ORIGINAL ARTICLE

Ioannis G. Fatouros á Allan H. GoldfarbAthanasios Z. Jamurtas á Theodore J. AngelopoulosJiaping Gao

Beta-endorphin infusion alters pancreatic hormone and glucose levelsduring exercise in rats

Accepted: 26 February 1997

Abstract b-Endorphin (BE) infusion at rest can in¯u-ence insulin and glucagon levels and thus may a�ectglucose availability during exercise. To clarify the e�ectof BE on levels of insulin, glucagon and glucose duringexercise, 72 untrained male Sprague-Dawley rats wereinfused i.v. with either: (1) BE (bolus 0.05 mg á kg)1

+0.05 mg á kg)1 á h)1, n = 24); (2) naloxone (N, bolus0.8 mg á kg)1 + 0.4 mg á kg)1, n = 24); or (3) volume-matched saline (S, n = 24). Six rats from each groupwere killed after 0, 60, 90 or 120 min of running at22 m á min)1, at 0% gradient. BE infusion resulted inhigher plasma glucose levels at 60 min [5.93 (0.32) mM]and 90 min [4.16 (0.29) mM] of exercise compared to S[4.62 (0.27) and 3.41 (0.26 mM] and N [4.97 (0.38) and3.44 (0.25) mM]. Insulin levels decreased to a greaterextent with BE [21.5 (0.9) and 18.3 (0.6) uIU á ml)1] at60 and 90 min compared to S [24.5 (0.5) and 20.6(0.6) uIU á ml)1] and N [24.5 (0.4) and 21.6 (0.7) uIU áml)1] groups. Plasma C-peptide declined to a greaterextent at 60 and 90 min of exercise with BE infusioncompared to both S and N. BE infusion increased glu-cagon at all times during exercise compared to S and N.These data suggest that BE infusion during exercise in-¯uences plasma glucose by augmenting glucagon levelsand attenuating insulin release.

Key words Naloxone á Insulin á Glucagon á C-peptide

Introduction

b-endorphin (BE) can be elevated in the circulation inresponse to stressful conditions including exercise ofsu�cient intensity and duration (Farrell et al. 1987;Goldfarb et al. 1990, 1991). The signi®cance of BE ele-vation during exercise remains unclear but BE has beenimplicated in the glucoregulatory response both at rest(Fatouros et al. 1995; Feldman et al. 1983; Giuglianoet al. 1987; Ipp et al. 1978; Reid et al. 1981) and duringexercise (Hickey et al. 1994).

Hypoglycemia induced by diet has been shown toincrease both hypothalamic as well as circulating BElevels (Fatouros et al. 1995). This response was accom-panied by a decreased insulin level and an increasedglucagon level. BE infusion into the periphery at restresulted in hyperglycemia in animals (Matsumura et al.1984; Rudman et al. 1983) and in humans (Feldmanet al. 1983; Giugliano et al. 1988; Reid et al. 1981). BE-induced hyperglycemia at rest was accompanied by anelevation in plasma glucagon (Giugliano et al. 1989;Paolisso et al. 1987; Reid et al. 1981) and either an in-crease (Feldman et al. 1983; Reid et al. 1981) or a de-crease (Giugliano et al. 1987; Matsumura et al. 1984) inthe level of insulin. The discrepancy of BE e�ects oninsulin has been postulated to result from di�erences inthe experimental species, types of infusion and dosesadministered, and, most importantly, the plasma glucoseconcentration at the time of the manipulation (Giugl-iano et al. 1989).

Opioid antagonism with naloxone (N) during exercisehas generally failed to lead to alterations in plasmaglucose, glucagon and insulin levels (Bramnert 1988;Farrell et al. 1988; Grossman et al. 1984; Staessen et al.1988). In contrast, N infusion was reported to inducehyperglycemia at 90 min of exercise despite a signi®cantelevation of the level of plasma C-peptide (Hickey et al.1994). It is unclear why hyperglycemia occurred in thisstudy. Not all of the e�ects of BE on glucose and pan-creatic hormones appear to be blocked by N (Feldman

Eur J Appl Physiol (1997) 76: 203±208 Ó Springer-Verlag 1997

I.G. Fatouros á A.H. Goldfarb (&) á A.Z. JamurtasUniversity of North Carolina at Greensboro,Department of Exercise and Sports Science,250 HHP Building, Greensboro, NC 27412, USA

T.J. AngelopoulosUniversity of Southern Mississippi,Laboratory of Applied Physiology, Hattiesburg,Miss. 39406 Jiaping Gao, Washington University Medical School,St. Louis, MO 63310, USA

et al. 1983; Giugliano et al. 1988, 1989; Rudman et al.1983). The e�ects of BE infusion on glucose, insulin andglucagon levels during prolonged exercise has not beenexamined; neither has a comparison between the e�ectsof BE infusion and N infusion during exercise beenmade.

