The Role of Glucose, Glucagon and Glucocorticoids

Transcription

1 European J. Biochem. 6 (1968) The Role of Glucose, Glucagon and Glucocorticoids in the Regulation of Liver Glycogen Synthesis H. DE WULF and H. G. HERS Laboratoire de Chimie Physiologique, Universith de Louvain (Received June;7 / July 3, 1968) The conversion of liver glycogen synthetase b into a in vivo, induced by glucose or by glucocorticoids, is further enhanced when both treatments are combined. Once activated, the enzyme is rapidly inactivated by an injection of glucagon, epinephrine or cyclic AMP. When given together with glucose, these substances prevent the usual activation of the enzyme. The sensitivity of the mouse to glucagon was studied by the inactivation of liver glycogen synthetase and by the activation of liver phosphorylase; the fist effect was obtained with lower doses than the second. On the other hand, about 1 times more glucagon was required to inactivate the synthetase previously activated by prednisolone than to prevent its activation by glucose. This difference in sensitivity can presumably not be explained by a different rate of destruction of cyclic AMP, as neither the total activity of the diesterase nor its affinity for the cyclic nucleotide is altered by the treatments. No evidence has been obtained suggesting that prednisolone lowers the sensitivity of the synthetase b ase towards cyclic AMP. In a recent communication, Bishop and Larner [l] have shown that glucagon cancels the activation of glycogen synthetase which has been induced in dog liver by the simultaneous administration of glucose and insulin. They concluded that glycogen synthetase is presumably inactivated [i.e. converted from a glucose 6-phosphate independent form (I) back to a glucose 6-phosphate dependent form (D)] by a specific kinase which is sensitive to adenosine 3,5 - cycl ophosphate. We have independently described the very large activation of liver glycogen synthetase which occurs in mice after the administration of glucose or glucocorticoids [2,3] ; this activation appears now to be the result of the conversion of a form of glycogen synthetase called b, which is inactive in the presence of the concentrations of phosphate, ATP and glucose 6-phosphate existing in the cell, into an active or a form [4]. The a and b enzymes are most probably identical to the I and D forms. As insulin, given either alone or simultaneously with glucose, has no activating effect on the b to a conversion [2] and as the glucose effect is still present after the administration of anti-insulin serum [5], the activation observed by Bishop and Larner [l] was presumably the result Enzymes. Glycogen synthetase or UDPG : a-1,4-glucan a-4-glucosyltransferase (EC ), glycogen phosphorylase or a-1,4-gluca,n : orthophosphate glucosyltransferase (EC ), AMP-5 -nucleotidase or 5 -ribonucleotide phosphohydrolase (EC ), phosphodiesterase or cyclic adenosine monophosphate phosphohydrolase (EC ). of the glucose load rather than of insulin and glucagon most probably suppressed the activation of glycogen synthetase achieved by glucose. It also appears that a normal insulin secretion is not required for the stimulation of glycogen synthesis by glucocorticoids. As shown by Tarnowski et al. [6], this stimulation is somewhat delayed in the alloxandiabetic adrenalectomized rat, but is otherwise normal. This delay explains why Kreutner and Goldberg [7] did not observe the effect within two hours after the administration of cortisol. In the present paper we report a study of the interactions of the three main factors which appear to control the rate of glycogen synthesis in the liver, namely glucose, glucocorticoids and glucagon. MATERIAL AND METHODS Unless otherwise stated, experiments were performed with normally fed male Naval Medical Research Institute mice, weighing about 2 g. F rednisolone, at the dose of 1 mgl2 g body weight, was injected subcutaneously 3 h before the start of the experiment. All other test substances were injected intravenously, while control animals received an equivalent volume of.15 M NaC1. The rate of liver glycogen synthesis in vivo was determined as previously described [2]. Except for the experiment reported in Fig. 1, glycogen synthetase was assayed with.5 ml of a

