Insulin regulation of hepatic glycogen metabolism in the dog€¦ · Insulin regulation of hepatic glycogen metabolism in the dog. Am. J. Physiol. 220(Z) : 499-506. 1971.-The effect

AMERICAN JOURNAL OF PHYSIOLOGY Vol. 220, No. 2, February 1971. Printed in U.S.A.

Insulin regulation of hepatic glycogen

metabolism in the dog

JONATHAN S. BISHOP, NELSON D. GOLDBERG, AND JOSEPH LARNER Departments of Biochemistry and Pharmacology, University of iMinnesota M’edical School,

Minneapolis, Minnesota 55455

BISHOP, JONATHAN S., NELSON D. GOLDBERG, AND JOSEPH LARNER. Insulin regulation of hepatic glycogen metabolism in the dog. Am. J. Physiol. 220(Z) : 499-506. 1971.-The effect of insulin on the activity of the glycogen-synthesizing and degrading enzymes was studied in serially collected biopsy samples of dog liver. In intact normal dogs infusion of glucose alone promoted activation of glycogen transferase and inactivation of phosphorylase as readily as did infusion of insulin with glucose. However, in pancreatectomized dogs these enzymic changes were essentially absent when hyperglycemia resulted either from withdrawal of daily insulin or from infusion of glucose when diabetes was “controlled” by daily insulin. A 30-min infusion of insulin into the pancreatectomized dog promoted less activation of transferase (threefold) in the hyper- glycemic “diabetic” condition than in the insulin-controlled condition (7- to lo-fold). The hepatic concentration of 3’) 5’cyclic AMP in some but not all intact normal and controlled pancreatectomized dogs decreased slightly following insulin infusion, but was not elevated in the diabetic condition. It was concluded that insulin is required for the activation of glycogen transferase in liver.

insulin and liver glycogen transferase; insulin and liver phosphorylase; insulin and 3’) 5’-CAMP in liver; glucagon and liver glycogen transferase; phosphorylase and hepatic glucose output; Leuconostoc mesenteroides strain 39 for blood glucose- 1 -14C analysis; insulin resistance and glycogen transferase; UDP-glucose level in liver; glucose- 6-P level in liver; 3’,5’-CAMP level in liver; pancreatectomized diabetic dogs; hyperglycemia and liver glycogen transferase

RAPID ACTIVATION of glycogen transferase and inactivation of glycogen phosphorylase in liver has been shown to result

after insulin infusion into the intact dog (3). In earlier experiments it was found that insulin infusions led to a cessation of glycogen loss and an accumulation of liver glycogen composed of glucosyl units which derived from 14C-labeled plasma glucose (4).

These observations in the whole animal indicated that insulin affected the hepatic cell directly. This conclusion was consistent with the work of others who have shown apparent hepatic effects of insulin in the isolated perfused ,organ ( 12, 16, 24, 25), even though the latter observations were related to the metabolism of glucose, nitrogen, and fat rather than specifically of glycogen.

Since glucose must be infused along with insulin in the whole animal to prevent the development of a counter- insulin reaction consequent to the induced hypoglycemia (7), it is reasonable to question whether the observed effects resulted from the action of these compounds sepa-

rately or in combination. It has been suggested by DeWulf and Hers (8) that a high glucose concentration in blood might decrease the basal secretion rate of glucagon, a known insulin antagonist. On the other hand, evidence has appeared indicating that in the isolated perfused liver the perfusate concentration of glucose alone may regulate metabolism of glycogen (6, 10).

This is a report of further studies of liver glycogen transferase and phosphorylase activity during insulin infusion which were carried out to correlate the enzymic changes with previously observed changes in liver glycogen and glucose metabolism in the whole animal. In addition, by using pancreatectomized dogs, an attempt was made to separate the effect of hyperglycemia alone from that of insulin. A possible role for 3’) 5’-cyclic AMP in the development of the altered enzyme activity was investigated by determining the liver concentrations of this cyclic nucleotide during insulin infusion.

METHODS

Care of Animals

Adult, mongrel, male dogs of the retriever type, weigh- ing about 20 kg, were obtained from the Mayo Animal Hospital, after being dewormed, vaccinated, and quaran- tined for at least 1 month. A dry commercial dog chow (Ralston Purina Co., St. Louis, MO.) was fed ad libitum or, for studies including measurements of hepatic glucose output, was weighed out to provide about 70 Cal/kg per day, of which carbohydrate accounted for no more than 48 % and protein no less than 29 %, according to the manu- facturer’s analysis. No food was given during the 18 hr prior to the test.

Pancreatectomized dogs were not used until at least 3 weeks after surgery or until all wounds had completely healed. Commercial chow, supplemented either with frozen raw pancreas or with a commercial pancreas extract (Cotazym, 0 g r anon, Inc., West Orange, N.J.) sufficient to prevent excessively large stools, promoted a weight gain to almost preoperative levels. In the ‘ccontrolled” pancreatectomized animals sufficient commercial insulin, about 1 U/kg twice daily, was given to limit glycosuria to about 5 g/24 hr as determined by the glucose method of Somogyi (26). For studies in the “diabetic” state, daily insulin was withheld for 2-4 days, and urine glucose output of more than 60 g/day was observed.

499

by 10.220.33.4 on July 13, 2017http://ajplegacy.physiology.org/

Dow

nloaded from

http://ajplegacy.physiology.org/

500 BISHOP, GOLDBERG, AND LARNER

Experimental Procedures

Serial liver biopsies were collected, at the intervals indicated during control and experimental periods, by percutaneous needle biopsy. (The Menghini needle was obtained from Societa Italiana (ICO), Bologna, Italy.) Local nerve-block anesthesia (1% mepivacaine hydrochloride, (4), or general anesthesia from intravenous sodium thi- amylal (Surital, Parke-Davis), was used without supplemental respiration and with the dog either standing or suspended in a Pavlov sling. In six experiments (three of which are included in Table 1 and three others, G, H, and 1, in Tables 2 and 3) general anesthesia from sodium pen- tobarbital was used for collection of larger liver samples through a laparotomy incision. The pieces of liver were quickly blotted and frozen in less than 15 set after removal by immersing them either in isopentane, at - 150 C or alternatively, by compressing them between aluminum blocks which had been precooled in liquid nitrogen.

