| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Articles |
Departments of Physiology (A.G., S.J.F., R.H.M., Z.Q.S., M.V.), Medicine (A.G., Z.Q.S., M.V.), and Surgery (Z.Q.S.), University of Toronto, Toronto, Ontario, Canada
Address all correspondence and requests for reprints to: Adria Giacca, M.D., Department of Physiology, University of Toronto, Medical Sciences Building, Room 3363, Toronto, Ontario, Canada M5S 1A8. E-mail: adria.giacca{at}utoronto.ca
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
550 ng/liter), achieved by a high rate
portal glucagon infusion (5 ng/kg·min). As in the previous study
(basal glucagon levels, 170 ng/liter), we used depancreatized dogs
and assessed GP with HPLC-purified [6-3H]glucose. After
obtaining constant basal hyperglycemia (10 mM) with
portal infusions of insulin (4.8 ± 0.5 pmol/kg·min) and
glucagon, an additional infusion of insulin was administered for 180
min, either portally (portal; n = 7) or peripherally (peripheral;
n = 8) at the same rate (5.4 pmol/kg·min) or at half that rate
peripherally (1/2 periph; n = 5). Plasma glucose and glucose
specific activities were clamped at basal levels. Systemic insulin
levels increased by 215 ± 16, 310 ± 26, and 184 ± 15
pM, and estimated hepatic insulin levels increased by
398 ± 20, 310 ± 26, and 184 ± 15 pM with
portal, peripheral, and 1/2 periph, respectively. GP was suppressed to
the same extent with portal and peripheral (53 ± 6% and 50
± 6%), but less with 1/2 periph (35 ± 5%). FFA levels were
suppressed to a greater extent with peripheral than portal or 1/2
periph, whereas the responses of lactate alanine and glycerol to
insulin infusion were similar in the three groups. Thus, in the present
report, unlike in our previous study, 1) suppression of GP was
proportional to the hepatic insulin levels; and 2) systemic insulin
levels did not dominate suppression of GP. We, therefore, conclude that
in hyperglycemic depancreatized dogs 1) glucagon, at concentrations
seen in poorly controlled diabetes, can unmask a direct effect of
hepatic insulin levels on GP; and 2) the suppression of glucagon may
play a role in the peripheral effect of exogenously delivered insulin
on GP. This is the first in vivo study to show that the
main direct effect of insulin on the liver is to counteract the effect
of glucagon. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The mechanism of insulin’s peripheral regulation of GP remains elusive. Recent studies in normal dogs suggest that the insulin-induced suppression of FFA plays an important role in insulin’s peripheral effect (7, 8). In our study in depancreatized dogs, full dose peripheral infusion resulted in a slightly greater decrease in FFA, glycerol, alanine, and glucagon than equidose portal infusion. We calculated that the differential suppression of the glucagon levels could have accounted for more than half of the differential suppression of GP between the two treatments.
The aim of the present study was to assess the role of glucagon in mediating the indirect regulation of GP by peripheral insulin levels in diabetes. We, therefore, infused insulin portally or peripherally at the same rate using methods identical to those used in the previous study, except we maintained glucagon at a constant level. This was performed to determine whether the previously observed difference in suppression of GP with equidose peripheral and portal insulin infusion is diminished when differences in glucagon between the two protocols are eliminated. To clamp the glucagon levels, we infused glucagon at a high rate to override the endogenous glucagon secretion and thus ensure that glucagon levels would not decrease during insulin infusion. The glucagon infusion rate resulted in 3-fold higher glucagon levels than those in normal dogs (9, 10) or in the previous study in moderately hyperglycemic insulin-infused depancreatized dogs (5). These high levels can be seen in poorly controlled alloxan-diabetic dogs (11) as well as during exercise (9) in normal dogs. We also compared portal and half-dose peripheral infusions resulting in equivalent peripheral insulin levels during high rate glucagon infusion. This was performed to determine whether the high glucagon level present in the current study per se could have altered the balance between