Am J Physiol Endocrinol Metab 282: E794–E801, 2002;
10.1152/ajpendo.00374.2001.
Glucose release from GLUT2-null hepatocytes:
characterization of a major and a minor pathway
MASAYA HOSOKAWA AND BERNARD THORENS
Institute of Pharmacology and Toxicology, University of Lausanne, CH-1005 Lausanne, Switzerland
Received 17 August 2001; accepted in final form 10 November 2001
gluconeogenesis; hepatic glucose output; intracellular traffic;
glucose-6-phosphatase
is an essential physiological function that is activated in the fasted state to
prevent development of hypoglycemia. In diabetes, this
function becomes progressively insensitive to inhibition by insulin and therefore contributes to increased
hyperglycemia. Hepatic glucose production can result
from activation of two metabolic pathways, glycogenolysis or gluconeogenesis, that converge at the level of
glucose 6-phosphate (G-6-P) production. Conversion to
glucose then requires G-6-P to enter the endoplasmic
reticulum (ER), a process catalyzed by a glucose-6phosphate translocase (4), followed by hydrolysis to
glucose and phosphate by the membrane-associated
GLUCOSE RELEASE FROM HEPATOCYTES
Address for reprint requests and other correspondence: B. Thorens,
Institute of Pharmacology and Toxicology, 27, rue du Bugnon, Ch-1005
Lausanne, Switzerland (E-mail: [email protected]).
E794
glucose-6-phosphatase, whose catalytic site is located
inside the ER lumen (9, 13). Release of glucose outside
the cells has been classically viewed as involving diffusion of glucose back into the cytosol and transport
across the plasma membrane by the glucose transporter GLUT2 (15). Recently, we reported that, in the
absence of this transporter, glucose could, however,
still be released at a normal rate, even though facilitated diffusion of 3-O-methylglucose (3-MG) across the
plasma membrane was reduced by ⬎95% (5). We presented evidence that glucose release in the absence of
GLUT2 could rely on a membrane traffic mechanism
issued from the ER. This pathway was characterized
by its sensitivity to low temperature and to the acute
effect of progesterone. It was, however, insensitive to
cytochalasin B or inhibitors of the classical intracellular transport pathway, brefeldin A and monensin (see
Fig. 9). Importantly, this membrane traffic pathway
appeared to coexist with the GLUT2-dependent pathway in normal hepatocytes.
The conclusions about the existence of a membrane
traffic pathway were thus inferred from the observation that release of neosynthesized glucose could be
inhibited by low temperature and progesterone and on
the parallel demonstration that facilitated diffusion of
3-MG across the plasma membrane was strongly reduced. However, the formal possibility remained that
glucose could still be released by a plasma membrane
carrier specific for glucose and unable to transport
3-MG (5).
In the present study, we investigated the mechanisms of glucose release from control and GLUT2deficient [GLUT2(⫺/⫺)] hepatocytes by means of a
glucose biosynthetic labeling protocol. We demonstrate
that, in the absence of GLUT2, there is an intracellular
accumulation of glucose even though there is a constant release of glucose into the culture medium. We
show that the rate of glucose release during a continuous labeling experiment can be impaired by microtubule disruption in addition to inhibition by progesterone and low temperature. We further show that the
intracellular pool of glucose is located in the cytosol
and is released from the cells with much slower kinetics than the bulk of newly synthesized glucose. The
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.
0193-1849/02 $5.00 Copyright © 2002 the American Physiological Society
http://www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.5 on June 16, 2017
Hosokawa, Masaya, and Bernard Thorens. Glucose
release from GLUT2-null hepatocytes: characterization of a
major and a minor pathway. Am J Physiol Endocrinol Metab
282: E794–E801, 2002; 10.1152/ajpendo.00374.2001.—We
previously reported that glucose can be released from
GLUT2-null hepatocytes through a membrane traffic-based
pathway issued from the endoplasmic reticulum. Here, we
further characterized this glucose release mechanism using
biosynthetic labeling protocols. In continuous pulse-labeling
experiments, we determined that glucose secretion proceeded
linearly and with the same kinetics in control and GLUT2null hepatocytes. In GLUT2-deficient hepatocytes, however,
a fraction of newly synthesized glucose accumulated intracellularly. The linear accumulation of glucose in the medium
was inhibited in mutant, but not in control, hepatocytes by
progesterone and low temperature, as previously reported,
but, importantly, also by microtubule disruption. The intracellular pool of glucose was shown to be present in the
cytosol, and, in pulse-chase experiments, it was shown to be
released at a relatively slow rate. Release was not inhibited
by S-4048 (an inhibitor of glucose-6-phosphate translocase),
cytochalasin B, or progesterone. It was inhibited by phloretin, carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone,
and low temperature. We conclude that the major release
pathway segregates glucose away from the cytosol by use of a
membrane traffic-based, microtubule-dependent mechanism
and that the release of the cytosolic pool of newly synthesized
glucose, through an as yet unidentified plasma membrane
transport system, cannot account for the bulk of glucose
release.
