Restoration of stable metabolic conditions during islet suppression in dogs DAVID
C. BRADLEY
AND
RICHARD
Department of Physiology and Biophysics, Los Angeles, California 90033
N. BERGMAN University
Bradley, David C., and Richard N. Bergman. Restoration of stable metabolic conditions during islet suppression in dogs. Am. J. Physiol. 262 (Endocrinol. Metab. 25): E532-E538, 1992.-These studies were undertaken to examine the stability of metabolic conditions during islet suppression with fixed-rate insulin and glucagon replacement. Somatostatin was infused peripherally at 0.8 pg* min-l . kg-l, insulin was infused intraportally at 200 cumin-’ kg-‘, and glucagon was infused intraportally at 0, 0.6, 1, 2, 5, or 20 nggmin-‘. kg-l in conscious overnight-fasted dogs. [ 3-3H]glucose was infused for measurement of glucose kinetics. During infusion, plasma insulin was 7.2 & 0.4 pU/ml. Plasma glucagon rose linearly with glucagon dose, achieving basal levels at 2 ng* min-l . kg-l infusion (164 t 18 vs. basal = 182 =f:57 pg/ml). Plasma glucose and hepatic glucose output (HGO) decreased from basal at doses 0,0.6, and 1 nge min-l . kg-‘, increased from basal at doses 5 and 20 ng* min-l* kg-l, and remained close to basal at dose 2 ng*min-l . kg-’ (92 * 20 vs. basal = 99 $- 3 mg/dl and 2.4 & 0.2 vs. basal = 2.7 * 0.2 mg*min-l gkg-’ for glucose and HGO, respectively; P > 0.47). Glucose clearance and blood lactate were also closely matched to basal at dose 2 ng.min-l kg? Coefficients of variation during 2 ng.min-’ kg-’ glucagon infusion (last hour) were 3.4, 4.6, 4.9, and 4.7% for glucose, HGO, clearance, and lactate, respectively. These findings indicate that fasting metabolic conditions, as inferred from blood glucose, lactate, insulin, and glucagon levels, and the rates of glucose production and uptake can be recreated in toto during fixed-rate islet l
l
l
hormone replacement.
of Southern
California
Medical School,
islet suppression with basal insulin replacement, presumably to compensate for a somatostatin-induced increase in glucose clearance (3). Thus, in our experiments, it was not possible to simultaneously restore basal glucose and basal glucose turnover. In general, the failure to reattain basal conditions could result from inappropriate insulin and glucagon doses or possibly from the failure to maintain constant infusion rates (3, 21). On the other hand, it is possible that insulin-glucagon replacement alone is insufficient to recreate the delicate balance of basal concentrations and fluxes. The goal of the present study was to determine whether constant replacement infusions of insulin and glucagon are sufficient to restore basal metabolic conditions during islet suppression in the overnight-fasted dog. To establish an optimal hormone infusion protocol, we tested a variety of insulin-glucagon combinations in separate experiments. However, infusion rates were always held constant during a given experiment. Thus it was possible to assess the feasibility of basal replacement without the need for on-line adjustments in hormone infusion rates (and consequently fluctuations in plasma hormone levels). Our results demonstrate that stable metabolic cond&ions can be achieved during islet suppression with fixed-rate (open-loop) islet hormone replacement.
islet clamp protocol; islet hormone replacement; somatostatin
METHODS
that insulin and glucagon play critical roles in the maintenance of basal glucose homeostasis (7,25,26,35). This is evident during islet suppression by exogenous somatostatin, which elicits an initial decline in plasma glucose due to glucagon lack, followed by a period of ascending glucose, as the effects of insulin lack on glucose production and uptake slowly develop (7). The preceding concept suggests an important physiological question. If, during islet suppression, the basal concentrations of insulin and glucagon were replaced by constant infusions, would basal metabolic conditions be restored? The assumption that they would is generally implicit in hormone replacement studies (2, 10, 11, 33). However, several groups have reported a failure to recreate basal conditions during islet hormone replacement, as judged by the ability to simultaneously restore the various parameters of glucose metabolism (e.g., hormone levels, plasma glucose, glucose turnover) to basal. Jackson et al. (21) were unable to maintain basal glucose production during islet suppression in humans with replacement infusions of insulin and glucagon, despite constant glucose uptake. In canine studies, we found that extremely elevated glucagon infusion rates (-20 ng min-’ +kg-‘) were required to maintain euglycemia during
Animals and surgical procedures. Five dogs were used for a total of 30 “islet clamp” protocols (see below). Dogs were housed under controlled kennel conditions (12:12-h light-dark cycle) in the University of Southern California Medical School Vivarium and were fed once a day with standard Chow (25% protein, 9% fat, and 49% carbohydrate; Wayne Dog Chow; Alfred Mills, Chicago, IL). Dogs were used for experiments only if judged to be in good health, as determined by body temperature, hematocrit, regularity of food intake, and direct observation. All protocols were approved in advance by the Institutional Animal Care and Use Committee. Surgery was performed on each dog at least 1 wk before its first experiment. After an overnight fast, anesthesia was induced with sodium thiamylal (Biotal; Bio-Ceutic Laboratories, St. Joseph, MO) and was maintained with halothane and nitrous oxide. A Tygon catheter was inserted in the right jugular vein through a lateral neck incision, and the catheter tip was advanced to the right atrium. This catheter was used for sampling of central venous blood. The catheter end was led subcutaneously to the back of the neck and exteriorized. The portal vein was accessedvia a midline incision, and a Silastic catheter (0.04 in. ID) was inserted through a pinhole in the vessel wall and sutured in place 3 cm from the porta hepatis. This second catheter was used for infusion of insulin and glucagon. When the peritoneal cavity was closed, the free end of the portal vein catheter was led subcutaneously to the back of the neck and exteriorized. Catheters were filled with heparin (100 U/ml)saline, coiled and capped, and placed in a small bag protected
IT
HAS
BEEN
FIRMLY
ESTABLISHED
l
E532
0193-1849/92 $2.00 Copyright 0 1992 the American Physiological
1
Society
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 14, 2019.
BASAL
HORMONE
with a heavy denim collar. Catheter location was confirmed at autopsy. &perimentaZ protocol. The general design of experiments was as follows. During peripheral somatostatin infusion, insulin was infused intraportally in all experiments at 200 PU kg-‘, a rate previously found to match plasma insulin levels to basal (3). To create a variety of metabolic responses, glucagon was infused intraportally at one of six rates (0, 0.6, 1, 2, 5, or 20 ng* min-l . kg”) but always held constant throughout a given experiment. Blood was sampled during the basal and infusion periods for measurement of glucose, lactate, insulin, and glucagon and for calculation of glucose production and uptake. Experiments were performed on surgically prepared male mongrel dogs (22-36 kg) in the conscious relaxed state. A minimum of 5 days elapsed between experiments. Food was removed at 1500 h the day before experiments, which began at 0900 h. On the day of each experiment, an intracatheter (19-gauge; Deseret Medical, Sandy, UT) was inserted in a cephalic vein for infusion of [ 3-3H]glucose (tracer) and somatostatin. At time -150 min, 28 &i of tracer were injected via the cephalic catheter, and a 0.25 &i/min tracer infusion was begun through the same catheter. This latter infusion continued throughout the experiment (tracer equilibration period 90 min, basal period 60 min, hormone infusion period 240 min). Seven basal blood samples were collected (every 10 min) from -60 min to time -1 min. At time 0, somatostatin (0.8 pg* mine1 . kg-l) was infused via the cephalic catheter, and insulin (200 pUo min. kg-‘) and glucagon (0, 0.6, 1, 2, 5, or 20 ng*min-’ kg-l) were infused via the portal catheter. Central venous blood samples (7 ml) were taken every 10 min during the basal and replacement periods. A 0.6-ml aliquot of this was added to a tube containing 0.9 ml 10.4% perchloric acid for the measurement of blood lactate (see below). The remaining 6.4 ml of blood were deposited in an ice-cold tube containing 350 ~1 Trasylol (aprotinin; FBA Pharmaceuticals, New York, NY), 10 U heparin, and 1 mg/ml sodium fluoride. A 0.5.ml vol of this mixture was immediately centrifuged for on-line measurement of plasma glucose. The remainder was kept on ice until centrifugation, after which supernatant plasma was divided into aliquots into separate microfuge tubes for each of the assays discussed below. Plasma for glucagon assay (0.7 ml) was transferred directly to 2 ml 95% ethanol, dried under vacuum at room temperature, redissolved in phosphate buffer (see Assays), and frozen. All samples were stored at -2OOC. Assays. For measurement of [3-3H]glucose concentration, samples were deproteinized with zinc sulfate and barium hydroxide. Supernatant was evaporated at 7O”C, redissolved in water, and counted in Ready Safe scintillation fluid (Beckman) in a Beckman liquid scintillation counter [intra-assay coefficient of variation (CV) -5%]. Tracer infusates and plasma samples were processed and counted in an identical manner. Lactate concentration in acid extracts of whole blood (see Experimental protocol) was measured with a kit from Sigma (726UV/826UV) after neutralization with 1.7 mM potassium hydroxide (CV -1% postextraction). Glucose was measured (CV -2%) with a glucose autoanalyzer (Yellow Springs Instruments, Yellow Springs, OH). Insulin was measured by radioimmunoassay using the dextran-charcoal separation technique (see Ref. 19, CV -7%). Materials for assay, including canine insulin standard, antibody, and 12’1-labeled insulin, were from Novo-Nordisk. Glucagon was assayed using a kit also obtained from Novo-Nordisk (CV w 11%). This assay uses antiserum K5563 and incorporates a step for ethanolic extraction of plasma. Other materials. Somatostatin was purchased from Bachem (Torrance, CA), porcine insulin was from Sigma (St. Louis, MO), and bovine glucagon was from Calbiochem (San Diego, l
l
min-l
E533
REPLACEMENT
.