insulin’s direct and indirect effects on GP. In vitro data (isolated rat hepatocytes and liver perfusion preparations) suggest that glucagon can enhance insulin’s direct effect on GP (12, 13). If this occurs in vivo, portal insulin infusion that did not suppress GP more than half-dose peripheral infusion despite greater hepatic insulinization when glucagon levels were basal (i.e., in our previous study) could prove more effective in suppressing GP under conditions of a high glucagon clamp. This was indeed the major finding of this current study, which, to our knowledge, is the first to demonstrate the importance of glucagon for insulin’s direct action on GP in vivo.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Experimental design
Three hyperinsulinemic glucose clamp protocols were carried out
in overnight fasted dogs under conditions of hyperglucagonemia and
moderate hyperglycemia as follows: protocol 1, insulin (54 pmol/kg
bolus plus 5.4 pmol/kg·min) was given intraportally (portal; n =
7); protocol 2, the same dose of insulin was infused into a peripheral
vein (peripheral; n = 8), resulting in much higher peripheral
insulin levels than with the equidose portal infusions; in this
protocol, portal levels of insulin were calculated to be lower than
with portal infusion, due to dilution of insulin in the greater
systemic than portal blood volume (the method for calculation of portal
insulin levels is described below); and protocol 3, peripheral
infusions of insulin were given to match increments in systemic
insulinemia as obtained with portal insulin infusions. Assuming a 50%
hepatic degradation of insulin, we infused peripherally half the dose
(half-periph; n = 5) given in the other protocols (27 pmol/kg
bolus plus 2.7 pmol/kg·min). In this protocol, portal insulin levels
were calculated to be lower than those in protocol 1 or 2. The
experiments were carried out at least 2 weeks after pancreatectomy, and
at least 7 days elapsed between successive experiments in the same dog.
Only healthy dogs that were weight-stable and had at least 5 days of
relatively well controlled diabetes (glycosuria, <1%) and an absence
of diarrhea and visible steatorrhea were used for experimental studies.
Of the 10 dogs studied, 3 were subjected to all 3 protocols, and 7
experiments were paired in protocols 1 and 2. The last insulin
injection (usual dose of regular insulin, one half to two thirds the
usual dose of NPH insulin) was given with the last meal at least
24 h before the experiment to ensure initial hyperglycemia (>16
mM) the following morning. Food was withdrawn for 18 h
before each experiment.
On the day of the study, an intraportal infusion of crystalline porcine
insulin (Iletin II, Eli Lilly Co., Indianapolis, IN) was started (20
pmol/kg·min) and gradually reduced. When blood glucose approached 14
mM, an intraportal infusion of glucagon (Eli Lilly Co.) was
initiated (5 ng/kg·min) and maintained throughout the experiment.
Intraportal insulin infusion was adjusted until a constant basal rate
(4.8 ± 0.5 pmol/kg·min) maintained moderate hyperglycemic
levels (10 mM) for 60 min or more before the clamp. When
plasma glucose approached 12 mM, a primed (155.4 x
106 dpm) continuous infusion (1.11 x 106
dpm/min) of a HPLC-purified tracer mixture containing 50%
2-[3H]glucose and 50% 6-[3H]glucose was
also started and maintained throughout the experiment. These two
tracers were used to determine hepatic glucose cycling (GC) in addition
to glucose production and utilization. We wished to measure GC for two
reasons: 1) glucagon has previously been shown to be an important
regulator of GC (15, 16); and 2) insulin-induced changes in hepatic GC
may be a consequence of insulin’s direct effect on hepatic glucose
metabolism. After 120 min or more of tracer equilibration and at least
30 min of steady state hyperglycemia, four arterial samples were taken
(from -30 to 0 min) for basal determinations. At time zero, a
suprabasal infusion of insulin was started and maintained for 180 min
through either the portal or peripheral venous route according to the
protocol design. During the suprabasal insulin infusions, plasma
glucose was clamped at the mean basal level by adjusting the rate of a
variable infusion of glucose according to frequent (every 5 min)
glycemic determinations.
