European Journal of Pharmacology, 41 (1977) 437--441 © Elsevier/North-Holland Biomedical Press, Amsterdam -- Printed in The Netherlands

437

THE CONTRIBUTION OF RENAL AND EXTRARENAL MECHANISMS TO HYPOKALEMIA INDUCED BY GLUCAGON * GEORGE W. PETTIT **, R O B E R T L. VICK and MICHAEL D. KASTELLO ***

Department of Physiology, Baylor College of Medicine, Houston, Texas 77025, U.S.A. Received 17 June 1976, revised MS received 29 October 1976, accepted 5 November 1976

G.W. PETTIT, R.L. VICK and M.D. KASTELLO, The contribution of renal and extrarenal mechanisms to hypokalemia induced by glucagon, European J. Pharmacol. 41 (1977) 437--441. Other investigators have shown that infusion of glucagon causes the concentration of potassium, [K*], in the arterial plasma to increase rapidly, then to decrease to less than the beginning value. In studies on anesthetized dogs, we found that the magnitude of the initial, rapid rise of [K +] was increased by nephrectomy but not affected by pancreatectomy. The subsequent decline of [K ÷] and the persistent hypokalemia were not affected significantly by nephrectomy. Plasma [K +] decreased in the nephrectomized-pancreatectomized dogs, as it did in the nephrectomized and the control groups, but the effect was temporary, and [K ÷] began to increase again, even though the infusion of glucagon continued; after the infusion was ended, plasma [K +] became Significantly higher than the beginning value. These data suggest that the hypokalemia caused by infusion of glucagon initially depends on extrarenal factors other than insulin, and, later, depends on insulin. Glucagon

Hypokalemia

Potassium

Nephrectomy

1. Introduction Wolfson and Ellis (1956) found that a single injection of glucagon, i.v., causes the concentration of potassium, [K ÷], in the arterial plasma to increase rapidly, then to decrease to less than the beginning value and remain depressed. They attributed the prolonged hypokalemia to insulin secreted secondary to the glucagon-induced hyperglycemia; other

* This work was supported by Public Health Service Grant HL 14315 from the National Heart and Lung Institute and by the General Clinical Research Center program of the Division of Research Resources, National Institutes of Health Grant RR-134. ** Present address: U.S. Army Medical Research Institute of Infectious Diseases, Fort Derrick, Frederick, Maryland 21701, U.S.A. *** With the technical assistance of Charley Roberson.

Pancreatectomy

investigators have shown that glucagon also causes the release of insulin by acting directly on the pancreas (Samols et al., 1965). In addition, it has been shown that glucagon exerts a kaliuretic action (Staub et al., 1957; Elrick et al., 1958; Pullman et al., 1967). Thus, renal excretion of K ÷ and insulin-induced tissue uptake of K ÷ both m a y be involved in the hypokalemic action of glucagon. To assess the contributions of each mechanism to hypokalemia, we have infused glucagon continuously in splenectomized {control) dogs, in splenectomized-nephrectomized dogs, and splenectomized-nephrectomized-pancreatectomized dogs. Administration of glucagon by continuous infusion is preferable to single injection in these studies, because it reduces the possibility that the decrease of plasma [K ÷] might be attributable, in part, to the declining effect of glucagon due to dilution and metabolism.

438

2. Materials and methods Healthy dogs not selected for sex or breed and weighing between 10 and 30 kg were anesthetized by morphine sulfate, 20 mg/kg, s.c., and pentobarbital sodium, 5 mg/kg, i.v. The trachea was cannulated to permit control of respiration by positive pressure ventilation, and lead II ECG and arterial pressure were recorded. The spleen was reached through a left paralumbar incision, emptied by compression, and tied off to prevent sequestration or release of red blood cells during the experimental procedure. In animals to be nephrectomized, both kidneys were exposed through bilateral, paralumbar incisions, and each renal artery, vein, and ureter was occluded in a mass ligature. Care was taken not to include the adrenal glands in the tie. In animals to be pancreatectomized, the pancreas was reached through a midline abdominal incision. The caudal pancreaticoduodenal artery and vein and the pancreatic branches of the splenic artery and vein were ligated and cut. The cranial pancreaticoduodenal artery and vein were left intact, and every fragment of the pancreas was removed carefully by the avulsion method of Markowitz et al. (1964). This procedure allowed removal of the pancreas without compromising blood flow in the duodenum. Other investigators have found the half-life of insulin in the dog to be of the other of 10 min (Arnould et al., 1967; Hommel et al., 1971). To assure that plasma insulin concentration, [Insulin], had fallen to negligible levels, we waited at least 40 min after the pancreatectomy was completed before beginning our experiments. After an additional 30-min control period, glucagon (crystalline porcine giucagon, certified insulin-free), 4 /~g/kg/min, or isotonic sodium chloride solution was infused at 0.5 ml/min, using a syringe-driver pump, through a non-occluding catheter placed in the right femoral vein. Each infusion lasted 30 min and was followefl by a 30-min recovery period.