Insulin (low in glucagon)l was freshly dissolved in 0.15 M

sodium chloride, pH 3.0, and infused either separately (Table 1) or upon addition to an 18 % solution of glucose, pH 3.0 (Tables 3, 4).

Hepatic glucose -12C was measured in some experiments (Table 1) by an isotope-dilution technique (4). For this, the body glucose pool was labeled to constant specific activity with trace amounts of glucose-l-l 4C2 (for instance 4.3 PC as an intravenous priming injection and 0.04 &min infused into a saphenous vein throughout the 5- to 6-hr experiment). After 60 min had elapsed to allow for mixing, samples of jugular venous blood were withdrawn into heparinized syringes at 15- to 30-min intervals through an indwelling polyethylene catheter. The plasma was immediately separated by centrifugation and frozen. The glucose concentration in a 1 : 15 diluted protein-free filtrate of the plasma (27) was determined with glucose oxidase. The amount of 14C specifically located in the carbon-l of the plasma glucose was determined by measuring the 14C02 evolved during fermentation of the sample using Leuconostoc mesenteroides, strain 39, according to the method described in the APPENDIX. During the 2nd and 3rd hr of the isotope infusion, control samples of blood and liver were collected and the insulin (0.04 to 0.20 U/kg per hour) was administered, first as a rapid injection (2.0 ml in 2 min) and then as a slower constant infusion (0.135 ml/min). In order to limit the degree of hypoglycemia brought about by the insulin, a continuous glucose infusion (400-600 mg/kg per hr) was started about 15 min after the start of the insulin infusion. Glucose-l -14C was added to the glucose infusion so that the specific radioactivity of the infusion approxi- mated that of the plasma glucose. In other experiments

(Tables 3, 4) the glucose concentration in whole blood, after treatment with oxalate and sodium fluoride, was determined by an automated ferricyanide method (Techni-

con). Hepatic glucose-6-P and UDPglucose concentrations

were determined microfluorometrically, using glucose-6-P

l Trypsin-treated amorphous insulin and crystalline glucagon were the gift of the Eli Lilly Co., Indianapolis, Ind.

2 Nuclear Research Chemicals, Inc., Orlando, Fla. Specific radioactivity 15.6 mc/mM.

dehydrogenase or UDPglucose dehydrogenase (Sigma Chemical Co.). For this purpose, neutralized perchloric acid extracts of frozen powdered liver were prepared according to Lowry et al. (2 1). The hepatic concentrations of 3’ ,5’-cyclic AMP (CAMP) were determined by the method of Goldberg et al. ( 11).

For assay of glycogen transferase and glycogen phosphorylase activity the frozen sample (about 20 mg) was homogenized for 2 min in 24 volumes of 250 mM sucrose, 5 mM EDTA, 50 mM KF, and 10 mM Na2S03, pH 7.4, and centrifuged for 20 min at 1,800 X g at 2 C. The superna- tant extract was immediately analyzed for transferase activity as the rate of transfer of glucosyl-14C units from labeled UDPglucose to glycogen according to either the method of Villar-Palasi et al. (3 l), or this method as modi- fied by Thomas, Schlender, and Larner (29). For transferase I activity, incubations were for 15 and 30 min in the test mixture minus glucose-6-P; and for total transferase (I + D), incubations were for 5 and 10 min in the test mixture plus 7.3 mM glucose-6-P. The 1: 25 sucrose extract of liver was then diluted with 5 (or 10) volumes of 33 mM

glycerol-P, 50 n?M KF, and 50 mM mercaptoethanol, pH 6.1. Phosphorylase activity was assayed in the direction of glycogen synthesis ( 13) immediately after dilution using incubations for 10 and 20 min at either 30 C (Tables 3 and 4) or at 37 C (Table 1). The protein content of the diluted extract was determined by the method of Lowry et al. (ZO), following precipitation in 5 % trichloroacetic acid to remove reducing substances.

RESULTS

Studies in Intact Dogs

It was reported earlier (3) that administration of insulin in relatively high doses (2.0 p/kg per hr) resulted in both transferase activation and phosphorylase inactivation. To determine whether the two enzyme systems were equally responsive to this hormone, the activities of these enzymes were determined after infusion of the lowest dose of insulin that would decrease hepatic glucose output (4). Table 1 summarizes the results from six dogs in which seven experiments with a low rate of insulin infusion (0.4 to 0.20 U/kg per hour) resulted in a 40-60 % decrease in hepatic glucose output, similar to that reported elsewhere (4). In these experiments a 3- to lo-fold increase in transferase I activity (from 2 & 1% to 9 =t 2 % I/total) was associated with a 25% decrease in phosphorylase activity (from 7.6 rfi 0.5 to 5.6 + 0.7 pmoles inorganic P liberated per hour per milligram protein). For comparison, in four experiments using three of these same dogs (reported in part previously (3) a higher rate of insulin infusion ( 1.3-2.7 U/kg per hr) resulted in a greater increase in transferase I (to 20 =t 6 %) and a greater decrease in phosphorylase (to 2.5 -+: 0.3 pmoles P/hr per mg). Measurements of 3’ ,5’-cyclic AMP concentration in these same liver samples showed no significant change when greater insulin effects were produced with the lo-fold greater rate of insulin infusion. Glucose (0.4-l .5 g/kg per hr) was infused along with the insulin, since it is known that severe hypoglycemia will obliterate at least some effects of insulin upon liver glucose metabolism (7). The average plasma glucose concentration during the infusions


Dow

nloaded from


INSULIN REGULATION OF HEPATIC GLYCOGEN METABOLISM 501

TABLE 1. Efect of insuhn-with-glucose infusion on liver on intact normal dogs, anesthetized and unanestheti<ed

Infusions, per kg per hr

Saline -20 to -5

Insulin, .04-.02 units 35 to 1oc

+ Glucose, 0.4-0.6 g

Insulin, 1.3-2.7 units

+ Glucose, 1.5 g

7 to 25

Trans- f erase, % I !

Total!

