143

TABLE II-HL-A TYPE OF LYMPHOCYTES WHICH REACTED WITHSERUM S9Hb40

* Bl = Blank.

reacted with lymphocytes from a further 15 subjects(8 N.P.C. patients and 7 controls). Of these 15, 12 hadW15 (6 with N.P.c. and 6 controls). 4 other controlshad W15 but did not react with S98a/28. Thusbesides an activity corresponding to that of serum

S98b/40, serum S98a/28 seemed to have an activityresembling W15.

DISCUSSION

Two approaches have been used to investigate thefindings 3,4 of an increased frequency of undetectablesecond-locus antigen(s) in N.P.C. patients. First, familystudies have indicated that the blank is inherited asan HL-A antigen, is not due to homozygosity ofdetected antigens, and cannot be ascribed to a disease-induced phenomenon. The second approach involvedan attempt to identify sera having anti-HL-A activityagainst additional second-locus antigens, and whichreact more often with lymphocytes from N.P.C. casesthan from controls. Two sera were identified whichseemed to specify a previously undetectable antigen,Sin-2. One serum (S98b/40) may be operationallymonospecific for Sin-2. This serum reacted with

lymphocytes from 11 subjects; 9 were confirmed ashaving N.P.C., 1 was a control, and the llth hadclinical symptoms and signs of N.P.C., but biopsy hadnot been performed. Of the 10 patients with a defi-nite diagnosis who reacted against S98b/40, 9 hadN.P.C., which gives a nominal one-sided significancelevel of 0008. The correction factor applied to thisnominal level should correspond to the number ofdifferent reactivity patterns found among the screen-ing sera which could be described as new HL-A

antigens. No other serum from the 612 subjectstested reacted with a sufficient number of lympho-cytes to enable the specificity of any new reactivityto be clearly identified. The nominal value of 0-006would therefore seem to be a reasonble, conservativeestimate.

Typing in families with more than one N.p.c. casefrom Hong Kong has confirmed that Sin-2 segregatesas an HL-A antigen.5 Correlation analyses of S98b/40reactivity with sera suspected of detecting second-locus antigens of low frequency in Caucasians mayenable Sin-2 to be further identified. Finally, thehigh risk of N.P.c. associated with Sin-2 needs to beconfirmed in a large series. Nevertheless, the resultsindicate the existence of a hitherto undetected second-locus HL-A antigen which is associated with a highrisk of N.P.C.

Supported by a collaborative research agreement betweenthe Biological Carcinogenesis Unit, I.A.R.C., Lyon, France,and the I.A.R.C. Research Centre, Department of Pathology,University of Singapore, under contract no. NIH-NCI-E-70-2076 with the special viral cancer programme of the NationalCancer Institute, National Institutes of Health, Bethesda,U.S.A.

Requests for reprints should be addressed to M. J. S.,W.H.O. Immunology Research and Training Centre, Facultyof Medicine, University of Singapore, Sepoy Lines, Singapore 3.

REFERENCES

1. Morris, P. J., Lawler, S., Oliver, R. T. Histocompatibility Testing1972 (edited by J. Dausset and J. Colombani); p. 669. Copenhagen,1973.

2. Rogentine, G. N., Yankee, R. A., Gart, J. J., Nam, J., Trapani,R. J. J. clin. Invest. 1972, 51, 2420.

3. Simons, M. J., Day, N. E., Wee, G. B., Shanmugaratnam, K.,Ho, H. C., Wong, S. H., Ti, T. K., Yong, N. K., Darmalingam,S., de Thé, G. B. Cancer Res. 1974, 34, 1192.

4. Simons, M. J., Wee, G. B., Day, N. E., Morris, P. J., Shanmugarat-nam, K., de Thé, G. B. Int. J. Cancer, 1974, 13, 122.