G.W. P E T T I T ET AL.

Samples of blood were taken from the left femoral artery at specified intervals throughout each experiment. A portion of each sample was collected in a heparinized tube for measurement of plasma [K÷], and, during some experiments, the remainder was collected in an EDTA tube for measurement of plasma [Insulin]. All samples were centrifuged immediately after collection. Assuming a blood volume equal to 7.9% of b o d y weight in the dog (Courtice, 1943), no more than 6% of any animal's blood was removed during =ny experiment. Plasma [K ÷] (mean + S.E.M.) was determined using an internal standard flame photometer. Plasma [Insulin] (mean + S.E.M.) was measured by radioimmunoassay using the double antibody system of Morgan and Lazarow (1963). The technique was modified by using porcine insulin as standard, 113~labelled porcine insulin as tracer, and dextrancoated charcoal instead of the second antib o d y (Herbert et al., 1965). Although measurements of [Insulin] obtained by this assay are porcine insulin equivalents and do not represent absolute concentrations of dog insulin, it has been shown that this assay gives a valid representation of changes in canine pancreatic H-cell activity (Hiatt et al., 1972). The Student's t-test was used for statistical analysis of data. Data are presented as means +S.E.M.

3. Results In 7 dogs, mean plasma [Insulin] after pancreatectomy was 2.4 + 0.6 pU/ml; after infusion of glucagon at a rate of 4 #g/kg/min for 10 min, the value-was 5.6 + 2.9 pU/ml. The increase was not statistically significant and we assumed that pancreatectomy had removed all important sources of insulin secretion. Fig. 1 illustrates the results of control studies in which isotonic sodium chloride solution, rather than glucagon, was infused into 7 splenectomized, nephrectomized (Sx,Nx) dogs and 5 splenectomized, nephrectomized,

GLUCAGON AND HYPOKALEMIA • Sx. Nx 0 Sx, Nx, Px

~) +0.5 .÷-o-

w

E

439

..... 0 ...........

o

~:

J.

...... 6

i 50

i 60

....?--' ....... 7

4-

~E"

",- BEFORE

"w - 0 . 5

i -30

"='

t"

Px

I s o ! o n i c

I -20

J -I0

i 0

N a CI

l I0

l 20

i 30

i 40

Fig. 1. Changes in arterial plasma [K +] (mean ± S.E.M.) against time (rain, abscissa) in 7 splenectomized-nephrectomized (Sx,Nx) dogs and in 5 splenectomized-nephrectomized-pancreatectomized (Sx,Nx,Px) dogs during control experiments in which isotonic NaCl solution was infused intravenously. Total elapsed time between sample taken before pancreatectomy and min --30 is time for operation plus 40-rain recovery period.

pancreatectomized (Sx,Nx,Px) dogs. Mean arterial plasma [K ÷] at min 0 in the Sx,Nx group was 3.71 + 0.05 mEq/1. In the Sx,Nx, Px group, mean arterial plasma [K ÷] before Px was 3.65 + 0.18 mEq/1; this value increased to 3.99 + 0.16 mEq/1 by min --30, after which no further change occurred. The increase of plasma [K ÷] associated with pancreatectomy is related to the morphine used

h-

• s,

+1.5

l

e Sx, Nx 0 Sx, Nx, Px

(D

-4"o_

in anesthesia and is discussed extensively in other work (Pettit and Vick, 1974). Variation within each group was reduced by calculating all values as differences from that at min 0. Fig. 2 illustrates the results of infusing glucagon, 4 #g/kg/min for 30 min, into 3 groups of 9 dogs each. In all 3 groups, glucagon caused temporary hyperkalemia followed by hypokalemia. In the Sx (control) group,

r~ i I .iIi,J.

cr +1.0 w

E +

x,- +0.5 O9 / 13.. "0

.... i ii, '

0

, ........ s........+.........+...........

