Glucagon and Catecholamine Secretion During Hypoglycemia in Normal and Diabetic Man JAMES W. BENSON, JR., DAVID G. JOHNSON, JERRY P. PALMER, PHILLIP L. WERNER, AND JOHN W. ENSINCK Department of Medicine, University of Washington, School of Medicine, Seattle, Washington 98195 ABSTRACT. To examine the role of catecholamines in glucopenia-induced glucagon secretion, urinary epinephrine and norepinephrine and plasma immunoreactive glucagon (IRG) were measured during insulin-induced hypoglycemia in normal and insulin-dependent diabetic man. Despite equivalent levels of hypoglycemia the mean plasma IRG increment in diabetics was only 15% of normals. The increases in epinephrine and norepinephrine were comparable in diabetics and normals; thus, the blunted response in plasma IRG to the stimulus of low blood sugar in diabetics cannot be explained by a generalized defect in catecholamine secretion. It is suggested that an alpha-cell glucoreceptor
defect may account for the abnormal response to glucopenia in diabetics. To evaluate the possibility that impaired sensitivity to circulating catecholamines could explain the alpha cell dysfunction in diabetics, exogenous epinephrine was infused in normals and insulindependent diabetics. Elevated plasma IRG during epinephrine infusion was observed in only 50% of normals. The diabetics were hyperresponsive; mean increment in plasma IRG 3 times that of normals. The data demonstrate enhanced rather than impaired sensitivity to catecholamines in diabetes mellitus. (J Clin Endocrinol Metab 44: 459, 1977)
H
YPOGLYCEMIA stimulates glucagon secretion in vitro in the isolated, perfused pancreas (1,2) and in vivo in experimental animals (3) and normal man (4). In insulin-dependent diabetic man, however, the alpha cell response to glucopenia is abolished (5). Although there is strong evidence in man (6-8) and in vitro (1,2) that glucopenia-induced glucagon secretion is independent of catecholamines, it has been suggested that the defect in alpha cell response to hypoglycemia in diabetes resides in the sympathetic nervous system (9). During infusion of catecholamines in the isolated rat pancreas (2) and man (10), plasma immunoreactive glucagon (IRG) levels are elevated. In addition, in stress
states such as acute myocardial infarction (11), burns (12), trauma (13), sepsis (14) and diabetic ketoacidosis (15), there is an elevation of both plasma catecholamines and plasma IRG. Neither of these lines of evidence, however, is proof of a physiologic causal relationship between catecholamines and glucagon secretion in man, since both epinephrine infusions and stressful states entail a multiplicity of metabolic changes. We have sought to dissect this relationship in normal and diabetic subjects by 1) directly measuring urinary catecholamines and plasma IRG during insulin-induced hypoglycemia and 2) measuring plasma IRG in response to exogenous catecholamines. Materials and Methods
Received July 9, 1976. Supported in part by the Clinical Research Center at the University of Washington (GCRC Grant RR-37) and by NIH grants AM-16008 and AM-17698. Dr. Benson is a recipient of a National Institutes of Health Research Fellowship (1-F22-AM02400-02). Dr. Johnson was supported by a Research Career Development Award (1-K04-AM-70727). Reprint requests should be addressed to: John W. Ensinck, M.D., Department of Medicine, University of Washington, Seattle, Washington 98195.
The control population consisted of 12 nonobese, healthy volunteers (11 male and 1 female) between the ages of 19 and 36 years (24.5 ± 1.2 years, mean ± SE) with fasting plasma glucose levels less than 102 mg/dl and no family history of diabetes. No subjects were taking medications at the time of the study. The insulindependent, non-obese diabetic population comprised 9 men between the ages of 18 and 42 years (25.0 ± 2.4 years) without intercurrent ill-
459
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 March 2016. at 23:26 For personal use only. No other uses without permission. . All rights reserved.
ness and taking no medications other than insulin. None had historical or clinical evidence of peripheral or autonomic neuropathy. All volunteers were informed of the nature, purpose and potential risks of the procedures. Tests were begun at 0800 h in the Clinical Research Center following an overnight fast. Indwelling catheters for administration of drugs and collection of blood samples were kept patent with 0.9% saline. Blood samples for plasma glucose and plasma IRG levels were obtained every 15 min during the 1 h control period and every 10 min thereafter for 120 min. Insulin-induced
hypoglycemia
Regular insulin was administered iv over 5-10 sec in 7 normals and 6 diabetics. All normal volunteers reached a nadir of plasma glucose less than 30 mg/dl after receiving 0.15 U/kg of insulin. In order to achieve the same degree of hypoglycemia, the diabetics required 0.12-0.6 U/kg (mean 0.34 U/kg). Urine collections for epinephrine and norepinephrine were obtained for the 1 h control period and for the 2 h following insulin injection. Epinephrine
JCE&M Vol 44
BENSON ETAL.
