Skip to main content
NCBI home page
Immunology logoLink to Immunology
. 2002 Aug;106(4):584–589. doi: 10.1046/j.1365-2567.2002.01449.x

Evidence from twins for acquired cellular immune hyperactivity in type 1 diabetes

Nikolai Petrovsky *,†, Kirsten O Kyvik ‡, Vagn Bonnevie-Nielsen §, Henning Beck-Nielsen ‡, Anders Green §, Leonard C Harrison *
PMCID: PMC1782755  PMID: 12153522

Abstract

Type 1 diabetes has been associated with an increased frequency of activated T cells and T-cell hyperactivity to non-specific and disease-specific stimuli including the islet autoantigen glutamic acid decarboxylase 65 (GAD). To address whether T-cell hyperactivity is genetic or acquired we measured whole blood cytokines in vitro in response to GAD or tetanus in 18 identical twin pairs, nine discordant for type 1 diabetes. In addition, the activity of 2′, 5′ oligoadenylate synthetase (OAS) in blood mononuclear cells was measured as a marker of viral infection. Interleukin-2 (IL-2) basally and IL-2 and interferon-γ (IFN-γ) in response to GAD, were detected more frequently and at higher levels in diabetic compared to non-diabetic twins. IL-10 was not different between groups. OAS activity was increased in diabetic compared to non-diabetic twins and showed a correlation with basal IL-2 and GAD-stimulated IFN-γ and IL-10. These findings suggest that T-cell hyperactivity in type 1 diabetes is an acquired trait and could reflect persisting virus expression.

Introduction

Discordance for type 1 diabetes in the majority of identical twins1–3 implies that environmental agents as well as genes contribute to pathogenesis. Epidemiological evidence is consistent with a pathogenic role for enteroviruses4 and rotavirus,5 but the only virus unequivocally associated with human type 1 diabetes is rubella acquired in utero.6 Evidence for persisting virus expression is suggested by the finding of increased activity of the virus- and interferon α/β-inducible enzyme 2′5′ oligoadenylate synthase (OAS) in blood mononuclear cells of people with type 1 diabetes.7 People with type 1 diabetes and prediabetic identical twins have been shown to have an increased frequency of activated T cells8–12 and T-cell hyperactivity to the islet autoantigen glutamic acid decarboxylase (GAD) as well as to non-specific stimuli.13 However, it is unknown whether this state of T-cell hyperactivity is genetic or acquired. The answer could lead to further insights into pathogenesis. Amino acid sequence similarities between GAD and Coxsackie viruses14–16 or rotavirus17,18 raise the possibility that T-cell hyperactivity to GAD could be acquired by molecular mimicry with viruses.

As an experiment of nature, identical twins offer the opportunity to study differences in individuals discordant for disease but with a similar genetic background. Islet cell antibodies (ICA) have been detected in the non-diabetic identical twin discordant for diabetes19 but ICA as well as activated T cells may remit without development of diabetes.20,21 Although the T-cell repertoire is subject to epigenetic influences, no major differences have yet been found in T-cell receptor genes in identical twins discordant for type 1 diabetes.22 Thus, it is reasonable to conclude that differences in disease expression between identical twins can be due to environmental agents. To investigate whether T-cell hyperactivity in type 1 diabetes is genetic or acquired, we analysed whole blood cytokine responses to GAD and tetanus control antigen, and the activity of blood mononuclear cell OAS, in identical twins concordant and discordant for disease.

Materials and methods

Subjects

Eighteen twin pairs in the Danish Twin Registry3 were studied: seven were concordant and 11 discordant for type 1 diabetes, giving a pair-wise concordance rate of 39% (Table 1). The median duration of diabetes was 13 years (range 2–32 years). Two of the 11 non-diabetic twins had impaired glucose tolerance and for the purposes of analysis were included in the group with diabetes. Thus, there was a total of 27 diabetic and nine non-diabetic twins, nine concordant and nine discordant. Seven of the nine discordant twins had been discordant for > 5 years. Subjects gave informed consent to the study, which was approved by the human research ethics committee. No twin reported symptoms of intercurrent infection at the time blood was taken for T-cell assay.

Table 1.