T ......~ "

-0.5

T

T--~

G L U C A G 0 N I

I

I

I

I

I

-50

-20

-I0

0

I0

20

I

50

I

40

I

I

50

60

Fig. 2. Changes in arterial plasma [K ÷] (mean -+ S.E.M.) during intravenous infusion of glucagon (4 ~g/kg/min) in 9 splenectomized (Sx)dogs (joined by solid lines), 9 splenectomized-nephrectomized dogs (joined by dashed lines) and in 9 splenectomized-nephrectomized-pancreatectomized dogs (joined by d o t t e d lines).

440

mean arterial plasma [K ÷] was 3.68 + 0.04 mEq/1 at min 0; this value was elevated significantly at min 2, 4 (p < 0.01) and 6 (p < 0.05), was depressed significantly at min 15 and 20 (p < 0.02), and continued at about the same level for the remainder of the experiment. In the Sx,Nx group, mean arterial plasma [K ÷] was 3.64 + 0.16 mEq/1 at min 0; glucagon increased this value significantly at min 2 (p < 0.01), 4 and 6 (p < 0.02) and decreased it significantly at min 15 and 20 (p < 0.05). Mean arterial plasma [K*] in this group also remained decreased after the infusion had ended; the difference was significant at min 40 (p < 0.02). In the Sx,Nx,Px group, mean arterial plasma [K ÷] increased from 3.65 + 0.08 mEq/1 before Px to 3.93 +- 0.15 mEq/1 at min --30 and continued to increase during the 30-min control period (to 4.13 + 0.15 mEq/l at rain 0), although the value at min --10 was not significantly different from that at min 0. Infusion of glucagon increased mean arterial plasma [K ÷] significantly at min 2 (p < 0.02) and 4 (p < 0.01) and decreased it significantly at min 15 and 20 (p < 0.01). Hypokalemia proved to be temporary in this group; plasma [K ÷] began to rise again, even before the infusion was finished, and was significantly increased (p < 0.01) at min 60. 4. Discussion

The initial hyperkalemic effect of glucagon in each of the 3 groups of animals represented in fig. 2 is similar. However, in comparison with the other 2 groups, the largest increase of [K ÷] is less, but not significantly so, in the group that underwent only splenectomy. This possible difference may be attributable to renal function in these non-nephrectomized animals; such a conclusion is consistent with that drawn from other work (Pettit et al., 1975) which showed that renal excretion of K ÷ increases while plasma [K ÷] is increasing but diminishes after plasma [K ÷] has begun to decline. The hypokalemic effect of glucagon in each of the 3 groups is indistinguishable at 10 and

G.W. PETTIT ET AL.

15 min after beginning infusion. Thus , neither renal nor pancreatic factors needed to be involved at this time. The hypokalemia in the Sx,Nx and the Sx,Nx,Px, groups may be related in part to the direct stimulation of Na ÷, K÷-ATPase by glucagon (Dambach and Friedman, 1974); however, epinephrine and growth hormone both cause uptake of K ÷ by tissues (Vick et al., 1972; Zierler and Rabinowitz, 1963), and glucagon also causes release of both of these hormones (Sarcione et al., 1963; Mitchell et al., 1969; Cain et al., 1970). Although plasma [K *] is decreased in the group with kidneys and pancreas removed, the hypokalemia is short-lived and followed by hyperkalemia. In contrast with these results, plasma [K ÷] remains below the beginning level in the Sx and the Sx,Nx groups for the remainder of the experiment. Thus, it appears that persistence of glucagon-induced hypokalemia may depend on insulin, although the initial development does not. Since the hypokalemic effect in nephrectomized animals is not significantly different from that seen in non-nephrectomized animals, one may assume that the hypokalemia depends mainly on extrarenal mechanisms. Thus, although glucagon increases net renal excretion of potassium (Staub et al., 1957; Elrick et al., 1958; Pullman et al., 1967), in our experiments, this kaliuretic action is not solely responsible for the hypokalemic effeCt of glucagon. A decrease of plasma [K ÷] increases the intracellular-to-extracellular chemical activity gradient acting on K ÷, causing egress of K ÷ from cells to plasma. Thus, if movement of K ÷ out of cells is not prevented, the kaliuretic action of glucagon would be expected to have little effect on plasma [K÷]:. This observation explains further w h y renal excretion of K ÷ attenuates glucagon-induced hyperkalemia but does n o t contribute significantly to glucagon-induced hypokalemia. Acknowledgements The authors gratefully acknowledge the valuable contributions of Ms. Anna M. Swander, who perform-

