Insulin-mediated Antilipolysis in Permeabilized Rat Adipocytes"

Vol. 259, No . 12, Issue of'June 25, pp. 7701-7704,1984 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1984 by The American Society of Biological Chemists, Inc.

Insulin-mediated Antilipolysis in PermeabilizedRat Adipocytes" (Received for publication, December 28, 1983)

Robert A. MooneyS, Richard D. Ebersohl, and Jay M. McDonald8 From t h e Diuision of Laboratory Medicine, Departments of Pathology und Medicine, Washington University SchooE of Medicine, St. Louis, Missouri 631 10

Elucidating the mechanism by which insulin inhibits lipolysis has been hampered by the unavailability of a broken cell preparation in which the intact cell responses to the hormone could be duplicated. Here we report, using digitonin-permeabilized rat adipocytes, that physiological concentrations ofinsulin inhibit cyclic AMP-activated lipolysis despite the absence of cytosolic and plasma membrane integrity. Cyclic AMP (1.0 mM) maximally activates lipolysis in permeabilized adipocytes greater than 10-fold. Insulin inhibits this activation in a biphasic manner with maximum inhibition of 59 f 8%( N = 7) at lo-' M. At the submaximal concentrations of cyclic AMP (1.0 to 10 pM), insulin (lo-' M) inhibits lipolysis 80 to 90%.Additionally, the antilipolytic effect of insulin is rapid (3 min) and it is specific, with the relatively inactive desoctapeptide analogue of insulin being threeorders of magnitude less inhibitory than native insulin. In contrast to permeabilized cells, intact cells demonstrate only a small lipolytic response to cyclic AMP which is insensitive to insulin. These findings suggest the following about insulin's antilipolytic effects: 1) an intact cell is not required; 2) the intracellular mechanism of action does not require physiological concentrations of thefreely diffusible cytosolic components; and 3) a site of insulin action independent of adenylate cyclase may play a major role.

investigations with isolated plasma membranes address one aspect of insulin action, z.e. plasma membrane events, no currently used broken cell preparation has adequately reproducedinsulin-dependent effects on intracellularprocesses such as lipolysis. Only the workof Seals et al. (14, 15), which demonstrated an insulin-dependentincrease in activity of pyruvate dehydrogenase (14), a mitochondrial enzyme, and a decrease in the phosphorylation of its a-subunit (15) using a combination of plasma membranes and mitochondria provided evidencethat such studies are possible. In this report, a recentlydescribedtechnique to study regulation of lipolysis in digitonin-permeabilized adipocytes (16) is used to demonstrate that the antilipolytic effect of insulin can be reproduced in a fully permeabilized cell. Additionally, the evidencefor a site of insulin action on the hormone-sensitive lipase which is independent of adenylate cyclase is suggested. EXPERIMENTALPROCEDURES