GLUCOSE RELEASE FROM GLUT2-NULL HEPATOCYTES
release in the culture medium of cytosolic glucose does
not require return to the ER but is by a mechanism
that requires ATP production and could be inhibited by
phloretin. There are therefore two pathways for glucose release from GLUT2(⫺/⫺) hepatocytes, a major
one that segregates glucose away from the cytosol and
a minor one that probably involves glucose diffusion
across the plasma membrane by a low-affinity transport mechanism but which cannot account for the rapid
rate of glucose output. These data therefore further
support the proposal that hepatic glucose release is
mostly by a membrane traffic-based pathway.
MATERIAL AND METHODS
AJP-Endocrinol Metab • VOL
AG1-X8 and 50W-X8, respectively; Bio-Rad, Hercules, CA)
exactly as described (5). Radioactivity was measured by liquid scintillation counting (Tri-Carb 2100 TR; Packard Bioscience).
RESULTS
We previously showed that the kinetics of release
of newly synthesized glucose was similar in control
and GLUT2-null hepatocytes (5). Here, to more precisely characterize the release pathways, we performed biosynthetic labeling experiments. Freshly
isolated hepatocytes obtained from 24-h-fasted mice
were incubated in the presence of [14C]pyruvate for
different periods of time, and the newly synthesized
[14C]glucose secreted by the cells or remaining intracellular was then quantitated. Figure 1A shows that,
in control hepatocytes, the newly synthesized glucose is released in the cell supernatant at a constant
rate (⬃6 nmol 䡠 mg protein⫺1 䡠 h⫺1) and that there is
no intracellular accumulation of glucose. In contrast,
in GLUT2(⫺/⫺) mice (Fig. 1B), there is an accumulation of glucose inside the cells, which reaches a
plateau after ⬃30 min. Release of glucose in the
culture medium, however, increases linearly over the
time of the experiment at a rate that is similar to
that observed in control hepatocytes. Figure 1C
shows the same data expressed as a percentage of
total glucose produced at 2 h.
We next attempted to interfere with the secretion of
glucose in the GLUT2(⫺/⫺) hepatocytes by using drugs
known to interfere with vesicular trafficking. In continuous 1-h pulse-labeling experiments, and as previously reported, progesterone reduced glucose release
by ⬃50% and incubation of the cells at low temperature
(12°C) by 70% (Table 1). In the presence of progesterone, glucose release from control hepatocytes was not
affected (data not shown and Ref. 5).
It has been reported that capacitative Ca2⫹ entry
after depletion of the ER Ca2⫹ stores may involve a
direct connection between the ER and the plasma
membrane. This was demonstrated, in particular, by
showing that actin overpolymerization by jasplakinolide or calyculin A could block this Ca2⫹ entry by forming a tight microfilament network below the plasma
membrane (10, 11). We therefore tested the effect of
both jasplakinolide and calyculin A on glucose release.
Even though we could confirm the action of these
molecules on induction of actin polymerization by phalloidin-Alexa red staining, no inhibition of glucose release could be observed (Table 1). We also evaluated
whether depleting the cells in cholesterol could reduce
glucose release. This was based on the premise that
transport of newly synthesized cholesterol from the ER
to the plasma membrane may proceed through the
same pathway as glucose release, since transport of
both cholesterol and glucose is reduced by low temperature and progesterone treatment (5, 14, 17). Thus
cholesterol could have been involved in this vesicular
traffic pathway. Incubation of the cells in the presence
of methyl-␤-cyclodextrin at concentrations that markedly reduce cellular cholesterol content, (12) however,
282 • APRIL 2002 •
www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.5 on June 16, 2017
Materials. The glucose-6-phosphate translocase inhibitor
S-4048 was a generous gift of Dr. A. W. Herling (Aventis,
Frankfurt, Germany). Progesterone was purchased from ICN
(Eschwege, Germany). Calyculin A was purchased from Calbiochem (Darmstadt, Germany). Jasplakinolide was purchased from Molecular Probes (Eugene, OR). Phloretin, carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone (FCCP),
cytochalasin B, streptolysin O, colchicine, methyl-␤-cyclodextrin, and nocodazole were purchased from Sigma. [14C]pyruvate was purchased from New England Nuclear (Boston,
MA). All other chemicals were of reagent grade.