Therefore, the purposes of the present investigationwere: (1) to examine the e�ect that BE infusion has onplasma glucose, insulin and glucagon levels duringprolonged exercise; and (2) to compare the e�ects of BE,N and S infusion on plasma glucose, insulin and glu-cagon levels during prolonged exercise.

Methods

Animal care and experimental design

Untrained, male Sprague-Dawley rats (n = 72), weighing 275±300 g, purchased from Harlan, Ind., were used in this study. Ratswere housed individually in hanging cages within a room controlledfor temperature 22 (1)°C, humidity 40 (2)%, and light-dark cycle(lights on for 0600±1800 hours). Food and water were providedadlibitum with the intake of food determined daily. This study wasapproved by the University Institutional Animal Care and UseCommittee prior to any experimentation and followed the guide-lines of the Institute of Animal Resources of the National ResearchCouncil.

Rats were randomly assigned to one of the three treatmentgroups: (1) saline infusion group (S), (2) BE infusion group (BE), or(3) naloxone infusion group (N). Animals were familiarized withthe treadmill, running daily for 10 min at 22 m á min)1 at 0° gra-dient for 4 days. One day after the last familiarization bout animalswere anesthetized by an intramuscular injection of a ketamine(60 mg á kg)1) and xylazine solution (2 mg á kg)1) (Sigma). Cath-eters were composed of 20-gauge polyethylene (PE 20) tubing andwere channeled subcutaneously through the skin into the rightjugular vein and advanced 5±10 mm into the vessel. Catheters werewashed with 0.2 ml of normal saline (0.9%) and ®lled with he-parinized saline (100 U á ml)1). The placement of the catheter wascon®rmed by withdrawing blood into it. Catheters were capped andpatency maintained daily with normal saline. Rats received60 mg á kg)1 i.m. penicillin after the surgery and were allowed torecover for 5 days. The rats then went through a second familiar-ization running period for 4 days.

Infusion experiments took place 1 day after the last familiar-ization bout. Animals were fasted for 7±8 h prior to infusion. Theinfusions took place between 0700 and 0900 hours in order to avoiddiurnal variations of the major glucoregulatory hormones. Cathe-ters were exposed and connected to 1-ml syringes with 27-gaugeneedles attached and placed within a Harvard Apparatus infusionpump. Rat synthetic BE (Sigma) was administered by constantintravenous (i.v.) infusion at 0.05 mg á kg)1 á h)1 [dissolved in0.280 ml of 0.9% saline solution (total volume)] after an initial i.v.bolus of 0.05 mg á kg)1 (in 0.1 ml of 0.9% saline). Volume-mat-ched N (Sigma), dissolved in 0.9% saline, was infused i.v. at0.4 mg á kg)1 á h)1 after an initial bolus of 0.8 mg á kg)1 into the Ngroup. Volume-matched 0.9% saline was infused into the S group.The bolus injection was administered 5±7 min prior to the initiationof the infusion. In each group, six animals were killed by decapi-tation at rest and after 60, 90, and 120 min of running. Also ani-mals were infused with either BE (n = 3) or N (n = 3) for 120 minand killed at rest and used as a nonexercise, infused controls.

Animals ran on a motor-driven treadmill at 22 m á min)1 at 0°gradient. During the last 15 min of the 2-h run the speed was in-creased to 28 m á min)1. The rats were motivated to run by a mildelectric current at the rear of the treadmill for the ®rst 15±20 min ofthe exercise bout.

