| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Department of Endocrinology (Å.L.L., U.-B.E., B.H., L.G., T.T.), University of Lund, S-205 02 Malmö, Sweden; and Department of Medicine (T.T.), Helsinki University Central Hospital, FIN-00029 Helsinki, Finland
Address all correspondence and requests for reprints to: Tiinamaija Tuomi, M.D., Wallenberg Laboratory, Department of Endocrinology, University of Lund, S-205 02 Malmö, Sweden. E-mail: Tiinamaija. . Tuomi{at}endo.mas.lu.se
Abstract
To study whether antibodies to glutamic acid decarboxylase (GADab) are associated with subclinical ß-cell damage and impaired insulin secretion, we screened 441 nondiabetic patients with autoimmune thyroiditis (AT) for GADab, and 15 (3.4%) were found positive. Antibodies to IA-2 were found in two GADab+ and one GADab- patients. We matched 11 GADab+ and 13 GADab- AT patients who were euthyroid on thyroxin supplementation, and 13 control subjects for sex, age, and body mass index and measured insulin, C-peptide, and glucagon response to glucose and arginine at three blood glucose concentrations (fasting, 14 mmol/liter, >25 mmol/liter). In the fasting state, all groups had similar blood glucose concentration and HbA1c level, but the serum insulin concentration was higher in the AT patients compared with the control subjects (P < 0.04). The acute insulin response to arginine was lower in GADab+ than in GADab- thyroiditis subjects at glucose concentration of 14 and >25 mmol/liter (AIR14: 76.8 ± 52.0 vs. 158.2 ± 118.2 mU/liter, P = 0.040; AIR>25: 84.3 ± 64.4 vs. 167.9 ± 101.5 mU/liter, P = 0.035). In conclusion, GADab were associated with a decreased insulin secretion capacity in nondiabetic subjects with thyroiditis, which suggests that GADab positivity could be a marker of subclinical insulitis.
THE ß-CELL DESTRUCTION that leads to insulin deficiency in type 1 diabetes (1) is associated with circulating antibodies to insulin (IAA), islet cell cytoplasmic antigens (ICA), glutamic acid decarboxylase (GADab), and protein tyrosine phosphatase-like protein (IA-2ab) (2, 3, 4, 5, 6, 7). Antibody positivity, especially positivity for multiple antibodies, is highly predictive of diabetes in nondiabetic first degree relatives of subjects with type 1 diabetes (8, 9). This risk is increased in siblings who share the HLA type of their type 1 diabetic sibling or have HLA DR3 and/or DR4 (10, 11, 12, 13). ICA positivity has been associated with an impaired first phase insulin response during an iv glucose tolerance test (IVGTT) even in first degree relatives who have not developed diabetes (14, 15, 16, 17).
GADab are the most sensitive marker for autoimmune diabetes in adults (18, 19). However, little is known about the insulin secretory capacity or risk for diabetes in nondiabetic GADab+ adults who do not have relatives with type 1 diabetes or polyendocrine diseases. No deterioration was seen in the insulin response during IVGTT in adult Scandinavian nondiabetic GADab+ relatives of type 2 diabetic patients compared with GADab- subjects (20). However, although IVGTT provides a sensitive measure of first-phase insulin secretion, it does not test the maximum capacity of the ß-cell. Using a highly sensitive iv glucose-arginine test, we could demonstrate a decreased maximal insulin secretory capacity in LADA patients with short-duration compared with GADab- type 2 diabetic patients, although the groups had similar fasting glucose, insulin, and C-peptide concentrations (21).
The aim of this study was to further investigate whether GADab in nondiabetic adult subjects with chronic autoimmune thyroiditis were associated with an impaired insulin response to maximal stimulation of the ß-cells with iv glucose and arginine. In addition, we studied the short-term risk for GADab+ patients with thyroiditis to develop diabetes.
