- Email: [email protected]
Glucose metabolism in insulin-producing tumoral cells
ARCHIVES OF BIOCHEMISTRY Vol. 241, No. 2, September,
AND BIOPHYSICS pp. 561-5’70, 1985
Glucose Metabolism MARIE-HaLfiNE FRANCINE Labcwatoq
GIROIX, ABDULLAH MALAISSE-LAGAE,
of Expekmental Received
in Insulin-Producing
January
Medicine,
Brussels
Tumoral
Cells’
SENER, SIMON P. DUFRANE, WILLY J. MALAISSE’
AND
Free
28, 1985, and in revised
University, form
Brussels May
115, Belgium
1’7, 1985
Homogenates of insulin-producing tumoral cells catalyzed the phosphorylation of glucose, mannose, and fructose. The kinetics of phosphorylation at increasing glucose concentrations, the inhibitory effect of glucose 6-phosphate, and the comparison of results obtained with distinct hexoses indicated the presence of both low-K, hexokinaselike and high-K, enzymatic activities, the results being grossly comparable to those collected in normal pancreatic islets. Relative to protein content, the glucose-phosphorylating enzymatic activity was higher in tumoral than normal islet cells. The activity of other enzymes was either lower (glutamate dehydrogenase), moderately higher (phosphoglucomutase, lactate dehydrogenase) or considerably greater (ornithine decarboxylase) in tumoral than in normal islet cells. In intact tumoral cells, incubated under increasing glucose concentrations, the oxidation of D-[U-‘4Clglucose and the output of lactic and pyruvic acids reached a close-to-maximal value at 2.8 mM glucose. The ratios for glucose oxidation/utilization and lactate/pyruvate output were much lower in tumoral than in normal islet cells. Although glucose caused a modest increase in insulin output from the tumoral cells, this effect was saturated at a low glucose concentration (2.8 mM) and less marked than that of other secretagogues (e.g., Lleucine, L-ornithine, or forskolin). Thus, despite a close-to-normal enzymatic equipment for glucose phosphorylation, the tumoral cells displayed severe abnormalities in the metabolism and secretory response to this hexose. These findings point to regulatory mechanisms distal to glucose phosphorylation in the control of glucose metabolism in insulin-prod.ucing cells. o 19% Academic PMS, IX
Insulin-producing tumoral cells of the RIN type are poorly responsive to glucose as compared to other nutrient or nonnutrient secretagogues (l-3). The biochemical determinant(s) of such a poor responsiveness remains to be elucidated. In one study, it was proposed that the failure of glucose to stimulate insulin release was attributable to an abnormal glucose metabolism in the tumoral cells, especially the lack of a high-Km glucokinase-like activity (4). Yet, in another study, gluco-
kinase was purified from the same tumor line (5). In the present study, we have investigated both the kinetics of glucose phosphorylation in cell homogenates and the oxidation of glucose and output of lactic and pyruvic acids in intact insulinproducing tumoral cells. Our results suggest that, by comparison with normal pancreatic islet cells, abnormal glucose metabolism in tumoral cells is attributable mainly to perturbation at a site distal to the phosphorylation of this hexose. MATERIALS
1 This investigation was supported by grants from the Belgian Fsoundation for Scientific Medical Research and Belgian Ministry of Scientific Policy. ’ To whom correspondence should be addressed.
AND
METHODS
Insulin-producing cells from RINm5F, were grown, harvested, previously described (6). Cells 561
0003-9861/85 Copyright All rights
the
rat cell line, and counted as were seeded at a
$3.00
Q 198.5 by Academic Press. Inc. of reproduction in any form reserved.
562
GIROIX
density of 0.2-0.4 X lo6 cells/ml in a total volume of 5 ml, and the cells were harvested 3-6 days after seeding. The glucagon and somatostatin contents of these cells represent, on a weight basis, less than 0.02% of their insulin content (2). Normal pancreatic islets were isolated from fed albino rats by the collagenase technique (7). The methods used to measure the activity of glucose-phosphorylating enzymes (S), phosphoglucomutase (9), lactate dehydrogenase (lo), and glutamate dehydrogenase (ll), and the insulin and protein contents of tissue homogenate (12) were identical to those described in the cited references. For the assay of ornithine decarboxylase, about 2 X 10’ tumoral cells were sonicated (3 times, 10 s) in 1.0 ml of a sodium cacodylate-HCl buffer (50 mM, pH 6.0) containing dithiotreitol (0.5 mM), and pyridoxal5’-phosphate (0.2 mM). Aliquots (30 ~1 each) of this homogenate were incubated for 60 min at 30°C in an assay medium (final volume, 100 pl) consisting of the same buffer and containing DL-[1-“C]ornithine (0.02 mM). The reaction was halted by addition of 0.1 ml HCl(0.5 M), and the “COr formed was trapped in hyamine hydroxide (250 ~1) over 60-min incubation at 20°C. Blank values were measured in the absence of tissue homogenate. The methods used to measure the oxidation of U“C-labeled nutrients (13), output of lactic (14) and pyruvic (15) acids, and release of insulin from intact tumoral cells were similar to those previously used in the study of normal pancreatic islets. For measuring insulin release, groups of about 250 X 103 cells (range, 150 to 320 X 108 cells) were each incubated for 90 min in 1.0 ml of our usual hicarbonatebuffered medium (14) in a series of nine separate experiments. All results are expressed as the means (*SE) together with the number of individual observations (in parentheses). Data obtained in tumoral cells are always expressed per 10e cells and, whenever required, compared to results obtained in normal pancreatic islets and expressed per islet. The protein content averaged 139 + 13 ng per 10’ cells and 745 + 88 ng per islet, respectively (n = 6-8). RESULTS
Glucose Phosphorglation Tissue Homogenates
in
At low glucose concentrations (10 to 250 PM), the phosphorylation of D-[U14Clglucose in tumoral cells was compatible with the participation of a hexokinaselike enzyme with a Km for glucose close to 0.05 mM and a V,,, close to 0.68 pmol min-’ lo3 cells-’ (Fig. 1). This low-K, enzyme was inhibited by glucose 6-phos-
ET
AL.
o’..
04
10
’
20
33 1 /[GL”c5ooSE]
100 (mm-~)
FIG. 1. Lineweaver-Burk plot of D-[U-“C]ghCOSe phosphorylation by homogenates of tumoral cells incubated at low concentrations of D-ghCOSe (10 to 250 PM). Mean values (*SE) refer to 6 to 12 measurements.
phate with a Ki close to 0.1 mM (data not shown). When the contribution of the low-K, enzyme was subtracted from the readings obtained at higher concentrations of D-glUCOS6 (Table I, upper panel), the Hofstee plot based on the residual values for glucose phosphorylation still suggested the participation of two enzymatic activities with distinct Km’s for glucose (Fig. 2, upper panel). Thus, in addition to an enzymatic activity with a K,,, for glucose close to 0.3 m&q a high-Km enzyme was obviously operative. In a double-reciprocal plot, this latter enzyme yielded a curvilinear relationship similar to that of purified glucokinase, with an apparent Km close to 25 mM. Comparable results were obtained when the phosphorylation of D-[U-‘“Clglucose was measured in the presence of glucose 6-phosphate (initial concentration, 1.0 mM) in order to inhibit the low-K, hexokinase-like enzyme (Fig. 2, lower panel). Thus, the experimental data were again compatible with the presence of two
GLUCOSE METABOLISM
IN INSULIN-PRODUCING
TUMORAL CELLS
563
TABLE I COMPARISON
OF D-[U-%]GLUCOSE
PHOSPHORYLATION
IN TUMORAL CELLS AND PANCREATIC ISLETS Reaction
D-[U-“C]GIucose (mM]I
Glucose
6-phosphate (mM)
1.00 5.00 10.00 20.00 30.00
-
0.25 2.50 5.00 10.00 20.00
1.00 1.00 1.00 1.00 1.00
a Mean values (t- SE, or range of separate experiments, triplicate
-
of individual variations measurements being
enzymes with apparent K,‘s for glucose close to 0.2 and 25 mM, respectively. The absolute rate of glucose phosphorylation attributable to these two enzymatic activities represented, in the presence of glucose 6-phosphate, about half of the value found in its absence. A number of further experiments were performed to assess the significance of these findings. First, the rate of D-[U14Clglucose phosphorylation was measured within the same experiments in tumoral cells and pancreatic islets removed from normal rats at glucose concentrations ranging from 1.0 to 30.0 mM, in order to compare the behavior of the high-K, enzyme(s). The pattern of glucose phosphorylation, as ;a function of the hexose concentration, was similar in tumoral cells and pancreatic islets (Table I, upper panel). Thm, the islet/tumoral cell ratio in reaction velocity was not significantly affected by the glucose concentration, and averaged l.!j6 f 0.03 (n = 5). Likewise, at low glucose concentrations, the Km for glucose and Ki for glucose 6-phosphate of the low-K, hexokinase-like enzyme was similar in tumoral cells and pancreatic islets (16). Moreover, the results illustrated in Table I (lower panel) indicate that, in the presence of glucose 6-phosphate, the pattern of glucose phosphory-
velocity
(fmoI/min)”
per lOa Cells 1040 1156 1244 1593 1908 132 241 270 324 443
per Islet
f 19 + 31 + 50 + 32 k 118
(2) (2) (2) (2) (2)
1695 1825 1992 2417 2786
t k + f
(6) (3) (4) (5) (1)
242 349 393 546 635
9 8 12 21
when n = 2) are shown together made within each experiment.