Analytical techniques

After decapitation, mixed arteriovenous blood was collected intochilled tubes containing Na2EDTA. Blood samples were immedi-ately centrifuged (1400 � g for 20 min) and plasma was collectedand stored at )90°C until analyzed. Plasma insulin concentrationswere determined using a commercially available radioimmunoassay(RIA) kit (ICN, Calif., USA). The sensitivity of the assay was5 uIU á ml)1 and the inter- and intra-assay coe�cients of variationwere <10%. The plasma C-peptide concentrations were deter-mined using a standard double-antibody RIA (Linco Research,Mo., USA). The sensitivity of the plasma C-peptide assay was30 pM and the intra-assay coe�cient of variation was <15%.Plasma glucagon concentrations were measured using a commer-cially available RIA kit (ICN). The sensitivity of the glucagon assaywas 20 pg á ml)1 and the intra-assay coe�cient of variation was<15%. Plasma BE concentrations were measured by a double-antibody RIA (INCSTAR, Calif., USA) after extraction usingcolumn chromatography procedures. The sensitivity of the assaywas 3 pmol á l)1 and the intra-assay coe�cient of variation was<5%. Plasma glucose was determined by a glucose oxidase method(Sigma Diagnostics, Mo., USA). Plasma lactate concentrationswere measured enzymatically (Sigma Diagnostics). All assays wereperformed in duplicate.

Statistical analysis

The data were analyzed using a three by four analysis of variance(treatment by time). Signi®cant e�ects were identi®ed by using aNeuman-Keuls post-hoc procedure. The level of signi®cance wasset at P < 0.05. All data are presented as means (SE).

Results

No signi®cance di�erences occurred in weight gain orfood intake comparing the groups at any time.

Plasma BE

BE infusion signi®cantly elevated plasma BE concen-tration at all times compared to those of S and Ngroups. Plasma BE concentration increased signi®cantlywith time in all three groups in response to exercise(Table 1) reaching peak levels at 120 min of exercise.

Table 1 Plasma Beta-Endorphin concentration in response to ex-ercise during infusion of either Saline, Beta-Endorphin, or Nalox-one

Treatment Time (Mins)

0 60 90 120

Saline 7.7 21.0b 28.4bc 38.3bcd

�0.9 �0.9 �0.8 �1.5b-endorphin 17.8a 36.7ab 43.6abc 57.6abcd

�1.0 �0.7 �1.3 �6.3Naloxone 7.2 20.2b 27.6bc 36.5bcd

�1.0 �0.8 �0.6 �1.2

Values are means � SE in pmol á L)1 with an N = 6 subjects pertreatment time. a = signi®cant di�erence P < 0.05 vs both Salineand Naloxone treatments. b = signi®cant P < 0.05 vs 0 time.c = a signi®cant di�erence P < 0.05 vs 60 min time, d = sig-ni®cant di�erence P < 0.05 vs 90 min time

204

The BE infusion group maintained 15±20 pmol á l)1

higher levels of plasma BE during exercise compared tothe other groups. The S and N groups had similar BElevels (Table 1). Infusion of BE for 120 min independentof exercise increased plasma BE to 45 pmol á l)1.

Plasma glucose

The statistical analysis indicated that there was a treat-ment and time e�ect but no interaction with plasmaglucose. The BE group tended to have higher plasmaglucose at rest but this was not a signi®cant di�erence.BE infusion resulted in signi®cantly higher plasma glu-cose at 60 and 90 min of exercise compared to the othergroups. Glucose levels decreased in all three groups overtime. The S and N groups showed similar decreases inplasma glucose over time during exercise [S = 5.57(0.14) mM at rest vs 2.91 (0.20 mM) at 120 min,N = 5.54 (0.14) mM at rest vs 3.01 (0.19) mM at120 min]. BE infusion delayed the plasma glucose de-cline until 90 min [6.14 (0.23) mM at rest vs 4.15(0.29) mM at 90 min]. There were no di�erences inplasma glucose among the groups at 120 min of exercise(Fig. 1A).

Plasma lactate

Lactate levels were similar at rest in all three groups.Plasma lactate values increased signi®cantly in allgroups during exercise (Fig. 1B) with the highest re-corded levels at 60 min. Plasma lactate levels in the BEgroup demonstrated signi®cantly higher lactate valuescompared to the S and N groups at 60, 90, and 120 minof exercise. There were no di�erences between the S andN groups.

Plasma insulin

Insulin levels were similar at rest in all three groups.Insulin levels decreased signi®cantly with time in allgroups (Fig. 2A). The BE group had signi®cantly lowerplasma insulin values compared to the other two groupsat 60 min, and the level was lower than that of the Ngroup at 90 min. By 120 min of exercise plasma insulinlevels were equally depressed in all three groups.