460
infusions
Epinephrine in normal saline was infused at 6 /Ag/min for 1 h in 7 diabetics and 8 normals. The infusate contained ascorbic acid, 2.5 mg/ml, to prevent oxidation of epinephrine (16). Determinations Plasma glucose was determined by the glucose oxidase method on the Beckman Glucose Analyzer (Beckman Instruments, Fullerton, Ca.). In our IRG assay, each plasma sample is extracted with acetone and benzamidine is added (final concentration 0.01M) to inhibit degradation during incubation with 30K antiserum (6). Acetone extraction has been found to remove IRG species of molecular weight greater than 3500 daltons (unpublished observation). Urinary epinephrine and norepinephrine were measured by a modification of the fluorimetric method of von Euler and Lishajko (17). Urine collections were immediately acidified to keep pH less than 3.0 with O.lM HC1 and frozen until assays were performed. To 15 ml aliquots of filtered urine were added 15 mg of NaS2O5 and 5 ml 2% NaEDTA. The aliquots were immedi-
1977 No 3
ately adjusted to pH 8.6 with 1.0N NaOH and passed over a column of 400 mg alumina. The column was washed three times with 0.2M sodium acetate buffer, pH 8.6 and twice with water. 0.2M catecholamines were eluted with 6 ml of 0.2 acetic acid. Catecholamine concentrations in the eluate were determined by the trihydroxyindole method. A 1.0 ml aliquot was brought to pH 6.2-6.5 by addition of 0.1 of phosphocarbonate buffer. The phosphocarbonate buffer consists of a 1.1 mixture of 0.5M potassium phosphate buffer, pH 6.5 and 5.0M potassium carbonate. Oxidation was started by 0.1 ml of 0.25% K3FeCN6. After 3 min, 2 ml of the sodium hydroxide-ascorbic acid mixture were added. Blanks consisted of samples to which the sodium hydroxide-ascorbic acid mixture was added before the 0.25% K3FeCN6. External standards of norepinephrine and epinephrine (100 ng) were run with each assay and the concentration of both catecholamines in the samples calculated using a Hewlett Packard Model 9830A computer. Statistical analysis employing Student's t test and paired t test as well as calculation of IRG areas were performed on a Wang Model 600 computer (Wang Laboratories, Inc., Tewksbury, Mass.). Values are expressed as mean ± SE.
Results Comparison of effects of insulin-induced hypoglycemia in normals and diabetics There was no significant difference in the mean basal plasma IRG between diabetics (52.0 ± 2.5 pg/ml) and normals (49.0 ± 3.7 pg/ml). As shown in Fig. 1, the nadir of plasma glucose in diabetics (21.2 ± 1.8 mg/dl) was not significantly different from that in normals (23.7 ± 2.4 mg/dl), but occurred later in diabetics. Despite equal and protracted hypoglycemia, all diabetics demonstrated a marked attenuation of the plasma IRG response. Normals showed a significant increase in plasma IRG beginning 20 min after insulin injection with peak increase of 131.7 ± 15.2 pg/ml above basal. In contrast, in all diabetics the plasma IRG initially decreased significantly to a mean of 9.4 ±1.7 pg/ml below basal, then showed little change from baseline. In diabetics, there was a significant increase above
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 March 2016. at 23:26 For personal use only. No other uses without permission. . All rights reserved.
461
CATECHOLAMINES AND GLUCAGON IN MAN basal only at 70 and 80 min with peak increase in plasma IRG of only 17.6 ± 2.3 pg/ml above basal. The difference in absolute increase in plasma IRG between diabetics and normals was significant at all points after 10 min. The area under the IRG curve was 985 ± 243 pg/ml/120 min in diabetics and 7310 ± 869 pg/ml/120 min in normals (P < .001). Urinary norepinephrine and epinephrine excretion for control and experimental periods is shown in Table 1. Only the ratio of EPIhyP0giyCemia/EPIC0ntI.0l was significantly different between the two populations and was, in fact, increased in diabetics. There was a trend toward enhanced excretion of both epinephrine and norepinephrine during hypoglycemia in diabetics, but this did not achieve statistical significance.
\
604020
"
o
1401301008060- \ 40200-
Comparison of effects of epinephrine infusion in normals and diabetics In normal subjects basal plasma glucose was 89.5 ± 2.8 mg/dl and basal plasma IRG was 62.2 ± 6.2 pg/ml; in diabetics these values were 246.4 ± 67 mg/dl and 52.0 ± 4.5 pg/ml, respectively. Of 8 normal subjects tested, only 4 showed increased plasma IRG during epinephrine infusion; in all normals there was suppression of plasma IRG to below basal during the last hour of observation. The mean plasma IRG in normals was significantly above basal during the first 40 min of infusion with peak increment
Insulin 160- Injection 140120100-
Normal n = 7 Diabetic n=6 p .05.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 March 2016. at 23:26 For personal use only. No other uses without permission. . All rights reserved.
