0021-972x/92/7505-1282$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright 0 1992 by The Endocrine Society
Vol. 15, No. 5 Printed in U.S.A
The Effects of Human Proinsulin on Glucose Turnover and Intermediary Metabolism in Insulin-DependentDiabetes Mellitus* S. N. DAVIS?, C. N. HALES,
M. ANSIFEROV, H. ORSKOV,
AND
C. HETHERINGTON, M. BROWN, K. G. M. M. ALBERT1
W. J. BRANCH,
Department of Medicine, The Medical School, University of Newcastle upon Tyne, Newcastle upon Tyne, England; Department of Clinical Biochemistry (W.J.B., C.N.H.), Addenbrookes Hospital, University of Cambridge, Cambridge, England; Second University Clinic of Internal Medicine (H.O.), Aarhus Kommunhospitalet, Denmark; and Department of Medicine, Vanderbilt University School of Medicine, Diabetes Research and Training Center, B-3307 Medical Center North, Nashville, Tennessee 37232-2230 ABSTRACT We have
comparedthe action of humanproinsulinand insulinon glucoseturnover, intermediarycarbohydrate,and lipid metabolismin insulin-dependent-diabetic (IDDM) subjects. Six, young, weightmatched(23 * 2 kg-‘) IDDM subjects underwent separate hyperinsulinemic
euglycemic
clamps.
Three,
low dose, iv infusions
of both
insulin
and proinsulin were usedto construct doseresponsecurves. The proinsulin
infusions
were
chosen
to give
steady
state
levels
approxi-
matelyor equalto 20.fold higher on a molar basis than insulin, based on previousfindingsthat proinsulinhasonly 5-10% the biological potency of insulin. Hepatic glucose production, measured using [6’6’“H,]glucose, was suppressed equally by proinsulin and insulin at the three dose levels: (I,) 2.8 t 0.7 (P,) 3.3 + 0.6. (I,)_ 2.3 + 0.9 (P.,)_, 3.3 & 1.1, (I:{) -2.0 f 1.7 (P:J -1.1 + 0.6 Fmol/kg min-’ Percentage elevation of glucose disposal was significantly increased during the insulininfusionscompared to proinsulin;(I,) 132 f 12 (P,) 78 + 4 p < 0.01; (I~)‘157 + 18 (P,) 104 + 14; P < 0.05; (1:J 242 + 23 (P:,) 159 + 24
P
ROINSULIN and insulin concentrations are significantly reduced in insulin-dependent-diabetic (IDDM) subjects. However, the effects of proinsulin deficiency remain unknown as its physiological role in normal man is unresolved. Proinsulin has been estimated to comprise 6% (1) of the total insulin-like immunoreactivity in the portal vein of normal man. Recent work, however, in both normal and noninsulindependent-diabetic man (NIDDM), using highly specific proinsulin assays,has shown that peripheral circulating levels are approximately or equal to 10% compared to insulin (2). This apparent reversal of the normal portal-peripheral gradient is due to the fact that proinsulin has a much lower metabolic clearance rate and longer half-life compared to insulin (3, 5). Previous in viva dose response studies in normal and NIDDM man have demonstrated that, on a molar basis, proinsulin has a much lower biological potency comReceived February 11, 1991. Address correspondence and requests for reprints to: S. N. Davis, Department of Medicine, Vanderbilt University School of Medicine, Diabetes Research and Training Center, B-3307 Medical Center North, Nashville, Tennessee 37232-2230. *Financial support was given by the British Diabetic Association, Mason Medical Research Foundation, and Eli Lilly Company. t Medical Research Council Training Fellow.
p = 0.02. Dose response curve analysis demonstrated that proinsulin stimulated glucose disposal approximately or equal to 3.7% whereas suppression of HGP was ~5.7% compared to insulin. Proinsulin had a significantly weaker effect than insulin, at the lowest infusion dose, in percent suppression of plasma nonesterified fatty acids (I1 34 + 4, P, 14 + 15%; P < 0.05), blood glycerol (I, 47 + 4, P, 30 ? 3%; P < 0.01) and 3-hydroxybutyrate levels (I, 81 + 7, P, 42 + 17%; P < 0.05). Proinsulin caused significant net reductions in blood lactate levels compared to insulin at each infusion dose; (PI) -130 + 34, (I,) -32 + 30 pmol/L (P < 0.05) (P2) -139 + 76 (12) +8 + 65 Fmol/L (P < 0.05) (P:,) 48 + 60 (I?) 230 + 64 rmol/L (P < 0.05). We conclude that in IDDM: 1) proinsulin has a preferential effect on the liver compared to muscle, in terms of glucose handling; 2) proinsulin may have a different effect on lactate metabolism compared to insulin; 3) proinsulin at the lowest dose resulted in an inability to suppress lipolysis and ketogenesis; 4) glucose turnover can be underestimated using [6’6”H2]glucose. (J Clin Endocrinol Metab 75: 12821288,1992)
pared to insulin (3, 4). The majority of studies have demonstrated that proinsulin has a relatively greater effect on suppressinghepatic glucose production (5-12%) (3-7) compared to stimulating peripheral glucose uptake (3-8%) (3-6). Proinsulin and insulin’s effects on intermediary carbohydrate and lipid metabolism do not appear to be identical. We have previously shown, during isoglycemic clamp studies, in both normal and NIDDM man that blood lactate concentrations can be significantly reduced by proinsulin infusions (5, 6). In addition, blood glycerol, 3-hydroxybutyrate and plasma nonesterified-fatty acids (NEFA) levels were higher after a low dose proinsulin infusion compared to a biologically equipotent dose of insulin in NIDDM subjects. There is, however, scant data concerning proinsulin’s metabolic affects in IDDM subjects. Bergenstal et al. (8) investigated the dose of proinsulin needed to maintain normoglycemia after acute insulin withdrawal in IDDM subjects. Whereas Cohen et al. (9) studied the comparative effects of insulin and proinsulin on the hepatic response to glucagon infusion. This latter study did not find a relatively hepatospecific action of proinsulin
after
a prolbngeh
(>iS
h)
infusion
&ed
to
render
subjects normoglycemic. We were, therefore, particularly interested in examining the acute dose response effects of 1282
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
THE
EFFECTS
OF HUMAN
proinsulin on suppression of hepatic glucose production and glucose use in IDDM. In addition we have compared the effects of three biologically equipotent low dose iv infusions of human proinsulin and insulin on different aspects of intermediary lipid and carbohydrate metabolism, known to be abnormal in IDDM subjects. This will enable a more complete profile of proinsulin action in IDDM to be obtained. Materials
and Methods
Subjects We studied six, male, IDDM subjects. Mean age 31 f 8 yr (Table 1) all of whom had documented episodes of diabetic ketoacidosis. Each subject had normal blood count, plasma electrolytes, and liver function. None had detectable micro- or macrovascular complications of diabetes. All subjects gave written informed consent. Subjects were asked to take their usual weight maintaining diet for 3 days before each study. The amount of exercise before studies was controlled, so that no subject performed any strenuous exercise 72 h before a study and arrived via organized transport. All studies were performed in the Clinical Research Centre and received local Ethical Committee approval.