2 vol.6. N.4, 19s H. DE WULF and H. G. HERS 559 I ACTIVITYOF GLYCOGEN SYNTHETASE (nmoles/min/g LIVER) Fig. 1. Correlation between the activity of liver glycogen synthetase measured in a concentrated hogenate and the rate of the conversion in vivo of glucose to liver glycogen in the same animal. The experiments carried out in vivo lasted 1 min and were performed by the injection of [6-3H]glucose to mice 5 min after the injection of glucose (1 mg/g) or of.15 M NaCI. Half of the mice had been treated with prednisolone. After the death of the animal, one half of the liver was used for the determination of the tritiated glycogen and the other for the assay of glycogen synthetase at " in a 5 /, homogenate [2].,(o) control mice; () mice injected with glucose; () prednisolone treated mice ; (H) prednisolone treated mice injected with glucose. Correlation coefficient:.94 (P <.1) rn of phosphorylase was the same whether the liver had been frozen according to Wollenberger et al. [lo] or in a solid C,-acetone mixture. Phosphodiesterase was assayed according to Butcher and Sutherland [ll] and the Pi released estimated by the Fiske-Subbarow method [12]. Snake venom (of Crotalus atrox) was purchased from Sigma Chemical Company (St. Louis, Mo., U.S.A.). Glucagon (Novo Laboratories, Copenhagen) was a gift of Dr. Hubert. All other techniques as well as the source of the chemicals have been described [2-41. RESULTS Interaction between Glucose and Glucocorticoids The experiments presented in Table 1 show the previously reported [2,3] stimulation of liver glycogen synthesis by a glucose load or by prednisolone ; when the two treatments were combined in the same animal, their effects were superimposed and the rate of liver glycogen synthesis reached very high values (1 to 2O/, of glycogen deposited in one hour). In these experiments, the increased rate of glycogen synthesis was also accounted for by a parallel activation of liver glycogen synthetase (Fig.l), of which a large proportion was converted into the a form [4]. In order to reach a better understanding of the stimulatory effects, we have also investigated to what extent the known latencies in the glucose (3 to 5 min) and in the corticoid (2-3 h) effects would be modified by the association of the two treatments. It is shown in Table 2 that in prednisolone treated mice, the latency in the action of glucose is shorter Table 1. Influence of a glucose load on the rate of liver glycogen synthesis by normal and prednisolone treated mice Glucose (1 mg/g) or.15 M NaCl was injected with a trace amount of [U-14C]glucose to mice which had or had not been treated with prednisolone. The animals were killed 1 rnin later. Values shown are means f standard error of the means with the number of animals in parentheses. When only two values were obtained, individual data are given Experiment Treatment Mean glucose concentration Glucose incorporated into glycogen mg/g liver mg/lo min/g liver Saline (3) 1.45 f Glucose (2) 2. ; no 1 Prednisolone + saline (3) 1.4 &.7.22 &.64 Prednisolone + glucose (3) 2.57 & a &.212 Prednisolone saline (9) 1.57 &.G8.596 &.92 no 2 Prednisolone + glucose (9) j=.247 a The reduced latency in the glucose effect after prednisolone administration (see Table 2) explains the fact that the effect of the combined treatments is greater than the sum of the two effects obtained separately. 25O/, liver homogenate, as described in the preceding paper [8]. Phosphorylase activity was determined in the presence of AMP, according to Hers [9] except that we used shorter incubation times (2, 4 and 6min) and a loo/, homogenate made in 15mM sucrose, 1 mm EDTA, 1 mm NaF. The activity than in untreated mice, in that a nearly maximal glucose effect is obtained within 3 min. In the reverse experiment, not reported in detail, the administration of glucose did not shorten the lag period (2 to 3 h) which is required to obtain the stimulation of glycogen synthetase by prednisolone.

