(Received for publication, February 27, 1997, and in revised form, April 16, 1997)
From the Diabetes Research Center, Faculty of Medicine, Vrije
Universiteit Brussel, Laarbeeklaan 103, B-1090 Brussels, Belgium, the
¶ Molecular Nutrition Unit, Department of Nutrition, University of
Montreal, Montreal, Quebec H2L 4M1, Canada, and the Centre de
Recherche L.-C. Simard, Institut du Cancer, Montreal, Quebec H2L 4M1,
Canada
Previous studies in rat islets have
suggested that anaplerosis plays an important role in the regulation of
pancreatic cell function and growth. However, the relative
contribution of islet
cells versus non-
cells to
glucose-regulated anaplerosis is not known. Furthermore, the fate of
glucose carbon entering the Krebs cycle of islet cells remains to be
determined. The present study has examined the anaplerosis of glucose
carbon in purified rat
cells using specific 14C-labeled
glucose tracers. Between 5 and 20 mM glucose, the oxidative production of CO2 from [3,4-14C]glucose
represented close to 100% of the total glucose utilization by the
cells. Anaplerosis, quantified as the difference between 14CO2 production from
[3,4-14C]glucose and [6-14C]glucose, was
strongly influenced by glucose, particularly between 5 and 10 mM. The dose dependence of glucose-induced insulin
secretion correlated with the accumulation of citrate and malate in
(INS-1) cells. All glucose carbon that was not oxidized to
CO2 was recovered from the cells after extraction in
trichloroacetic acid. This indirectly indicates that lactate output is
minimal in
cells. From the effect of cycloheximide upon the
incorporation of 14C-glucose into the acid-precipitable
fraction, it could be calculated that 25% of glucose carbon entering
the Krebs cycle via anaplerosis is channeled into protein synthesis. In
contrast, non-
cells (approximately 80% glucagon-producing
cells) exhibited rates of glucose oxidation that were
to
those of the total glucose utilization and no detectable
anaplerosis from glucose carbon. This difference between the two cell
types was associated with a 7-fold higher expression of the anaplerotic
enzyme pyruvate carboxylase in
cells, as well as a 4-fold lower
ratio of lactate dehydrogenase to FAD-linked glycerol phosphate
dehydrogenase in
cells versus
cells. Finally,
glucose caused a dose-dependent suppression of the activity
of the pentose phosphate pathway in
cells. In conclusion, rat
cells metabolize glucose essentially via aerobic glycolysis, whereas
glycolysis in
cells is largely anaerobic. The results support the
view that anaplerosis is an essential pathway implicated in
cell
activation by glucose.
Pancreatic cells are equipped with a sensing device that
measures the levels of circulating nutrients by processes requiring cellular uptake and metabolism (Refs. 1 and 2; reviewed in Refs. 3-5).
D-Glucose elicits insulin secretion only when extracellular
levels exceed the basal threshold value of 3 mM (6). This
feature has been largely attributed to the enzyme glucokinase, which is
rate-limiting for overall glucose consumption in
cells from rat (7)
and human (8) islets of Langerhans. Although targeted gene disruption
in mice (9) and mutations in human diabetes (10) have strengthened the
concept that glucokinase is a glucose-sensing protein, the following
evidence indicates that (post)mitochondrial events are important for
glucose signaling (4, 5). First, up to 80% of glucose carbon is
oxidized in
cells; this is a very high fraction when compared with
other cell types (11, 12). Moreover, the fraction of total glucose utilization that is further oxidized to CO2 increases in
rat
cells (13) or isolated islets (14) when glycolysis accelerates, whereas this ratio decreases in most cell types (the Crabtree effect;
see Ref. 14). Second, glucose utilization in
cells does not
accelerate under anaerobic conditions (the Pasteur effect), which has
been explained by low lactate dehydrogenase
(LDH)1 and high mitochondrial FAD-linked
glycerol-3- phosphate dehydrogenase (mGPDH) expression (12, 15). Third,
the dimethyl ester of succinate, which is converted to succinate and
oxidized in the Krebs cycle, is a potent "anaplerotic" secretagogue
(4, 16). Fourth, islets cultured at low glucose levels (17) or in the presence of palmitate (18) exhibited parallel suppression of glucose-induced insulin release and down-regulation of pyruvate dehydrogenase (PDH). Moreover, PDH activity is acutely responsive to
glucose (19). Fifth, glucose signaling in the mouse insulinoma cell
line MIN6 is severely blunted by knocking out its mitochondrial genome
(20).
The nature of the metabolic signals required for the activation of
insulin release remains largely unknown. An increase in glucose
metabolism causes a rise in islet [ATP]/[ADP] (2). This results in
the closure of KATP channels (21), membrane depolarization,
and influx of Ca2+, which is required for exocytosis (3).
