Interaction of Fat-stimulated Gastric InhibitoryPolypeptide on Pancreatic Alpha and Beta Cell Function

C. A. VERDONK,R. A. RIZZA, R. L. NELSON, V. L. XV. Go, J. E. GERICH, andF. J. SERVICE, Endocrine Research Unit, Gastroenterology Research Unit,Departments of Medicine and Physiology, Mayo Medical School andMayo Clinic, Rochester, Minnesota 55901

A B S T RA C T Gastric inhibitory polypeptide (GIP) isconsidered to be the principal mediator of the entero-insular axis. A glucose-insulin clamp technique wasused to study the effects of differing blood glucoselevels on the insulinotropic and glucagonotropicactions of fat-stimulated GIP in seven healthy subjects,as well as the effect of physiologic hyperinsulinemiaon GIP secretion. Blood glucose levels were clampedfor 4 h at 43+2 mg/dl (hypoglycemic clamp), 88+1mg/dl (euglycemic clamp), and 141+2 mg/dl (hypergly-cemic clamp) in the presence of a constant insulin in-fusion (100 mU/kg per h).

Under hypoglycemic clamp conditions there was noincrease in C-peptide nor glucagon after Lipomul in-gestion, despite an increase of GIP of 51.7±8.7 ng/mlper 120 min. Under euglycemic clamp conditions,small and inconsistent increases in C-peptide andglucagon were observed after fat ingestion and a con-comitant increase of GIP of 35.2±9.4 ng/ml per 120min. Under hyperglycemic clamp conditions after fatingestion and a GIP increase of 24.0±5.7 ng/ml per 120min, C-peptide increased from 6.4±5 ng/ml to 11.0±1.1ng/ml (P < 0.01) but glucagon did not change. Thesefindings confirm that in healthy man GIP exerts itsinsulinotropic properties only under hyperglycemicconditions and indicate that GIP is not glucagonotropic.

Under euglycemic clamp conditions (plasma glu-cose, 89±1 mg/dl) and physiologic hyperinsulinemia(serum immunoreactive insulin, 137±3 ,uU/ml) GIPresponses to fat ingestion (39.7±9.8 ng/ml per 120 min)were not different from the GIP responses to fat inges-tion in the absence of hyperinsulinemia (39.7±11.1ng/ml per 120 min). Therefore, insulin under normogly-

This work was presented in part at the 39th Annual Meetingof the American Diabetes Association, Los Angeles, Calif., 11June 1979.

Address reprint requests to Dr. Service at the Mayo Clinic.Receivedfor publication 14 September 1979 and in revised

form 14 January 1980.

cemic conditions does not exert an inhibitory effecton fat-stimulated GIP secretion. The higher GIP re-sponse to oral fat in the hypoglycemic clamp, and thelower GIP response in the hyperglycemic clamp com-pared to the response in the euglycemic clamp suggestsan effect of glycemia itself on GIP secretion in thepresence of hyperinstulinemia.

INTRODUCTION

Gastric inhibitory polypeptide (GIP)' is considered tobe the gastrointestinal factor primarily responsible forthe greater plasma insulin response to oral compared toparenteral nutrient administration (1-3). Althoughingestion of carbohydrate, fat (4), and some amino acids(5) results in increased levels of GIP, increased levelsof insulin are observed only after oral carbohydrate(4) and amino acid (5) and not after oral fat (4). Theseobservations suggest that the insulin secretory respon-siveness of the beta cell to GIP is influenced bysubstrate or hormonal factors. There is in vitro evidencethat the insulinogenic effect of GIP is glucose depend-ent (2, 6). Studies in humans, however, have shownconflicting results. There is evidence, using a glucoseclamp technique, that the insulinotropic action of oralglucose-stimulated GIP (7) occurs only during hyper-glycemia, but also evidence that amino acid-stimulatedGIP is insulinotropic in the absence of hyperglycemia(5,8). The involvement of a glucose-dependent mecha-nism for the insulinotropic action of fat-stimulated GIPhas been reported from nonsteady-state conditions (8-10), but has not been investigated using a glucoseclamp.

Whether insulin released after nutrient ingestionalso regulates the secretion of GIP as part of a negativefeedback system is presently controversial. The re-duced GIP responses to oral fat observed after an intra-

'Abbreviation uised in this )ai)er: GIP, gastric inhibitorypolypeptide.