f 22 (2) + 45 (2) * 139 (2) f 179 (2) 2 64 (2) f f 31 f + with
11 27 38 55 69
(10) (6) (10) (8) (6)
the number
lation at high glucose concentrations was also similar in tumoral cells and pancreatic islets. Again, the islet/tumoral cell ratio in reaction velocity was not significantly affected by the glucose concentration and averaged 1.57 + 0.08 (n = 5), in close agreement with the value for such a ratio in the absence of glucose 6-phosphate (1.56 f 0.03). Last, the phosphorylation of D-glucose, whether at low (0.25 mM) or high (10.0 mM) glucose concentrations, yielded comparable affinity for ATP in tumoral cells and pancreatic islets, respectively (data not shown), the rate of phosphorylation averaging, in the presence of 0.5 and 1.7 XIIM ATP, 18.0 + 2.8 and 62.1 + 4.4% (n = 4 in each case) of the paired value recorded at the high concentration of ATP (5.0 InM) used in all other experiments. Second, when D-[U-‘4C]mannose was used instead of D-[U-14C]g]UCOSe, the rate of phosphorylation by tumoral cell homogenates incubated in the presence of 10.0 mM mannose was, like in the case of glucose, about twice higher than that recorded at a low concentration of mannose (0.25 mM), whether in the absence or presence of glucose 6-phosphate (Table II). A different picture was obtained with D-[U-‘“Clfructose, in which case the ratio in reaction velocity at [lO.O mM]/[0.25 mM]
564
GIROIX
ET
AL.
F NO
GLUCOSE
6-PHOSPHATE
Y 0 T .-c E -. L !
10
5
9
5 3
0
0
l/[GLUCOSE]@M-;’ 4033 a05
P
I
0.1
0.2
0.3
0.4
0.5
0.6
a1
FIG. 2. Hofstee plots for D-[U-“C]ghCOSe phosphorylation by homogenates of tumoral cells in the absence (upper panel) or presence (lower panel) of glucose 6-phosphate (1.0 mM). In the upper panel, the data refer to the residual rate of glucose phosphorylation after subtraction of the value attributable to a low-K,,, hexokinase-like enzyme (see text). Mean values (+SE) refer to 630 (upper panel) and 3-15 (lower panel) measurements. The concentrations of D-ghCOSe amounted (from right to left) to 0.25, 1.0, 5.0, 10.0, 20.0, and 30.0 mM (upper panel) and 0.05, 0.25, 2.5, 5.0, 10.0, and 20.0 mM (lower panel). The insets illustrate Lineweaver-Burk plots of the data attributable to a high-Km enzyme after subtraction of the contribution attributable to a low-K, enzyme.
averaged 16.00 -t 1.15 and 12.91 f 1.94 in the absence and presence of glucose 6phosphate, respectively. These ratios, which were not significantly different from one another (P > 0.2), would yield a Km for fructose averaging 5.3 + 0.9 mM, close
to that previously found in islet homogenates (6.0 mM), in which the phosphorylation of fructose is catalyzed by hexokinase, but not fructokinase (17). Last, control experiments provided the following information. In erythrocyte or
GLUCOSE
METABOLISM
IN
INSULIN-PRODUCING TABLE
PHOSPHORYLATIONOF
Glucose &phosphate MM)
Hexose
(mM) 0.25 0.25 10.00 10.00 ’ Mean
GLIJCOSE,MANNOSE,
1.00 1.00 values
(&SE)
refer
AND
TUMORAL
565
II FRUCI-OSE
Phosphorylation
BY TUMORAL rate
(fmol
D-[U-‘*C]Glucose
D-[U-‘%]Mannose
722 202 1253 400
665 118 1285 290
+ 5 + 14 f 13 + 26
CELLS
CELLHOMOGENATES
min-’
lo3 cells-‘)” D-[U-“C]Fructose
+ 46 + 2 f 77 f 32
9’7k 6 352 2 1552 -t 56 452 f 63
to six measurements.
parotid gland homogenates examined by the same procedure as that used for the study of tumoral cells, the phosphorylation of D-[U-14Cjghcose over a wide range of concentrations (25 GM to 20 mM) yielded, in a double-reciprocal plot, a single, straight line with a K, for glucose close to 0.05 mrvtr (8, 18). When increasing amounts of the parotid gland homogenate (corresponding to 25, 50, and 75 pg wet wt per assay) were mixed with a constant amount of liver homogenate (corresponding to 75 pg Twet wt per assay), the activity of liver glucokinase (as judged from the difference in glucose phosphorylation rate at 0.25 and 110.0mM D-ghCOSe, respectively) was not affected by either the amount of parotid homogenate or concentration of glucose 6-ph.osphate (nil or 1.0 I’IIM) present in the assay medium. Thus, at the three concentrations of parotid homogenates, the activity of liver glucokinase averaged 0.74 k 0.06 and 0.70 + 0.01 pmol min-’ pg wet wt-l in the absence and presence of glucose 6-phosphate, respectively. This indicates that the presence of increasing contributions of hexokinase did not hamper the characterization of glucokinase activity. Such an activity failed to be affected by glucose 6-phosphate in these experiments, as well as in experiments conducted solely with liver homogenates (data not shown). Likewise, when homogenates derived from insulin-producing cells (corresponding to about 17 islets per assay) and liver (corresponding to 75 pg wet weight per assay) were mixed together, the phosphorylation rate of D[U-‘*Clglucose (0.5 and 10.0 mM) yielded
results compatible with a mere additive effect. There was no obvious indication, therefore, that the homogenate prepared from insulin-producing cells contained any endogenous modulator of glucokinase activity. Other Enzymatic
Data
The data illustrated in Table I indicate that the rate of glucose phosphorylation per lo3 tumoral cells represented 64 -t 2% (n = 10) of the corresponding value per pancreatic islet. For the purpose of comparison, the activity of four other enzymes was measured in tumoral cell homogenates (Table III). Such homogenates displayed phosphoglucomutase activity, which was considerably enhanced by glucose l,gbiphosphate; lactate dehydrogenase activity; glutamate dehydrogenase activity, which was increased by ADP and L-leucine; and ornithine decarboxylase activity, which was inhibited by a-(difluoromethyl)ornithine. When compared to data previously obtained in normal pancreatic islets (9, 10, 12, 19) the activity of phosphoglucomutase and lactate dehydrogenase per lo3 tumoral cells again represented about 60% of that measured per islet. The activity of glutamate dehydrogenase in tumoral cells was much lower, representing, per lo3 cells, no more than one-tenth of that measured per islet (11, 19, 20). On the contrary, the activity of ornithine decarboxylase per lo3 tumoral cells was about five times higher than that measured per islet (1.23 & 0.39 amol min-’ islet-l; n = 5).