CA). Isotopic [3-3H]glucose (high-performance liquid chromatography-determined purity >95%) was from Amersham (Arlington, IL). Data analysis and calculations. Hepatic glucose output (HGO) and glucose uptake were calculated from Steele’s equation (37). Glucose and specific activity data were smoothed with the optimal segments technique (14). “Steady state” was defined for all data as the average value over the last 120 min of the “replacement” (hormone-infusion) period. Although seven basal samples were taken from time -60 to time -1 min, plasma tracer concentration was not always stable by time -60 min; therefore “basal” was defined for all data as the average value at times -20, -10, and -1 min Data are reported as means t SE. Classical statistics (Student’s t test, two-way analysis of variance) were performed using the Statistical Analysis Systems statistical package on a personal computer. Some of the data reported herein are published in a companion paper (6) that is conceptually distinct from the present analysis. RESULTS
Hormones. Constant intraportal insulin replacement at 200 ~U~rnin-l . kg-’ resulted in plasma insulin concentrations that were somewhat lower than basal (7.2t 0.4 vs. 11.2 t 0.9 pU/ml; P < 0.001) but well within the typical fasting range (2, 22). During infusion, plasma insulin levels did not differ among glucagon doses (P = 0.71; Table 1). Basal plasma glucagon averaged 153 t 16 pg/ml overall. During hormone replacement, glucagon concentration was linearly dependent on glucagon dose, varying from 79 t 22 pg/ml at dose 0 ng*min-‘* kg-’ to 1,186 t 209 pg/ml at dose 20 ng= min-’ kg-’ (Fig. 1). At dose 2 ng min-l kg-‘, plasma glucagon was closely matched to basal (steady state = 164 t 18 vs. basal = 182 & 57 pg/ l
l
ml;
l
P = 0.62).
Glucose. Basal plasma glucose averaged 100 t 1 mg/dl (n = 30). During replacement, glucose increased from basal at glucagon doses 5 and 20 nggmin-‘. kg-‘, remained close to basal at dose 2 ngomin-‘. kg-‘, and decreased from basal at doses 0, 0.6, and 1 ng*min” kg-’ (Fig. 2A). Steady-state glucose (Fig. 2C) varied from 73 t 7 mg/dl at dose 0 ng mine1 kg-’ to 198 t 19 mg/dl at dose 20 ng*min-lo kg-‘. Steady-state glucose differed significantly from basal (P < 0.05) at all glucagon doses except 2 ngemin-l . kg-’ (steady state = 92 & 20 vs. basal = 99 t 3 mg/dl; P= 0.74). Glucose kinetics. Basal HGO averaged 2.6 t 0.1 mg* min-’ kg-’ (n = 30). Similar to glucose, HGO increased from basal at doses 5 and 20 ngomin-l* kg-‘, remained close to basal at dose 2 ng min-’ . kg-’ (steady state = 2.4 t 0.2 vs. basal = 2.7 t 0.2 mg* min-‘. kg-‘), and decreased from basal at doses 0, 0.6, and 1 ng*min-lo kg-’ (Fig. 2B). Steady-state HGO (Fig. 20) was different from basal (P < 0.05) at all doses except 0.6 and 2 ngm min-‘0 kg-’ (P = 0.13 and 0.47, respectively). At the highest doses, HGO rose sharply during the first 20-30 min then declined toward steady state over the next 2-3 h. This transient pattern is consistent with a number of studies reporting evanescence of glucagon-stimulated HGO (8, 13), which may involve inhibition by the ascending glucose concentration (4, 27). Whichever the l
l
l
l
l
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 14, 2019.
E534
BASAL
HORMONE
REPLACEMENT
Table 1. Insulin, glucose clearance, and lactate data Glucagon
Infusion
Rate (ng min-’ l
l
kg-l)
Basal 0
0.6
1
2
5
20
Plasma insulin, ll.ZkO.9 7.2&1.0* 8.2t0.6 6.2&1.2* 7.2tl.l 7.2t0.9 7.4kO.5 d/ml Glucose clearance, 2.6kO.l 2.9t0.4 3.1kO.4 3.0t0.2 2.7t0.2 1.9zko.4 1.8kO.l" ml. min” kg” Blood lactate, mM 0.48t0.05 0.49kO.03 0.56kO.15 0.47t0.10 0.46kO.04 0.38kO.06 0.34t0.03 Values are means & SE for basal (n = 30 experiments) or steady-state averages (n = 5). * Difference from corresponding basal (P < 0.05 by two-tailed paired t test; steady-state value was lower than corresponding basal value in all 3 cases where significance was observed). l
z
t -------
k-
-------------s
BASAL
3 f J---cL 0 1 GLUCAGON hg/min
60 IO INFUSION per kg)
B In
RATE
/
Fig. 1. Plasma glucagon dose response. Points represent average (n = 5) steady-state plasma glucagon concentration during hormone replacement at glucagon doses 0.6, 1, 2, 5, and 20 ngomin-‘okg-’ (dose 0 ng. min”. kg-’ excluded to allow logarithmic abscissa). Error bars, 1 SE. Dashed horizontal line, mean (n = 30) basal concentration (153 t 16 Pg/ml)~
case, the apparent dose-response relationship between HGO and glucagon was markedly different, depending on whether HGO was taken as steady state (averaged over last 2 h) or as “extreme” HGO (maximum or minimum). Both relationships are shown in Fig. 20. Basal glucose clearance averaged 2.6 t 0.1 ml. min-’ kg” (n = 30). Steady-state clearance declined with increasing glucagon dose, averaging slightly higher than basal at doses 0, 0.6, and 1 ng* min-lo kg-’ (Table 1) and markedly lower than basal at doses 5 and 20 (P = 0.001 at dose 20 ng.min-l . kg-‘). The decline in clearance observed at the higher doses is consistent with previous studies demonstrating that clearance decreases with increasing glucose (5, 39). At glucagon dose 2 ng* min-’ -l, steady-state clearance (2.7 t 0.2 ml. min-’ kg-‘) was indistinguishable (P = 0.84) from basal clearance (2.8 t 0.2 ml~min-lokg-l). Lactate. Basal blood lactate concentration averaged 0.48 t 0.05 mM (n = 30). During replacement, steadystate lactate levels did not differ among glucagon doses (P = 0.47) and were not different from basal at any dose (P > 0.08; Table 1). Some evidence of overreplacement of glucagon was nevertheless apparent at doses 5 and 20 ng min-’ kg-‘, where lactate increased from time 0 to -20 min and subsequently declined toward basal levels (Fig. 3), probably as a result of a surge in net hepatic lactate output (11). At dose 2 ng. min-’ . kg-‘, blood lactate concentration remained stable throughout the basal and infusion periods. l
l
l
kg
l
l
0 u I
-90
0
90
180
270
0.6
1
2
5
/
30
TIME GLUCAGON DOSE (ng/min per kg) (mid Fig. 2. Plasma glucose and glucose production data. A and B: time courses. q , dose = 0; A, dose = 0.6; 0, dose = 1; n , dose = 2; A, dose = 5; l , dose = 20 (glucagon doses are in ng. mine1 kg-‘). Each time point represents average of 5 animals. For clarity, error bars are not shown. Hormones were infused from time O-240 min. C and D: dose-response profiles. Each point represents 1 dose and is average of 5 animals. Error bars, 1 SE. Horizontal lines, mean (n = 30) basal values. Values at 0 glucagon replacement are excluded to allow logarithmic abscissa. D: open symbols, extreme (maximum or minimum) values; closed symbols, steady-state values. HGO, hepatic glucose output. l
Summary of dose-response data. Steady-state glucagon, glucose, and HGO data, as well as the extreme HGO data, are plotted in Fig. 4 vs. glucagon dose (data are expressed as percent mean basal). All four curves intersect at ~2 ng*min-l *kg-’ glucagon infusion. At higher doses, the curves diverge and are ordered (highest to lowest) steady-state glucagon, extreme HGO, steadystate glucose, and steady-state HGO, whereas at lower doses their order is reversed. None of the four variables differed significantly from 100% (P > 0.41)at dose 2 ng min-‘* kg-’ (107 2 11, 92 t 16, 92 t 20, and 93 t 8% mean basal for glucagon, extreme HGO, glucose, and steady-state HGO, respectively). Stability of metabolic variables. To assess metabolic stability during hormone replacement, CVs were calculated for the last hour of the replacement periods (180l
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 14, 2019.
BASAL
HORMONE
E535
REPLACEMENT
amsad 20
-
-90
0 TIME
90
180
270
-90
0
90
TIME
(mid
180
270
-90
wmin
0
90
TIME
(mid
Fig. 3. Blood lactate time courses at glucagon doses of 2 (left), 5 (midcUe), and 20 (right) ngemin-’ were infused from time O-240 min. Error bars, 1 SE. * Difference from basal (P c 0.05).
per
kg
180
I
270
(mid
kg? Hormones
l
identical results were obtained when basal and steadystate were defined as the last three points (last 20 min) of the period in question (data not shown). Thus metabolic variables were not destabilized by constant islet hormone replacement. DISCUSSION
GLUCAGON INFUSION (ng/min per kg)
RATE
Fig. 4. Steady-state plasma glucagon (A), extreme HGO (0), steadystate plasma glucose (A), and steady-state HGO (0) vs. glucagon replacement rate. Each point represents 1 dose and is average of 5 animals (values at dose 0 nge min-’ . kg-’ are excluded to allow logarithmic abscissa). For clarity, error bars are not shown.
