after a 5-h fast in the evening, whereas the measurements during the experimental periods were performed
after a 12-h overnight fast. Therefore, it cannot be
excluded that the different duration of fasting had an
influence on the results of the present study; however,
the main conclusions were derived from the data of the
experimental periods that were obtained after identical
periods of fasting.
Endogenous glucose Ra (representing mainly hepatic
glucose production) and plasma glucose concentrations
were increased during hyperosmolality compared with
isoosmolality. In contrast, plasma glucose concentrations were decreased in the hypoosmolality study compared with the isoosmolality study. The increase in
glucose metabolic clearance rate during clamping was
diminished during hypoosmolal conditions, indicating
diminished insulin sensitivity of peripheral glucose
metabolism.
Concerning the effects of hypoosmolality, these data
should be discussed in the light of previous findings
obtained in vitro, suggesting that short-term modulation of cell volume within a narrow range acts as a
potent signal to modify cellular metabolism and gene
expression (20). Hypoosmolal liver cell swelling increases protein synthesis, glycogen synthesis, and amino
acid uptake and decreases proteolysis, glycogenolysis,
and glycolysis, whereas opposite metabolic effects are
triggered by cell shrinkage (18, 20, 41). It has been
demonstrated that in isolated perfused livers and hepatocytes, insulin increases and glucagon decreases cellular volume within minutes (13), and these effects may
explain in part their metabolic effects. The Na1dependent amino acid transport systems in the plasma
membrane may act as a transmembrane signaling
system, triggering cellular function by altering cellular
hydration in response to substrate delivery (18–20).
Regulation of phosphatidylinositol 3-kinase in skeletal
muscle may be an important component of the signaling mechanisms involved in cell volume-mediated control of membrane transport (34).
The present finding may be important to understand
the pathogenesis of protein catabolism in various diseases. Indeed, a close relationship between cellular
hydration of skeletal muscle and nitrogen balance has
been demonstrated in severely ill patients (21). Regarding the question whether the observed metabolic effects
of changes in osmolality are in fact related to changes
in cell volume, no direct measurements of cell volume
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.2 on June 15, 2017
Because of crossreactivity of the administered vasopressin analog Minirin with antidiuretic hormone, plasma
antidiuretic hormone concentrations were not measured under hypoosmolal conditions.
Indirect calorimetry. Resting energy expenditure decreased from baseline to the experimental period during the isoosmolality (P , 0.05) and hypoosmolality
studies (P , 0.005 vs. baseline) but remained unchanged during the hyperosmolality study (Table 3).
V̇O2 and V̇CO2 decreased from baseline to the experimental period during the isoosmolality and hypoosmolality
studies (P , 0.05 or less). V̇CO2 during the experimental
period was lower during the hypoosmolality study than
during the isoosmolality study (P , 0.05). V̇CO2 increased during clamping in all studies (P , 0.05 or
less). Utilization of carbohydrates as percentage of
nonprotein energy expenditure was lower, and utilization of fat was higher during the hypoosmolality study
than during the isoosmolality study (P , 0.05 or less).
E193
E194
EFFECTS OF HYPOOSMOLALITY ON PROTEIN AND GLUCOSE METABOLISM
acids are available for oxidation, and fatty acid oxidation is inversely related to plasma glucose concentrations (40).
It is concluded that moderate states of hyperosmolality and hypoosmolality influence protein, glucose, and
fat metabolism; these effects may be linked to hypoosmolal cell swelling and hyperosmolal cell shrinkage.
Hypoosmolality in the present study exerted a proteinand glucose-sparing effect with increased utilization of
fat. In contrast, hyperosmolality exerted opposite effects on glucose metabolism with increased hepatic
glucose production, resulting in modestly increased
plasma glucose concentrations.
The results of the study therefore suggest that alterations in whole body water and ionic balance exert
metabolic effects, which may be important in clinical
situations of altered extracellular osmolality.
We gratefully acknowledge the excellent technical assistance of S.
Vosmeer, K. Dembinski, and S. Sansano, and we thank those who
measured the plasma antidiuretic hormone concentrations in the
laboratory of Prof. M. Vallotton, University Hospital, Geneva.
This work was supported by Swiss National Science Foundation
Grant 32–39747.93.