Animals. RIPGLUT1 ⫻ GLUT2(⫺/⫺) mice were from our
own colony (6, 16). As control animals, we used wild-type
C57BL/6J mice purchased from BRL (Basel, Switzerland).
Hepatocyte preparation. Livers of 8-wk-old mice were perfused through the inferior vena cava with a buffer consisting
of (in mM) 140 NaCl, 2.6 KCl, 0.28 Na2HPO4, 5 glucose, and
10 HEPES (pH 7.4). The perfusion was first for 5 min with
the buffer supplemented with 0.1 mM EGTA and then for 15
min with the buffer containing 5 mM CaCl2 and 0.2 mg/ml
collagenase type 2 (Worthington, Lakewood, NJ). All of the
solutions were prewarmed at 37°C and gassed with a mixture
of 95% O2-5% CO2, resulting in a pH of 7.4. The isolated
hepatocytes were filtered on nylon mesh (0.75 ␮m in diameter), washed two times with the above-mentioned buffer
without collagenase, and suspended in a small volume of
DMEM (GIBCO, Rockville, MD) without glucose or pyruvate
and counted. The viability of hepatocytes was measured by
Trypan blue staining. The preparations with viability ⬍90%
were discarded.
Biosynthetic labeling. For pulse-labeling experiments,
hepatocytes (7.5 ⫻ 105) were incubated at 37 or 12°C in 0.5
ml of DMEM containing 1 mM pyruvate, 0.24 mM 3-isobutyl1-methylxanthine (IBMX), and 0.05 ␮Ci of [14C]pyruvate in
the presence of either various inhibitors or vehicle. Incubations were stopped by placing the cells on ice followed by
centrifugation at 4°C for 60 s at a speed of 1,000 rpm. The
supernatant was removed, and the cells were lysed in 0.2%
sodium deoxycholate. An aliquot was kept for protein determination (bicinchoninic acid kit; Pierce, Rockville, IL), and
the rest was used for determination of radioactivity.
For pulse-chase experiments, hepatocytes were pulse labeled for 15 or 30 min as described above. They were then
washed two times using 1-min centrifugations and were
returned to 0.5 ml of DMEM containing 1 mM pyruvate and
0.24 mM IBMX in the presence of various inhibitors or
vehicle, but without [14C]pyruvate, and kept for the indicated
periods of time at 37°C.
[14C]glucose measurement and analysis. [14C]glucose was
separated from charged metabolites by passage of lysates or
supernatants on anion and cation exchangers (Dowex
E795
E796
GLUCOSE RELEASE FROM GLUT2-NULL HEPATOCYTES
Table 1 Effect of various inhibitors on [14C]glucose
release measured at the end of the 60-min
continuous pulse labeling
Inhibitors
Progesterone
Low temperature
Calyculin A
Methyl-␤-cyclodextrin
Jasplakinoride
Colchinie
Nocodazole
(10 ␮g/ml)
(12°C)
(100 nM)
(10 mM)
(3 ␮M)
(80 ␮M)
(50 ␮M)
Without
With
58 ⫾ 2.4
55 ⫾ 1.1
41 ⫾ 2
53.1 ⫾ 6.8
58.4 ⫾ 1.6
54 ⫾ 1.2
50.6 ⫾ 1.4
33 ⫾ 4*
16.4 ⫾ 0.6*
47 ⫾ 3.7
55.5 ⫾ 3.3
58.5 ⫾ 1.3
46 ⫾ 1.1*
35.9 ⫾ 2.0*
Fig. 1. Kinetics of [14C]glucose synthesis and release from hepatocytes of control or GLUT2-deficient [GLUT2(⫺/⫺)] mice. Freshly
isolated hepatocytes were incubated in the presence of 1 mM pyruvate and 0.05 ␮Ci [14C]pyruvate for the indicated periods of time,
and [14C]glucose present in the supernatants (SN) and cell was
quantitated. A: control hepatocytes. There is no intracellular accumulation of glucose. B: GLUT2(⫺/⫺) hepatocytes. There is an accumulation of glucose inside the cells, which reaches its maximum at
⬃30 min. Glucose released in the medium proceeds at a constant
rate. C: data in A and B are replotted as a percentage of total glucose
produced at 2 h. Data are means ⫾ SE for n ⫽ 4 experiments.
AJP-Endocrinol Metab • VOL
did not reduce the rate of glucose release. Finally,
experiments were carried out to evaluate the role of
microtubules in glucose release. As shown in Table 1,
colchicine reduced the rate of secretion by ⬃15% and
nocodazole by ⬃30%.