Nonetheless, electrophysiological studies have suggested that other
glucose-derived signals are required for exocytosis as well (22). This
study explores the concept that anaplerosis is one of the
KATP-independent signal-generating pathways for
glucose-induced insulin release (4, 23, 24). Anaplerosis,
i.e. filling up the Krebs cycle with intermediates that are
channeled into anabolic pathways (4), requires conversion of pyruvate
into oxaloacetate by pyruvate carboxylase (PC). Metabolic flux through
this enzyme is quantitatively important in rat islets (25) due to the
abundant expression of PC (23, 24). Stimulated flow of intermediates
from the Krebs cycle into synthesis of coupling factors or
macromolecules is believed to sustain stimulus-secretion coupling and
cell growth. One of the hypothetical coupling factors is
malonyl-CoA, which is produced by acetyl-CoA carboxylase (ACC). Glucose
rapidly increases citrate content of the cells, which in turn may
activate ACC (23). Chronic stimulation with glucose increases ACC
mRNA abundance in INS-1 cells (26). Because ACC activity in islets
is ten times higher than fatty acid synthase activity (23), malonyl-CoA
accumulates rapidly in the cells (27). Thus, malonyl-CoA, the
physiological inhibitor of carnitine palmitoyltransferase I (28), may
act as a metabolic coupling factor in
cell signaling (4, 5, 29,
30). Anaplerosis might also be instrumental in accelerating the
pyruvate/malate shuttle (24), which results in the formation of
cytosolic NADPH. Finally, because glucose acutely enhances both
transcription and translation in
cells, anaplerosis is expected to
be required for de novo synthesis of metabolic precursors
for regulated gene expression and growth.
Our recent work (31, 32) has shown that glycolysis accelerates
proportionally to extracellular glucose between 1 and 10 mM
in purified cells and non-
cells. Moreover, when differences in
cellular volume are taken into account, the rate of glycolysis is
similar in
cells and non-
cells (31). This similarity can be
explained by the presence of glucokinase in both
cells and
cells (32) and by the fact that glucose transport is not rate-limiting
for glucose metabolism (31). How can glucose, which is metabolized
similarly in glycolysis, activate
cells while it inhibits
cell
function? Part of the answer to this question may be found by
considering anaplerosis in
cells and
cells. Indeed, it is
intriguing that about 40% of glucose-derived carbon entering the
citric acid cycle is carboxylated in rat islets (33). This extent of
anaplerotic input is unusually high for a non-gluconeogenic (34) and
non-lipogenic (23) tissue. Thus, approximately 40% of the carbons of
glucose that enter the krebs cycle must leave it for a non-oxidative
fate, which remains to be assessed. This consideration and the fact
that the major anaplerotic enzyme PC is expressed at very high levels
in rat islets (24, 35) and
(INS-1) cells (23) suggest an important
function for pyruvate carboxylation in the
cell.
To further extend our understanding of the role of anaplerosis in islet
tissue, we have endeavored to provide answers to the following three
questions: (i) is glucose-induced anaplerosis a unique property of cells or is it also present in islet non-
cells? (ii) Can the
magnitude of glucose-induced anaplerosis in
cells be quantified?
(iii) What is the fate of the glucose carbon entering the mitochondria
as pyruvate but escaping the Krebs cycle? We addressed these questions
by comparing glucose regulation of anaplerosis in
fluorescence-activated cell sorter-purified
cells and non-
cells
(approximately 80% glucagon-producing
cells; Ref. 36). The present
study demonstrates that glucose metabolism distal from pyruvate differs
markedly in
cells versus non-
cells.
Labeled compounds were purchased from Amersham Corp. (D-[5-3H]glucose (13 Ci/mmol), D-[U-14C]glucose (292 mCi/mmol), D-[1-14C]glucose (54 mCi/mmol), D-[6-14C]glucose (56 mCi/mmol), L-[3,5-3H]tyrosine (52 Ci/mmol), NaH14CO3 (54 mCi/mmol), and 3H2O (5 mCi/ml)) or from NEN Life Science Products (D-[3,4-14C]glucose (54 mCi/mmol)). Other chemicals were from Sigma or Merck (Darmstadt, Germany).
Islet Cell Preparation and Culture of INS-1 CellsRat islet
and non-
cells were purified from adult male Wistar rats by
autofluorescence-activated cell sorting, as has been described
previously (36). Purity of the
cell preparations was more than
90%, as determined by immunocytochemistry for insulin or by electron
microscopical analysis. Non-
cells were composed of >80%
cells, 5-10%
cells, and 10-15% other cells. INS-1 cells (37)
were cultured at 37 °C in 95% air-5% CO2 in RPMI 1640 medium supplemented with 10% heat-inactivated fetal calf serum (Life
Technologies, Inc., Paisley, United Kingdom), 2 mM L-glutamine, 1 mM sodium pyruvate, 10 mM Hepes, 50 µM
-mercaptoethanol, 100 units/ml penicillin and 100 µg/ml streptomycin. The attached cells
were harvested from culture dishes (Nunc, Roskilde, Denmark) by a 5-min
treatment with Mg2+/Ca2+-free
phosphate-buffered saline, pH 7.4, containing 0.025% (w/v) trypsin
(Boehringer Mannheim).