Plasma C-peptide

No di�erences in plasma C-peptide levels were apparentat rest comparing the three groups. Plasma C-peptidedeclined signi®cantly after 60 min of running in all threegroups and continued to decrease with time (Fig. 2B).BE infusion resulted in signi®cantly lower plasmaC-peptide values compared to the other two groups at 60and 90 min of exercise.

Plasma C-peptide/insulin

No di�erences occurred in the C-peptide/insulin ratio atrest, independent of treatment. The C-peptide/insulinratio decreased signi®cantly in all three groups until90 min of exercise (Table 2). Thereafter, the ratios re-mained depressed independent of treatment. The BEinfusion group had a signi®cantly lower C-peptide/in-sulin ratio compared to the other groups at 90 and120 min of exercise (Table 2).

Plasma glucagon

There were no signi®cant di�erences in plasma glucagonat rest independent of treatment. Plasma glucagon sig-ni®cantly increased by 60 min of exercise and continuedto rise over time in all groups (Fig. 2C). BE infusionexacerbated the increase in plasma glucagon levelscompared to the other two groups at 60 and 90 min ofexercise. BE infusion resulted in higher plasma glucagon

Fig. 1 The data for glucose (A) and lactate (B) concentration for thethree treatment groups over time. Values are means (SE). *Signi®cantdi�erence between b-endorphin-infused group (BE) group comparedto saline- and naloxone-infused groups

205

values compared to S but not N and 120 min of exercise(Fig. 2C).

Discussion

This investigation reports the ®rst attempt to examinethe role of BE in the regulation of glucose during exer-cise by directly infusing BE into the circulation andcomparing the results to those obtained from bothS-infused and N-infused animals. Previous studies that

have examined the role of BE in determining glucoselevels during exercise have only infused opioid antago-nists such as naloxone (Bramnert 1988; Farrell et al.1986; Grossman et al. 1984; Hickey et al. 1994) andcompared the results to controls. These studies generallyreport that opioid antagonism does not signi®cantlyin¯uence glucose levels. Our results using N con®rm thatopioid antagonism does not appreciably alter glucoseconcentration and neither does it a�ect insulin or glu-cagon levels compared to those levels of an S-infusedcontrol rat. However, this study clearly demonstratesthat BE infusion, which elevates plasma BE concentra-tion to a greater extent than exercise alone, in¯uencesglucose, insulin, and glucagon concentrations.

Continuous infusion of BE at a rate similar to that ofthis study (0.05 mg á h)1) in humans was shown toslightly increase plasma glucose levels, not to a�ect in-sulin and to increase glucagon levels (Giugliano et al.1987). In this report, infusion of BE at a higher dose of0.5 mg á h)1 increased insulin, glucagon and glucoselevels, but no report of the BE level was mentioned. Wechose the lower dose of BE to infuse in the present studyto prevent the rise in insulin. In the present study, BEinfused for 120 min without exercise (n = 3) resulted ina sixfold to sevenfold increase from the basal restinglevels in our animals. A similar response was reported tooccur in humans (Giugliano et al. 1987). During exer-cise, BE increased in the S and N groups approximately®ve fold. This is similar to the increase reported to occurin humans in response to moderate-intensity exercise(Goldfarb et al. 1990, 1991).

BE infusion during exercise resulted in an eight foldincrease in BE compared to resting levels. Despite thefact that the BE-infused group constantly demonstrateda 15±20 pmol á l)1 higher BE level than the S and Ngroups, the exercise plus the BE infusion response didnot appear to be completely additive. This suggests thateither there was some inhibition of endogenous BE re-lease or enhanced BE degradation or a combination ofboth factors in the BE-infused, exercising animals. It wasreported that an infusion of BE from 0.3 to3.0 lg á kg)1 á min)1 in humans inhibited the release ofadrenocorticotrophin (ACTH) (Taylor et al. 1983). This

Fig. 2 The data for insulin (A), C-peptide (B), and glucagon (C) forthe three treatment groups over time. Values are means (SE).*Signi®cant di�erence between BE group compared to saline- andnaloxone-infused groups

Table 2 C-Peptide/Insulin response during exercise in response toeither infusion of Saline, Beta-Endorphin or Naloxone

Treatment Time (Mins)