Materials-Male Sprague-Dawley rats, weighing 120 g, were purchased from Eldridge Laboratory Animals, Antonia, MO. Collagenase (Type I) from Clostridium histolyticumwas purchased from Worthington. Bovine serum albumin (Cohn Fraction V), cyclic AMP, enzymes, and digitonin were obtained from Sigma. Porcine insulin and desoctapeptide insulin were generous gifts from Lilly. Methods-Isolated rat adipocytes were prepared by collagenase digestion of epididymal fat pads according to theprocedure of Rodbell (17). Adipocytes were suspended in KRP' buffer with 30 mg/ml of bovine serum albumin and 2.0 mg/ml of glucose and brought to a cell concentration of 1 X IO6 cells/ml. Aliquots of 0.5 ml were distributed The mechanism by whichthe antilipolytic effect of insulin into 1.5-ml microfuge tubes at 37 "C. When preparing permeabilized in adipocytesis mediated is not understood. It is controversial, adipocytes, digitonin was then added and thecells were incubated for for example, whether insulin regulates adenylate cyclase ac- 30 min. The characterization of this preparation has been published tivity (1-5) and whether the modulation of intracellular cyclic (16). The effect of digitonin on plasma membrane integrity was AMP concentrations bythe hormone playsa primary role (6- assessed by four parameters: 1 ) complete leakage of cellular lactate 9). One obstacle to a further understanding of the effects of dehydrogenase without release of membrane-bound cytochrome c insulin on lipolysis specifically and of insulin action in general reductase; 2) complete inhibition of glucose oxidation; 3) rapid efflux of intracellular [ssRb]rubidium; and 4) activation of lipolysis with is the unavailability of a broken cell system in which the effects of insulin on intracellular processescan be reproduced. exogenous cyclic AMP (16). Following the 30-min permeabilization period, cyclic AMP and Several reports demonstrate the effectsof insulin on plasma insulin were added, when appropriate, and lipolysis was monitored membrane events such as the (Ca2+ M?)-ATPase of rat over a subsequent 30-min assay period. Lipolysis was routinely asadipocytes (lo), and more recently, the phosphorylation of sayed in intact and digitonin-treated cell preparations by measurethe insulin receptor in several cells (11-13). Though these ment of the netrelease of glycerol over the 30-min assay period during which the rateof lipolysis was linear. Glycerol content was determined * This work was supported by United States Public Health Service at thebeginning and end of the assay by the method of Wieland (18). Grant AM25897 and a grant from the Juvenile Diabetes Foundation The lipolysis assays which contained 0.5 X lo6 cells per 0.5 mlof International. The costs of publication of this article were defrayed KRP were terminated by addition of 3 N HClO, to the suspensions. in part by the payment of page charges. This article must therefore Extracts were neutralized with 2 N KHC03, and theglycerol content be hereby marked "advertisement" in accordance with 18 U.S.C. was determined fluorometrically utilizing a coupled enzyme reaction Section 1734 solely to indicate this fact. requiring ATP, NAD+, a-glycerophosphate dehydrogenase, and was $ Supported by a fellowship from the Juvenile Diabetes Foundation initiated with glycerokinase. The assays were performed in a hydraInternational. Current address, Department of Pathology and Labo- zine-glycine buffer at pH 9.8. ratory Medicine, University of Rochester Medical Center, Box 626, 601 Elmwood Avenue, Rochester, NY 14642. To whom reprint requests are to be sent. The abbreviation used is: KRP, Krebs-Ringer phosphate.

+

7701

7702

Insulin-mediated Antilipolysis in Permeabilized Adipocytes

fold, respectively, by the cyclic nucleotide. Most significant, The digitonin-permeabilized rat adipocytes used in these however, was the absence of any antilipolytic effect of insulin studies are a unique preparation in which the intracellular in theintact cells a t any hormone concentrations tested (Fig. 1). regulatory mechanisms of hormone-sensitive lipase are accesThe antilipolytic effect of insulin, when expressed as per sible to the extracellular milieu and thus may be regulated cent of inhibition of cyclic AMP-activated lipolysis, was most directly by exogenous cyclic AMP (see Ref. 16 for details). The cells respond to exogenous cyclic AMP with increases in pronounced a t low cyclic AMP concentrations. In the cyclic AMPconcentration range of 1.0 to10 p ~ , M insulin the rate of lipolysis which are demonstrable a t cyclic AMP inhibited the cyclic nucleotide-dependent activity by 80 to concentrationsas low as 1.0 p ~ A. t 30 p~ cyclic AMP, 90% (data not shown). The per cent of inhibition decreased approximately 4-fold activation of lipolysis is observed and a to approximately 50 to 60% as thecyclic AMP concentration maximum activation of approximately 10-fold is observed a t increased to 25 p ~ Further . increases in cyclic AMP concen1.0 mM cyclic AMP. In contrast, cyclic AMP is a relatively tration to 1.0 mM did not produce any further significant ineffective activator of lipolysis in intact cells with concentrachange in the per cent of inhibition of lipolysis (data not tions of 10 pM or less being without effect and 1.0 mM shown). activating lipolysis only 2- to 3-fold. This is in agreement The susceptibility of adipocytes to the antilipolytic effects with the general experience that exogenous cyclic AMP is a of insulin was closely correlated to the permeabilization of poor lipolytic agent with intact cells (19). Thus, the permeathese cells. As shown in Fig. 2, both thelarge lipolytic response bilized rat adipocyte preparation is a brokencell system which to exogenous cyclic AMP and the antilipolytic effect of insulin has rates of lipolysis approaching those of catecholamine- occurred a t a digitonin concentration between 5 and 10 pg/ treated intact cells while responding to low exogenous con- ml, the concentration range in which the cells become comcentrations of the intracellularlipolytic mediator, cyclic AMP, pletely permeabilized (16). At a digitonin concentrationof 20 that are ineffective with intact cells. pg/ml, the lipolytic effect of cyclic AMP and theantilipolytic Using this permeabilized preparation, insulin inhibited the effect of insulin were essentially equal to those at 10 pg/ml. lipolytic effect of cyclic AMP in a concentration-dependent At lower concentrations of digitonin(less than 5pg/ml), manner between and M when both were added where plasma membrane permeability isminimal, exogenous simultaneously (Fig. 1). The largest antilipolytic effect was cyclic AMP elevated the lipolytic rate only 2- to %fold, and observed a t lo-' M insulin a t both maximal (1.0mM) and there was no effect of insulin. submaximal (0.1 mM) cyclic AMP concentrations with lipolThe potentialsignificance of these observations withregard ysis being inhibited by 59 f 8% (range 35-89, n = 7) and 59 to the mechanism of insulin action was investigated further. & 11%(range 24-100, R = 6), respectively. Against submaxiFirst, the inhibition of cyclic AMP-activated lipolysis occurs mal cyclic AMP (0.1 mM), however, 10"' M insulin was nearly rapidly (Fig. 3). Following a 3-min lag which is present even as inhibitory as M . As has been reported previously for in the absence of insulin, the rate of cyclic AMP-activated lipolysis (20, 21) and other insulin-dependentmechanisms lipolysis was linear for 30 min in the presence or absence of (22, 23), a biphasic response to insulin concentrations was insulin with insulin-dependentsuppression of lipolysis apparobserved, with much of the antilipolytic effect being lost at ent at theearliest time point measured ( 5 min) after the lag M insulin or greater. In contrast to the permeabilized period. Second, the specificity of the response was confirmed cells, where 0.1 and 1.0 mM cyclic AMP activated lipolysis 8- by comparing the concentration dependence of the antilipoand 12-fold, lipolysis in intactcells was activated 2.3- and 2.5- lytic response of native insulinto thatof the relatively inactive desoctapeptideinsulin analogue (Fig. 4). Consistent with other insulin-sensitive pathways in intact cells (24, 25), the 2 120, 1 effective antilipolytic concentration of desoctapeptide insulin RESULTS