GLUCAGON AND HYPOKALEMIA ed the insulin assays, Mrs. Rhonda Weikert and Mrs. Regina Staley who provided secretarial support, and Mrs. Phebe W. Summers for editorial assistance. We appreciate complimentary samples of glucagon provided by Dr. J. Hosley of Lilly Research Laboratories, Indianapolis, Indiana, U.S.A.

References Arnould, Y., F. Cantraine, H.A. Ooms, C. Delcroix and J.R.M. Franckson, 1967, Kinetics of plasma disappearance of labelled iodoinsulins following intravenous injection, Arch. Intern. Pharmacodyn. Therap. 166, 225. Cain, J.P., G.H. Williams and R.G. Dluhy, 1970, Glucagon stimulation of human growth hormone, J. Clin. Endocrinol. Metab. 31,222. Courtice, R.C., 1943, The blood volume of normal animals, J. Physiol. (London) 102, 290. Dambach, G. and N. Friedmann, 1974, The effects of varying ionic composition of the perfusate on liver membrane potential, gluconeogenesis and cyclic AMP responses, Biochim. Biophys. Acta 332, 374. Elrick, H., E.R. Huffman, C.J. Hlad, Jr., N. Whipple and A. Staub, 1958, Effects of glucagon on renal function in man, J. Clin. Endocrinol. Metab. 18, 813. Herbert, V., K.-S. Lau, C.W. Gottlieb and S.J. Bleicher, 1965, Coated charcoal immunoassay of insulin, J. Clin. Endocrinol. Metab. 25, 1375. Hiatt, N., M.B. Davidson and G. Bonorris, 1972, The effect of potassium chloride infusion on insulin secretion in vivo, Horm. Metab. Res. 4, 64. Hommel, H., U. Fischer and H. Kansy, 1971, Insulin in the pancreatic blood and insulin half-life in an Alsatian bitch with spontaneous diabetes, Horm. Metab. Res. 3,213.

441 Markowitz, J., J. Archibald and H.G. Downie, 1964, Experimental surgery, 5th ed. (Williams & Wilkins, Baltimore). Mitchell, M.L., M.J. Bryne and J. Silver, 1969, Growth-hormone release by glucagon, Lancet 1, 289. Morgan, C.R. and A. Lazarow, 1963, Immunoassay of insulin: two antibody system, Diabetes 12, 115. Pettit, G.W. and R.L. Vick, 1974, An analysis of the contribution of the endocrine pancreas to the kalemotropic actions of catecholamines, J. Pharmacol. Exptl. Therap. 190,234. Pettit, G.W., R.L. Vick and A.M. Swander, 1975, Plasma K ÷ and insulin: changes during KC1 infusion in normal nephrectomized dogs, Amer. J. Physiol. 228, 107. Pullman, T.N., A.R. Lavender and A. Impi, 1967, Direct effects of glucagon on renal hemodynamics and excretion of inorganic ions, Metabolism 16, 358. Samols, E., G. Marri and V. Marks, 1965, Promotion of insulin secretion by glucagon, Lancet 2, 415. Sarcione, E.J., N. Back, J.E. Sokal, B. Mehlman and E. Knoblock, 1963, Elevation of plasma epinephrine levels produced by glucagon in vivo, Endocrinology 72, 523. Staub, A., V. Springs, F. Stoll and H. Elrick, 1957, A renal action of glucagon, Proc. Soc. Exptl. Biol. Med. 94, 57. Vick, R.L., E.P. Todd and D.W. Luedke, 1972, Epinephrine-induced hypokalemia: relation to liver and skeletal muscle, J. Pharmacol. Exptl. Therap. 181, 139. Wolfson, S.K., Jr. and S. Ellis, 1956, Effects of glucagon on plasma potassium, Proc. Soc. Exptl. Biol. Med. 9 1 , 2 2 6 . Zierler, K.L. and D. Rabinowitz, 1963, Roles of insulin and growth hormone based on studies of forearm metabolism in man, Medicine (Baltimore) 42, 385.