Proc. Nati. Acad. Sci. USAVol. 88, pp. 1296-1300, February 1991Physiology/Pharmacology

Mechanism of glucoregulatory responses to stress and theirdeficiency in diabetes

(glucose metabolism/carbachol/third ventricle)

PHILIP D. G. MILES*, KEIICHI YAMATANI*t, H. LAVINA A. LICKLEY*t§¶, AND MLADEN VRANIC*tDepartments of *Physiology, tMedicine, and §Surgery, University of Toronto, and $Women's College Hospital, Toronto, ON M5S 1A8, Canada

Communicated by Viktor Mutt, October 15, 1990 (received for review January 24, 1990)

ABSTRACT During exercise, increased energy demandsare met by increased glucose production that occurs simulta-neously with the increased glucose uptake. We had previouslyobserved that, during exercise, metabolic clearance rate ofglucose (MCR) increases markedly in normal, but only mar-ginally in poorly controlled diabetic dogs. We wished todetermine (i) whether in a more general model of stressmatched increases in rate of appearance of glucose and MCRalso occur, or if MCR is suppressed, as during catecholamineinfusion; and (ii) whether diabetes affects stress-inducedchanges in rate ofglucose appearance and MCR. Therefore, weinjected carbachol (27 nmol/50 pl), an analog of acetylcholine,intracerebroventricularly in seven conscious dogs before andafter induction of alloxan diabetes. In normal dogs, plasmaepinephrine and cortisol increased 4- to 5-fold, whereas nor-epinephrine and glucagon doubled. Plasma insulin, however,remained unchanged. Tracer-determined hepatic glucose pro-duction increased rapidly, but transiently, by 2.5-fold. Thisincrement can be fully explained by the observed increments inthe counterregulatory hormones. Surprisingly, however, MCRalso promptly increased, and therefore, plasma glucosechanged only marginally. After induction of diabetes, theanimals were given intracerebroventricular carbachol whileplasma glucose was maintained at moderate hyperglycemia(9.0 ± 0.4 mM). Increments in counterregulatory hormoneswere similar to those seen in normal dogs, except for exagger-ated norepinephrine release. Peripheral insulin levels werehigher in diabetic than in normal dogs; however, MCR wasmarkedly reduced and the lipolytic response to stress in-creased, indicating insulin resistance. Interestingly, the hyper-glycemic response to stress was 6-fold greater in diabetic thannormal animals, relating mainly to the failure ofMCR to rise.Plasma lactate increased equivalently in diabetic and normalanimals despite suppression ofMCR in the diabetics, indicatingeither greater muscle glycogenolysis and/or impairment inglucose oxidation. We conclude that in this stress model MCRincreases as in exercise in normal but not in diabetic dogs. Wespeculate that glucose uptake in stress could be mediatedthrough an insulin-dependent neural mechanism.

The obligatory requirement of the brain for glucose necessi-tates a precise mechanism for glucose homeostasis. Thehormonal and metabolic responses to stress have been ex-amined under a variety of stress conditions such as severeinjury (1, 2), major surgery (3, 4), myocardial infarction (4, 5),and emotional stress (6); however, the glucoregulatory mech-anisms are not fully understood. In exercise, another form ofstress, the peripheral energy requirements necessitatechanges in glucose metabolism, whereby the augmented rateofglucose utilization (glucose disappearance; Rd) is preciselymatched with an elevated rate of hepatic glucose production