An aliquot of the 2.22 x 107 dpm/ml mixture of 2-[3H]- and [6-3H]glucose was added to the glucose infusate (25% dextrose; Abbott Laboratories, Montreal, Canada) to minimize the decline in specific activity during the clamp (HOT Ginf method). The specific activity of the infusate was calculated based on estimations of the parameters of the formula of Finegood et al. (17), modified to allow for the incomplete suppression of GP: SAGinf = I x [(GinfSS/GPb) - F]/(GinfSS x BW), where SAGinf is the specific activity of the glucose infusate (disintegrations per min/µmol), I is the constant tracer infusion rate (disintegrations per min), GinfSS is the steady state glucose infusion rate (micromoles per kg/min), GPb is the basal glucose production (micromoles per kg/min), BW is the body weight (kilograms), and F (fractional suppression) = (GPb - GPSS)/GPb. The following estimates, based on previous (5) and pilot experiments, were used in all experiments: GPb = 28 µmol/kg·min, GPSS = 0.50 · GPb, and GinfSS = 27 µmol/kg·min. Although two tracers were used, we based our estimates primarily on glucose production (rate of appearance of glucose determined with [6-3H]glucose), as this was the main variable of interest in our study.
During the clamp, arterial samples were taken every 10 min in the first and third hours and every 15 min in the second hour. The total amount of blood sampled in each experiment was 150–160 ml. The procedures for the collection of samples have been described previously (18).
Laboratory methods
All laboratory methods are identical to those used in the
previous study (5). Plasma glucose concentrations were measured by the
glucose oxidase method on a glucose analyzer (Glucose Analyzer II,
Beckman Instruments, Fullerton, CA). Insulin and glucagon were analyzed
by RIA (19, 20). Plasma samples of six normal control dogs were
analyzed together with those in the present study. On these normal
control samples, the insulin assay reading was 56.3 ± 3.6
pM, and the glucagon assay reading was 174 ± 29
ng/liter. FFAs were determined by a radiochemical technique (21);
lactate, alanine, and glycerol were determined by enzymatic
fluorometric methods (22). Specific activities for
[6-3H]- and [2-3H]glucose were determined
with the dimedone precipitation technique (23), as previously described
(5).
Calculations
GP and glucose output (GO) were calculated as the endogenous
rate of glucose appearance measured with [6-3H]glucose
and [2-3H]glucose, respectively; glucose utilization (GU)
was calculated as the rate of glucose disappearance (Rd) measured with
[6-3H]glucose. Rd corresponded to GU, and the plasma
clearance rate of glucose (Rd/glycemia) corresponded to the glucose MCR
because plasma glucose levels were below the canine renal threshold for
glucose (18). To account for the exogenously infused mixture of labeled
and unlabeled glucose, a modified one-compartment model (17) was used
for the calculation of GP = I/SA + SAGinf x
Ginf/SA - pVG(dSA/dt)/SA - GINF and GU = I/SA +
SAGinf x Ginf/SA - pVG(dSA/dt)/SA - pVdG/dt,
where G is the plasma glucose concentration, p (pool fraction) = 0.65,
and V (distribution volume of glucose) = 25% of body weight as
described previously (17). Data were smoothed with the optimized
optimal segments routine (24). With the HOT Ginf method, the
monocompartmental assumption becomes minor, because the nonsteady state
component of Steele’s equation is close to zero. GC was calculated as
the difference between GO and GP. Portal insulin concentrations were
estimated by a modified Fick principle (2). The equation
IPOR = IPER + INFI/PPF (2) was
used, where IPOR and IPER are portal and
peripheral insulin concentrations, INFI is the portal
insulin infusion rate, and PPF is the portal plasma flow. Hepatic
insulin concentrations were calculated based on a weighted average of
72% supply from the portal vein and 28% from the hepatic artery (25).
A portal plasma flow rate of 500 ml/min (3) was used for all
calculations. It was also assessed whether weight standardization of
portal plasma flow led to different results. If portal flow rates were
standardized by weight either using 20.8 ml kg/min (500 ml/min divided
by 24 kg, which was the average weight of dogs) as portal plasma flow
or using 19.2 ml/kg·min as portal plasma flow [assuming a hepatic
blood flow of 44.5 ml/kg·min (25), 72% contribution of portal flow
to hepatic flow (25), and 40% hematocrit], the resulting hepatic
insulin levels differed from our reported values by less than 20
pM.