Glucose
metabolism was measured in batches of freshly isolated cells (5 × 104 cells and 1 × 105 non-
cells) over
2-h incubations at 37 °C in 100 µl of Earle's-Hepes buffer (36)
containing the indicated concentrations of glucose. Conversion of
[5-3H]glucose (50 µCi/ml; specific activities, 1.3-25
Ci/mol) to 3H2O was measured simultaneously
with production of CO2 from 14C-labeled
D-glucose tracers (11). Total radioactivity added to the
cells was 50 µCi/ml for [1-14C]glucose and
[6-14C]glucose, 25 µCi/ml for
[U-14C]glucose, and 2-5 µCi/ml for
[3,4-14C]glucose, resulting in specific activities of
0.1-50 Ci/mol. Cellular metabolism was stopped by the addition of 20 µl of 0.4 mol/liter citrate buffer, pH 4.9, containing 5 mM KCN, 10 µM antimycin A, and 10 µM rotenone. Hydroxyhyamine (Hewlett-Packard) was used to
capture the produced 14CO2. Tritiated water and
14CO2 production were measured via liquid
scintillation counting. Calculations of glucose oxidation took into
account the CO2 recovery (89 ± 1%), which was
assessed by adding 0.05 µCi NaH14CO3 to
separate incubation vials without cells. The contribution of the
pentose phosphate pathway to total glucose utilization was calculated
from the specific yields of 14CO2 from
[1-14C]glucose and [6-14C]glucose oxidation
as has been explained in detail in Ref. 38. The flux of glucose-related
anaplerosis was quantified as the difference between
[3,4-14C]glucose oxidation and
[6-14C]glucose oxidation, which reflects the amount of
glucose carbon entering the Krebs cycle without being oxidized to
CO2 (25). Cellular radioactivity accumulating during a 2-h
incubation of
cells with [U-14C]glucose was measured
after the cells were washed four times with 0.8 ml of incubation
medium, sonication in either 1 M perchloric acid or 10%
TCA, and liquid scintillation counting of radioactivity in the
acid-soluble and acid-precipitable fractions.
Cells were grown in
21-cm2 Petri dishes in regular RPMI medium at 11 mM glucose. They were then preincubated for 3 days in culture medium at 5 mM glucose because we have previously
shown that such protocol allows a robust secretory response to high glucose (26). The culture medium was removed, and attached cells (8 × 106) were washed twice with phosphate-buffered
saline and preincubated for 30 min at 37 °C in Krebs-Ringer
bicarbonate medium (KRB) containing 10 mM Hepes (pH 7.4),
0.07% bovine serum albumin, and 4 mM glucose. Cells were
then washed twice with phosphate-buffered saline and incubated for 30 min in KRB-Hepes medium containing 0.07% bovine serum albumin and
different glucose concentrations. Incubation media were collected to
determine insulin release (23). For measurements of citrate and malate
accumulation, cells were scraped from the dishes after addition of 10%
TCA. Precipitated proteins were removed by centrifugation, and the
supernatants were extracted five times with ether. Samples were
lyophilized and stored at 80 °C. The assays of citrate and malate
are described in detail in Ref. 39.
The expression level of PC was measured as described
previously (23) by direct streptavidin blotting. Samples corresponding to 25 µg of protein from INS-1 cells, 2 × 105
flow-sorted cells, and 6.6 × 105 non-
cells
were loaded per lane on 5% SDS-polyacrylamide gels. As verified in
each experiment by Ponceau red staining, the amount of total protein
load per lane was similar for the three cell types.
The effect of cycloheximide (40 µM) on total protein synthesis was assessed by incubating
purified cells for 2 h at 10 mM glucose in the
presence of L-[3,5-3H]tyrosine (250 µCi/ml;
4.8 µM). The amount of newly synthesized protein was
assessed by precipitating cellular extracts in 10% TCA followed by
liquid scintillation counting (40). For the incorporation of
[U-14C]glucose into proteins, batches of 1.5 × 105
cells were incubated for 2 h at 37 °C in
100 µl of Earle's-Hepes buffer containing 25 µCi/ml
[U-14C]glucose (concentration, 10 mM;
specific activity, 2.5 Ci/mol) either without or with 40 µM cycloheximide. Cells were washed four times with
incubation medium to remove the excess of extracellular tracer and
homogenized in 10% TCA. The amount of newly synthesized protein was
assessed by liquid scintillation counting of the TCA precipitates (40)
and calculated as the difference between control cells and cells
incubated with cycloheximide.