To evaluate whether the membrane traffic-based
and GLUT2-dependent pathways could be evidenced in
control hepatocytes, pulse-labeling experiments were
carried out for 30 min in the presence of nocodazole
and/or cytochalasin B. Figure 2 shows that the release
of [14C]glucose from control hepatocytes could be reduced significantly by nocodazole (⬃15% reduction)
and the GLUT2 inhibitor cytochalasin B (⬃28% reduction). Importantly, a combination of cytochalasin B and
nocodazole produced an additive inhibitory effect
(⬃48% reduction), suggesting that both pathways indeed coexist in normal hepatocytes.
Next, we studied the rate of release of the pool of
glucose that accumulates inside the GLUT2(⫺/⫺)
hepatocytes and whether we could interfere pharmacologically with this release. For these experiments,
freshly isolated hepatocytes were pulse labeled with
[14C]pyruvate for 30 min, washed, and then returned to
a culture medium containing an excess of cold pyruvate. Intracellular and secreted [14C]glucose was then
quantitated. Figure 3 shows that the rate of glucose
release is relatively slow, with a half-time of ⬃30 min
and an absolute rate of ⬃0.25 nmol 䡠 mg protein⫺1 䡠 h⫺1.
The glucose leaving the cells during this experiment
was quantitatively recovered in the supernatant (Fig.
3), indicating that the decrease in intracellular glucose
was the result of secretion and not of metabolism. This
rate of release is slower than that measured during
continuous pulse labeling (6 nmol 䡠 mg protein⫺1 䡠 h⫺1).
This therefore suggested that the intracellularly accumulated [14C]glucose is not in an intermediate compartment of the major glucose release pathway.
On the basis of the model presented in Fig. 9, the
intracellularly accumulated glucose could be either in a
membrane compartment in transit between the ER to
the plasma membrane or in the cytosol. If this glucose
were in a closed vesicular compartment, permeabilization of the plasma membrane at the end of the pulse
282 • APRIL 2002 •
www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.5 on June 16, 2017
Data are expressed as means ⫾ SE; n ⫽ 4⫺6. Data represent
glucose secreted from GLUT2-deficient hepatocytes and are expressed as percentage of total newly synthesized glucose. *P ⬍ 0.01
by Student’s t-test.
GLUCOSE RELEASE FROM GLUT2-NULL HEPATOCYTES
E797
labeling with streptolysin O would preserve the retention of sugar in the cells. In contrast, if this glucose
were in the cytosol, it would diffuse out of the cells
rapidly. We therefore pulse labeled the cells for 15 min,
exposed them to streptolysin O for 10 min at 37°C, and
determined the amount of [14C]glucose retained intracellularly and present in the supernatant. The concentration of streptolysin O used was the minimal concentration that led to a complete permeabilization of the
cells, as measured by Trypan blue staining and by
measuring the release of lactate dehydrogenase (data
not shown). Figure 4 shows that streptolysin O led to
the release of ⬃90% of intracellular [14C]glucose. The
same cellular depletion of glucose was observed when
the cells were permeabilized with Triton X-100 (0.1%),
which should also permeabilize the intracellular mem-
Fig. 3. Kinetics of release of intracellularly accumulated [14C]glucose. GLUT2(⫺/⫺) hepatocytes were pulse labeled for 30 min with
[14C]pyruvate, washed, and returned to a nonradioactive culture
medium. The amount of [14C]glucose remaining inside the cells or
secreted in the cell culture medium was then measured at the
indicated time and expressed as nmol/mg protein (prot). The total
amount of glucose radioactively labeled remained constant during
the time of the experiment. Data are means ⫾ SE for n ⫽ 4.
AJP-Endocrinol Metab • VOL
brane. We take this as an indication that the glucose
that accumulates during this period of pulse labeling is
present in the cytosol.
To evaluate whether intracellular glucose release
required phosphorylation into G-6-P, entry into the
ER, and exit by a membrane traffic pathway, we
pulse-labeled the cells for 30 min and chased them in
the presence of S-4048, an inhibitor of the glucose6-phosphate translocase (1, 7). The data presented in
Fig. 5 show that there was no inhibition of glucose
release induced by S-4048. To check that the concentration of inhibitor was effective in blocking the
glucose-6-phosphate translocase, we separately showed
that it completely inhibited [14C]glucose formation
from [14C]pyruvate (data not shown). If the ER-plasma
membrane pathway was used, it should also have been
inhibited by progesterone treatment. Figure 6 shows
that release was not affected by incubation with pro-
Fig. 4. The intracellular pool of glucose is located in the cytosol.
