Gluconeogenesis from Alanine in Normal Postabsorptive Man Intrahepatic Stimulatory Effect of Glucagon /. L. Chiasson, M.D., ]. E. Liljenquist, M.D., B. C. Sinclair-Smith, M.D., and W. W. Lacy, M.D., Nashville
SUMMARY Although the stimulatory effect of glucagon on gluconeogenesis has been well demonstrated in certain systems in vitro, this effect has never been established in man. The present study was undertaken, therefore, to determine whether glucagon could stimulate gluconeogenesis from alanine in normal fasting man. Glucagon might stimulate this process by increasing the hepatic alanine uptake and/or by shunting the extracted alanine within the liver into the gluconeogenic pathway. In order to be able to examine these two aspects of gluconeogenesis, we combined the hepatic veinbrachial artery catheterization technic with an isotopic infusion of alanine-14C. Alanine-14C specific activity was measured in whole blood and plasma by use of a rapid chromatographic technic. Since plasma contributed 93 per cent of the alanine extracted by the splanchnic bed with a specific activity three times that of the red
blood cells, plasma alanine specific activity was used to study the conversion of alanine to glucose. A constant infusion of alanine-14C achieved a relatively stable arterial specific activity by forty minutes. The administration of glucagon by constant infusion (15-50 ng./kg./min.) had no effect on the splanchnic extraction of alanine. Net splanchnic glucose-1 *C production, however, doubled during the glucagon infusion, and the conversion of alanine to glucose increased from 30 ± 2 to 58 ± 9 fimol/nun. These data (1) demonstrate that in normal man fasted twelve to fourteen hours, glucagon at supraphysiblogic levels can double the rate of gluconeogenesis from alanine and (2) indicate that this stimulatory effect of glucagon is exerted within the liver by shunting the extracted alanine toward new glucose formation rather than by increasing the hepatic extraction of alanine. DIABETES 24:574-84, June, 1975.
The importance of amino acids as a source of fuel, particularly for glucose production during periods of fasting, is generally recognized.1 In recent years several studies employing organ catheterization technic and column chromatography for plasma amino acid analysis have contributed significantly to our understanding of the role of amino acids in gluco^ neogenesis.2"5 These studies have clearly documented the flow of amino acids from muscle to liver under various metabolic conditions. Of all the amino acids, alanine has been assigned a primary role
as a gluconeogenic precursor based on the observations that its hepatic uptake represents nearly 50 per cent of the total amino acids extracted by the liver in man and that its hepatic extraction increases markedly after a thirty-six to forty-eight-hour fast.2 While these studies in man have helped define the interorgan flux of amino acids from muscle to liver, they have provided little information on the intrahepatic fate of the amino acids once extracted. Without this information, meaningful conclusions concerning the process of gluconeogenesis itself are difficult to derive. Using tracer technics, both in vitro6"8 and in vivo49"1 x studies have demonstrated that alanine could be converted to glucose by the liver. While the tracer technic can provide information concerning the fate of an amino acid extracted by the liver, the use of a tracer alone gives little information concerning the total uptake of the amino acid by the liver. Since gluconeogenesis from alanine could be controlled
This work was presented in part at the Canadian Society for Clinical Investigation held in Montreal, Canada, on January 23, 1974. From the Department of Medicine, Vanderbilt University School of Medicine, Nashville, Tennessee 37232. Address reprint requests to: J. L. Chiasson, Department of Medicine, Vanderbilt University School of Medicine, Nashville, Tennessee 37232. Accepted for publication March 12, 1975.
574
DIABETES, VOL. 2 4 , NO. 6
J. L. CHIASSON, M.D., AND ASSOCIATES
by altering the hepatic extraction of alanine and/ or by affecting the intrahepatic disposition of the extracted alanine, a measurement of both is essential in a study on the regulation of gluconeogenesis. For this purpose we combined an isotopic infusion of alanine- 14 C (to trace the intrahepatic conversion of alanine- 14 C to glucose-14C) with our hepatic veinbrachial artery catheterization technic (to measure the splanchnic extraction of alanine and alanine- 14 C and the release of glucose-14C). Accumulated evidence indicates that glucagon may be a key hormone in regulating the flow of amino acids to glucose. It is well established that glucagon can stimulate gluconeogenesis in certain systems in vitro, especially in the isolated perfused rat liver. 6p7il2i15 We have recently shown that glucagon can stimulate the conversion of alanine- 1 4 C to glucose- 14 C in the intact dog.51 Direct evidence, however, that this hormone plays a major role in regulating gluconeogenesis in man is still lacking. Circumstantial evidence supporting this view derives from the observations that glucagon levels are increased during fasting, 1 6 s t a r v a t i o n , 1 7 1 8 and hypoglycemia, 1 9 all conditions wherein gluconeogenesis is thought to be markedly stimulated. Further indirect evidence comes from a case report of Levy et al. 2 0 describing a patient with isolated glucagon deficiency in whom fasting hypoglycemia was a prominent symptom. The present study was designed to determine whether glucagon can stimulate gluconeogenesis from alanine in normal man and to examine its action on both hepatic uptake and intrahepatic disposition of alanine. The ideal setting to demonstrate any stimulatory effect of glucagon on gluconeogenesis is a state wherein glucagon is relatively low and gluconeogenesis is minimal. For this purpose, we choose to study the effect of glucagon on gluconeogenesis in normal man after an overnight fast (twelve to fourteen hours), a state fulfilling these two criteria. Since the catheterization procedure entailed a small but definite risk and since glucagon has never been shown to stimulate gluconeogenesis in man, we have used supraphysiologic doses of glucagon, hoping to get a maximum stimulative effect on this process. MATERIALS AND PREPARATIONS Sterile, pyrogen-free L-alanine-U- 1 4 C (256 mCi/mmol) (New England Nuclear, Boston, Mass.) was used as a tracer. The glucagon used in this study was purchased from Eli Lilly Laboratories (InJUNE,