3 Y 56 Action of Glucose, Glucagon and Glucocorticoids on Glycogen Synthesis European J. Biochem. Table 2. Influence of prednisolone on the latency of the glucose effect Glucose (1 mglg) was injected to 24 h fasted mice which had or had not been treated with prednisolone; 2 or 5 min later, a trace amount of [U-14C]glucose was administered and the mice were killed 1 rnin later. Values shown are means f standard error of the means, with the number of animals in parentheses Treatment Glucose incorporated into glycogen Control pg/min/g liver Preduisolone Control 2.4 f.76 (5) (5) 2 min after glucose 14.7 & 4.48 (6) 62.8 f 7.7 (5) 5 min after glucose 42.6a f 1.44 (6) 87. f (5) a Significantly different from the value obtained at 2min, at the 5 level. - LT 25 - w 2 -I -E"._ W -I K w In F 15 - W I +- z > In z 1 - W 8 3 W LL 5-5 s I I I 1 I I TIME AFTER GLUCAGON (min) Fig.2. The kinetics of the inactivation of liver glycogen synthetase a by glucagon. Glucagon (1 pg/2 g) was injected to mice which had received prednisolone 3 h earlier. Vertical bars represent f the standard error of the means and the number of experiments is given in parentheses The Inactivation of Glycogen Xynthetase by Glucagon, Epinephrine and Cyclic AMP The intravenous administration of glucagon, epinephrine or cyclic AMP to normal mice caused a decrease in the already low activity of liver glycogen synthetase measured in IOmM sulfate, lowering it sometimes to values as low as 2 to 3O/, of the activity measured in 5mM glucose 6-phosphate. When the same substances were administered to animals whose synthetase had been activated either by glucocorticoids or by an intravenous load of glucose, they induced a rapid and profound inactivation of the enzyme ; when given simultaneously with glucose, they prevented its activation. The inactivation is a reconversion of the a form into 6 ; with cyclic AMP, it is complete in 1 min and is therefore more rapid than the b to a transformation which is induced by glucose [2]. These effects are illustrated in Table 3 and in Fig. 2 and 3. (5) I I I I P (5) TIME AFTER CYCLIC AMP (min) Fig.3. The kinetics of the inactivation of liver glycogen synthetase a by cyclic AMP. Same procedure and representation as in Fig. 2 except that cyclic AMP (1 pmole or.329 mg/2 g) was injected Table 3. The effect of glucagon, epinephrine and cyclic AMP on liver glycogen synthetase in three separate experiments, either glucagon (1 pg/2 g) or epinephrine (5 pg/2 g) or cyclic AMP (.6 mg/2 g) was injected to mice, simultaneously with glucose (1 mg/g) or 3 h after the administration of prednisolone, or to control animals. The mice were killed 6 min later. Values shown are means & standard error of the means, with the number of animals in parentheses. When less than 3 measurements were obtained, individual data are given Experiment no 1 no 2 no 3 Treatment Control Activity of glycogen synthetase Glucose nmoles/min/g liver Prednisoloue Control 22.5 f 9 (3) 8 f 21 (3) 125 & 34 (3) Glucagon 4 i 1 (3) 12 f 5.5 (3) 16 f 5 (3) Control 13 Xk 4 (5) 111 f 32 (4) 23 f 46 (5) Epinephrine 5 x (1) 13 i 6 (5) 14 i 3 (5) Control 28 ; 3 (2) 164 & 6 (3) 163 f 22.5 (3) Cyclic AMP 7 f 1(3) 6 ; 8 (2) 11 f 2 (3)

4 Vol.6, N.4, 1968 H. DE WULF and H. G. HERS 561 Table 4. The sensitivity towards glucagon after the administration of glucose orland glucocortiwids to mice Mice were injected with various doses of glucagon, given either together with a glucose load (1 mg/g) or after the administration of prednisolone; in a third series of animals, the two treatments were combined. The mice were killed 6 min later. Values shown are means f standard error of the means, with the number of animals in parentheses. When only two measurements were made, individual data are given Amount Activity of glycogen synthetase Activity of glycogen phosphorylase of glucagon given Prednisolone Glucose Glucose Prednisolone Glucose Glucose (rg/?o 8 body weight nmoleslminlg liver Prednisolone + Prednisolone + pmoles/min/g liver 115 & 9 (24) 15 f 1 (37) 146 f 17 (1) 3.4 f 1.7 (2) 35. f 