GLUT2(⫺/⫺) hepatocytes were labeled for 15 min with [14C]pyruvate and incubated in control (Ctr) conditions with streptolysin O
(SLO, 3,300 IU/ml in the presence of 1 mM dithiothreitol) or with
0.1% Triton X-100 (Tx) for 10 min. Intracellular glucose was then
determined. Permeabilization of the plasma membrane (SLO) and
the intracellular membranes (Tx) leads to an ⬃90% release of glucose. Data indicate intracellular glucose expressed as a percentage of
total synthesized glucose. Data are means ⫾ SE for n ⫽ 4.
282 • APRIL 2002 •
www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.5 on June 16, 2017
Fig. 2. Glucose is released from normal hepatocytes by both the GLUT2-dependent and
membrane traffic-based pathways. Freshly
isolated hepatocytes were incubated in the
presence of 1 mM pyruvate and 0.05 ␮Ci
[14C]pyruvate for 30 min, and [14C]glucose
present in the supernatants and cell lysate
was quantitated. The pulse-labeling experiments were performed in the presence of cytochalasin B (cytoB; 50 ␮M), nocodazole, (50
␮M), or a combination of both inhibitors. The
presence of both inhibitors showed an additive inhibitory effect. Data are means ⫾ SE
for n ⫽ 6.
E798
GLUCOSE RELEASE FROM GLUT2-NULL HEPATOCYTES
gesterone. These data therefore suggested that the
intracellularly accumulated glucose was released by a
mechanism that did not require phosphorylation and
entry into the ER. It rather involved diffusion through
the plasma membrane.
To determine whether the release of accumulated
[14C]glucose involved passive diffusion across the
plasma membrane or diffusion through a membrane
transporter with low affinity for glucose, we tested the
effect of transport inhibitors. Pulse-labeled hepatocytes were therefore incubated during the chase period
with cytochalasin B, a specific inhibitor of glucose
transporters, or phloretin, an inhibitor of several types
of transporters, including glucose and monocarboxylate transporters. Figure 7, A and B, shows that cytochalasin B did not affect the rate of glucose efflux,
Fig. 7. Release of the intracellular store of glucose is not sensitive to
cytochalasin B (50 ␮M; A) but is impaired by phloretin (0.3 mM; B).
GLUT2(⫺/⫺) hepatocytes were pulse labeled for 30 min with
[14C]pyruvate, washed, and then chased in cold medium in the
presence of the glucose transporter inhibitor cytochalasin B or the
less specific carrier inhibitor phloretin for the indicated periods of
time, and intracellular [14C]glucose was quantitated. Data indicate
intracellular glucose expressed as a percentage of total synthesized
glucose. Data are means ⫾ SE for n ⫽ 4.
whereas phloretin significantly slowed down this release.
To evaluate whether energy was required for this
release, we inhibited oxidative phosphorylation by the
uncoupler FCCP during the chase period and measured the rate of glucose release. Figure 8A shows that
FCCP significantly reduced the rate of glucose release.
Finally, we evaluated the effect of low temperature
incubation (12°C) on the rate of glucose release. Figure
8B shows that glucose release was strongly suppressed
at this temperature.
DISCUSSION
Fig. 6. Release of the intracellular store of glucose is not sensitive to
progesterone. GLUT2(⫺/⫺) hepatocytes were pulse labeled for 30
min with [14C]pyruvate, washed, and then chased in cold medium in
the presence of progesterone (10 ␮g/ml). No inhibition of glucose
release could be observed. Data indicate intracellular glucose expressed as a percentage of total synthesized glucose. Data are
means ⫾ SE for n ⫽ 6.
AJP-Endocrinol Metab • VOL
This study presents novel evidence that glucose release from hepatocytes in the absence of GLUT2 proceeds differently than in its presence. Our pulse-labeling experiments show that neosynthesized glucose can
be secreted at a constant high rate but that there is
also formation of an intracellular pool of glucose. Extracellular release of this intracellular glucose pool
282 • APRIL 2002 •
www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.5 on June 16, 2017
Fig. 5. The intracellular pool of glucose is released without reentering the endoplasmic reticulum. GLUT2(⫺/⫺) hepatocytes were pulse
labeled for 30 min with [14C]pyruvate, washed, and then chased in
cold medium in the presence of the glucose-6-phosphate translocase
inhibitor S-4048 (10 ␮M) for the indicated periods of time. Blocking
the translocase did not reduce the rate of glucose release. Data
indicate intracellular glucose expressed as a percentage of total
synthesized glucose. Data are means ⫾ SE for n ⫽ 4.