240 min) and compared with CVs during fasting (-60 to -1 min). CVs for plasma glucose, HGO, glucose clearance, and blood lactate data are shown in Table 2 for all glucagon doses. Last-hour CVs varied from 0.5 to 9.6,0.3 to 14.6, 0.7 to 17.6, and 0.4 to 13.4% for glucose, HGO, clearance, and lactate, respectively. In no case was the CV during the last hour of replacement significantly greater than that during the basal period (P > 0.32). In fact, with the exception of the glucose data, CVs during the last hour of replacement were generally less than those during the corresponding basal periods. Nearly Table 2. Coefficients of variation for glucose,
Despite widespread use of somatostatin as an experimental tool (17,20,34,36,41), the feasibility of restoring basal conditions with constant insulin and glucagon has not been rigorously assessed. This is because virtually all experiments incorporating a period of basal replacement have utilized on-line adjustment of insulin (2,10,11,33) or glucagon (3, 21) to maintain euglycemia. The effects of fluctuations in islet hormone levels remain controversial (1529-31). Abrupt changes in plasma glucagon have been shown to alter hepatic sensitivity to glucagon (12, 15), an effect that could significantly influence results from protocols using variable glucagon infusion. On the other hand, variable insulin designs suffer from a potentially more important drawback. The effect of insulin to suppress HGO, like its effect on glucose uptake, requires several hours to develop (1, 42). Thus, although insulin is usually fixed before commencement of the test period (2, 1 l), it is not clear how much time must elapse after a period of insulin adjustment before the insulin effect may be considered constant.
lactate, hepatic glucose output, and glucose clearance Glucagon
Basal 0
0.6
Infusion 1
rate (ng min-’ l
2
l
kg-‘) 5
20
Plasma glucose 2.3k0.2 3.4kl.O 2.9k0.6 4.7k1.4 3.4k1.4 2.OkO.5 2.3k0.6 Hepatic glucose output 7.5t0.7 4.4t2.5 3.9k1.8 4.1k2.6 4.6k1.6* 3.9kl.l 5.Ok1.3 4.3kl.O Glucose clearance 7.4kO.8 5.0k1.2 3.9t0.7 3.6kO.9 4.9t1.5" 8.2k2.9 6.8k0.4 Blood lactate 14.7k2.0 10.9kl.l 7.4k1.2" 8.5k0.7 4.7k1.6 6.4k1.3 Values are means ,t SE in % for averages of coefficients of variation (CVs) from individual experiments, each calculated from variance among individual samples (7 samples included in each CV calculation); n = 30 (basal) and 5 (steady state). * Significant difference from corresponding basal (P > 0.05 by two-tailed paired t test; last hour value was lower than corresponding basal value in all 3 cases where significance was observed). Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 14, 2019.
E536
BASAL
HORMONE
REPLACEMENT
With these considerations in mind, we chose to test the present hypothesis (that replacement infusions of insulin and glucagon are sufficient to restore basal conditions during islet suppression) in the absence of any on-line adjustments in hormone infusion rates. Instead, we infused insulin at a basal rate in all experiments and set different glucagon infusion rates in separate experiments in an effort to define an optimal insulin-to-glucagon infusion ratio. Our data indicate that, during basal insulin infusion, glucagon infusion at rates 2 nge min-’ kg-’ result in hyperglycemia, subbasal glucose clearance, and surges in glucose production and blood lactate levels. However, at 2 ng min-’ . kg-’ glucagon infusion, plasma glucose, glucose production, glucose clearance, and blood lactate remain closely matched to their basal values. Moreover, metabolic variables were not discernably less stable and were in fact frequently more stable during hormone replacement than during the basal (fasting) periods. These results demonstrate that, on average, the principal concentrations and fluxes of glucose metabolism can be simultaneously restored to their basal levels during islet suppression with constant intraportal insulin and glucagon replacement. It should be understood, however, that individual animals could vary substantially in their response to exogenous hormone administration (6). In addition, the replacement doses defined as optimal under specific experimental conditions may not be generally applicable. It has been controversial not only how to replace islet hormones during somatostatin infusion, but whether indeed such replacement can reestablish basal metabolic conditions (3, 21). Cherrington and associates (2, 10, 11, 33) have employed a replacement protocol using constant glucagon and variable insulin infusion to maintain euglycemia. With the use of this design, they have reported constancy of plasma glucose as well as glucose turnover. However, there is reason to examine this approach more carefully. The assumption underlying the variable-insulin protocol is that the effects of insulin on metabolic parameters occur rapidly and will remain constant. At the time the protocol was first used, these assumptions appeared justifiable because 1) fasting glucose uptake is largely insulin independent (5, 39), such that insulin adjustments modulate plasma glucose primarily via HGO, and 2) HGO suppression by insulin was thought to be rapid. However, recent experiments using improved methods of HGO assessment have established that the effects of insulin on glucose production under normal conditions develop slowly and are not fully manifest for several hours (42). A potential explanation for this delay may be surmised from recent studies by us (1) and others (32), which indicate that hepatic insulin action may be mediated via a peripheral mechanism (23) for which transcapillary insulin transport is rate limiting (42). Therefore, during insulin adjustment, one might expect that it would be difficult to pinpoint the requisite insulin infusion rate unless prolonged periods of observation were possible. Thus, in the present study, it was decided to maintain constant insulin infusion at all times. One alternative to insulin adjustment is to vary the l
l
l
glucagon infusion rate while holding insulin constant. Jackson et al. (21) used such a protocol to maintain euglycemia during islet suppression in humans. They reported considerable instability in plasma glucose following a period of glucagon adjustment. We used a similar approach in dogs and found that an unacceptably high glucagon infusion rate was required to restore basal glucose (3), possibly due to waning of the glucagon effect (8, 13). Thus, while glucagon action is rapid (16), its well-documented evanescence seems to limit the usefulness of glucagon adjustments for the maintenance of euglycemia. The present studies, in which both islet hormones were administered at fixed rates, support the contention of Cherrington and associates that a constant metabolic state may be reacheived during islet suppression by replacing basal plasma insulin and glucagon levels. However, one inconsistency between our results and results previously reported (2, 7) is the glucagon infusion rate needed for basal reestablishment. Thus the Vanderbilt group has typically used 0.6-l ng min-’ . kg-’ intraportal glucagon (2, 7), whereas we required -2 ng* min-lo kg-l in the present study. The reason for this discrepancy is not clear, although it should be noted that the Vanderbilt group has not reported a complete glucagon dose response at constant insulin. Nevertheless, results from both laboratories are consistent, in the sense that metabolic parameters are restored to basal concurrently with a matching of plasma immunoreactive glucagon to basal. This suggests the possibility that our animals regulate their blood glucose at higher glucagon levels, perhaps as a result of differences in diet, exercise, or other variables. It is noteworthy that fasting metabolic conditions were closely recreated at the 2 ng min-’ . kg-’ glucagon dose, despite some (-30%) underreplacement of plasma insulin. Based on immunoreactive glucagon measurements (Fig. l), it seems unlikely that the insulin underreplacement was compensated by a decrease in plasma glucagon. Alternatively, it is possible that the expected consequences of insulin underreplacement (increased HGO and decreased glucose clearance) were masked by extrapancreatic effects of somatostatin. O’Brien et al. (28) have shown increased hepatic insulin extraction and decreased HGO in response to somatostatin; such an effect would tend to maintain basal HGO levels despite hypoinsulinemia. Also, a potential decrease in glucose clearance may have been averted by somatostatin, which is known to augment glucose clearance in some cases (3). It is conceivable, therefore, that the (unintentional) underreplacement of insulin in the present experiments was fortuitously compensated by extrapancreatic somatostatin effects. In studies where potential extrapancreatic somatostatin effects could be confounding, the depancreatized dog model (40) may be preferable to somatostatin-based protocols. In summary, the present data demonstrate that stable metabolic conditions closely resembling the fasting state can be recreated during islet suppression by means of constant intraportal replacement infusions of insulin and glucagon (200 PU. min-l kg-’ and 2 ng* min-’ . kg-‘, respectively). This result supports the validity of islet clamp protocols and suggests that the primary effect of l
l
l
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 14, 2019.