1975
dianapolis, Ind.). Hepatic blood and plasma flow were determined by using indocyanine green (Hynson, Westcott, and Dunning, Inc., Baltimore, Md.). Phadebas Insulin Radioimmunoassay Kit was purchased from Pharmacia Fine Chemicals, Inc. (Piscataway, N J . ) and Trasylol from FBS Pharmaceuticals, Inc. (New York, N.Y.). All enzymes and cofactors used in the determinations of lactate, pyruvate, and glycerol were obtained from Boehringer Mannheim Corp. (New York, N.Y.). For infusion, glucagon was mixed with saline in a sterile 100-ml. volumetric flask containing 2-3 ml. of the subject's blood to eliminate binding of glucagon to glassware. Alanine-U- 14 C, 100 ;uCi, was mixed with saline in a sterile 50-ml. volumetric flask to which 1 ml. of the subject's plasma had been added. Three milliliters of plasma was also added to freshly prepared indocyanine green in order to increase its stability. 21
SUBJECTS Eight normal male volunteers participated in this study (table 1). All subjects were screened prior to the study and were accepted only when found to fulfill the following criteria: None of the subjects had any personal history of diabetes mellitus or any other endocrine or major disease. Each subject had a normal standard three-hour oral glucose tolerance test (40 gm. of glucose per square meter of body surface).22 Hepatic function assessed by measurement of bromsulfthalein extraction, serum glutamic-oxaloacetic transaminase, serum bilirubin, and alkaline phosphatase was normal. Each subject had a normal serum urea nitrogen and urinalysis, and cardiovascular function was normal by history, physical examination, chest x-ray, and electrocardiogram. Complete blood counts including a sedimentation rate were normal in all subjects. The subjects were all placed on a 300-gm. carbohydrate diet to which they adhered for three days prior to the study. All subjects were fasted twelve to fourteen hours prior to the study. The protocol for the present study was reviewed and approved by the Vanderbilt University Clinical Investigation Committee. This committee has the responsibility for protecting the rights of human volunteers participating in research studies. The nature, purpose, and possible risks of the procedures were fully explained to each subject before obtaining his voluntary consent. 575
EFFECT OF GLUCAGON ON GLUCONEOGENESIS FROM ALANINE IN MAN
TABLE 1 Clinical data on subjects
Age Glucagon group (n=5) Mean S.E.M. Control group (n= 3) Mean S.E.M.
Height
Weight
S.A.*
cm.
kg-
m.2
EHPFf ml./min.
FBSt mg./lOOml.
27.8
179.6
2.9
1.9
82 4.4
1.99 0.05
960
95.2
24.3
183.2
74.2
2.8
4.9
5.3
1.98 0.07
1,156 76.9
62.4
2.5 89.6 0.88
*Body surface area in square meters. fMean estimated hepatic plasma flow as milliliters per minute. ^Fasting blood glucose at the beginning of the study.
PROCEDURES
In the Vanderbilt University Hospital Cardiac Catheterization Laboratory, a Teflon catheter was inserted percutaneously in the left brachial artery. A 1-cm. cutdown was then performed over a small tributary vein distal to the right antecubital fossa. A no. 8F Cournand catheter was guided through this vein to a right-sided hepatic vein and under fluoroscopy was positioned 3-4 cm. from the wedge position. A continuous saline infusion maintained the patency of the hepatic vein catheter without added anticoagulant. In all subjects a trace amount of L-alanine-U-14C was administered in a peripheral vein as a constant infusion at a rate of 0.86 /uCi/min. throughout the 100-minute study period. After a sixty-minute basal period, a glucagon (15 ng./kg./min., n = \; 25 ng./kg./min., n — \\ 50 ng./kg./min., »=3) or saline (» = 3) infusion was started and continued over the last forty minutes. In all subjects, indocyanine green was administered in a peripheral vein as a constant infusion throughout the study for determination of hepatic blood flow. Simultaneous blood samples were obtained from the brachial artery and hepatic vein throughout the study. Approximately 500 ml. of blood was withdrawn during the study, the volume of which was replaced with saline infused throughout the study at a rate of 7 ml. per minute. ANALYTIC METHODS The system for determination of alanine concentration and specific activity consisted of two identical chromatographic columns, 0.9 cm. internal diameter, packed with Beckman Spherical Resin, Type AA15 (Beckman Instrument Co., Palo Alto, Calif.). Separate buffer pumps, with identical flow rates, were used on each column and were fed from common buf576
fer and NaOH reservoirs through three-way stopcocks. The effluent from the two columns led to a four-way valve arranged so that when one column was directed to a fraction collector through a "T" fitting, the other column was directed to waste. The effluent stream directed to the fraction collector was continuously sampled from the "T" fitting by a Technicon AutoAnalyzer, which reacted a small portion of the column effluent with ninhydrin for quantification of the alanine peak. The remainder of the effluent stream went directly to the fraction collector, where an appropriately timed fraction collected only the alanine peak. The radioactivity in this fraction was determined by counting a 2-ml. aliquot in 10 ml. of Aquasol (New England Nuclear, Boston, Mass.) for twenty minutes in a Packard Tri-Carb Scintillation Spectrometer (Packard Instrument Company, Dowers Grove, 111.). The following chromatographic conditions were used: column height 18.5 cm., temperature 65°C, buffer flow rate 2.2 ml./min., buffer pH 2.75 to 2.98, and Na concentration 0.2N. Under these conditions, glycine and alanine are well separated, with an elution time for alanine of about 23.5 minutes. Glucose- 14 C, lactate- 14 C, and pyruvate- 14 C eluted in the first five minutes. In this system, glutathione elutes with a group of unseparated amino acids well ahead of glycine and alanine. Plasma or whole-blood samples were prepared for analysis by the addition of an equal volume of 6 per cent sulfosalicylic acid. After chilling for forty-five minutes at 4° C., the precipitate was separated at 12,000 X g in a refrigerated centrifuge. One milliliter of the clear filtrate was applied to the chromatographic column. When plasma was compared with whole blood, the samples were processed in parallel, with whole blood added to the 6 per cent sulfosalicylic acid immediately after it was drawn. The plasma was DIABETES, VOL. 2 4 , NO. 6