566
GIROIX TABLE ENZYMATIC
Enzyme Phosphoglucomutase (EC 2.7.5.1)
ACTIVITIES
Glutamate dehydrogenase (EC 1.4.1.3)
2-Ketoglutarate + NH4 (50.0) NADH (0.3)
NADPH
Ornithine decarboxylase (EC 4.1.1.1’7)
Metabolic
CELL HOMOGENATES Reaction (pm01 mini’
or inhibitor (m@
Glucose 1,6bisphosphate
l-phosphate
Pyruvate (2.0) + NADH (0.3)
III
Activator
(0.2)
Lactate dehydrogenase (EC 1.1.1.27)
AL.
IN TUMORAL
Substrate(s) W-0 Glucose
ET
velocity ld ceils-i)
2.14 -t 0.13 (3) 46.39 f 1.97 (3) (0.025) 189.17
+ 5.56 (3)
(7.0)
(0.3)
DL-[1-‘%]Ornithine (0.02)
-
2.75 17.94 13.88 1.28 12.81 7.10
ADP (1.0) L-Leueine (10.0) ADP (1.0) L-Leucine (10.0) or-(Difluoromethyl)ornithine (1.0)
0.42 0.89 0.47 0.07 0.36 0.49
(3) (3) (3) (3) (3) (3)
6.92 f 0.93 X 1O-6 (3) 0.62 f 0.19 x 1o-6 (2)
Data
Relative to basal value, the output of lactic and pyruvic acid was considerably increased when 2.8 mM D-gh.Icose was present in the incubation medium (Fig. 3). However no further increase (P > 0.2) in the output of these organic acids was observed when the glucose concentration was raised from 2.8 to 7.0 or 16.7 mM. A comparable pattern characterized the oxidation of D-[U-‘4C]ghCOSe by the tumoral cells (Fig. 3). A concentration-dependent relationship for D-[U-14Clglucose oxidation was observed, however, at low glucose concentrations in the 0.05 to 2.8 mM range, with a half-maximal response at about 0.3 mM D-glucose (Fig. 4). The oxidation of L-[U-14Clglutamine was concentration-related in the 0.1 to 10.0 mM range, with a half-maximal response at about 0.2 mM L-glutamine (Fig. 4). At all concentrations, the oxidation of Lglutamine was higher than that of Dglucose. The oxidation of L-[U-‘4C]leucine averaged 1.08 f 0.03 and 1.49 + 0.17 pmol 60 min-’ lo3 cells-’ at L-leucine concen-
+ + f f f *
LACTATE
OUTPUT P
PYRUVATE
0 i ii 4 ki I
10.
GLUCOSE
Ott, 1.7 m28 587.0a3 GLUCOSE
I
OUTPUT
OXIDATION
187
@i)
FIG. 3. Output of lactic and pyruvic acids and oxidation of D-[D-“C&iucose at increasing concentrations of D-ghCOSe. Mean values (*SE, whenever it exceeds the size of the mean symbol) refer to 2% 44 (lactate output), 16-32 (pyruvate output), and 828 (glucose oxidation) individual measurements.
GLUCOSE 12.5
METABOLISM
IN
r
INSULIN-PRODUCING
P
-“Cl
GLUTAMINE
/
TUMORAL
CELLS
567
lease, which reached a value higher than that observed, within the same experiments, in the presence of 7.0 mM D-glucose (P < 0.001; d.f. = 57). L-Glutamine failed to affect L-leucine-stimulated insulin output. DISCUSSION
oh0
cl055
0.1
”
a203 1.0 1.7 3.0 NUTRIENT(mM)
FIG. 4. Oxidation of L-[U-“Clglutamine “Clglucose at increasing concentrations scale) of these nutrients. Mean values to 8-28 individual measurements.
trations of ‘1.0 and 20.0 (n = 8 in both cases). Insulin
IrIM,
56
rao
and D-[U(logarithmic (*SE) refer
respectively
At variance with a prior report (4), the present results indicate the presence, in tumoral insulin-producing cells, of a highK, glucose-phosphorylating enzymatic activity in addition to low-K, hexokinaselike activity. The latter activity itself consisted of both a typical hexokinase, with a Km for glucose close to 0.05 mM and a Ki for glucose 6-phosphate close to 0.1 mM, and a second isoenzyme with a Km for glucose close to 0.2-0.3 mM and an apparently lesser sensitivity to glucose 6phosphate (half inhibition at approximately 1.0 mM glucose 6-phosphate). An uncommon hexokinase (peak III,) was also noticed by Meglasson and Matschinsky on chromatography of insulinoma homogenates (21). The presence of a high-K, glucokinase-like enzyme in the tumoral cells is also in good agreement with the data obtained by these authors (5). The
Content and Release
The insulin content of the cells averaged, in a series of seven preparations, 345 f 66 nU/103 cells. In the absence of exogenous nutrient, the basal release of insulin, relative to the insulin content, was much Ihigher than in normal pancreatic islets, amounting to 113 + 9 nU 90 min-’ lo3 cells-’ (Table IV). In the presence of 2.8 mM D-glucose, the secretory rate was slightly higher than basal value (P < 0.02; d.f. = 38). However, no significant difference in insulin output was observed when the glucose concentration was raised from 2.8 to 16.7 mM. At the latter glucose concentration, both forskolin and theophylline markedly augmented insulin output (P < 0.05 or less; d.f. = 18 or more). In the presence of glucose, Lornithine also augmented insulin release (P < 0.001; d.f. = 57). In the absence of glucose, L-leucine stimulated insulin re-
TABLE INSULIN
D-Glucose b-4 2.8 5.6 7.0 16.7 16.7 16.7 7.0 -
RELEASE
IV
BY TUMORAL
Other secretagogues bw Forskolin (0.005) Theophylline (1.4) L-Ornithine (10) L-Leucine (10) L-Leucine (10) + Lglutamine (10)
o Mean values (*SE) are shown number of individual observations
CELLS Insulin output= (nU 90 min-’ 10s cells-‘) 112.5 133.4 137.7 147.4 138.4 226.1 223.0 197.5 230.5 232.7
f 9.0 (30) f 7.4 (20) zt 8.2 (20) + 6.2 (20) f 3.5 (44) f 13.0 (23) k 15.2 (10) + 11.2 (30) + 7.5 (30) f 6.5 (30)
together with the (in parentheses).
568
GIROIX
activity of this enzyme was documented here with either D-glucose or D-mannose as substrate, both in the absence or presence of glucose 6-phosphate. Unexpectedly, this high-Km enzyme appeared, in insulinproducing cells, not totally unresponsive to glucose 6-phosphate. Whatever their precise identity, the glucose-phosphorylating enzymes, as well as phosphoglucomutase and lactate dehydrogenase, displayed an activity three to four times higher in tumoral than nontumoral islet cells, when related to the protein content of these cells. The metabolism of glucose in intact tumoral cells, when compared to normal islet cells, displayed several and severe abnormalities. First, relative to protein content, the rate of glucose utilization in the range 2.8-16.7 mM of glucose concentrations was 3 to 13 times higher in tumoral cells than in normal islets. Indeed, after correction for basal values, the output of lactate and pyruvate and the oxidation of glucose averaged 67 pmol glucose residue 60 min-’ lo3 tumoral cells-’ as compared to mean values for glucose utilization of 27 and 129 pm0160 mine1 islet-’ in normal pancreatic islets incubated at 2.8 and 16.7 mM D-glucose, respectively (22). Second, at low glucose concentrations (2.8-3.0 mM), the rate of glucose utilization by intact tumoral cells (1.07 pmol min-’ lo3 cells-‘, at 37’C) was not vastly different from its rate of phosphorylation in cell homogenates (1.10 pmol min-’ lo3 cells-’ at 30°C). This suggests that, at variance with the situation documented in normal islet cells (16), hexokinase is little inhibited by endogenous glucose 6phosphate in tumoral cells exposed to a low concentration of glucose. Third, in tumoral cells, the oxidation of glucose and the output of lactic and pyruvic acids reached a close-to-maximal value at a low concentration of glucose (2.8 mM), in fair agreement with the data obtained by Halban et aZ. (4) for the production of 3Hz0 from D-[5-3H]glucose in the same cell line. This behavior contrasts sharply with the sigmoidal pattern of glucose utilization in
ET
AL.