7.6 zto.5

(9)

5.6 zto.7

(6)

20 2.5 zt6 zto.3

(4) (3 >

- Phos-

phorylase (-AMP)

-

kg

1.18 zto.ll

(6)

1.34 &O. 14

(6)

1.22

(2)

Glucose

Hepatic output, w/kg per hr

192 zt12

(6)

105 zt22 (6 >

Plasma zoncn, w/;oo

100 zt4

(11)

101 A13

(7)

101 zt23

(4)

“Min- Values are means ZIZ SE with no. of observations given in parentheses. utes from start of insulin infusion; the longer infusion time before biopsy was allowed for a more complete development of the insulin effect with low concentrations of insulin (with glucose). (See Fig. 1:) t Transferase I activity averaged 0.15 AZ 0.04 pmoles UDPglucose transferred to glycogen per minute per gram protein during the control period, and 0.77 =t 0.29 and 1.80 A 0.15 during infusion of low and high concentrations of insulin (with glucose), respectively. Numbers in parentheses represent the total number of samples analyzed.

was the same as in the control period ( 100 mg/ 100 ml). In three experiments, in which hypoglycemia (45-55 % of control value) was observed, slightly less change in enzyme activity was seen than in five experiments in which the plasma glucose concentration was elevated an average of 30 mg/ 100 ml above the control value. Figure 1 shows in detail one of the three experiments in which a glucose infusion (0.6 g/kg per hr), started 17 min after the beginning of a low-rate of insulin infusion (0.20 U/kg per hr), kept the plasma glucose at essentially the control level. There was a progressive increase in the I form of transferase during the 90-min infusion period (from 2 to 16 %). The small increase in the activity of transferase measured in the presence of glucose-62 (i.e., total) under the assay conditions employed may be ex- plained, at least in part, by the previously described changes in the kinetic characteristics of the enzyme (3) that accom- pany the conversion of the D to the I form. A decrease in phosphorylase activity was more clearly seen when the assay was done without added 5’-AMP (1.5 mM) in the test mixture because it is known that this nucleotide stimulates both the active and the inactive form of liver phosphorylase an equal amount (33). Th e d ecrease in UDPglucose concentration (from 0.3 1 to 0.16 mmoles/kg) along with transferase activation has been shown to be associated with increased glycogen synthesis ( 15, 18). The glucose-6-P levels (0.24 mmoles/kg) were unchanged or increased during insulin infusion. When the insulin-induced enzyme activity changes were reversed, as when insulin was stopped and a glucagon infusion was started, glucose-62 concentration increased to 1.05 mmoles/kg. It was concluded that the systems for activating transferase and for inactivating phosphorylase were probably equally sensitive to insulin.

In order to determine the extent to which these two enzyme changes, brought on by the administration of insulin (Table 1) correlated with observed rates of glycogen breakdown and synthesis in vivo, these parameters have been calculated from the observed enzyme activities. Table 2 is a

summary of such catalytic capacities and compares them to glucose output measurements made in these experiments and in 11 previous experiments in which changes in liver glycogen content and in glycogen labeling with plasma glucose- 14C were measured directly (4). The potential of liver phosphorylase to catalyze glycogen breakdown (9,120 pmoles/hr per 30 g liver) was 9 times greater than the observed rate of glucoseJ2C output from the liver ( 1,066 pmoles/hr per kg body wt). The rate of glucoseJ2C loss from glycogen (870 pmoles/hr per 30 g liver) nearly equaled the rate of hepatic glucose-12C output ( 1,000 pmoles/hr per kg body wt). The smaller amounts of infused insulin (0.04-0.20 U/hr per kg) reduced glucose output (to 55 % of the control) and also phosphorylase activity (to 73 % of the control), whereas with 10 times this amount of insulin, phosphorylase activity decreased further (to 33 % of the control). Transferase I activity increased severalfold after about 70 min of the infusion with lower insulin dose, as did the incorporation of plasma glucoseJ4C into glycogen (from 24 and 6 to 118 and 116 pmoles/hr per 30 g liver, respectively). After a longer period (about 137 min) of the infusion with lower insulin dose, the rate of glucose- 14C incorporation into glycogen had increased further (to 495 ,umoles/hr), but with the higher insulin dose such a greater increase in transferase I activity (to 275 pmoles/hr per 30 g liver) was observed more quickly (after about 15 min).

Glucose alone ( 1.5 g/kg per hr), infused for 30 min (Fig. 2A and Table 3 experiment B-5), promoted rapid transferase

my moles - mln x mg prot.

PHOSPHORYLASE 6 _ pmoles Pi hr x mg prot.

”

1 1.0 GLUCOSE-6-PO,+ O-6?

UDP-GLUCOSE

150 PUWAGUJCOSE

ma/lOOml “L-?w-- 100 d

50. i

’ CONTROL1 I N GLUCOS t

s u L I N -120-60 0 20 40 a 80

TIME IN MINUTES

‘-0 ? -12

-10

-8 -6

-4

-2 SO

Q ,y - 0.6

-0.4 -0.2

’ 0 -200

-150 - 100 -50

,O - 150

- IO0

-50

FIG. 1. Effect of insulin infusion on liver. An anesthetized intact dog was infused with trace amounts of labeled glucose for measurement of hepatic glucose -12C output, starting 170 min before infusion of insulin (0.2 U/kg per hour). Serial liver biopsy samples, obtained through an abdominal incision, were collected at the times indicated (see METHODS). Infusion of glucose (0.6 g/kg per hr) was started 17 rnin after insulin. After insulin and glucose infusions were stopped glucagon was infused (0.010 mg/kg per hr).


Dow

nloaded from


502

TABLE 2. Changes in hepatic glycogen and glucose metabolism during insulin administration

Insulin Infusion Rate W/kg per W

Biopsy time, min” a

b Transferase I (- Glc-

6-P) t (pmoles/hr per 30 9) a

Phosphorylase (-AMP) t pmoles/hr per 30 g a

Glucose J2C output, pmoles/hr per kg a

b

Glucose-12C from glycogen, pmoles/hr per 30 g b

Glucose-l*C into glycogen, pmoles/hr per 30 g b

Control

24&6 (11)

3,120%6OC (11)

1,066*66 (6)

1 ,OlOZt70 (11)

87O~t88 (11)

6*6 (11)

Experimental Periods

35-100

69& 14

118&30 (7)

i, 720&84C (6)

583& 12: (6)

561&77 (11)

148~#~llC (11)

116*16 (11)

.20 1.3 - 2.7

137ztll

303&20 (11)

236&248 (11)

495rt 104 (11)

7-27

!75&23 (4)

