- Email: [email protected]
Insulin antibody does not cause insulin resistance during glucose clamping in rats
Dkzberes Reseearch and C/in&z1 Practice, 18 (1992)
143- 15 1 0 1992 Elsevier Science Publishers B.V. All rights reserved 0168-8227/92/$05.00
143
DIABET 00689
Insulin antibody does not cause insulin resistance during glucose clamping in rats M. Tominaga,
M. Matsumoto,
The Third Department of Internal Medicine,
M. Igarashi, H. Eguchi, A. Sekikawa
and H. Sasaki
Yamagata Universit.v School of Medicine, 2-2 Iida-Nishi 2-Chome,
Yamagata Cit.v,
Yamagata 990-23, Japan
(Received 16 March 1992) (Revision accepted 9 July 1992)
Summary
Although it has often been stated that insulin antibodies cause insulin resistance, this concept is still controversial. The effect of insulin antibody GP30, commonly used in insulin radioimmunoassay, on insulin action was investigated in Wistar rats in vivo by the euglycemic glucose clamp technique. As a preliminary experiment, the equilibrium time required for insulin antibody to bind with endogenous insulin was examined. One hundred pi/kg insulin antibody took 60 min or more to attain equilibrium, but 10 pi/kg insulin antibody almost immediately equilibrated with endogenous insulin. During a 60-min glucose clamp study, 2 mU/kg/min porcine insulin was infused with 100 pi/kg insulin antibody. At steady state, during the last 20-min period, the mean glucose infusion rate was 2.10 f 0.85 mg/kg/min (n = 5, mean ? SD), significantly lower than the 5.77 k 1.61 mg/kg/min of the control, indicating insulin resistance before equilibrium was reached. However, the glucose infusion rates during the clamp with 10 &kg insulin antibody and 100 pi/kg insulin antibody infused 75 min before the insulin were 6.10 k 1.44 and 7.12 + 1.19 mg/kg/min, respectively, no different from the control. In these instances, free insulin levels measured by radioimmunoassay using the polyethyleneglycol method were 43.8 + 20.4 and 15.4 t 6.1 pU/ ml, respectively, lower than the control (77.0 2 16.1 pU/ml). Furthermore, by increasing the insulin infusion rate to 5 mU/kg/min, the glucose infusion rate during the clamp with 100 pi/kg insulin antibody concomitantly infused was 11.79 & 1.14 mg/kg/min, no different from the 14.50 f 2.82 mg/kg/min of the control, while free insulin levels were 62.5 k 29.4 ,uU/ml, still lower than the control (152.7 + 69.5 pU/ ml). In conclusion, this in vivo study suggested that even blocking antibodies against insulin did not cause insulin resistance once binding reached equilibrium. Key words: Insulin antibody;
Insulin resistance;
Glucose clamp
Correspondence to: M. Tominaga, The Third Department of Internal Medicine. Yamagata University Iida-Nishi 2-Chome, Yamagata City, Yamagata 990-23, Japan.
School of Medicine, 2-2
144
Introduction In the past, patients with diabetes mellitus receiving impure insulin preparations were sometimes reported to have very high titers of insulin antibodies and to need tremendously high daily requirements of insulin [l-4], so it has commonly been believed that insulin antibodies induce insulin resistance. However, Armitage et al. recently reported that the titers of insulin antibodies measured by their own method did not correlate with the daily requirement of insulin [ 51. From this investigation, they suspected that insulin antibody was not the causal factor of insulin resistance, and that insulin antibody behaved as a reservoir after equilibrium with insulin. On the other hand, the euglycemic glucose clamp technique originally described by Andres et al. [6] has given us the most precise method for estimating insulin action in vivo. In the present study, insulin action was investigated in the presence of insulin antibody by this technique in rats in order to elucidate the contribution of insulin antibody to insulin resistance before and after equilibrium with insulin.
Materials and Methods Equilibrium time of insulin antibody GP30 with endogenous insulin Under 0.35 g/kg i.p. chloral hydrate anesthesia, Silastic catheters were cannulated into the left carotid artery and the left femoral vein of male Wistar rats weighing 265 k 19 g (mean k SD). Ten and 100 $/kg insulin antibody GP30 (BioMaker), commonly used in insulin radioimmunoassay [7], were injected through the femoral vein catheter. At three random times from 0, 7, 15, 30, 60, and 120 min after injection, 1 ml blood samples were obtained from the carotid artery catheter in order to minimize the total blood volume sampled to less than 3 ml. Then, free and total insulin concentrations were measured by radioimmunoassay to ascertain the equilibrium time required for insulin antibody GP30 to bind with the endogenous insulin of rats.
Glucose clamp technique Male Wistar rats were also used for glucose clamp after 24-h fasting. Under the same anesthesia using chloral hydrate, two Silastic catheters were cannulated into the left femoral vein for 3H-3glucose and 20% glucose infusion, and another catheter into the right femoral vein for the infusion of porcine insulin (Actrapid MC insulin, Novo) diluted with saline containing 0.25% bovine albumin (Sigma). A doubled lumen catheter was inserted into the right jugular vein for sampling of the heparinized blood to measure blood glucose levels by Glucose Monitor (Kyoto Daiichi Kagaku). Another Silastic catheter was cannulated into the left carotid artery for blood sampling to determine the specific activity of 3H-3glucose as well as free and total insulin levels and insulin-binding capacity. Initially, 5 PCi 3H-3-glucose was infused as a bolus followed by continuous infusion at a rate of 0.05 &i/min using a peristaltic rolling pump. Sixty minutes later, after blood sampling for determination of the specific activity of 3H-3glucose, porcine insulin solution was continuously infused at 2 and 5 mU/kg/min using a syringe pump after a lo-n-tin priming infusion. In one protocol, 100 pi/kg insulin antibody GP30 was infused 75 min before the insulin, but in others, 10 and 100 pi/kg insulin antibody was concomitantly infused at the beginning of insulin infusion. Changes in blood glucose levels were maintained within 10% of the basal levels for 60 min by variable infusion rates of 20% glucose which were calculated from the levels of, and changes in, glucose concentration measured by Glucose Monitor every 2 min. A modified algorithm [ 81 to determine the glucose infusion rate (GIR), originally described by DeFronzo et al. [ 91, was employed. The calculation done using a programmable calculator is as follows: GIR = 0.2 x (GD - BG) + GC GC = CC _ , - [BG-GD x 4 (BG-BG _ ,)I/64 where GD is the target blood glucose level, BG the blood glucose level, and GC a constant determined during the clamp.
145
Blood samples were taken at the end of the glucose clamp to measure free insulin levels and specific activity of 3H-3-glucose. Glucose disappearance rate (Gd) and hepatic glucose output (HGO) were determined by Steele’s method [ lo]. These are the parameters for insulin action along with the mean GIR during the last 20 min clamp. Determination offree and total insulin Free and total insulin levels were determined by radioimmunoassay according to Nakagawa et al. [ 1 I]. Serum samples were immediately mixed with ice-chilled polyethylene glycol and centrifuged. The supernatant was kept at -20 “C until assay. Other serum samples for assay of total insulin and insulin-binding capacity were stored at -20 “C until assay. Free insulin levels in the supernatant were determined by radioimmunoassay. Total insulin levels of serum samples after acidification with HCl were also measured by radioimmunoassay. Insulin-binding capacities of the same samples with ‘?-insulin were estimated by the polyethyleneglycol method.