462
BENSON ETAL. EPINEPHRINE 6//g/min
•—• Normal n=7 o—o Diabetic n=8
60\
p
defect may account for the abnormal response to glucopenia in diabetics. To evaluate the possibility that impaired sensitivity to circulating catecholamines could explain the alpha cell dysfunction in diabetics, exogenous epinephrine was infused in normals and insulindependent diabetics. Elevated plasma IRG during epinephrine infusion was observed in only 50% of normals. The diabetics were hyperresponsive; mean increment in plasma IRG 3 times that of normals. The data demonstrate enhanced rather than impaired sensitivity to catecholamines in diabetes mellitus. (J Clin Endocrinol Metab 44: 459, 1977)
H
YPOGLYCEMIA stimulates glucagon secretion in vitro in the isolated, perfused pancreas (1,2) and in vivo in experimental animals (3) and normal man (4). In insulin-dependent diabetic man, however, the alpha cell response to glucopenia is abolished (5). Although there is strong evidence in man (6-8) and in vitro (1,2) that glucopenia-induced glucagon secretion is independent of catecholamines, it has been suggested that the defect in alpha cell response to hypoglycemia in diabetes resides in the sympathetic nervous system (9). During infusion of catecholamines in the isolated rat pancreas (2) and man (10), plasma immunoreactive glucagon (IRG) levels are elevated. In addition, in stress
states such as acute myocardial infarction (11), burns (12), trauma (13), sepsis (14) and diabetic ketoacidosis (15), there is an elevation of both plasma catecholamines and plasma IRG. Neither of these lines of evidence, however, is proof of a physiologic causal relationship between catecholamines and glucagon secretion in man, since both epinephrine infusions and stressful states entail a multiplicity of metabolic changes. We have sought to dissect this relationship in normal and diabetic subjects by 1) directly measuring urinary catecholamines and plasma IRG during insulin-induced hypoglycemia and 2) measuring plasma IRG in response to exogenous catecholamines. Materials and Methods
Received July 9, 1976. Supported in part by the Clinical Research Center at the University of Washington (GCRC Grant RR-37) and by NIH grants AM-16008 and AM-17698. Dr. Benson is a recipient of a National Institutes of Health Research Fellowship (1-F22-AM02400-02). Dr. Johnson was supported by a Research Career Development Award (1-K04-AM-70727). Reprint requests should be addressed to: John W. Ensinck, M.D., Department of Medicine, University of Washington, Seattle, Washington 98195.
The control population consisted of 12 nonobese, healthy volunteers (11 male and 1 female) between the ages of 19 and 36 years (24.5 ± 1.2 years, mean ± SE) with fasting plasma glucose levels less than 102 mg/dl and no family history of diabetes. No subjects were taking medications at the time of the study. The insulindependent, non-obese diabetic population comprised 9 men between the ages of 18 and 42 years (25.0 ± 2.4 years) without intercurrent ill-
459
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 March 2016. at 23:26 For personal use only. No other uses without permission. . All rights reserved.
ness and taking no medications other than insulin. None had historical or clinical evidence of peripheral or autonomic neuropathy. All volunteers were informed of the nature, purpose and potential risks of the procedures. Tests were begun at 0800 h in the Clinical Research Center following an overnight fast. Indwelling catheters for administration of drugs and collection of blood samples were kept patent with 0.9% saline. Blood samples for plasma glucose and plasma IRG levels were obtained every 15 min during the 1 h control period and every 10 min thereafter for 120 min. Insulin-induced
hypoglycemia
Regular insulin was administered iv over 5-10 sec in 7 normals and 6 diabetics. All normal volunteers reached a nadir of plasma glucose less than 30 mg/dl after receiving 0.15 U/kg of insulin. In order to achieve the same degree of hypoglycemia, the diabetics required 0.12-0.6 U/kg (mean 0.34 U/kg). Urine collections for epinephrine and norepinephrine were obtained for the 1 h control period and for the 2 h following insulin injection. Epinephrine
JCE&M Vol 44
BENSON ETAL.