Measurement of LDH activity in purified and
non-
cells was done as described by Sekine et al. (12)
with minor modifications. Pellets from 3 × 105 cells
were suspended in 10 mM Tris-HCl buffer, pH 7.0, containing 2 mM EDTA, 1 mM DTT and lysed by three cycles
of freeze-thawing. Enzymatic activity was determined in the supernatant
fraction after centrifugation (2 min at 104 × g) at 4 °C in 20 mM Hepes buffer, pH 7.2, in
the presence of 0.05% (w/v) bovine serum albumin (fraction V; Sigma),
20 µM NADH, and 2 mM pyruvate (monosodium
salt, freshly prepared solution); the consumption of NADH was measured
in a Hitachi F-2000 fluorometer (Hitachi, Tokyo, Japan) at 340/460 nm.
Determination of mGPDH (EC 1.1.99.5) was performed as described by
MacDonald (15) using batches of 2 × 105
cells and
non-
cells.
Glucose utilization in cells, as
assessed from the production of 3H2O from
[5-3H]glucose, was markedly dependent on the
extracellular glucose concentration (Fig.
1A). Confirming our previous observations in
flow-sorted
cell subsets (11), the fraction of total glucose utilization that was further oxidized to CO2 was high using
[U-14C]-labeled glucose as substrate: 69% at 1 mM glucose, 75% at 5 mM glucose, and 82% at
10 mM glucose. Because data obtained with this tracer
reflect mean oxidation rates of all six carbon atoms of glucose, no
information can be retrieved about oxidative pathways that are
preferentially used by the cells. We therefore incubated
cells with
glucose tracers that were labeled at specific carbon atoms. Oxidation
of [6-14C]glucose to 14CO2,
reflecting Krebs cycle activity at the level of isocitrate dehydrogenase and
-ketoglutarate dehydrogenase, increased 30 times
when medium glucose was raised from 1 to 20 mM (Fig.
1B). However, the specific yield of
14CO2 from [6-14C]glucose was
much lower than that of [U-14C]glucose. To explain this
difference, we considered two possibilities. First, oxidative pathways
in addition to the Krebs cycle, in particular the pentose phosphate
pathway, contribute to total glucose oxidation in
cells. Second,
high anaplerotic input into the cycle provides a possible explanation
for this phenomenon. Thus, with respect to glucose labeled in position
6, anaplerosis allows dilution of the [14C] tracer among
Krebs cycle intermediates at a carbon position that does not yield
14CO2 during the first and second round in the
cycle. In addition, some labeled intermediates escape the cycle
(cataplerosis; Fig. 2A) during the first two
rounds. By contrast, uniformly labeled glucose yields labeled
CO2 during the first two cycle rounds.
To examine the first possibility, we compared [1-14C]glucose and [6-14C]glucose oxidation. Whereas both [1-14C]glucose and [6-14C]glucose enter the citric acid cycle as [2-14C]acetyl-CoA, only [1-14C]glucose produces 14CO2 during the oxidative part of the pentose phosphate pathway (38). Fig. 1B shows that the specific yield of 14CO2 obtained after incubating the cells with [1-14C]glucose and [6-14C]glucose was almost the same over the whole tested range of glucose concentrations. This indicates that the oxidative flux through the pentose phosphate pathway is very low and is detectable only when cells were exposed to 5 mM substrate or less (range between 0.3 and 0.6 pmol of glucose equivalents/2 h/103 cells). The contribution of the pentose phosphate pathway to overall glucose utilization declined with rising glucose levels: 18 ± 4%, 10 ± 3%, and 6 ± 5% at 1, 2.5, and 5 mM glucose, respectively, and below detection limit at 7.5 mM glucose or higher. Low contribution of the pentose phosphate pathway to overall glucose utilization has been described in rodent islets (25, 41), but acute suppression of this pathway by elevated glucose has not, to our knowledge, been reported before.
A consequence of a low contribution of the pentose phosphate pathway to overall glucose utilization is that [3,4-14C]glucose will be converted to [1-14C]pyruvate (Fig. 2B). [1-14C]Pyruvate enters the mitochondria, where it is either decarboxylated by PDH, yielding 14CO2, or carboxylated by PC, yielding [1-14C]oxaloacetate (Fig. 2B). This intermediate enters the Krebs cycle, where it produces 14CO2 during the first cycle or exits from the cycle to reconvert to [1-14C]pyruvate via the citrate-pyruvate or pyruvate-malate shuttles (Fig. 2B). Therefore, even in the presence of anaplerosis, almost all of the label in [3,4-14C]glucose that is converted to [1-14C]pyruvate is recovered in the form of 14CO2 (Fig. 2B).