J. L. CHIASSON, M.I
separated by centrifugation and added to sulfosalicylic acid within five minutes. A filtrate of plasma and whole blood from a single sample of venous blood to which was added a known amount of alanine-14C was analyzed along with unknown samples as a quality control. Replicate analyses over a period of six weeks gave the following statistics on the reproducibility of the method: the mean plasma alanine concentration was 455 /nmol/L., with a S.D. of 15 fimollL., and the mean per cent recovery of added 14C counts was 96.9 per cent, with a S.D. of 1.9 per cent (n=30); the mean whole-bk>od alanine concentration was 423 /umol/L., with a S.D. of 13 /amol/L., and the mean per cent recovery of added alanine-14C was 97 per cent, with a S.D. of 2.2 per cent (n=15). Glucose-14C was determined as described in a previous paper.i] Since glucose equilibrates almost instantly with the intracellular water of erythrocytes,23'24 the net splanchnic glucose-14C production rate was calculated by multiplying the arterial-hepatic venous differences by the plasma flow plus 70 per cent (intracellular water distribution) of the red blood cell flow. Whole-blood glucose was measured within four minutes by the Hoffman ferricyanide reaction using the Technicon AutoAnalyzer.25 Immunoreactive glucagon was assayed by Dr. Roger Unger, Dallas, Texas, using the 30K antiserum.26 Immunoreactive insulin was assayed by the Sephadex bound-antibody procedure. 27 The plasma concentrations of indocyanine green were determined in a Beckman spectrophotometer at 815 m/x and calculated according to the method of Leevy et al. 28 Whole blood was deproteinized with 30 per cent perchloric acid immediately after sampling and kept on ice for determination of lactate, pyruvate, and glycerol by a modified enzymatic fluorometric method.29"31 The statistical significance was determined by the Student's test or paired t test whenever applicable.32
AND ASSOCIATES
RESULTS
Glucagon, Blood Glucose, Insulin, and Estimated Hepatic Plasma and Blood Flow
The basal arterial plasma glucagon level was stable at a mean of 134 ± 2 . 5 pg./ml. and was unaffected by saline infusion. Administration of glucagon at 50 ng./kg./min. (n = 3) resulted in a rapid increase from basal levels to 5,066 ± 536 pg./ml. Glucagon infused at 15 ng./kg./min. (n = l)and25 ng./kg./min. (n = l) produced arterial plasma concentrations of 1,700 and 2,450 pg./ml., respectively, within ten to fifteen minutes. Since the response of the parameters measured was maximum with the 15 ng./kg./min. glucagon infusion, the five glucagon-treated subjects were considered as a single group for analytic purposes. The mean net splanchnic glucose production in the controls was stable over the 100-minute study period at 159-3 ± 5 . 7 mg./min. (table 2), maintaining an arterial blood glucose of 90.3 ± 0.5 mg./lOO ml. In the glucagon-treated group the net splanchnic glucose production increased from a mean basal level of 142.9 ± 7 . 3 mg./min. to a peak of 504.2 ± 80.6 mg./min. ten minutes after glucagon infusion was started. This peak was then followed by a gradual decline to 253-5 ± 10.8 mg./min. by 100 minutes despite continuing glucagon infusion. The mean arterial blood glucose level increased during glucagon infusion from a basal level of 94.5 ± 0.5 mg./lOO ml. to a maximum of 173.2 ± 11.7 mg./lOO ml. by the end of the study. The mean arterial plasma immunoreactive insulin concentration of 5.7 ± 0.4 ju,U./ml. in the controls did not change during saline infusion. In the glucagon-treated subjects, insulin gradually rose from the mean basal levels of 6.9 ± 0 . 1 /u.U./ml. at sixty minutes to 24.4 ± 4 . 2 /u,U./ml. by" 100 minutes. The estimated mean hepatic blood and plasma flow measured by the indocyanine green technic were 1,781 ± 104 ml./min. and 1,034 ± 58 ml./min.,
TABLE 2 Effect of glucagon and saline on net splanchnic glucose production
Glucagon group Mean S.E.M. Control group Mean S.E.M.
Net splanchnic glucose production (mg./min.) Basal Period 30' 40' 50' 60'* 70'
Glucagon or Saline 80' 90'
100'
0'
10'
20'
109 27
130 11
132 15
160 2
152 23
155 31
140 12
504 81
446 57
348 42
253 11
161 34
163 29
160 10
148 6
153 20
177 42
131 16
142 19
160 17
149 65
137 28
*Glucagon or saline infusion was begun after the sixty-minute sample. JUNE, 1975
577
EFFECT OF GLUCAGON ON GLUCONEOGENESIS FROM ALANINE IN MAN
respectively. The flow was very stable throughout the study and was not affected by the glucagon infusion. Effect of Glucagon on the Dynamics of Alanine-14C Across the Splanchnic Bed
Flux
To determine whether whole blood, plasma, or both whole blood and plasma concentrations and specific activities would be needed to estimate the splanchnic extraction of alanine and the conversion of circulating alanine into glucose, we measured the alanine concentration and radioactivity in both whole blood and plasma before and during glucagon administration in four subjects. Eight comparisons were made before and eight during the glucagon infusion. Since glucagon had no effect on the arterial-hepatic venous differences, the means of all sixteen comparisons are shown in table 3. It is apparent that under the conditions of our study, most of the net splanchnic extraction of both total alanine and 14C-alanine was from the plasma compartment, with red cells contributing less than 7 per cent. Table 4 shows the mean alanine-14C specific activity in plasma, whole blood, and red blood cells. Of particular note is the fact that the specific activity was different in the plasma and cell compartments. Since most of the alanine is extracted from plasma at a higher specific activity than in whole blood, we chose to measure alanine in plasma rather than in whole blood for the remainder of the studies reported below. The mean alanine specific activities observed in arterial and hepatic venous plasma during a constant infusion of alanine-14C in control and glucagontreated subjects are shown in figures 1A and B. Although a completely flat plateau is not reached, the change in specific activity between forty and 100 minutes was less than 6 per cent. The ratio of the
TABLE 4 Arterial and hepatic venous alanine-14C specific activity in whole blood, plasma, and blood cells*
Mean S.E.M.