Subjects and Methods
From 1995–1999, we analyzed GADab in 441 patients attending two outpatient clinics in southern Sweden for follow-up of their chronic autoimmune thyroiditis (AT). At diagnosis of AT, most patients had clinical or subclinical hypothyroidism, and all had antibodies to thyroglobulin, thyroid peroxidase, or thyroid microsomal antigens. In the majority, the diagnosis was verified by a needle biopsy of the thyroid. At the time of the study, all patients were clinically euthyroid and had normal serum concentrations of TSH and free T3. (Table 1
). Most of the patients were on thyroxin replacement therapy.
|
. The study was approved by the Ethics Committee of the University of Lund. Methods
An iv glucose-arginine test was performed after an overnight fast as previously described (21, 22) (Fig. 1). Briefly, blood samples were drawn for measurement of glucose, insulin, C-peptide and glucagon 5 and 2 min before as well as 2, 3, 4, and 5 min after an iv injection of 5 g arginine hydrochloride. This procedure was carried out at basal state and after the blood glucose concentration had been raised to 14 mmol/liter and >25 mmol/liter [coefficient of variation (CV), 8.1% and 7.0%, respectively) by a variable iv infusion of 20% glucose. A 2.5-h resting period was allowed before raising the blood glucose to >25 mmol/liter (over 25–30 min). The acute insulin response to arginine (AIR) was calculated as the mean of the +2 to +5 min values after subtraction of the mean of the prestimulus values. The acute C-peptide (ACR) and glucagon (AGR) responses to arginine were calculated in the same manner.
|
AIR/glucose). From this equation, the ß cell sensitivity to glucose (BG50) was calculated as the blood glucose level at which half-maximal AIR to arginine (AIRmax/2) occurred. An insulin resistance index was calculated from fasting blood glucose and insulin concentration according to the homeostasis model assessment (HOMA) (23).
A double antibody ELISA (DAKO Corp., Cambridgeshire, UK) was used to measure serum insulin concentrations with an interassay CV of 8.9%. Serum C-peptide concentrations were measured using RIA (Linco Research, Inc., St. Charles, MO) with an interassay CV of 9.8%. Plasma glucagon concentrations were measured at -2 min, 3 min, and 4 min using RIA (Linco, Linco Research, Inc., St. Charles, MO) with a CV of 3.6% and a cross-reactivity with enteric glucagon of 0.1%. Blood glucose was measured on an Accutrend blood glucose analyzer.
GAD and IA-2 antibodies were measured from frozen serum samples by a radioimmunoprecipitation assay employing 35S-labeled recombinant human IA-2 [intracellular part, (24)] or GAD65 produced by in vitro transcription/translation as described previously for GADab (25). The results were expressed as relative units, RU = (sample cpm - mean cpm of three negative controls)/(cpm of a positive internal references serum - mean cpm of three negative controls) x 100. The cut-off limit for positivity was 5 RU for GADab and 2.5 RU for IA-2ab, which represents the mean + 3 SD of 296 (GADab) or 155 (IA-2ab) in Finnish healthy control subjects. In the Combined Autoantibody Workshop, the specificity of the GADab assay was 99% and the sensitivity was 75% (26); in the Diabetes Antibody Standardization Program, the specificity/sensitivity was 94%/76% for the GADab assay and 100%/57% for the IA-2ab assay. Five RU equals 32 IU/ml of GADab according to the standardized international units.
IAA were measured in sera of the 24 patients participating in the glucose-arginine test, using a radioimmunoprecipitation microassay employing mono-[125 I] TyrA (14)-human insulin (27). The IAA levels were expressed in RU based on a standard curve. Cut-off limit for positivity was 1.55 RU (99th percentile in 374 nondiabetic subjects). The sensitivity of the microassay was 35%, and the specificity was 100% in the Combined Autoantibody Workshop (26).
Statistical analyses were performed using the BMDP new system, version 1.12, for Windows (BMDP Statistical Software, Inc., Los Angeles, CA). Data are given as mean ± SD or as median [interquartile range] unless indicated otherwise. Statistical significance of the difference between groups was tested by the Mann-Whitney Individual Rank Sum Test.