(glucose appearance; Ra), thus averting the threat of hypo-glycemia (7). There is also a safeguard against hyperglycemiadependent on insulin. With insulin deficiency or resistancethe exercise-induced increment in metabolic clearance rate ofglucose (MCR) is diminished, whereas, depending on thediabetic state, the increment in Ra is often normal-henceleading to further deterioration of glycemic control (7, 8). Wepreviously attempted to reproduce a portion of the stressresponse in dogs by epinephrine infusion (9-11): progressivehyperglycemia resulted because of enhanced glucose pro-duction and suppressed glucose clearance. In diabetes thehyperglycemia was further exacerbated with epinephrineinfusion (10, 11).A general model of stress was developed in anesthetized

rats (12). The intracerebroventricular (i.c.v.) injection ofcarbachol, an acetylcholine analog, elicits hormonal re-sponses similar to those seen in clinical stress and in exercise.In studies in the rat, glucose fluxes were not examined and,therefore, the mechanism ofthe ensuing hyperglycemia couldnot be determined. In addition, it is well known that anes-thesia can affect both hormonal responses and their periph-eral effects (13).The aim of the present study is to ascertain the glucoreg-

ulatory responses to this moderate general stress (i.c.v.carbachol) in physiology and diabetes. II Specifically, wewished to determine (i) whether the expected increase inglucose production induced by release of counterregulatoryhormones would also be accompanied by suppression ofMCR, as during epinephrine infusions, or if the increasedenergy demands would lead to an increase in MCR as seen inexercise; and (ii) whether excessive hyperglycemia occurs indiabetic dogs, and if so, whether the hyperglycemia resultsfrom excessive glucose production and/or suppressed glu-cose clearance. Therefore, we injected a small dose ofcarbachol into the third ventricle via a chronic indwellingi.c.v. cannula in conscious dogs before and after induction ofalloxan diabetes. Non-steady-state tracer methods were usedto measure glucose production and utilization. In diabeticdogs, basal glucose was clamped at =9.0mM with a subbasalperipheral infusion to mimic the hyperglycemia frequentlyseen in postabsorptive diabetic patients.

MATERIALS AND METHODSExperimental Animals. Studies were conducted in seven

mongrel dogs, 18-26 kg, that were housed under controlledtemperature and light cycle conditions and fed once daily adiet ofdry chow (lab canine diet 5006; Purina Mills, St. Louis,

Abbreviations: Ra, rate of glucose appearance; Rd, rate of glucosedisappearance; MCR, metabolic clearance rate of glucose; i.c.v.,intracerebroventricular.tPresent address: The Third Department of Internal Medicine,Yamagata University, School of Medicine, Japan."This work was presented in part at the 24th Annual Meeting of theEuropean Association of the Study of Diabetes (36).

1296

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

May

3, 2

021

Proc. Nati. Acad. Sci. USA 88 (1991) 1297

MO) mixed with canned meat (Friskies, Don Mills, ONCanada). Food intake, body weight, body temperature, andhematocrit were monitored regularly. Because extendedmaintenance of animals with chronic indwelling cathetersrequires use of antibiotics (14), either penicillin G (PenlongXL; Rogar/STB, Montreal) or gentamicin (Gentocin; Scher-ing) was administered as needed.

Surgery was done under general anesthesia to implant ani.c.v. guide cannula into the third ventricle (15) and vascularsampling and infusion catheters. Two to three weeks elapsedbefore the first experiment was performed. The patency ofthe head cannula was verified by aspirating a small volume ofcerebrospinal fluid before each experiment. Experimentswere done before and after the induction of alloxan diabetes(65 mg/kg; Sigma). Diabetic dogs were maintained on regularand NPH pork insulin (Eli Lilly) injected s.c. daily duringmorning feeding to keep glucosuria between 0.1 and 1%. Atleast 6 days was allowed between experiments. All experi-ments were performed on 16- to 18-hr fasted, conscious dogstrained to lie quietly.

Experimental Protocol. Normal dogs. Experiments werebegun with a priming injection (25 ktCi per 5 ml; 1 Ci = 37GBq) and constant infusion (0.25 puCi per min) of D-[3-3H]glucose (New England Nuclear). Tracer glucose wasdiluted to 2.5 ,uCi/ml in a solution containing unlabeledD-glucose at 100 mg/deciliter (BDH) as carrier and sodiumbenzoate at 200 mg/deciliter (Fisher) as preservative. Car-bachol (carbamoylcholine; Aldrich) at 27 nmol per 50 p1 ofwater was injected i.c.v. after a 2-hr tracer equilibriumperiod. Injections were made through a 24-gauge injectioncannula connected to a 100-ttl microsyringe (Hamilton) viapolyethylene tubing (Intramedic PE 50; Clay Adams). Thehead-cannula retaining screw was loosened to allow theneedle to pierce the septum and extend =1 mm beyond theguide cannula.