BASAL
HORMONE
somatostatin on fasting glucose metabolism is to reduce the plasma concentrations of insulin and glucagon. We thank Donna Banks, Ginger Hayes, Paul Kirkman, and Dr. Linda Kirkman for surgical and experimental assistance and Elsa Demirchyan and Lena Roumian for technical assistance. This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-27619. D. C. Bradley is a predoctoral trainee supported by National Institute on Aging Trainee Grant T32-AG-00093. Address for reprint requests: R. N. Bergman, Dept. of Physiology and Biophysics, Univ. of Southern California Medical School, 2025 Zonal Ave., 128 Mudd Bldg., Los Angeles, CA 90033. Received 5 July 1991; accepted in final form 4 November 1991. REFERENCES 1. Ader, M., and R. N. Bergman. Peripheral effects of insulin dominate suppression of fasting hepatic glucose production. Am. J. Physiol. 258 (Endocrinol. Metab. 21): E1020-E1032,1990. 2. Adkins-Marshall, B. A., S. R. Myers, G. K. Hendrick, P. E. Williams, K. Triebwasser, B. Floyd, and A. D. Cherrington. Interaction between insulin and glucose-delivery route in regulation of net hepatic glucose uptake in conscious dogs. Diabetes 39: 87-95,199O. 3. Bergman, R. N., M. Ader, D. T. Finegood, and G. Pacini. Extrapancreatic effect of somatostatin infusion to increase glucose clearance. Am. J. Physiol. 247 (Endocrinol. Metab. 10): E370-E379, 1984. 4. Bergman, R. N., and R. J. Bucolo. Interaction of insulin and glucose in the control of hepatic glucose balance. Am. J. Physiol. 227: 1314-1322,1974. 5. Best, J. D., G. J. Taborsky, J. B. Halter, and D. Porte, Jr. Glucose disposal is not proportional to plasma glucose level in man. Diabetes 30: 847-850, 1981. 6. Bradley, D. C., and R. N. Bergman. Hepatic glucagon sensitivity and fasting glucose concentration in normal dogs. Am. J. Physiol. 262 (Endocrinol. Metab. 25): E539-E545,1992 7. Cherrington, A. D., J. L. Chiasson, J. E. Liljenquist, A. S. Jennings, U. Keller, and W. W. Lacy. The role of insulin and glucagon in the regulation of basal glucose production in the postabsorptive dog. J. Clin. Invest. 58: 1407-1418, 1976. 8. Cherrington, A. D., M. P. Diamond, D. R. Green, and P. E. Williams. Evidence for an intrahepatic contribution to the waning effect of glucagon on glucose production in the conscious dog. Diabetes 31: 917-922, 1982. 9. Cherrington, A. D., W. W. Lacy, and J. L. Chiasson. Effect of glucagon on glucose production during insulin deficiency in the dog. J. Clin. Invest. 62: 664-677, 1978. 10. Cherrington, A. D., W. W. Lacy, P. E. Williams, and K. E. Steiner. Failure of somatostatin to modify effect of glucagon on carbohydrate metabolism in the dog. Am. J. Physiol. 244 (Endocrinol. Metab. 7): E596-E602,1983. 11. Davis, M. A., P. E. Williams, and A. D. Cherrington. Effect of glucagon on hepatic lactate metabolism in the conscious dog. Am. J. Physiol. 248 (Endocrinol. Metab. 11): E463-E470,1985. 12. Deri, J. J., P. E. Williams, K. E. Steiner, and A. D. Cherrington. Altered ability of the liver to produce glucose following a period of glucagon deficiency. Diabetes 30: 490-495, 1981. 13. El Refai, M., and R. N. Bergman. Glucagon-stimulated glycogenolysis: time-dependent sensitivity to insulin. Am. J. Physiol. 236 (Endocrinol. Metab. Gastrointest. Physiol. 5): E246-E254,1979. 14. Finegood, D. T., and R. N. Bergman. Optimal segments: a method for smoothing tracer data to calculate metabolic fluxes. Am. J. Physiol. 244 (Endocrinol. Metab. 7): E472-E479,1983. 15. Fradkin, J., H. Shamoon, P. Felig, and R. S. Sherwin. Evidence for an important role of changes in rather than absolute concentrations of glucagon in the regulation of glucose production in humans. J. Clin. Endocrinol. Metab. 50: 698-703, 1980. 16. Gerich, J., P. Cryer, and R. Rizza. Hormonal mechanisms in acute glucose counterregulation: the relative roles of glucagon, epinephrine, norepinephrine, growth hormone, and cortisol. Met& Clin. Exp. 29: 1164-1175, 1980. 17. Hansen, I., R. Firth, M. Haymond, P. Cryer, and R. Rizza.
REPLACEMENT
E537
The role of autoregulation of the hepatic glucose production in man: response to a physiologic decrement in plasma glucose. Diabetes 35: 186-191, 1986. G. K., R. T. Frizzell, and A. D. Cherrington. 18. Hendrick, Effect of somatostatin on nonesterified fatty acid levels modifies glucose homeostasis during fasting. Am. J. Physiol. 253 (Endocrinol. Metab. 16): E443-E452,1987. 19. Herbert, V., K. Law, C. Gottlieb, and S. Bleicher. Coated charcoal immunoassay of insulin. J. Clin. Endocrinol. Metab. 25: 1375-1384,1965. 20. Hoelzer, D. R., G. P. Dalsky, W. E. Clutter, S. D. Shah, J. 0. Holloszy, and P. E. Cryer. Glucoregulation during exercise: hypoglycemia is prevented by redundant glucoregulatory systems, sympathochromaffin activation, and changes in islet hormone secretion. J. Clin. Invest. 77: 212-221, 1986. 21. Jackson, R. A., J. B. Hamling, B. M. Sim, P. M. Blix, J. B. Jaspan, S. Markanday, and J. D. N. Nabarro. Basal glucose homeostasis with and without fixed concentrations of glucagon and insulin. Metab. Clin. Exp. 36: 131-136, 1987. 22. Lavelle-Jones, M., M. H. Scott, 0. Kolterman, A. H. Rubenstein, J. M. Olefsky, and A. R. Moossa. Selective suppression of hepatic glucose output by human proinsulin in the dog. Am. J. Physiol. 252 (Endocrinol. Metab. 15): E230-E236, 1987. 23. Levine, R., and I. B. Fritz. The relation of insulin to liver metabolism. Diabetes 5: 209-219, 1956. 24. Li, C. H. Human growth hormone: 1974-1981. Mol. Cell. Biochem. 46: 31-41,1982. H. L. A., G. G. Ross, and M. Vranic. Effects of 25. Lickley, selective insulin or glucagon deficiency on glucose turnover. Am. J. Physiol. 236 (Endocrinol. Metab. Gastrointest. Physiol. 5): E255E262,1979. J. E., G. L. Mueller, A. D. Cherrington, U. 26. Liljenquist, Keller, J. L. Chiasson, J. M. Perry, W. W. Lacy, and D. Rabinowitz. Evidence for an important role of glucagon in the regulation of hepatic glucose production in normal man. J. Clin. Invest. 59: 369-374, 1977. J. E., G. L. Mueller, A. D. Cherrington, J. M. 27. Liljenquist, Perry, and D. Rabinowitz. Hyperglycemia per se (insulin and glucagon withdrawn) can inhibit hepatic glucose production in man. J. Clin. Endocrinol. Metab. 48: 171-175, 1979. D. W., R. V. Rajotte, G. D. Molnar, K. Toth, and 28. O’Brien, T. D. Petersen. Somatostatin increases hepatic insulin extraction. Diubetes Res. Clin. Pratt. 7: 171-178, 1988. 29. Paolisso, G., A. J. Scheen, A. S. Luyckx, and P. J. Lefebvre. Pulsatile hyperglucagonemia fails to increase hepatic glucose production in normal man. Am. J. Physiol. 252 (Endocrinol. Metab. 15): El-E7,1987. 30. Paolisso, G., A. J. Scheen, E. M. Verdin, A. S. Luyckx, and P. J. Lefebvre. Insulin oscillations per se do not affect glucose turnover parameters in normal man. J. Clin. Endocrinol. Metab. 63: 520-525,1986. A. Scheen, F. 31. Paolisso, G., S. Sgambato, N. Passariello, D’Onofrio, and P. J. Lefebvre. Greater efficacy of pulsatile insulin in type I diabetics critically depends on plasma glucagon levels. Diabetes 36: 566-570, 1987. 32. Prager, R., P. Wallace, and J. M. Olefsky. Direct and indirect effects of insulin to inhibit hepatic glucose output in obese subjects. Diabetes 36: 607-611, 1987. 33. Radosevich, P. M., P. E. Williams, J. R. McRae, W. W. Lacy, D. N. Orth, and N. N. Abumrad. &Endorphin inhibits glucose production in the conscious dog. J. Clin. Invest. 73: 12371241,1984. 34. Sherwin, R. S., W. Tamborlane, R. Hendler, L. Sacca, R. A. DeFronzo, and P. Felig. Influence of glucagon replacement on the hyperglycemic and hyperketonemic response to prolonged somatostatin infusion in normal man. J. Clin. Endocrinol. Metab. 45: 1104-1107,1977. 35. Shulman, G. I., J. E. Liljenquist, P. E. Williams, W. W. Lacy, and A. D. Cherrington. Glucose disposal during insulinopenia in somatostatin-treated dogs: the roles of glucose and glucagon. J. Clin. Invest. 62: 487-491, 1978. 36. Starke, A., T. Imamura, and R. H. Unger. Relationship of glucagon suppression by insulin and somatostatin to the ambient glucose concentration. J. Clin. Invest. 79: 20-24, 1987. 37. Steele, R., J. S. Wall, R. C. deBodo, and N. Altszuler.
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 14, 2019.
E538
BASAL
HORMONE
Measurement of size and turnover rate of body glucose pool by the isotope dilution method. Am. J. Physiol. 187: 15-24, 1956. 38. Vasquez, B., M. Nagulesparan, and R. H. Unger. Fasting and postprandial plasma glucagon and glucagon-like immunoreactivity are normal in obese Pima Indians. Horm. Metab. Res. 15: 218-220, 1983. 39. Verdonk, C. A., R. A. R&a, and J. E. Gerich. Effects of plasma glucose concentration on glucose utilization and glucose clearance in normal man. Diabetes 30: 535-537,198l.
REPLACEMENT
40. Wasserman, D. H., H. L. A. Lickley, and M. Vranic. Important role of glucagon during exercise in diabetic dogs. J. A&. Physiol. 59: 1272-1281, 1985. 41. Wolfe, R. R., E. R. Nadel, J. H. F. Shaw, L. A. Stephenson, and M. H. Wolfe. Role of changes in insulin and glucagon in glucose homeostasis in exercise. J. Clin. Invest. 77: 900-907, 1986. 42. Yang, Y. J., I. D. Hope, M. Ader, and R. N. Bergman. Insulin transport across capillaries is rate limiting for insulin action in dogs. J. Clin. Invest. 84: 1620-1628, 1989.