The production of 14CO2 from
[3,4-14C]glucose rose 25-fold upon increasing glucose
from 1 to 20 mM (Fig. 1B). The rate of
[3,4-14C]glucose oxidation was higher than
[U-14C]glucose oxidation. Second, in the range of 5-20
mM glucose, almost all glucose that was utilized was
further oxidized as can be seen when glucose usage is compared with
[3,4-14C]glucose oxidation (Fig. 1, A and
B). Third, the rate of [3,4-14C]glucose
oxidation was about 2-fold higher than oxidation of [1-14C]glucose or [6-14C]glucose. As
discussed above, anaplerosis may explain why the production of
14CO2 from [6-14C]glucose is
relatively low in comparison to that observed with [3,4-14C]glucose. Thus, glycolytic conversion of
[6-14C]glucose to [3-14C]pyruvate results
in the production of [2-14C]acetyl-CoA, which requires
two complete cycles before any 14CO2 is
produced (Fig. 2C). When no anaplerosis occurs, the specific activity of all Krebs cycle intermediates rises, until the input of
radioactivity as [2-14C]acetyl-CoA equals that of
14CO2-production. In that case, no difference
with 14CO2-production from
[3,4-14C]glucose is detected. However, when anaplerosis
occurs, extra glucose carbon entering the cycle as
[3-14C]oxaloacetate will be diverted into anabolic
pathways (cataplerosis; Fig. 2). In that case, radioactivity entering
both as [2-14C]acetyl-CoA and
[3-14C]oxaloacetate has ample time to escape oxidation.
Therefore, the difference between 14CO2
production from [3,4-14C]glucose and
[6-14C]glucose is an index of the anaplerotic flux of
glucose carbon into the Krebs cycle. As indicated in Fig.
3, differences between [3,4-14C] glucose
oxidation and [6-14C]glucose oxidation exhibited a
sigmoidal substrate concentration dependence with half-maximal effect
at 7.5 mM glucose. The quantitative importance of this
difference and the glucose dependence between 5 and 10 mM
are both apparent (Fig. 3).
In summary, two conclusions emerge from this series of experiments.
First, the ratio of glucose oxidation to glucose utilization (in the
range of 5-20 mM glucose) is close to 1 in cells, when the production of 3H2O from
[5-3H]glucose is compared with
[3,4-14C]glucose oxidation. Second, an important
difference exists between the degree of oxidation of carbons 3 and 4 of
glucose versus that of carbons 1 and 6. This difference can
be explained by a high anaplerotic input of glucose carbon into the
citric acid cycle.
Fully compatible with the view that pyruvate carboxylation
is high in cells are the measurements of the INS-1 cell content of
citrate and malate (Fig. 4). INS-1 cells were used for
these experiments because the low amount of available material obtained precluded the same measurements in flow-sorted islet
cells. It
should be emphasized that the rates of glucose utilization and
oxidation of INS-1 cells are similar to fluorescence-activated cell
sorter-purified
cells (data not shown; see also Ref. 12). Fig. 4
shows that glucose caused a dose-dependent rise in the cellular contents of citrate and malate, which closely correlated to
the dose dependence of glucose-induced insulin release. Half-maximal and maximal effects of glucose on insulin secretion and the
accumulation of both metabolites were observed at 10 and 16 mM glucose, respectively.
Comparison of D-Glucose Utilization and Oxidation in Purified
We next assessed the cell
specificity of the metabolic organization in cells by comparing
glucose oxidation and utilization in islet
cells and non-
cells
(approximately 80%
cells; Ref. 36). When normalized for the
differences in cellular volume, glycolytic activity at 10 mM glucose was the same in
cells and non-
cells
(Table I). This similarity has been described before (31) and was explained by the presence of glucokinase in both
cell
and non-
cell preparations (32). In contrast to the comparability in
rates of total glucose utilization, three important differences between
cells and non-
cells were noted at the level of mitochondrial
glucose metabolism. First, rates of glucose oxidation were different in
the two cell types, particularly at high glucose levels (Table I). For
[3,4-14C]glucose oxidation, this difference was already
noted at 1 mM substrate; increasing the glucose level to 10 mM accelerated [3,4-14C]glucose oxidation
20-fold in
cells as compared with only 7-fold in non-
cells.
Using [6-14C]glucose as a substrate, no differences
between
cells and non-
cells were noted at 1 mM, but
oxidation was again more accelerated at 10 mM glucose in
cells than in non-
cells. Second, the ratios of glucose
oxidation to total glucose utilization were high in
cells and
increased with rising substrate concentrations, whereas in non-
cells these ratios were much lower and were independent of the
substrate concentration. The difference was the largest for the ratio
of [3,4-14C]glucose oxidation to total glucose
utilization, which was up to 6 times higher in
cells than in
non-
cells. Third, rates of [3,4-14C]glucose oxidation
exceeded those of [6-14C]glucose oxidation in
cells
approximately by a factor of 2, whereas in non-
cells, no
differences were observed. This indicates that pyruvate carboxylation
in non-
cells was below the detection limit of our experimental
system.
|
Together, these data indicate that glucose metabolism in islet and
cells is similar at the level of glycolysis but diverges markedly
beyond pyruvate formation. In
cells, aerobic glycolysis is followed
almost exclusively by pyruvate channeling to the mitochondrion. A major
extent of the pyruvate carbons enter the citric acid cycle by
anaplerosis in
cells. By contrast, glycolysis in non-
cells is
mostly anaerobic, and anaplerosis with glucose-derived carbons is
extremely low in these cells.