0 60 90 120

Saline 2.16 1.32b 0.95bc 0.90bc

�0.20 �0.10 �0.10 �0.10b-endorphin 2.11 1.12b 0.44abc 0.49abc

�0.10 �0.10 �0.10 �0.10Naloxone 2.09 1.33b 0.93bc 0.91bc

�0.10 �0.10 �0.10 �0.10

Values are means � SE with an N = 6 subjects per each treatmenttime. a = signi®cant di�erence P < 0.05 vs both Saline and Na-loxone treatments. b = signi®cant di�erence P < 0.05 comparedto 0 time. c = signi®cant di�erence P < 0.05 vs 60 minute time

206

suggests that the elevation of BE may have had an in-hibitory e�ect on the exercise-induced stimulation ofBE. It was not the aim of this study to determine the fateof BE during exercise. It is possible that exercise mayalso in¯uence BE pharmacokinetics as compared tothose under resting conditions.

The plasma glucose concentration was at a higherlevel after BE infusion at 60 and 90 min compared to theother two treatments. The hyperglycemic e�ect of BEhas been noted to occur at rest in numerous human andanimal studies (Feldman et al. 1983; Ipp et al. 1978;Matsumura et al. 1984; Rudman et al. 1983). This is the®rst study to our knowledge that has reported that BEinfusion during exercise can manifest a higher glucoselevel than that in control S-infused animals. Addition-ally, the N-infused group failed to demonstrate changesin glucose concentration compared to the S-infusedgroup at any time point of exercise. Opioid antagonismwith N has primarily failed to demonstrate a role for BEin glucose regulation in most cases (Bramnert 1988;Farrell et al. 1986; Grossman et al. 1984; Staessen et al.1985, 1988). In contrast, a recent study noted that Nadministration resulted in a relative hyperglycemiaduring a 90-min running bout (Hickey et al. 1994). Thepresent study does not support this ®nding. It has beenpostulated that N may not block all of the e�ects of BE(Feldman et al. 1983).

In most cases, plasma glucose during prolonged ex-ercise declines after 60±90 min (Galbo et al. 1975). BEinfusion during exercise resulted in a greater decline ininsulin and an enhanced glucagon concentration. Thecombination of these factors would normally help toprevent plasma glucose levels from declining during ex-ercise. BE infusion at rest in humans was shown to de-crease the insulin concentration and to reduce the rate ofdisappearance of glucose from the blood (Giuglianoet al. 1987). It is probable that this same e�ect wouldhave occurred in our exercised rats.

The e�ect of BE to give rise to a relatively higherglucose concentration during exercise, as observed inthis study, and at rest, as observed in previous studies,can be attributed to either an increase in glucose ap-pearance, a decrease in glucose disappearance or acombination of these factors. Glucose production ratherthan glucose utilization was more likely to be stimulatedby either central (Matsumura et al. 1984; Paolisso et al.1987) or peripheral (Allan et al. 1983; Leach et al. 1985;Matsumura et al. 1984) BE administration. However,two other studies could not con®rm these results (Bru-baker et al. 1987; El-Tayeb et al. 1985). If BE stimulatesglucose production, then this could be achieved througheither hepatic glycogenolysis and/or gluconeogenesis.Opioid peptides have been shown to induce glycogen-olysis in the liver and this e�ect was cAMP related(Allan et al. 1983; Matsumura et al. 1984). Additionally,BE was able to stimulate gluconeogenesis in liver fromL-lactate (Matsumura et al. 1984). Plasma lactate wassigni®cantly elevated in the BE group compared to the Sand N groups in the present study. This may have

contributed, in part, to the elevated plasma glucose levelin the BE group. Further research is needed to con®rm ifthis was a contributory factor.

The rate of glycogenolysis in the liver could havecontributed to the enhanced glucose levels observed inthis study. Liver glycogenolysis is mainly regulated bythe combined drop in insulin and the increase in gluca-gon (Farrell 1992; Galbo 1992). In this study BE infu-sion augmented glucagon levels compared to the othertwo treatments during exercise and exacerbated the re-duction in insulin levels at 60 and 90 min of exercise.This combined action on these two hormones probablycontributed to the higher glucose levels observed at 60and 90 min of exercise in the BE group.