normal pancreatic islets (23). Since the pattern of glucose phosphorylation at increasing glucose concentrations was not vastly different in tumoral and nontumoral islet cell homogenates, the disparity in the concentration-dependent relationships for glucose utilization in intact cells points to a difference between tumoral and normal cells at a site distal to the phosphorylation of this hexose. Fourth, after correction for basal values, the production of 14C02 from D-[U-14C]ghCOSe in tumoral cells accounted for no more than 12.3% of the rate of glucose utilization, whereas in normal pancreatic islets the ratio of glucose oxidation/utilization is close to 28.0% (23). Thus, in the insulinproducing tumoral cells, like in other tumoral cells (24), the rate of glucose oxidation was abnormally low relative to the rate of pyruvate generation. Last, in the tumoral cells, the lactate/pyruvate output ratio was close to 2.4, as distinct from values of 6.7 and 10.0 in normal islets exposed to 2.8 and 16.7 mM D-glucose, respectively (12). This comparison suggests that the cytosolic NADH/NAD+ ratio was maintained at a lower value in tumoral than in normal islet cells. This would represent a favorable situation for the maintenance of a high rate of glycolysis in tumoral cells, but may adversely affect insulin release, to the extent that the latter process depends on the induction of a more reduced state in cytosolic redox couples (25). Our secretory data were in fair agreement with the metabolic data in that insulin release was slightly increased by D-glucose above basal value, but failed to be significantly affected when the glucose concentration was raised from 2.3 to 16.7 mM. Under suitable conditions, the tumoral cells were duly responsive, however, to other secretagogues such as L-leucine, forskolin, theophylline, or ornithine. These findings complement prior observations (l-3). L-Glutamine failed to enhance Lleucine-stimulated insulin release. This unexpected situation and the low rate of L-leucine oxidation relative to that of
IGLUCOSE
METABOLISM
IN
INSULIN-PRODUCING
other nutrients could be related to the fact that thle tumoral cells are cultured in a medium containing 2.0 lllM L-glutamine and, hence, may have a high basal glutamate clontent (26). Indeed, in islet cells, a rise in glutamate content is associated with a decreased oxidation of Lleucine (27). At this point it should be stressed that, although the responsiveness of glutamate dehydrogenase to ADP and L-leucine appeared similar in tumoral and normal islet cells, its activity relative to protein content was about twice lower in tumoral than normal cells (20). Such was not the case for ornithine decarboxylase, the activity of which was, relative to protein content, about 30 times higher in tumoral than normal islet cells. Incidentally, the novel demonstration of ornithine decarboxylase activity in insulin-producing cells, taken together with the known insulinotrop’ic capacity of this amino acid (28), calls, by analogy with studies performed in o,ther tissues (29), for further investigation on the possible role of polyamines in the process of stimulus-secretion coupling in the pancreatic B cell (30). In conclusion, the present work illustrates the existence of dramatic alterations in the metabolism of glucose and secretory response to this hexose in insulin-producing tumoral cells, as compared to normal pancreatic islets. These abnormalities were apparently not attributable to any gross difference in the hexosephosphorylating enzymatic activities of tumoral cells and normal islets, respectively. Therefore, the present data reinforce the view (31) the regulatory mechanisms located distally to the site of glucose phosphorylation participate in the control of glucose utilization and, hence, glucose-induced insulin release in insulinproducing cells. ACKNOWLEDGMENTS The authors
TUMORAL
CELLS
569
Schoonheydt and M. Urbain for technical assistance; and to C. Demesmaeker for secretarial help.
REFERENCES 1. SOPWITH, A. M., HUTTON, J. C., NABER, S. P., CHICK, W. C., AND HALES, C. N. (1981) Diabetologia 21,224-229. 2. PRAZ, G. A., HALBAN, P. A., WOLLHEIM, C. B., BLONDEL, B., STRAUSS, A. J., AND RENOLD, A. E. (1983) B&&em. J. 210, 345-352. 3. HOENIG, M., FERGUSON, D. C., ANDMATSCHINSKY, F. M. (1984) Diabetes 33, l-7. 4. HALBAN, P. A., PRAZ, G. A., AND WOLLHEIM, C. B. (1983) B&hem J. 212,439-443. 5. MEGLASSON, M. D., AND MATSCHINSKY, F. M. (1983) J. Biol. Chem. 258, 6’705-6708. 6. GIROIX, M.-H., MALAISSE-LAGAE, F., SENER, A., AND MALAISSE, W. J. (1985) Mol. Cell. B&hem. 66, 61-64. 7. LACY, P. E., AND KOSTIANOVSKY, M. (1967) Diabetes 16, 35-39. 8. MALAISSE, W. J., GIROIX, M.-H., DUFRANE, S. P., MALAISSE-LAGAE, F., AND SENER, A. (1985) B&hem. Int. 10,233-240. 9. MALAISSE-LAGAE, F., SENER, A., AND MALAISSE, W. J. (1982) Biochimie 64, 1059-1063. 10. MALAISSE, W. J., KAWAZU, S., HERCHUELZ, A., HUTTON, J. C., SOMERS, G., DEVIS, G., AND SENER, A. (1979) Arch. B&hem. Biophys. 194, 49-62. 11. HUTTON, J. C., SENER, A., AND MALAISSE, W. J. (1979) B&hem. J. 184, 291-301. 12. GIROIX, M.-H., DUFRANE, S. P., MALAISSE-LAGAE, F., SENER, A., AND MALAISSE, W. J. (1984) B&hem. Biophys. Res. Cmnmun. 119,543-548. 13. CARPINELLI, A. R., SENER, A., HERCHUELZ, A., AND MALAISSE, W. J. (1980) Metab. Clin. Exp. 29,540-545. 14. SENER, A., AND MALAISSE, W. J. (1976) B&hem. Med. 15, 34-41. 15. MALAISSE, W. J., HUTTON, J. C., KAWAZU, S., AND SENER, A. (1978) Eur. J B&hem. 87,121-130. 16. GIROIX, M.-H., SENER, A., PIPELEERS, D. G., AND MALAISSE, W. J. (1984) B&hem. J 223, 447453. 17. SENER, A., GIROIX, M.-H., AND MALAISSE, W. J. (1984) Eur. J. Biochem. 144, 223-226. 18. GIROIX, M.-H., DUFRANE, S. P., SENER, A., AND MALAISSE, W. J., Submitted for publication. 19. MALAISSE, W. J., SENER, A., WELSH, M., MALAISSE, W. J., HELLERSTR~M, C., AND CHRISTOPHE, J. (1983) B&hem. J. 210, 921-927. 20. SENER, A., AND MALAISSE, W. J. (1980) Nature (London) 288, 187-189.
GIROIX 21. MEGLASSON, M. D., AND MATSCHINSKY, F. M. (1984) in Methods in Diabetes Research (Larner, J., and Pohl, S. L., eds.), Vol. 1, pp. 213226, Wiley, New York. 22. MALAISSE, W. J., SENER, A., LEVY, J., AND HERCHUELZ, A. (1976) Acta Diabetol. Lot. 13, 202215. 23. SENER, A., LEVY, J., AND MALAISSE, W. J. (1976) Biochem J. 156.521-525. 24. REITZER, L. J., WICE, B. M., AND KENNELL, D. (1979) J. BioL C&m. 254, 2669-2676. 25. MALAISSE, W. J., MALAISSE-LAGAE, F., AM) SENER, A. (1984) Experientia 40, 1035-1043. 26. MALAISSE, W. J., SENER, A., CARPINELLI, A. R., ANJANEYULU, K., LEBRUN, P., HERCHUELZ, A.,
ET
27.
28. 29. 30.
31.
AL. AND CHRISTOPHE, J. (1989) Mel Cell Endock&. 20,171-x39. MALAISSE, W. J., SENER, A., MALAISSE-LAGAE, F., WELSH, M., MATTHEWS, D. E., BIER, D. M., AND HELLERSTR~M, C. (1982) J. Biol Chem 257,8731-8737. CHARLES, S., TAMAGAWA, T., AND HENQUIN, J. C. (1982) B&hem J. 208, 301-308. KOENIG, H., GOLDSTONE, A., AND Lu, C. Y. (1933) Nature (London) 305,530-534. MALAISSE, W. J., LEBRUN, P., AND HERCHUELZ, A. (1985) in Calcium Physiology (MarmB, D., ed.), Springer-Verlag, Berlin, in press. SENER, A., AND MALAISSE, W. J. (1984) Experientie 40, 1026-1035.