3,0001t360

P 1 . Values are means & SE with no. or observations givenin paren-

theses. *Minutes from start of insulin infusion, the data re- corded in line a are calculated from the data of Table 1, this paper. The data of line b are from Bishop et al. (4), Table 4, in which it was found that less than 50 mg plasma glucose equivalents/30 g liver per hr were incorporated into liver glycogen during early time periods after the start of the insulin infusion (mean 69 rt 14 min). More than 50 mg plasma glucose equivalents per 30 g liver per hr were incorporated in late time periods (mean 137 =t 11 min). tFactors used to convert specific enzyme activity (Table l), measured in 25-fold diluted extracts, which averaged 4 mg protein/ml, to conditions in undiluted tissue were: X 153,000 for transferase I, and X 1,200 for phosphorylase. The following rate-modifying assumptions are included in these factors: For transferase I : a) activity at 30 C X 1.7 = activity at 37 C; b) activity at 4.5 mM UDPG X 0.5 = activity at 0.2 mM UDPG, assuming that the intracellular UDPG concentration (during insulin) is 0.2 mM (Fig. l), that the K, for UDPG of transferase I is 0.2 mM (3), and that the concentration of Glc-6-P in liver (0.3 mM) is maximally effective in stimulating transferase I at this nonsaturating concentration of UDPG. Sample calculation :

.15 mpmoles 4 mg protein 25 ml extract min X mg protein

X ml extract

X g liver

X. 30 g liver

kg body wt X 6y x 1.7 X 0.5 =

pmoles 24 X ~

hr X 30 g

For phosphorylase : a) the velocity in the direction of glycogen breakdown is 40y0 of the velocity in the direction of glycogen synthesis (22). No adjustment is made for kinetic influences of concentrations of glycogen and orthophosphate since liver Pi is 7-9 mM (22) and K, for Pi (at glycogen 0.5 9%) is 1.1 mM (23). The calculated rate of glycogen synthesis ascribed to transferase D was ac- tually decreased by insulin (e+ 1.5 of ref. 3, data not shown separately here but included in Table 1) from a control value of 41 to

BISHOP, GOLDBERG, AND LARNER

activation and phosphorylase inactivation in the intact dog. To consider the hypothesis ( 10) that these effects of glucose infusion were brought about more because of an increase in the circulating glucose concentration than by an increase in the secretion rate of endogenous insulin, comparable infusions were administered to pancreatectomized dogs. It was recognized that an alternative hypothesis (8) would not be tested with this preparation, namely that infused glucose promoted these reciprocal changes in enzyme activity indirectly by promoting a decrease in basal glucagon secretion, since the glucagon-secreting alpha cells as well as the insulin- secreting beta cells would be absent in the pancreatectomized animal.

Studies in Pancreatectomized Dogs

Dogs con trolled by daily maintenance insulin injections. The effects of insulin administration in four pancreatectomized dogs, given sufficient daily insulin to control excessive glycosuria, were compared with those observed in three intact dogs (Table 3). During the control period ( 18 hr after the last meal and the last insulin injection) transferase I activity was low (less than 8 % of total), but phosphorylase activity, which was much more variable, was about twice that found in the intact dogs. Infusion of glucose alone (1.5 g/kg per hr) for 30 min promoted little or no change in transferase I activity in three pancreatectomized dogs, compared to the 15-fold increase in the one intact dog (B-5) tested in this series of experiments (Fig. 2). Inclusion of insulin (3.2 U/kg per hr) in the subsequent 30 min of glucose infusion (2.5 g/kg per hr) produced a 7- to lo-fold activation of transferase in these experiments. A comparable degree of transferase activation was seen in both the intact and the pancreatectomized dogs when the same amount of insulin was included in the first 30 min of a glucose infusion (experiments A-6, G, and 1). Whole-blood glucose level in the pancreatectomized dogs was increased from about 250 mg/ 100 ml in the control period to about 450 mg/ 100 ml during the glucose infusion and to about 630 mg/ 100 ml during the infusion of insulin with glucose. Phosphorylase activity during the infusion of glucose alone was variable but somewhat lower than in the control period and was decreased further when insulin was added to the glucose infusion. No explana- tion can be offered either for the variability in phosphorylase activity, which was especially apparent in the control

less than 22 pmoles/hr per 30 g liver. The conversion factor (X 3,940) included the following rate-modifying assumptions : a) activity at V,,, + 13 = activity at 0.3 mM UDPG. This relationship was reached by using the Michaelis-Menten equation: v = V,,,/ (1 + Km/s) and assuming K, for UDPG to be 3.6 mM (control sample) and V,,, (control and insulin samples) 10.6 mpmoles/min per mg protein, as observed in this experiment (see Fig. 1 of ref. 3), and assuming the intracellular UDPG concentration (control) to be 0.3 mM. b) Stimulation of transferase D by Glc-6-P to be V,,, + 4, assuming intracellular Glc-6-P concentration to be 0.22 mM as observed (Fig. 1) and assuming K, for Glc-6-P = 0.67 mM (31). After insulin, not only did the UDPG level decrease (Fig. l), but also the amount of the D form of transferase (because of the conversion of D to I), so that the rate of glycogen synthesis ascribable to this form of transferase would have been correspondingly lower.


Dow

nloaded from



TABLE 3. E$ect of glucose, and insuhz-with-glucose infusions on liver qf intact dogs and of insulin-controlled pancreatectomi<ed dogs

’ Glu- Saline Glucose Insulin + Glucose ca-

Cond i- &Pn tion Exp. No.

-20 -5 5 5 30 6 min min min min min min

Intact

Mean

Pan-x

Mean

Intact

Mean

Pan-x

Mean

Intact

Mean

Pan-x

Mean

B-5 B-6 G

I 5 (C-l) 6 (C-l) 7 (C-l)

B-5 B-6 G

I 5 (C-I) 6 (C-l) 7 (C-l)

B-5 B-6 G

I 5 (C-l) 6 (C-I) 7 (C-I)

5 5 3 2 8 7

5 &l(3)

3 2 1 3 4 2 2

2 rto.5 (4)

Transferase, 70 I/ Tota I

2

70 80 1

75(l)

31 40 43 63

45W

2 2 5 8 2 3

4&l (3)

15 31 17 31 70 55 49 70

5 21 11

42&11(4)

Phosphorylase, pmoleslhr per mg

9.0 7.7 3.0 3.5 6.4 7.6

. 3.6

6.3&l .4(3)

;;;I;:! 165

3.3(l)

. .