Digitonin- treated calls

-

h

=Ic-"-m"

y

?

e-

1.0 mM CAMP

+

*20O I

I I

.Intact

a

04-4 0

' la2

10'"

1d-10

li-9

ld-8

1;-7

1i-6

INSULIN (M)

M insulin 0

g

4

0;lO

5

1I0

I5

280

DIGITONIN (pg/rnl)

FIG. 1. Effect of insulin concentrationson CAMP-activated FIG. 2. Effect of digitonin concentration on the antilipolytic lipolysis. Rat adipocytes (1 X IO6 cells/ml) were preincubated for 30 min a t 37 "C in KRP (0,A) or KRP plus 10 pg/ml of digitonin (0, activity of insulin. Rat adipocytes (1 X lo6 cells/ml) were preinA). Cyclic AMPand insulin were then added, and lipolysis was cubated for 30 min at 37 "C in KRP containing digitonin at the monitored over the subsequent 30 min. Basal rates of lipolysis were concentrations indicated. Cyclic AMP (1 mM) alone (0)or cyclic M) (0)were then added, and lipolysis was 13 f 2 for intact cells and 9 & 1 for digitonin-treated cells. Results AMPplusinsulin represent the mean of four determinations performed in two experiments.

monitored over the subsequent 30 min. Results represent the mean of six determinations performed in three experiments.