Materials Biosynthetic human proinsulin was prepared and characterized as previously described (lo), and kindly supplied by Eli Lilly Co. (Indianapolis, IN), as was regular human insulin. However due to adverse findings in a prospective proinsulin clinical trial performed elsewhere, Eli Lilly Co. acutely withdrew proinsulin before all the planned studies were completed. Therefore, two high dose and one low dose proinsulin infusion protocol, to be described below, could not be completed. 6’6Dideuterated glucose (99.3% mol/excess) was purchased from K. K. Grief (London, UK).
Glucose clamp studies Approximately 16 h before arriving at the Clinical Research Centre patients reduced their usual doses of regular and intermediate-acting insulins by approximately or equal to 70%. At 0800 h, after a 10-h overnight fast, a low dose iv infusion of regular insulin was started to render subjects normoglycemic. This infusion was continued until blood glucose had been stable for 3 h. We have three pieces of information which demonstrate that this method provided a similar metabolic profile to overnight glucose normalization: first, the basal free insulin values in this study (0.1 + 0.02 nmol/L) were similar to previous values of 0.08 f 0.1 nmol/L reported by Hother-Nielsen et al. (11). Second, mean basal hepatic glucose production (HGI’) values in the present study (12.8 + 1.1 pmol/kg min-’ were similar to previously reported values of 11.7 + 1.1 pmol/kg min-’ after overnight normalization (12) and third, the basal values of intermediary metabolites were all within the normal reference range of our laboratory. In vim sensitivity to insulin or proinsulin, relative changes in glucose turnover and intermediary metabolism were assessed by a modification
TABLE
1. Clinical
Characteristics
PROINSULIN
IN
IDDM
1283
of the euglycemic clamp technique (13) as previously described (5). Basal hepatic glucose production was determined for all subjects during each clamp. A primed infusion of 5 mg/kg [6’6’Hz]glucose was administered in logarithmically decreasing doses over 10 min with a constant infusion of 60 pg/kg.min [6’6’*Hz]glucose. A 3-h isotope equilibration period was used. During this time blood glucose was maintained at a constant level by a low dose iv infusion of insulin. During the last 30 min of this period (150-180 min) the mean coefficient of variation (cv) of blood glucose was 1.4 + 0.3% and APE 2.0 + 0.3%. Therefore under these conditions steady state isotope kinetics can be considered to exist. Consequently Steele’s equation for steady state conditions (14) modified for stable isotopes (15) was used. During the hormone infusions glucose rate or appearance (R,) and glucose utlization (Rd) were determined by continuing the infusion of [6’6’2H2]glucose and measuring glucose APE every 30 min until the last hour whereupon measurements were made every 15 min. Glucose turnover was calculated from the nonsteady state Steele equations in their modified form (15, 16). A pool correction factor of 0.65 and a volume of distribution of 20% were assumed. However as described elsewhere, it has now become apparent that this model is not fully quantitative as underestimates of glucose appearance and utilization can be obtained (17-18). Insulin was infused for 3 h and proinsulin for 5 h. These infusion periods were selected because it has been previously demonstrated that proinsulin takes 2 h longer than insulin to achieve steady state rates of glucose disposal (4). Insulin and proinsulin’s biological effects were, thus, quantified from the last hour of each respective infusion. Insulin infusion rates of 120,210, and 360 pmol/kg h-’ (I,, I,, and I3 respectively) were used to construct a dose response curve for insulin action in the physiological range. Three comparable, biologically equipotent doses of proinsulin were calculated as proposed by Bergenstal etal. (8) Proinsulin infusion rates of 260, 450, and 770 pmol/kg h-’ (I’,, PZ, and I’, respectively) were therefore used. At time zero each subject received in a randomized order, a primed continuous infusion of either insulin or proinsulin. The priming dose was given over the first 10 min in a logarithmic decreasing manner to acutely raise the hormone concentration to the desired level (19). The hormones were suspended in a plasma like colloid (Haemacel, Hoescht, W. Germany) to prevent any adsorption onto the plastic syringe and tubing. Arterialized blood was measured every 5 min by a glucose oxidase method analyzer (YSI model 23A, Yellow Springs Instruments, Clandon Scientific London, UK). Blood glucose was maintained at 0.3 mmol/L below the time zero level by infusing a variable amount of 10 or 20% glucose. Twenty millimoles per L KC1 were added to the glucose infusate during the glucose clamp to maintain blood potassium levels. The concentration of the glucose infusate was measured after each glucose clamp.