1119J. Clin. Invest. (© The American Society for Clinical Investigation, Inc. 0021-9738/80/05/1119/07 $1.00Volume 65 May 1980 1119-1125

Venous bolus of insulin or durinlg a coIncomitant infuisionof gIlucose supported an inhibitorx action of insulinon the secretioni of GIP (4, 8- 11). However, Andersenet al. (7), using a glucose-insulin clacmp technique ateuglycem-nia and hyperglycemia, fouind no feedbackinhib)ition of insulin on glucose-stimultlated GIP secre-tioll.

The eurrent studies were undertaken to examinewhether the level of glycemia moduilates the insulino-tropic effect of fat-stimulated GIP, to determinewhether insulin inhibits the secretion of fat-stimulatedGIP, anid to determine wvhether the level of glycemiaitself miav inifluence the GIP responise to oral fat. GIP,C-peptide, and glucagon responsies to fat ingestionNere imieastured in healthy subjects while glycemiia was

miiaintainied by a glucose-insulin clamiip techni(lue in thehypoglycemic, euglycemic, and hyperglycemic rangesdurinig a concomitant inftusion of insulin at a rate suf-ficienit to achieve physiologic hvperinssulinemnia.

METHODS

inormed conisenit was obtained fromil seveniioriiail nonobesesul)ects (three males, four females) ages 36±5 y r (mean ± SEM).All were within 10%of their ideal bodx weight and n1on1e had afhimilv history of'diabetes mellitis.

Each subject was stu(die(l in the ov-ernight f:aste(d state at eachglycemic clamiip level, and six of the seven subjects were stud-iedl dIurinig saline infusion in the absence of' glucose-insulillclamiip with and xvithout the ingestion of Lipomutl. Each studywas sep)arated by 1-2 wk.

For each glucose clamp studv 18-gauge indwelling catheterswere inserted into contralateral antecuibital veins, one forthe conttinuous inf'usioni of crystalline insulini (pork U100, EliLilly & Co., Indianapolis, Ind.), at the rate of 100 mU/kg per h1v imleanls of a Harvard pump (Harvard Apparatus, millis,NMass.) and one for the intermittent (every 20 min) wvithdrawalof' blood for the determination of hormonies. Distal to theinisilini inf'usion site in a separate f'orearmii vein, a double-lumiien caltheter was inserted for continutiouis withdrawal of)lood(at a rate of 2 ml/h for glucose analyN-sis by the Biostator.

The glucose clamp was achieved using the Biostator GCIIS(Life Science Instruments, Elkhart, Ind.), wvhich permits con-tinuous analysis and minuite-by-minute recording of plasmaglucose levels (glucose-oxidase) as well as the infuision ofglucose according to predetermined computer-containedalgorithlmls (12). Glucose (50 g/dl) was inf'used through theinsuilin infuision access site at rates determined by theBiostator (mode 7:1). In each subject the glucose clamp wasmaintainled at 45 mg/dl (hypoglycemic clamp), 140 mg/dl(hyperglycemic clamp), and at the basal overnight fastingplasmiia gluicose level (euiglycemic clamp). The gltucose in-fuision rate at the desired plasma gltucose level, determinedin prelininary studies, w as 0.97+0.06 mg/kg per min,6.0(.1 miig/kg per min, and 7.4±.1 mg/kg per mimi and the inverseof'the static gain for glucose inf'usion was 18, 45, and 45 for thehypoglycemic, euiglycemic, and hyperglycemiiic clampis, respec-tively. Becauise the maximal infusion rate of gltucose that canbe infuised by the Biostator is 1 g/min, additional glutcose forthe euglycemic and hyperglycemic clamiips was given by avariable-speed infusion Harvard puimp.

Each gltucose clamp study was conducted for a total of'240min. The first 120 min w.ere devoted to obtaining stableplasma glulcose and insulin conicentrations. Emutlsified cornoil

(Lipomul, Upjohn Co., Kalamazoo, Mich.), 67 g, was adminiis-tered orally at 120 min At 20-min intervals duiring the aclamps,glucose levels obtained by the Biostator w ere checked againistthe reference method, YSI 23A glucose analyzer, YellowSprings Instrumient Co., Yellow. Springs, Ohio. The Bio-stator glutcose values for the hyperglycemic anld euglycemicclamps were found to be consistently within ±O10% of'the YSIreadings. The mediani percent difference between the twomethods xwas 3.5%. For the hypoglycemic clamp the YSIglucose readinigs wvere consistently greater than the Biostatorvalues with a mnedian (lifference of 15%.

For the two stuidies wvithout the insulini-gluicose clampii), withand without the Lipomutl ingestion, blood samiples were oh-tained for 20-mimi intervals f'or glucose, insulin, and GIP deter-miniationis for 140 mi.