Statistical analysis Statistical significance was assessed by Student’s t-test.
Results Equilibrium time of insulin antibody GP30 with endogenous insulin As shown in Fig. 1, total insulin levels increased gradually in rats injected with 100 PI/kg insulin antibody GP30 and reached a plateau 60 min after injection. However, a small peak was seen in the 10 $/kg insulin antibody injected group 7 to 15 min after administration (Fig. 1). Free insulin levels were suppressed by both doses of insulin antibody during the entire experimental Control
Insulin Antibody
6.
-2
P _,
%=
4
E” 2-
F-RKlOOuVkg)
o0
30 Time(minl
0715
30
60
120
Timdmin)
Fig. 1. Free and total insulin levels after 10 and 100 $/kg insulin antibody GP30. After 10 and 100 PI/kg insulin antibody GP30 were injected into rats, free and total insulin levels were measured by radioimmunoassay using the polyethyleneglycol method of Nakagawa. Total insulin levels in the 100 PI/kg insulin antibody group (open circle) increased gradually and reached a plateau at 60 min after injection. Total insulin levels in the 10 pi/kg injected group (closed circle) reached a small peak 7 to 15 min after administration. Free insulin levels in both 10 and 100 pi/kg injected rats (open and closed squares, respectively) were almost suppressed.
60
0
30
60
Time(mm)
Fig. 2. Rat euglycemic glucose clamp: control experiment of 2 mU/kg/min insulin infusion and 100 pi/kg insulin antibody concomitantly infused. Left: euglycemic glucose clamp with 2 mU/kg/min porcine insulin infusion as a control. Blood glucose level (BG) was maintained within log,, range of the fasting glucose level by variable glucose infusion rates (GIR) shown in the lower columns. Steady-state GIR, estimated by the mean of GIR during the last 20 min clamp, represents the insulin action. Steady-state GIR of this group was 5.77 f 1.61 mg/kg/min. Right: euglycemic glucose clamp, 100 &kg insulin antibody infused at the beginning of 2 mU/ kg/min porcine insulin infusion. Steady-state GIR was 2.20 f 0.85 mg/kg/min, significantly lower than the control. In this figure, data for BG and GIR are shown as mean k SE.
146
period. From these experiments, it appears that 100 PI/kg insulin antibody took 60 min or more to reach equilibrium, but 10 PI/kg insulin antibody almost immediately equilibrated with endogenous insulin. Insulin resistance before binding of insulin antibody> with insulin reached equilibrium Steady-state GIR of the glucose clamp with 2 mU/kg/min insulin infusion was 5.77 + 1.61 mg/kg/ min in the control, as shown on the left of Fig. 2. GIR of the clamp with 100 ,nl/kg insulin antibody (Fig. 2, right) was 2.10 t 0.85 mg/kg/min, significantly lower than the control, indicating insulin resistance. However, it can be seen that GIR gradually increased towards the end-point of the clamp suggesting that the binding of the insulin antibody with the infused porcine insulin was reaching equilibrium during this 60-min clamp, since the binding of
TABLE Data
100 pi/kg insulin antibody to reach equilibrium is a time-consuming process. G, and HGO, other parameters for insulin action, are shown in Table 1. Like GIR, G, during the clamp with concomitant infusion of 100 pi/kg insulin antibody was significantly lower than the control. HGO during the clamp with 100 pi/kg insulin antibody was not suppressed, although HGO was almost completely suppressed by 2 mU/kg/min insulin infusion in the control rats. These results suggested that insulin resistance caused by insulin antibody was present in both peripheral tissue and the liver before the binding of insulin antibody with insulin reached equilibrium. No insulin resistance after binding of insulin antibody with insulin reached equilibrium As shown on the left of Fig. 3, in the clamp with 10 pi/kg insulin antibody, GIR was
1
of rat glucose
clamp Control
Insulin
antibody”
Insulin
infusion
Insulin
antibody”
(100 PI/kg)
(10 PI/kg)
Insulin
antibodyb
(100 PI/kg)
rate 2 mU/kg/min
Number
of experiments
Glucose
infusion
01)
5
rate (mg/kg/min)
5
5
5
5.77 f 1.61
6.10& 1.44
2.10 + 0.85*
7.12& 1.19
Glucose
disappearance
rate (mg/kg/min)
6.66 & 0.89
7.52 * 0.80
4.87 + 1.27%
7.00 + 1.04
Hepatic
glucose
(mg/kg/min)
0.89 & 1.97
1.42i
2.77 f 1.85
0.11
77.0 + 16.1
43.8 * 20.4*
21.3 + 3.0*
15.4 + 6.1*
Free insulin Total
insulin
Insulin
output
level (p U/ml) level (p U/ml)
binding
capacity
130.9 + 24.2
(“,)
9.9 * 3.3
Insulin infusion rate 5 mLJ/kg/min Number of experiments (n) Glucose
infusion
Glucose
disappearance
rate (mg/kg/min)
Hepatic
glucose
(mg/kg/min)
Free insulin Total Insulin Glucose
insulin
5
rate (mg/kg/min) output
level (p U/ml) capacity
disappearance
5
10.56 k 2.53
9.73 * 0.52
8.38 + 1.13
- 3.94 & 2.65
- 5.42 & 2.55
- 3.16+_ 1.01
74.4 * 27.7* 210.9 k 86.8 5.1 f- 1.9
glucose
33.4 f 7.8
5
(“,)
rate and hepatic
254.0 k 62.3
41.7 + 5.6
15.15 & 2.76
152.7 + 69.5
output
during
steady
state of glucose
f 0.34
701.0 * 123.1
14.5 + 2.82
level @U/ml)
binding
1.31
11.79-t
1.14
62.1 + 29.4* 517.0 2 157.3 20.9 & 8.4 clamp
were estimated
by Steele’s method
using ‘H-3-glucose. Free and total insulin levels were measured by radioimmunoassay using polyethyleneglycol. a Insulin antibody administered concomitantly with insulin infusion. b Insulin antibody administered 75 min before insulin infusion. Data are shown as mean + SD.
*p
147 lnslin
haulin
Antibody lOO~Vkg(75min w
Antibody
huh
prior)
Anttbody
100Wkg
15 -
e
.E
5g
lo-
E 5-
0
30 Timetmin)
60
0
30
60
Timahin)
Fig. 3. Rat euglycemic glucose clamp: 10 PI/kg insulin antibody concomitantly infused group and 75 min prior 100 PI/kg insulin antibody infused group both with 2 mU/kg/min insulin infusion. Left: euglycemic glucose clamp of 2 mU/kg/min insulin infusion with 10 pi/kg insulin antibody concomitantly infused. GIR was 6.10 k 1.44 mg/kg/min, not different from that of the control. Right: euglycemic glucose clamp of 100 nl/ kg insulin antibody 75 min before the beginning of 2 mu/kg/ min insulin infusion. GIR was 7.12 f 1.19 mg/kg/min, not different from the control. Data are shown as mean t SE.