Analytical
methods
Plasma free insulin concentrations were measured, after polyethylene glycol extraction, by double-antibody RIA (20), interassay cv, 7.5% at 7.0 mIJ/L, and 6.8% at 44 mU/L. Proinsulin was measured, after removal of endogenous insulin antibodies by polythylene glycol extraction, by a specific two-site immunoradiometric assay using ‘251-labeled mouse monoclonal antirabbit immunoglobulin with an interassay cv of 7.0% at 1.0 nmol/L (21). Blood lactate, alanine, glycerol, and 3-hydroxybutyrate were determined by an automated fluorimetric assay as described by Lloyd et al. (22) NEFA were measured by an enzymatic
of Subjects
Patient
Age
BMI
(Yd
kg/M-'
1 2 3 4 5 6 Mean Normal
28 34 37 43 22 22 31 + 8
26 23 25 22 19 22 23 + 2
Duration of Diabetes (yr)
HBA, (%I
Usual daily insulin dose (U)
Insulin binding antibodies (pg/L)
10 12 5 24 8 9 11 f 6
9.2 a.2 10.1 8.6 10.0 9.3 9.6 + 0.6 (5-7.5%)
60 42 36 44 44 48 46 + 7
2.1 1.0 1.7 5.2 1.3 4.8 2.7 f 1.7 (2.0-10)
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
1284
DAVIS
calorimetric method (23). Glucagon concentrations an chromatographic immunoassay method (24) 7% at 25 q/L. Hormone samples were stored at metabolites at -40 C until assay. Samples for [6’6’2H2]glucose were derivatized and measured
ET
JCE & M. 1992 Vol 76. No 5
AL.
were determined by with an interassay cv of -70 C and intermediary isotopic enrichment of as previously described
U=I1 Lz3=P1
(6). Statistical
analyses
Data will be expressed as mean f SEM unless otherwise stated, and analyzed using standard two-way parametric analysis of variance with a repeated measures design and by Student’s t test for paired and unpaired data where appropriate. Blood glucose concentrations obtained during the glucose clamp studies were converted to plasma glucose values by the following formula: plasma glucose concentration = whole blood glucose concentration/l-O.3 hematocrit (25) and these results were used in the calculation for non steady state glucose turnover. Calculation of glucose turnover rates were performed with PDP-11 software (University of Newcastle upon Tyne).
p(O.02
p(O.01 7 m
=Pa
Results Euglycemic
clamp studies
Mean blood glucose values were similar during proinsulin and insulin infusions (4.4 * 0.1, 4.2 f 0.1 mmol/L, respectively). The cv of blood glucose was similar during proinsulin and insulin infusions (1.6 + 0.2, 1.7 + O.l%, respectively). Mean basal and steady state hormone levels obtained during the clamp studies are shown in Fig. 1. Basal plasma free
160
120
g 27
B z
4 I
12.0
A
k
I
,A .. I
9.0 I
i
.._ i . .. ..-.
I
I
I
T
1
tD3
oh----0
0
1 12 0
-:0
II
=I3
Eta-P,
240
300
TlME (MIN)
2
FIG. 2. Comparison of the effects of proinsulin and insulin on suppression of hepatic glucose production. Insulin was infused for 180 min, proinsulin was infused for 300 min. Calculated negative values of HGP have been included in the analysis as percentage suppression greater than 100%.
o.oY 0
60
120
180
240
300
TIME (MIN)
0.67
B $ 8’
z:
0.04 0
60
120
180
TIME (MN)
FIG. 1. Plasma free hormone levels. a) Proinsulin levels (nmol/L) 0 0 (Pi) M (P2) A - A (P:J measured during each of the proinsulin infusion. b) Insulin levels (nmol/L). 0 0 (Ii) I (I,) A - - - A (L) measured during each insulin infusion.