Diabetic dogs. Experiments were done only when themorning plasma glucose was between 8.9 and 13.4 mM. Aninsulin infusion was begun to lower the plasma glucose levels,which were clamped between 8.9 and 10.0 mM (moderatehyperglycemia). This level of glycemia during the carbacholexperiments was below the urinary threshold in dogs, andtherefore, a correction for glycosuria in diabetic dogs was notnecessary. Once the desired level was achieved, the insulininfusion rate was held constant, and the experimental pro-tocol proceeded as outlined above for the normal animals.The infusate was prepared from regular pork insulin (100units per ml, Iletin II) in 50 ml ofnormal saline containing 2-3ml of the dog's own plasma.

Control studies. Five normal dogs also received i.c.v.injection of 100 ,ul ofwater to ascertain that any effects of thei.c.v. carbachol were not merely a reflection of the pressureeffect of the volume of fluid injected.Blood Sampling and Assay Methods. Blood samples for

determination of plasma glucose, glucose specific activity,insulin, cortisol, lactate, and glycerol were collected in tubescontaining heparin (50 units per ml ofblood) and NaF (1-2 mgper tube). Samples for determination of plasma glucagon andfree fatty acid concentration were collected in tubes contain-ing 0.1 ml of aprotinin (Trasylol; FBA Pharmaceuticals, NewYork) and 0.1 ml of EDTA (24 mg/ml; BDH). For determi-nation of norepinephrine and epinephrine, 1-ml blood sam-ples were collected in polyethylene tubes containing 2.5 mgof glutathione (Boehringer Mannheim) and 10 IlI of EGTA(Sigma). All blood was stored at 4°C and centrifuged within60 min. The resultant plasma was stored at -20°C, except theplasma for catecholamine measurements, which was depro-teinized with 2 M HCI04 and stored at -70'C.Glucose concentration, specific activity, and concentra-

tion of metabolites (glycerol, free fatty acid, lactate) and

hormones (glucagon, epinephrine, norepinephrine) were de-termined as described (16).

Calculations. Glucose concentration and specific activitydata were systematically smoothed using the optimal seg-ments technique (17). The Ra and Rd of glucose were calcu-lated with an equation for non-steady-state turnover (18) thatwas subsequently validated (19). MCR was calculated bydividing Rd by the prevailing plasma glucose concentration.MCR represents an estimate of glucose utilization, partiallycorrected for the mass action effect of glucose. The validityand meaning of this measurement have been reviewed (20).

Statistical Analysis. Values presented are means ± SEMs.Statistical analysis to assess responses to i.c.v. carbachol anddifferences in response in the diabetic as opposed to thenormal state was performed by using a two-way analysis ofvariance for unbalanced data (Statistical Analysis System;SAS Institute, Carey, NC), run on a microcomputer (IBMP/S 2). Significance was assumed at P < 0.05.

RESULTSNormal Dogs. As seen in Fig. 1, the i.c.v. injection of

carbachol in normal dogs resulted in a marked rapid, buttransient, increase in Ra (P < 0.0001) from a basal value of15.8 ± 2.1 to a peak value of 35.8 ± 6.4 ,umol/min per kg at

15

E 13

Q)0

g

C 9aEU) 70

540

co -,

a) ono E

,- Ew-60

30

. Carbachol

*II*II

e-*NormalI 0o oDiabetic

II I~~~-Vt _ts

20 [

10 L

c0-.1

L

DU)U,00

5

m

c

E11-

U)0C - s)0 )

0-'a) C

U, -

gE(D -',Jn0

30 -

20 -

10 .4-

I

2

1 1 T

-50 -25 0 25 50 75 100 125

Time (min)

FIG. 1. The effect of carbachol (27 nmol per 50 Al of water)injected into the third ventricle (t = 0 min) of normal (e) andmoderately controlled diabetic (o) dogs (n = 7) (mean + SEM).