6.10 & 1.44 mg/kg/min, no different from the control, indicating no insulin resistance even with insulin antibody. On the right of Fig. 3 is shown the clamp with 100 pi/kg insulin antibody, in which insulin antibody was infused 75 min before the infusion of 2 mU/kg/min porcine insulin. GIR was 7.12 + 1.19 mg/kg/min, also no different from the control. Furthermore, since with the more insulin infused, the faster equilibrium is reached, a glucose clamp with 100 /A/kg insulin antibody concomitantly infused with increased doses of insulin to 5 mU/kg/min was carried out. GIR was 11.79 k 1.14 mg/kg/min, as shown on the right of Fig. 4, no different from the control (14.50 k 2.82 mg/kg/min) as shown on the left of Fig. 4. GTR of 10 &kg insulin antibody with 5 mU/kg/min insulin infusion was also no different from the control (not shown in figure, see Table 1). G,s during the clamp of 2 mU/kg/min insulin
n. v
0
30 Time(min)
60
0
30
60
Time(min)
Fig. 4. Rat euglycemic glucose clamp: control experiment of 5 mU/kg/min insulin infusion and 100 pi/kg insulin antibody concomitantly infused. Left: euglycemic glucose clamp of 5 mU/kg/min insulin infusion as a control. GIR was 14.50 k 2.82 mg/kg/min. Right: euglycemic glucose clamp of 5 mU/kg/min insulin infusion concomitantly infused with 100 PI/kg insulin antibody. GIR was 11.79 & 1.14 mg/kg/min, not different from the control. Data are shown as mean f SE.
infusion with 10 pi/kg insulin antibody and 100 pi/kg insulin antibody infused 75 min before insulin infusion were not different from the control, as shown in Table 1. Upon increasing the insulin infusion rate to 5 mU/kg/min, there was no difference from the control in either G, or HGO during the clamp with 100 pi/kg insulin antibody. Free insulin levels during the clamp of 2 mU/ kg/min insulin infusion with 10 $/kg insulin antibody were 43.8 2 20.4 pU/ml, significantly lower when compared to 77.0 + 16.1 pU/ml during the control clamp as shown in Table 1, although GIR, G, and HGO during the two clamps did not differ. Furthermore, during the clamp of 5 mu/kg/ min insulin infusion with 10 and 100 $/kg insulin antibody, free insulin levels were 74.4 rf-27.7 and 62.1 k 29.4 pU/ml, respectively, significantly lower than the 152.7 & 69.5 pU/ml of the control. Total insulin levels and insulin binding capacities found in these glucose clamp studies are also shown in Table 1.
148
20
15
10
5
0
200
100 Free Insulin
0
Level
1
1
2
5
Insulin
(pU/ml)
Infusion
Rate
(mU/kg/min)
Fig. 5. Relationship of free insulin levels and insulin infusion rates to glucose infusion rates during glucose clamp. Left: relationship between free insulin levels and glucose infusion rates. Closed circle indicates control; closed rectangle indicates glucose clamp with 10 PI/kg insulin antibody, and open rectangle indicates glucose clamp with 100 pi/kg insulin antibody which was administered 75 min before the beginning of 2 mU/kg/min insulin infusion and given concomitantly with 5 mU/kg/min insulin infusion. Despite the significantly lower free insulin levels of the antibody infused groups, the glucose infusion rates were almost the same as those of the control. Right: relationship between insulin infusion rates and glucose infusion rates during clamps. It was noted that the dose-response obtained by 10 and 100 pi/kg insulin antibody almost overlapped that of the control. Data are shown as mean k SE.
On the left of Fig. 5, the relationship between GIR representing insulin action during the clamp with insulin antibody, and free insulin levels is shown. It can be seen that free insulin levels were significantly lower than the control, while GIR is comparable to the control clamp at the same insulin rate. In other words, free insulin levels did not seem to be responsible for insulin action, but insulin infusion rates seemed to have more effect. This relationship is more clearly demonstrated on the right of Fig. 5 in which the dose-response obtained from the clamps with 10 and 100 pi/kg insulin antibody was almost overlapped by that of the control clamps.
Discussion It is well known that improvement in insulin preparations has been achieved to avoid several clinical problems caused by ‘dirty’ preparation, such as lipodystrophy, lipoatrophy and insulin allergy. After the development of highly purified or monocomponent insulin and recombinant human insulin preparations, which have less antigenicity and are now widely available for clinical practice, the problem of insulin resistance due to high titers of insulin antibody [l-4] seems to have faded away. However, it is also known that insulin antibodies are produced, though the titers of these antibodies are low, in patients given these im-
149
proved preparations [ 121, even human insulin preparations [ 131. The question of whether or not these weak insulin antibodies are responsible for insulin resistance has apparently not been resolved, because it is still often stated in textbooks or reviews [ 141 that insulin antibodies cause insulin resistance at the prereceptor or prebinding level. In addition, insulin autoantibodies have sometimes been found in newly diagnosed patients with insulin-dependent diabetes mellitus before treatment [ 15-171. Based on the rigid concept that insulin antibody causes insulin resistance, although nobody has mentioned it before, the authors think that it is theoretically possible to consider that these autoantibodies against endogenous insulin are partly responsible for insulin resistance, also frequently observed in newly diagnosed patients with insulin-dependent diabetes mellitus [ 18,191, and they participate in the further deterioration of glucose metabolism. Taken altogether, uncertainty has resulted from the unsolved question of whether or not insulin antibodies cause insulin resistance. Is it fact or hallowed fiction? Armitage et al. [ 51 argued that insulin antibodies are not responsible for insulin resistance after observing no relationship between titers of insulin antibodies produced by even ‘dirty’ insulin preparations and daily requirements of insulin in patients with insulin-dependent diabetes mellitus. From these earlier results, it is suggested as an hypothesis that insulin antibodies cause insulin resistance before attaining equilibrium, but are not responsible for insulin resistance in the steady state of equilibrium of insulin antibody binding with insulin. So far, to our knowledge, there have not been any reports on the investigation of insulin resistance induced by insulin antibodies using the glucose clamp technique, the most exact method of demonstrating insulin resistance in vivo [ 6,8,9]. From our data, it is clearly shown that insulin antibodies did not take part in insulin resistance in accordance with the argument advanced by Armitage et al. [ 51 using a different method.