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 14, 2019.
C. BRADLEY
AND
RICHARD
Department of Physiology and Biophysics, Los Angeles, California 90033
N. BERGMAN University
Bradley, David C., and Richard N. Bergman. Restoration of stable metabolic conditions during islet suppression in dogs. Am. J. Physiol. 262 (Endocrinol. Metab. 25): E532-E538, 1992.-These studies were undertaken to examine the stability of metabolic conditions during islet suppression with fixed-rate insulin and glucagon replacement. Somatostatin was infused peripherally at 0.8 pg* min-l . kg-l, insulin was infused intraportally at 200 cumin-’ kg-‘, and glucagon was infused intraportally at 0, 0.6, 1, 2, 5, or 20 nggmin-‘. kg-l in conscious overnight-fasted dogs. [ 3-3H]glucose was infused for measurement of glucose kinetics. During infusion, plasma insulin was 7.2 & 0.4 pU/ml. Plasma glucagon rose linearly with glucagon dose, achieving basal levels at 2 ng* min-l . kg-l infusion (164 t 18 vs. basal = 182 =f:57 pg/ml). Plasma glucose and hepatic glucose output (HGO) decreased from basal at doses 0,0.6, and 1 nge min-l . kg-‘, increased from basal at doses 5 and 20 ng* min-l* kg-l, and remained close to basal at dose 2 ng*min-l . kg-’ (92 * 20 vs. basal = 99 $- 3 mg/dl and 2.4 & 0.2 vs. basal = 2.7 * 0.2 mg*min-l gkg-’ for glucose and HGO, respectively; P > 0.47). Glucose clearance and blood lactate were also closely matched to basal at dose 2 ng.min-l kg? Coefficients of variation during 2 ng.min-’ kg-’ glucagon infusion (last hour) were 3.4, 4.6, 4.9, and 4.7% for glucose, HGO, clearance, and lactate, respectively. These findings indicate that fasting metabolic conditions, as inferred from blood glucose, lactate, insulin, and glucagon levels, and the rates of glucose production and uptake can be recreated in toto during fixed-rate islet l
l
l
hormone replacement.
of Southern
California
Medical School,
islet suppression with basal insulin replacement, presumably to compensate for a somatostatin-induced increase in glucose clearance (3). Thus, in our experiments, it was not possible to simultaneously restore basal glucose and basal glucose turnover. In general, the failure to reattain basal conditions could result from inappropriate insulin and glucagon doses or possibly from the failure to maintain constant infusion rates (3, 21). On the other hand, it is possible that insulin-glucagon replacement alone is insufficient to recreate the delicate balance of basal concentrations and fluxes. The goal of the present study was to determine whether constant replacement infusions of insulin and glucagon are sufficient to restore basal metabolic conditions during islet suppression in the overnight-fasted dog. To establish an optimal hormone infusion protocol, we tested a variety of insulin-glucagon combinations in separate experiments. However, infusion rates were always held constant during a given experiment. Thus it was possible to assess the feasibility of basal replacement without the need for on-line adjustments in hormone infusion rates (and consequently fluctuations in plasma hormone levels). Our results demonstrate that stable metabolic cond&ions can be achieved during islet suppression with fixed-rate (open-loop) islet hormone replacement.
islet clamp protocol; islet hormone replacement; somatostatin
METHODS
that insulin and glucagon play critical roles in the maintenance of basal glucose homeostasis (7,25,26,35). This is evident during islet suppression by exogenous somatostatin, which elicits an initial decline in plasma glucose due to glucagon lack, followed by a period of ascending glucose, as the effects of insulin lack on glucose production and uptake slowly develop (7). The preceding concept suggests an important physiological question. If, during islet suppression, the basal concentrations of insulin and glucagon were replaced by constant infusions, would basal metabolic conditions be restored? The assumption that they would is generally implicit in hormone replacement studies (2, 10, 11, 33). However, several groups have reported a failure to recreate basal conditions during islet hormone replacement, as judged by the ability to simultaneously restore the various parameters of glucose metabolism (e.g., hormone levels, plasma glucose, glucose turnover) to basal. Jackson et al. (21) were unable to maintain basal glucose production during islet suppression in humans with replacement infusions of insulin and glucagon, despite constant glucose uptake. In canine studies, we found that extremely elevated glucagon infusion rates (-20 ng min-’ +kg-‘) were required to maintain euglycemia during
Animals and surgical procedures. Five dogs were used for a total of 30 “islet clamp” protocols (see below). Dogs were housed under controlled kennel conditions (12:12-h light-dark cycle) in the University of Southern California Medical School Vivarium and were fed once a day with standard Chow (25% protein, 9% fat, and 49% carbohydrate; Wayne Dog Chow; Alfred Mills, Chicago, IL). Dogs were used for experiments only if judged to be in good health, as determined by body temperature, hematocrit, regularity of food intake, and direct observation. All protocols were approved in advance by the Institutional Animal Care and Use Committee. Surgery was performed on each dog at least 1 wk before its first experiment. After an overnight fast, anesthesia was induced with sodium thiamylal (Biotal; Bio-Ceutic Laboratories, St. Joseph, MO) and was maintained with halothane and nitrous oxide. A Tygon catheter was inserted in the right jugular vein through a lateral neck incision, and the catheter tip was advanced to the right atrium. This catheter was used for sampling of central venous blood. The catheter end was led subcutaneously to the back of the neck and exteriorized. The portal vein was accessedvia a midline incision, and a Silastic catheter (0.04 in. ID) was inserted through a pinhole in the vessel wall and sutured in place 3 cm from the porta hepatis. This second catheter was used for infusion of insulin and glucagon. When the peritoneal cavity was closed, the free end of the portal vein catheter was led subcutaneously to the back of the neck and exteriorized. Catheters were filled with heparin (100 U/ml)saline, coiled and capped, and placed in a small bag protected
IT
HAS
BEEN
FIRMLY
ESTABLISHED
l
E532
0193-1849/92 $2.00 Copyright 0 1992 the American Physiological
1
Society
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 14, 2019.
BASAL
HORMONE
with a heavy denim collar. Catheter location was confirmed at autopsy. &perimentaZ protocol. The general design of experiments was as follows. During peripheral somatostatin infusion, insulin was infused intraportally in all experiments at 200 PU kg-‘, a rate previously found to match plasma insulin levels to basal (3). To create a variety of metabolic responses, glucagon was infused intraportally at one of six rates (0, 0.6, 1, 2, 5, or 20 ng* min-l . kg”) but always held constant throughout a given experiment. Blood was sampled during the basal and infusion periods for measurement of glucose, lactate, insulin, and glucagon and for calculation of glucose production and uptake. Experiments were performed on surgically prepared male mongrel dogs (22-36 kg) in the conscious relaxed state. A minimum of 5 days elapsed between experiments. Food was removed at 1500 h the day before experiments, which began at 0900 h. On the day of each experiment, an intracatheter (19-gauge; Deseret Medical, Sandy, UT) was inserted in a cephalic vein for infusion of [ 3-3H]glucose (tracer) and somatostatin. At time -150 min, 28 &i of tracer were injected via the cephalic catheter, and a 0.25 &i/min tracer infusion was begun through the same catheter. This latter infusion continued throughout the experiment (tracer equilibration period 90 min, basal period 60 min, hormone infusion period 240 min). Seven basal blood samples were collected (every 10 min) from -60 min to time -1 min. At time 0, somatostatin (0.8 pg* mine1 . kg-l) was infused via the cephalic catheter, and insulin (200 pUo min. kg-‘) and glucagon (0, 0.6, 1, 2, 5, or 20 ng*min-’ kg-l) were infused via the portal catheter. Central venous blood samples (7 ml) were taken every 10 min during the basal and replacement periods. A 0.6-ml aliquot of this was added to a tube containing 0.9 ml 10.4% perchloric acid for the measurement of blood lactate (see below). The remaining 6.4 ml of blood were deposited in an ice-cold tube containing 350 ~1 Trasylol (aprotinin; FBA Pharmaceuticals, New York, NY), 10 U heparin, and 1 mg/ml sodium fluoride. A 0.5.ml vol of this mixture was immediately centrifuged for on-line measurement of plasma glucose. The remainder was kept on ice until centrifugation, after which supernatant plasma was divided into aliquots into separate microfuge tubes for each of the assays discussed below. Plasma for glucagon assay (0.7 ml) was transferred directly to 2 ml 95% ethanol, dried under vacuum at room temperature, redissolved in phosphate buffer (see Assays), and frozen. All samples were stored at -2OOC. Assays. For measurement of [3-3H]glucose concentration, samples were deproteinized with zinc sulfate and barium hydroxide. Supernatant was evaporated at 7O”C, redissolved in water, and counted in Ready Safe scintillation fluid (Beckman) in a Beckman liquid scintillation counter [intra-assay coefficient of variation (CV) -5%]. Tracer infusates and plasma samples were processed and counted in an identical manner. Lactate concentration in acid extracts of whole blood (see Experimental protocol) was measured with a kit from Sigma (726UV/826UV) after neutralization with 1.7 mM potassium hydroxide (CV -1% postextraction). Glucose was measured (CV -2%) with a glucose autoanalyzer (Yellow Springs Instruments, Yellow Springs, OH). Insulin was measured by radioimmunoassay using the dextran-charcoal separation technique (see Ref. 19, CV -7%). Materials for assay, including canine insulin standard, antibody, and 12’1-labeled insulin, were from Novo-Nordisk. Glucagon was assayed using a kit also obtained from Novo-Nordisk (CV w 11%). This assay uses antiserum K5563 and incorporates a step for ethanolic extraction of plasma. Other materials. Somatostatin was purchased from Bachem (Torrance, CA), porcine insulin was from Sigma (St. Louis, MO), and bovine glucagon was from Calbiochem (San Diego, l
l
min-l
E533
REPLACEMENT
.