Statistics
The data were expressed as the mean ± SE.
ANOVA for repeated measures was carried out to determine differences
between experimental protocols and between periods of the experiment
within each protocol (basal, -30 to 0 min; clamp, 0–180 min). In the
case of unequal variances, data were logarithmically transformed. The
clamp values reported refer to the last 2 h of the clamp.
Pearson’s correlation coefficient was calculated by linear regression
analysis. Calculations were performed with SAS software (Statistical
Analysis System, Cary, NC), and unless otherwise indicated,
significance was presumed at P < 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Specific activities of
[6-3H]- and [2-3H]glucose did not change
significantly from the respective basal levels for each protocol.
|
) in systemic insulin levels of
310 ± 26 pM with peripheral insulin infusion that was
higher (P < 0.001) than the increment achieved with
both equidose portal insulin infusion (215 ± 16 pM)
and half-periph insulin infusion (184 ± 15 pM; Fig. 2). The half-periph insulin infusion resulted in
increments in plasma insulin levels comparable to those obtained with
portal infusion (P = NS). The calculated basal hepatic
insulin levels were 273 ± 39 with peripheral, 350 ± 84 with
portal, and 448 ± 67 pM with half-periph
(P < 0.05, peripheral vs. half-periph). As
shown in Fig. 2, the rise in hepatic insulin levels with portal
(398 ± 20 pM) was higher (P < 0.001)
than that with peripheral (310 ± 26 pM) insulin
infusion, both of which were higher (P < 0.001) than
that with half-periph (184 ± 15 pM) insulin
infusions. Thus, by experimental design, portal insulin infusion
resulted in the highest level of hepatic insulinization, whereas
peripheral infusion resulted in the highest level of systemic
insulinemia, and portal and half-periph infusion resulted in equal
increments in systemic insulinemia. The rates of exogenous glucose
infusion (Fig. 2) required to clamp plasma glucose at the basal
hyperglycemic levels were higher (P < 0.01) with
peripheral insulin infusion (31.5 ± 3.4 µmol/kg·min) than
with either portal (26.1 ± 5.1) or half-periph insulin infusions
(21.2 ± 3.1).
|
). Basal FFA levels were similar between protocols
(Fig. 3). FFA levels were suppressed during the clamp to lower levels
with peripheral insulin infusion (P < 0.05) than with
either portal or half-periph insulin infusions.
|
shows the individual comparisons between
glucose turnover in the portal and equidose peripheral protocol
(left side) and glucose turnover in the portal and half-dose
peripheral protocol (right side). When systemic insulin
levels rose higher with peripheral than equidose portal infusion, as
expected GU (Fig. 4, top left) was higher
(P < 0.01) with peripheral insulin (45.4 ± 4.5
vs. 39.4 ± 4.3 µmol/kg·min). As observed in Fig. 4
(bottom left), GP was suppressed to similar levels
with either peripheral (from 28.0 ± 1.9 to 14.5 ± 2.7;
GP, -14.0 ± 0.6 µmol/kg·min) or portal (from 26.7 ±
2.6 to 13.0 ± 2.2; GP = -13.7 ± 0.5
µmol/kg·min) insulin infusions, which represent equivalent
suppressions of 50 ± 6% and 53 ± 6%, respectively. The
time course of GP suppression was similar with equidose peripheral or
portal insulin infusions (Fig. 4, bottom left). These
results are in contrast with those of our previous study, in which
basal glucagon levels declined and a greater suppression of GP by
peripheral than equidose portal insulin infusion was observed (5) (Fig. 5, top).