5. Ho, H. C., Simons, M. J., Day, N. E., Chan, J., Wee, G. B., Kwan,H. C., Chan, S. H., de Thé, G. B. Unpublished.

Hypothesis

MODEL FOR EXTRUSION OF

INSULIN &bgr; GRANULES

NORMAN R. LAZARUS BRIAN DAVIS

Diabetes Research Laboratory, Wellcome Foundation,Dartford, Kent DA1 5AH

Summary Considerable controversy surroundsthe mechanisms of insulin secretion.

This hypothesis attempts to explain the events whichmay lead to insulin extrusion through the &bgr;-cellmembrane. The role of various effectors of insulinsecretion on the activities of membrane-bound enzymesis discussed. Glucose, without effect on these

enzymes, did induce insulin secretion from insulin

granules and plasma membranes in vitro. The model

proposes a mechanism whereby glucose may triggerfirst-phase insulin secretion; reinforcing mechan-

isms, such as raised levels of cyclic adenosine mono-phosphate, could augment subsequent secretion.

THE two areas of controversy which surround themechanism of glucose-induced insulin release are therole of cyclic 3",5’-adenosine monophosphate(CA.M.P.) and the question of whether glucose has tobe metabolised or whether it reacts with a " gluco-receptor ".1 It might be possible to resolve these

questions by dividing insulin secretion, for the pur-poses of investigation, into two mechanistically dis-tinct processes :

Emiocytosis.-This would involve the fusion of the 6granular membrane with the cellular plasma membranesuoh that secretory openings can be formed. The pro-cesses governing these reactions are unknown but are

probably located in both the /3 granular and plasma mem-branes and would be linked to cellular metabolism by beingenergy dependent.

Granular movement.-This process moves the granulesto the plasma membrane. This movement is also energydependent.While both processes are mechanistically distinct

144

they must be closely integrated and synchronised notonly with each other but with cellular metabolismOne of the difficulties in setting up experiments is

trying to obtain conditions which will distinguistthose reactions which are directly related to the pro-cess under study and those that are concerned withthe general metabolism of the cell. We are not con-vinced that published experiments have resolved thisdifficulty.We have tried to understand the mechanisms

underlying granular extrusion by studying purifiedplasma membranes. Thus far the following facts havecome to light.2-4 There is in the plasma membrane aprotein kinase that is stimulated by CA.M.P., glucose-6-phosphate, and phosphoenolpyruvate. It is also pos-sible to increase kinase activity by inhibiting the Na+/ /K+ A.T.P.ase in the membrane with ouabain. Ouabain

potentiates glucose-induced insulin secretion.5s Thiseffect is probably mediated by allowing more substrate(i.e., A.T.P.) to be presented to the kinase and suggests apossible link between the CA.M.P. system and the

Na+/K+ A.T.p.ase.7 Glucose has no effect on thiskinase. Phosphodiesterase is also present in the mem-brane. The close relationship between the adenylatecyclase and the phosphodiesterase suggests thatCA.M.P. levels can be regulated at the membranewithout necessarily affecting CA.M.P. levels in the cell.To investigate the role of the above described reac-

tions in insulin secretion we devised a system thatwould liberate insulin in vitro. This system consistsof incubating, under precise conditions, purifiedplasma membranes with isolated /3 granules and

determining whether insulin could be released fromthe granules. Preliminary (unpublished) experimentssuggest that insulin granules are stable under ourconditions of incubation, and that plasma membranesare necessary for insulin release. This release is

dependent upon the presence of Ca++ and A.T.P.