AND BIOPHYSICS pp. 561-5’70, 1985
Glucose Metabolism MARIE-HaLfiNE FRANCINE Labcwatoq
GIROIX, ABDULLAH MALAISSE-LAGAE,
of Expekmental Received
in Insulin-Producing
January
Medicine,
Brussels
Tumoral
Cells’
SENER, SIMON P. DUFRANE, WILLY J. MALAISSE’
AND
Free
28, 1985, and in revised
University, form
Brussels May
115, Belgium
1’7, 1985
Homogenates of insulin-producing tumoral cells catalyzed the phosphorylation of glucose, mannose, and fructose. The kinetics of phosphorylation at increasing glucose concentrations, the inhibitory effect of glucose 6-phosphate, and the comparison of results obtained with distinct hexoses indicated the presence of both low-K, hexokinaselike and high-K, enzymatic activities, the results being grossly comparable to those collected in normal pancreatic islets. Relative to protein content, the glucose-phosphorylating enzymatic activity was higher in tumoral than normal islet cells. The activity of other enzymes was either lower (glutamate dehydrogenase), moderately higher (phosphoglucomutase, lactate dehydrogenase) or considerably greater (ornithine decarboxylase) in tumoral than in normal islet cells. In intact tumoral cells, incubated under increasing glucose concentrations, the oxidation of D-[U-‘4Clglucose and the output of lactic and pyruvic acids reached a close-to-maximal value at 2.8 mM glucose. The ratios for glucose oxidation/utilization and lactate/pyruvate output were much lower in tumoral than in normal islet cells. Although glucose caused a modest increase in insulin output from the tumoral cells, this effect was saturated at a low glucose concentration (2.8 mM) and less marked than that of other secretagogues (e.g., Lleucine, L-ornithine, or forskolin). Thus, despite a close-to-normal enzymatic equipment for glucose phosphorylation, the tumoral cells displayed severe abnormalities in the metabolism and secretory response to this hexose. These findings point to regulatory mechanisms distal to glucose phosphorylation in the control of glucose metabolism in insulin-prod.ucing cells. o 19% Academic PMS, IX
Insulin-producing tumoral cells of the RIN type are poorly responsive to glucose as compared to other nutrient or nonnutrient secretagogues (l-3). The biochemical determinant(s) of such a poor responsiveness remains to be elucidated. In one study, it was proposed that the failure of glucose to stimulate insulin release was attributable to an abnormal glucose metabolism in the tumoral cells, especially the lack of a high-Km glucokinase-like activity (4). Yet, in another study, gluco-
kinase was purified from the same tumor line (5). In the present study, we have investigated both the kinetics of glucose phosphorylation in cell homogenates and the oxidation of glucose and output of lactic and pyruvic acids in intact insulinproducing tumoral cells. Our results suggest that, by comparison with normal pancreatic islet cells, abnormal glucose metabolism in tumoral cells is attributable mainly to perturbation at a site distal to the phosphorylation of this hexose. MATERIALS
1 This investigation was supported by grants from the Belgian Fsoundation for Scientific Medical Research and Belgian Ministry of Scientific Policy. ’ To whom correspondence should be addressed.
AND
METHODS
Insulin-producing cells from RINm5F, were grown, harvested, previously described (6). Cells 561
0003-9861/85 Copyright All rights
the
rat cell line, and counted as were seeded at a
$3.00
Q 198.5 by Academic Press. Inc. of reproduction in any form reserved.
562
GIROIX
density of 0.2-0.4 X lo6 cells/ml in a total volume of 5 ml, and the cells were harvested 3-6 days after seeding. The glucagon and somatostatin contents of these cells represent, on a weight basis, less than 0.02% of their insulin content (2). Normal pancreatic islets were isolated from fed albino rats by the collagenase technique (7). The methods used to measure the activity of glucose-phosphorylating enzymes (S), phosphoglucomutase (9), lactate dehydrogenase (lo), and glutamate dehydrogenase (ll), and the insulin and protein contents of tissue homogenate (12) were identical to those described in the cited references. For the assay of ornithine decarboxylase, about 2 X 10’ tumoral cells were sonicated (3 times, 10 s) in 1.0 ml of a sodium cacodylate-HCl buffer (50 mM, pH 6.0) containing dithiotreitol (0.5 mM), and pyridoxal5’-phosphate (0.2 mM). Aliquots (30 ~1 each) of this homogenate were incubated for 60 min at 30°C in an assay medium (final volume, 100 pl) consisting of the same buffer and containing DL-[1-“C]ornithine (0.02 mM). The reaction was halted by addition of 0.1 ml HCl(0.5 M), and the “COr formed was trapped in hyamine hydroxide (250 ~1) over 60-min incubation at 20°C. Blank values were measured in the absence of tissue homogenate. The methods used to measure the oxidation of U“C-labeled nutrients (13), output of lactic (14) and pyruvic (15) acids, and release of insulin from intact tumoral cells were similar to those previously used in the study of normal pancreatic islets. For measuring insulin release, groups of about 250 X 103 cells (range, 150 to 320 X 108 cells) were each incubated for 90 min in 1.0 ml of our usual hicarbonatebuffered medium (14) in a series of nine separate experiments. All results are expressed as the means (*SE) together with the number of individual observations (in parentheses). Data obtained in tumoral cells are always expressed per 10e cells and, whenever required, compared to results obtained in normal pancreatic islets and expressed per islet. The protein content averaged 139 + 13 ng per 10’ cells and 745 + 88 ng per islet, respectively (n = 6-8). RESULTS
Glucose Phosphorglation Tissue Homogenates
in
At low glucose concentrations (10 to 250 PM), the phosphorylation of D-[U14Clglucose in tumoral cells was compatible with the participation of a hexokinaselike enzyme with a Km for glucose close to 0.05 mM and a V,,, close to 0.68 pmol min-’ lo3 cells-’ (Fig. 1). This low-K, enzyme was inhibited by glucose 6-phos-
ET
AL.
o’..
04
10
’
20
33 1 /[GL”c5ooSE]
100 (mm-~)
FIG. 1. Lineweaver-Burk plot of D-[U-“C]ghCOSe phosphorylation by homogenates of tumoral cells incubated at low concentrations of D-ghCOSe (10 to 250 PM). Mean values (*SE) refer to 6 to 12 measurements.
phate with a Ki close to 0.1 mM (data not shown). When the contribution of the low-K, enzyme was subtracted from the readings obtained at higher concentrations of D-glUCOS6 (Table I, upper panel), the Hofstee plot based on the residual values for glucose phosphorylation still suggested the participation of two enzymatic activities with distinct Km’s for glucose (Fig. 2, upper panel). Thus, in addition to an enzymatic activity with a K,,, for glucose close to 0.3 m&q a high-Km enzyme was obviously operative. In a double-reciprocal plot, this latter enzyme yielded a curvilinear relationship similar to that of purified glucokinase, with an apparent Km close to 25 mM. Comparable results were obtained when the phosphorylation of D-[U-‘“Clglucose was measured in the presence of glucose 6-phosphate (initial concentration, 1.0 mM) in order to inhibit the low-K, hexokinase-like enzyme (Fig. 2, lower panel). Thus, the experimental data were again compatible with the presence of two
GLUCOSE METABOLISM
IN INSULIN-PRODUCING
TUMORAL CELLS
563
TABLE I COMPARISON
OF D-[U-%]GLUCOSE
PHOSPHORYLATION
IN TUMORAL CELLS AND PANCREATIC ISLETS Reaction
D-[U-“C]GIucose (mM]I
Glucose
6-phosphate (mM)
1.00 5.00 10.00 20.00 30.00
-
0.25 2.50 5.00 10.00 20.00
1.00 1.00 1.00 1.00 1.00
a Mean values (t- SE, or range of separate experiments, triplicate
-
of individual variations measurements being
enzymes with apparent K,‘s for glucose close to 0.2 and 25 mM, respectively. The absolute rate of glucose phosphorylation attributable to these two enzymatic activities represented, in the presence of glucose 6-phosphate, about half of the value found in its absence. A number of further experiments were performed to assess the significance of these findings. First, the rate of D-[U14Clglucose phosphorylation was measured within the same experiments in tumoral cells and pancreatic islets removed from normal rats at glucose concentrations ranging from 1.0 to 30.0 mM, in order to compare the behavior of the high-K, enzyme(s). The pattern of glucose phosphorylation, as ;a function of the hexose concentration, was similar in tumoral cells and pancreatic islets (Table I, upper panel). Thm, the islet/tumoral cell ratio in reaction velocity was not significantly affected by the glucose concentration, and averaged l.!j6 f 0.03 (n = 5). Likewise, at low glucose concentrations, the Km for glucose and Ki for glucose 6-phosphate of the low-K, hexokinase-like enzyme was similar in tumoral cells and pancreatic islets (16). Moreover, the results illustrated in Table I (lower panel) indicate that, in the presence of glucose 6-phosphate, the pattern of glucose phosphory-
velocity
(fmoI/min)”
per lOa Cells 1040 1156 1244 1593 1908 132 241 270 324 443
per Islet
f 19 + 31 + 50 + 32 k 118
(2) (2) (2) (2) (2)
1695 1825 1992 2417 2786
t k + f
(6) (3) (4) (5) (1)
242 349 393 546 635
9 8 12 21
when n = 2) are shown together made within each experiment.