1.7 (3) 35.6 f 6.5 (4).3 71 f 2 ( 5).1 87 f 16 ( 4) 18 f 7 (1) 32.5 & 5 ( 5) 42.6 f 2.3 ( 5) ( 8) 115 It 11 ( 5) 38.8 f 1.1 ( 4).6 6 f 3 (5) 41.4 f 3.9 ( 5) f 1 ( 9) 11.5 f 2.5 ( 9) 122 i 5 8 ( 5) 37.9 & 2.4 ( 8) 47.8 & 2.8 ( 9) f 1 ( 5) 48.3 f 3.5 ( 5) f 8 (11) 18 f 4.5 (19) 1 i & 1 (11) 32.4 & 3.5 (1) 37.1 f 2.7 (11) 28.6 * 3.2 (5) f.3 ( 5) 37.4 f 5.2 ( 5) 1 15 f 3 ( 9) 5 f 1 (13) 52.5 f 18 ( 3) 42.2 & 2.4 ( 9) 4.6 f 2.6 (13) 3 5 (8) 44.6 f 2.9 ( 5) 46.2 & 4. ( 8) 1 12 f 3 (7) 4 & 1 (8) 52.1 f 3.3 ( 4) 48.2 f 1.5 ( 6) 2 12 f 2.5 ( 5) 3 f.5 ( 9) 8 f 3 ( 5) 48.9 f 5.6 ( 5) 46.8 f 2.9 (1) 31 ; 47 (2) Control 16.5 f 2 (45) 4.8 & 1.9 (36) f - The dose of glucagon required to inactivate the synthetase of mice previously treated by corticoids or to prevent its activation by glucose has been investigated in detail (Table 4). More than 25 mice were used in this investigation which, in spite of a marked variability in the results, indicates that the inactivation of the enzyme previously activated by prednisolone requires higher doses of glucagon than the prevention of its activation by glucose. It is evident that the glucose effect was completely inhibited by a dose of glucagon as low as.1 pg/ 2 g whereas 1 pg was required to lower the synthetase activity of the prednisolone treated animals to the value of the controls. A statistical analysis shows that, for the steroid injected mice, the degree of inhibition was significantly less than for the glucose treated animals, at almost each dose assayed (.1 pg/ 2 g being the only major exception). The difference between the effects of the two treatments was particularly evident in paired experiments performed on the same group of mice (see Table 5). When the glucose and glucocorticoids treatments were combined in the same animal, the sensitivity towards glucagon was also lower than that after glucose alone. It appears therefore that the prednisolone treatment has induced a certain degree of resistance to glucagon, at least as far as the inactivation of glycogen synthetase is concerned. We could only observe a small activation of phosphorylase by glucagon in the mouse, mostly because the activity of this enzyme was already very high in a liver homogenate obtained from an untreated animal. It appears that prednisolone and glucose induced a small but highly significant decrease in the activity and that glucagon reversed this effect; however, the variability of the results prevented an accurate estimation of the difference in their sensitivity towards glucagon between the mice treated by glucose or by glucocorticoids (Tables 4 and 5). One has however the general impression that more glucagon was required to substantially activate the phosphorylase than to inactivate the synthetase. This is particularly true in the case of the glucose treated mice who displayed an inactivation of their synthetase at.1 pg/2 g and only a statistically significant increase of their phosphorylase at.1 pg/2 g; however, for the prednisolone treated animals, a reproducible effect of glucagon on phosphorylase could only be obtained at a dose of 1 pg/2 g, but an effect has also been obtained at as low a dose as.1 pg/2og. When the effects glucagon on the synthetase and on the phosphorylase in the same group of animals are plotted against each other, it clearly appears (Fig.4) that the first effect has reached its maximum when the second is still barely detectable. We may conclude that the amount of glucagon which is required to inactivate glycogen synthetase is 1 to 1 times lower than that which induces a dehite activation of phosphorylase. A similar experiment was also performed in vivo with cyclic AMP (Table 6) ; a dose of.1 mg/2 g of the cyclic nucleotide caused approximately the same effect on the synthetase as.3 pg/2 g ofglucagon in the glucose treated animals. A clear-cut difference in sensitivity between the glucose and the glucocorticoid treated animal could not be demonstrated. The greater sensitivity of the glucose injected mice to.1 mg/2og of cyclic AMP is not statistically significant.