|
|
, top right) to
39.4 ± 4.3 µmol/kg·min with portal insulin and 39.2 ±
3.6 µmol/kg·min with half-periph insulin infusion. The greater
hepatic insulinization with portal insulin infusion during the clamp
resulted in greater (P < 0.001) suppression of GP
(from 26.7 ± 2.7 to 13.0 ± 2.2; GP = -13.7 ±
0.5 µmol/kg·min) than with the half-periph insulin infusion (from
27.7 ± 2.0 to 18.4 ± 1.3; GP = -9.8 ± 0.6
µmol/kg·min; Fig. 4, bottom right), which
represents 53 ± 6% vs. 35 ± 5% suppression
(portal and half-periph insulin infusions, respectively). This is in
contrast with our previous study (5), which showed equal suppression of
GP with the same portal and half-periph insulin infusion protocols
(Fig. 5, top). Of note, the magnitude of the decline in FFA
levels was less than that in the previous study in all three protocols
(Fig. 5, bottom), because basal FFA levels were lower (Table 1); however, it is of note that the patterns of FFA
suppression were similar in the studies, i.e. FFA still
declined to a greater extent with full-dose peripheral infusion (Fig. 5, bottom). In Table 1, the basal and clamp averages of
insulin, glucagon, FFA, GP, GU, and GC in the present and previous
studies are summarized. Values for both studies refer to the 60- to
180-min period of the clamp. In the present study, basal GC was lower
(P < 0.01) in the peripheral protocol than the portal
protocol. As indicated in Table 1, the rates of GC did not change from
basal in response to suprabasal insulin infusions.
|
shows metabolite levels. The values of the
previous study are given for comparison. Basal lactate levels (Table 2)
were similar between the portal and peripheral protocols, but were
lower (P < 0.05) in the half-periph protocol group.
During the clamp there was a trend for lactate levels to be suppressed
in all protocols, but this suppression was not significant and not
different between protocols. Basal alanine levels were similar between
the portal and peripheral protocols, but were lower (P
< 0.05) in the half-periph protocol group. All insulin treatment
protocols resulted in a similar suppression () of alanine. Basal
glycerol levels were similar between the portal and peripheral
protocols, but were lower (P < 0.05) in the
half-periph protocol group. During the clamp, glycerol levels were
suppressed with both portal and peripheral insulin treatment. Glycerol
levels during the clamp were similar between all three protocols (Table 2).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
It might be argued that we are comparing the present results with historical controls. We elected not to repeat the protocol of the previous study, when glucagon levels declined, for two reasons: 1) all the methods in the laboratory have remained the same since the previous study; and 2) an additional study was performed, which was identical to the previous study except for the lower doses of insulin infused. The latter study, which was performed simultaneously with the present study, confirmed the results obtained with the higher doses (i.e. that glucagon and GP declined proportionally to peripheral insulin levels) (26).
To obtain the same preclamp plasma glucose levels as in the previous
study (5), the basal intraportal insulin infusion rates had to be
higher in the present study to offset the hyperglycemic action of
glucagon. Therefore, basal insulin levels were also higher. In the
presence of more insulin and glucagon, basal GP and GU were greater
than those in the previous study. Basal GC (glucose 
glucose-6-phosphate) was not different, which supports our previous
observations (27) that acutely, GC is substrate dependent in
depancreatized dogs (i.e. similar GC rate at similar
glycemia). Basal alanine, lactate, and FFA were lower than those in the
previous study, presumably due to the higher basal insulin (FFA) or
basal insulin and glucagon (alanine and lactate) levels. Whereas it is
possible that the higher basal insulin might have influenced
comparisons between the present and our previous study, data in the
literature suggest that the direct effect of insulin on the liver is
not more evident, but is, rather, diminished when insulin levels are
high (28).
Although it might be argued that systemic hyperinsulinism might indirectly modify hepatic hemodynamics (via activation of the sympathetic nervous system), differences in insulin levels of the magnitude observed between the previous and the present study have been shown not to affect hepatic hemodynamic parameters, such as portal venous, arterial hepatic, or hepatic blood flow (25).
In the present study, there were small intergroup differences in basal insulin and metabolite levels. Basal insulin levels were highest in the half-periph group and lowest in the peripheral group, although the difference did not reach statistical significance. Basal GC was lowest in the peripheral group, which might indicate that insulin sensitivity was greatest. The half-periph group had the lowest lactate, alanine, and glycerol levels, which might be in accordance with the highest insulin levels. Of course, these differences might also be due to random intergroup variability. One might argue that suppression of GP was greater in the portal than in the half-periph infusion group, because this group was more insulin resistant. In addition, the rise in insulin levels, although not significantly different, was not completely superimposable in the two groups. However, glucose utilization, which is a more sensitive indicator of impaired insulin action than GP in diabetes (29), was equivalent in the portal and half-periph groups, suggesting equal peripheral insulin action during the clamp.