Physiology/Pharmacology: Miles et A

I

i

.

Dow

nloa

ded

by g

uest

on

May

3, 2

021

1298 Physiology/Pharmacology: Miles et al.

10 min. Ra then fell to 60% of the initial response andgradually approached basal levels over 2 hr. There was aconcomitant increase in Rd (P < 0.0001), which rose from14.8 ± 2.3 to 26.6 ± 6.5 mg per min per kg. Because Ra andRd both rose, there was only a marginal increase in plasmaglucose concentration (5.9 ± 0.2 to 6.7 ± 0.3 mM; P <0.0001), and consequently, MCR mirrored Rd. Concurrentlywith Rd (P < 0.001), plasma lactate increased 3-fold from 1.16± 0.15 to 3.55 ± 1.07 mM. Plasma glycerol concentrationdoubled from 95 ± 15 to 187 ± 46 ,AM (P < 0.001), and freefatty acids levels also rose (815 ± 138 to 1321 ± 265microequivalents per liter; P < 0.01) (Fig. 2).

Fig. 3 shows a rapid 2-fold increase over basal in plasmanorepinephrine concentration (1.06 ± 0.33 to 1.88 ± 0.37 nM;P < 0.0001) and a 3- to 4-fold delayed increase in epinephrine(0.66 ± 0.18 to 2.44 ± 0.64 nM; P < 0.0001). There was amarked delayed increase in plasma cortisol (34 ± 13 to 188 +42 nM; P < 0.0001) and a modest increase in plasma glucagon(151 ± 21 to 247 ± 65 ng/liter; P < 0.05). In contrast, insulinremained unchanged from a basal level of 85 ± 13 pM.

Diabetic Dogs. Results seen after the induction of alfoxandiabetes are presented together with those in the nondiabeticstate. The desired basal plasma glucose concentration of 9.0± 0.4 mM (moderate hyperglycemia) was achieved with anaverage peripheral insulin infusion rate of 137 ± 23 micro-units per min/kg, which caused moderate peripheral hyper-insulinemia (117 ± 22 pM). This infusion rate is -80%/ thatrequired to achieve normoglycemia in depancreatized dogs(11). Basal MCR in these moderately hyperglycemic dogswas markedly suppressed from that in normal dogs (1.58 +0.18 ml per min/kg compared with 2.59 ± 0.33; P < 0.0001),indicating insulin resistance.The i.c.v. administration of carbachol to diabetic dogs

caused a marked and sustained hyperglycemic response that

5 r

E0)

0

0-a

4

3

2

500

400

3000

0 200

100 F

2000 r

LLJ

Carbachol

*-* Normalo0o Diabetic

/ \~~~~~~~~~~~

1500 F

1000

500' I

-50 -25 0 25 50 75 100 125

Time (min)

FIG. 2. The effect of carbachol (27 nmol per 50 jl of water)injected into the third ventricle (t = 0 min) of normal (0) andmoderately controlled diabetic (o) dogs (n = 7) (mean ± SEM). FFA,free fatty acids.

c

I.-_Q)C:

0).L_

C0)

0z

cC_

-c

0)

L-

3

2

2

0

--200

c

-6* 100 .

oC)

0-

>-300 -

c

c 20000)

o 100200

0

c 100M._

Ca- O

Proc. Natl. Acad. Sci. USA 88 (1991)

Carbachol

II LLI

o-Normol0- o Diabetic

-50 -25 0 25 50 75 100 125

Time (min)

FIG. 3. The effect of carbachol (27 nmol per 50 1,u of water)injected into the third ventricle (t = 0 min) of normal (0) andmoderately controlled diabetic (o) dogs (n = 7) (mean + SEM).