in Permeabilized Adipocytes

Insulin-mediated Antilipolysis

7703

vated lipolysis suggests the following concerning this effect of insulin: 1) an intact impermeable plasma membrane is not required; 2) the intracellular mechanism of action does not require physiological concentrations of the freely diffusible cytosolic components; and 3) a site of insulin action, independent of adenylate cyclase, may play a major role in mediating the effect. Not only is the antilipolytic effect of insulin observable in permeabilized adipocytes, but pemeabilization is required in order to demonstrate this effect of insulin (Fig. 2). The antilipolytic effect of insulin cannot be explained, for example, by the presence of a population of nonpermeabilized cells, since insulin has no effect on lipolysis in intact cells that have been exposed to cyclic AMP. Furthermore, the association of insulin’s antilipolytic effect with permeabilization cannot be explained solely bythe increased lipolytic rates observed when 0 5 IO 15 20 25 30 cyclic AMP is added to thepermeabilized cell. At cyclicAMP concentrations of 1 to 30 KM, ratesof lipolysis in permeabilTIME (min) ized cells were similar to the rates obtained when intact cells FIG. 3. Rate of change in lipolysis in response to insulin. Cyclic AMP (1.0 mM) alone (0)or cAMP plus lo-’ M insulin (0)were are exposed to 1 mM cyclic AMP. However, under these lipolysis in thepermeabilized cells added to permeabilized cells as described in Fig. 1. Rates of lipolysis conditions insulin inhibited were monitored at several intervals over the next 30 min. Each value by 80% but had no effect on intact cells. represents the average of duplicate samples in a representative exThere aretwo previous studies in which digitonin was used periment. during the investigation of adipocyte functions. Cuatrecasas (26) reportedthat the treatmentof isolated adipocytes with 6 to 20 pg/ml of digitonin increased the specific binding of lZ5Iinsulin by 9 and 63%, respectively. They report no attempts, however, to correlate these changes with changes in insulinmediated cellular events. In contrast, Akhtar and Perry (27) have reported that 5 to 10 pg/ml of digitonin reduced insulinstimulated glucose uptake in rat adipocytes without affecting basal uptake. A t higher digitonin concentrations both basal and insulin-stimulated glucose uptake were inhibited, and plasma membrane permeability was demonstrated. Based on the present findings, these latter observations do not reflect a general inhibition of insulin-mediated effects by digitonin but rather a more specific effect on the as yet unidentified coupling of the insulin-receptor complex to glucose transport. The exact site at which the antilipolytic effect of insulin is mediated remains unknown. Especially controversial is the INSULIN DESOCTAPEPTIDE apparent inhibition of adenylate cyclase by insulin which has (M) INSULIN (M) been reported by Hepp (1) and Cuatrecasas and co-workers FIG. 4. Comparison of the antilipolytic effects of desocta- (2, 3) but not confirmed by others (4, 5). Inthe present peptide and native insulin. Insulin and desoctapeptide insulin, at experiments, adenylate cyclase apparently contributed a mithe concentrations indicated, were added to permeabilized adipocytes nor role, if any, in mediating the antilipolytic effect of insulin along with cAMP (1.0 mM). Rates of lipolysis were determined as described in Fig. 1,and results were expressed as aper cent of lipolysis since this enzyme was essentially bypassed when the lipolytic activated by cAMP alone. Results represent the mean S.E. of four pathway was activated by cyclic AMP. In the permeabilized determinations performed in two experiments. adipocyte, concentrations of this cyclic nucleotide in the freely exchangeable intracellular pool(s) would be nearly equal to was three orders of magnitude higher than that of native extracellular concentrations. With most cyclic AMP conceninsulin. trations used in these experiments, the resulting cytosolic The contribution of adenylate cyclase to both the lipolytic cyclic AMP concentrations would far exceed the contribution and antilipolytic response in this preparation would appear of endogenous adenylate cyclase, especially since the enzyme to be minimal. An assessment of the contribution of adenylate was not activated by any lipolytic hormone. Conclusive evicyclase to lipolysis under these experimental conditions can be obtained by measuring the ratesof lipolysis in theabsence dence for this, however, cannot be obtained unless adenylate of added cyclic AMP. Basal rates of lipolysis were 13 & 2 mol cyclase activity is totally eliminated from these preparations. of glycerol/30 min/mg of protein for intact cells and 9 f 1for Thus, without that ability, an insulin-sensitive synergistic permeabilized cells. Thus, in permeabilized cells, this rate is relationship between adenylate cyclase and exogenous cyclic less than 10% of that observed with 1.0 mM cyclic AMP. AMP, possibly involving compartmentalized cyclic AMP, canAdditionally, insulin had little effect on this basal lipolytic not be ruled out. These findings have other interesting implications with rate, inhibiting it by 6.5 & 3.9% (n= 6) in intactcells and by 2.6 f 1.7% (n = 5 ) in permeabilized cells. regard to therole played by other proposed soluble mediators of insulin action inthe regulation of hormone-sensitive lipase. DISCUSSION The permeabilized insulin-sensitive adipocyte preparation The demonstration that insulin can exert an antilipolytic would seemingly not permit effective concentrations of ions effect in permeabilized adipocytes against cyclic AMP-acti- (e.g. calcium, magnesium, hydrogen, potassium), small molec3