Results
Among the 441 nondiabetic patients with AT, fifteen (3.4%) had GADab. In addition, two of the GADab+ patients and one GADab- patient had IA-2ab. IAA were found in 1/11 GADab+ patients and in none of the 13 GADab- patients participating in the glucose arginine test. The GADab measurement was repeated at the time of investigation. After a mean period of 8.2 (± 2.3) months, only two patients with initially marginally increased GADab levels seroconverted to GADab- (Table 2). Their insulin and glucagon responses during the iv glucose-arginine test did not differ from the other GADab+ patients (Table 2).
|
). Irrespective of GADab positivity the patients with thyroiditis had similar fasting serum (fS) C-peptide (GADab+ vs. GADab-: 0.47 [0.27] vs. 0.43 [0.52] nmol/liter) and insulin concentrations (5.5 [7.3] vs. 6.5 [4.3]) as well as HOMA insulin resistance indices (51.2 [1.6] vs. 1.4 [1.3]). The corresponding values were lower for the control subjects; fS-C-peptide 0.27 [0.22] nmol/liter, fS-insulin 1.8 [4.6] mU/liter, HOMA insulin resistance index 0.4 [1.1] (P < 0.04 vs. both thyroiditis groups) (Table 1). Insulin and C-peptide responses
The fasting insulin concentration and the insulin response to glucose was similar in GADab+ and GADab- patients (Fig. 1 and 2A, Table 3). Also, the insulin response to arginine at fasting was similar in the two AT groups. However, at the higher glucose concentrations the GADab+ patients had lower insulin response to arginine compared with the GADab- patients (AIR14: 63.4 [59.5] vs. 98.7 [176.2] mU/liter, P = 0.040; AIR>25: 58.5 [89.5] vs. 152.0 [139.3] mU/liter, P = 0.035) (Fig. 2B, Table 3). The insulin concentration peaked 4 min after the arginine injection. This peak concentration was significantly lower in GADab+ than in GADab- patients, both at blood glucose of 14 mmol/liter (87.2 [103.0] vs. 144.7 [199.5] mU/liter, P = 0.026) and >25 mmol/liter (108.2 [155.2] vs. 197.1 [172.8] mU/liter, P = 0.029) (Fig. 1). There was no significant difference in the insulin response to 14 or >25 mmol/liter glucose and arginine between the control subjects and the thyroiditis patients (Fig. 1).
|
|
). The acute C-peptide response to glucose and arginine was lower in GADab+ than in GADab- patients at fasting (ACRf, 0.40 [0.51] vs. 0.70 [0.82] nmol/liter, P = 0.040) (Table 3). The glucose potentiation of ß-cell function was lower in GADab+ compared with GADab- patients (SLOPEAIR: 5.3 [5.3] vs. 9.6 [17.2], P = 0.030), while the ß-cell sensitivity to glucose was unaffected (BG50: GADab+: 7.9 [6.2] vs. GADab-: 7.6 [2.0] mmol/liter). The respective figures for the control group were 7.2 [7.5] (SLOPEAIR) and 9.1 [2.9] mmol/liter (BG50).
Glucagon response
The glucagon concentration decreased with increasing blood glucose concentration without any difference between the three groups. The glucagon response to arginine peaked 3 min after the injection (Fig. 3). This peak concentration was lower in GADab+ than in GADab- patients at all glucose concentrations (fasting glucose: 108.5 [30.7] vs. 140.6 [65.7] ng/liter, P = 0.022; 14 mmol/liter glucose: 68.8 [35.3] vs. 90.1 [48.8] ng/liter, P: ns; >25 mmol/liter glucose: 55.0 [15.7] vs. 71.7 [20.8] ng/liter, P = 0.042), whereas the total glucagon response to arginine was not significantly different between either the AT groups or the control subjects (GADab+ AT vs. GADab- AT vs. control subjects: AGRf: 62.2 [19.8] vs. 85.4 [59.6] vs. 80.8 [86.6]; AGR14: 34.8 [27.5] vs. 47.8 [63.1] vs. 51.8 [63.1]; AGR>25: 25.3 [23.5] vs. 21.3 [42.8] vs. 47.3 [31.3] ng/liter) (Fig. 2B, Table 3).