Whole blood Artery Hepatic vein cpm//i,mol 3365 2195 174 176
Plasma Artery Hepatic vein cpm/ftmol 4967 3422 226 268
Blood cells Artery Hepatic vein cpm//i,mol 1685 1491 152 148
*n= 16 (four observations in four subjects).
hepatic venous specific activity to the arterial specific activity (figure 1) is a measure of the dilution of plasma alanine by unlabeled alanine from the splanchnic bed. The figure shows that the ratio rose during the initial period of the constant infusion of alanine-14C, but reached a plateau by forty minutes and remained constant thereafter, with a mean value of 0.65, indicating a 35 per cent dilution of the specific activity. Glucagon had no demonstrable effect on this ratio. Effect of Glucagon on Uptake of Gluconeogenic Precursors
Figure 2 shows the mean arterial plasma alanine concentration and mean net splanchnic alanine extracALANINE- I4 C
6
ALANINE SPEC. ACT. 4 CPM X I0"3 PER 2
!'
INFUSION
0.86/JCL/MIN
ortery(n=4)
>JMOLE
0 HV:A
l 0
RATIO
0.5
1-1
I
I
0
TABLE 3 Contribution of plasma and red blood cell to alanine arterial-hepatic venous differences* Plasma Whole blood Cell contentf cone. contents^ Alanine in yuxnol/L. of whole blood A§ 128 ± 5 249 ± 10 121 ± 8 A-HV 64 ± 3 69 ± 3 5±2 Alanine-14C in cpm x 1(T3/L. of whole blood 194 ± 14 622 ± 19 817 ± 23 33 ± 9 417 ± 11 450 ± 12 A-HV •All values subjects). •(•Calculated ^Calculated §A=artery;
578
Plasma contribution % 92.5 ± 3 % 93.1 ± 2
are mean ± S.E.M., n= 16 (four observations in four as plasma concentration x (1-Hct). as whole blood minus plasma content. A-HV= arterial-hepatic venous difference.
ALANINE SPEC. ACT. 4 CPM X I 0 3 PER 2 juMOLE 0 HV:A RATIO
'° o.5
20
40 60 MINUTES
80
100
FIG. 1. The effects of glucagon (A) and saline (B) on the arterial and hepatic venous alanine specific activity and the hepatic venous:arterial (HV:A) alanine specific activity ratio. Values are the means ± S.E.M. DIABETES, VOL. 2 4 , NO. 6
J. L. CHIASSON, M . D . , A N D ASSOCIATES
ALANINE-WC
ARTERIAL ALANINE /iMOLES/L
200
6LUCAG0N
®
300
I
\
I
I
100
NET
from the product of hepatic plasma flow times the arterial-hepatic venous differences. The mean net splanchnic extraction of lactate was 119 — 53 /xmoles/min. before and 118 ± 58 jumoles/min. after glucagon infusion. Net splanchnic extraction of glycerol was 46 ± 9 /imoles/min. before and 49 ± 9 /umoles/min. after glucagon infusion. The net splanchnic balance of pyruvate was very small and was not affected by glucagon. Thus, glucagon had no significant effect on the net splanchnic balance of the four major gluconeogenic precursors—alanine, lactate, pyruvate, and glycerol.
INFUSION 0.86juCi/MIN
I
-*—g-
(n = 4)
°
SPLANCHNIC ? n n ALANINE ^ U U UPTAKE ,00
s—*-
>JMOLES/MIN
0
SALINE
Aguilar-Parada, E., Eisentraut, A. M., and Unger, R. H . : Pancreatic glucagon secretion in normal and diabetic subjects. Am. J. Med. Sci. 257:415, 1969. 2 'Wide, L., and Porath, J.: Radioimmunoassay of proteins with the use of Sephadex-coupled antibodies. Biochem. Biophys. Acta 730:257, 1966. 28 Leevy, C. M., Mendenhall, C. L., Lesko, W . , and Howard,
584
M. M.: Estimation of hepatic blood flow with indocyanine green. J. Clin. Invest. 47:1169, 1962. 2!l Hohorst, H. J.: L-lactate determination with lactate dehydrogenase and DPN. In Methods of Enzymatic Analysis, Bergmeyer, H. V., Ed. New York, Academic Press. 1963, p. 266. 30 Butcher, T., Rudolf, C , Lamprecht, W . , and Latzko, E.: Pyruvate. In Methods of Enzymatic Analysis. Bergmeyer, H. V., Ed. New York, Academic Press. 1963, p. 253. 31 Chernick, S. S.: Determination of glycerol in acyl glycerols. In Methods of Enzymology, vol. 14. Lowenstein, John M., Ed. 5th ed. New York, Academic Press, 1969, p. 627. 32 Snedecor, G. W.: Statistical Methods, 5th ed. Ames, Iowa, State University Press, 1965, p. 35. 33 Elwyn, D. H.: Distribution of ami no acids between plasma and red blood cells in the dog. Fed. Proc. 25:854, 1966. 34 Elwyn, D. H . , Launder, W. J . , Pariksh, H. C , and Wise, E. M., Jr.: Role of plasma and erythrocytes in interorgan transport of ami no acids in dogs. Am. J. Physiol. 222:1333, 1972. li5 Felig, P., Wahren, J . , and Raf, L.: Evidence of interorgan amino acid transport by blood cells in humans. Proc. Nat. Acad. Sci. U.S.A. 70:1775, 1973. 3f> Aoki, T. T., Muller, W. A., Brennan, M. F., and Cahill, G. F., Jr.: Blood cell and plasma amino acid levels across forearm muscle during a protein meal. Diabetes 22:768, 1973. 37 Winter, C. G., and Christensen, H. N . : Migration of amino acids across membrane of the human erythrocyte. J. Biol. Chem. 239:872, 1964. 38 Haft, D. E., Tennen, E., and Mehtalia, S.: Alanine metabolism in perfused livers of normal and adrenalectomized rats. Am. J. Physiol. 222:365, 1972. 3! 'Krebs, H. A., Hemo, R., Weidemann, M. J., and Speake, R. N . : The fate of isotopic carbon in kidney cortex synthesizing glucose from lactate. Biochem. J. 101:242, 1966. 40 Rocha, D. M., Santeusanio, F., Faloona, G. R., Unger, R. H.: Abnormal pancreatic alpha-cell function in bacterial infections. N. Engl. J. Med. 288:700, 1973. 41 Unger, R. H.: Glucagon and the insulin:glucagon ratio in diabetes and other catabolic illnesses. Diabetes 20:834, 1971.