With regard to the significantly low GIR during the clamp with 2 mU/kg/min insulin infusion accompanied by 100 pi/kg insulin antibody, insulin antibody GP30, used in our investigation, was confirmed to be a blocking type. Although it has been reported that 0.5 ml GP30 antiserum at a dilution of 1:80000 has the capacity to bind 54% of 2 PU porcine insulin, and that the K, was 1.8 x 10” [7], to our knowledge no in vitro study has revealed whether GP30 is a blocking type or not. There are two types of insulin antibodies, one of which binds to the B-chain of the insulin molecule and produces a blocking effect [20]. If the epitope of the antibody used in this investigation was not toward the binding site of the insulin molecule with the insulin receptor, in other words if the insulin-insulin antibody complex could bind to the insulin receptor, these glucose clamp studies would be meaningless, because insulin action would be normally revealed with or without insulin antibody. Fortunately, GP30 was a blocking-type antibody. The GIR during this clamp gradually increased as shown on the right of Fig. 2, suggesting that the binding of the antibody with infused porcine insulin had not reached equilibrium. If the clamp were continued for a longer time, GIR would be increased to the normal range. It is important that insulin antibodies produced gradually in patients given insulin preparations are different from these model experiments in which antibodies were given in a one-shot manner. Since even in this model insulin resistance was not revealed, it is suggested that weak insulin antibodies, produced in patients treated with monocomponent insulin or human insulin preparations or found in newly diagnosed IDDM patients, have no adverse effect on the cause of insulin resistance, because insulin antibodies produced by insulin preparations are usually present in a steady state of equilibrium with insulin, except for those which have a tremendously high binding capacity. It is also noted from Fig. 5 that free insulin levels did not represent insulin action in the presence of insulin antibody, but the insulin infusion
150
rates correlated more accurately with the insulin action. The free fraction of administered insulin binds to the insulin receptor and generates biological action, but the other fractions of administered insulin binding insulin antibodies might become free, bind the insulin receptor and generate biological action or recombine with the antibodies. Free insulin fraction is yielded successively in this manner. Considering the time course, all the insulin administered worked normally without insulin resistance even in the presence of insulin antibody, unless the binding capacity of insulin antibody was tremendously high. This might be important when one reads the data revealed by glucose clamp studies carried out on patients having insulin antibody. The manner adopted in most papers, in which free insulin levels were compared to normal subjects, could lead to underestimation of the tissue sensitivity of insulin in patients with insulin antibody. In clinical practice, however, since insulin antibodies have other properties, for example causing increased risk for delayed hypoglycemia [ 21,221, we of course agree that improved insulin preparations should be recommended to patients with diabetes mellitus. What we would like to emphasize is that weak insulin antibodies clinically found in patients receiving improved preparations have no adverse effect. In COI~CZUSZO~, our results suggested that even blocking antibodies against insulin do not cause insulin resistance, but they behave only as reservoirs once binding reaches equilibrium.
References Patterson. R., O’Rourke, J., Roberts, M. and Suszko, I. (1977) Immunologic reactions against insulin. 1. IgG antiinsulin and insulin resistance. J. Immunol. 110, 11261134. Witters. L. A., Ohman. J. L., Weir, G. C., Raymond, L. W. and Lowell, F. (1977) Insulin antibodies in the pathogenesis of insulin allergy and resistance. Am. J. Med. 63, 703-709. Kurtz. A.B. and Nabarro, J.D.N. (1980) Circulating insulin-binding antibodies. Diabetologia 19, 329-334. Nathan, D.M., Axelrod, L., Flier, J.S. and Carr, D.B.
5
6
7
8
9
10
11
12
13
14
15
16
17
(198 1) U-500 insulin in the treatment of antibody-mediated insulin resistance. Ann. Intern. Med. 94, 653-656. Armitage, M., McCaughey, E., Brooks, A., Cove, D., Betts, P., Casey, C. and Wilkin, T. (1988) Insulin resistance and insulin antibodies: fact or hallowed fiction? Pratt. Diabetes 5, 200-202. Andres, R., Swerdloff, R., Pozefsky, T. and Coleman D. (1966) Manual feedback technique for the control of blood glucose concentration. In: L.T. Skeggs Jr. (Ed.), Automation in Analytical Chemistry. Mediad, New York, pp. 486-491. Odell, W.D., Abraham, G.E., Skowsky, W.R., Hescox, M.A. and Fisher, D.A. (1971) Production of antisera for radioimmunoassay. In: W.D. Odell and W.H. Daughaday (Eds.), Principles of Competitive Protein-binding Assays. Lippincott, Philadelphia, pp. 57-76. Iochida, L.C., Tominaga, M., Matsumoto, M., Sekikawa, A. and Sasaki, H.,( 1989) Insulin resistance in septic rats - a study by the euglycemic clamp technique. Life Sci. 45, 1567-1573. DeFronzo, R., Tobin, J.D. and Andres, R. (1979) Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am. J. Physiol. 237, E214-E233. Steele, R. (1959) Influence of glucose loading and of injected insulin on hepatic glucose output. Ann. N.Y. Acad. Sci. 82, 420-430. Nakagawa, S., Nakayama, H., Sasaki, T., Yoshini. K., Yu, Y.Y., Shinozaki, K.,Aoki, S. and Mashimo. K. (1973) A simple method for determination of serum free insulin levels in insulin-treated patients. Diabetes 22, 590-600. Heding, L.G., Larsson, Y. and Ludvigssen, J. (1980) The immunogenicity of insulin preparation. Antibody levels before and after transfer to highly purified porcine insulin. Diabetologia 19. 5 1 l-5 15. Heding, L.G., Marshall, M.O., Persson, B., Dahlquist, G., Thalme, B., Lindgren, F., Akerblom, H.K., Rilva, A., Knip, M., Ludvigsson, J., Stenhammar, L., Stromberg, L., Sovik, O., Bmvre, H., Wefring, K., Vidnes, J., Kjagkd, J.J., Bro, P. and Kaad, P.H. (1984) Immunogenicity of monocomponent human and porcine insulin in newly diagnosed type 1 (insulin-dependent) diabetic children. Diabetologia 27, 96-98. Moller, D.E. and Flyer, J.S. (1991) Insulin resistance mechanisms, syndromes, and implications. N. Engl. J. Med. 325, 938-948. Palmer, J.P., Asplin. C.M., Clemens, P., Lyen, K.. Tatpati, 0.. Raghu, P.K. and Paquette, T.L. (1983) Insulin antibodies in insulin dependent diabetic before insulin treatment. Science 222, 1337-1339. Wilkin, T., Hoskins, P.J.. Armitage, M., Rodier, M., Casey, C., Diaz, J-L., Pyke, D.A. and Leslie, R.D.G. (1985) Value of insulin auto-antibodies as serum markers for insulin-dependent diabetes mellitus. Lancet i, 480482. Atkinson, M.A., MaClaren. N.K., Riley, W.J.. Winter,
151 W.E., Fisk, D.D. and Spillar, R.P. (1985) Are insulin autoantibodies markers for insulin-dependent diabetes mellitus? Diabetes 35, 894-898. 18 Del Prato, S., Noradini, R., Tiengo, A., Tessari, P., Avagaro, A., Travisan, R., Valerio, A., Muggeo, M., Cobelli, C. and Toffolo, G. (1983) Insulin-mediated glucose disposal in type I diabetes: evidence for insulin resistance. J. Clin. Endocrinol. Metab. 57, 904-910. 19 Nankervis, A., Proietto, J., Aitken, P. and Alford, F. (1984) Impaired insulin action in newly diagnosed type 1 (insulin-dependent) diabetes mellitus. Diabetologia 27. 497-503.
20 Kelso, J.M., Tamai, I.Y., Roth, M.D., Valdes, I. and Arquilla, E.R. (1980) Induction of hyperglycemia with insulin antibodies to B-chain determinants. Diabetes 29. 383390. 21 Bolli, G.B., Dimitriadis, G.D., Pehling, G.B.. Baker, B.A., Haymond, M.W., Cryer, P.E. and Gerich, J.E. (1984) Abnormal glucose counterregulation after subcutaneous insulin-dependent diabetes mellitus. N. Engl. J. Med. 310, 1706-1711. 22 Van Haerften, T.W. (1989) Clinical significance of insulin antibody in insulin-treated diabetic patients. Diabetes Care 12. 641-648.