460
infusions
Epinephrine in normal saline was infused at 6 /Ag/min for 1 h in 7 diabetics and 8 normals. The infusate contained ascorbic acid, 2.5 mg/ml, to prevent oxidation of epinephrine (16). Determinations Plasma glucose was determined by the glucose oxidase method on the Beckman Glucose Analyzer (Beckman Instruments, Fullerton, Ca.). In our IRG assay, each plasma sample is extracted with acetone and benzamidine is added (final concentration 0.01M) to inhibit degradation during incubation with 30K antiserum (6). Acetone extraction has been found to remove IRG species of molecular weight greater than 3500 daltons (unpublished observation). Urinary epinephrine and norepinephrine were measured by a modification of the fluorimetric method of von Euler and Lishajko (17). Urine collections were immediately acidified to keep pH less than 3.0 with O.lM HC1 and frozen until assays were performed. To 15 ml aliquots of filtered urine were added 15 mg of NaS2O5 and 5 ml 2% NaEDTA. The aliquots were immedi-
1977 No 3
ately adjusted to pH 8.6 with 1.0N NaOH and passed over a column of 400 mg alumina. The column was washed three times with 0.2M sodium acetate buffer, pH 8.6 and twice with water. 0.2M catecholamines were eluted with 6 ml of 0.2 acetic acid. Catecholamine concentrations in the eluate were determined by the trihydroxyindole method. A 1.0 ml aliquot was brought to pH 6.2-6.5 by addition of 0.1 of phosphocarbonate buffer. The phosphocarbonate buffer consists of a 1.1 mixture of 0.5M potassium phosphate buffer, pH 6.5 and 5.0M potassium carbonate. Oxidation was started by 0.1 ml of 0.25% K3FeCN6. After 3 min, 2 ml of the sodium hydroxide-ascorbic acid mixture were added. Blanks consisted of samples to which the sodium hydroxide-ascorbic acid mixture was added before the 0.25% K3FeCN6. External standards of norepinephrine and epinephrine (100 ng) were run with each assay and the concentration of both catecholamines in the samples calculated using a Hewlett Packard Model 9830A computer. Statistical analysis employing Student's t test and paired t test as well as calculation of IRG areas were performed on a Wang Model 600 computer (Wang Laboratories, Inc., Tewksbury, Mass.). Values are expressed as mean ± SE.
Results Comparison of effects of insulin-induced hypoglycemia in normals and diabetics There was no significant difference in the mean basal plasma IRG between diabetics (52.0 ± 2.5 pg/ml) and normals (49.0 ± 3.7 pg/ml). As shown in Fig. 1, the nadir of plasma glucose in diabetics (21.2 ± 1.8 mg/dl) was not significantly different from that in normals (23.7 ± 2.4 mg/dl), but occurred later in diabetics. Despite equal and protracted hypoglycemia, all diabetics demonstrated a marked attenuation of the plasma IRG response. Normals showed a significant increase in plasma IRG beginning 20 min after insulin injection with peak increase of 131.7 ± 15.2 pg/ml above basal. In contrast, in all diabetics the plasma IRG initially decreased significantly to a mean of 9.4 ±1.7 pg/ml below basal, then showed little change from baseline. In diabetics, there was a significant increase above
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 March 2016. at 23:26 For personal use only. No other uses without permission. . All rights reserved.
461
CATECHOLAMINES AND GLUCAGON IN MAN basal only at 70 and 80 min with peak increase in plasma IRG of only 17.6 ± 2.3 pg/ml above basal. The difference in absolute increase in plasma IRG between diabetics and normals was significant at all points after 10 min. The area under the IRG curve was 985 ± 243 pg/ml/120 min in diabetics and 7310 ± 869 pg/ml/120 min in normals (P < .001). Urinary norepinephrine and epinephrine excretion for control and experimental periods is shown in Table 1. Only the ratio of EPIhyP0giyCemia/EPIC0ntI.0l was significantly different between the two populations and was, in fact, increased in diabetics. There was a trend toward enhanced excretion of both epinephrine and norepinephrine during hypoglycemia in diabetics, but this did not achieve statistical significance.
\
604020
"
o
1401301008060- \ 40200-
Comparison of effects of epinephrine infusion in normals and diabetics In normal subjects basal plasma glucose was 89.5 ± 2.8 mg/dl and basal plasma IRG was 62.2 ± 6.2 pg/ml; in diabetics these values were 246.4 ± 67 mg/dl and 52.0 ± 4.5 pg/ml, respectively. Of 8 normal subjects tested, only 4 showed increased plasma IRG during epinephrine infusion; in all normals there was suppression of plasma IRG to below basal during the last hour of observation. The mean plasma IRG in normals was significantly above basal during the first 40 min of infusion with peak increment
Insulin 160- Injection 140120100-
Normal n = 7 Diabetic n=6 p .05.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 March 2016. at 23:26 For personal use only. No other uses without permission. . All rights reserved.
462
BENSON ETAL. EPINEPHRINE 6//g/min
•—• Normal n=7 o—o Diabetic n=8
60\
p