f 22 (2) + 45 (2) * 139 (2) f 179 (2) 2 64 (2) f f 31 f + with
11 27 38 55 69
(10) (6) (10) (8) (6)
the number
lation at high glucose concentrations was also similar in tumoral cells and pancreatic islets. Again, the islet/tumoral cell ratio in reaction velocity was not significantly affected by the glucose concentration and averaged 1.57 + 0.08 (n = 5), in close agreement with the value for such a ratio in the absence of glucose 6-phosphate (1.56 f 0.03). Last, the phosphorylation of D-glucose, whether at low (0.25 mM) or high (10.0 mM) glucose concentrations, yielded comparable affinity for ATP in tumoral cells and pancreatic islets, respectively (data not shown), the rate of phosphorylation averaging, in the presence of 0.5 and 1.7 XIIM ATP, 18.0 + 2.8 and 62.1 + 4.4% (n = 4 in each case) of the paired value recorded at the high concentration of ATP (5.0 InM) used in all other experiments. Second, when D-[U-‘4C]mannose was used instead of D-[U-14C]g]UCOSe, the rate of phosphorylation by tumoral cell homogenates incubated in the presence of 10.0 mM mannose was, like in the case of glucose, about twice higher than that recorded at a low concentration of mannose (0.25 mM), whether in the absence or presence of glucose 6-phosphate (Table II). A different picture was obtained with D-[U-‘“Clfructose, in which case the ratio in reaction velocity at [lO.O mM]/[0.25 mM]
564
GIROIX
ET
AL.
F NO
GLUCOSE
6-PHOSPHATE
Y 0 T .-c E -. L !
10
5
9
5 3
0
0
l/[GLUCOSE]@M-;’ 4033 a05
P
I
0.1
0.2
0.3
0.4
0.5
0.6
a1
FIG. 2. Hofstee plots for D-[U-“C]ghCOSe phosphorylation by homogenates of tumoral cells in the absence (upper panel) or presence (lower panel) of glucose 6-phosphate (1.0 mM). In the upper panel, the data refer to the residual rate of glucose phosphorylation after subtraction of the value attributable to a low-K,,, hexokinase-like enzyme (see text). Mean values (+SE) refer to 630 (upper panel) and 3-15 (lower panel) measurements. The concentrations of D-ghCOSe amounted (from right to left) to 0.25, 1.0, 5.0, 10.0, 20.0, and 30.0 mM (upper panel) and 0.05, 0.25, 2.5, 5.0, 10.0, and 20.0 mM (lower panel). The insets illustrate Lineweaver-Burk plots of the data attributable to a high-Km enzyme after subtraction of the contribution attributable to a low-K, enzyme.
averaged 16.00 -t 1.15 and 12.91 f 1.94 in the absence and presence of glucose 6phosphate, respectively. These ratios, which were not significantly different from one another (P > 0.2), would yield a Km for fructose averaging 5.3 + 0.9 mM, close
to that previously found in islet homogenates (6.0 mM), in which the phosphorylation of fructose is catalyzed by hexokinase, but not fructokinase (17). Last, control experiments provided the following information. In erythrocyte or
GLUCOSE
METABOLISM
IN
INSULIN-PRODUCING TABLE
PHOSPHORYLATIONOF
Glucose &phosphate MM)
Hexose
(mM) 0.25 0.25 10.00 10.00 ’ Mean
GLIJCOSE,MANNOSE,
1.00 1.00 values
(&SE)
refer
AND
TUMORAL
565
II FRUCI-OSE
Phosphorylation
BY TUMORAL rate
(fmol
D-[U-‘*C]Glucose
D-[U-‘%]Mannose
722 202 1253 400
665 118 1285 290
+ 5 + 14 f 13 + 26
CELLS
CELLHOMOGENATES
min-’
lo3 cells-‘)” D-[U-“C]Fructose
+ 46 + 2 f 77 f 32
9’7k 6 352 2 1552 -t 56 452 f 63
to six measurements.
parotid gland homogenates examined by the same procedure as that used for the study of tumoral cells, the phosphorylation of D-[U-14Cjghcose over a wide range of concentrations (25 GM to 20 mM) yielded, in a double-reciprocal plot, a single, straight line with a K, for glucose close to 0.05 mrvtr (8, 18). When increasing amounts of the parotid gland homogenate (corresponding to 25, 50, and 75 pg wet wt per assay) were mixed with a constant amount of liver homogenate (corresponding to 75 pg Twet wt per assay), the activity of liver glucokinase (as judged from the difference in glucose phosphorylation rate at 0.25 and 110.0mM D-ghCOSe, respectively) was not affected by either the amount of parotid homogenate or concentration of glucose 6-ph.osphate (nil or 1.0 I’IIM) present in the assay medium. Thus, at the three concentrations of parotid homogenates, the activity of liver glucokinase averaged 0.74 k 0.06 and 0.70 + 0.01 pmol min-’ pg wet wt-l in the absence and presence of glucose 6-phosphate, respectively. This indicates that the presence of increasing contributions of hexokinase did not hamper the characterization of glucokinase activity. Such an activity failed to be affected by glucose 6-phosphate in these experiments, as well as in experiments conducted solely with liver homogenates (data not shown). Likewise, when homogenates derived from insulin-producing cells (corresponding to about 17 islets per assay) and liver (corresponding to 75 pg wet weight per assay) were mixed together, the phosphorylation rate of D[U-‘*Clglucose (0.5 and 10.0 mM) yielded
results compatible with a mere additive effect. There was no obvious indication, therefore, that the homogenate prepared from insulin-producing cells contained any endogenous modulator of glucokinase activity. Other Enzymatic
Data
The data illustrated in Table I indicate that the rate of glucose phosphorylation per lo3 tumoral cells represented 64 -t 2% (n = 10) of the corresponding value per pancreatic islet. For the purpose of comparison, the activity of four other enzymes was measured in tumoral cell homogenates (Table III). Such homogenates displayed phosphoglucomutase activity, which was considerably enhanced by glucose l,gbiphosphate; lactate dehydrogenase activity; glutamate dehydrogenase activity, which was increased by ADP and L-leucine; and ornithine decarboxylase activity, which was inhibited by a-(difluoromethyl)ornithine. When compared to data previously obtained in normal pancreatic islets (9, 10, 12, 19) the activity of phosphoglucomutase and lactate dehydrogenase per lo3 tumoral cells again represented about 60% of that measured per islet. The activity of glutamate dehydrogenase in tumoral cells was much lower, representing, per lo3 cells, no more than one-tenth of that measured per islet (11, 19, 20). On the contrary, the activity of ornithine decarboxylase per lo3 tumoral cells was about five times higher than that measured per islet (1.23 & 0.39 amol min-’ islet-l; n = 5).