When the suprabasal insulin infusions were given, GP decreased less
than in the previous study during the peripheral and half-dose
peripheral protocols, probably because the glucagon decline was absent.
However, in the portal protocol, the percent decrease in GP was similar
to that observed in the previous study, and the value (differences
of GP from basal) was actually greater. Glucose cycling did not change
during insulin infusion by either route, as in the previous study.
In the previous study, peripheral insulin infusion decreased alanine and glycerol more than portal infusion. However, we calculated that the small difference in gluconeogenic precursors could only have accounted for a minimal part of the observed difference in GP. In the present study, no such differences were observed. FFA decreased more with peripheral than portal infusion, as in the previous study. Despite this difference, GP was similarly suppressed, which suggests that suppression of FFA is not the sole mechanism of GP suppression in depancreatized dogs.
In insulin-withdrawn depancreatized dogs, up to 70% of the decrease in GP induced by insulin is due to suppression of glucagon (30). In depancreatized dogs, extrapancreatic glucagon is secreted by the gastric mucosa (31) and maintains normal to mildly elevated plasma glucagon levels during insulin treatment (32, 33). Insulin inhibits extrapancreatic as well as pancreatic glucagon. Exogenous insulin infusions, whether given portally or peripherally, inhibit glucagon in proportion to systemic insulin concentrations, because the site of portal insulin infusion is downstream from the pancreas. However, the route by which insulin inhibits glucagon levels in normal physiology is thought to be different, because endogenous insulin can affect glucagon secretion in proportion to the local pancreatic insulin levels (34, 35).
In the previous study, peripheral insulin infusion resulted in a slightly greater suppression of the peripheral glucagon levels than equidose portal infusion. The difference just failed to reach statistical significance. However, when we later reanalyzed the plasma samples from the experiments on the same dogs in the same assay, the difference proved significant (P < 0.05; unpublished data). Such small differences in peripheral glucagon levels are difficult to detect in glucagon immunoassays that also measure cross-reacting material (36) and underestimate differences in hepatic glucagon concentrations. Even small differences in hepatic glucagon levels are important (36, 37), because glucagon drives as much as 75% of basal GP (38). We calculated that the previously observed difference in glucagon levels could have sustained more than half of the difference in GP suppression between equidose portal and peripheral insulin infusions.
In the present study GP suppression was similar with equidose portal or peripheral insulin infusion at similar glucagon levels, which suggests that 1) the effect of peripheral insulin was diminished when the difference in glucagon was eliminated; or 2) the effect of hepatic insulin was enhanced at high glucagon levels; or 3) both 1 and 2 are correct. That high glucagon levels can potentiate the effect of hepatic insulin is demonstrated by the different results of the present compared to those of our previous study when portal and half-dose peripheral infusions are considered. In both studies, hepatic insulin levels were greater with portal than half-dose peripheral infusion, whereas peripheral insulin levels were matched (i.e. a selective difference in hepatic insulin levels was generated). However, only in the present study at high glucagon levels, and not in the previous study when glucagon levels were basal (and declined to the same extent with the two protocols), was suppression of GP greater with portal infusion than half-dose peripheral infusion (and was, therefore, dependent on hepatic insulin levels).
It should be considered that with peripheral insulin infusion, both portal and peripheral insulin levels increase, and the same occurs with portal insulin infusion, although with the latter, the elevation in peripheral insulin levels is less than that in portal levels. With both half-dose and full-dose peripheral infusions, the percent suppression of GP was reduced (52% and 35%) compared to that in the previous study (71% and 59%), whereas with portal infusion, the absolute decrease in glucose GP was greater; however, the percent GP suppression was not increased (53% vs. 56%). Thus, despite the effect of portal insulin levels being potentiated in the present study (see the above discussion), the effect of portal insulin infusion (which elevates both portal and peripheral insulin levels) was not markedly increased compared to that in the previous study, and the effect of peripheral insulin infusions was decreased. Therefore, although the effect of peripheral insulin levels was not directly assessed, it should have been reduced. We take these results to indicate that glucagon suppression mediates part of the peripheral effect of insulin in our model. However, we cannot exclude that in the present study, glucagon desensitized GP to the effect of FFA, presumably by altering the proportion between glycogenolysis and gluconeogenesis.