Address for reprint requests: U. Keller, Depts. of Research and
Internal Medicine, Univ. Hospital Basel, Petersgraben 4, 4031 Basel,
Switzerland.
Received 9 April 1998; accepted in final form 29 September 1998.
REFERENCES
1. Allsop, J. R., R. R. Wolfe, and J. F. Burke. Tracer priming the
bicarbonate pool. J. Appl. Physiol. 45: 137–139, 1978.
2. Baquet, A., L. Hue, A. J. Meijer, G. M. Van Woerkom, and
P. J. A. M. Plomp. Swelling of rat hepatocytes stimulates
glycogen synthesis. J. Biol. Chem. 265: 955–959, 1990.
3. Castellino, P., L. Luzi, D. C. Simonson, M. Haymond, and
R. A. Defronzo. Effect of insulin and plasma amino acid
concentrations on leucine metabolism in man. Role of substrate
availability on estimates of whole body protein synthesis. J. Clin.
Invest. 80: 1784–1793, 1987.
4. Chen, K. S., and H. A. Lardy. Multiple requirements for
glycogen synthesis by hepatocytes isolated from fasted rats. J.
Biol. Chem. 260: 14684–14688, 1985.
5. Chernick, S. Determination of glyercol in acyl glycerols. Methods Enzymol. 14: 627–630, 1969.
6. Cobelli, C., G. Toffolo, D. Bier, and R. Nosadini. Models to
interpret kinetic data in stable isotope tracer studies. Am. J.
Physiol. 253 (Endocrinol. Metab. 16): E551–E564, 1987.
7. DeFronzo, R., J. Tobin, and R. Andres. Glucose clamp
technique: a method for quantifying insulin secretion and resistance. Am. J. Physiol. 237 (Endocrinol. Metab. Gastrointest.
Physiol. 6): E214–E223, 1979.
8. Ferranini, E., E. Barrett, S. Bevilacqua, and S. Defronzo.
Effects of fatty acids on glucose production and utilization in
man. J. Clin. Invest. 72: 1737–1747, 1983.
9. Finn, J., D. Lindsay, M. Clark, A. Connolly, and G. Hill.
Progressive cellular dehydration and proteolysis in critically ill
patients. Lancet 347: 654–656, 1996.
10. Fukagawa, N., K. Minaker, J. Rowe, M. Goodman, D.
Matthews, D. Bier, and V. Young. Insulin-mediated reduction
of whole body protein breakdown. Dose-response effects on
leucine metabolism in postabsorptive man. J. Clin. Invest. 76:
2306–2311, 1985.
11. Graf, J., P. Haddad, D. Häussinger, and F. Lang. Cell volume
regulation in liver. Renal Physiol. Biochem. 3–5: 202–220, 1988.
12. Graf, J., and D. Häussinger. Ion transport in hepatocytes:
mechanism and correlations to cell volume, hormone actions and
metabolism. J. Hepatol. 24: 53–77, 1996.
13. Häussinger, D. Regulation of metabolism by changes in cellular
hydration. Clin. Nutr. 14: 4–12, 1995.
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.2 on June 15, 2017
could be obtained in our study. In vitro regulatory
mechanisms are activated within minutes in response
to hypoosmolality or hyperosmolality exposure. However, these volume-regulatory mechanisms do not completely restore cell hydration. This has been studied
extensively in liver (11, 22, 42). In fact, a decrease of
extracellular osmolality in perfused rat livers by 20
mosmol/kgH2O resulted in a persistent 4.7 6 1.3%
increase in cell hydration (S. von Dahl and D. Häeussinger, unpublished data). The extent of this cell
volume deviation after completion of volume regulatory
ion fluxes determines the metabolic changes (14, 20, 31)
that persist as long as the anisotonicity is maintained.
The longest time period studied in vitro was about 2 h.
Therefore, we believe that alterations of cell volume
were likely to explain the present findings. The present
alterations may be comparable to those in vitro from
above. However, such small changes of cell volume are
difficult to measure with sufficient precision.
The reason that acute hyperosmolality did not result
in protein catabolism in the present study is not
obvious; it is possible that 17 h was an inappropriate
time frame over which to detect any changes in protein
breakdown and/or that the increase in plasma osmolality was too modest to exert sufficient cell shrinking.