DIABETES, VOL. 2 4 , NO. 6
SUMMARY Although the stimulatory effect of glucagon on gluconeogenesis has been well demonstrated in certain systems in vitro, this effect has never been established in man. The present study was undertaken, therefore, to determine whether glucagon could stimulate gluconeogenesis from alanine in normal fasting man. Glucagon might stimulate this process by increasing the hepatic alanine uptake and/or by shunting the extracted alanine within the liver into the gluconeogenic pathway. In order to be able to examine these two aspects of gluconeogenesis, we combined the hepatic veinbrachial artery catheterization technic with an isotopic infusion of alanine-14C. Alanine-14C specific activity was measured in whole blood and plasma by use of a rapid chromatographic technic. Since plasma contributed 93 per cent of the alanine extracted by the splanchnic bed with a specific activity three times that of the red
blood cells, plasma alanine specific activity was used to study the conversion of alanine to glucose. A constant infusion of alanine-14C achieved a relatively stable arterial specific activity by forty minutes. The administration of glucagon by constant infusion (15-50 ng./kg./min.) had no effect on the splanchnic extraction of alanine. Net splanchnic glucose-1 *C production, however, doubled during the glucagon infusion, and the conversion of alanine to glucose increased from 30 ± 2 to 58 ± 9 fimol/nun. These data (1) demonstrate that in normal man fasted twelve to fourteen hours, glucagon at supraphysiblogic levels can double the rate of gluconeogenesis from alanine and (2) indicate that this stimulatory effect of glucagon is exerted within the liver by shunting the extracted alanine toward new glucose formation rather than by increasing the hepatic extraction of alanine. DIABETES 24:574-84, June, 1975.
The importance of amino acids as a source of fuel, particularly for glucose production during periods of fasting, is generally recognized.1 In recent years several studies employing organ catheterization technic and column chromatography for plasma amino acid analysis have contributed significantly to our understanding of the role of amino acids in gluco^ neogenesis.2"5 These studies have clearly documented the flow of amino acids from muscle to liver under various metabolic conditions. Of all the amino acids, alanine has been assigned a primary role
as a gluconeogenic precursor based on the observations that its hepatic uptake represents nearly 50 per cent of the total amino acids extracted by the liver in man and that its hepatic extraction increases markedly after a thirty-six to forty-eight-hour fast.2 While these studies in man have helped define the interorgan flux of amino acids from muscle to liver, they have provided little information on the intrahepatic fate of the amino acids once extracted. Without this information, meaningful conclusions concerning the process of gluconeogenesis itself are difficult to derive. Using tracer technics, both in vitro6"8 and in vivo49"1 x studies have demonstrated that alanine could be converted to glucose by the liver. While the tracer technic can provide information concerning the fate of an amino acid extracted by the liver, the use of a tracer alone gives little information concerning the total uptake of the amino acid by the liver. Since gluconeogenesis from alanine could be controlled
This work was presented in part at the Canadian Society for Clinical Investigation held in Montreal, Canada, on January 23, 1974. From the Department of Medicine, Vanderbilt University School of Medicine, Nashville, Tennessee 37232. Address reprint requests to: J. L. Chiasson, Department of Medicine, Vanderbilt University School of Medicine, Nashville, Tennessee 37232. Accepted for publication March 12, 1975.
574
DIABETES, VOL. 2 4 , NO. 6
J. L. CHIASSON, M.D., AND ASSOCIATES
by altering the hepatic extraction of alanine and/ or by affecting the intrahepatic disposition of the extracted alanine, a measurement of both is essential in a study on the regulation of gluconeogenesis. For this purpose we combined an isotopic infusion of alanine- 14 C (to trace the intrahepatic conversion of alanine- 14 C to glucose-14C) with our hepatic veinbrachial artery catheterization technic (to measure the splanchnic extraction of alanine and alanine- 14 C and the release of glucose-14C). Accumulated evidence indicates that glucagon may be a key hormone in regulating the flow of amino acids to glucose. It is well established that glucagon can stimulate gluconeogenesis in certain systems in vitro, especially in the isolated perfused rat liver. 6p7il2i15 We have recently shown that glucagon can stimulate the conversion of alanine- 1 4 C to glucose- 14 C in the intact dog.51 Direct evidence, however, that this hormone plays a major role in regulating gluconeogenesis in man is still lacking. Circumstantial evidence supporting this view derives from the observations that glucagon levels are increased during fasting, 1 6 s t a r v a t i o n , 1 7 1 8 and hypoglycemia, 1 9 all conditions wherein gluconeogenesis is thought to be markedly stimulated. Further indirect evidence comes from a case report of Levy et al. 2 0 describing a patient with isolated glucagon deficiency in whom fasting hypoglycemia was a prominent symptom. The present study was designed to determine whether glucagon can stimulate gluconeogenesis from alanine in normal man and to examine its action on both hepatic uptake and intrahepatic disposition of alanine. The ideal setting to demonstrate any stimulatory effect of glucagon on gluconeogenesis is a state wherein glucagon is relatively low and gluconeogenesis is minimal. For this purpose, we choose to study the effect of glucagon on gluconeogenesis in normal man after an overnight fast (twelve to fourteen hours), a state fulfilling these two criteria. Since the catheterization procedure entailed a small but definite risk and since glucagon has never been shown to stimulate gluconeogenesis in man, we have used supraphysiologic doses of glucagon, hoping to get a maximum stimulative effect on this process. MATERIALS AND PREPARATIONS Sterile, pyrogen-free L-alanine-U- 1 4 C (256 mCi/mmol) (New England Nuclear, Boston, Mass.) was used as a tracer. The glucagon used in this study was purchased from Eli Lilly Laboratories (InJUNE,