143- 15 1 0 1992 Elsevier Science Publishers B.V. All rights reserved 0168-8227/92/$05.00
143
DIABET 00689
Insulin antibody does not cause insulin resistance during glucose clamping in rats M. Tominaga,
M. Matsumoto,
The Third Department of Internal Medicine,
M. Igarashi, H. Eguchi, A. Sekikawa
and H. Sasaki
Yamagata Universit.v School of Medicine, 2-2 Iida-Nishi 2-Chome,
Yamagata Cit.v,
Yamagata 990-23, Japan
(Received 16 March 1992) (Revision accepted 9 July 1992)
Summary
Although it has often been stated that insulin antibodies cause insulin resistance, this concept is still controversial. The effect of insulin antibody GP30, commonly used in insulin radioimmunoassay, on insulin action was investigated in Wistar rats in vivo by the euglycemic glucose clamp technique. As a preliminary experiment, the equilibrium time required for insulin antibody to bind with endogenous insulin was examined. One hundred pi/kg insulin antibody took 60 min or more to attain equilibrium, but 10 pi/kg insulin antibody almost immediately equilibrated with endogenous insulin. During a 60-min glucose clamp study, 2 mU/kg/min porcine insulin was infused with 100 pi/kg insulin antibody. At steady state, during the last 20-min period, the mean glucose infusion rate was 2.10 f 0.85 mg/kg/min (n = 5, mean ? SD), significantly lower than the 5.77 k 1.61 mg/kg/min of the control, indicating insulin resistance before equilibrium was reached. However, the glucose infusion rates during the clamp with 10 &kg insulin antibody and 100 pi/kg insulin antibody infused 75 min before the insulin were 6.10 k 1.44 and 7.12 + 1.19 mg/kg/min, respectively, no different from the control. In these instances, free insulin levels measured by radioimmunoassay using the polyethyleneglycol method were 43.8 + 20.4 and 15.4 t 6.1 pU/ ml, respectively, lower than the control (77.0 2 16.1 pU/ml). Furthermore, by increasing the insulin infusion rate to 5 mU/kg/min, the glucose infusion rate during the clamp with 100 pi/kg insulin antibody concomitantly infused was 11.79 & 1.14 mg/kg/min, no different from the 14.50 f 2.82 mg/kg/min of the control, while free insulin levels were 62.5 k 29.4 ,uU/ml, still lower than the control (152.7 + 69.5 pU/ ml). In conclusion, this in vivo study suggested that even blocking antibodies against insulin did not cause insulin resistance once binding reached equilibrium. Key words: Insulin antibody;
Insulin resistance;
Glucose clamp
Correspondence to: M. Tominaga, The Third Department of Internal Medicine. Yamagata University Iida-Nishi 2-Chome, Yamagata City, Yamagata 990-23, Japan.
School of Medicine, 2-2
144
Introduction In the past, patients with diabetes mellitus receiving impure insulin preparations were sometimes reported to have very high titers of insulin antibodies and to need tremendously high daily requirements of insulin [l-4], so it has commonly been believed that insulin antibodies induce insulin resistance. However, Armitage et al. recently reported that the titers of insulin antibodies measured by their own method did not correlate with the daily requirement of insulin [ 51. From this investigation, they suspected that insulin antibody was not the causal factor of insulin resistance, and that insulin antibody behaved as a reservoir after equilibrium with insulin. On the other hand, the euglycemic glucose clamp technique originally described by Andres et al. [6] has given us the most precise method for estimating insulin action in vivo. In the present study, insulin action was investigated in the presence of insulin antibody by this technique in rats in order to elucidate the contribution of insulin antibody to insulin resistance before and after equilibrium with insulin.
Materials and Methods Equilibrium time of insulin antibody GP30 with endogenous insulin Under 0.35 g/kg i.p. chloral hydrate anesthesia, Silastic catheters were cannulated into the left carotid artery and the left femoral vein of male Wistar rats weighing 265 k 19 g (mean k SD). Ten and 100 $/kg insulin antibody GP30 (BioMaker), commonly used in insulin radioimmunoassay [7], were injected through the femoral vein catheter. At three random times from 0, 7, 15, 30, 60, and 120 min after injection, 1 ml blood samples were obtained from the carotid artery catheter in order to minimize the total blood volume sampled to less than 3 ml. Then, free and total insulin concentrations were measured by radioimmunoassay to ascertain the equilibrium time required for insulin antibody GP30 to bind with the endogenous insulin of rats.
Glucose clamp technique Male Wistar rats were also used for glucose clamp after 24-h fasting. Under the same anesthesia using chloral hydrate, two Silastic catheters were cannulated into the left femoral vein for 3H-3glucose and 20% glucose infusion, and another catheter into the right femoral vein for the infusion of porcine insulin (Actrapid MC insulin, Novo) diluted with saline containing 0.25% bovine albumin (Sigma). A doubled lumen catheter was inserted into the right jugular vein for sampling of the heparinized blood to measure blood glucose levels by Glucose Monitor (Kyoto Daiichi Kagaku). Another Silastic catheter was cannulated into the left carotid artery for blood sampling to determine the specific activity of 3H-3glucose as well as free and total insulin levels and insulin-binding capacity. Initially, 5 PCi 3H-3-glucose was infused as a bolus followed by continuous infusion at a rate of 0.05 &i/min using a peristaltic rolling pump. Sixty minutes later, after blood sampling for determination of the specific activity of 3H-3glucose, porcine insulin solution was continuously infused at 2 and 5 mU/kg/min using a syringe pump after a lo-n-tin priming infusion. In one protocol, 100 pi/kg insulin antibody GP30 was infused 75 min before the insulin, but in others, 10 and 100 pi/kg insulin antibody was concomitantly infused at the beginning of insulin infusion. Changes in blood glucose levels were maintained within 10% of the basal levels for 60 min by variable infusion rates of 20% glucose which were calculated from the levels of, and changes in, glucose concentration measured by Glucose Monitor every 2 min. A modified algorithm [ 81 to determine the glucose infusion rate (GIR), originally described by DeFronzo et al. [ 91, was employed. The calculation done using a programmable calculator is as follows: GIR = 0.2 x (GD - BG) + GC GC = CC _ , - [BG-GD x 4 (BG-BG _ ,)I/64 where GD is the target blood glucose level, BG the blood glucose level, and GC a constant determined during the clamp.
145
Blood samples were taken at the end of the glucose clamp to measure free insulin levels and specific activity of 3H-3-glucose. Glucose disappearance rate (Gd) and hepatic glucose output (HGO) were determined by Steele’s method [ lo]. These are the parameters for insulin action along with the mean GIR during the last 20 min clamp. Determination offree and total insulin Free and total insulin levels were determined by radioimmunoassay according to Nakagawa et al. [ 1 I]. Serum samples were immediately mixed with ice-chilled polyethylene glycol and centrifuged. The supernatant was kept at -20 “C until assay. Other serum samples for assay of total insulin and insulin-binding capacity were stored at -20 “C until assay. Free insulin levels in the supernatant were determined by radioimmunoassay. Total insulin levels of serum samples after acidification with HCl were also measured by radioimmunoassay. Insulin-binding capacities of the same samples with ‘?-insulin were estimated by the polyethyleneglycol method.
