The In Vitro Inhibition of Insulin

Secretion by Diphenyihydantoin

J. STEPHENKIZER, MARIOVARGAS-CORDON,KLAus BRENDEL, and RUBIN BRESSLER

From the Departments of Medicine and Physiology and Pharmacology, DukeUniversity Medical Center, Durham, North Carolina 27706

A B S T R A C T Glucose intolerance has been observedfollowing diphenylhydantoin (DPH) intoxication. Be-cause of this association between DPH and hypergly-cemia, the effect of DPHon insulin release in vitro usingpreparations of isolated islets of Langerhans and pan-creatic pieces was examined. In concentrations identicalwith those considered necessary for adequate anti-convulsant therapy in man, DPHmarkedly decreases theinsulin secretory response of pancreatic pieces to metha-choline, 1 ig/ml, tolbutamide, 250 I-g/ml, and glucose,200 mg/100 ml, without any demonstrable alteration inthe oxidative conversion of glucose-1-"C or glucose-6-14Cto 14CO2 by isolated islets. This DPH-induced inhibitionof insulin secretion is not reversed by higher concentra-tions of glucose (600 mg/100 ml) or by increasing con-centrations of extracellular calcium ion (4-6 mmoles/liter). 0.1 mmpotassium and 10' M ouabain, however,effectively restore the DPH-induced block of insulinsecretion in pancreatic pieces. 60 mmpotassium ion, onthe other hand, not only restores the insulin secretoryresponse to glucose (200 mg/100 ml) but results in anadded stimulation of insulin secretion in the presence ofDPH. In the presence of DPH, 2'Na accumulation by iso-lated islets is decreased by 26-40% as compared withcontrols. Such evidence is considered to indirectly sup-port the postulate that the electrophysiological propertiesof DPHon the pancreas are due to a stimulation of themembrane sodium-potassium-magnesium ATPase.

INTRODUCTIONIn 1965 an association between hyperglycemia and di-phenylhydantoin (DPH) was described in rabbits (1).This observation was soon followed by the clinical rec-ognition of hyperglycemia and nonketotic hyperosmolar

Dr. Bressler's present address is Department of Pharma-cology, College of Medicine, University of Arizona, Tucson,Ariz. 85721.

Received for publication 31 March 1970 and in revisedform 2 June 1970.

coma resulting from DPHintoxication (2, 3), particu-larly in the critically ill or azotemic patient (3). Theetiology of the DPH-induced hyperglycemia has beenclarified somewhat by the demonstration of a transientlyabnormal glucose tolerance test and low insulinogenic in-dex in a patient with DPHintoxication which fully re-solved after cessation of anticonvulsant therapy withDPHand reinstitution of treatment with phenobarbital(4). Millichap has studied the hyperglycemic effects ofDPHand phenobarbital in rabbits and has found signifi-cant elevations of blood glucose within 1-2 hr after in-traperitoneal administration of these agents (5), DPHbeing the more potent of the two drugs (5). From thesedata, Millichap has postulated that anticonvulsant therapymay stimulate the hypothalamus, resulting in hyperac-tivity of the sympathoadrenal system and subsequenthypoinsulinemia (5).

In this paper, studies are presented which furtherelucidate the mechanism of DPH-induced hyperglycemiaand hypoinsulinemia. Studies were carried out on the ef-fects of DPH on insulin secretion in preparations ofisolated islets of Langerhans and pancreatic pieces. Wehave found that DPHand mesantoin, another hydantoinanticonvulsant, are potent inhibitors of pancreatic in-sulin release, whereas phenobarbital and hydroxydi-phenylhydantoin, a metabolite of DPH lacking anticon-vulsant properties, have no insulin inhibitory effects.

METHODSAnimals. Adult male Sprague-Dawley rats were obtained

from Cammi Research, Norfolk, Va. Animals were fed adlibitum a diet consisting of water and Purina Lab Chowup to the time of their sacrifice. For the isolation of islets,rats weighing more than 450 g were used, whereas animalsweighing approximately 300 g were used for experimentsinvolving whole pancreatic pieces.