insulin concentrations were similar at the start of insulin and proinsulin infusions (0.1 + 0.02, 0.1 f 0.03 nmol/L respectively). Steady state levels were obtained between 60-120 min during the insulin infusions and 210 min for proinsulin (Fig. 1). Basal glucagon levels were similar at the start of the proinsulin and insulin infusions (24.0 + 3.3, 25.2 + 4.5 ng/ L, respectively). There was a trend for a greater reduction in glucagon levels during I1 US. I’, (-7.5 + 2, -4.4 f 2 rig/L) and I2 US. P2 (-9.8 f 3, -5 + 2.5 rig/L), respectively. During the highest infusion dose (13,P3) the reduction in plasma glucagon was equivalent (-7.1 + 2, -8.3 + 2.5 rig/L). Glucose turnover Basal values of HGP were similar at the start of each insulin and proinsulin infusion (12.6 +. 1.7, 13.3 + 1.5 pmol/ kg min-‘, respectively). The effects on suppression of HGP
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
THE
EFFECTS
OF HUMAN
by insulin and proinsulin are shown in Fig. 2. At the lowest hormone infusion (IJ’,), apart from the first hour, equivalent suppression of HGP was observed at all time points. HGP was not maximally suppressed in any individual at any incremental time point during the I1,P1 infusion. The two highest infusions (12, P2, 13, P3) did produce negative HGP values, in some subjects at some incremental time points. These negative values have been included in the analysis as percentage suppression values greater than 100%. During the 12,P2 infusions HGP was suppressed by a significantly greater amount during the first 2 h of the I2 infusion compared to P2 (Fig. 2). However, comparison of subsequent time points and the final hour of each respective infusion demonstrated equivalent affects 83.0 f 8.2%. Significant differences in percentage suppression of HGP occurred at the first to second hour incremental time points for both insulin (76 + 7 - 94 + 5; P = 0.03) and proinsulin (36.6 - 59 + 5 P = 0.03). There was equivalent suppression of HGP by IJ and P3 at each successive time point. Final hour HGP was suppressed Is: -2.0 + 1.7 pmol/kg min-’ (116 + 13%), PJ: -1.1 + 0.6 pmol/kg min-’ (109 + 5%). No significant differences occurred at successive incremental time points during the I3 infusion, but significant differences were observed during the P3 infusion, first to second 56 + 8 - 83 f 6; P < 0.01 and second to third h 84 + 6 - 102 + 9%; P < 0.02 (Fig. 2). The hormonal effect on glucose utilization (Rd) is presented as percentage increase of baseline. In this way an easier comparison with percentage suppression of HGP can be made. Rd was significantly increased at all time points during each insulin compared to proinsulin infusion (Fig. 3). The Rd values from the last hour of the clamp studies were as follows: Pi VS. I1 9.9 f 0.6, 15.4 f 1.7 pmol/kg.min-’ (P < O.Ol), Pz ‘us. I*; 15.1 + 2.7, 19.8 + 2.5 pmol/kg min (P < 0.05) and P3 VS. 13: 19.8 + 2.5, 31 * 1.1 ymol/kgamin-’ (p = 0.02 respectively). Dose response curves for the two hormones effects on suppression of HGP and elevation of Rd are shown in Fig. 4. Analysis of the dose response curves demonstrate a nonsignificant difference of regression line slopes. Thus an approximation of the potencies of the two hormones, can be obtained from these curves. Higher doses of proinsulin are needed to produce equivalent suppression of HGP compared to insulin. Comparison of different insulin and proinsulin concentrations on the suppression of HGP curve demonstrates that, on a molar basis, proinsulin had about 6% the biological effect of insulin. Comparison of different proinsulin and insulin concentrations on the increase of Rd curve demonstrated that, on a molar basis, proinsulin had approximately or equal to 3.7% the biological effect of insulin. Intermediary
metabolism
(Table 2)
Peripheral blood levels of lactate, alanine, glycerol, 3hydroxybutyrate, and NEFA were similar at the start of each infusion. There was a significant net decreaseof blood lactate levels during the P1 compared to I, (-130 + 34, -32 f 30 pmol/L; P < 0.05) and P2 compared to I2 infusion doses (-139 + 76, +8 + 65 pmol/L; P < 0.05 respectively). During
PROINSULIN
1285
IN IDDM p(O.01
p(O.02
p(0.01
insulin proinsulin
1"
1
insulin proinsulin
insulin proinsulin
p(O.01
p(O.03
infusion infusion 0
-I,
eo
‘P,
infusion infudon
18Omin JOOmin
)pco,ol
180min 3OOmin
) p(0,~5
infusion infurion
lBOmin) 300min
p(~,02
O-I3 ez
0 TIME
FIG. 3. Comparison centage stimulation
240
=P3
300
(MN)
of the effects of proinsulin and of R,+ Basal is depicted as 100%.
insulin
on per-
the highest dose blood lactate rose by a significantly greater amount during I3 compared to P3 [230 + 164, 48 + 60 pmol/ L (P < 0.05), respectively]. Blood alanine levels fell by similar, small amounts during Ii, P1, and 12,P2infusions. During the highest infusion dose (I,,P,) there were comparable, small nonsignificant increases. Blood glycerol, 3-hydroxybutyrate, and plasma NEFA levels were suppressedby a significantly greater percentage during the I1 compared to Pi infusions. During the two highest infusion dosesthese metabolites were suppressedby similar amounts.
Discussion
In the present study we have used the hyperinsulinemic euglycemic clamp technique to compare the effects of human proinsulin and insulin on intermediary carbohydrate and lipid metabolism in IDDM subjects.We have alsoconstructed dose response curves for proinsulin and insulin action on suppression of HGP and stimulation of peripheral glucose disposal. At the start of each hormone infusion a priming dose of insulin or proinsulin was administered concurrent with a
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
DAVIS
1286
I50[A a
(3 =
100
I
x4
Insulin
a+
Proinsulin
0.1
Concentration
300-B
s
x--9(
Insulin
-
Proinsulin
1.0
of Hormones
(nmol/L)
v
zoo-
s ‘Z B7 2
10.0 I
4 IOO-
8
OLi
0.1 Concentration
I
I.0
of
Hormones
10.0
(nmol/L)
FIG. 4. Dose response curves for suppression of hepatic glucose duction and percentage elevation of R,, during insulin (x-x) proinsulin (04) infusions.