Statistical analysis Statistical significance was assessed by Student’s t-test.
Results Equilibrium time of insulin antibody GP30 with endogenous insulin As shown in Fig. 1, total insulin levels increased gradually in rats injected with 100 PI/kg insulin antibody GP30 and reached a plateau 60 min after injection. However, a small peak was seen in the 10 $/kg insulin antibody injected group 7 to 15 min after administration (Fig. 1). Free insulin levels were suppressed by both doses of insulin antibody during the entire experimental Control
Insulin Antibody
6.
-2
P _,
%=
4
E” 2-
F-RKlOOuVkg)
o0
30 Time(minl
0715
30
60
120
Timdmin)
Fig. 1. Free and total insulin levels after 10 and 100 $/kg insulin antibody GP30. After 10 and 100 PI/kg insulin antibody GP30 were injected into rats, free and total insulin levels were measured by radioimmunoassay using the polyethyleneglycol method of Nakagawa. Total insulin levels in the 100 PI/kg insulin antibody group (open circle) increased gradually and reached a plateau at 60 min after injection. Total insulin levels in the 10 pi/kg injected group (closed circle) reached a small peak 7 to 15 min after administration. Free insulin levels in both 10 and 100 pi/kg injected rats (open and closed squares, respectively) were almost suppressed.
60
0
30
60
Time(mm)
Fig. 2. Rat euglycemic glucose clamp: control experiment of 2 mU/kg/min insulin infusion and 100 pi/kg insulin antibody concomitantly infused. Left: euglycemic glucose clamp with 2 mU/kg/min porcine insulin infusion as a control. Blood glucose level (BG) was maintained within log,, range of the fasting glucose level by variable glucose infusion rates (GIR) shown in the lower columns. Steady-state GIR, estimated by the mean of GIR during the last 20 min clamp, represents the insulin action. Steady-state GIR of this group was 5.77 f 1.61 mg/kg/min. Right: euglycemic glucose clamp, 100 &kg insulin antibody infused at the beginning of 2 mU/ kg/min porcine insulin infusion. Steady-state GIR was 2.20 f 0.85 mg/kg/min, significantly lower than the control. In this figure, data for BG and GIR are shown as mean k SE.
146
period. From these experiments, it appears that 100 PI/kg insulin antibody took 60 min or more to reach equilibrium, but 10 PI/kg insulin antibody almost immediately equilibrated with endogenous insulin. Insulin resistance before binding of insulin antibody> with insulin reached equilibrium Steady-state GIR of the glucose clamp with 2 mU/kg/min insulin infusion was 5.77 + 1.61 mg/kg/ min in the control, as shown on the left of Fig. 2. GIR of the clamp with 100 ,nl/kg insulin antibody (Fig. 2, right) was 2.10 t 0.85 mg/kg/min, significantly lower than the control, indicating insulin resistance. However, it can be seen that GIR gradually increased towards the end-point of the clamp suggesting that the binding of the insulin antibody with the infused porcine insulin was reaching equilibrium during this 60-min clamp, since the binding of
TABLE Data
100 pi/kg insulin antibody to reach equilibrium is a time-consuming process. G, and HGO, other parameters for insulin action, are shown in Table 1. Like GIR, G, during the clamp with concomitant infusion of 100 pi/kg insulin antibody was significantly lower than the control. HGO during the clamp with 100 pi/kg insulin antibody was not suppressed, although HGO was almost completely suppressed by 2 mU/kg/min insulin infusion in the control rats. These results suggested that insulin resistance caused by insulin antibody was present in both peripheral tissue and the liver before the binding of insulin antibody with insulin reached equilibrium. No insulin resistance after binding of insulin antibody with insulin reached equilibrium As shown on the left of Fig. 3, in the clamp with 10 pi/kg insulin antibody, GIR was
1
of rat glucose
clamp Control
Insulin
antibody”
Insulin
infusion
Insulin
antibody”
(100 PI/kg)
(10 PI/kg)
Insulin
antibodyb
(100 PI/kg)
rate 2 mU/kg/min
Number
of experiments
Glucose
infusion
01)
5
rate (mg/kg/min)
5
5
5
5.77 f 1.61
6.10& 1.44
2.10 + 0.85*
7.12& 1.19
Glucose
disappearance
rate (mg/kg/min)
6.66 & 0.89
7.52 * 0.80
4.87 + 1.27%
7.00 + 1.04
Hepatic
glucose
(mg/kg/min)
0.89 & 1.97
1.42i
2.77 f 1.85
0.11
77.0 + 16.1
43.8 * 20.4*
21.3 + 3.0*
15.4 + 6.1*
Free insulin Total
insulin
Insulin
output
level (p U/ml) level (p U/ml)
binding
capacity
130.9 + 24.2
(“,)
9.9 * 3.3
Insulin infusion rate 5 mLJ/kg/min Number of experiments (n) Glucose
infusion
Glucose
disappearance
rate (mg/kg/min)
Hepatic
glucose
(mg/kg/min)
Free insulin Total Insulin Glucose
insulin
5
rate (mg/kg/min) output
level (p U/ml) capacity
disappearance
5
10.56 k 2.53
9.73 * 0.52
8.38 + 1.13
- 3.94 & 2.65
- 5.42 & 2.55
- 3.16+_ 1.01
74.4 * 27.7* 210.9 k 86.8 5.1 f- 1.9
glucose
33.4 f 7.8
5
(“,)
rate and hepatic
254.0 k 62.3
41.7 + 5.6
15.15 & 2.76
152.7 + 69.5
output
during
steady
state of glucose
f 0.34
701.0 * 123.1
14.5 + 2.82
level @U/ml)
binding
1.31
11.79-t
1.14
62.1 + 29.4* 517.0 2 157.3 20.9 & 8.4 clamp
were estimated
by Steele’s method
using ‘H-3-glucose. Free and total insulin levels were measured by radioimmunoassay using polyethyleneglycol. a Insulin antibody administered concomitantly with insulin infusion. b Insulin antibody administered 75 min before insulin infusion. Data are shown as mean + SD.
*p
147 lnslin
haulin
Antibody lOO~Vkg(75min w
Antibody
huh
prior)
Anttbody
100Wkg
15 -
e
.E
5g
lo-
E 5-
0
30 Timetmin)
60
0
30
60
Timahin)
Fig. 3. Rat euglycemic glucose clamp: 10 PI/kg insulin antibody concomitantly infused group and 75 min prior 100 PI/kg insulin antibody infused group both with 2 mU/kg/min insulin infusion. Left: euglycemic glucose clamp of 2 mU/kg/min insulin infusion with 10 pi/kg insulin antibody concomitantly infused. GIR was 6.10 k 1.44 mg/kg/min, not different from that of the control. Right: euglycemic glucose clamp of 100 nl/ kg insulin antibody 75 min before the beginning of 2 mu/kg/ min insulin infusion. GIR was 7.12 f 1.19 mg/kg/min, not different from the control. Data are shown as mean t SE.