|
During the median follow-up of 48 months, two of the 15 (13.3%) GADab+ compared with 11 of the 426 (2.6%) GADab- patients were diagnosed with diabetes (P = 0.08). None of them was treated with insulin. One GADab+ patient was diagnosed with diabetes 18 months after the investigation. Another GADab+ patient who was also IA-2ab+ had a slightly elevated fasting blood glucose (6.8 mmol/liter) at the time of investigation despite a normal HbA1c level (4.2%). She was diagnosed with diabetes 4 weeks after the initial test with an oral glucose tolerance test and reexamined 32 months later when her HbA1c level was still within the normal range on diet treatment. Her insulin response to arginine was only about 50% of that in the first test (AIRf: 21.6 vs. 12.9, AIR14: 35.1 vs. 20.4, AIR>25: 57.1 vs. 25.2 mU/liter). Also the C-peptide response was halved (ACR14: 0.82 vs. 0.35, ACR>25 1.04 vs. 0.55 nmol/liter) , whereas there was no difference in her fasting C-peptide concentration (ACRf: 0.40 vs. 0.34 nmol/liter). The glucagon response was 8–10% lower at glucose of 14 and > 25 mmol/liter (AGR14: 82.4 vs. 75.5, AGR>25: 82.8 vs.69.5 ng/liter), with no difference at fasting (AGRf: 92.8 vs. 90.6).
Discussion
In this study, we showed that GADab positivity in nondiabetic patients with autoimmune thyroiditis was significantly associated with a decreased maximal insulin secretory capacity. Despite similar fasting glucose and insulin concentrations and HbA1c level the GADab+ patients had significantly lower insulin response compared with the GADab- patients after arginine injection at glucose concentrations of 14 and >25 mmol/liter. The present results are thus in line with our previous study on LADA patients with short diabetes duration (21).
The cause of the decreased insulin secretion in the GADab+ patients in this study is unlikely to result from glucose toxicity or differences in insulin sensitivity as we studied nondiabetic AT subjects with similar degree of insulin sensitivity as estimated with the HOMA index. However, in keeping with some previous reports, all AT patients seemed to be more insulin resistant than the control subjects without AT (28, 29, 30, 31, 32). The reason remains unexplained but could reflect differences in the sensitivity of different tissues to T3 and T4 despite clinical and biochemical euthyroidism on T4 substitution therapy (33).
A possible explanation for the decreased maximal insulin secretory capacity in the GADab+ patients could be decreased ß-cell mass, which is supported by the almost dose-dependent relationship between the insulin-response and the strength of the stimulus employed. Further, the association between autoimmune markers and decreased ß-cell mass could be compatible with subclinical insulitis. Of note, we have previously shown that nondiabetic GADab+ patients with autoimmune polyendocrine syndrome 1 (APS1) had lower fasting C-peptide concentrations and a lower first phase insulin response to iv glucose than GADab negative patients, despite the fact that only a minority tend to develop clinical diabetes (34). The nondiabetic APS1 patients also showed T cell proliferation in response to GAD (35). Only histological evidence could prove the hypothesis on subclinical insulitis, but such data are scarce. In the only reported study, no signs of insulitis could be demonstrated in the postmortem pancreata of three nondiabetic ICA and GADab positive APS patients who had died because of adenocarcinoma or kidney failure (36).
The fact that the insulin secretion decreased by half in one GADab+ subject who was prospectively followed by glucose-arginine tests, and that more GADab+ than GADab- AT patients tended to develop diabetes during the short follow-up, supports an association between autoimmunity to GAD and the development of diabetes in AT. Although the patients did not require insulin therapy at diagnosis, we do not know whether they will become insulin requiring later. We have previously shown a positive predictive value of 27% and a negative predictive value of 91% for GADab and development of type 1 diabetes in APS1 patients (34). Also, between 14 to 40% of ICA+ patients with nonspecified organ-specific autoimmune disease developed type 1 diabetes, whereas only 2.7% of ICA- patients did (37).