Sertum samples were f'rozen for insuilini assay. Blood samplesfor GIP and C-peptide were collectedl on ice in tubes conttain-ing EDTA andl Trasvlol (500 kallikreini inhibitor tuniits/Il;Sigmla Chemiiical Co., St. Louis, Io.) cenitrif'ugedl at 4°C afterwhich the plasma was frozen until assay. Blood for glucagonlwas processed similarly except for the use of benzamidine(0.1 M) instead of Trasvlol.

Hormonie assaq s. Plasmna GIP was measure(d bv the miiethodof' Kuzio et al. (13). Purified GIP, obtained fromii Dr. J. C.Brown (University of' British Columbia, \ancouver, BritishColtumbia, Canca(la) was used as standardl and( tracer. Anlti-serumii R4817 wxas used at a final dilution of 1:100,000. Thisantiserumil detects the tw ro molecular forms of imminuinoreactiveGIP (5,000 and 7,500 mol wt) present in postprandial blood.The limit of' detectioni, intraassav and interassav coefficientof variation for the plasmiia internal reference stanldardls were50 pg/ml, 7 anld 13%, respectively. No cross-reactivity wasdetected with glucagon (crystalline porcine glucagon, EliLilly & Co.), humiican pancreatic polvpeptide (Eli Lillx & Co.),highly purified cholecvstokinin, secretin, vasoactive intes-tinal peptide (all gifts from Dr. XI. Mlutt, Stockholmii, Sw,veden),motilin (Dr. J. C. Brow.n), and gastrin (Imperial ChemiiicalIndustries LTD, Londoni, England) in concenitrations uip to10 ng/ml.

Insulini and( gluicagoni were measutred 1w the miietho(ds ofHerbert et al. (14) and Faloona et al. (15), respectively.

Plasmaiit C-peptide wvas measured using the reagenits andprocedure obtained from Calbiochem-Behri ng Corp., Amiieri-can Hoechst Corp. (San Diego, Calif.). The characteristics ofthe rabbit aintiserumiii against synthetic hulman C-peptide havebeen descril)ed (16). 1211 synthetic tvrosvl C-peptide was tiusedas tracer, a.nd synthetic human C-peptide as staindatrd.

Arialhjtic nmethods. Data in the text and figtures are givenias mean±SE-M. The integrated plasma GIP response afteringestion of Lipomntul was calculated from the area eircuim-scribed by the curve (using the menan of the 100 and 120 mifor basal) in eaclh person and expressed as ig/mIl per 120 min.The coefficient of variation of the glucose claimps was caltcu-lated wvith the Biostator glucose vallues at 5-mimi intervalsduiring the 60-240-mimi period of each clamp.

Statistical evaluation was performed by m11eans of' the two-tailed paired t tests. The rank sum test was ised to compatrethe plasma GIP responses after Lipomutl ingestion tinder thevarious claimp conditionis, since these responses were nlotnormally distributed (17).

RESULTS

Hypoglycemic clamp. Fig. 1 shows the glucose,GIP, C-peptide, insulin, and glucagon levels beforeand after Lipomul ingestion during the hypoglycemicclamp. Glucose levels decreased progressively and

1120 Verdonik, Rizza, Nelsoni, Go, Gerich, atnd Service

HORMONERESPONSESDURINGHYPOGLYCEMICCLAMPBEFOREANDAFTERLIPOMUL (67g) INGESTION

INSULIN (O. IU/kq/h) +VARIABLE GLUCOSE

100 - GLUCOSE

i 80 LIPOMUL p.o.N\ ffO- t

6 40

20L

GASTRIC INHIBITORY ,POLYPEPTIDE

Soo.00 k

N400k-

200- INSULIN

N 500

-20 0 40 80 120 160 200 240

MEAN± SEM

n - 7

*P

HORMONERESPONSESDURINGEUGLYCEMICCLAMPBEFOREANDAFTERLIPOMUL (67g) INGESTION

INSULIN (0.1U/kg/h) + MEANt SEMVARIABLE GLUCOSE na 7

GLUCOSE *P

EFFECTOF INSULIN ONGASTRIC INHIBITORYPOLYPEPTIDERESPONSESTO FAT INGESTION

INSULIN 10.1 U/kg/h) +VARIABLE GLUCOSEor SALINE

LIPOMUL- 67 g p.o.GLUCOSE t

% 10090

E280180

0135

90_::

45

OL_960

720

480Q~

240

oL_

MEAN± SEMn-6

--o CONTROL*_ 'CLAMP'

* P

at euglycemia. Although the GIP levels 120 min afterLipomul were not returniing to base-line, they wereclose to a plateau configuirationi in the last 60 min. Theobserved levels likely represent maximal responses asthey are similar to those reported by others after fatingestion (8, 9, 11). In addition, further sampling be-yond 120 min was unlikely to have shown a differencein GIP between control and euglycemic clamp studiesbecause differences observed in GIP by others be-tween control aind glucose infusion studies occuirredbefore 120 min after fat or galactose ingestion. Theslight decreases in basal GIP during the 120-inin pre-Lipomul period in the presence aind absence of hyper-insulinemia may represent the effect of fasting on GIP.