. . . 11.1 10.1

. . 16.1 ~ 17.8

8.3 5.2 18.7 18.3

18.7 &2.2(4) 11.4 zt2.4(3)

3’) 5’-Cyclic AMP, pmoles/kg

1.701 1.70 1.63 1.58 0.78 1.09 1.63 2.89

1.63 A. 12(3) 1.60( 1)

; ; ! I I#;; 1.60 1.54 1 ;iii

1.87 =f=.28(4) 1.67zt.10(3)

3.8

5.5 2.7 2.4

0.91 2.38 1.17

3.0 3.4 9.2 6.6

5.6(2)

9.3 9.1 4.7 5.5 7.5 2.2 1.0 2.3

4.6&1.7(4)

0.82 0.78 1.45 1.94

1.25(2)

1.31 1.45 0.75

3.02 1.65 1.22 0.86

1.41 zt. lO(4)

17.9

54.2 47.5

Numbers in parentheses are numbers of experiments. Serial liver biopsies were collected at the indicated times during infusions of glucose (2.5 g/kg per hr) or of insulin (3.2 U/kg per hr) with glucose (2.5 g/kg per hr). Glucagon (.020 mg/kg) was injected after insulin-with-glucose infusion was stopped.

periods, or for the increase with insulin in one dog (dog G), that had abdominal surgery.3

The concentration of 3’, 5’-cyclic AMP in the livers of the six animals, comparably tested (Table 3, except experiment A-S), was decreased from 1.6 1 + .19 pmoles/kg in the control (and glucose infusion) periods to 1.36 + 25 pmoles/kg during the infusion of insulin (3.2 U/kg per hr) with glucose.

3 Whereas transferase I activity consistently increased, phosphorylase activity was unchanged (2 experiments of Table 1) or increased (exp G, Table 3) during insulin infusion in some early experiments in which liver samples were collected directly from the liver exposed by abdominal laparotomy. It was observed that a more consistent pattern of reciprocal changes in enzyme activity was seen if liver samples were collected by percutaneous needle biopsy, using general or local anesthesia. Thus in order to avoid adverse influences possibly related to the surgery, and because repeated studies could be done easily in the same dog, the percutaneous needle biopsy technique was used for most experiments.

This was an average decrease of 16 s+ 8 % from control values in the individual experiments. However, with smaller amounts of infused insulin (Table l), or, in one instance in which endogenous insulin was evoked by a glucose infusion

INTACT PANCREATECTOMIZED

1.0 r

t 1 r t

GLUCAGON GLUCAGON 1

MINUTES

FIG. 2. Effect of glucose infusion on liver of intact and of pancreatectomized dogs. Infusion of glucose (1.5 g/kg per hr) into anesthetized dogs, either a) intact or b) pancreatectomized and maintained with- daily insulin injections. Serial liver sampling by percutaneous needle biopsy was done at times indicated (see METHODS). After 30 min of glucose infusion, in the pancreatectomized dog (b), insulin (3.2 U/kg per hr) was added to the glucose infusion (increased to 2.5 g/kg per hr). After glucose and insulin-with-glucose infusions were stopped, glucagon (.020 mg/kg) was injected.

TABLE 4. E$ect of insulin-with-glucose infusion on her of diabetic pancrea tectomised dogs

Saline or Glucose Insulin + Glucose

Exp No.

-10 min 15 min 1 30min IISminl 90 min

Transf erase, yO I/ Total

H 4 9 12 14 14 5 (D-I) 5 [ll 7 8 4 (D-1) 2 17

Mean ZJZ SE 4&l 12 & 3

Phosphorylase, pmoles/hr per mg

H 22.2 4.8 4.5 7.3 6.0 5 (D-I) 3.5 [7.2] 4.3 4.5 6.6 6 (D-l) 3.5 1.6

Mean & SE 10 * 5 4&l

3’) 5t-Cyclic AMP, pmoles/kg

H 1.17, 0.90 1.37 1.05 1.17 0.80 5 (D-I) 0.75, 1.15 6 (D-I) 2.62, 2.91

Mean rt SE 1.59 It 0.47

Serial liver biopsies were collected at the indicated times during infusion of insulin (3.2 U/kg per hr) with glucose (2.5 g/kg per hr). Inoneexperiment, 5 (D-I), glucose (2.5 g/kg per hr) was also infused for 30 min before the insulin was added; the data are within the brackets. Daily insulin was withheld for 4 days except for experiment H (2 days only).


Dow

nloaded from


504

alone (Table 3, experiment A-5), no decrease in CAMP concentration was observed. A 30-fold rise in CAMP concentration was observed when glucagon ( lo-20 pg/kg) was injected after stopping the insulin infusion (Table 3, experiments 5 (C-l) and 6 (C-l)), w h ereas, if a comparable insulin infusion was continued in two other experiments (Table 1), only a two- to threefold rise was observed (namely, from 1.27 and 1.16 to 2.27 and 3.20 pmoles/kg, respectively).

Diabetic dogs after withdrawal of maintenance insulin. Three pancreatectomized dogs were allowed to become insulin deficient, with high glycosuria (more than 60 g/day), by withholding daily insulin injections for 2 or 4 days (Table 4). A comparable infusion of insulin (3.2 U/kg per hr) plus glucose (2.5 g/kg per hr) promoted an increase in activity of transferase (from 4 =t 1% I/total in control periods to 12 & 3 % within 30 min), which was less than the increase (3 % to 45 %) observed with insulin administration in the controlled state (Table 3). The phosphorylase activity in the control periods was too variable to provide for meaningful inter- pretation but could be shown to decrease more than half in two of the three cases. Whole-blood glucose concentration was elevated (440 mg/ 100 ml) and remained high during the infusion (640 mg/ 100 ml).

Because the hepatic concentration of CAMP has been found to be elevated in the diabetic state induced with alloxan in the rat ( 17), it seemed likely that the apparent insulin resistance of the diabetic dog may have been the result of a similar phenomenon. However, this possibility was not borne out by the finding that the concentrations of CAMP observed in liver were no higher in the diabetic ( 1.59 & .47 pmoles/kg) than in the controlled pancreatectomized dog (or in the intact dog) (1.87 r+ .28 and 1.63 + .12 pmoles/kg, respectively). This relationship is more clearly shown in the case of dogs 5 and 6 where the two conditions were observed in the same animal.