566
GIROIX TABLE ENZYMATIC
Enzyme Phosphoglucomutase (EC 2.7.5.1)
ACTIVITIES
Glutamate dehydrogenase (EC 1.4.1.3)
2-Ketoglutarate + NH4 (50.0) NADH (0.3)
NADPH
Ornithine decarboxylase (EC 4.1.1.1’7)
Metabolic
CELL HOMOGENATES Reaction (pm01 mini’
or inhibitor (m@
Glucose 1,6bisphosphate
l-phosphate
Pyruvate (2.0) + NADH (0.3)
III
Activator
(0.2)
Lactate dehydrogenase (EC 1.1.1.27)
AL.
IN TUMORAL
Substrate(s) W-0 Glucose
ET
velocity ld ceils-i)
2.14 -t 0.13 (3) 46.39 f 1.97 (3) (0.025) 189.17
+ 5.56 (3)
(7.0)
(0.3)
DL-[1-‘%]Ornithine (0.02)
-
2.75 17.94 13.88 1.28 12.81 7.10
ADP (1.0) L-Leueine (10.0) ADP (1.0) L-Leucine (10.0) or-(Difluoromethyl)ornithine (1.0)
0.42 0.89 0.47 0.07 0.36 0.49
(3) (3) (3) (3) (3) (3)
6.92 f 0.93 X 1O-6 (3) 0.62 f 0.19 x 1o-6 (2)
Data
Relative to basal value, the output of lactic and pyruvic acid was considerably increased when 2.8 mM D-gh.Icose was present in the incubation medium (Fig. 3). However no further increase (P > 0.2) in the output of these organic acids was observed when the glucose concentration was raised from 2.8 to 7.0 or 16.7 mM. A comparable pattern characterized the oxidation of D-[U-‘4C]ghCOSe by the tumoral cells (Fig. 3). A concentration-dependent relationship for D-[U-14Clglucose oxidation was observed, however, at low glucose concentrations in the 0.05 to 2.8 mM range, with a half-maximal response at about 0.3 mM D-glucose (Fig. 4). The oxidation of L-[U-14Clglutamine was concentration-related in the 0.1 to 10.0 mM range, with a half-maximal response at about 0.2 mM L-glutamine (Fig. 4). At all concentrations, the oxidation of Lglutamine was higher than that of Dglucose. The oxidation of L-[U-‘4C]leucine averaged 1.08 f 0.03 and 1.49 + 0.17 pmol 60 min-’ lo3 cells-’ at L-leucine concen-
+ + f f f *
LACTATE
OUTPUT P
PYRUVATE
0 i ii 4 ki I
10.
GLUCOSE
Ott, 1.7 m28 587.0a3 GLUCOSE
I
OUTPUT
OXIDATION
187
@i)
FIG. 3. Output of lactic and pyruvic acids and oxidation of D-[D-“C&iucose at increasing concentrations of D-ghCOSe. Mean values (*SE, whenever it exceeds the size of the mean symbol) refer to 2% 44 (lactate output), 16-32 (pyruvate output), and 828 (glucose oxidation) individual measurements.
GLUCOSE 12.5
METABOLISM
IN
r
INSULIN-PRODUCING
P
-“Cl
GLUTAMINE
/
TUMORAL
CELLS
567
lease, which reached a value higher than that observed, within the same experiments, in the presence of 7.0 mM D-glucose (P < 0.001; d.f. = 57). L-Glutamine failed to affect L-leucine-stimulated insulin output. DISCUSSION
oh0
cl055
0.1
”
a203 1.0 1.7 3.0 NUTRIENT(mM)
FIG. 4. Oxidation of L-[U-“Clglutamine “Clglucose at increasing concentrations scale) of these nutrients. Mean values to 8-28 individual measurements.
trations of ‘1.0 and 20.0 (n = 8 in both cases). Insulin
IrIM,
56
rao
and D-[U(logarithmic (*SE) refer
respectively
At variance with a prior report (4), the present results indicate the presence, in tumoral insulin-producing cells, of a highK, glucose-phosphorylating enzymatic activity in addition to low-K, hexokinaselike activity. The latter activity itself consisted of both a typical hexokinase, with a Km for glucose close to 0.05 mM and a Ki for glucose 6-phosphate close to 0.1 mM, and a second isoenzyme with a Km for glucose close to 0.2-0.3 mM and an apparently lesser sensitivity to glucose 6phosphate (half inhibition at approximately 1.0 mM glucose 6-phosphate). An uncommon hexokinase (peak III,) was also noticed by Meglasson and Matschinsky on chromatography of insulinoma homogenates (21). The presence of a high-K, glucokinase-like enzyme in the tumoral cells is also in good agreement with the data obtained by these authors (5). The
Content and Release
The insulin content of the cells averaged, in a series of seven preparations, 345 f 66 nU/103 cells. In the absence of exogenous nutrient, the basal release of insulin, relative to the insulin content, was much Ihigher than in normal pancreatic islets, amounting to 113 + 9 nU 90 min-’ lo3 cells-’ (Table IV). In the presence of 2.8 mM D-glucose, the secretory rate was slightly higher than basal value (P < 0.02; d.f. = 38). However, no significant difference in insulin output was observed when the glucose concentration was raised from 2.8 to 16.7 mM. At the latter glucose concentration, both forskolin and theophylline markedly augmented insulin output (P < 0.05 or less; d.f. = 18 or more). In the presence of glucose, Lornithine also augmented insulin release (P < 0.001; d.f. = 57). In the absence of glucose, L-leucine stimulated insulin re-
TABLE INSULIN
D-Glucose b-4 2.8 5.6 7.0 16.7 16.7 16.7 7.0 -
RELEASE
IV
BY TUMORAL
Other secretagogues bw Forskolin (0.005) Theophylline (1.4) L-Ornithine (10) L-Leucine (10) L-Leucine (10) + Lglutamine (10)
o Mean values (*SE) are shown number of individual observations
CELLS Insulin output= (nU 90 min-’ 10s cells-‘) 112.5 133.4 137.7 147.4 138.4 226.1 223.0 197.5 230.5 232.7
f 9.0 (30) f 7.4 (20) zt 8.2 (20) + 6.2 (20) f 3.5 (44) f 13.0 (23) k 15.2 (10) + 11.2 (30) + 7.5 (30) f 6.5 (30)
together with the (in parentheses).
568
GIROIX
activity of this enzyme was documented here with either D-glucose or D-mannose as substrate, both in the absence or presence of glucose 6-phosphate. Unexpectedly, this high-Km enzyme appeared, in insulinproducing cells, not totally unresponsive to glucose 6-phosphate. Whatever their precise identity, the glucose-phosphorylating enzymes, as well as phosphoglucomutase and lactate dehydrogenase, displayed an activity three to four times higher in tumoral than nontumoral islet cells, when related to the protein content of these cells. The metabolism of glucose in intact tumoral cells, when compared to normal islet cells, displayed several and severe abnormalities. First, relative to protein content, the rate of glucose utilization in the range 2.8-16.7 mM of glucose concentrations was 3 to 13 times higher in tumoral cells than in normal islets. Indeed, after correction for basal values, the output of lactate and pyruvate and the oxidation of glucose averaged 67 pmol glucose residue 60 min-’ lo3 tumoral cells-’ as compared to mean values for glucose utilization of 27 and 129 pm0160 mine1 islet-’ in normal pancreatic islets incubated at 2.8 and 16.7 mM D-glucose, respectively (22). Second, at low glucose concentrations (2.8-3.0 mM), the rate of glucose utilization by intact tumoral cells (1.07 pmol min-’ lo3 cells-‘, at 37’C) was not vastly different from its rate of phosphorylation in cell homogenates (1.10 pmol min-’ lo3 cells-’ at 30°C). This suggests that, at variance with the situation documented in normal islet cells (16), hexokinase is little inhibited by endogenous glucose 6phosphate in tumoral cells exposed to a low concentration of glucose. Third, in tumoral cells, the oxidation of glucose and the output of lactic and pyruvic acids reached a close-to-maximal value at a low concentration of glucose (2.8 mM), in fair agreement with the data obtained by Halban et aZ. (4) for the production of 3Hz0 from D-[5-3H]glucose in the same cell line. This behavior contrasts sharply with the sigmoidal pattern of glucose utilization in
ET
AL.