1975
dianapolis, Ind.). Hepatic blood and plasma flow were determined by using indocyanine green (Hynson, Westcott, and Dunning, Inc., Baltimore, Md.). Phadebas Insulin Radioimmunoassay Kit was purchased from Pharmacia Fine Chemicals, Inc. (Piscataway, N J . ) and Trasylol from FBS Pharmaceuticals, Inc. (New York, N.Y.). All enzymes and cofactors used in the determinations of lactate, pyruvate, and glycerol were obtained from Boehringer Mannheim Corp. (New York, N.Y.). For infusion, glucagon was mixed with saline in a sterile 100-ml. volumetric flask containing 2-3 ml. of the subject's blood to eliminate binding of glucagon to glassware. Alanine-U- 14 C, 100 ;uCi, was mixed with saline in a sterile 50-ml. volumetric flask to which 1 ml. of the subject's plasma had been added. Three milliliters of plasma was also added to freshly prepared indocyanine green in order to increase its stability. 21
SUBJECTS Eight normal male volunteers participated in this study (table 1). All subjects were screened prior to the study and were accepted only when found to fulfill the following criteria: None of the subjects had any personal history of diabetes mellitus or any other endocrine or major disease. Each subject had a normal standard three-hour oral glucose tolerance test (40 gm. of glucose per square meter of body surface).22 Hepatic function assessed by measurement of bromsulfthalein extraction, serum glutamic-oxaloacetic transaminase, serum bilirubin, and alkaline phosphatase was normal. Each subject had a normal serum urea nitrogen and urinalysis, and cardiovascular function was normal by history, physical examination, chest x-ray, and electrocardiogram. Complete blood counts including a sedimentation rate were normal in all subjects. The subjects were all placed on a 300-gm. carbohydrate diet to which they adhered for three days prior to the study. All subjects were fasted twelve to fourteen hours prior to the study. The protocol for the present study was reviewed and approved by the Vanderbilt University Clinical Investigation Committee. This committee has the responsibility for protecting the rights of human volunteers participating in research studies. The nature, purpose, and possible risks of the procedures were fully explained to each subject before obtaining his voluntary consent. 575
EFFECT OF GLUCAGON ON GLUCONEOGENESIS FROM ALANINE IN MAN
TABLE 1 Clinical data on subjects
Age Glucagon group (n=5) Mean S.E.M. Control group (n= 3) Mean S.E.M.
Height
Weight
S.A.*
cm.
kg-
m.2
EHPFf ml./min.
FBSt mg./lOOml.
27.8
179.6
2.9
1.9
82 4.4
1.99 0.05
960
95.2
24.3
183.2
74.2
2.8
4.9
5.3
1.98 0.07
1,156 76.9
62.4
2.5 89.6 0.88
*Body surface area in square meters. fMean estimated hepatic plasma flow as milliliters per minute. ^Fasting blood glucose at the beginning of the study.
PROCEDURES
In the Vanderbilt University Hospital Cardiac Catheterization Laboratory, a Teflon catheter was inserted percutaneously in the left brachial artery. A 1-cm. cutdown was then performed over a small tributary vein distal to the right antecubital fossa. A no. 8F Cournand catheter was guided through this vein to a right-sided hepatic vein and under fluoroscopy was positioned 3-4 cm. from the wedge position. A continuous saline infusion maintained the patency of the hepatic vein catheter without added anticoagulant. In all subjects a trace amount of L-alanine-U-14C was administered in a peripheral vein as a constant infusion at a rate of 0.86 /uCi/min. throughout the 100-minute study period. After a sixty-minute basal period, a glucagon (15 ng./kg./min., n = \; 25 ng./kg./min., n — \\ 50 ng./kg./min., »=3) or saline (» = 3) infusion was started and continued over the last forty minutes. In all subjects, indocyanine green was administered in a peripheral vein as a constant infusion throughout the study for determination of hepatic blood flow. Simultaneous blood samples were obtained from the brachial artery and hepatic vein throughout the study. Approximately 500 ml. of blood was withdrawn during the study, the volume of which was replaced with saline infused throughout the study at a rate of 7 ml. per minute. ANALYTIC METHODS The system for determination of alanine concentration and specific activity consisted of two identical chromatographic columns, 0.9 cm. internal diameter, packed with Beckman Spherical Resin, Type AA15 (Beckman Instrument Co., Palo Alto, Calif.). Separate buffer pumps, with identical flow rates, were used on each column and were fed from common buf576
fer and NaOH reservoirs through three-way stopcocks. The effluent from the two columns led to a four-way valve arranged so that when one column was directed to a fraction collector through a "T" fitting, the other column was directed to waste. The effluent stream directed to the fraction collector was continuously sampled from the "T" fitting by a Technicon AutoAnalyzer, which reacted a small portion of the column effluent with ninhydrin for quantification of the alanine peak. The remainder of the effluent stream went directly to the fraction collector, where an appropriately timed fraction collected only the alanine peak. The radioactivity in this fraction was determined by counting a 2-ml. aliquot in 10 ml. of Aquasol (New England Nuclear, Boston, Mass.) for twenty minutes in a Packard Tri-Carb Scintillation Spectrometer (Packard Instrument Company, Dowers Grove, 111.). The following chromatographic conditions were used: column height 18.5 cm., temperature 65°C, buffer flow rate 2.2 ml./min., buffer pH 2.75 to 2.98, and Na concentration 0.2N. Under these conditions, glycine and alanine are well separated, with an elution time for alanine of about 23.5 minutes. Glucose- 14 C, lactate- 14 C, and pyruvate- 14 C eluted in the first five minutes. In this system, glutathione elutes with a group of unseparated amino acids well ahead of glycine and alanine. Plasma or whole-blood samples were prepared for analysis by the addition of an equal volume of 6 per cent sulfosalicylic acid. After chilling for forty-five minutes at 4° C., the precipitate was separated at 12,000 X g in a refrigerated centrifuge. One milliliter of the clear filtrate was applied to the chromatographic column. When plasma was compared with whole blood, the samples were processed in parallel, with whole blood added to the 6 per cent sulfosalicylic acid immediately after it was drawn. The plasma was DIABETES, VOL. 2 4 , NO. 6