DISCUSSION

The results of these experiments indicate that activation of liver glycogen transferase when the intact dog is infused with insulin and supplemental glucose (3) is promoted by the insulin itself, rather than by the glucose which was in- tended to prevent the development of hypoglycemia. This effect of the hormone was shown under two conditions which greatly reduce the possibility that a decrease in basal glucagon secretion, promoted by the infused glucose, was more important than the infused insulin in activating transferase. In one condition, the amount of glucose infused with the insulin was not sufficient to produce hyperglycemia in 6 of the 11 studies in intact animals (Fig. 1 and Table 1). Thus, in these experiments, it would not seem reasonable to expect that a decrease in glucagon secretion occurred.

In the second condition extra insulin was capable of activating transferase under circumstances when hyperglycemia alone was ineffective, namely, when both the insulin-secreting and the glucagon-secreting cells were removed by pancreatectomy (Fig. 2B and Table 3). This insulin-controlled pancreatectomized animal was not severely insulin deficient, since near-normal blood glucose levels were maintained by the daily insulin injections before the daily hormone administration was terminated. How- ever, there was a chronic deficiency of pancreatic glucagon.


In these experiments the activation of glycogen transferase was apparently mediated more directly through the action of insulin itself than through a pathway which would in- volve an influence of glucagon via 3’, 5’-CAMP on transferase I kinase. Nevertheless, it is conceivable that in the intact animal (Fig. 2A) a decrease in basal glucagon secretion could potentiate, but not necessarily initiate, the activation of transferase, which was shown here to be promoted by extra circulating insulin, whether secreted or infused. Simi- lar demonstrations of an insulin requirement for transferase activation following hydrocortisone administration to ad- renalectomized rats ( 18) and following puromycin administration to intact tadpoles (5) have been reported.

In the light of these experiments three possible mecha- nisms by which insulin may act on liver directly to promote transferase activation could be considered.

Decrease in Basal Transferase I Kinase Activity Secondary to a Decrease in Cellular 3’ , ~/-CAMP Concentration

That insulin promotes reciprocal changes in the activity of transferase and phosphorylase in liver makes attractive the hypothesis that both enzyme changes are promoted by a common mechanism, such as a lowering of liver concentrations of 3’, 5’-cyclic AMP. Cyclic AMP is a known stimula- tor in vitro of ki .nases in liver and in muscle which promote (indirectly) the activation of phosphorylase and inactivation of transferase (i.e., a decrease in transferase activity in the 1 form).

The implication of this hypothesis is that insulin affects the adenyl cyclase and/or phosphodiesterase which regu- lates the tissue concentration of CAMP. The total experience reported here ( 14 experiments) indicated no significant change in CAMP concentrations when transferase was activated to varying degrees by amounts of insulin extending over more than a 20-fold range of concentrations in the infusion ( 1.39 + .36 and 1.3 1 + .30 ,umoles/kg, respectively). However, in the six experiments with the largest insulin infusion (3.2 p/kg per hr) with glucose, a small progressive decrease in liver CAMP concentration was seen which was of a similar degree to that observed by Exton and Park ( 17) following perfusion of a normal rat liver with insulin. These investigators consider such a small decrease to be physio- logically relevant, since a major portion of the acid extract- able cyclic AMP may be biologically inert (see discussion by Larner ( 19)). These small decreases in CAMP concentration that were sometimes noted after insulin administration could conceivably influence the system on the basis of the compart- mental concept referred to above. The present experiments support an action of insulin on some aspect of cyclic AMP metabolism, since insulin administered at the same time as glucagon (Table 1, footnote) resulted in a much smaller rise in CAMP concentration than occurred when glucagon was given alone (Table 3). However, the possibility still remains that other factors consequent to insulin administration may be more directly related to the changes in transferase I and phosphorylase activity when these are brought about under more physiological conditions (i.e., not in con- junction with large doses of glucagon).

Furthermore, the relevance to insulin action of a change in cellular CAMP level has been questioned more recently by

the experiments of Nichols and Goldberg (unpublished


Dow

nloaded from



data), as quoted by Villar-Palasi et al. (30). These investigators found, in the early periods following insulin administration to rats, that liver transferase I activity apparently increased minutes before a measurable decrease in liver CAMP concentration was detectable.

Alteration of Activity State of One of the Interconuerting Enzymes for Glycogen Transferase

Inactivation of transferase I kinase. These experiments did not test, but do not exclude, the possibility that in liver, as in muscle (32), insulin administration promotes the trans- formation of the kinase of transferase I to a less active state, as discussed in detail by Larner et al. ( 19). This would con- stitute an action stemming directly from insulin itself on this interconverting enzyme of the transferase system. However, it appears (vide infra) that an activity of insulin can influence the other interconverting component (i.e., the phosphatase) of this system in liver, even though such an influence has not yet been apparent in muscle ( 19).

Activation of transferase D phosfihatase. A 2- to 4-fold increase in the activity of transferase D phosphatase was observed during the infusion of insulin with glucose, when the activity of this enzyme was measured in some of the sample shown in Table 3 (1). Th e insulin dependence of the activation of phosphatase was apparent from the fact that phosphatase activation did not occur during comparable insulin-with- glucose infusions in animals that were in the diabetic state not controlled with exogenous insulin for over 96 hr, even tholqh small increases in transferase I activity were observed in this condition (Table 4).

High cellular glucose concentration alone may sometimes result in transferase activation in the intact dog, as found by Glinsmann et al. ( 10) and others (6) in the isolated perfused rat liver. For instance, a twofold increase in transferase activity was observed during the 30-min glucose infusion in one experiment (Table 3, experiment 6 C-1)). This partial activation of transferase was associated with the gradual development of an increase in the activity of transferase D phosphatase, which is reported elsewhere ( 1). Addition of insulin to the glucose infusion in the subsequent 30 min resulted in no further increase in phosphatase activity, but did result in further activation of transferase. From the sequence of the development of the changes in activity of these two enzymes in this experiment, it is clear that phosphatase activation alone is not sufficient for maximal transferase activation. On the other hand, in two other experiments (Table 3, experiment 7 (C-1) and Table 3, dog 6 in ref. 1) no such activation of phosphatase (or transferase) was seen with glucose infusion until insulin was added. The mechanism of this occasional effect of glucose to promote an increase in the activity of transferase D phosphatase is not known.