proand
continuous infusion. This was done in an attempt to acutely raise the hormone to the desired level (13). As can be seen from Fig. 1 insulin infusions reached steady state levels between 60 and 120 min. The proinsulin infusions, needed longer to reach steady state (60-210 min). This is qualitively similar to previous work in normal (3, 5) and NIDDM man (6). This present finding in normoglycemic IDDM subjects further supports the suggestion made by Revers et al. (3) that proinsulin and insulin may have different volumes of distribution. Which, further emphasizes the need for extending the duration of glucose clamp studies during proinsulin infusions. TABLE
2. Mean
values
of intermediary
metabolites
during
hormone
I1
Lactate
Basal Steady state Basal Steady state Basal Steady state % Suppression Basal Steady state % Suppression Basal Steady state % Suppression
Alanine Glycerol
3-OH NEFA
’ Incremental
h (P
Vol. 15, No. 5 Printed in U.S.A
The Effects of Human Proinsulin on Glucose Turnover and Intermediary Metabolism in Insulin-DependentDiabetes Mellitus* S. N. DAVIS?, C. N. HALES,
M. ANSIFEROV, H. ORSKOV,
AND
C. HETHERINGTON, M. BROWN, K. G. M. M. ALBERT1
W. J. BRANCH,
Department of Medicine, The Medical School, University of Newcastle upon Tyne, Newcastle upon Tyne, England; Department of Clinical Biochemistry (W.J.B., C.N.H.), Addenbrookes Hospital, University of Cambridge, Cambridge, England; Second University Clinic of Internal Medicine (H.O.), Aarhus Kommunhospitalet, Denmark; and Department of Medicine, Vanderbilt University School of Medicine, Diabetes Research and Training Center, B-3307 Medical Center North, Nashville, Tennessee 37232-2230 ABSTRACT We have
comparedthe action of humanproinsulinand insulinon glucoseturnover, intermediarycarbohydrate,and lipid metabolismin insulin-dependent-diabetic (IDDM) subjects. Six, young, weightmatched(23 * 2 kg-‘) IDDM subjects underwent separate hyperinsulinemic
euglycemic
clamps.
Three,
low dose, iv infusions
of both
insulin
and proinsulin were usedto construct doseresponsecurves. The proinsulin
infusions
were
chosen
to give
steady
state
levels
approxi-
matelyor equalto 20.fold higher on a molar basis than insulin, based on previousfindingsthat proinsulinhasonly 5-10% the biological potency of insulin. Hepatic glucose production, measured using [6’6’“H,]glucose, was suppressed equally by proinsulin and insulin at the three dose levels: (I,) 2.8 t 0.7 (P,) 3.3 + 0.6. (I,)_ 2.3 + 0.9 (P.,)_, 3.3 & 1.1, (I:{) -2.0 f 1.7 (P:J -1.1 + 0.6 Fmol/kg min-’ Percentage elevation of glucose disposal was significantly increased during the insulininfusionscompared to proinsulin;(I,) 132 f 12 (P,) 78 + 4 p < 0.01; (I~)‘157 + 18 (P,) 104 + 14; P < 0.05; (1:J 242 + 23 (P:,) 159 + 24
P
ROINSULIN and insulin concentrations are significantly reduced in insulin-dependent-diabetic (IDDM) subjects. However, the effects of proinsulin deficiency remain unknown as its physiological role in normal man is unresolved. Proinsulin has been estimated to comprise 6% (1) of the total insulin-like immunoreactivity in the portal vein of normal man. Recent work, however, in both normal and noninsulindependent-diabetic man (NIDDM), using highly specific proinsulin assays,has shown that peripheral circulating levels are approximately or equal to 10% compared to insulin (2). This apparent reversal of the normal portal-peripheral gradient is due to the fact that proinsulin has a much lower metabolic clearance rate and longer half-life compared to insulin (3, 5). Previous in viva dose response studies in normal and NIDDM man have demonstrated that, on a molar basis, proinsulin has a much lower biological potency comReceived February 11, 1991. Address correspondence and requests for reprints to: S. N. Davis, Department of Medicine, Vanderbilt University School of Medicine, Diabetes Research and Training Center, B-3307 Medical Center North, Nashville, Tennessee 37232-2230. *Financial support was given by the British Diabetic Association, Mason Medical Research Foundation, and Eli Lilly Company. t Medical Research Council Training Fellow.
p = 0.02. Dose response curve analysis demonstrated that proinsulin stimulated glucose disposal approximately or equal to 3.7% whereas suppression of HGP was ~5.7% compared to insulin. Proinsulin had a significantly weaker effect than insulin, at the lowest infusion dose, in percent suppression of plasma nonesterified fatty acids (I1 34 + 4, P, 14 + 15%; P < 0.05), blood glycerol (I, 47 + 4, P, 30 ? 3%; P < 0.01) and 3-hydroxybutyrate levels (I, 81 + 7, P, 42 + 17%; P < 0.05). Proinsulin caused significant net reductions in blood lactate levels compared to insulin at each infusion dose; (PI) -130 + 34, (I,) -32 + 30 pmol/L (P < 0.05) (P2) -139 + 76 (12) +8 + 65 Fmol/L (P < 0.05) (P:,) 48 + 60 (I?) 230 + 64 rmol/L (P < 0.05). We conclude that in IDDM: 1) proinsulin has a preferential effect on the liver compared to muscle, in terms of glucose handling; 2) proinsulin may have a different effect on lactate metabolism compared to insulin; 3) proinsulin at the lowest dose resulted in an inability to suppress lipolysis and ketogenesis; 4) glucose turnover can be underestimated using [6’6”H2]glucose. (J Clin Endocrinol Metab 75: 12821288,1992)
pared to insulin (3, 4). The majority of studies have demonstrated that proinsulin has a relatively greater effect on suppressinghepatic glucose production (5-12%) (3-7) compared to stimulating peripheral glucose uptake (3-8%) (3-6). Proinsulin and insulin’s effects on intermediary carbohydrate and lipid metabolism do not appear to be identical. We have previously shown, during isoglycemic clamp studies, in both normal and NIDDM man that blood lactate concentrations can be significantly reduced by proinsulin infusions (5, 6). In addition, blood glycerol, 3-hydroxybutyrate and plasma nonesterified-fatty acids (NEFA) levels were higher after a low dose proinsulin infusion compared to a biologically equipotent dose of insulin in NIDDM subjects. There is, however, scant data concerning proinsulin’s metabolic affects in IDDM subjects. Bergenstal et al. (8) investigated the dose of proinsulin needed to maintain normoglycemia after acute insulin withdrawal in IDDM subjects. Whereas Cohen et al. (9) studied the comparative effects of insulin and proinsulin on the hepatic response to glucagon infusion. This latter study did not find a relatively hepatospecific action of proinsulin
after
a prolbngeh
(>iS
h)
infusion
&ed
to
render
subjects normoglycemic. We were, therefore, particularly interested in examining the acute dose response effects of 1282
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
THE
EFFECTS
OF HUMAN
proinsulin on suppression of hepatic glucose production and glucose use in IDDM. In addition we have compared the effects of three biologically equipotent low dose iv infusions of human proinsulin and insulin on different aspects of intermediary lipid and carbohydrate metabolism, known to be abnormal in IDDM subjects. This will enable a more complete profile of proinsulin action in IDDM to be obtained. Materials
and Methods
Subjects We studied six, male, IDDM subjects. Mean age 31 f 8 yr (Table 1) all of whom had documented episodes of diabetic ketoacidosis. Each subject had normal blood count, plasma electrolytes, and liver function. None had detectable micro- or macrovascular complications of diabetes. All subjects gave written informed consent. Subjects were asked to take their usual weight maintaining diet for 3 days before each study. The amount of exercise before studies was controlled, so that no subject performed any strenuous exercise 72 h before a study and arrived via organized transport. All studies were performed in the Clinical Research Centre and received local Ethical Committee approval.