Buffer. Fresh buffers were made daily and containedthe following, unless otherwise specified: 100 mm sodiumchloride, 5 mmpotassium chloride, 2 mmcalcium chloride,1.25 mmKH2PO4, 1.25 mmmagnesium sulfate, 26.2 sodiumbicarbonate, 5 mmsodium pyruvate, 5.5 mmsodium fuma-

1942 The Journal of Clinical Investigation Volume 49 1970

rate, S mmsodium i-glutamate, and dextrose, 200 mg/100ml. After preparation buffers were gassed 10 min with a95: 5 mxture of oxygen and carbon dioxide. In those ex-periments requiring the measurement of insulin secretion,bovine serum albumin was added to a final concentration of0.2%. In those experiments requiring potassium in alteredconcentrations, NaCl and NaH2P04 were either substitutedfor or replaced by KC1 and KH2PO4as necessary. All solu-tions were kept stoppered and on ice prior to use.

Piece experiments. For all experiments with pancreaticpieces except experiment set II, three to four rats werestunned by a blow on the head followed by cervical disloca-tion. The splenic portions of the pancreases were immedi-ately removed and transferred to a glass Petri dish ofbuffer over ice. All visible fat and lymph nodes wereremoved, and the pancreas was cut into approximately 100pices, each weighing about 5 mg. We have found that theuse of smaller pieces markedly enhances the insulin outputper milligram of tissue. The pancreas pieces were random-ized, and approximately six to seven pieces were transferredto 10-cc Erlenmeyer flasks containing 2 ml of buffer andthe appropriate test drug After being gassed with a 95: 5mixture of oxygen and carbon dioxide, the flasks werestoppered and incubated in a Dubnoff shaker bath at 37'Cfor 90 min at 100 cycles per min. Anti-insulin antibody wasadded to the flasks to prevent the breakdown of insulinaccording to the method of Malaisse, Malaisse-Lague, andWright (6).

Isolated islets. Islets were isolated after the method ofLacey and Kostianovsky (7). For experiments in whichinsulin was to be measured, five islets were transferred toapothecary dram vials containing 0.75 ml of buffer and theappropriate test drug. After being gassed with a 95: 5 mix-ture of oxygen and carbon dioxide, the vials were stopperedwith serum stoppers and incubated for 90 min in a Dubnoffshaker bath at 37'C at 100 cycles per min. In these experi-ments, anti-insulin antibody was not added to the incubationmedium.

Oxidation studies. For the oxidation studies, 10 isletswere transferred to apothecary dram vials containing 0.5ml of buffer, a small glass bead, the appropriate test drug,and either glucose-1-'C or glucose-6-14C with a final specificactivity of 1.5 X 10' disintegrations per 1 mg of glucose.Each vial was gassed with a mixture of 95: 5 oxygen andcarbon dioxide and stoppered tightly with a serum stopperfrom which a polyethylene center well was suspended.Incubations were carried out in a Dubnoff shaker bath for20 min at 37'C and 150 cycles per min. The reactions wereterminated by the immediate injection of 0.2 ml of 4 NH2SO4, and the evolved "CO2 was collected at 30'C for60 min in a Dubnoff shaker bath in 0.1 ml of hydroxide ofhyamine (Packard) injected into the center well. All sam-ples were counted in a Packard liquid scintillation counterusing automatic external standardization. Results were ex-pressed as counts evolved as "CO2 per 10 islets per 20 min.

'Na accumulation studies. In three separate experiments,50 islets were transferred to each of two apothecary dramvials containing 0.75 ml of buffer and 'Na ranging in spe-cific activity from 107 to 2 X 107 disintegrations per minper 112 uEq of sodium. In each experiment, one vial servedas a control, and the second served as a test vehicle con-taining DPH in a concentration of 7.5 gg/ml. Each vialwas gassed with a 95: 5 mixture of oxygen and carbondioxide, stoppered tightly, and incubated in a Dubnoff shakerbath for 90 min at 37'C at 150 cycles per min. At the endof the incubation, each vial was rapidly filtered through a

Millipore filter, and the islets were rinsed with 4 ml ofbuffer. The Millipore filter disc and the adherent isletswere transferred with forceps to a Baird Atomic poly-ethylene counting tube (Baird Atomic, Inc., Bedford, Mass.)and counted in a Nuclear-Chicago gamma counter. The spe-cific activity and total counts in the original incubationmedium were determined by counting a 0.5 ml sample ofthe original 0.75 ml of medium plus 4.0 ml of rinse.Accumulation was then expressed as the fraction of thetotal 22Na counts remaining bound per 50 islets.

Insulin assay. Insulin was measured using a single anti-body, cellulose column assay developed in our laboratory'as a modification of the method of Malaisse et al (6).Standard curves were run with each day's determinations.