Although the coexistence of autoimmune thyroid disease and type 1 diabetes is widely acknowledged, data on the incidence or prevalence of diabetes or islet autoantibodies in AT are scarce. Hypothyroid AT has been found in 5–15%, Graves' disease in 9.3% (38, 39, 40, 41) and thyroid antibodies in 5–22% (38, 39, 40, 41, 42, 43, 44, 45, 46) of type 1 diabetic patients. In those having both diabetes and thyroid disease, the diagnosis of type 1 diabetes frequently precedes that of hypothyroidism and either succeeds or coincides with thyroiditis (47). Thus, the prevalence of diabetes or islet autoantibodies is dependent upon the age of the study population. We found GADab in 3.4% of 441 nondiabetic Swedish AT patients, which is three times higher than the 1.1% prevalence reported for the Swedish adult population (48) but comparable to previous reports in autoimmune thyroid disease. GADab have been reported in 7.9% (6/76) of Japanese patients with Hashimoto’s disease (49) and in 3.6% (1/26) of German patients with any autoimmune thyroid disease (50). Similarly, 6% (13/212) of Japanese and 13% (11/85) of Swedish patients with Graves' disease had GADab (49, 51).
The GADab+ and GADab- patients also differed in their glucagon response to arginine, whereas the response to glucose was similar. Patients with diabetes show relative hyperglucagonemia to ambient glucose concentration and an exaggerated response to arginine (52, 53, 54). Increased insulin and glucose concentrations are considered to inhibit glucagon secretion from the 
-cells (55). Accordingly, the fact that the glucagon response to glucose was similar in the GADab+ and GADab- patients points to similar -cell sensitivity to glucose in the two groups. Despite the lower insulin response, arginine stimulated glucagon secretion less in GADab+ compared with the GADab- patients, which suggests -cell dysfunction in the GADab+ patients. Autoantibodies to -cells have been described, but they have not been associated with -cell dysfunction, diabetes, or diabetes associated autoantibodies (56).
In conclusion, we found that GADab positivity in nondiabetic patients with thyroiditis was associated with decreased maximal insulin secretory capacity and decreased glucagon response to arginine. This finding supports the hypothesis that GADab is a pancreatic marker for subclinical insulitis, which could lead to insulin deficiency and diabetes.
Acknowledgments
Gertrud Ahlqvist, Marianne Lundberg, Britt Bruveris-Svenburg, Anita Nilsson, and Mona Hansson are acknowledged for skillful technical assistance and our colleagues at the Department of Endocrinology at Malmö University Hospital, Sweden, for recruiting subjects. The IAA measurements were performed in the laboratory of Dr. Mikael Knip, University of Oulu, Finland.
Footnotes
This study was financially supported by the Påhlsson Foundation, the Medical Faculty of the Lund University, the Malmö University Hospital, the Swedish Society of Medicine, the Crafoord Foundation, the Swedish Medical Doctors Association, and the Swedish Diabetes Association.
Abbreviations: ACR, Acute C-peptide response; AGR, acute glucagon response; AIR, acute insulin response; APS, autoimmune polyendocrine syndrome; AT, autoimmune thyroiditis; BMI, body mass index; CV, coefficient of variation; fS, fasting serum; GADab, glutamic acid decarboxylase antibodies; HLA, human leukocyte antigen; HOMA, homeostasis model assessment; IAA, insulin antibodies; IA-2ab, protein tyrosine phosphatase-like protein antibodies; ICA; islet cell antibodies; IVGTT, iv glucose tolerance test; LADA, latent autoimmune diabetes in adults; OGTT, oral glucose tolerance test; RU, relative units.
Received March 7, 2001.
Accepted December 11, 2001.
References
cell autoantibodies and glucagon response to arginine. Diabetes 33:435–437[Abstract]
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
; n = 13) patients with thyroiditis. Shaded areas represent S-insulin response (mean ± SEM) in control subjects (n = 13). *, P = 0.026 GADab+ vs. GADab- thyroiditis.
; n = 13). AIR and AGR were calculated as the mean of the 3–5 min (3–4 min for glucagon) values after subtraction of the mean of the prestimulus values. Data are shown as mean ± SEM. *, P = 0.040 GADab+ vs. GADab- thyroiditis. **, P = 0. 034 GADab+ vs. GADab-thyroiditis.