GLUCOSE RELEASE FROM GLUT2-NULL HEPATOCYTES
proceeds at a slower rate than release of the bulk of
glucose and is sensitive to different pharmacological
interference. Our data are compatible with the existence of a major pathway for glucose release that compartmentalizes glucose away from the cytosol and that
is based on a membrane traffic mechanism. Separately,
a smaller pool of cytosolic glucose is formed that can be
released slowly from the cells, probably by a facilitated
diffusion process through the plasma membrane, but
that is unable to account for the rapid rate of glucose
output from the hepatocytes. This is shown in Fig. 9.
A major difference in the release of glucose from
control and GLUT2(⫺/⫺) hepatocytes is revealed by
our continuous pulse-labeling experiment. Although no
accumulation of glucose in the cytosol can be observed
in the control cells, in the absence of GLUT2, the
neosynthesized glucose is found predominantly inside
the cell for ⬃30 min, and then the intracellular pool
remains constant for the rest of the pulse labeling. At
the same time, however, a linear accumulation of glucose is measured in the supernatant. Because there is
AJP-Endocrinol Metab • VOL
no delay in appearance of glucose in the culture medium, this suggests that the intracellular pool of glucose does not represent the filling of an obligatory
intermediate compartment in the normal glucose secretory pathway; it may be a side compartment not
involved in the major release pathway.
We previously presented evidence that the major
pathway for glucose release in GLUT2(⫺/⫺) hepatocytes was through membrane traffic issued from the
ER and reaching the plasma membrane without transiting through the Golgi complex (5). We showed that
this pathway was sensitive to low temperature (12°C)
and to the acute effect of progesterone, two conditions
that also slow down the appearance of newly synthesized cholesterol to the plasma membrane (14, 17).
Here, to get further evidence for the involvement of a
membrane traffic mechanism, we evaluated the effect
of substances interfering with actin and microtubule
polymerization. The polymerization of the microfilament was stimulated by jasplakinolide or the phosphatase inhibitor calyculin A. However, in contrast to the
interference of this treatment with capacitative Ca2⫹
entry (10, 11), no impairment of glucose release could
be observed. This may suggest that there is no direct
interaction between the ER and the plasma membrane.
Alternatively, this may be because of a relatively lower
level of actin expression in hepatocytes compared with
the fibroblasts studied in the mentioned reports and,
therefore, to an insufficient density of the subplasma
membrane microfilament mesh to prevent an ERplasma membrane interaction.
The effect of depolymerizing microtubules was, however, very significant, reaching ⬃30% with nocodazole.
Microtubule disruption has been shown in many situations to block membrane traffic and is therefore an
additional evidence that glucose release in GLUT2(⫺/⫺) hepatocytes is through a vesicular pathway.
Similar treatment of control hepatocytes with nocodazole reduced the rate of neosynthesized glucose release
by ⬃15%. Inhibition of glucose transporter by cytochalasin B reduced glucose release by ⬃28%, and a combination of both the membrane traffic and facilitated
diffusion inhibitors reduced release by ⬃48%. This
therefore indicates that, in control hepatocytes, both
pathways participate in glucose release.
The presence of an internal pool of glucose that
reached a maximal level after ⬃30 min of pulse labeling provided the basis for subsequent studies to evaluate how it was released from the cells. First, we could
demonstrate that this pool was most probably within
the cytosol, since it could be completely released from
the cells by streptolysin O permeabilization. Indeed, if
it were partly in a membrane compartment en route to
the plasma membrane, we would have expected it to
remain associated with the cells after this treatment.
The kinetics of glucose release into the culture medium from this cytosolic pool, as studied in pulse-chase
experiments, was much slower than the rate of glucose
release measured in the continuous pulse. This indicates that this intracellular pool was released through
a minor pathway distinct from that taken by the ma-
282 • APRIL 2002 •
www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.5 on June 16, 2017
Fig. 8. The rate of intracellular glucose release is reduced by mitochondrial uncoupling (A) and low temperature (B). GLUT2(⫺/⫺)
hepatocytes were pulse labeled for 30 min with [14C]pyruvate,
washed, and then chased in cold medium in the presence of the
mitochondrial uncoupler carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone (FCCP; 30 ␮M) or at low temperature (12°C) for the
indicated periods of time, and intracellular [14C]glucose was quantitated. Significant inhibition of release was observed in both conditions. Data indicate intracellular glucose expressed as a percentage
of total synthesized glucose. Data are means ⫾ SE for n ⫽ 4.
E799
E800
GLUCOSE RELEASE FROM GLUT2-NULL HEPATOCYTES
jority of glucose. This was further confirmed by assessing the sensitivity of its release to several compounds.
First, blocking the glucose-6-phosphate translocase
with S-4048 did not slow down this glucose release,
indicating that there was no need for glucose reentry in
the ER. The fact that progesterone also did not impair
release further indicated that release did not follow the
same path as the bulk of secreted glucose. Second, in
contrast to the release of most of the glucose, release of
the intracellular pool could be blocked by phloretin,
FCCP, and low temperature but not cytochalasin B.