In summary, we believe that in addition to potentiating the effect of portal insulin, the high glucagon clamp also diminished the effect of peripheral insulin. However, definite evidence that a glucagon clamp diminishes the peripheral effect of insulin on GP can only be obtained by inducing a selective difference in peripheral insulin levels at matched hepatic insulin levels.
We and others have shown that both hepatic and systemic insulin levels regulate GP in nondiabetic dogs (4, 39) and humans (6). Previously, we could not observe any direct effect of the hepatic insulin levels in suppressing GP in depancreatized dogs (5). However, in the present study, glucagon restored GP dependence on hepatic insulin levels. The potentiation or unmasking of insulin’s direct effect on GP by glucagon is consistent with Unger’s hypothesis that the main role of insulin in hepatic glucose metabolism is to counteract the effects of glucagon (40). In rat hepatocytes, insulin by itself had no effect on glycogen synthesis, but it overcame glucagon’s induced inhibition of glycogen accumulation (12). The effects of insulin on glycogen synthase and phosphorylase were enhanced in hepatocytes incubated with glucagon (41). Insulin was more potent in inhibiting the release of glucose by the perfused liver in the presence than in the absence of glucagon (42, 43). Presumably, insulin’s effect on GP is easier to observe when glycogenolysis is stimulated and tissue cAMP concentrations are high (43). Previous in vivo studies suggest that insulin can be very potent in inhibiting the stimulatory effect of elevated glucagon levels on GP (44). The present study shows that it is insulin’s direct action in suppressing GP that is potentiated by high glucagon levels, as seen in poorly controlled diabetes. This study does not answer the questions of whether 1) less elevated glucagon levels, as observed after a mixed meal under nondiabetic conditions, can also potentiate insulin’s direct effects; and 2) whether the insulin-induced suppression of glucagon can diminish insulin’s direct effect on GP and thus increase the dependence of GP on insulin’s peripheral effects. However, our preliminary results in nondiabetic humans suggest that even a slight elevation of glucagon levels can potentiate insulin’s direct effect (45).
Ader and Bergman (2) and Rebrin et al. (3) suggested that in normal dogs, GP suppression during hyperinsulinemia is dominated by systemic, rather than hepatic, insulin independent of glucagon. In one study, the researchers infused somatostatin and 5 ng/kg·min glucagon during insulin infusion by either route (2); in another study, they used somatostatin to suppress glucagon and did not replace it (3). In both studies, GP was suppressed in proportion to the peripheral insulin levels. However, more recently, the same researchers have found that glucagon potentiates the effect of portal insulin to suppress GP in somatostatin-infused normal dogs (46).
In conclusion, we found that the level of glucagon is an important determinant of the balance between the hepatic and the peripheral regulation of GP by insulin in hyperglycemic depancreatized dogs. Glucagon suppression may both mediate part of the peripheral effect of exogenous insulin on GP and diminish the hepatic effect of insulin on GP, both of which would accentuate the dependence of GP on insulin’s peripheral effects. The present study is, to our knowledge, the first demonstration that glucagon can potentiate the direct effects of insulin on GP in vivo.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 A.G. and S.J.F. are to be considered first authors of this
manuscript.
3 Supported by a JDFI Career Development Award.
4 Supported by a Medical Research Council studentship.
5 Supported by a scholarship from the Banting and Best Diabetes
Center, University of Toronto.