Clinical and laboratory data in twins

Whole blood plasma
GAD65 10 mg/ml Tetanus 10 L.f.u./ml
Twin pair number Status* Diabetes duration (years) GADAb N <5 U/ml IA-2Ab N <2 U/ml ICA (N < 10 JDFU) IAA (N < 40 nU/ml) Basal IL-2 IL-2 IFN-γ IL-10 IL-2 IFN-γ IL-10 PBMC OAS
1.1 D 17 0 17 NA NA 25 868 0 0 NA 3 NA NA
1.2 D 21 0 2 NA NA 59 58 0 NA NA 0 NA NA
2.1 D 2 29 0 80 30 35 136 0 10 0 1 24 3·6
2.2 IGT 2 18 91 85 1 0 186 77 863 NA 35 NA 4·0
3.1 D 5 50 12 9 3 0 520 52 27 59 604 41 7·3
4.1 D 29 0 4 0 34 98 214 258 190 37 30 107 4·4
5.1 D 5 6 53 12 0 78 354 0 296 72 115 30 4·6
6.1 D 8 16 0 6 14 0 32 0 340 0 102 24 4·3
7.1 D 16 69 0 7 10 108 214 0 559 102 12 25 3·8
7.2 D 12 11 0 15 6 60 96 0 81 70 52 24 8·8
8.1 D 9 32 0 6 4 0 103 183 2780 NA 24 NA 6·8
8.2 D 20 2 1 0 2 0 440 626 809 18 31 92 6·4
9.1 D 13 12 92 0 11 138 0 80 332 136 17 117 8·9
9.2 IGT 2 1 0 34 1 207 0 254 267 184 30 215 8·7
10.1 D 9 0 3 NA NA 110 399 58 30 112 62 266 17·6
11.1 D 32 4 0 0 4 0 0 0 30 0 271 79 4·9
12.1 D 19 0 0 0 22 155 71 156 254 158 31 598 8·6
12.2 D 20 0 0 0 29 262 144 239 787 253 53 1079 7·7
13.1 D 17 5 1 0 1 35 180 0 172 27 752 19 NA
13.2 D 21 8 2 0 2 33 150 100 90 68 910 19 NA
14.1 D 8 55 1 0 2 159 100 1344 942 159 418 1432 5·5
14.2 D 16 8 3 0 1 324 355 262 1194 348 77 198 8·2
15.1 D 3 67 0 0 12 39 0 192 77 67 382 32 5·2
15.2 D 6 64 0 0 3 0 0 245 85 0 367 74 4·3
16.1 D 11 0 0 0 NA 0 4 278 531 0 50 67 13·6
17.1 D 13 57 0 0 1 0 122 0 288 0 NA 45 2·7
18.1 D 16 0 0 0 6 0 0 0 90 0 0 21 NA
Mean 13·0 19·0 10·4 10·6 8·6 71·30 175·8 163·1 445·0 81·30 170·2 201·2 6·80
SD 8·00 24·23 25·66 23·46 10·35 87·66 202·7 277·2 589·5 91·38 251·5 360·3 3·49
3.2 ND 0 0 0 1 0 0 0 42 0 369 33 6·1
4.2 ND 0 0 0 1 20 74 0 139 0 104 74 2·1
5.2 ND 1 0 0 1 0 53 0 644 0 245 24 3·5
6.2 ND 2 0 0 1 0 0 0 247 0 59 23 3·9
10.2 ND 0 0 NA NA 0 0 0 77 0 19 43 4·9
11.2 ND 0 0 0 0 0 0 0 14 0 247 54 3·4
16.2 ND 12 11 16 0 94 227 621 745 61 89 110 11·0
17.2 ND 0 0 0 1 66 0 37 49 0 NA 37 3·2
18.2 ND 0 0 0 0 0 7 0 104 0 84 23 NA
Mean 20·0 40·1 73·1 229·0 6·8 151·9 46·8 4·8
SD 35·34 75·32 205·82 273·69 20·33 121·00 29·02 2·80
Open in a new tab

D=diabetic; IGT=impaired glucose tolerant; ND = non-diabetic; N=normal range; NA = not assayed.

Blood sampling and whole blood assay

A heparinized venous blood sample was obtained from each twin pair and aliquots (260 µl/well in quadruplicate) immediately transferred into 96-well tissue culture plates (Nunc, Roskilde, Denmark), preloaded with or without GAD (10 µg/ml) or tetanus toxoid (10 Lyons flocculating units [L.f.u.]/ml) as control antigen. After incubation for 48 hr at 37° in 5% CO2, 95% air, the plasma supernatant was removed from each well, and quadruplicates pooled and stored at −20° for subsequent measurement of cytokines. Blood mononuclear cells were purified by centrifugation of blood on Ficoll-Hypaque and 5 × 106 cells frozen at −70° for measurement of OAS. Plasma was stored at −20° for measurement of islet antibodies.