6.10 & 1.44 mg/kg/min, no different from the control, indicating no insulin resistance even with insulin antibody. On the right of Fig. 3 is shown the clamp with 100 pi/kg insulin antibody, in which insulin antibody was infused 75 min before the infusion of 2 mU/kg/min porcine insulin. GIR was 7.12 + 1.19 mg/kg/min, also no different from the control. Furthermore, since with the more insulin infused, the faster equilibrium is reached, a glucose clamp with 100 /A/kg insulin antibody concomitantly infused with increased doses of insulin to 5 mU/kg/min was carried out. GIR was 11.79 k 1.14 mg/kg/min, as shown on the right of Fig. 4, no different from the control (14.50 k 2.82 mg/kg/min) as shown on the left of Fig. 4. GTR of 10 &kg insulin antibody with 5 mU/kg/min insulin infusion was also no different from the control (not shown in figure, see Table 1). G,s during the clamp of 2 mU/kg/min insulin
n. v
0
30 Time(min)
60
0
30
60
Time(min)
Fig. 4. Rat euglycemic glucose clamp: control experiment of 5 mU/kg/min insulin infusion and 100 pi/kg insulin antibody concomitantly infused. Left: euglycemic glucose clamp of 5 mU/kg/min insulin infusion as a control. GIR was 14.50 k 2.82 mg/kg/min. Right: euglycemic glucose clamp of 5 mU/kg/min insulin infusion concomitantly infused with 100 PI/kg insulin antibody. GIR was 11.79 & 1.14 mg/kg/min, not different from the control. Data are shown as mean f SE.
infusion with 10 pi/kg insulin antibody and 100 pi/kg insulin antibody infused 75 min before insulin infusion were not different from the control, as shown in Table 1. Upon increasing the insulin infusion rate to 5 mU/kg/min, there was no difference from the control in either G, or HGO during the clamp with 100 pi/kg insulin antibody. Free insulin levels during the clamp of 2 mU/ kg/min insulin infusion with 10 $/kg insulin antibody were 43.8 2 20.4 pU/ml, significantly lower when compared to 77.0 + 16.1 pU/ml during the control clamp as shown in Table 1, although GIR, G, and HGO during the two clamps did not differ. Furthermore, during the clamp of 5 mu/kg/ min insulin infusion with 10 and 100 $/kg insulin antibody, free insulin levels were 74.4 rf-27.7 and 62.1 k 29.4 pU/ml, respectively, significantly lower than the 152.7 & 69.5 pU/ml of the control. Total insulin levels and insulin binding capacities found in these glucose clamp studies are also shown in Table 1.
148
20
15
10
5
0
200
100 Free Insulin
0
Level
1
1
2
5
Insulin
(pU/ml)
Infusion
Rate
(mU/kg/min)
Fig. 5. Relationship of free insulin levels and insulin infusion rates to glucose infusion rates during glucose clamp. Left: relationship between free insulin levels and glucose infusion rates. Closed circle indicates control; closed rectangle indicates glucose clamp with 10 PI/kg insulin antibody, and open rectangle indicates glucose clamp with 100 pi/kg insulin antibody which was administered 75 min before the beginning of 2 mU/kg/min insulin infusion and given concomitantly with 5 mU/kg/min insulin infusion. Despite the significantly lower free insulin levels of the antibody infused groups, the glucose infusion rates were almost the same as those of the control. Right: relationship between insulin infusion rates and glucose infusion rates during clamps. It was noted that the dose-response obtained by 10 and 100 pi/kg insulin antibody almost overlapped that of the control. Data are shown as mean k SE.
On the left of Fig. 5, the relationship between GIR representing insulin action during the clamp with insulin antibody, and free insulin levels is shown. It can be seen that free insulin levels were significantly lower than the control, while GIR is comparable to the control clamp at the same insulin rate. In other words, free insulin levels did not seem to be responsible for insulin action, but insulin infusion rates seemed to have more effect. This relationship is more clearly demonstrated on the right of Fig. 5 in which the dose-response obtained from the clamps with 10 and 100 pi/kg insulin antibody was almost overlapped by that of the control clamps.
Discussion It is well known that improvement in insulin preparations has been achieved to avoid several clinical problems caused by ‘dirty’ preparation, such as lipodystrophy, lipoatrophy and insulin allergy. After the development of highly purified or monocomponent insulin and recombinant human insulin preparations, which have less antigenicity and are now widely available for clinical practice, the problem of insulin resistance due to high titers of insulin antibody [l-4] seems to have faded away. However, it is also known that insulin antibodies are produced, though the titers of these antibodies are low, in patients given these im-
149
proved preparations [ 121, even human insulin preparations [ 131. The question of whether or not these weak insulin antibodies are responsible for insulin resistance has apparently not been resolved, because it is still often stated in textbooks or reviews [ 141 that insulin antibodies cause insulin resistance at the prereceptor or prebinding level. In addition, insulin autoantibodies have sometimes been found in newly diagnosed patients with insulin-dependent diabetes mellitus before treatment [ 15-171. Based on the rigid concept that insulin antibody causes insulin resistance, although nobody has mentioned it before, the authors think that it is theoretically possible to consider that these autoantibodies against endogenous insulin are partly responsible for insulin resistance, also frequently observed in newly diagnosed patients with insulin-dependent diabetes mellitus [ 18,191, and they participate in the further deterioration of glucose metabolism. Taken altogether, uncertainty has resulted from the unsolved question of whether or not insulin antibodies cause insulin resistance. Is it fact or hallowed fiction? Armitage et al. [ 51 argued that insulin antibodies are not responsible for insulin resistance after observing no relationship between titers of insulin antibodies produced by even ‘dirty’ insulin preparations and daily requirements of insulin in patients with insulin-dependent diabetes mellitus. From these earlier results, it is suggested as an hypothesis that insulin antibodies cause insulin resistance before attaining equilibrium, but are not responsible for insulin resistance in the steady state of equilibrium of insulin antibody binding with insulin. So far, to our knowledge, there have not been any reports on the investigation of insulin resistance induced by insulin antibodies using the glucose clamp technique, the most exact method of demonstrating insulin resistance in vivo [ 6,8,9]. From our data, it is clearly shown that insulin antibodies did not take part in insulin resistance in accordance with the argument advanced by Armitage et al. [ 51 using a different method.