Materials Biosynthetic human proinsulin was prepared and characterized as previously described (lo), and kindly supplied by Eli Lilly Co. (Indianapolis, IN), as was regular human insulin. However due to adverse findings in a prospective proinsulin clinical trial performed elsewhere, Eli Lilly Co. acutely withdrew proinsulin before all the planned studies were completed. Therefore, two high dose and one low dose proinsulin infusion protocol, to be described below, could not be completed. 6’6Dideuterated glucose (99.3% mol/excess) was purchased from K. K. Grief (London, UK).
Glucose clamp studies Approximately 16 h before arriving at the Clinical Research Centre patients reduced their usual doses of regular and intermediate-acting insulins by approximately or equal to 70%. At 0800 h, after a 10-h overnight fast, a low dose iv infusion of regular insulin was started to render subjects normoglycemic. This infusion was continued until blood glucose had been stable for 3 h. We have three pieces of information which demonstrate that this method provided a similar metabolic profile to overnight glucose normalization: first, the basal free insulin values in this study (0.1 + 0.02 nmol/L) were similar to previous values of 0.08 f 0.1 nmol/L reported by Hother-Nielsen et al. (11). Second, mean basal hepatic glucose production (HGI’) values in the present study (12.8 + 1.1 pmol/kg min-’ were similar to previously reported values of 11.7 + 1.1 pmol/kg min-’ after overnight normalization (12) and third, the basal values of intermediary metabolites were all within the normal reference range of our laboratory. In vim sensitivity to insulin or proinsulin, relative changes in glucose turnover and intermediary metabolism were assessed by a modification
TABLE
1. Clinical
Characteristics
PROINSULIN
IN
IDDM
1283
of the euglycemic clamp technique (13) as previously described (5). Basal hepatic glucose production was determined for all subjects during each clamp. A primed infusion of 5 mg/kg [6’6’Hz]glucose was administered in logarithmically decreasing doses over 10 min with a constant infusion of 60 pg/kg.min [6’6’*Hz]glucose. A 3-h isotope equilibration period was used. During this time blood glucose was maintained at a constant level by a low dose iv infusion of insulin. During the last 30 min of this period (150-180 min) the mean coefficient of variation (cv) of blood glucose was 1.4 + 0.3% and APE 2.0 + 0.3%. Therefore under these conditions steady state isotope kinetics can be considered to exist. Consequently Steele’s equation for steady state conditions (14) modified for stable isotopes (15) was used. During the hormone infusions glucose rate or appearance (R,) and glucose utlization (Rd) were determined by continuing the infusion of [6’6’2H2]glucose and measuring glucose APE every 30 min until the last hour whereupon measurements were made every 15 min. Glucose turnover was calculated from the nonsteady state Steele equations in their modified form (15, 16). A pool correction factor of 0.65 and a volume of distribution of 20% were assumed. However as described elsewhere, it has now become apparent that this model is not fully quantitative as underestimates of glucose appearance and utilization can be obtained (17-18). Insulin was infused for 3 h and proinsulin for 5 h. These infusion periods were selected because it has been previously demonstrated that proinsulin takes 2 h longer than insulin to achieve steady state rates of glucose disposal (4). Insulin and proinsulin’s biological effects were, thus, quantified from the last hour of each respective infusion. Insulin infusion rates of 120,210, and 360 pmol/kg h-’ (I,, I,, and I3 respectively) were used to construct a dose response curve for insulin action in the physiological range. Three comparable, biologically equipotent doses of proinsulin were calculated as proposed by Bergenstal etal. (8) Proinsulin infusion rates of 260, 450, and 770 pmol/kg h-’ (I’,, PZ, and I’, respectively) were therefore used. At time zero each subject received in a randomized order, a primed continuous infusion of either insulin or proinsulin. The priming dose was given over the first 10 min in a logarithmic decreasing manner to acutely raise the hormone concentration to the desired level (19). The hormones were suspended in a plasma like colloid (Haemacel, Hoescht, W. Germany) to prevent any adsorption onto the plastic syringe and tubing. Arterialized blood was measured every 5 min by a glucose oxidase method analyzer (YSI model 23A, Yellow Springs Instruments, Clandon Scientific London, UK). Blood glucose was maintained at 0.3 mmol/L below the time zero level by infusing a variable amount of 10 or 20% glucose. Twenty millimoles per L KC1 were added to the glucose infusate during the glucose clamp to maintain blood potassium levels. The concentration of the glucose infusate was measured after each glucose clamp.