was six times greater than that in normal dogs. Fig. 1indicates the main reason for exaggerated hyperglycemia wasthat, unlike the situation in normal animals, the increase in Rawas not matched by a comparable increase in Rd, and MCRremained unchanged from the already greatly diminishedbasal rate. The more prolonged peak in Ra may also havecontributed to hyperglycemia. In diabetic dogs Rd rose from14.1 + 1.0 to 21.4 ± 2.3 (P < 0.0001). This increment in Rdof the diabetic group was significantly less than that of thenormal group (P < 0.0001). Fig. 2 shows that the increase inlactate (1.34 ± 0.28 to 3.97 ± 0.83 mM; P < 0.0001) wassimilar to that in normal dogs, despite a much smaller anddelayed increase in Rd. There was a rapid and marked 4-foldincrease in plasma glycerol from 113 ± 25 to 444 ± 81 ,M (P< 0.0001) that was much larger than in normal dogs. Therewas also an increase in free fatty acids from 1124 ± 308 to1718 ± 284 microequivalents per liter (P < 0.001).The magnitude and dynamics of the responses of plasma

epinephrine, cortisol, glucagon, and insulin in the diabetics(Fig. 3) were remarkably similar to that of the normal dogs.However, the norepinephrine response was almost double indiabetic (1.06 + 0.21 to 2.67 ± 0.20; P < 0.0001).

IIIIII

I

Dow

nloa

ded

by g

uest

on

May

3, 2

021

Proc. NatL. Acad. Sci. USA 88 (1991) 1299

Control Studies. A control study in which water wasinjected i.c.v. was done in five normal dogs; there was nosignificant change in any of the hormonal or metabolicparameters measured (data not shown).

DISCUSSIONThis study was designed to investigate the hormonal andglucoregulatory responses to carbachol injection into thebrain third ventricle, which provides a model of moderatestress, and to assess alterations in these responses in diabeticdogs. Stress can be classified as a state in which physiologicalhomeostasis is perturbed. However, there are many differenttypes of stress, each characterized by different hormonal andmetabolic responses. The activation of the sympathetic ner-vous system and the anterior pituitary-adrenocortical axis iscommon to all types of stress. It has been proposed thatstimulation of corticotropin-releasing factor can activate theadrenocorticotropin (ACTH)-cortisol system and the auto-nomic nervous system, concurrently (21). In different typesof stress, however, the ratio of the various counterregulatoryhormones released varies. Elevated epinephrine and norepi-nephrine levels are seen during major surgery (3, 4), trauma(1, 2), severe infections (1, 2), myocardial infarctions (4, 5),ketoacidosis (1), and also emotional stress (6). However,under certain stress conditions, the adrenomedullary activitycan be stimulated more than the sympathetic outflow (22, 23).We have observed that in insulin-induced hypoglycemia,plasma epinephrine increased 40-fold, whereas norepineph-rine and cortisol increased only 4-fold (24).

Exercise can also be considered as a type of stress. Duringmoderate exercise in normoglycemic dogs, there are 2- to3-fold proportional increases in epinephrine, norepinephrine,and cortisol (25). However, even with very mild hypoglyce-mia, increments of epinephrine and cortisol, but not norepi-nephrine, are generally exaggerated (25). In alloxan diabeticdogs (partial insulin deficiency), responses of all counterreg-ulatory hormones (norepinephrine, epinephrine, cortisol,and glucagon) were increased (26, 27).Normal Dogs. The i.c.v. administration of a small dose of