The observation of a reduced GIP response after fatingestion during hyperglycemnia compared to theresponse after the same stimulus in the presence ofeuglycemia is consistent with previous reports (8-10,25), but is open to an interpretation different from anlinhibition of GIP by insulin alone. The increased GIPresponse to oral fat during hypoglycemia coimpared toeuglycemia couipled with the reduced GIP response inhyperglycemia (Fig. 3) in the presence of similar serumninsulin levels (144+3 ,uU/ml for hypoglycemiia, 141+2,uU/ml for etuglycemia, and 183±5 ,uU/ml for hypergly-cemia) stiggests that the glucose level itself in thepresence of hyperinsulinemia affects the GIP responseto oral fat. Whether this effect is on one or bothmolecuilar species remains to be determined.

In summary, fat-stimulated GIP has insulinotropicactivity that is glucose dependent, is not gluicagonotro-pic, is not inhibited by physiologic hyperinsuilinemia ateuglycemia, but is influenced by the ambient gluicoselevel in the presence of hyperinsulinemia.

ACKNOWLEDGMENTS

The excellent technical assistance of R. Westland, N. Reed,J. King, K. Greene, B. Brick, I). Stenner, and D. Nash is grate-fully acknowledged.

This investigation was supported in part by research grantsAM-20973, AM 20411, AMX 5827, and RR 00036 from theNational Institutes of Health, U. S. Public Health Service,and from the Mayo Foundation.

REFERENCES

1. Dupre, J., S. A. Ross, D. Watson, anid J. C. Brown. 1973.Stimulation of insulin secretion by gastric inhibitory poly-peptide in man.J. Clin. Endocrinol. Metab. 37: 826-828.

2. Pedersoni, R. A., J. C. Brown. 1976. The insulinotropicaction of gastric inhibitory polypeptide in the perftusedisolated rat pancreas. Endocrinologyv. 99: 780-785.

3. Creutzfeldt, W. 1979. The ineretini concept today. Diabet-ologia. 16: 75-85.

4. Brown, J. C., J. R. Dryburgh, S. A. Ross, and J. Dtupre.1975. Identification and actions of gastric inhibitory poly-peptide. Recent Prog. Horin. Res. 31: 487-532.

5. Thomlas, F. B., D. Sinar, E. L. Mazzaferri, S. Catalanid,H. S. M\rekhjian, J. H. Caldwell, and J. J. Fromkes. 1978.

Selective release of gastric inhibitory polypeptide b)yintra(illodenal aimino acid perfiIsion in man. G itstroeter-ologyJ. 74: 1261-1265.

6. Schauider, P., J. C. Brown, II. Frerichs, aId \XV. Creitzi-f'eldt. 1975. Gastric inhibitory polypeptide: eff'ect onglutcose-indcice(d insuilin release fromil isolaited rait paincre-atic islets in vitro. Diabetologia. 11: 483-484.

7. Andersen, D. K., 1. Elathi, J. C. Brown, J. D. Tobin, andeR. Anidres. 1978. Oral glucose augmentation of instilinsecretion. Interactions of gastric inhibitory polypepti(lewith ambiient glucose iand insuiiin levels. J. Cliti. In uest.62: 152-161.

8. Cleator, I. G. MI., (nd R. 1I. Gouirlav. 1975. Release ofimiminoreactive gaistric inhibitory polypeptide (IR-(GIP)by orail ingestioni of food substances. Arrm. J. Surg. 1:30:128-135.

9. Ross, S. A., and J. Dumpre. 1978. Eff'ects of inigestioni oftriglvceride or galactose on secretioni of'gastric inhibitorxp)olypeptide and(I un responises to intravenous glucose innormiial and dlial)etic suiibjects. Diabetes. 27: 327-333.

10. Crockett, S. E., S. Cataland, J. MI. Fatlko, and( E. L. Nlazza-f'erri. 1976. The insuilinotropic effect otf en(logenousgastric inhibitory p)olypeptide in normllal subjects.J. Cliii.Euidoctrituol. Metabl. 42: 1098-1103.