Analytical
methods
Plasma free insulin concentrations were measured, after polyethylene glycol extraction, by double-antibody RIA (20), interassay cv, 7.5% at 7.0 mIJ/L, and 6.8% at 44 mU/L. Proinsulin was measured, after removal of endogenous insulin antibodies by polythylene glycol extraction, by a specific two-site immunoradiometric assay using ‘251-labeled mouse monoclonal antirabbit immunoglobulin with an interassay cv of 7.0% at 1.0 nmol/L (21). Blood lactate, alanine, glycerol, and 3-hydroxybutyrate were determined by an automated fluorimetric assay as described by Lloyd et al. (22) NEFA were measured by an enzymatic
of Subjects
Patient
Age
BMI
(Yd
kg/M-'
1 2 3 4 5 6 Mean Normal
28 34 37 43 22 22 31 + 8
26 23 25 22 19 22 23 + 2
Duration of Diabetes (yr)
HBA, (%I
Usual daily insulin dose (U)
Insulin binding antibodies (pg/L)
10 12 5 24 8 9 11 f 6
9.2 a.2 10.1 8.6 10.0 9.3 9.6 + 0.6 (5-7.5%)
60 42 36 44 44 48 46 + 7
2.1 1.0 1.7 5.2 1.3 4.8 2.7 f 1.7 (2.0-10)
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
1284
DAVIS
calorimetric method (23). Glucagon concentrations an chromatographic immunoassay method (24) 7% at 25 q/L. Hormone samples were stored at metabolites at -40 C until assay. Samples for [6’6’2H2]glucose were derivatized and measured
ET
JCE & M. 1992 Vol 76. No 5
AL.
were determined by with an interassay cv of -70 C and intermediary isotopic enrichment of as previously described
U=I1 Lz3=P1
(6). Statistical
analyses
Data will be expressed as mean f SEM unless otherwise stated, and analyzed using standard two-way parametric analysis of variance with a repeated measures design and by Student’s t test for paired and unpaired data where appropriate. Blood glucose concentrations obtained during the glucose clamp studies were converted to plasma glucose values by the following formula: plasma glucose concentration = whole blood glucose concentration/l-O.3 hematocrit (25) and these results were used in the calculation for non steady state glucose turnover. Calculation of glucose turnover rates were performed with PDP-11 software (University of Newcastle upon Tyne).
p(O.02
p(O.01 7 m
=Pa
Results Euglycemic
clamp studies
Mean blood glucose values were similar during proinsulin and insulin infusions (4.4 * 0.1, 4.2 f 0.1 mmol/L, respectively). The cv of blood glucose was similar during proinsulin and insulin infusions (1.6 + 0.2, 1.7 + O.l%, respectively). Mean basal and steady state hormone levels obtained during the clamp studies are shown in Fig. 1. Basal plasma free
160
120
g 27
B z
4 I
12.0
A
k
I
,A .. I
9.0 I
i
.._ i . .. ..-.
I
I
I
T
1
tD3
oh----0
0
1 12 0
-:0
II
=I3
Eta-P,
240
300
TlME (MIN)
2
FIG. 2. Comparison of the effects of proinsulin and insulin on suppression of hepatic glucose production. Insulin was infused for 180 min, proinsulin was infused for 300 min. Calculated negative values of HGP have been included in the analysis as percentage suppression greater than 100%.
o.oY 0
60
120
180
240
300
TIME (MIN)
0.67
B $ 8’
z:
0.04 0
60
120
180
TIME (MN)
FIG. 1. Plasma free hormone levels. a) Proinsulin levels (nmol/L) 0 0 (Pi) M (P2) A - A (P:J measured during each of the proinsulin infusion. b) Insulin levels (nmol/L). 0 0 (Ii) I (I,) A - - - A (L) measured during each insulin infusion.