In the diabetic pancreatectomized dog (Table 4) in which there was low activity of transferase D phosphatase (l), transferase was not activated by the high basal glucose concentration. In this condition infused insulin produced only partial activation of transferase in 30-120 min. However, this activation of transferase is brought about without the associated activation of phosphatase which occurs in the normal. The mechanism of the activation of transferase in this condition (i.e., chronic insulin deprivation) is not

known. On the other hand, it has been suggested elsewhere (1) that the sluggish activation of transferase during insulin infusion in the diabetic pancreatectomized dog represents a type of insulin resistance which may be due to the inability of the phosphatase to become rapidly activated, since no abnormal elevation of cyclic AMP concentration in the liver was observed.

In summary, it appears that activation of liver glycogen transferase by insulin (and in the presence of glucose) may operate through one mechanism or a combination of mecha- nisms, either individually or in concert. Further study will be required to understand the molecular events in this com- plex regulatory system.

APPENDIX

Rapid Method for Determining 14C Content of Samples Containing Glucose-I-C by Fermentation of Glucose with L. Mesenteroides, Strain 39

Procedure. To 5.0 ml of a 1: 15 Somogyi filtrate of plasma (containing 0.1-0.5 mg glucose) in a 125 ml Erlenmeyer flask, 7.5 mg glucose-12C and 0.6 mmoles phosphate buffer, pH 5.9, were added in a total volume of 13 ml. The flask was flushed with nitrogen for 1 min before addition of 0.5 ml of a 20y0 resuspension of twice-washed cells of L. mesenteroides, strain 39,4 harvested in the active fermenting phase (9); the washed pellet of bacteria can be stored frozen under distilled water at -90 C for at least 9 months without loss of activity. Hyamine hydroxide (0.15 ml of 0.8 M solution in methanol) was added to a small glass cup in a wire basket suspended from a rubber serum stopper which seals the flask. After incubation at 30 C with shaking for 4 hr, the cup was removed, wiped free of condensed water vapor, and placed in a counting vial. The hyamine, having solidi- fied with loss of methanol and absorption of water, was redissolved in 1.0 ml absolute ethanol overnight, and then 15 ml of 0.5% PPO-

toluene were added and the 14C was counted in a liquid scintillation spectrometer. The 14C measured was corrected for the completeness of fermentation, usually 93-96(%, observed with a sample of authentic glucose- 1 -14C of known 14C content, determined as the glucosotriazole. Analyses have proven unreliable, and are discarded, if the bacteria were able to accomplish no more than 90% fermentation of the standard.

Evaluation. In a separate experiment the specific activities of 12 plasma glucose samples and two authentic glucose- 1 -14C solutions were determined both by this method and by the method of liquid scintillation counting of plasma glucose isolated as the glucosotriazole derivative (28). The average range of 14C content determined in triplicate fermentations (AZ. 1%) compared favorably with that for triplicate weighings of the singly isolated glucosotriazole (& 3.0%). Thus, compared with the previous method, the fermentation method was at least 3 times as rapid and no less accurate. Furthermore, because of the specificities of the bacterial enzymes (9) the only 14C02 produced comes from C-l of the glucose. Therefore, the values for glucose -12C output were not lowered by measurement of whatever 14C would recycle into the other five carbons of plasma glucose, such as results from glucose catabolism and resynthesis from labeled 3-carbon compounds.

The excellent technical assistance of Miss Helene Sasko, Mrs. Anne O’Toole, and Miss Juelle Johnson is gratefully recognized. In the development of the method for determination of the radioactivity in plasma glucose labeled with glucose-l-W, using L. mesenteroides, the advice of Dr. Robert Steele and the assistance of Miss Clara Bjerknes of the Brookhaven National Laboratory, Upton, L.I. N.Y., and Miss Barbara Matarese, formerly at New York University, were invaluable. The authors are grateful to Professor Francisco Grande, Department of Physiology, not only for instruction in pancreatectomy, but also for the donation of several pancreatectomized dogs, both for preliminary studies and for three which are included in this paper (dogs G, H, and 1).

4 Culture No. 1229 1 from American Type Culture Collection, Rockville, Md.


Dow

nloaded from


506

This study was supported in part by Public Health Service Research Grants AM-09774, AM- 133 19, and HE-079 14.

J. S. Bishop was formerly a Special Research Fellow of the Public Health Service.

Present address of J. S. Bishop: Dept. of Internal Medicine, Univ. of Minnesota, College of Health Sciences, Minneapolis, Minn.

REFERENCES

1. BISHOP, J. S. Inability of insulin to activate liver glycogen transferase D phosphatase in the diabetic pancreatectomized dog. Biochim. Biophys. Acta 208 : 208-2 18, 1970.

2. BISHOP, J. S., N. D. GOLDBERG, F. GRANDE, AND J. LARNER. Decreased activity of glycogen synthetase D phosphatase in insulin-insensitive diabetic dog liver. Diabetes 18 : 337, 1969.

3. BISHOP, J. S., AND J. LARNER. Rapid activation-inactivation of liver uridine diphosphate, glucose-glycogen transferase and phosphorylase by insulin and glucagon, in vivo. J. Biol. Chem. 242: 1355-1356, 1967.

4. BISHOP, J. S., R. STEELE, N. ALTSZULER, A. DUNN, C. BJERKNES, AN*D R. C. DEEODO. Effects of insulin on liver glycogen synthesis and breakdown in the dog. Am. J. Physiol. 208 : 307-3 16, 1965.

5. BLATT, L. M., J. 0. SCAMAHORN, AND K. H. KIM. Mechanism of puromycin activation of tadpole liver glycogen synthetase. Biochim. Biophys. Acta 177 : 553-559, 1969.

6. BUSCHIAZZO, H., J. H. EXTON, AND C. R. PARK. Effects of glucose on glycogen synthetase, phosphorylase, and glycogen deposition in the perfused rat liver. Proc. Natl Acad. Sci. U.S. 65: 383-387, 1970.

7. DEBODO, R. C., R. STEELE, N. ALTSZULER, A. DUNN, AND J. S. BISHOP. Effects of insulin on hepatic glucose metabolism and glucose utilization by tissues. Diabetes 12 : 16-30, 1963.

8. DEWULF, H., AND H. G. HERS. The role of glucose, glucagon and glucocorticoids in the regulation of liver glycogen synthesis. European J. Biochem. 6 : 558-564, 1968.