carbachol seems to provide a reasonable model for a gener-alized, moderate stress because most of the generally ac-cepted stress hormones were released. Temporally the peakin plasma norepinephrine preceded that of epinephrine, cor-tisol, and glucagon. This relationship is not unexpectedbecause norepinephrine release reflects increased sympa-thetic neural activity (28), which, in turn, stimulates therelease of the other hormones from the adrenal medulla andcortex and pancreatic alpha cells, respectively. Interestingly,no change occurred in plasma insulin, which could reflect abalance between the opposing effects on the pancreatic betacell of (i) the alpha- and beta-adrenergic effects of thecatecholamines or (ii) the adrenergic or cholinergic effects(29). The possibility of parasympathetic involvement, inaddition to sympathetic involvement, arises from the fact thatpupil dilation and salivation were often observed with car-bachol administration. Concurrently, with the increase incounterregulatory hormones there was a 2-fold increase inglucose production that then gradually declined toward basal.We suggest that this increase in glucose production resultedfrom the 4-fold increase of epinephrine that was accompaniedby a 2-fold increase in glucagon. This result compares well toour previous experiments in depancreatized dogs that weremaintained normoglycemic with a constant insulin infusion(11). During epinephrine infusion in these dogs, the incre-ments in glucagon and in glucose production were similar tothe responses occurring after i.c.v. carbachol. It is unlikelythat increments in cortisol played a role during the 2-hrexperimental period because corticol effects are usually notmanifested for 3-4 hr (30). It was surprising, however, that

the increase in glucose production was accompanied by amatched increment of glucose uptake, particularly becausethere was no change in plasma insulin, the only hormoneknown to increase glucose uptake. Therefore, the concomi-tant increase in Ra and Rd resulted in only a minimal, butsignificant, increase in plasma glucose concentration. Undersimilar conditions mimicked by epinephrine infusion, a 2-foldrise in plasma insulin elicited only a minimal increase in Rd(9). Indeed, in early studies (31) we have demonstrated thata 2-fold increase in MCR requires a 5- to 6-fold increase inperipheral insulin concentration. Our insulin assay can detectmuch smaller changes than that and therefore, the increase inMCR cannot be related to the change in peripheral insulinconcentration. It is known that peripheral insulin does notaccurately reflect the changes in portal insulin, but forperipheral MCR it is peripheral and not portal insulin that isrelevant. With epinephrine infusion in human (32, 33) and dog(11), MCR decreased, while in the present study, i.c.v.carbachol injection increased MCR. As a consequence,plasma glucose increased markedly during epinephrine infu-sion but only marginally during carbachol injection. Thisincrease in MCR after carbachol is similar to that seen duringexercise in normal dogs, where insulin exerts only a permis-sive effect (25). The rapid increment in MCR after carbacholimplies the possibility of a neural mechanism that couldincrease glucose uptake independently of any increment inplasma insulin.Carbachol injection not only mobilized glucose, but it also

increased lipolysis, as evidenced by increase of plasmaglycerol and free fatty acids.

Diabetic Dogs. The metabolic effects of i.c.v. carbacholwere greatly exaggerated in diabetic dogs. The increment inplasma glucose was six times, and that of glycerol was twoand one-half times larger than in normal dogs. However, ofthe hormones measured, only the response of norepinephrinewas exaggerated in diabetes. Because the diabetic dogs weregiven a constant insulin infusion, that insulin concentrationdid not change is not surprising. The initial rise of plasmaglucose could have been, in part, from a more prolonged peakresponse of hepatic glucose production. However, the mainreason for hyperglycemia is that the increment in Rd wasdelayed and much smaller than in normal dogs, and MCRremained suppressed. Therefore, the glucoregulatory effectsof i.c.v. carbachol were comparable to those with epineph-rine infusion in diabetic, hyperglycemic dogs (11). Rd prob-ably rose due only to the mass action of glucose becauseMCR remained unchanged. Plasma lactate rose similarly inboth diabetic and normal dogs. Because in diabetic dogs theRd rise was much smaller than in normal dogs, more lactatecould have been produced from greater adrenergic stimula-tion ofmuscle glycogenolysis and/or glucose oxidation couldbe impaired in diabetic animals. The latter is consistent witha defect in glucose oxidation apparently from decreasedpyruvate dehydrogenase activity, as seen in longstandingdiabetes (34, 35). The subbasal insulin infusion that main-tained moderate hyperglycemia in diabetic dogs led to pe-ripheral insulin levels higher than in normal dogs. This is notunexpected because insulin was infused peripherally and notintraportally. However, despite such peripheral insulin lev-els, MCR was markedly reduced, and stress-induced lipolysiswas enhanced, indicating insulin resistance. Because of in-sulin resistance, carbachol did not increase glucose clearancein diabetic dogs, as it did in normal dogs. Thus, normalincreases in glucose uptake during both stress and exerciseappear to be insulin dependent. Our experimental design doesnot permit us to delineate the site ofincreased glucose uptakein normal dogs. One can surmise that, in addition to muscle,there would also be some uptake in adipose tissue as illus-trated in Fig. 2 because during carbachol stimulation, in