Low LDH activity and high
expression of mGPDH in flow-sorted rat cells has been described
before (12) and was considered important for the strictly aerobic mode
of glycolysis in these cells. Because flow-sorted non-
cells exhibit
a much more anaerobic glycolysis than
-cells (Table I) and because
mGPDH has not yet been determined in these cells, we felt that the data
of Sekine et al. (12) needed confirmation, with the
extension of measuring mGPDH activity in non-
cells. As Table
II shows, LDH activity was approximately 2-fold higher
in non-
cells than in
cells. Furthermore, mGPDH activity was
2-fold higher in
cells. Consequently, the LDH/mGPDH activity ratio
was 4 times more elevated in non-
cells as compared with
cells
(p < 0.01), favoring anaerobic glycolysis in non-
cells and and aerobic glycolysis in
cells.
|
PC is the major anaplerotic enzyme in mammalian cells (4) and is very
abundant in islet tissue, accounting for up to 0.4% of total islet
protein (35). However, the distribution of PC among different islet
cells is not known. The results in Fig. 5 indicate that
PC is expressed at a high level in both INS-1 cells and purified rat
cells, whereas on the other hand, the enzyme is not abundant in
flow-sorted rat non-
cells. Densitometric analysis of the results
from three independent cell preparations in each group indicated that
PC abundance is about 7-fold higher in
cells than in non-
cells.
Taking into account that the average volume of
cells is 2.5 times
that of non-
cells and that 3-fold more non-
cells were loaded on
the polyacrylamide gels, it can be estimated that the
cells
contain, on the average, about 20 times (7 × 3) more PC than
non-
cells. Knowing that non-
cells are contaminated to the
extent of 5% by
cells, the difference between the expression level
of PC between
cells and
cells is likely much larger. Thus, the
data in Fig. 5 are in accordance with the observed high anaplerosis of
glucose-derived carbon in
cells only.
Protein Synthesis from Glucose Carbon
To investigate the
metabolic fate of glucose carbon that is utilized during glycolysis but
escapes mitochondrial oxidation to CO2, purified cells
were labeled for 2 h with [U-14C]glucose, washed,
and homogenized in either 1 M perchloric acid or 10% TCA.
The effect of glucose upon the accumulation of 14C-labeled
molecules was more pronounced in the acid precipitate than in the
acid-soluble fraction: mean values ± S.E. at 1 and 10 mM glucose from 3 experiments were 0.5 ± 0.1 and
4.2 ± 0.4 pmol glucose equivalents/103 cells,
respectively, for the perchloric acid supernatants (8-fold difference)
and 0.1 ± 0.02 and 5.1 ± 0.6 pmol glucose
equivalents/103 cells for the perchloric acid pellet
(50-fold difference). Total accumulation of 14C-labeled
molecules (pmol glucose equivalents) in the cellular extracts (TCA
supernatant plus pellet) was the same as the difference between glucose
utilization and CO2 production (Table III).
This indirectly confirms the concept (12) that very little lactate is
produced and exported by these cells. Because D-glucose has a profound stimulatory effect on protein synthesis in pure
cells, even in the absence of exogenous amino acids (11, 40), a fraction of
the anaplerotic/cataplerotic carbon flux from
[14C]glucose into the TCA precipitable material may
represent newly synthesized proteins. To estimate the size of this
anabolic pathway, purified
cells were incubated at 10 mM glucose with cycloheximide, which inhibited 95% of
total protein synthesis (Table III). Glucose utilization and oxidation
remained unchanged in the presence of cycloheximide. However, the
14C radioactivity recovered from the TCA pellet was reduced
by about 50% of the control value in the presence of the protein
synthesis inhibitor (p < 0.01) (Table III). Thus,
glucose-derived carbon entering the citric acid cycle via anaplerosis
is channeled into glucose-stimulated synthesis of proteins. This
anabolic pathway accounts for up to 50% of the acid-precipitable
14C radioactivity accumulating in
cells and 25% of the
glucose carbon entering the Krebs cycle via anaplerosis.
|
Our results show that glucose metabolism in rat
cells is essentially aerobic, supporting the idea (12) that lactate
output is minimal in these cells. Studies using isolated islets (25, 42) reported lower ratios of glucose oxidation over glucose utilization
(between 20 and 30% in most studies). We considered the possibility of
oxidative glucose metabolism being an artifact of isolated
cells
unlikely. First, isolated
cells are responsive to acute glucose
stimulation, both in terms of glucose-regulated proinsulin biosynthesis
(11) and glucose-induced insulin release (40). Second, glucose
metabolism and glucose sensing are very well correlated in these cells
(11). Third, whereas glucagon, somatostatin and other agents that alter
cellular cyclic AMP levels markedly influence the amplitude of
glucose-induced insulin secretion from isolated
cells (43), these
islet hormones do not affect the rate of glucose oxidation in purified
cells (44). Fourth, oxidative glucose metabolism is not a feature
of freshly isolated
cells because it is present in
cells
maintained for 10 days in culture (45).