This further indicated a differential pharmacological
sensitivity of this minor release pathway. It also supported the hypothesis that release was not by nonspecific diffusion across the plasma membrane but involved some specific membrane component. The
sensitivity to phloretin and to energy depletion could
suggest the possible involvement of a membrane protein belonging to the class of ATP-binding cassette
(ABC) transporters. Importantly, because the kinetics
of glucose release from the cytosol of GLUT2-null hepatocytes are not rapid enough to account for the bulk of
glucose output, this therefore excludes the possibility
that glucose is released by an as yet uncharacterized
plasma membrane carrier that would be specific for
glucose and unable to transport 3-MG (see Introduction).
Our previous study on hepatic glucose metabolism in
GLUT2(⫺/⫺) mice during the fed-to-fasted transition
AJP-Endocrinol Metab • VOL
indicated that the abnormal control of glycogen metabolism and paradoxical regulation of glucose-sensitive
genes was associated with, and partly caused by, persistently elevated intracellular G-6-P levels (2). We
postulated that this G-6-P originated from continuous
phosphorylation of glucose reentering the cytosol from
the ER lumen and on the way back to the ER, thereby
forming a futile cycle. Our present data, however, show
that the cytosolic pool of glucose issued from hydrolysis
of G-6-P in the ER leaves the cells without reentering
the ER. This suggests that, even though part of the
cytosolic glucose could be phosphorylated back to
G-6-P, it may not return to the ER lumen. This could be
possible if distinct pools of G-6-P existed with different
capability to eventually enter the ER. This is actually
compatible with a previous report by Christ and
Jungermann (3), demonstrating the existence of separate pools of G-6-P issued from gluconeogenesis and
glycogenolysis.
Together these data bring important new information about the mechanism by which glucose is released
from hepatocytes when GLUT2 is no longer present.
Two pathways can be identified. The minor one consists of the release of a cytosolic pool of glucose by
diffusion through the plasma membrane by a mechanism catalyzed by a so far unidentified carrier. This
pathway is relatively inefficient, as determined by its
very slow kinetics. The major pathway is through a
mechanism that compartmentalizes glucose away from
282 • APRIL 2002 •
www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.5 on June 16, 2017
Fig. 9. Glucose (Glc) release from hepatocytes. Gluconeogenesis and glycogen degradation pathways converge at
the level of G-6-P production. G-6-P enters the endoplasmic reticulum (ER) to be hydrolyzed in glucose and
phosphate, a mechanism that requires the presence of a glucose-6-phosphate translocase (G6T) and of glucose-6phosphatase (G-6-Pase). Glucose release from the cells can take two major pathways that are both present in
control hepatocytes as follows: 1) return of glucose into the cytosol for its release out of the cells by facilitated
diffusion through GLUT2 and 2) release by a membrane traffic pathway, issued from the ER and reaching the
plasma membrane without transiting through the Golgi complex. This pathway is sensitive to progesterone, low
temperature (5), and microtubule depolymerization (present study). In the absence of GLUT2, facilitated diffusion
across the plasma membrane is reduced by at least 95%, but the rate of glucose release from hepatocytes is the
same as from control hepatocytes, indicating that route 2 can quantitatively replace route 1. Here we show that a
part of the newly formed glucose reenters the cytosol to form a pool that can slowly diffuse out the cells (route 3).
This minor release pathway does not involve reentry into the ER, since the G6T inhibitor S-4048 does not block it.
It is, however, inhibited by phloretin, low temperature, and mitochondrial uncoupling by FCCP.
GLUCOSE RELEASE FROM GLUT2-NULL HEPATOCYTES
We gratefully acknowledged Dr. Kaethi Geering for critical reading of this manuscript.
This work was supported by Swiss National Science Foundation
Grant 31-46958.96 to B. Thorens.
REFERENCES
1. Arion WJ, Canfield WK, Ramos FC, Su ML, Burger HJ,
Hemmerle H, Schubert G, Below P, and Herling AW. Chlorogenic acid analogue S3488: a potent competitive inhibitor of
the hepatic and renal glucose-6-phosphatase system. Arch Biochem Biophys 351: 279–285, 1998.
2. Burcelin R, Muñoz MC, Guillam MT, and Thorens B. Liver
hyperplasia and paradoxical regulation of glycogen metabolism
and glucose-sensitive gene expression in GLUT2-null hepatocytes. Further evidence for the existence of a membrane-based
glucose release pathway. J Biol Chem 275: 10930–10936, 2000.