11. Ebert, R., H. Frerichs, atnd XV. Cretzfteldt. 1979. Impairedf`eedback contr-ol of'fat iniduiced gastric inh ibitorv I)olypeI)-tide secretion by instulin in obesity and(i glucose intolerance.Eur.J. Cliii. Iincest. 9: 129-135.

12. Pfeiffer, E. F., Ch. Thumln, and A. Hi. Clemiiens. 1974. Theartificial beta cell-a contintuouis control of'1)lood suigar 1yexternial regtulattioni of insutilin infuisioni (gltucose controlle(dinstilini infuisioni system). Horiii. Metab. Res. 6: 339-342.

13. Kuizio, NI., J. R. Dryburgh, K. MI. Mlalloy, and J. C. Brown.1974. Radioil.inmntlloassav for gastric inhibitorv polypep-tide. Gastroetnterology. 66: 357-364.

14. Herbert, N., K-S. Lati, C. XV7. Gottlieb, and S. J. Bleichelr.1965. Coated charcoal i mmnmnoassay of' inisuilin. J. CliGu.Endocrinol. Metal). 25: 1375-1384.

15. FalooIia, G., and R. Unger. 1974. Gltucagoni. Ini Mlethodsof'Hormone Radioimimunoassay. B. Jaffe and(I H. Behrmiiani,ec(litors. Academiiic Press, Inc. Newv York. 317-330.

16. Faber, 0. K., C. Binider, J. Mlarkussen, L. G. Heding,V'. K. N'aithani, fI. Kuiztuya, P. Blix, D. L. Hlorwitz, andA. H. Rtubenstei n. 1978. C:haracterizati on of'seven C-pep-tide anitiserat. Diabetes. 27(Suppl. 1): 170-177.

17. Dixoni, NV. J., F. J. Mlassey, 1969. Introdtuction to statistictalanalysis. NMcGraw-Hill Book Co., Inc. New 'York. 488 pp.

18. NIutller, XV. A., G. R. Faloonia, E. Aguilar-Parada, and(l R. H.Unger. 1970. Abnormal alpha cell futnctioni in diabetes:response to carbohydrate and proteini ingestion. N. Euigl.

J. Med. 283: 109-115.19. Bottger, I., R. Dobbs, G. R. Faloona, and R. H. Unger.

1973. The effects of triglyceride absorption tupoon gluca-goun, inisuilin anid guit gluicagoni-like imnmunoreactivitv. J.Clini. In;est. 52: 2532-2541.

20. Pederson, R. A., II. E. Schubert, and J. C. Browin. 1975.The instulinotropic ctetion of gastric inhibitory polypep-tide. Can. J. Physiol. Pharmlacol. 53: 217-223.

21. Elahi, D., D. K. Andersen, J. C. Brown, H. T. Debas,R. J. Hershcopf, C. S. Raizes, J. D. Tobin, aind R. Andres.1979. Pancreatic alpha and beta cell responses to (C11infuision in niormial mnani. Am. J. Phy.siol. 237(2): E1185-E191.

22. Pederson, R. A., and J. C. Brown. 1978. Interactiont ofgastric inhibitorxn polypeptide, glucose, and arginine onlinsuilin and glucagon secretion f'romn the perfuiserl ratpancreas. Entdocrinology 103: 610-615.

23. NleDonatld, T. J., J. Duipre, Y. Catissignate, J. Radzitik, and(

1124 Verdonik, Rizza, Nelsoni, Go, Geric h, anid Service

S. Van Vliet. 1979. Hyperglucagonemia in liver cirrhosiswith portal-systemic venous anastomoses: response ofp)lasina glutcagon and gastric inhibitory polypeptide to oralor intravenous glucose in eirrhotics with normnal or ele-vated plasma glucose levels. Mfetab. Clin. Exp. 28: 300-307.

24. Service, F. J., R. L. Nelson, A. H. Rubenstein, VI. L. W. Go.

1978. Direct effect of insulin on secretion of insulin, glu-cagon, gastric inhibitory polypeptide and gastrin duringmainteinanice of normoglycemia. J. Clinl. Endocrinology.Metab. 47: 488-493.

25. Morgan, L. MI., J. W. Wright, and V. NMarks. 1979. Theeffect of oral galactose on GIP and insulin secretion in man.Diabetologia. 16: 235-239.

Interaction of Fat-stimulated GIP otl Alpha and Beta Cell Fuinction 1125