What may provide an explanation for the much lower oxidative fraction
of glucose metabolism in whole islets? Anaerobic glycolysis is a likely
factor, because lactate output from whole islets is much higher (2 pmol/min/islet; Ref. 42) than in purified cells (12). Islet non-
cells (endocrine or non-endocrine) with elevated rates of anaerobic
glycolysis may be responsible for this difference. On the basis of the
present study, we believe that isolated islet non-
cells contribute
little to lactate output from whole islets. Indeed, they represent 20%
of the total islet mass and exhibit rates of glucose utilization that
are very similar to those of
cells (31), whereas their oxidative
capacity is indeed lower and their LDH activity is 2-fold higher than
that of
cells (Table II). Contaminating cells in islets with very
high rates of anaerobic glycolysis (e.g. exocrine cells) may
be considered. Alternatively, high rates of lactate output may also
originate from
cells in the centers of isolated islets that are
prone to oxygen depletion and necrosis (46). Taken together, the large
differences in the ratio of glucose oxidation to glucose utilization
between whole islets on the one hand and flow-sorted
cells or INS-1 cells on the other hand stress the importance of the choice of the
experimental model in metabolic studies related to glucose sensing.
Anaplerosis ensures the maintenance of the pool of Krebs
cycle intermediates when these are simultaneously used for biosynthetic purposes, such as synthesis of proteins, lipids, or heme (47). When
fueled by glucose metabolism, anaplerosis is initiated by PC, which
catalyzes the formation oxaloacetate from pyruvate. Glucose has been
reported to induce the expression of the PC gene in rat islets (17).
The islet abundance of PC is only equaled by gluconeogenic tissues such
as the liver and kidney (23, 24). MacDonald (24) provided evidence for
a new function of islet PC in cell signaling, i.e. the
rapid formation of oxaloacetate serving as a substrate for the
pyruvate-malate shuttle, which provides cytosolic NADPH, a putative
coupling factor. The present study provides support for this
hypothesis. Thus, high PC expression and glucose-regulated anaplerosis
have been observed in
cells but not in non-
cells. In addition,
the dose dependence of malate accumulation in INS cells correlated
closely with that of insulin release.
Measurements in purified cells indicate that anaplerosis is
quantitatively important and exquisitely regulated by glucose. In
agreement with this observation is the fact that in cultured islets,
approximately 60% of pyruvate entering the mitochondria is directly
oxidized via PDH, whereas the remaining 40% is carboxylated by PC
(33). The relative pyruvate use via PC and PDH could not be calculated
in the present study because we have not measured the yields of
14CO2 from [1,4-14C]
versus [2,3-14C]succinate,
[1-14C] versus [2-14C]acetate
and [2-14C] versus
[6-14C]glucose (33). Instead, the magnitude of
anaplerosis was estimated by a radiometric method using
[6-14C]glucose and [3,4-14C]glucose. The
estimated rates of 14CO2 production from
glucose were highest when [3,4-14C]glucose was used as a
tracer, reaching 90-100% of total glucose consumption when glucose
levels exceeded 5 mM (Fig. 1 and Table I). On the other
hand, 14CO2 production from
[6-14C]glucose was much lower. As is explained in detail
in Fig. 2 and under "Results," anaplerosis can account for such
difference. Because the estimated value of this flux is markedly
regulated by glucose (Fig. 3), it can be postulated that the
anaplerotic flux is correlated and perhaps even responsible for the
glucose-dependent synthesis of metabolic coupling
factor(s). Indeed, the concentration dependence of this flux is clearly
sigmoidal and exhibits the most pronounced glucose dependence between 5 and 10 mM substrate, the range at which insulin secretion
is physiologically regulated.
Interestingly, using the same radiometric technique, no anaplerotic
flux of glucose carbon was detectable in purified islet non- cells
(Table I). This observation is consistent with the much lower PC
protein abundance in these cells and contrasts with the similarity in
total glucose utilization (31) and glucokinase expression (32) between
and non-
cells. These data are also consistent with the
observation that glucose increases NADPH autofluorescence in rat
cells but not in rat non-
cells, a phenomenon that has been the
basis for autofluorescence-activated separation of islet cells (36).
However, the possibility cannot be discounted that glucose-induced
anaplerosis and high expression of PC are present in
somatostatin-secreting
cells, which are stimulated by glucose and
which constitute a small (about 10%) fraction in the sorted non-
cell preparations (36).