With regard to the significantly low GIR during the clamp with 2 mU/kg/min insulin infusion accompanied by 100 pi/kg insulin antibody, insulin antibody GP30, used in our investigation, was confirmed to be a blocking type. Although it has been reported that 0.5 ml GP30 antiserum at a dilution of 1:80000 has the capacity to bind 54% of 2 PU porcine insulin, and that the K, was 1.8 x 10” [7], to our knowledge no in vitro study has revealed whether GP30 is a blocking type or not. There are two types of insulin antibodies, one of which binds to the B-chain of the insulin molecule and produces a blocking effect [20]. If the epitope of the antibody used in this investigation was not toward the binding site of the insulin molecule with the insulin receptor, in other words if the insulin-insulin antibody complex could bind to the insulin receptor, these glucose clamp studies would be meaningless, because insulin action would be normally revealed with or without insulin antibody. Fortunately, GP30 was a blocking-type antibody. The GIR during this clamp gradually increased as shown on the right of Fig. 2, suggesting that the binding of the antibody with infused porcine insulin had not reached equilibrium. If the clamp were continued for a longer time, GIR would be increased to the normal range. It is important that insulin antibodies produced gradually in patients given insulin preparations are different from these model experiments in which antibodies were given in a one-shot manner. Since even in this model insulin resistance was not revealed, it is suggested that weak insulin antibodies, produced in patients treated with monocomponent insulin or human insulin preparations or found in newly diagnosed IDDM patients, have no adverse effect on the cause of insulin resistance, because insulin antibodies produced by insulin preparations are usually present in a steady state of equilibrium with insulin, except for those which have a tremendously high binding capacity. It is also noted from Fig. 5 that free insulin levels did not represent insulin action in the presence of insulin antibody, but the insulin infusion
150
rates correlated more accurately with the insulin action. The free fraction of administered insulin binds to the insulin receptor and generates biological action, but the other fractions of administered insulin binding insulin antibodies might become free, bind the insulin receptor and generate biological action or recombine with the antibodies. Free insulin fraction is yielded successively in this manner. Considering the time course, all the insulin administered worked normally without insulin resistance even in the presence of insulin antibody, unless the binding capacity of insulin antibody was tremendously high. This might be important when one reads the data revealed by glucose clamp studies carried out on patients having insulin antibody. The manner adopted in most papers, in which free insulin levels were compared to normal subjects, could lead to underestimation of the tissue sensitivity of insulin in patients with insulin antibody. In clinical practice, however, since insulin antibodies have other properties, for example causing increased risk for delayed hypoglycemia [ 21,221, we of course agree that improved insulin preparations should be recommended to patients with diabetes mellitus. What we would like to emphasize is that weak insulin antibodies clinically found in patients receiving improved preparations have no adverse effect. In COI~CZUSZO~, our results suggested that even blocking antibodies against insulin do not cause insulin resistance, but they behave only as reservoirs once binding reaches equilibrium.
References Patterson. R., O’Rourke, J., Roberts, M. and Suszko, I. (1977) Immunologic reactions against insulin. 1. IgG antiinsulin and insulin resistance. J. Immunol. 110, 11261134. Witters. L. A., Ohman. J. L., Weir, G. C., Raymond, L. W. and Lowell, F. (1977) Insulin antibodies in the pathogenesis of insulin allergy and resistance. Am. J. Med. 63, 703-709. Kurtz. A.B. and Nabarro, J.D.N. (1980) Circulating insulin-binding antibodies. Diabetologia 19, 329-334. Nathan, D.M., Axelrod, L., Flier, J.S. and Carr, D.B.
5
6
7
8
9
10
11
12
13
14
15
16
17
(198 1) U-500 insulin in the treatment of antibody-mediated insulin resistance. Ann. Intern. Med. 94, 653-656. Armitage, M., McCaughey, E., Brooks, A., Cove, D., Betts, P., Casey, C. and Wilkin, T. (1988) Insulin resistance and insulin antibodies: fact or hallowed fiction? Pratt. Diabetes 5, 200-202. Andres, R., Swerdloff, R., Pozefsky, T. and Coleman D. (1966) Manual feedback technique for the control of blood glucose concentration. In: L.T. Skeggs Jr. (Ed.), Automation in Analytical Chemistry. Mediad, New York, pp. 486-491. Odell, W.D., Abraham, G.E., Skowsky, W.R., Hescox, M.A. and Fisher, D.A. (1971) Production of antisera for radioimmunoassay. In: W.D. Odell and W.H. Daughaday (Eds.), Principles of Competitive Protein-binding Assays. Lippincott, Philadelphia, pp. 57-76. Iochida, L.C., Tominaga, M., Matsumoto, M., Sekikawa, A. and Sasaki, H.,( 1989) Insulin resistance in septic rats - a study by the euglycemic clamp technique. Life Sci. 45, 1567-1573. DeFronzo, R., Tobin, J.D. and Andres, R. (1979) Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am. J. Physiol. 237, E214-E233. Steele, R. (1959) Influence of glucose loading and of injected insulin on hepatic glucose output. Ann. N.Y. Acad. Sci. 82, 420-430. Nakagawa, S., Nakayama, H., Sasaki, T., Yoshini. K., Yu, Y.Y., Shinozaki, K.,Aoki, S. and Mashimo. K. (1973) A simple method for determination of serum free insulin levels in insulin-treated patients. Diabetes 22, 590-600. Heding, L.G., Larsson, Y. and Ludvigssen, J. (1980) The immunogenicity of insulin preparation. Antibody levels before and after transfer to highly purified porcine insulin. Diabetologia 19. 5 1 l-5 15. Heding, L.G., Marshall, M.O., Persson, B., Dahlquist, G., Thalme, B., Lindgren, F., Akerblom, H.K., Rilva, A., Knip, M., Ludvigsson, J., Stenhammar, L., Stromberg, L., Sovik, O., Bmvre, H., Wefring, K., Vidnes, J., Kjagkd, J.J., Bro, P. and Kaad, P.H. (1984) Immunogenicity of monocomponent human and porcine insulin in newly diagnosed type 1 (insulin-dependent) diabetic children. Diabetologia 27, 96-98. Moller, D.E. and Flyer, J.S. (1991) Insulin resistance mechanisms, syndromes, and implications. N. Engl. J. Med. 325, 938-948. Palmer, J.P., Asplin. C.M., Clemens, P., Lyen, K.. Tatpati, 0.. Raghu, P.K. and Paquette, T.L. (1983) Insulin antibodies in insulin dependent diabetic before insulin treatment. Science 222, 1337-1339. Wilkin, T., Hoskins, P.J.. Armitage, M., Rodier, M., Casey, C., Diaz, J-L., Pyke, D.A. and Leslie, R.D.G. (1985) Value of insulin auto-antibodies as serum markers for insulin-dependent diabetes mellitus. Lancet i, 480482. Atkinson, M.A., MaClaren. N.K., Riley, W.J.. Winter,
151 W.E., Fisk, D.D. and Spillar, R.P. (1985) Are insulin autoantibodies markers for insulin-dependent diabetes mellitus? Diabetes 35, 894-898. 18 Del Prato, S., Noradini, R., Tiengo, A., Tessari, P., Avagaro, A., Travisan, R., Valerio, A., Muggeo, M., Cobelli, C. and Toffolo, G. (1983) Insulin-mediated glucose disposal in type I diabetes: evidence for insulin resistance. J. Clin. Endocrinol. Metab. 57, 904-910. 19 Nankervis, A., Proietto, J., Aitken, P. and Alford, F. (1984) Impaired insulin action in newly diagnosed type 1 (insulin-dependent) diabetes mellitus. Diabetologia 27. 497-503.
20 Kelso, J.M., Tamai, I.Y., Roth, M.D., Valdes, I. and Arquilla, E.R. (1980) Induction of hyperglycemia with insulin antibodies to B-chain determinants. Diabetes 29. 383390. 21 Bolli, G.B., Dimitriadis, G.D., Pehling, G.B.. Baker, B.A., Haymond, M.W., Cryer, P.E. and Gerich, J.E. (1984) Abnormal glucose counterregulation after subcutaneous insulin-dependent diabetes mellitus. N. Engl. J. Med. 310, 1706-1711. 22 Van Haerften, T.W. (1989) Clinical significance of insulin antibody in insulin-treated diabetic patients. Diabetes Care 12. 641-648.