Received July 30, 1996.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. Girard The Inhibitory Effects of Insulin on Hepatic Glucose Production Are Both Direct and Indirect Diabetes, December 1, 2006; 55(Supplement_2): S65 - S69. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Gupta, E. Park, H. Sandhu, T. Goh, V. Tchipashvili, and A. Giacca Direct and indirect effects of insulin on hepatic glucose production in diabetic depancreatized dogs during euglycemia. J. Endocrinol., September 1, 2006; 190(3): 695 - 702. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Ozawa, M. Miyazaki, M. Matsuhisa, K. Takano, Y. Nakatani, M. Hatazaki, T. Tamatani, K. Yamagata, J.-i. Miyagawa, Y. Kitao, et al. The Endoplasmic Reticulum Chaperone Improves Insulin Resistance in Type 2 Diabetes Diabetes, March 1, 2005; 54(3): 657 - 663. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Suzuki, K. Tobe, M. Aoyama, A. Inoue, K. Sakamoto, T. Yamauchi, J. Kamon, N. Kubota, Y. Terauchi, H. Yoshimatsu, et al. Both Insulin Signaling Defects in the Liver and Obesity Contribute to Insulin Resistance and Cause Diabetes in Irs2-/- Mice J. Biol. Chem., June 11, 2004; 279(24): 25039 - 25049. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. C. Camacho, R. R. Pencek, D. B. Lacy, F. D. James, and D. H. Wasserman Suppression of Endogenous Glucose Production by Mild Hyperinsulinemia During Exercise Is Determined Predominantly by Portal Venous Insulin Diabetes, February 1, 2004; 53(2): 285 - 293. [Abstract] [Full Text]
|
|
|
|
 |
|
 L. Majdoub, M. Vermorel, and I. Ortigues-Marty Ryegrass-based diet and barley supplementation: Partition of energy-yielding nutrients among splanchnic tissues and hind limbs in finishing lambs J Anim Sci, April 1, 2003; 81(4): 1068 - 1079. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Gupta, H. Sandhu, T. Goh, K. Shah, S. R. Wiesenthal, H. Yoshii, V. Chong, T. K. T. Lam, C. A. Haber, W. Williams, et al. Insulin inhibits glucose production by a direct effect in diabetic depancreatized dogs during euglycemia Am J Physiol Endocrinol Metab, November 1, 2002; 283(5): E1002 - E1007. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Satake, M. C. Moore, K. Igawa, M. Converse, B. Farmer, D. W. Neal, and A. D. Cherrington Direct and Indirect Effects of Insulin on Glucose Uptake and Storage by the Liver Diabetes, June 1, 2002; 51(6): 1663 - 1671. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Z. Zhao, M. M. Shinohara, D. Huang, M. Shimizu, H. Eldar-Finkelman, E. G. Krebs, J. A. Beavo, and K. E. Bornfeldt Leptin Induces Insulin-like Signaling That Antagonizes cAMP Elevation by Glucagon in Hepatocytes J. Biol. Chem., April 6, 2000; 275(15): 11348 - 11354. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. K. Sindelar, K. Igawa, C. A. Chu, J. H. Balcom, D. W. Neal, and A. D. Cherrington Effect of a selective rise in hepatic artery insulin on hepatic glucose production in the conscious dog Am J Physiol Endocrinol Metab, April 1, 1999; 276(4): E806 - E813. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. F. Lewis, M. Vranic, and A. Giacca Role of free fatty acids and glucagon in the peripheral effect of insulin on glucose production in humans Am J Physiol Endocrinol Metab, July 1, 1998; 275(1): E177 - E186. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. H. McCall, S. R. Wiesenthal, Z. Q. Shi, K. Polonsky, and A. Giacca Insulin acutely suppresses glucose production by both peripheral and hepatic effects in normal dogs Am J Physiol Endocrinol Metab, February 1, 1998; 274(2): E346 - E356. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. K. Sindelar, C. A. Chu, D. W. Neal, and A. D. Cherrington Interaction of equal increments in arterial and portal vein insulin on hepatic glucose production in the dog Am J Physiol Endocrinol Metab, November 1, 1997; 273(5): E972 - E980. [Abstract] [Full Text] [PDF]
|
|
| ||||
), or peripheral insulin infusion of 27 pmol/kg + 2.7 pmol/kg·min
(half-periph; n = 5; 
) in moderately hyperglycemic (