Physiology/Pharmacology: Miles et al.

Dow

nloa

ded

by g

uest

on

May

3, 2

021

1300 Physiology/Pharmacology: Miles et al.

addition to lipolysis, some reesterification for free fatty acidsis indicated as well.

In conclusion, in normal dogs, moderate stress induced byi.c.v. carbachol resulted in the release of all counterregula-tory hormones, but the increases in epinephrine and cortisolwere most prominent. Increments in glucose production andclearance were matched, and therefore, plasma glucose in-creased only marginally. The increase in glucose productionreflects the changes in the counterregulatory hormones. Wespeculate, however, that the increment of glucose clearancecould reflect a hitherto unknown neural mechanism. It isinteresting that glucoregulation during stress and exercise isremarkably similar, which could reflect increased fuel de-mands occurring under both sets ofconditions. In moderatelycontrolled diabetes there was insulin resistance, evidencedby suppressed metabolic glucose clearance and by excessive,stress-induced lipolysis. After i.c.v. carbachol, only therelease of norepinephrine was exaggerated. However, theglycemic response is augmented 6-fold, as compared tonormal dogs. This result occurred mainly because glucoseclearance did not increase. We suggest that, as in exercise,regulation of glucose uptake during stress requires a permis-sive effect of insulin.

The authors thank D. Bilinski, L. Cook, L. Lam, and M. VanDe-langeryt for excellent technical assistance, the Core Laboratory ofBanting and Best Diabetes Centre of the University of Toronto foranalyses of plasma lactate and glycerol, and Phillip E. Williams,Vanderbilt University School of Medicine, Nashville, TN, for sur-gical consultation. This work was supported by grants from theMedical Research Council (Canada), Canadian Diabetes Associa-tion, and Juvenile Diabetes Foundation International. K.Y. wassupported by a Canadian Diabetes Association Research Fellowship.M.V. is a Killam Scholar of Canada Council.

1. Wolfe, R. R., Durkot, M. J., Allsop, J. R. & Burke, J. F. (1979)Metabolism 28, 1031-1039.

2. Clifton, G. L., Ziegler, M. G. & Grossman, R. G. (1981) Neu-rosurgery 8, 10-14.

3. Halter, J. B., Pflug, A. E. & Porte, D., Jr. (1977) J. Clin.Endocrinol. Metab. 45, 936-944.

4. Cryer, P. E. (1980) N. Engl. J. Med. 303, 436-444.5. Christensen, N. J. & Videbaek, J. (1974) J. Clin. Invest. 54,

278-286.6. Kemmer, F. W., Bisping, R., Steingruber, H. J., Baar, H.,

Hartmann, F., Schlagheche, R. & Berger, M. (1986) N. Engl.J. Med. 314, 1078-1084.

7. Vranic, M., Kawamori, R., Pek, S., Kovacevic, N. & Wren-shall, G. A. (1976) J. Clin. Invest. 57, 245-255.

8. Vranic, M. & Wrenshall, G. A. (1969) Endocrinology 85, 165-171.

9. Gray, D., Lickley, L. & Vranic, M. (1980) Diabetes 29,600-609.

10. Perez, G., Kemmer, F. W., Lickley, H. L. A. & Vranic, M.(1981) Am. J. Physiol. 241, E328-E335.

11. Kemmer, F. W., Lickley, H. L. A., Gray, D. E., Perez, G. &Vranic, M. (1982) Am. J. Physiol. 242, E428-E436.

12. Brown, M. R., Rivier, J. & Vale, W. (1979) Endocrinology 140,1709-1715.

13. Sebel, P. S. (1987) in Hazards and Complications of Anaes-thesia, eds. Taylor, T. H. & Major, E. (Churchill Livingstone,New York), pp. 101-106.