Despite the fact that it is a difficult task when
studying islet cells, the present paper has attempted to delineate some pathways into which anaplerosis of glucose carbon is channeled. For
this purpose the cells were labeled with [U-14C]glucose
and acid extracted. Whereas glucose increased the radioactivity in both
the acid-soluble and acid-insoluble pools, the effect on the
acid-insoluble pool was much larger. This difference may be explained
as follows. The acid-soluble pool might represent a whole range of
metabolites whose specific activities and concentrations rise as a
consequence of accelerated glucose metabolism. In contrast, the
increase in radioactivity in the acid-insoluble fraction might reflect
not only the specific activities of the precursors but also the
activation of biosynthetic processes by glucose. The best studied
glucose-regulated biosynthesis in islet cells is that of insulin and
other islet cell proteins (11, 40, 48). The results with cycloheximide
indicate that glucose carbon is channeled into de novo
synthesis of amino acids required for glucose-stimulated protein
synthesis in cells: cycloheximide prevented incorporation of 2.6 pmol/103 cells/2 h of glucose equivalents into
TCA-precipitable material, which corresponds to 45% of radioactivity
accumulating in the acid-insoluble pool. Freshly isolated
cells
synthesize 67 fmol of preproinsulin/103 cells/2 h at 10 mM glucose (40), requiring 7.5 pmol of aminoacyl tRNA.
Using [3H]tyrosine and [3H]histidine as
markers, it was further estimated that at least the same amount of
amino acid is required by purified
cells to synthesize non-insulin
proteins (40), bringing the total cellular need to 15 pmol of amino
acids/2 h/103 cells. Thus, it appears that glucose carbon
contributes significantly to the pool of nonessential amino acids
required for protein synthesis.
The remaining 55% of glucose carbon accumulating in the acid-insoluble
fraction of the cell extracts may represent labeling of other
macromolecules, such as aminoacyl tRNAs, mRNAs, lipids, and
glycogen. Because chronic exposure of rat cells to 10 mM glucose leads to modest accumulation of glycogen in the
cells (45), the amount accumulating during 2 h of incubation may
be small compared with overall glucose utilization. Similarly, previous studies (23, 49), as well as lipid extractions of
cells labeled
with [U-14C]glucose, indicate that glucose conversion
into lipids represents a small fraction (5% or less) of total glucose
utilization (data not shown). Furthermore, because accelerated
glycolysis and glycerol phosphate production are likely to enhance
esterification of unlabeled fatty acids to
[U-14C]glycerol phosphate (49), most of the radioactivity
in the lipid extracts may be due to the glycerol part of the molecules (49) and not related to anaplerosis. With respect to the concept proposing that an "anaplerotic malonyl-CoA" pathway and lipid esterification are implicated in the short term or long term control of
insulin release (4, 5, 23, 30), the present study has added one
supporting piece of evidence: both the anaplerotic flux of
glucose-derived carbon and the accumulation of cellular citrate, the
carbon precursor of malonyl-CoA and allosteric activator of ACC (50),
correlate with insulin secretion.
The contribution of the
pentose phosphate pathway to total glucose utilization was low, and it
decreased with rising glucose levels. Previous studies (25, 41) have
already pointed out the low activity of this pathway in islets. This
study extends these observations by showing that low pentose phosphate
pathway activity is a metabolic feature of cells. Furthermore, the
present data suggest that acceleration of glucose metabolism acutely
suppresses flux through this pathway. The suppression can be explained
by the following chain of events: glucose-accelerated metabolism increases anaplerosis (Ref. 25 and this study), accelerating the
pyruvate/malate shuttle (24) and resulting in an increased cytosolic
[NADPH]/[NADP+] ratio (36, 48); this condition is
likely to inhibit glucose-6-phosphate dehydrogenase, the
flux-generating enzyme of the pentose phosphate pathway, as has been
described in most cells (47). Thus, the dose dependence of the glucose
carbon flux through the pentose pathway does not correlate with insulin
release, indicating that pentose phosphate pathway intermediates are
not directly implicated in the short term control of insulin
secretion.
The organization of glucose metabolism distal
from glycolysis is different in pancreatic cells versus
non-
cells. Glycolysis is aerobic in
cells and an important part
of glucose carbon is diverted into anabolic pathways, in particular the
de novo synthesis of amino acids and proteins. Anaplerosis
is regulated by glucose in
cells only, and this correlates with the
high expression level of PC in this cell type. In contrast, glycolysis is mainly anaerobic in
cells, and the low expression of PC and high
LDH content explain why glucose-dependent anaplerosis
is undetectable in these cells.
We thank Erik Quartier for technical assistance, An Gielen for secretarial help, and Lutgard Heylen and Geert Stangé for islet cell purification. We acknowledge Drs. Claes Wollheim and Maryam Asfari (University of Geneva, Geneva, Switzerland) for generously providing us with the INS-1 cell line.