insulin concentrations were similar at the start of insulin and proinsulin infusions (0.1 + 0.02, 0.1 f 0.03 nmol/L respectively). Steady state levels were obtained between 60-120 min during the insulin infusions and 210 min for proinsulin (Fig. 1). Basal glucagon levels were similar at the start of the proinsulin and insulin infusions (24.0 + 3.3, 25.2 + 4.5 ng/ L, respectively). There was a trend for a greater reduction in glucagon levels during I1 US. I’, (-7.5 + 2, -4.4 f 2 rig/L) and I2 US. P2 (-9.8 f 3, -5 + 2.5 rig/L), respectively. During the highest infusion dose (13,P3) the reduction in plasma glucagon was equivalent (-7.1 + 2, -8.3 + 2.5 rig/L). Glucose turnover Basal values of HGP were similar at the start of each insulin and proinsulin infusion (12.6 +. 1.7, 13.3 + 1.5 pmol/ kg min-‘, respectively). The effects on suppression of HGP
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
THE
EFFECTS
OF HUMAN
by insulin and proinsulin are shown in Fig. 2. At the lowest hormone infusion (IJ’,), apart from the first hour, equivalent suppression of HGP was observed at all time points. HGP was not maximally suppressed in any individual at any incremental time point during the I1,P1 infusion. The two highest infusions (12, P2, 13, P3) did produce negative HGP values, in some subjects at some incremental time points. These negative values have been included in the analysis as percentage suppression values greater than 100%. During the 12,P2 infusions HGP was suppressed by a significantly greater amount during the first 2 h of the I2 infusion compared to P2 (Fig. 2). However, comparison of subsequent time points and the final hour of each respective infusion demonstrated equivalent affects 83.0 f 8.2%. Significant differences in percentage suppression of HGP occurred at the first to second hour incremental time points for both insulin (76 + 7 - 94 + 5; P = 0.03) and proinsulin (36.6 - 59 + 5 P = 0.03). There was equivalent suppression of HGP by IJ and P3 at each successive time point. Final hour HGP was suppressed Is: -2.0 + 1.7 pmol/kg min-’ (116 + 13%), PJ: -1.1 + 0.6 pmol/kg min-’ (109 + 5%). No significant differences occurred at successive incremental time points during the I3 infusion, but significant differences were observed during the P3 infusion, first to second 56 + 8 - 83 f 6; P < 0.01 and second to third h 84 + 6 - 102 + 9%; P < 0.02 (Fig. 2). The hormonal effect on glucose utilization (Rd) is presented as percentage increase of baseline. In this way an easier comparison with percentage suppression of HGP can be made. Rd was significantly increased at all time points during each insulin compared to proinsulin infusion (Fig. 3). The Rd values from the last hour of the clamp studies were as follows: Pi VS. I1 9.9 f 0.6, 15.4 f 1.7 pmol/kg.min-’ (P < O.Ol), Pz ‘us. I*; 15.1 + 2.7, 19.8 + 2.5 pmol/kg min (P < 0.05) and P3 VS. 13: 19.8 + 2.5, 31 * 1.1 ymol/kgamin-’ (p = 0.02 respectively). Dose response curves for the two hormones effects on suppression of HGP and elevation of Rd are shown in Fig. 4. Analysis of the dose response curves demonstrate a nonsignificant difference of regression line slopes. Thus an approximation of the potencies of the two hormones, can be obtained from these curves. Higher doses of proinsulin are needed to produce equivalent suppression of HGP compared to insulin. Comparison of different insulin and proinsulin concentrations on the suppression of HGP curve demonstrates that, on a molar basis, proinsulin had about 6% the biological effect of insulin. Comparison of different proinsulin and insulin concentrations on the increase of Rd curve demonstrated that, on a molar basis, proinsulin had approximately or equal to 3.7% the biological effect of insulin. Intermediary
metabolism
(Table 2)
Peripheral blood levels of lactate, alanine, glycerol, 3hydroxybutyrate, and NEFA were similar at the start of each infusion. There was a significant net decreaseof blood lactate levels during the P1 compared to I, (-130 + 34, -32 f 30 pmol/L; P < 0.05) and P2 compared to I2 infusion doses (-139 + 76, +8 + 65 pmol/L; P < 0.05 respectively). During
PROINSULIN
1285
IN IDDM p(O.01
p(O.02
p(0.01
insulin proinsulin
1"
1
insulin proinsulin
insulin proinsulin
p(O.01
p(O.03
infusion infusion 0
-I,
eo
‘P,
infusion infudon
18Omin JOOmin
)pco,ol
180min 3OOmin
) p(0,~5
infusion infurion
lBOmin) 300min
p(~,02
O-I3 ez
0 TIME
FIG. 3. Comparison centage stimulation
240
=P3
300
(MN)
of the effects of proinsulin and of R,+ Basal is depicted as 100%.
insulin
on per-
the highest dose blood lactate rose by a significantly greater amount during I3 compared to P3 [230 + 164, 48 + 60 pmol/ L (P < 0.05), respectively]. Blood alanine levels fell by similar, small amounts during Ii, P1, and 12,P2infusions. During the highest infusion dose (I,,P,) there were comparable, small nonsignificant increases. Blood glycerol, 3-hydroxybutyrate, and plasma NEFA levels were suppressedby a significantly greater percentage during the I1 compared to Pi infusions. During the two highest infusion dosesthese metabolites were suppressedby similar amounts.
Discussion
In the present study we have used the hyperinsulinemic euglycemic clamp technique to compare the effects of human proinsulin and insulin on intermediary carbohydrate and lipid metabolism in IDDM subjects.We have alsoconstructed dose response curves for proinsulin and insulin action on suppression of HGP and stimulation of peripheral glucose disposal. At the start of each hormone infusion a priming dose of insulin or proinsulin was administered concurrent with a
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
DAVIS
1286
I50[A a
(3 =
100
I
x4
Insulin
a+
Proinsulin
0.1
Concentration
300-B
s
x--9(
Insulin
-
Proinsulin
1.0
of Hormones
(nmol/L)
v
zoo-
s ‘Z B7 2
10.0 I
4 IOO-
8
OLi
0.1 Concentration
I
I.0
of
Hormones
10.0
(nmol/L)
FIG. 4. Dose response curves for suppression of hepatic glucose duction and percentage elevation of R,, during insulin (x-x) proinsulin (04) infusions.
proand
continuous infusion. This was done in an attempt to acutely raise the hormone to the desired level (13). As can be seen from Fig. 1 insulin infusions reached steady state levels between 60 and 120 min. The proinsulin infusions, needed longer to reach steady state (60-210 min). This is qualitively similar to previous work in normal (3, 5) and NIDDM man (6). This present finding in normoglycemic IDDM subjects further supports the suggestion made by Revers et al. (3) that proinsulin and insulin may have different volumes of distribution. Which, further emphasizes the need for extending the duration of glucose clamp studies during proinsulin infusions. TABLE
2. Mean
values
of intermediary
metabolites
during
hormone
I1
Lactate
Basal Steady state Basal Steady state Basal Steady state % Suppression Basal Steady state % Suppression Basal Steady state % Suppression
Alanine Glycerol
3-OH NEFA
’ Incremental
h (P