5 Action of Glucose, Glucagon and Glucocorticoids on Glycogen Synthesis European J. Biochem ACTIVATION OF GLYCOGEN PHOSPHORYLASE (pmoles/min/g LIVER) Fig.4. Comparison between the effects of glucagon on liver glycogen synthetase and on liver glycogen phosphorylase. For each dose of glucagon given to glucose or prednisolone treated animals (see Table 4), the mean decrease in glycogen synthetase is plotted against the mean increase in phosphorylase Table 6. The sensitivity towards cyclic AMP In a single experiment, mice were injected with various doses of cyclic AMP, together with a glucose load (1 mg/g) or after the administration of prednisolone. The animals were killed 6 min later. Values shown are means & standard error of the means, with the number of animals in parentheses Activity of glycogen synthetase Amount of cyclic AMP given Control Prednisolone Glucose mg/2 g body weight nmoles/min/g livcr 11A5.5 (3) 114&24 (6) 86f25 (4).3 6 f 15 (6) 12 f 1 (6).1 83f24 (6) 24& 9 (5) I Properties of the Phosphodiesterase As a change in the properties of the phosphodiesterase could explain the lowered sensitivity of the prednisolone treated animals towards glucagon, we have assayed this enzyme in the liver of control and treated mice. It was found that the treatment by prednisolone, 3 h before death, did not alter the total activity of the enzyme (mean value: 1.2 pmole/min/g liver) or its affinity for cyclic AMP (K, = DISCUSSION Glycogen metabolism in the liver is controlled by two key enzymes, glycogen synthetase and glycogen phosphorylase ; both of them exist in two forms, one active, the other inactive, which can apparently be interconverted by phosphorylation and dephosphorylation under the action of specific kinases and phosphatases (see [S]). The administration of glucose or glucocorticoids to the animal cakes

6 Vol.6, N.4,1968 H. DE WULF and H. G. HERS 563 the conversion in vivo of glycogen synthetase b (inactive) into a (active) [2,3] simultaneously with a small but significant decrease in the activity of phosphorylase (Table 4). The further administration of glucagon, epinephrine or cyclic AMP induces a rapid reconversion of glycogen synthetase a into b and again a relatively small activation of phosphorylase. The stimulating effects of a glucose load and of glucocorticoids on liver glycogen synthesis are to great extent additive when the two treatments are combined in the same animal (Table 1) and a similar observation has been previously done by other authors [ The same additivity of the effects was observed at the level of glycogen synthetase and it is now clear that the very high rate of glycogen synthesis which is reached under these conditions is entirely explained by a parallel increase in the amount of glycogen synthetase a present in the tissue (Fig. 1). The effects of glucagon and epinephrine can be explained by their known action on the liver adenylate cyclase [17], as cyclic AMP accelerates not only the activation of phosphorylase [17] but also the inactivation of glycogen synthetase [S]. The variations in the activity of phosphorylase induced by glucagon are much smaller than those of glycogen synthetase. This is mostly due to the already very high activity of phosphorylase in a liver homogenate obtained from an untreated mouse. It is now clear that this situation cannot be attributed to a premortern epinephrine or glucagon release, as the synthetase, which is much more sensitive to the hormonal effect, is not inactivated in the same conditions. It is therefore probable that phosphorylase is rapidly activated during the homogeneisation process, and that the activities recorded in vitro are not a true picture of the in vivo situation. One may however assume that the changes in the activity of phosphorylase observed after the administration of glucagon, epinephrine or cyclic AMP, reflect similar modifications which have occurred in vivo. The effect of glucagon on glycogen synthetase k also obtained with a much smaller dose than on phosphorylase. The difference in the sensitivity of the two systems which is apparent in Fig.4 would allow glucagon to stop the synthesis of glycogen before stimulating its degradation. The slow turnover of liver glycogen which is known to occur in normal conditions [18,19] requires a low activity of both synthetase and phosphorylase. From the date of Table 4, it appears that in the mouse this situation would be obtained with about.3 pg of glucagon per 2g of body weight. It is somewhat contradictory, however, that a half maximal effect on dog liver phosphorylase base [17] and