However, more pronounced experimental dehydration
in living human volunteers would be too risky. It also
should be noted that net water balance was only
affected in the hypoosmolality group, whereas the
hyperosmolality group experienced no change of water
balance.
Postabsorptive plasma insulin concentrations were
lower during hypoosmolality, and C-peptide suppression during clamping was increased. Compared with
isoosmolality, this is a yet undescribed phenomenon.
The antiproteolytic effect of insulin is well described
(10); therefore, lowered insulin levels during hypoosmolal conditions may partly counteract the antiproteolytic effect in humans, thereby masking, in part, the
antiprotein catabolic effects of hypoosmolality. On the
other hand, decreased insulin concentrations may contribute to the observed increase in fat utilization as
observed by indirect calorimetry and to the increase in
plasma glycerol concentrations. Increased oxidation of
lipid substrates and increased ketone body concentrations may explain, in part, the protein-sparing effect
during hypoosmolality (38). Additionally, the increase
in glycerol concentrations during hypoosmolal conditions may exert a nitrogen-sparing effect by conserving
amino acids as gluconeogenic precursors, thus making
amino acids available for reincorporating into protein
and reducing urea production (26). The mechanism by
which hypoosmolality increased lipolysis remained unclear; plasma catecholamines were not increased during hypoosmolality (S. Bilza and U. Keller, unpublished
data). Oxidation of carbohydrates was decreased in the
present study. Several reasons may explain this finding. Increased utilization of fat reduces the need for
glucose as a fuel (8). This may be the result of decreased
plasma insulin levels, insulin being a potent inhibitor
of lipolysis (27). When lipolysis is increased, more fatty
EFFECTS OF HYPOOSMOLALITY ON PROTEIN AND GLUCOSE METABOLISM
31. Lang, F., G. Busch, M. Ritter, H. Völkl, S. Waldegger, E.
Gulbins, and D. Häussinger. Functional significance of cell
volume regulatory mechanisms. Physiol. Rev. 78: 247–306, 1998.
32. Liard, J., W. Dolci, and M. Vallotton. Plasma vasopressin
levels after infusions of hypertonic saline solutions into the
renal, portal, carotid, or systemic circulation in conscious dogs.
Endocrinology 114: 986–991, 1984.
33. Low, S., M. Rennie, and P. Taylor. Involvement of integrins
and the cytoskeleton in modulation of skeletal muscel glycogen
synthesis by changes in cell volume. FEBS Lett. 417: 101–103,
1997.
34. Low, S., M. Rennie, and P. Taylor. Signaling elements involved in amino acid transport responses to altered muscle cell
volume. FASEB J. 11: 1111–1117, 1997.
35. McGuire, E., J. Helderman, J. Tobin, R. Andres, and M.
Berman. Effects of arterial versus venous sampling on analysis
of glucose kinetics in man. J. Appl. Physiol. 41: 565–573, 1976.
36. Miles, J., R. Glasscock, J. Aikens, J. Gerich, and M. Haymond. A microfluorometric method for the determination of free
fatty acids in plasma. J. Lipid Res. 24: 96–99, 1983.
37. Mortimore, G. E., and A. R. Pösö. Mechanism and control of
deprivation induced protein degradation in liver: role of glucogenic amino acids. In: Glutamine Metabolism in Mammalian
Tissues, edited by D. Häussinger and H. Sies. Heidelberg,
Germany: Springer-Verlag, 1984, p. 138–157.
38. Nair, K., S. Welle, D. Halliday, and R. Campbell. Effect of
beta-hydroxybutyrate on whole-body leucine kinetics and fractional mixed skeletal muscle protein synthesis in humans. J.
Clin. Invest. 82: 198–205, 1988.
39. Schwenk, W., P. Berg, B. Beaufrère, J. Miles, and M.
Haymond. t-Butyldimethylsylation in the gas chromatographicmass spectrometric analysis of physiologic compounds found in
plasma using electron-impact ionization. Anal. Biochem. 141:
101–109, 1984.
40. Sidossis, L. S., and R. R. Wolfe. Glucose and insulin-induced
inhibition of fatty acid oxidation: the glucose-fatty acid cycle
reversed. Am. J. Physiol. 270 (Endocrinol. Metab. 33): E733–
E738, 1996.