9. GIBBS, M., P. K. KINDEL, AND M. BUSSE. Determination of isotopic carbon distribution in hexoses, using Leuconostoc mesenteroides. In: kfethods in Carbohydrate Chemistry, edited by R. L. Whistler and M. L. Wolfrom. New York: Academic, 1963, vol. II, p. 496-509.

10. GLINSMANN, W. H., E. P. HERN, AND A. LYNCH. Intrinsic regulation of glucose output by rat liver. Am. J. Physiol. 216: 698-703, 1968.

11. GOLDBERG, N. D., .J. LARNER, H. SASKO, AND A. G. O’TOOLE. Enzymic analysis of cyclic 3’ ,5’-AMP in mammalian tissues and urine. Anal. Biochem. 28: 523-544, 1969.

12. HAFT, D. E., AND L. L. MILLER. Alloxan diabetes and demon- strated direct action of insulin on metabolism of isolated perfused liver. Am. J. Phxsiol. 192 : 33-42, 1958.

13. HERS, H. G. Etudes enzymatiques sur fragments hgpatiques. Application a la classification des glycogenoses. Rev. Intern. Hepatol. 9: 35-55, 1959.

14. HIZUKURI, S., AND J. LARNER. Studies on UDPG: a-1,4-glucan a-4-glucosyl transferase. VII. Conversion of enzyme from glucose- 6-phosphate-dependent to independent form in liver. Biochem- istry 3 : 1783-l 788, 1964.

15. HORNBROOK, K. R., H. B. BURCH, AND 0. H. LOWRY. The effects of adrenalectomy and hydrocortisone on rat liver me- tabolites and glycogen synthetase activity. Mol. Pharmclcol. 2: 106-l 16, 1966.

16. HUSTON, C. J. W., L. C. ADAMS, R. A. FIELD, AND W. V. Mc- DERMOTT, JR. Effect of insulin on glucose uptake by the isolated nerfused canine liver. Lancet 1 : 357-359. 1962.


Present address of J. Larner: Dept. of Pharmacology, Univ. of Virginia, Charlottesville, Va.

This investigation was presented in part in a short communication (3), and in part in abstract form (2).

Received for publication 9 April 1970.

17. JEFFERSON, L. S., J. H. EXTON, R. W. BUTCHER, E. W. SUTHER- LAND, AND C. R. PARK. Role of adenosine 3’,5’-monophosphate in the effects of insulin and anti-insulin serum on liver metabolism. J. Biol. Chem. 243: 1031-1038, 1968.

18. KREUTNER, W., AND N. D. GOLDBERG. Dependence on insulin of the apparent hydrocortisone activation of hepatic glycogen synthetase. Proc. Natl. Acad. Sci. U.S. 58 : 15 15-15 19, 1967.

19. LARNER, J., C. VILLAR-PALASI, N. D. GOLDBERG, J. S. BISHOP, F. HUIJING, J. I. WENGER, H. SASKO, AND N. B. BROWN. Hor- monal and non-hormonal control of glycogen synthesis-control of transferase phosphatase and transferase I kinase. In: Control of Glycogen Metabolism. Oslo: Universitets Forlaget, 1968, p. l-18.

20. LOWRY, 0. H., J. V. PASSONEAU, F. X. HASSELBERGER, AND D. W. SCHULZ. Effect of ischemia on known substrates and cofactors of the glycolytic pathway in brain. J. Biol. Chem. 239 : 18-30, 1964.

21. LOWRY, 0. H., N. J. ROSEBROUGH, A. L. FARR, AND R. J. RAN- DALL. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193: 265-275, 1951.

22. MADDAIAH, V. T., AND N. B. MADSEN. Studies on the biological control of glycogen metabolism in liver. I. State and activation pattern of glycogen phosphorylase. Biochim. Biophys. Acta 121: 261-268, 1966.

23. MADDAIAH, V. T., AND N. B. MADSEN. Kinetics of purified liver phosphorylase. J. Biol. Chem. 241: 3873-3881, 1966.

24. MORTIMORE, G. E. Effect of insulin on release of glucose and urea by isolated perfused liver. Am. J. Physiol. 204 : 699-704, 1963.

25. MORTIMORE, G. E., E. KING, JR., C. E. MONDON, AND W. H. GLINSMANN. Effects of insulin on net carbohydrate alteration in perfused rat liver. Am. J. Physiol. 2 12 : 179-183, 1967.

26. SOMOGYI, M. Determination of blood sugar. J. Biol. Chem. 160: 69-73, 1945.

27. SOMOGYI, M. A rapid method for the estimation of urine sugar. J. Lab. Clin. Med. 26: 1220-1223, 1941.

28. STEELE, R., W. BERNSTEIN, AND C. BJERKNES. Single phototube liquid scintillation counting of C14: application to an easily isolated derivative of blood glucose. J. Appk Physiol. 10 : 3 19-326, 1957.

29. THOMAS, J. A., K. SCHLENDER, ,QND J. LARNER. A rapid filter paper assay for UDPGlucose-glycogen glucosyltransferase, including an improved biosynthesis of UDP-14C-glucose. Anal. Biochem. 25 : 486- 499, 1968.

30. VILLAR-PALASI, C., N. D. GOLDBERG, J. S. BISHOP, F. Q. NUTTAL, AND J. LARNER. Hormonal control of glycogen synthetase inter- conversions. Fed. European Biochem. Sot. Symp. 19 : 149-160, 1969.

3 1. VILLAR-PALASI, C., M. ROSELL-PEREZ, S. HIZUKURI, F. HUIJING, AND J. LARNER. Muscle and liver UDPGlucose: a-l, 4-glucan a-4-glucosyl transferase (glycogen synthetase). Methods Enzymol. 8: 374-384, 1966.

32. WENGER, J. I., AND C. VILLAR-PALASI. In vivo effect of insulin on muscle glycogen synthetase. Identification of the action pathway. Federation Proc. 26 : 563, 1967.

33. WOSILAIT, W. D., AND E. W. SUTHERLAND. The relationship of epinephrine and glucagon to liver phosphorylase. II. Enzymatic inactivation of liver phosphorylase. J. Biol. Chem. 2 18: 469-48 1,


Dow

nloaded from


Documents

Insulin regulation of hepatic glycogen metabolism in the dog€¦ · Insulin regulation of hepatic glycogen metabolism in the dog. Am. J. Physiol. 220(Z) : 499-506. 1971.-The effect