CA). Isotopic [3-3H]glucose (high-performance liquid chromatography-determined purity >95%) was from Amersham (Arlington, IL). Data analysis and calculations. Hepatic glucose output (HGO) and glucose uptake were calculated from Steele’s equation (37). Glucose and specific activity data were smoothed with the optimal segments technique (14). “Steady state” was defined for all data as the average value over the last 120 min of the “replacement” (hormone-infusion) period. Although seven basal samples were taken from time -60 to time -1 min, plasma tracer concentration was not always stable by time -60 min; therefore “basal” was defined for all data as the average value at times -20, -10, and -1 min Data are reported as means t SE. Classical statistics (Student’s t test, two-way analysis of variance) were performed using the Statistical Analysis Systems statistical package on a personal computer. Some of the data reported herein are published in a companion paper (6) that is conceptually distinct from the present analysis. RESULTS
Hormones. Constant intraportal insulin replacement at 200 ~U~rnin-l . kg-’ resulted in plasma insulin concentrations that were somewhat lower than basal (7.2t 0.4 vs. 11.2 t 0.9 pU/ml; P < 0.001) but well within the typical fasting range (2, 22). During infusion, plasma insulin levels did not differ among glucagon doses (P = 0.71; Table 1). Basal plasma glucagon averaged 153 t 16 pg/ml overall. During hormone replacement, glucagon concentration was linearly dependent on glucagon dose, varying from 79 t 22 pg/ml at dose 0 ng*min-‘* kg-’ to 1,186 t 209 pg/ml at dose 20 ng= min-’ kg-’ (Fig. 1). At dose 2 ng min-l kg-‘, plasma glucagon was closely matched to basal (steady state = 164 t 18 vs. basal = 182 & 57 pg/ l
l
ml;
l
P = 0.62).
Glucose. Basal plasma glucose averaged 100 t 1 mg/dl (n = 30). During replacement, glucose increased from basal at glucagon doses 5 and 20 nggmin-‘. kg-‘, remained close to basal at dose 2 ngomin-‘. kg-‘, and decreased from basal at doses 0, 0.6, and 1 ng*min” kg-’ (Fig. 2A). Steady-state glucose (Fig. 2C) varied from 73 t 7 mg/dl at dose 0 ng mine1 kg-’ to 198 t 19 mg/dl at dose 20 ng*min-lo kg-‘. Steady-state glucose differed significantly from basal (P < 0.05) at all glucagon doses except 2 ngemin-l . kg-’ (steady state = 92 & 20 vs. basal = 99 t 3 mg/dl; P= 0.74). Glucose kinetics. Basal HGO averaged 2.6 t 0.1 mg* min-’ kg-’ (n = 30). Similar to glucose, HGO increased from basal at doses 5 and 20 ngomin-l* kg-‘, remained close to basal at dose 2 ng min-’ . kg-’ (steady state = 2.4 t 0.2 vs. basal = 2.7 t 0.2 mg* min-‘. kg-‘), and decreased from basal at doses 0, 0.6, and 1 ng*min-lo kg-’ (Fig. 2B). Steady-state HGO (Fig. 20) was different from basal (P < 0.05) at all doses except 0.6 and 2 ngm min-‘0 kg-’ (P = 0.13 and 0.47, respectively). At the highest doses, HGO rose sharply during the first 20-30 min then declined toward steady state over the next 2-3 h. This transient pattern is consistent with a number of studies reporting evanescence of glucagon-stimulated HGO (8, 13), which may involve inhibition by the ascending glucose concentration (4, 27). Whichever the l
l
l
l
l
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 14, 2019.
E534
BASAL
HORMONE
REPLACEMENT
Table 1. Insulin, glucose clearance, and lactate data Glucagon
Infusion
Rate (ng min-’ l
l
kg-l)
Basal 0
0.6
1
2
5
20
Plasma insulin, ll.ZkO.9 7.2&1.0* 8.2t0.6 6.2&1.2* 7.2tl.l 7.2t0.9 7.4kO.5 d/ml Glucose clearance, 2.6kO.l 2.9t0.4 3.1kO.4 3.0t0.2 2.7t0.2 1.9zko.4 1.8kO.l" ml. min” kg” Blood lactate, mM 0.48t0.05 0.49kO.03 0.56kO.15 0.47t0.10 0.46kO.04 0.38kO.06 0.34t0.03 Values are means & SE for basal (n = 30 experiments) or steady-state averages (n = 5). * Difference from corresponding basal (P < 0.05 by two-tailed paired t test; steady-state value was lower than corresponding basal value in all 3 cases where significance was observed). l
z
t -------
k-
-------------s
BASAL
3 f J---cL 0 1 GLUCAGON hg/min
60 IO INFUSION per kg)
B In
RATE
/
Fig. 1. Plasma glucagon dose response. Points represent average (n = 5) steady-state plasma glucagon concentration during hormone replacement at glucagon doses 0.6, 1, 2, 5, and 20 ngomin-‘okg-’ (dose 0 ng. min”. kg-’ excluded to allow logarithmic abscissa). Error bars, 1 SE. Dashed horizontal line, mean (n = 30) basal concentration (153 t 16 Pg/ml)~
case, the apparent dose-response relationship between HGO and glucagon was markedly different, depending on whether HGO was taken as steady state (averaged over last 2 h) or as “extreme” HGO (maximum or minimum). Both relationships are shown in Fig. 20. Basal glucose clearance averaged 2.6 t 0.1 ml. min-’ kg” (n = 30). Steady-state clearance declined with increasing glucagon dose, averaging slightly higher than basal at doses 0, 0.6, and 1 ng* min-lo kg-’ (Table 1) and markedly lower than basal at doses 5 and 20 (P = 0.001 at dose 20 ng.min-l . kg-‘). The decline in clearance observed at the higher doses is consistent with previous studies demonstrating that clearance decreases with increasing glucose (5, 39). At glucagon dose 2 ng* min-’ -l, steady-state clearance (2.7 t 0.2 ml. min-’ kg-‘) was indistinguishable (P = 0.84) from basal clearance (2.8 t 0.2 ml~min-lokg-l). Lactate. Basal blood lactate concentration averaged 0.48 t 0.05 mM (n = 30). During replacement, steadystate lactate levels did not differ among glucagon doses (P = 0.47) and were not different from basal at any dose (P > 0.08; Table 1). Some evidence of overreplacement of glucagon was nevertheless apparent at doses 5 and 20 ng min-’ kg-‘, where lactate increased from time 0 to -20 min and subsequently declined toward basal levels (Fig. 3), probably as a result of a surge in net hepatic lactate output (11). At dose 2 ng. min-’ . kg-‘, blood lactate concentration remained stable throughout the basal and infusion periods. l
l
l
kg
l
l
0 u I
-90
0
90
180
270
0.6
1
2
5
/
30
TIME GLUCAGON DOSE (ng/min per kg) (mid Fig. 2. Plasma glucose and glucose production data. A and B: time courses. q , dose = 0; A, dose = 0.6; 0, dose = 1; n , dose = 2; A, dose = 5; l , dose = 20 (glucagon doses are in ng. mine1 kg-‘). Each time point represents average of 5 animals. For clarity, error bars are not shown. Hormones were infused from time O-240 min. C and D: dose-response profiles. Each point represents 1 dose and is average of 5 animals. Error bars, 1 SE. Horizontal lines, mean (n = 30) basal values. Values at 0 glucagon replacement are excluded to allow logarithmic abscissa. D: open symbols, extreme (maximum or minimum) values; closed symbols, steady-state values. HGO, hepatic glucose output. l
Summary of dose-response data. Steady-state glucagon, glucose, and HGO data, as well as the extreme HGO data, are plotted in Fig. 4 vs. glucagon dose (data are expressed as percent mean basal). All four curves intersect at ~2 ng*min-l *kg-’ glucagon infusion. At higher doses, the curves diverge and are ordered (highest to lowest) steady-state glucagon, extreme HGO, steadystate glucose, and steady-state HGO, whereas at lower doses their order is reversed. None of the four variables differed significantly from 100% (P > 0.41)at dose 2 ng min-‘* kg-’ (107 2 11, 92 t 16, 92 t 20, and 93 t 8% mean basal for glucagon, extreme HGO, glucose, and steady-state HGO, respectively). Stability of metabolic variables. To assess metabolic stability during hormone replacement, CVs were calculated for the last hour of the replacement periods (180l
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 14, 2019.
BASAL
HORMONE
E535
REPLACEMENT
amsad 20
-
-90
0 TIME
90
180
270
-90
0
90
TIME
(mid
180
270
-90
wmin
0
90
TIME
(mid
Fig. 3. Blood lactate time courses at glucagon doses of 2 (left), 5 (midcUe), and 20 (right) ngemin-’ were infused from time O-240 min. Error bars, 1 SE. * Difference from basal (P c 0.05).
per
kg
180
I
270
(mid
kg? Hormones
l
identical results were obtained when basal and steadystate were defined as the last three points (last 20 min) of the period in question (data not shown). Thus metabolic variables were not destabilized by constant islet hormone replacement. DISCUSSION
GLUCAGON INFUSION (ng/min per kg)
RATE
Fig. 4. Steady-state plasma glucagon (A), extreme HGO (0), steadystate plasma glucose (A), and steady-state HGO (0) vs. glucagon replacement rate. Each point represents 1 dose and is average of 5 animals (values at dose 0 nge min-’ . kg-’ are excluded to allow logarithmic abscissa). For clarity, error bars are not shown.