Expression of results. All data were compared using thestudent's t test for comparison of two samples, and 0.05was chosen as the accepted level of significance.

Sources. Diphenylhydantoin sodium was the gift ofParke, Davis and Co., and the guinea pig antiporcine insulinantiserum was provided by Dr. Peter Wright of the Phar-macology Department, University of Indiana. Porcine insu-lin for standards in the insulin assay was donated by EliLilly and Co. Pentobarbital sodium was purchased fromAbbott Laboratories, North Chicago, Ill. Mesantoin was agift from Sandoz, Inc., Hanover, N. J. Hydroxydiphenyl-hydantoin was synthesized by Dr. M. Bush of the Pharma-cology Department, Vanderbilt University, using the methodof Butler (8). Glucose-1-"`C and glucose-6-"C and 'Nawere purchased from New England Nuclear Corp., Boston,Mass. Insulin ('I) was purchased from Cambridge NuclearCorp., Cambridge, Mass. Radioactive 'Na was purchasedfrom New England Nuclear Corp.

RESULTS

Inhibition of insulin secretion by DPH. Usual thera-peutic levels of DPHin man range from 10 /Ag/ml plasmafor chronic anticonvalsant therapy (9) to 10-20 izg/mlfor acute termination of cardiac arrhythmia (10). Theeffect of DPH on glucose stimulated insulin secretionby pancreatic pieces is shown in Table I. 5 Pg/ml ofDPH did not significantly depress insulin secretion;however, in concentrations of DPH of 7.5, 20, and 40tug/ml, insulin secretion was markedly inhibited. Al-though both of the higher concentrations of DPHwere20% more effective than the 7.5 ,sg/ml concentration ofDPH, there was no significant difference among thethree as determined for a sample size of eight (P>0.05). On the basis of these studies, therefore, a concen-tration of 7.5 jsg/ml of DPH was selected for all fur-ther experimentation since it was the lowest concentra-tion which would routinely give a significant inhibitionof insulin release.

The effect of pentobarbital on insulin secretion. Be-cause of the possibility that depression of insulin releasemight be a common property of all anticonvulsants, theeffect of pentobarbital on insulin secretion was examinedsince it was routinely used to anesthetize rats from whichisolated iselets were obtained. Insulin secretion was as-

1 Brendel, K., B. Stocks, and R. Bressler. In preparation.

DPH Inhibition of Insulin Secretion 1943

TABLE IInhibition of Insulin Secretion by DPHin Pancreas Pieces*

Insulin secretionDPH mean ASE n Inhibition P

pg/ml IAU/100 mg of tissue %per 90 min

0 22,930 ±2783 8 - -5.0 16,792 41476 8 25 0.17.5 12,529 41788 8 40 0.5Pentobarbital 30 pg/ml

+glucose loo mg/100 ml 7442 :1:2071(5) - -

* Numbers in brackets indicate the number of incubations carried out. Theanesthetized rat received 6 mg/100 g i.p. pentobarbital 20 min beforesacrifice.

TABLE I I IComparison of the Effects of DPH, Hydroxy-DPH, andMesantoin on Insulin Secretion by Pancreatic Pieces*

Insulin secretionExperiment mean 4sr n P

AU/100 mg tissue/90 min

Control 20,658 ±1330 6DPH, 7.5,gg/ml 13,033 ±637 6 0.5

* Glucose concentration in media 200 mg/100 ml.

did 10 ug/ml of mesantoin, whereas hydroxydiphenyl-hydantoin in concentrations 10 times those of DPHfailed to inhibit insulin release, suggesting similar struc-tural requirements for the anticonvulsant and insulin in-hibitory properties of the DPHmolecule.

Effect of DPHon insulin secrtion by pancreatic islets.DPH markedly insulin secretion by isolated pancreaticislets. These data are shown in Table IV.

The effect of DPHon glucose oxidation by pancreaticislets. The data in Table V show that the DPH-inducedinhibition of insulin secretion by pancreatic islets is notaccompanied by a depression of glucose metabolism.DPH failed to diminish the oxidative conversion ofeither glucose-1-14C or glucose-6-14C to 14CO2 by theisolated islets.

The effect of DPHon the action of some insulin se-cretagogues. Increasing concentrations of glucose resultin augmented output of insulin by pancreatic pieces.However, the inhibition of insulin secretion by DPHat 200 mg/100 ml glucose was not overcome by increas-ing the glucose concentrations of the incubation media.These results are shown in Table VI.