on the inactivation of mouse liver glycogen synthetase [S] was 38 European J. Biochem., Vol.8 obtained with nearly the same concentration of cyclic AMP (.2 pm). The extreme hormonal sensitivity of the mice treated with glucose indicates that in these animals the amount of endogenous circulating glucagon has been reduced to a low level; similarly, the high activity of the synthetase in the liver of the animals treated by glucose or glucocorticoids indicates that concentration of cyclic AMP in this tissue is below.1 pm. The stimulation of glycogen synthesis by a load of glucose could therefore be simply explained by a reduction of the secretion of glucagon. Such a reduction has effectively been observed by Unger et al. [2] but has been put in doubt by other authors [21]. It is however important to recall that the activity of glycogen synthetase and presumably also the inhibition of glucagon secretion is by no means proportional to the glycemia but only to its variations and that a fed animal has no more synthetase a than a fasted one [2]. It is of great importance to take these facts into consideration when measuring the effect of glucose on the level of circulating glucagon. Catecholamines could also play a similar role in the regulation of the activity of glycogen synthetase. The animals treated by glucocorticoids were obviously less sensitive to glucagon than those treated by glucose, even when the two treatments were combined. One could assume that the steroid reduces the level of cyclic AMP in the liver (as indicated by the high activity of the synthetase) through an inhibition of the adenylate cyclase or an activation of the diesterase; it could also inhibit the action of the cyclic nucleotide. We have checked that the treatment with prednisolone does not change the total amount of diesterase in the liver nor its affinity for cyclic AMP; there is also no indication that it lowers the sensitivity of the synthetase kinase towards cyclic AMP, neither in vivo (Table 6) or in vitro [S]. If our interpretation is correct, glucose and glucocorticoids would act in a different way on the level of cyclic AMP in the liver, and secondarily on the rate of glycogen synthesis. This would explain why the two treatments have additive effects, glucose lowering the level of glucagon in an animal made already less sensitive to its action by glucocorticoids. A more formal demonstration of our interpretation must await the measurement of blood glucagon and of tissue cyclic AMP after the administration of glucose and of glucocorticoids. This work was supported by the Fonds de la Recherche Scientifique MBdicale and by the U. S. Public Health Service (Grant AM 9235). H. De Wulf is Aangesteld Navorser and W. Stalmans is Aspirant of the Nationaal Fonds voor Wetenschappelij k Onderzoek.

7 564 DE WULF and HERS: Action of Glucose, Glucagon and Glucocorticoids on Glycogen Synthesis European J. Biochem. REFERENCES 12. Fiske, C. H., and Subbarow, Y., J. Biol. Chem. 66 (1925) Glenn, E. M., Bowman, B. J., Bayer, R. B., and Meyer, C. E., Endocrinology, 68 (1961) Rav. P. D.. Foster. D... and Lardv...- H. A.. J. Bid. 6hem. 239 (1964) ' 15. Niemever. H.. Clark-Turri. L.. PBrez, N., and Rabaiille. ", E., 2rch. Biochem. Biophys: 19 (1965) Bishop, J. S., and Lamer, J., J. Biol. Chem. 242 (1967) Do Wulf, H., and Hers, H. G., European J. Biochem. 2 (1967) De Wulf, H., and Hers, H. G., European J. Biochem. 2 (1967) De Wulf, H., Stalmans, W., and Hers, H. G., European J. Biochem. 6 (1968) Hers, H. G., and De Wulf, H., In Control of Glycogen Metabolism (edited bv W. J. Whelan). Universitetsforlaget, Osl; 1968, p Tarnowski, W., Kittler, M., and Hilz, H., Biochem (1964) Kreutner, W., and Goldberg, N. D., Proc. Aratl. Acad. Sci. U. S. 58 (1967) De Wulf, H., and Hers, H. G., European J. Biochem. 6 (1968) Hers, H. G., In Advances in Metabolic Disorders (edited by R. Levine and R. Luft), Academic Press, NewYork, London 1964, Vol. 1, p Wollenberger, A., Ristau, O., and Schoffa, G., Arch. Ges. Physiol. 27 (196) Butcher, R. W., and Sutherland, E. W., J. Bid. Chem. 237 (1962) Foster, D. O., Ray, P. D., and Lardy, H. A., Biochemistry, 5 (1966) Rall, T. W., and Sutherland, E. W., J. Bid. Chem. 232 (1958) Stetten, D., Jr., and Boxer, G. E., J. Biol. Chem. 155 (1944) Shull, K. H., and Mayer, J., J. Bid. Chem. 218 (1956) Unger, R. H., Eisentraut, A. M., and Madison, L. L., J. Clin. Invest. 42 (1963) Lawrence, A. M., Proc. Natl. Ad. Sci. U. S. 55 (1966) 316. H. De Wulf and H. G. Hers Laboratoire de Chimie Physiologique Universite de Louvain Dekenstraat 6, Louvain, Eelgium