3. Christ B and Jungermann K. Sub-compartmentation of the
“cytosolic” glucose-6-phosphate pool in cultured rat hepatocytes.
FEBS Lett 221: 375–380, 1987.
4. Gerin I, Veiga-da-Cunha M, Achouri Y, Collet J-F, and Van
Schaftingen E. Sequence of a putative glucose 6-phosphate
translocase, mutated in glycogen storage disease type Ib. FEBS
Lett 419: 235–238, 1997.
AJP-Endocrinol Metab • VOL
5. Guillam MT, Burcelin R, and Thorens B. Normal hepatic
glucose production in the absence of GLUT2 reveals an alternative pathway for glucose release from hepatocytes. Proc Natl
Acad Sci USA 95: 12317–12321, 1998.
6. Guillam MT, Hümmler E, Schaerer E, Yeh JY, Birnbaum
MJ, Beermann F, Schmidt A, Dériaz N, and Thorens B.
Early diabetes and abnormal postnatal pancreatic islet development in mice lacking GLUT2. Nat Genet 17: 327–330, 1997.
7. Hemmerle H, Burger HJ, Below P, Schubert G, Rippel R,
Schindler PW, Paulu E, and Herling AW. Chlorogenic acid
and synthetic chlorogenic acid derivatives: Novel inhibitors of
hepatic glucose-6-phosphate translocase. J Med Chem 40: 137–
145, 1997.
8. Ichai C, Guignot L, El-Mir MY, Nogueira V, Guigas B,
Taine E, Mithieux G, and Leverve XM. Glucose-6-phosphate
hydrolysis is activated by glucagon in a low temperature-sensitive manner. J Biol Chem 276: 28126–28133, 2001.
9. Mithieux G. New knowledge regarding glucose-6-phosphatase
gene and protein and their roles in the regulation of glucose
metabolism. Eur J Endocrinol 136: 137–145, 1997.
10. Patterson RL, van Rossum DB, and Gill DL. Store-operated
Ca⫹⫹ entry: evidence for a secretion-like coupling model. Cell 98:
487–499, 1999.
11. Putney JW. “Kissin’ cousins”: intimate plasma membrane-ER
interactions underlie capacitative calcium entry. Cell 99: 5–8,
1999.
12. Rodal SK, Skretting G, Garred O, Vilhardt F, van Deurs B,
and Sandvig K. Extraction of cholesterol with methyl cyclodextrin perturbs formation of clathrin-coated endocytic vesicles. Mol
Biol Cell 10: 961–974, 1999.
13. Shelly LL, Lei KJ, Pan CJ, Sakata SF, Ruppert S, Schutz
G, and Chou JY. Isolation of the gene for murine glucose-6phosphatase, the enzyme deficient in glycogen storage disease
type 1a. J Biol Chem 268: 21482–21485, 1993.
14. Smart EJ, Ying YS, Donzell WC, and Anderson RGW. A role
for caveolin in transport of cholesterol from endoplasmic reticulum to plasma membrane. J Biol Chem 271: 29427–29435, 1996.
15. Thorens B. Glucose transporters in the regulation of intestinal,
renal and liver glucose fluxes. Am J Physiol Gastrointest Liver
Physiol 270: G541–G553, 1996.
16. Thorens B, Guillam MT, Beermann F, Burcelin R, and
Jaquet M. Transgenic reexpression of Glut1 or Glut2 in pancreatic ␤ cells rescues Glut2-null mice from early death and
restores normal glucose-stimulated insulin secretion. J Biol
Chem 275: 23751–23758, 2000.
17. Urbani L and Simoni RD. Cholesterol and vesicular stomatitis
virus G protein take separate routes from the endoplasmic reticulum to the plasma membrane. J Biol Chem 265: 1919–1923,
1990.
282 • APRIL 2002 •
www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.5 on June 16, 2017
the cytosol. It is probably based on vesicular traffic
from the ER to the plasma membrane, which is sensitive to low temperature, progesterone, and microtubule
depolymerization. This pathway is sufficient for the
rate of liver glucose output in the fasted state to reach
the same high level in control and mutant mice. This
pathway coexists in normal hepatocytes with the
GLUT2-dependent pathway. The relative contribution
of each pathway in hepatic glucose output is not
known. However, a recent study on the kinetics of
glucose synthesis from dihydroxyacetone and its secretion from rat hepatocytes showed that the accelerating
effect of glucagon on glucose release was mostly suppressed at 21°C. This could not be explained by
changes in the activity of the involved enzymes (8). It
was suggested that, instead, the temperature-sensitive
effect of glucagon was on regulating the membrane
traffic-based pathway we previously described. This
pathway may therefore represent an important physiological target of glucagon action.
E801