J. L. CHIASSON, M.I
separated by centrifugation and added to sulfosalicylic acid within five minutes. A filtrate of plasma and whole blood from a single sample of venous blood to which was added a known amount of alanine-14C was analyzed along with unknown samples as a quality control. Replicate analyses over a period of six weeks gave the following statistics on the reproducibility of the method: the mean plasma alanine concentration was 455 /nmol/L., with a S.D. of 15 fimollL., and the mean per cent recovery of added 14C counts was 96.9 per cent, with a S.D. of 1.9 per cent (n=30); the mean whole-bk>od alanine concentration was 423 /umol/L., with a S.D. of 13 /amol/L., and the mean per cent recovery of added alanine-14C was 97 per cent, with a S.D. of 2.2 per cent (n=15). Glucose-14C was determined as described in a previous paper.i] Since glucose equilibrates almost instantly with the intracellular water of erythrocytes,23'24 the net splanchnic glucose-14C production rate was calculated by multiplying the arterial-hepatic venous differences by the plasma flow plus 70 per cent (intracellular water distribution) of the red blood cell flow. Whole-blood glucose was measured within four minutes by the Hoffman ferricyanide reaction using the Technicon AutoAnalyzer.25 Immunoreactive glucagon was assayed by Dr. Roger Unger, Dallas, Texas, using the 30K antiserum.26 Immunoreactive insulin was assayed by the Sephadex bound-antibody procedure. 27 The plasma concentrations of indocyanine green were determined in a Beckman spectrophotometer at 815 m/x and calculated according to the method of Leevy et al. 28 Whole blood was deproteinized with 30 per cent perchloric acid immediately after sampling and kept on ice for determination of lactate, pyruvate, and glycerol by a modified enzymatic fluorometric method.29"31 The statistical significance was determined by the Student's test or paired t test whenever applicable.32
AND ASSOCIATES
RESULTS
Glucagon, Blood Glucose, Insulin, and Estimated Hepatic Plasma and Blood Flow
The basal arterial plasma glucagon level was stable at a mean of 134 ± 2 . 5 pg./ml. and was unaffected by saline infusion. Administration of glucagon at 50 ng./kg./min. (n = 3) resulted in a rapid increase from basal levels to 5,066 ± 536 pg./ml. Glucagon infused at 15 ng./kg./min. (n = l)and25 ng./kg./min. (n = l) produced arterial plasma concentrations of 1,700 and 2,450 pg./ml., respectively, within ten to fifteen minutes. Since the response of the parameters measured was maximum with the 15 ng./kg./min. glucagon infusion, the five glucagon-treated subjects were considered as a single group for analytic purposes. The mean net splanchnic glucose production in the controls was stable over the 100-minute study period at 159-3 ± 5 . 7 mg./min. (table 2), maintaining an arterial blood glucose of 90.3 ± 0.5 mg./lOO ml. In the glucagon-treated group the net splanchnic glucose production increased from a mean basal level of 142.9 ± 7 . 3 mg./min. to a peak of 504.2 ± 80.6 mg./min. ten minutes after glucagon infusion was started. This peak was then followed by a gradual decline to 253-5 ± 10.8 mg./min. by 100 minutes despite continuing glucagon infusion. The mean arterial blood glucose level increased during glucagon infusion from a basal level of 94.5 ± 0.5 mg./lOO ml. to a maximum of 173.2 ± 11.7 mg./lOO ml. by the end of the study. The mean arterial plasma immunoreactive insulin concentration of 5.7 ± 0.4 ju,U./ml. in the controls did not change during saline infusion. In the glucagon-treated subjects, insulin gradually rose from the mean basal levels of 6.9 ± 0 . 1 /u.U./ml. at sixty minutes to 24.4 ± 4 . 2 /u,U./ml. by" 100 minutes. The estimated mean hepatic blood and plasma flow measured by the indocyanine green technic were 1,781 ± 104 ml./min. and 1,034 ± 58 ml./min.,
TABLE 2 Effect of glucagon and saline on net splanchnic glucose production
Glucagon group Mean S.E.M. Control group Mean S.E.M.
Net splanchnic glucose production (mg./min.) Basal Period 30' 40' 50' 60'* 70'
Glucagon or Saline 80' 90'
100'
0'
10'
20'
109 27
130 11
132 15
160 2
152 23
155 31
140 12
504 81
446 57
348 42
253 11
161 34
163 29
160 10
148 6
153 20
177 42
131 16
142 19
160 17
149 65
137 28
*Glucagon or saline infusion was begun after the sixty-minute sample. JUNE, 1975
577
EFFECT OF GLUCAGON ON GLUCONEOGENESIS FROM ALANINE IN MAN
respectively. The flow was very stable throughout the study and was not affected by the glucagon infusion. Effect of Glucagon on the Dynamics of Alanine-14C Across the Splanchnic Bed
Flux
To determine whether whole blood, plasma, or both whole blood and plasma concentrations and specific activities would be needed to estimate the splanchnic extraction of alanine and the conversion of circulating alanine into glucose, we measured the alanine concentration and radioactivity in both whole blood and plasma before and during glucagon administration in four subjects. Eight comparisons were made before and eight during the glucagon infusion. Since glucagon had no effect on the arterial-hepatic venous differences, the means of all sixteen comparisons are shown in table 3. It is apparent that under the conditions of our study, most of the net splanchnic extraction of both total alanine and 14C-alanine was from the plasma compartment, with red cells contributing less than 7 per cent. Table 4 shows the mean alanine-14C specific activity in plasma, whole blood, and red blood cells. Of particular note is the fact that the specific activity was different in the plasma and cell compartments. Since most of the alanine is extracted from plasma at a higher specific activity than in whole blood, we chose to measure alanine in plasma rather than in whole blood for the remainder of the studies reported below. The mean alanine specific activities observed in arterial and hepatic venous plasma during a constant infusion of alanine-14C in control and glucagontreated subjects are shown in figures 1A and B. Although a completely flat plateau is not reached, the change in specific activity between forty and 100 minutes was less than 6 per cent. The ratio of the
TABLE 4 Arterial and hepatic venous alanine-14C specific activity in whole blood, plasma, and blood cells*
Mean S.E.M.