It was reported that the fall in insulin relieves theinhibition of hepatic glycogenolysis, and that this canaccount for as much as 65% of the increased hepaticglucose output (Wasserman et al. 1992). In addition, thefall in insulin sensitizes hepatic tissue to glucagon, suchthat the same absolute level of this hormone elicits amore potent response (Wasserman et al. 1992). It hasbeen found that the elevation in glucagon primarily in-creases the e�ciency of hepatic gluconeogenesis, andthat this accounts for almost 40% of the rise in hepaticglucose output (Farrell 1992; Galbo 1992; Kjaer et al.1987). The results in this study appear to suggest thatthese processes were primarily involved since plasmaglucose increased signi®cantly in the BE group at a timewhen plasma insulin and C-peptide levels decreased andglucagon increased compared to the other two groups.

Another possibility for the BE-induced hyperglyce-mia during exercise is a decreased utilization of glucoseby the working muscles. BE may have a lipolytic e�ecton adipose tissue at rest (Richter et al. 1983) and duringexercise (Vettor et al. 1987). Increased utilization of fattyacids during prolonged exercise may have spared glucoseutilization (Vranic et al. 1976). However, the increasedplasma lactate concentrations with BE infusion observedin this study suggest a higher glycolytic ¯ux.

In conclusion, the results from this study suggest thatBE is somewhat involved in the glucoregulatory mech-anism during prolonged exercise. BE probably a�ectsboth insulin and glucagon to alter glucose levels. How-ever, further research is needed in order to determine theexact mechanism(s) of BE action during exercise.

Acknowledgements This paper was supported in part by a SusanStout Fellowship and a grant from Sigma Xi as well as funds fromthe Dept of Exercise and Sport Sciences. We would like to thankDr. T.P. Hicks who graciously donated several pieces of equipmentto enable this study to come to fruition.

References

Allan EH, Green IC, Titheradege MA (1983) The stimulation ofglycogenolysis and neoglucogenesis in isolated hepatocytes byopiate peptides. Biochem J 216:507±510

Bramnert M (1988) The e�ect of naloxone on blood pressure, heartrate, plasma catecholamines, renin activity and aldosterone,following exercise in healthy males. Regul Pept 22:295±301

207

Brubaker PL, Sun A, Vranic M (1987) Lack of e�ect of b-endor-phin on basal and glucagon-stimulated hepatic glucose pro-duction in virtro. Metabolism 36:432±437

El-Tayeb KMA, Brubaker PL, Vranic M, Lickley HLA (1985)Beta-endorphin, modulation of the glucoregulatory e�ects ofrepeated epinephrine infusion in normal dogs. Diabetes34:1293±1300

Farrell PA (1992) Exercise e�ects on regulation of energy metab-olism by pancreatic and gut hormones. In: Lamb DR, Gisol®CV (eds) Perspectives in exercise and sport science, vol 5. En-ergy metabolism in exercise and sport, Brown and BenchmarkDubuque, pp 383±434

Farrell PA,Mikines KJ, Bach FW, Sonne B, GalboH (1986) Plasmab-endorphin immunoreactivity: e�ects of sustained hyperglyce-mia with and without prior exercise. Life Sci 39:965±971

Farrell PA, Kjaer M, Bach FW, Galbo H (1987) b-Endorphin andadrenocorticotropin response to supramaximal treadmill exer-cise in trained and untrained males. Acta Physiol Scand130:619±625

Farrell PA, Sonne B, Mikines K, Galbo H (1988) Stimulatory rolefor endogenous opioid peptides on postexercise insulin secre-tion in rats. J Appl Physiol 65:744±749

Fatouros JG, Goldfarb AH, Jamurtas AZ (1995) Low carbohy-drate diet induces changes in central and peripheral beta-en-dorphins. Nutr Res 15:1683±1694

Feldman M, Kiser R, Unger R, Li C (1983) b-Endorphin and theendocrine pancreas. Studies in healthy and diabetic humanbeings. N Engl J Med 308:349±353

Galbo H (1992) Exercise physiology: humoral function. Sports SciRev 1:65±93

Galbo H, Holst JJ, Christensen NJ (1975) Glucagon and plasmacatecholamine response to graded and prolonged exercise inman. J Appl Physiol 38:70±74

Giugliano D, Cozzolino D, Salvatore T, Ceriello A, Torella R(1987) Dual e�ect of beta-endorphin on insulin secretion inman. Horm Metab Res 19:502±503