14. DaRif, C. A. & Rush, H. G. (1983) Lab. Anim. Sci. 33, 90-94.15. Radosevich, P. M., Lacy, D. B., Brown, L. L., Williams,

P. E. & Abumrad, N. N. (1989) Am. J. Physiol. 256, E322-E330.

16. Bjorkman, O., Miles, P., Wasserman, D., Lickley, L. &Vranic, M. (1988) J. Clin. Invest. 81, 1759-1767.

17. Finegood, D. T. & Bergman, R. N. (1983) Am. J. Physiol. 244(7), E472-E479.

18. Steele, R. (1959) Ann. N. Y. Acad. Sci. 82, 420-430.19. Radziuk, J., Norwich, K. H. & Vranic, M. (1978) Am. J.

Physiol. 234, E84-E93.20. Radziuk, J. & Lickley, H. L. (1985) Diabetologia 28, 315-322.21. Brown, M. R. & Fisher, L. A. (1983) Brain Res. 280, 75-79.22. Brown, M. R. (1981) Diabetologia 20, 299-304.23. Brown, M. R. & Fisher, L. A. (1984) Am. J. Physiol. 247,

E41-E46.24. El-Tayeb, K. M. A., Brubaker, P. L., Lickley, H. L. A., Cook,

E. & Vranic, M. (1986) Am. J. Physiol. 250, E236-E242.25. Wasserman, D. H., Lickley, H. L. A. & Vranic, M. (1984) J.

Clin. Invest. 74, 1404-1413.26. Wasserman, D. H., Lickley, H. L. A. & Vranic, M. (1985) J.

Appl. Physiol. 59 (4), 1272-1281.27. Wasserman, D. H., Lickley, H. L. A. & Vranic, M. (1985) J.

Appl. Physiol. 59 (4), 1282-1289.28. Esler, M., Jennings, G., Leonard, P., Sacharias, N., Burke, F.,

Johns, J. & Blombery, P. (1984) Acta Physiol. Scand. 527,11-16.

29. Woods, S. C., Taborsky, G. J. & Porte, D., Jr. (1986) in TheNervous System, eds. Mountcastle, V. B., Bloom, F. E. &Geiger, S. R. (Am. Physiol. Soc., Bethesda, MD), Vol. 4, pp.365-411.

30. Gauthier, C., El-Tayeb, E., Vranic, M. & Lickley, H. L. A.(1986) Am. J. Physiol. 250, E393-E401.

31. Vranic, M., Radziuk, J. & Cherrington, A. (1973) in Regulationand Control in Physiological Systems, eds. Iberall, A. S. &Guyton, A. C. (Instrum. Soc. Am., Pittsburgh), pp. 487-490.

32. Rizza, R. A., Cryer, P. E., Haymond, M. W. & Gerich, J. E.(1980) J. Clin. Invest. 65, 682-689.

33. Rizza, R., Haymond, H., Cryer, P. & Gerich, J. (1979) Am. J.Physiol. 237, E356-E362.

34. Challis, R. A. J., Vranic, M. & Radda, G. K. (1989) Am. J.Physiol. 256, E129-E137.

35. Hagg, S. A., Taylor, S. I. & Ruderman, N. B. (1976) Biochem.J. 158, 203-210.

36. Miles, P. D. G., Yamatani, K., Lickley, H. L. A. & Vranic,M. (1988) Diabetologia 31 (7), 522A (abstr. no. 339).

Proc. Natl. Acad. Sci. USA 88 (1991)

Dow

nloa

ded

by g

uest

on

May

3, 2

021