Whole blood Artery Hepatic vein cpm//i,mol 3365 2195 174 176
Plasma Artery Hepatic vein cpm/ftmol 4967 3422 226 268
Blood cells Artery Hepatic vein cpm//i,mol 1685 1491 152 148
*n= 16 (four observations in four subjects).
hepatic venous specific activity to the arterial specific activity (figure 1) is a measure of the dilution of plasma alanine by unlabeled alanine from the splanchnic bed. The figure shows that the ratio rose during the initial period of the constant infusion of alanine-14C, but reached a plateau by forty minutes and remained constant thereafter, with a mean value of 0.65, indicating a 35 per cent dilution of the specific activity. Glucagon had no demonstrable effect on this ratio. Effect of Glucagon on Uptake of Gluconeogenic Precursors
Figure 2 shows the mean arterial plasma alanine concentration and mean net splanchnic alanine extracALANINE- I4 C
6
ALANINE SPEC. ACT. 4 CPM X I0"3 PER 2
!'
INFUSION
0.86/JCL/MIN
ortery(n=4)
>JMOLE
0 HV:A
l 0
RATIO
0.5
1-1
I
I
0
TABLE 3 Contribution of plasma and red blood cell to alanine arterial-hepatic venous differences* Plasma Whole blood Cell contentf cone. contents^ Alanine in yuxnol/L. of whole blood A§ 128 ± 5 249 ± 10 121 ± 8 A-HV 64 ± 3 69 ± 3 5±2 Alanine-14C in cpm x 1(T3/L. of whole blood 194 ± 14 622 ± 19 817 ± 23 33 ± 9 417 ± 11 450 ± 12 A-HV •All values subjects). •(•Calculated ^Calculated §A=artery;
578
Plasma contribution % 92.5 ± 3 % 93.1 ± 2
are mean ± S.E.M., n= 16 (four observations in four as plasma concentration x (1-Hct). as whole blood minus plasma content. A-HV= arterial-hepatic venous difference.
ALANINE SPEC. ACT. 4 CPM X I 0 3 PER 2 juMOLE 0 HV:A RATIO
'° o.5
20
40 60 MINUTES
80
100
FIG. 1. The effects of glucagon (A) and saline (B) on the arterial and hepatic venous alanine specific activity and the hepatic venous:arterial (HV:A) alanine specific activity ratio. Values are the means ± S.E.M. DIABETES, VOL. 2 4 , NO. 6
J. L. CHIASSON, M . D . , A N D ASSOCIATES
ALANINE-WC
ARTERIAL ALANINE /iMOLES/L
200
6LUCAG0N
®
300
I
\
I
I
100
NET
from the product of hepatic plasma flow times the arterial-hepatic venous differences. The mean net splanchnic extraction of lactate was 119 — 53 /xmoles/min. before and 118 ± 58 jumoles/min. after glucagon infusion. Net splanchnic extraction of glycerol was 46 ± 9 /imoles/min. before and 49 ± 9 /umoles/min. after glucagon infusion. The net splanchnic balance of pyruvate was very small and was not affected by glucagon. Thus, glucagon had no significant effect on the net splanchnic balance of the four major gluconeogenic precursors—alanine, lactate, pyruvate, and glycerol.
INFUSION 0.86juCi/MIN
I
-*—g-
(n = 4)
°
SPLANCHNIC ? n n ALANINE ^ U U UPTAKE ,00
s—*-
>JMOLES/MIN
0
SALINE
Aguilar-Parada, E., Eisentraut, A. M., and Unger, R. H . : Pancreatic glucagon secretion in normal and diabetic subjects. Am. J. Med. Sci. 257:415, 1969. 2 'Wide, L., and Porath, J.: Radioimmunoassay of proteins with the use of Sephadex-coupled antibodies. Biochem. Biophys. Acta 730:257, 1966. 28 Leevy, C. M., Mendenhall, C. L., Lesko, W . , and Howard,
584
M. M.: Estimation of hepatic blood flow with indocyanine green. J. Clin. Invest. 47:1169, 1962. 2!l Hohorst, H. J.: L-lactate determination with lactate dehydrogenase and DPN. In Methods of Enzymatic Analysis, Bergmeyer, H. V., Ed. New York, Academic Press. 1963, p. 266. 30 Butcher, T., Rudolf, C , Lamprecht, W . , and Latzko, E.: Pyruvate. In Methods of Enzymatic Analysis. Bergmeyer, H. V., Ed. New York, Academic Press. 1963, p. 253. 31 Chernick, S. S.: Determination of glycerol in acyl glycerols. In Methods of Enzymology, vol. 14. Lowenstein, John M., Ed. 5th ed. New York, Academic Press, 1969, p. 627. 32 Snedecor, G. W.: Statistical Methods, 5th ed. Ames, Iowa, State University Press, 1965, p. 35. 33 Elwyn, D. H.: Distribution of ami no acids between plasma and red blood cells in the dog. Fed. Proc. 25:854, 1966. 34 Elwyn, D. H . , Launder, W. J . , Pariksh, H. C , and Wise, E. M., Jr.: Role of plasma and erythrocytes in interorgan transport of ami no acids in dogs. Am. J. Physiol. 222:1333, 1972. li5 Felig, P., Wahren, J . , and Raf, L.: Evidence of interorgan amino acid transport by blood cells in humans. Proc. Nat. Acad. Sci. U.S.A. 70:1775, 1973. 3f> Aoki, T. T., Muller, W. A., Brennan, M. F., and Cahill, G. F., Jr.: Blood cell and plasma amino acid levels across forearm muscle during a protein meal. Diabetes 22:768, 1973. 37 Winter, C. G., and Christensen, H. N . : Migration of amino acids across membrane of the human erythrocyte. J. Biol. Chem. 239:872, 1964. 38 Haft, D. E., Tennen, E., and Mehtalia, S.: Alanine metabolism in perfused livers of normal and adrenalectomized rats. Am. J. Physiol. 222:365, 1972. 3! 'Krebs, H. A., Hemo, R., Weidemann, M. J., and Speake, R. N . : The fate of isotopic carbon in kidney cortex synthesizing glucose from lactate. Biochem. J. 101:242, 1966. 40 Rocha, D. M., Santeusanio, F., Faloona, G. R., Unger, R. H.: Abnormal pancreatic alpha-cell function in bacterial infections. N. Engl. J. Med. 288:700, 1973. 41 Unger, R. H.: Glucagon and the insulin:glucagon ratio in diabetes and other catabolic illnesses. Diabetes 20:834, 1971.
DIABETES, VOL. 2 4 , NO. 6