That the inhibition of insulin secretion by DPNis notonly resistant to reversal by increased concentrations ofglucose but also to reversal by pharmacological insulinsecretagogues is evidenced by the failure of methacholineand tolbutamide to reverse the DPH-induced block. Thesedata are shown in Table VII.

TABLE IVInhibition of Insulin Secretion by Isolated

Pancreatic Islets by DPH*

Insulin secretionDPH mean 4sr n P

pg/mi pU/5 islets/90 min

0 3539 ±250 57.5 516 ±124 5

TABLE VEffect of DPHon '4CO2 Production from Glucose-1-'4C and

Glucose-6-14C by Isolated Islets*

14CO2 production from:

Glucose-1-'4C Glucose-6-14CDPH mean 4SE mean 4SE n

Ag/ml dpm/5 islets/20 min0 1822 ±40 185 ±10 47.5 2043 ±65 245 417 4

* Each reaction mixture contained 5.5 ,uM glucose (1.5 X 105dpm glucose-1-'4C, 1.5 X 106 dpm glucose-6-'4C). Incubationswere carried out for 20 min at 370C. Final reaction volumeswere 0.5 ml.

The secretion of insulin by the pancreas is dependenton the presence of extracellular calcium (11). Becauseof the possibility that DPHmight be depressing the ac-tivity of the pancreatic f8-cell by virtue of a local anes-thetic effect similar to that described by Jaanus, Miele,and Rubin for procaine in the adrenal medulla (12), anattempt was made to overcome the DPH inhibition ofinsulin secretion with increasing concentrations of cal-cium in the incubation media. The data of Table VIIIshow that the DPH-induced inhibition of insulin secre-tion was not reversed by increasing concentrations ofcalcium in the incubation media.

The effects of ouabain and K+ concentrations on DPHinhibition of insulin secretion. Recent studies on thesynaptosome by Festoff and Appel have suggested thatthe neurophysiological properties of DPH can be at-tributed to its stimulation of the membrane sodium-po-tassium-magnesium ATPase and subsequent alterationof resting membrane potential (13).

The possibility that such a mechanism might be op-erative in the pancreas was assessed by testing the in-hibitory activity of DPH in the presence of ouabain(10-' mole/liter) and low extracellular K+ (0.1 mmole/

TABLE VIFailure of High Concentrations of Glucose to Reverse DPH

Inhibition of Insulin Secretion by Pancreas Pieces

Insulin secretionDPH Glucose mean 4sE n P

pg/ml mg/100 ml pU/100 mglissue/90 min

0 200 21,972 -1843 50 300 32,465 -2013 50 600 33,009 --1635 57.5 200 11,197 -1336 5

TABLE IXReversal of DPH-Induced Inhibition of Insulin Secretion in

Pancreas Pieces by Ouabain and Low K+ Concentrations*

Insulin secretionDPH Ouabain K+ mean 4SE; P

pg/ml moles/liter mmoles/ pU/100 mgliter tissue/90 min

o 0 5.0 23,259 411563 50 0 0.1 27,885 :1:2937 5 >0.20 10-4 5.0 30,775 4347 5 0.5

* Glucose concentration in media 200 mg/100 ml.

and in each instance, there was a reduction in 'Na ac-cumulation by the DPH-treated islets of 26-40%.

In a further attempt to delineate the qualitative natureof the DPH-induced inhibition of insulin release, thepotassium concentration in the medium bathing the pan-creatic pieces was raised to 60 mmoles/liter, a maneuvercalculated to antagonize the postulated effect of DPHonthe sodium-potassium ATPase and reduce the resting.membrane potential of the f-cell, thereby leading to itsexcitation. As shown in Table XI, 60 mmpotassiumnot only stimulated insulin secretion in the absence ofDPHas expected, but also it effectively stimulated insulinsecretion in the presence of DPHas well.