Pituitary plasma membranes and liver plasma mem-branes are ineffective; release is augmented by phos-phoenolpyruvate and glucose-6-phosphate; glucosecan cause release, but only in the presence of A.T.P.;and insulin secretion is inhibited by diazoxide.As a result of these findings we wish to postulate

the following model for insulin extrusion.Glucose acts as a trigger for insulin release, but

does so only in the presence of sufficient A.T.P. Thisaction of glucose is not mediated via any of the

enzymes described above but is distinct from them.The link between glucose and A.T.P. is unclear. Theaction of glucose can be reinforced by a number ofmechanisms probably mediated through the proteinkinase. These include CA.M.P., produced by those

compounds which stimulate the adenylate cyclase-namely, glucose-6-phosphate and phosphoenol-pyruvate, both metabolites of glucose. The synergisticeffects of these metabolites on glucose-mediatedinsulin release may be regarded as a modified" cascade " system providing a mechanism for con-tinued release once the initial triggering mechanismhas been set in motion by glucose itself. It may bethat the glucose trigger is associated with first-phaseinsulin release and that the reinforcing mechanismsare related principally to second-phase release.We are aware that some areas of this hypothesis are

highly speculative but believe there is merit in our

suggestions for the following reasons. They helpexplain why glucose-mediated insulin release may notbe CA.M.P. dependent, yet suggest a mechanism where-by CA.M.P. can augment the glucose response and whyraising CA.M.P. levels in the absence of glucose need notcause insulin release. It also suggests that glucosereleases insulin both by itself and by being metabolised,and that while the pathways of this effect are differentthey end at a common terminus. Probably the

greatest utility of the model is that it suggests manyideas that can be tested and allows for differentiatingthose insulin-secreting mechanisms that are mem-

brane associated and those that are not.

Requests for reprints should be addressed to N. R. L.

REFERENCES

1. Hellman, B., Idahl, L.-A., Lernmark, A., Taljedal, I.-B. Proc.natn. Acad. Sci. U.S.A. 1974, 71, 3405.

2. Lazarus, N. R., Davis, B. Proc. R. Soc. Med. 1974, 67, 841.3. Davis, B., Lazarus, N. R. J. Membrane Biol. (in the press).4. Davis, B., Lazarus, N. R. Biochem. Soc. Trans. 1974, 2, 409.5. Burr, I. M., Marliss, E. B., Stauffacher, W., Renold, A. E. Am. J.

Physiol. 1971, 221, 943.6. Gutman, R. A., Lazarus, N. R. Excerpta med. int. Congr. Ser. 1973,

no. 280, p. 23.7. Tria, E., Luly, P., Tomasi, V., Trevisani, A., Barnabei, O. Biochim.

biophys. Acta, 1974, 343, 297.

MODE OF INSULIN ACTION

A. H. KISSEBAH * B. R. TULLOCH †H. HOPE-GILL P. V. CLARKEN. VYDELINGUM T. R. FRASER

Department of Clinical Endocrinology, HammersmithHospital, London W12

Summary A unifying hypothesis is proposed forthe mechanism of insulin action in

adipose tissue. Insulin both induces displacementof Ca++ from a membrane-bound pool and inhibitsefflux of the ion, thereby facilitating a rise in intra-cellular free Ca++ concentration. The former effectcould enhance the transport of substrates and ionsinto the cell, while the latter modulates the activityof some intracellular enzymes to stimulate glyco-genesis, lipogenesis, and decrease lipolysis and

glycogenolysis. The calcium ion might act as the

missing second messenger for insulin action.

INTRODUCTION

EVIDENCE from a wide spectrum of sources suggeststhat the intracellular effects of adrenaline, glucagon,and many other hormones are mediated by the secondmessenger, 3; 5’-cyclic adenosine monophosphate(CA.M.P.). The mechanism whereby insulin evokesits cellular effects, however, remains an enigma.Initial reports 1.2 suggested that insulin operated by amechanism which involved a reduction in tissueCA.M.P., but later work has revealed at most a poorcorrelation between the extent of CA.M.P. changes andthe cellular responses to insulin. 3-6 Yet insulin exertsits cellular effects while remaining extracellular,7 and

* Present address: St. Mary’s Hospital, London W2.† Present address: Department of Medicine, Manchester Royal

Infirmary.