Giugliano D, Torella R, Lefebve PJ, D'Onofrio F (1988) Opioidpeptides and metabolic regulation. Diabetologia 31:3±15

Giugliano D, Cozzolino, D, Ceriello A, Salvatore T, Paolisso G,Torella R (1989) b-Endorphin and islet hormone release inhumans: evidence for interference with cAMP. Am J Physiol257:E361±E366

Goldfarb AH, Hat®eld B, Armstrong D, Potts J (1990) Plasmabeta-endorphin concentration: response to intensity and dura-tion of exercise. Med Sci Sports Exerc 22:241±244

Goldfarb AH, Hat®eld BD, Potts, J, Armstrong D (1991) Beta-endorphin time course response to intensity of exercise: e�ect oftraining status. Int J Sports Med 12:264±268

Grossman A, Bouloux P, Price P, Drury PL, Lam KSL, Turner T,Thomas J, Besser M, Sutton J (1984) The role of opioid pep-tides in the hormonal response to acute exercise in man. Clin Sci67:483±491

Hickey MS, Trappe SW, Blostein AC, Edwards BA, Goodpaster B,Craig BW (1994) Opioid antagonism alters blood glucose ho-meostasis during exercise in humans. J Appl Physiol 76:2452±2460

Ipp E, Dobbs RE, Unger RH (1978) Morphine and b-endorphinin¯uence the secretion of the endocrine pancreas. Nature276:190±191

Kjaer M, Secher NH, Bach FW, Galbo H (1987) Role of motorcenter activity for hormonal changes and substrate mobilizationin humans. Am J Physiol 253:R687±R695

Leach RP, Allan EH, Titheradge MA (1985) The stimulation ofglycogenolysis in isolated hepatocytes by opioid peptides. Bio-chem J 227:191±197

Matsumura M, Fukushima T, Saito H, Saito S (1984) In vivo andin vitro e�ects of b-endorphin on glucose metabolism in the rat.Horm Metab Res 16:27±31

Paolisso G, Guigliano D, Scheen A, Franchimont P, D'Onofrio F,Lefebvre P (1987) Primary role of glucagon release in the e�ectof b-endorphin on glucose homeostatis in normal man. ActaEndocrinol 115:161±169

Reid R, Sandler J, Yen S (1981) b-Endorphin stimulates the se-cretion of insulin and glucagon in humans. J Clin EndocrinolMetab 52:592±594

Richter W, Kerscher P, Schwandt P (1983) b-Endorphin stimulatesin vivo lipolysis in the rabbit. Life Sci 33:743±746

Rudman D, Bery C, Riedeburg C, Hollins B, Kutner M, DynnM, Chawla R (1983) E�ects of opioid peptides and opiatealkaloids on insulin secretion in the rabbit. Endocrinology112:1702±1710

Staessen J, Fiocchi R, Bouillon R, Fagard R, Lunen E, MoermanA, De Schaepdryver A, Amery A (1985) The nature of opioidinvolvement in the hemodynamic respiratory and humoral re-sponses to exercise. Circulation 72:982±990

Staessen J, Fiocchi R, Bouillon R, Fagard R, Hespel P, Linjen P,Moerman E, Amery A (1988) E�ects of opioid antagonism onthe hemodynamic and hormonal responses to exercise. Clin Sci75:292±300

Taylor T, Dluhy RG, Williams GH (1983) b-Endorphin suppressesadrenocorticotropin and cortisol levels in normal human sub-jects. J Clin Endocrinol Metab 57:592±596

Vettor R, Manno M, De Carlo E, Federspil G (1987) Evidence foran involvement of opioid peptides in exercise-induced lipolysisin rats. Horm Metab Res 19:282±283

Vranic M, Kawamori R, Wrenshall GA (1975) The role of insulinand glucagon in regulating glucose turnover in dogs duringexercise. Med Sci Sports Exerc 7:27±33

Vranic M, Ross G, Doi K, Lickley L (1976) The role of glucagon-insulin interactions in control of glucose turnover and its sig-ni®cance in diabetes. Metabolism [Suppl] 25:1375±1380

Wasserman DH, Mohr T, Kelly P, Lacy DB, Bracy B (1992) Im-pact of insulin de®ciency on glucose ¯uxes and muscles me-tabolism during exercise. Diabetes 41:1229±1238

208