DIS CUSSIONThe data presented in this report demonstrate that thera-peutic concentrations of DPHand mesantoin directly in-hibit the insulin secretory response to glucose, tolbuta-mide, and methacholine without a demonstrable altera-tion in the islets' ability to utilize glucose. Furthermore,

TABLE XEffect of DPHon Accumulation of 22Na by Isolated Islets

Fraction of totalcounts accumulation

Total counts 13.9 X 106 dpmControl 80.8 X 103 dpm 5.8 X 10-3DPH 72.0 X 103 dpm 4.8 X 10-3

Total counts 13.0 X 106 dpmControl 60.8 X 103 dpm 4.68 X 10-3DPH 44 X 103 dpm 3.38 X 10-3

Total counts 12.5 X 106 dpmControl 38.4 X 103 dpm 3.1 X 10-3DPH 23.2 X 103 dpm 1.8 X 10-3

22Na accumulation in presence and,ug/ml) (glucose 200 mg/100 ml).

absence of DPH (7.5

TABLE XIReversal of DPH-Induced Inhibition of Insulin

Secretion of Pancreas Pieces by 60 mMK+*

Insulin secretionDPH K+ mean ±sE n P

pg/ml mmoles/ pU/100 mgliter tissue/90 minz

0 5.0 17,275 ±:1512 57.5 5.0 9165 ±415 5

Matthews measured the resting membrane potential ofislets and found it to be approximately -20 mv (19).In the neuron, the resting membrane potential of - 90mv can be attributed almost entirely to the greaterpermeability (100-fold) of the neuronal membrane topotassium than to sodium (20). The significantly lowerresting membrane potential of the pancreatic f-cell ascompared with the neuron could result from either a de-creased potassium permeability or an increased sodiumpermeability of the membrane, assuming these ionic spe-cies to be the major determinants of the resting mem-brane potential. If the lower resting potential of the pan-creatic f-cell were, in fact, partially the result of an in-crease in the membrane sodium permeability, then it isconceivable that islet sodium influx would be largerthan that of the neuron. Such a rise in intracellular so-dium might result in an intracellular sodium to extra-cellular potassium concentration ratio significantlygreater than 10: 1 at the membrane, and DPH would,therefore, be expected to stimulate the sodium-potassium-magnesium ATPase of the f-cell even under normal con-ditions, since there is apparently a constant baselineelectrical activity of the islet at all times (19).

Such an interpretation of our data is also supportedby the observation that 60mM potassium not only re-verses the DPH-induced inhibition of insulin secretionbut also stimulates insulin secretion in the presence ofDPH. It appears likely that elevation of the extra-cellular potassium ion concentration reduces the sodium-potassium ratio to approximately 1: 1, well within therange of the sodium-potassium concentration ratioswherein Festoff and Appel demonstrated no stimulationof the sodium-potassium ATPase by DPH (13). 60 mMpotassium, therefore, would essentially neutralize thedepression of 8-cell excitability by DPH and in addi-tion, because of its reduction of transmembrane poten-tial, would result in just as great a stimulation ofinsulin secretion in the presence of DPH as in itsabsence.

A number of criticisms can be raised to such aninterpretation of our data. First, it may be argued thatDPH might block the passive flux of sodium sinceaccumulation of 'Na is dependent upon passive as wellas active transport. If the blocking of the passive dif-fusion of sodium were an important determinant of itsaction, however, ouabain, 0.1 mmpotassium, and 60 mmpotassium would not be expected to reverse the inhibi-tion of insulin release by DPH since they have noknown direct effect on passive sodium diffusion. Sec-ondly, it could be argued that DPH increases intra-cellular water such that the effective concentration ofsodium within the cell is subliminal. An increase ofintracellular water, however, could not give rise to theobserved increase in 'Na accumulation, although an

increase in intracellular water might well be observedas a result of a reduced intracellular sodium ion con-centration. Thirdly, one could argue that we have notdemonstrated that the intracellular concentration ofsodium is a determinant of the insulin secretory re-sponse of the f-cell to glucose, even assuming thatDPH were to accelerate the sodium-potassium-magne-sium ATPase since it is possible that ATP is a rate-limiting substrate for the process of insulin secretion.If ATP were being drawn away from other cellularprocesses by the DPH accelerated ATPase, a decreasein the utilization of glucose might be expected becauseof the necessary phosphorylation of glucose prior to itsfurther oxidation. From the glucose utilization data,however, it is apparent that C02 production from glu-cose was not diminished.

Admittedly, such evidence implicating the sodium-potas-sium ATPase as the main locus of the insulin inhibitoryeffects of DPH is largely indirect, and further studiesutilizing rates of 'Na efflux will be necessary to deter-mine the validity of such an interpretation of the data.