240 min) and compared with CVs during fasting (-60 to -1 min). CVs for plasma glucose, HGO, glucose clearance, and blood lactate data are shown in Table 2 for all glucagon doses. Last-hour CVs varied from 0.5 to 9.6,0.3 to 14.6, 0.7 to 17.6, and 0.4 to 13.4% for glucose, HGO, clearance, and lactate, respectively. In no case was the CV during the last hour of replacement significantly greater than that during the basal period (P > 0.32). In fact, with the exception of the glucose data, CVs during the last hour of replacement were generally less than those during the corresponding basal periods. Nearly Table 2. Coefficients of variation for glucose,
Despite widespread use of somatostatin as an experimental tool (17,20,34,36,41), the feasibility of restoring basal conditions with constant insulin and glucagon has not been rigorously assessed. This is because virtually all experiments incorporating a period of basal replacement have utilized on-line adjustment of insulin (2,10,11,33) or glucagon (3, 21) to maintain euglycemia. The effects of fluctuations in islet hormone levels remain controversial (1529-31). Abrupt changes in plasma glucagon have been shown to alter hepatic sensitivity to glucagon (12, 15), an effect that could significantly influence results from protocols using variable glucagon infusion. On the other hand, variable insulin designs suffer from a potentially more important drawback. The effect of insulin to suppress HGO, like its effect on glucose uptake, requires several hours to develop (1, 42). Thus, although insulin is usually fixed before commencement of the test period (2, 1 l), it is not clear how much time must elapse after a period of insulin adjustment before the insulin effect may be considered constant.
lactate, hepatic glucose output, and glucose clearance Glucagon
Basal 0
0.6
Infusion 1
rate (ng min-’ l
2
l
kg-‘) 5
20
Plasma glucose 2.3k0.2 3.4kl.O 2.9k0.6 4.7k1.4 3.4k1.4 2.OkO.5 2.3k0.6 Hepatic glucose output 7.5t0.7 4.4t2.5 3.9k1.8 4.1k2.6 4.6k1.6* 3.9kl.l 5.Ok1.3 4.3kl.O Glucose clearance 7.4kO.8 5.0k1.2 3.9t0.7 3.6kO.9 4.9t1.5" 8.2k2.9 6.8k0.4 Blood lactate 14.7k2.0 10.9kl.l 7.4k1.2" 8.5k0.7 4.7k1.6 6.4k1.3 Values are means ,t SE in % for averages of coefficients of variation (CVs) from individual experiments, each calculated from variance among individual samples (7 samples included in each CV calculation); n = 30 (basal) and 5 (steady state). * Significant difference from corresponding basal (P > 0.05 by two-tailed paired t test; last hour value was lower than corresponding basal value in all 3 cases where significance was observed). Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 14, 2019.
E536
BASAL
HORMONE
REPLACEMENT
With these considerations in mind, we chose to test the present hypothesis (that replacement infusions of insulin and glucagon are sufficient to restore basal conditions during islet suppression) in the absence of any on-line adjustments in hormone infusion rates. Instead, we infused insulin at a basal rate in all experiments and set different glucagon infusion rates in separate experiments in an effort to define an optimal insulin-to-glucagon infusion ratio. Our data indicate that, during basal insulin infusion, glucagon infusion at rates 2 nge min-’ kg-’ result in hyperglycemia, subbasal glucose clearance, and surges in glucose production and blood lactate levels. However, at 2 ng min-’ . kg-’ glucagon infusion, plasma glucose, glucose production, glucose clearance, and blood lactate remain closely matched to their basal values. Moreover, metabolic variables were not discernably less stable and were in fact frequently more stable during hormone replacement than during the basal (fasting) periods. These results demonstrate that, on average, the principal concentrations and fluxes of glucose metabolism can be simultaneously restored to their basal levels during islet suppression with constant intraportal insulin and glucagon replacement. It should be understood, however, that individual animals could vary substantially in their response to exogenous hormone administration (6). In addition, the replacement doses defined as optimal under specific experimental conditions may not be generally applicable. It has been controversial not only how to replace islet hormones during somatostatin infusion, but whether indeed such replacement can reestablish basal metabolic conditions (3, 21). Cherrington and associates (2, 10, 11, 33) have employed a replacement protocol using constant glucagon and variable insulin infusion to maintain euglycemia. With the use of this design, they have reported constancy of plasma glucose as well as glucose turnover. However, there is reason to examine this approach more carefully. The assumption underlying the variable-insulin protocol is that the effects of insulin on metabolic parameters occur rapidly and will remain constant. At the time the protocol was first used, these assumptions appeared justifiable because 1) fasting glucose uptake is largely insulin independent (5, 39), such that insulin adjustments modulate plasma glucose primarily via HGO, and 2) HGO suppression by insulin was thought to be rapid. However, recent experiments using improved methods of HGO assessment have established that the effects of insulin on glucose production under normal conditions develop slowly and are not fully manifest for several hours (42). A potential explanation for this delay may be surmised from recent studies by us (1) and others (32), which indicate that hepatic insulin action may be mediated via a peripheral mechanism (23) for which transcapillary insulin transport is rate limiting (42). Therefore, during insulin adjustment, one might expect that it would be difficult to pinpoint the requisite insulin infusion rate unless prolonged periods of observation were possible. Thus, in the present study, it was decided to maintain constant insulin infusion at all times. One alternative to insulin adjustment is to vary the l
l
l
glucagon infusion rate while holding insulin constant. Jackson et al. (21) used such a protocol to maintain euglycemia during islet suppression in humans. They reported considerable instability in plasma glucose following a period of glucagon adjustment. We used a similar approach in dogs and found that an unacceptably high glucagon infusion rate was required to restore basal glucose (3), possibly due to waning of the glucagon effect (8, 13). Thus, while glucagon action is rapid (16), its well-documented evanescence seems to limit the usefulness of glucagon adjustments for the maintenance of euglycemia. The present studies, in which both islet hormones were administered at fixed rates, support the contention of Cherrington and associates that a constant metabolic state may be reacheived during islet suppression by replacing basal plasma insulin and glucagon levels. However, one inconsistency between our results and results previously reported (2, 7) is the glucagon infusion rate needed for basal reestablishment. Thus the Vanderbilt group has typically used 0.6-l ng min-’ . kg-’ intraportal glucagon (2, 7), whereas we required -2 ng* min-lo kg-l in the present study. The reason for this discrepancy is not clear, although it should be noted that the Vanderbilt group has not reported a complete glucagon dose response at constant insulin. Nevertheless, results from both laboratories are consistent, in the sense that metabolic parameters are restored to basal concurrently with a matching of plasma immunoreactive glucagon to basal. This suggests the possibility that our animals regulate their blood glucose at higher glucagon levels, perhaps as a result of differences in diet, exercise, or other variables. It is noteworthy that fasting metabolic conditions were closely recreated at the 2 ng min-’ . kg-’ glucagon dose, despite some (-30%) underreplacement of plasma insulin. Based on immunoreactive glucagon measurements (Fig. l), it seems unlikely that the insulin underreplacement was compensated by a decrease in plasma glucagon. Alternatively, it is possible that the expected consequences of insulin underreplacement (increased HGO and decreased glucose clearance) were masked by extrapancreatic effects of somatostatin. O’Brien et al. (28) have shown increased hepatic insulin extraction and decreased HGO in response to somatostatin; such an effect would tend to maintain basal HGO levels despite hypoinsulinemia. Also, a potential decrease in glucose clearance may have been averted by somatostatin, which is known to augment glucose clearance in some cases (3). It is conceivable, therefore, that the (unintentional) underreplacement of insulin in the present experiments was fortuitously compensated by extrapancreatic somatostatin effects. In studies where potential extrapancreatic somatostatin effects could be confounding, the depancreatized dog model (40) may be preferable to somatostatin-based protocols. In summary, the present data demonstrate that stable metabolic conditions closely resembling the fasting state can be recreated during islet suppression by means of constant intraportal replacement infusions of insulin and glucagon (200 PU. min-l kg-’ and 2 ng* min-’ . kg-‘, respectively). This result supports the validity of islet clamp protocols and suggests that the primary effect of l
l
l
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 14, 2019.