normal pancreatic islets (23). Since the pattern of glucose phosphorylation at increasing glucose concentrations was not vastly different in tumoral and nontumoral islet cell homogenates, the disparity in the concentration-dependent relationships for glucose utilization in intact cells points to a difference between tumoral and normal cells at a site distal to the phosphorylation of this hexose. Fourth, after correction for basal values, the production of 14C02 from D-[U-14C]ghCOSe in tumoral cells accounted for no more than 12.3% of the rate of glucose utilization, whereas in normal pancreatic islets the ratio of glucose oxidation/utilization is close to 28.0% (23). Thus, in the insulinproducing tumoral cells, like in other tumoral cells (24), the rate of glucose oxidation was abnormally low relative to the rate of pyruvate generation. Last, in the tumoral cells, the lactate/pyruvate output ratio was close to 2.4, as distinct from values of 6.7 and 10.0 in normal islets exposed to 2.8 and 16.7 mM D-glucose, respectively (12). This comparison suggests that the cytosolic NADH/NAD+ ratio was maintained at a lower value in tumoral than in normal islet cells. This would represent a favorable situation for the maintenance of a high rate of glycolysis in tumoral cells, but may adversely affect insulin release, to the extent that the latter process depends on the induction of a more reduced state in cytosolic redox couples (25). Our secretory data were in fair agreement with the metabolic data in that insulin release was slightly increased by D-glucose above basal value, but failed to be significantly affected when the glucose concentration was raised from 2.3 to 16.7 mM. Under suitable conditions, the tumoral cells were duly responsive, however, to other secretagogues such as L-leucine, forskolin, theophylline, or ornithine. These findings complement prior observations (l-3). L-Glutamine failed to enhance Lleucine-stimulated insulin release. This unexpected situation and the low rate of L-leucine oxidation relative to that of
IGLUCOSE
METABOLISM
IN
INSULIN-PRODUCING
other nutrients could be related to the fact that thle tumoral cells are cultured in a medium containing 2.0 lllM L-glutamine and, hence, may have a high basal glutamate clontent (26). Indeed, in islet cells, a rise in glutamate content is associated with a decreased oxidation of Lleucine (27). At this point it should be stressed that, although the responsiveness of glutamate dehydrogenase to ADP and L-leucine appeared similar in tumoral and normal islet cells, its activity relative to protein content was about twice lower in tumoral than normal cells (20). Such was not the case for ornithine decarboxylase, the activity of which was, relative to protein content, about 30 times higher in tumoral than normal islet cells. Incidentally, the novel demonstration of ornithine decarboxylase activity in insulin-producing cells, taken together with the known insulinotrop’ic capacity of this amino acid (28), calls, by analogy with studies performed in o,ther tissues (29), for further investigation on the possible role of polyamines in the process of stimulus-secretion coupling in the pancreatic B cell (30). In conclusion, the present work illustrates the existence of dramatic alterations in the metabolism of glucose and secretory response to this hexose in insulin-producing tumoral cells, as compared to normal pancreatic islets. These abnormalities were apparently not attributable to any gross difference in the hexosephosphorylating enzymatic activities of tumoral cells and normal islets, respectively. Therefore, the present data reinforce the view (31) the regulatory mechanisms located distally to the site of glucose phosphorylation participate in the control of glucose utilization and, hence, glucose-induced insulin release in insulinproducing cells. ACKNOWLEDGMENTS The authors
TUMORAL
CELLS
569
Schoonheydt and M. Urbain for technical assistance; and to C. Demesmaeker for secretarial help.
REFERENCES 1. SOPWITH, A. M., HUTTON, J. C., NABER, S. P., CHICK, W. C., AND HALES, C. N. (1981) Diabetologia 21,224-229. 2. PRAZ, G. A., HALBAN, P. A., WOLLHEIM, C. B., BLONDEL, B., STRAUSS, A. J., AND RENOLD, A. E. (1983) B&&em. J. 210, 345-352. 3. HOENIG, M., FERGUSON, D. C., ANDMATSCHINSKY, F. M. (1984) Diabetes 33, l-7. 4. HALBAN, P. A., PRAZ, G. A., AND WOLLHEIM, C. B. (1983) B&hem J. 212,439-443. 5. MEGLASSON, M. D., AND MATSCHINSKY, F. M. (1983) J. Biol. Chem. 258, 6’705-6708. 6. GIROIX, M.-H., MALAISSE-LAGAE, F., SENER, A., AND MALAISSE, W. J. (1985) Mol. Cell. B&hem. 66, 61-64. 7. LACY, P. E., AND KOSTIANOVSKY, M. (1967) Diabetes 16, 35-39. 8. MALAISSE, W. J., GIROIX, M.-H., DUFRANE, S. P., MALAISSE-LAGAE, F., AND SENER, A. (1985) B&hem. Int. 10,233-240. 9. MALAISSE-LAGAE, F., SENER, A., AND MALAISSE, W. J. (1982) Biochimie 64, 1059-1063. 10. MALAISSE, W. J., KAWAZU, S., HERCHUELZ, A., HUTTON, J. C., SOMERS, G., DEVIS, G., AND SENER, A. (1979) Arch. B&hem. Biophys. 194, 49-62. 11. HUTTON, J. C., SENER, A., AND MALAISSE, W. J. (1979) B&hem. J. 184, 291-301. 12. GIROIX, M.-H., DUFRANE, S. P., MALAISSE-LAGAE, F., SENER, A., AND MALAISSE, W. J. (1984) B&hem. Biophys. Res. Cmnmun. 119,543-548. 13. CARPINELLI, A. R., SENER, A., HERCHUELZ, A., AND MALAISSE, W. J. (1980) Metab. Clin. Exp. 29,540-545. 14. SENER, A., AND MALAISSE, W. J. (1976) B&hem. Med. 15, 34-41. 15. MALAISSE, W. J., HUTTON, J. C., KAWAZU, S., AND SENER, A. (1978) Eur. J B&hem. 87,121-130. 16. GIROIX, M.-H., SENER, A., PIPELEERS, D. G., AND MALAISSE, W. J. (1984) B&hem. J 223, 447453. 17. SENER, A., GIROIX, M.-H., AND MALAISSE, W. J. (1984) Eur. J. Biochem. 144, 223-226. 18. GIROIX, M.-H., DUFRANE, S. P., SENER, A., AND MALAISSE, W. J., Submitted for publication. 19. MALAISSE, W. J., SENER, A., WELSH, M., MALAISSE, W. J., HELLERSTR~M, C., AND CHRISTOPHE, J. (1983) B&hem. J. 210, 921-927. 20. SENER, A., AND MALAISSE, W. J. (1980) Nature (London) 288, 187-189.
GIROIX 21. MEGLASSON, M. D., AND MATSCHINSKY, F. M. (1984) in Methods in Diabetes Research (Larner, J., and Pohl, S. L., eds.), Vol. 1, pp. 213226, Wiley, New York. 22. MALAISSE, W. J., SENER, A., LEVY, J., AND HERCHUELZ, A. (1976) Acta Diabetol. Lot. 13, 202215. 23. SENER, A., LEVY, J., AND MALAISSE, W. J. (1976) Biochem J. 156.521-525. 24. REITZER, L. J., WICE, B. M., AND KENNELL, D. (1979) J. BioL C&m. 254, 2669-2676. 25. MALAISSE, W. J., MALAISSE-LAGAE, F., AM) SENER, A. (1984) Experientia 40, 1035-1043. 26. MALAISSE, W. J., SENER, A., CARPINELLI, A. R., ANJANEYULU, K., LEBRUN, P., HERCHUELZ, A.,
ET
27.
28. 29. 30.
31.
AL. AND CHRISTOPHE, J. (1989) Mel Cell Endock&. 20,171-x39. MALAISSE, W. J., SENER, A., MALAISSE-LAGAE, F., WELSH, M., MATTHEWS, D. E., BIER, D. M., AND HELLERSTR~M, C. (1982) J. Biol Chem 257,8731-8737. CHARLES, S., TAMAGAWA, T., AND HENQUIN, J. C. (1982) B&hem J. 208, 301-308. KOENIG, H., GOLDSTONE, A., AND Lu, C. Y. (1933) Nature (London) 305,530-534. MALAISSE, W. J., LEBRUN, P., AND HERCHUELZ, A. (1985) in Calcium Physiology (MarmB, D., ed.), Springer-Verlag, Berlin, in press. SENER, A., AND MALAISSE, W. J. (1984) Experientie 40, 1026-1035.