250

I

I

*

7704

Insulin-mediated Antilipolysis

ular weight proteins, cyclic GMP, or hydrogen peroxide to be generated from the plasma membrane and carried via the soluble cytosolic compartment to an internal effector site. The exception would be a n effector site possessing very high binding affinityfor the putative insulin mediator. Under these conditions, the unbound mediator concentration would presumably be very low, but effector site occupancy might still approach the physiological range. Additionally, our findings do not rule out directedsignaling via membrane channels or an internal siteof insulin binding asa mechanism of regulation. Intracellular sitesof insulin binding have been identified in a variety of cells (28). The significance of thesesites, however, has yet to be elucidated. Finally,it will be of interest to determine whether other intracellular targets of insulin action such as pyruvate dehydrogenase and glycogen synthase can be similarly regulated by insulin in this permeabilized adipocyte preparation to determine whether these observations are universal or unique to the regulation of hormonesensitive lipase. REFERENCES 1. Hepp, K. (1972) Eur. J. Biochem. 31, 266-276 2. Illiano, G., and Cuatrecasas, P., (1972) Science 175, 906-908 3. Torres, H., Flawia, M., Hernaez, L., and Cuatrecasas, P. (1978) J. Membr. Biol. 43, 1-18 4. Pilkis, S., Claus, T., Johnson, R., and Park, C. (1975) J . Biol. Chem. 250,6328-6336 5. Pohl, S., Birnbaumer, L., and Rodbell, M. (1971) J. Biol. Chem. 246,1849-1856 6. Fain, J., and Rosenberg, L. (1972) Diabetes 21, Suppl. 2,414425 7. Siddle, K., and Hales, C. (1974) Biochern. J. 142,97-103

in Permeabilized Adipocytes 8. Fain, J., Li, S-Y., and Moreno, F. (1979) J. Cyclic Nucleotide Res. 5,189-196 9. Khoo, J., Steinberg, D., Thompson, B., and Mayer, S. (1973) J . Biol. Chem. 248,3823-3830 10. Pershadsingh, H., and McDonald, J. (1979) Nature (Lond.)281, 495-497 11. Haring, H-U., Kasuga, M., and Kahn, R. (1982) Biochem. Biophys. Res. Commun. 108, 1538-1545 12. Avruch, J., Nemenoff, R., Blackshear, P., Pierce, M., and Osathanondh, R. (1982) J . Biol. Chem. 257,15162-15166 13. Zick, Y., Kasuga, M., Kahn, R., and Roth, J. (1983)J. Bwl. Chem. 258,75-80 14. Seals, J., and Jarett, L. (1980) Proc. Natl. Acad. Sci. U S A . 77, 77-81 15. Seals, J., McDonald, J., and Jarett,L. (1979) J. Biol. Chem. 254, 6991-6996 16. Mooney, R., Ebersohl, R., and McDonald, J. (1983) Eur. J. Biochem. 136, 603-608 17. Rodbell, M. (1964) J. Bwl. Chem. 239,375-380 18. Wieland, 0.(1957) Biochem. 2. 329,313-319 19. Fain, J. N. (1973) Pharmucol. Reu. 25, 67-118 20. Kono, T.,and Barham, F. (1973) J. Biol. Chem. 248, 7417-7426 21. Desai, K., Li, K. C., and Angel, A. (1973) J. Lipid Res. 1 4 , 647655 22. Kono, T., Robinson, F., and Sarver, J. (1975) J . Biol. Chem. 250, 7826-7835 23. Wong, E., and Loten, E. (1981) Eur. J. Biochem. 115, 17-22 24. Gliemann, J., and Gammeltoft, S. (1974) Diabetologia 10, 105113 25. Lamer, J., Kikuchi, K., Freer, R., and Morris, H. (1979) Diabetes 28,390 26. Cuatrecasas, P. (1971) J. Biol. Chem. 246,6532-6542 27. Akhtar, R., and Perry, M. (1975) Biochim. Biophys. Acta 411, 30-40 28. Goldfine, I. (1981) in Biochemical Actions of Hormones (Litwack, G., ed) Vol. VIII, pp. 273-305, Academic Press, New York