BASAL
HORMONE
somatostatin on fasting glucose metabolism is to reduce the plasma concentrations of insulin and glucagon. We thank Donna Banks, Ginger Hayes, Paul Kirkman, and Dr. Linda Kirkman for surgical and experimental assistance and Elsa Demirchyan and Lena Roumian for technical assistance. This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-27619. D. C. Bradley is a predoctoral trainee supported by National Institute on Aging Trainee Grant T32-AG-00093. Address for reprint requests: R. N. Bergman, Dept. of Physiology and Biophysics, Univ. of Southern California Medical School, 2025 Zonal Ave., 128 Mudd Bldg., Los Angeles, CA 90033. Received 5 July 1991; accepted in final form 4 November 1991. REFERENCES 1. Ader, M., and R. N. Bergman. Peripheral effects of insulin dominate suppression of fasting hepatic glucose production. Am. J. Physiol. 258 (Endocrinol. Metab. 21): E1020-E1032,1990. 2. Adkins-Marshall, B. A., S. R. Myers, G. K. Hendrick, P. E. Williams, K. Triebwasser, B. Floyd, and A. D. Cherrington. Interaction between insulin and glucose-delivery route in regulation of net hepatic glucose uptake in conscious dogs. Diabetes 39: 87-95,199O. 3. Bergman, R. N., M. Ader, D. T. Finegood, and G. Pacini. Extrapancreatic effect of somatostatin infusion to increase glucose clearance. Am. J. Physiol. 247 (Endocrinol. Metab. 10): E370-E379, 1984. 4. Bergman, R. N., and R. J. Bucolo. Interaction of insulin and glucose in the control of hepatic glucose balance. Am. J. Physiol. 227: 1314-1322,1974. 5. Best, J. D., G. J. Taborsky, J. B. Halter, and D. Porte, Jr. Glucose disposal is not proportional to plasma glucose level in man. Diabetes 30: 847-850, 1981. 6. Bradley, D. C., and R. N. Bergman. Hepatic glucagon sensitivity and fasting glucose concentration in normal dogs. Am. J. Physiol. 262 (Endocrinol. Metab. 25): E539-E545,1992 7. Cherrington, A. D., J. L. Chiasson, J. E. Liljenquist, A. S. Jennings, U. Keller, and W. W. Lacy. The role of insulin and glucagon in the regulation of basal glucose production in the postabsorptive dog. J. Clin. Invest. 58: 1407-1418, 1976. 8. Cherrington, A. D., M. P. Diamond, D. R. Green, and P. E. Williams. Evidence for an intrahepatic contribution to the waning effect of glucagon on glucose production in the conscious dog. Diabetes 31: 917-922, 1982. 9. Cherrington, A. D., W. W. Lacy, and J. L. Chiasson. Effect of glucagon on glucose production during insulin deficiency in the dog. J. Clin. Invest. 62: 664-677, 1978. 10. Cherrington, A. D., W. W. Lacy, P. E. Williams, and K. E. Steiner. Failure of somatostatin to modify effect of glucagon on carbohydrate metabolism in the dog. Am. J. Physiol. 244 (Endocrinol. Metab. 7): E596-E602,1983. 11. Davis, M. A., P. E. Williams, and A. D. Cherrington. Effect of glucagon on hepatic lactate metabolism in the conscious dog. Am. J. Physiol. 248 (Endocrinol. Metab. 11): E463-E470,1985. 12. Deri, J. J., P. E. Williams, K. E. Steiner, and A. D. Cherrington. Altered ability of the liver to produce glucose following a period of glucagon deficiency. Diabetes 30: 490-495, 1981. 13. El Refai, M., and R. N. Bergman. Glucagon-stimulated glycogenolysis: time-dependent sensitivity to insulin. Am. J. Physiol. 236 (Endocrinol. Metab. Gastrointest. Physiol. 5): E246-E254,1979. 14. Finegood, D. T., and R. N. Bergman. Optimal segments: a method for smoothing tracer data to calculate metabolic fluxes. Am. J. Physiol. 244 (Endocrinol. Metab. 7): E472-E479,1983. 15. Fradkin, J., H. Shamoon, P. Felig, and R. S. Sherwin. Evidence for an important role of changes in rather than absolute concentrations of glucagon in the regulation of glucose production in humans. J. Clin. Endocrinol. Metab. 50: 698-703, 1980. 16. Gerich, J., P. Cryer, and R. Rizza. Hormonal mechanisms in acute glucose counterregulation: the relative roles of glucagon, epinephrine, norepinephrine, growth hormone, and cortisol. Met& Clin. Exp. 29: 1164-1175, 1980. 17. Hansen, I., R. Firth, M. Haymond, P. Cryer, and R. Rizza.
REPLACEMENT
E537
The role of autoregulation of the hepatic glucose production in man: response to a physiologic decrement in plasma glucose. Diabetes 35: 186-191, 1986. G. K., R. T. Frizzell, and A. D. Cherrington. 18. Hendrick, Effect of somatostatin on nonesterified fatty acid levels modifies glucose homeostasis during fasting. Am. J. Physiol. 253 (Endocrinol. Metab. 16): E443-E452,1987. 19. Herbert, V., K. Law, C. Gottlieb, and S. Bleicher. Coated charcoal immunoassay of insulin. J. Clin. Endocrinol. Metab. 25: 1375-1384,1965. 20. Hoelzer, D. R., G. P. Dalsky, W. E. Clutter, S. D. Shah, J. 0. Holloszy, and P. E. Cryer. Glucoregulation during exercise: hypoglycemia is prevented by redundant glucoregulatory systems, sympathochromaffin activation, and changes in islet hormone secretion. J. Clin. Invest. 77: 212-221, 1986. 21. Jackson, R. A., J. B. Hamling, B. M. Sim, P. M. Blix, J. B. Jaspan, S. Markanday, and J. D. N. Nabarro. Basal glucose homeostasis with and without fixed concentrations of glucagon and insulin. Metab. Clin. Exp. 36: 131-136, 1987. 22. Lavelle-Jones, M., M. H. Scott, 0. Kolterman, A. H. Rubenstein, J. M. Olefsky, and A. R. Moossa. Selective suppression of hepatic glucose output by human proinsulin in the dog. Am. J. Physiol. 252 (Endocrinol. Metab. 15): E230-E236, 1987. 23. Levine, R., and I. B. Fritz. The relation of insulin to liver metabolism. Diabetes 5: 209-219, 1956. 24. Li, C. H. Human growth hormone: 1974-1981. Mol. Cell. Biochem. 46: 31-41,1982. H. L. A., G. G. Ross, and M. Vranic. Effects of 25. Lickley, selective insulin or glucagon deficiency on glucose turnover. Am. J. Physiol. 236 (Endocrinol. Metab. Gastrointest. Physiol. 5): E255E262,1979. J. E., G. L. Mueller, A. D. Cherrington, U. 26. Liljenquist, Keller, J. L. Chiasson, J. M. Perry, W. W. Lacy, and D. Rabinowitz. Evidence for an important role of glucagon in the regulation of hepatic glucose production in normal man. J. Clin. Invest. 59: 369-374, 1977. J. E., G. L. Mueller, A. D. Cherrington, J. M. 27. Liljenquist, Perry, and D. Rabinowitz. Hyperglycemia per se (insulin and glucagon withdrawn) can inhibit hepatic glucose production in man. J. Clin. Endocrinol. Metab. 48: 171-175, 1979. D. W., R. V. Rajotte, G. D. Molnar, K. Toth, and 28. O’Brien, T. D. Petersen. Somatostatin increases hepatic insulin extraction. Diubetes Res. Clin. Pratt. 7: 171-178, 1988. 29. Paolisso, G., A. J. Scheen, A. S. Luyckx, and P. J. Lefebvre. Pulsatile hyperglucagonemia fails to increase hepatic glucose production in normal man. Am. J. Physiol. 252 (Endocrinol. Metab. 15): El-E7,1987. 30. Paolisso, G., A. J. Scheen, E. M. Verdin, A. S. Luyckx, and P. J. Lefebvre. Insulin oscillations per se do not affect glucose turnover parameters in normal man. J. Clin. Endocrinol. Metab. 63: 520-525,1986. A. Scheen, F. 31. Paolisso, G., S. Sgambato, N. Passariello, D’Onofrio, and P. J. Lefebvre. Greater efficacy of pulsatile insulin in type I diabetics critically depends on plasma glucagon levels. Diabetes 36: 566-570, 1987. 32. Prager, R., P. Wallace, and J. M. Olefsky. Direct and indirect effects of insulin to inhibit hepatic glucose output in obese subjects. Diabetes 36: 607-611, 1987. 33. Radosevich, P. M., P. E. Williams, J. R. McRae, W. W. Lacy, D. N. Orth, and N. N. Abumrad. &Endorphin inhibits glucose production in the conscious dog. J. Clin. Invest. 73: 12371241,1984. 34. Sherwin, R. S., W. Tamborlane, R. Hendler, L. Sacca, R. A. DeFronzo, and P. Felig. Influence of glucagon replacement on the hyperglycemic and hyperketonemic response to prolonged somatostatin infusion in normal man. J. Clin. Endocrinol. Metab. 45: 1104-1107,1977. 35. Shulman, G. I., J. E. Liljenquist, P. E. Williams, W. W. Lacy, and A. D. Cherrington. Glucose disposal during insulinopenia in somatostatin-treated dogs: the roles of glucose and glucagon. J. Clin. Invest. 62: 487-491, 1978. 36. Starke, A., T. Imamura, and R. H. Unger. Relationship of glucagon suppression by insulin and somatostatin to the ambient glucose concentration. J. Clin. Invest. 79: 20-24, 1987. 37. Steele, R., J. S. Wall, R. C. deBodo, and N. Altszuler.
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 14, 2019.
E538
BASAL
HORMONE
Measurement of size and turnover rate of body glucose pool by the isotope dilution method. Am. J. Physiol. 187: 15-24, 1956. 38. Vasquez, B., M. Nagulesparan, and R. H. Unger. Fasting and postprandial plasma glucagon and glucagon-like immunoreactivity are normal in obese Pima Indians. Horm. Metab. Res. 15: 218-220, 1983. 39. Verdonk, C. A., R. A. R&a, and J. E. Gerich. Effects of plasma glucose concentration on glucose utilization and glucose clearance in normal man. Diabetes 30: 535-537,198l.
REPLACEMENT
40. Wasserman, D. H., H. L. A. Lickley, and M. Vranic. Important role of glucagon during exercise in diabetic dogs. J. A&. Physiol. 59: 1272-1281, 1985. 41. Wolfe, R. R., E. R. Nadel, J. H. F. Shaw, L. A. Stephenson, and M. H. Wolfe. Role of changes in insulin and glucagon in glucose homeostasis in exercise. J. Clin. Invest. 77: 900-907, 1986. 42. Yang, Y. J., I. D. Hope, M. Ader, and R. N. Bergman. Insulin transport across capillaries is rate limiting for insulin action in dogs. J. Clin. Invest. 84: 1620-1628, 1989.
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 14, 2019.