41. Stoll, B., W. Gerok, F. Lang, and D. Häussinger. Liver cell
volume and protein synthesis. Biochem. J. 287: 217–222, 1992.
42. Vom Dahl, S., C. Hallbrucker, F. Lang, W. Gerok, and D.
Häussinger. A non-invasive technique for cell volume determination in perfused rat liver. Biol. Chem. Hoppe Seyler 372:
411–418, 1991.
43. Williamson, D., and J. Mellanby. D-(2)-Hydroxybutyrat. In:
Methoden der enzymatischen Analyse, edited by H. Bergmeyer.
Berlin, Germany: Springer-Verlag, 1974, p. 1883–1886.
44. Young, D., and A. Renold. A fluorimetric procedure for the
determination of ketone bodies in very small quantities of blood.
Clin. Chim. Acta 13: 791–793, 1966.
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.2 on June 15, 2017
14. Häussinger, D. The role of cell hydration for the regulation of
cell function. Biochem. J. 313: 697–710, 1996.
15. Häussinger, D., C. Halbrucker, S. vom Dahl, F. Lang, and
W. Gerok. Cell swelling inhibits proteolysis in perfused rat liver.
Biochem. J. 272: 239–242, 1990.
16. Häussinger, D., C. Hallbrucker, S. vom Dahl, U. Schweizer,
F. Lang, and W. Gerok. Cell volume is a major determinant of
proteolysis control in liver. FEBS Lett. 283: 70–72, 1991.
17. Häussinger, D., and F. Lang. The mutual interaction between
cell volume and cell function: a new principle of metabolic
regulation. Biochem. Cell Biol. 69: 1–4, 1991.
18. Häussinger, D., and F. Lang. Cell volume in the regulation of
hepatic function: a new mechanism for metabolic control. Biochim. Biophys. Acta 1071: 331–350, 1991.
19. Häussinger, D., and F. Lang. Cell volume and hormone action.
Trends Pharmacol. Sci. 13: 371–373, 1992.
20. Häussinger, D., F. Lang, and W. Gerok. Regulation of cell
function by the cellular hydration state. Am. J. Physiol. 267
(Endocrinol. Metab. 30): E343–E355, 1994.
21. Häussinger, D., E. Roth, F. Lang, and W. Gerok. Cellular
hydration state: an important determinant of protein catabolism
in health and disease. Lancet 341: 1330–1332, 1993.
22. Häussinger, D., T. Stehle, and F. Lang. Volume regulation in
liver; further characterization by inhibitors and ionic substitutions. Hepatology 11: 243–254, 1990.
23. Heding, L. Radioimmunological determination of human
C-peptide in serum. Diabetologia 11: 541–548, 1975.
24. Hellerstein, M. K. Isotopic studies of carbohydrate metabolism
in non-insulin-dependent diabetes mellitus. Curr. Op. Endocrinol. Diab. 2: 518–529, 1995.
25. Horber, F. F., C. M. Horber-Feyder, S. Krayer, W. F. Schwenk, and M. W. Haymond. Plasma reciprocal pool specific
activity predicts that of intracellular free leucine for protein
synthesis. Am. J. Physiol. 257 (Endocrinol. Metab. 20): E385–
E399, 1989.
26. Jahoor, F., E. J. Peters, and R. R. Wolfe. The relationship
between gluconeogenic substrate supply and glucose production
in humans. Am. J. Physiol. 258 (Endocrinol. Metab. 21): E288–
E296, 1990.
27. Jensen, M., M. Caruso, V. Heiling, and J. Miles. Insulin
regulation of lipolysis in nondiabetic and IDDM subjects. Diabetes 38: 1595–1601, 1989.
28. Jequier, E., K. Acheson, and Y. Schutz. Assessement of
energy expenditure and fuel utilization in man. Ann Rev Nutr:
187–208, 1987.
29. Katz, J., S. Golden, and P. A. Wals. Stimulation of hepatic
glycogen synthesis by amino acids. Proc. Natl. Acad. Sci. USA 73:
3433–3437, 1976.
30. Laager, R., R. Ninnis, and U. Keller. Comparison of the effects
of recombinant human insulin-like growth factor-I and insulin on
glucose and leucine kinetics in humans. J. Clin. Invest. 92:
1903–1909, 1993.
E195