ACKNOWLEDGMENTSWe would like to thank Mrs. Brenda Stocks and Mrs.Linda Rice for their expert technical assistance.

This work was supported in part by U. S. Public HealthService Grants HE-07061 and AM-12706.

REFERENCES1. Belton, N. R., J. E. Etheridge, Jr., and J. G. Millichap.

1965. Effect of convulsions and anticonvulsants on bloodsugar in rabbits. Epilepsia. 6: 243

2. Said, D. M., J. R Fraga, and T. E. Reichelderfer. 1968.Hyperglycemia associated with diphenylhydantoin (dilan-tin) intoxication. Med. Ann. D. C. 37: 170.

3. Goldberg, E. M., and S. S. Sanbar. 1969. Hypergly-cemic, hyperosmolar coma following administration ofdilantin diphenylhydantoin. Diabetes. 18(Suppl. 1): 356.

4. Peters, B. H., and N. A. Samaan. 1969. Hyperglycemiawith relative hypoinsulinemia in diphenylhydantoin tox-icity. N. Engl. J. Med. 281: 91.

5. Millichap, J. G. 1969. Hyperglycemic effect of diphenyl-hydantoin. N. Engl. J. Med. 281: 447.

6. Malaisse, W. J., F. Malaisse-Lagae, and P. H. Wright.1967. A new method for the measurement in vitro ofpancreatic insulin secretion. Endocrinology. 80: 99.

7. Lacy, P. E., and M. Kostianovsky. 1967. Method forthe isolation of intact islets of Langerhans from the ratpancreas. Diabetes. 16: 35.

8. Butler, T. C. 1957. The metabolic conversion of 5,5-di-phenyl hydantoin to 5-(p-hydroxyphenyl)-5-phenyl hy-dantoin. J. Pharmacol. Exp. Ther. 119: 1.

9. Lund, M., R. S. Jorgenson, and V. Kuhl. 1964. Serumdiphenylhydantoin (phenytoin) in ambulant patients withepilepsy. Epilepsia. 5: 51.

10. Conn, R. D., J. W. Kennedy, and J. R. Blackman. 1967.The hemodynamic effects of diphenylhydantoin. Amer.Heart J. 73: 500.

DPH Inhibition of Insulin Secretion 1947

11. Curry, D. L., L. L. Bennett, and G. M. Grodsky. 1968.Requirement for calcium ion in insulin secretion by theperfused rat pancreas. Amer. J. Physiol. 214: 174.

12. Jaanus, S. D., E. Miele, and R. P. Rubin. 1967. Theanalysis of the inhibitory effect of local anaesthetics andpropranolol on adreno-medullary secretion evoked bycalcium or acetylcholine. Brit. J. Pharmacol. Chemother.31: 319.

13. Festoff, B. W., and S. H. Appel. 1968. Effect of di-phenylhydantoin on synaptosome sodium-potassium-ATPase. J. Clin. Invest. 47: 2752.

14. Skou, J. C. 1965. Enzymatic basis for active transportof Na+ and K+ across cell membrane. Physiol. Rev. 45:586.

15. Muchnik, S., and R. A. Venosa. 1969. Role of sodiumions in the response of the frequency of miniature end-plate potentials to osmotic changes in the neuromuscularjunction. Nature (London). 222: 169.

16. Birks, R. I., and M. W. Cohen. 1968. The action of so-dium pump inhibitors on neuromuscular transmission.Proc. Roy. Soc. Ser. B. Biol. Sci. 170: 381.

17. Birks, R. I., and M. W. Cohen. 1968. The influence ofinternal sodium on the behaviour of motor nerve endings.Proc. Roy. Soc. Ser. B. Biol. Sci. 170: 401.

18. Toman, J. E. P. 1965. Drugs effective in convulsive dis-orders. In The Pharmacological Basis of Therapeutics.L. S. Goodman and A. Gilman, editors. The MacMillanCompany, New York. 3rd edition. Chap. 13. 215.

19. Dean, P. M., and E. K. Matthews. 1968. Electrical ac-tivity in pancreatic islet cells. Nature (London). 219:389.

20. Woodbury, J. W. 1965. The cell membrane: ionic po-tential gradients and active transport. In Physiology andBiophysics. T. C. Ruch and H. D. Patton, editors. W.B. Saunders Company, Philadelphia, Pa. 19th edition. 1:12.

1948 J. S. Kizer, M. Vargas-Cordon, K. Brendel, and R. Bressler