Antigens

Tetanus toxoid (2780 L.f.u./ml) was a gift from CSL, Melbourne Australia. cDNA for GAD65 was cloned into a hexahistidine vector and expressed in Escherichia coli. The protein was extracted from E. coli in 6 m guanidine–HCl for 1 hr at room temperature followed by centrifugation at 100 000 g for 30 min. The supernatant was adjusted to pH 8·0 with sodium hydroxide and applied to a nickel agarose column pre-equilibrated in 6 m guanidine–HCl, pH 8·0. The bound protein was equilibrated into 8 m urea with 10 column volumes prior to elution with stepwise pH decrements of 6·3, 5·9 and 4·5 in 8 m urea. Recombinant protein was recovered at the pH 4·5 step. Removal of urea by dialysis against 50 mm KH2PO4 pH 7·2, resulted in precipitation of protein, which was solubilized in sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) sample buffer and applied in 1 ml at 4 mg/ml to a SDS–PAGE Bio-Rad 491 preparative cell under reducing conditions. Protein was recovered by electroelution from the bottom of the gel and demonstrated to be essentially pure by migration as a single band after analytical one-dimensional SDS–10% PAGE and silver staining. The endotoxin content of the final preparation determined by the Limulus lysate assay (BioWhittaker, Walkersville, MD) was < 5 IU/mg GAD/ml.

Cytokines

Enzyme-linked immunosorbent assay (ELISA) kits were used to measure interleukin-2 (IL-2; Amersham International, Amersham, UK) and interferon-g (IFN-γ; CSL Melbourne). In the IL-10 ELISA, monoclonal antibody (mAb) JES3-9D7 was used for capture and biotinylated mouse mAb JES3-12G8 for detection. Both antibodies were from PharMingen (San Diego, CA). JES3-9D7 (5 µg/ml) was coated overnight at 4° onto Nunc Maxisorb plates. The plates were then washed with phosphate-buffered saline (PBS) and blocked with 10% bovine serum albumin in PBS for 1 hr at room temperature. Samples (50 µl) were added to wells and incubated overnight at 4° followed by washing and incubation with JES3-12G8 (1 µg/ml) for 1 hr at room temperature. After washing, streptavidin–peroxidase, 100 µl 1 : 500 in PBS, was added for 1 hr. The wells were washed again, then incubated with 100 µl tetramethylbenzadine peroxidase substrate and the reaction stopped after 30 min by addition of 100 µl of 1 m phosphoric acid. Colour development was read at 450 nm. The lower limits of detection for IL-2, IFN-γ and IL-10 were, respectively, 25 pg/ml, 20 pg/ml and 100 pg/ml.

OAS assay

OAS activity was measured in homogenates of blood mononuclear cells as the rate of incorporation of 32P-labelled ATP (U/min/ml) into diester bonds of newly synthesized oligoadenylate dimers and trimers detected by thin layer chromatography, as previously described.7

Islet antibodies

ICA were detected by indirect immunofluorescence on frozen sections of human group O pancreas and calibrated with JDF serum standards. GAD antibodies and IA-2 antibodies were measured by precipitation of 35S-methionine-labelled recombinant proteins generated in an in vitro transcription-translation system (Promega Corp., Annandale, NSW, Australia). IAA were measured by precipitation of 125I-insulin. Assays achieved optimal performance in all International Workshops and have been described previously.23

Statistics

Group frequencies were compared with Fisher's exact test and group means with the Mann–Whitney test (95% confidence intervals). Discordant twin pairs were compared with the Wilcoxon matched pairs test. Intra-twin correlation was determined with the Spearman rank test. Significance was defined as P < 0·05.

Results

Clinical and laboratory data are summarized in Table 1.

Basal plasma cytokines

In the absence of antigen, IL-2 was detected in whole blood plasma in 17/27 (63%) of diabetic compared to 2/9 (22%) non-diabetic twins (P = 0·04), and at higher levels in the diabetic than non-diabetic twins (mean ± SD: 71 ± 88 versus 20 ± 35 pg/ml, P = 0·05) (Fig. 1). Basal plasma IFN-γ and IL-10 were undetectable in all twins.

Figure 1.

Figure 1

Open in a new tab

IL-2 production during 48 hr in whole blood from non-diabetic (ND) and diabetic (D) twins.

Cytokine responses to antigens

In response to GAD, IL-2 was detected in 20/27 (74%) of diabetic compared to 3/9 (33%) non-diabetic twins (P = 0·04), and at higher levels in the diabetic than non-diabetic twins (176 ± 203 versus 40 ± 75 pg/ml, P = 0·02) (Fig. 2). Likewise, in response to GAD, IFN-γ was detected in 16/27 (59%) of diabetic compared to 1/9 (11%) non-diabetic twins (P = 0·02), and at higher levels in the diabetic than non-diabetic twins (163 ± 277 versus 73 ± 206 pg/ml, P = 0·04) (Fig. 2). IL-10 was detected in a similar proportion of diabetic (16/25, 64%) and non-diabetic (5/9, 56%) twins and was not different between groups (Fig. 2).

Figure 2.

Figure 2

Open in a new tab

Cytokine production during 48 hr in whole blood from non-diabetic (ND) and diabetic (D) twins, in the presence of glutamic acid decarboxylase 65 (GAD 65) at 10 µg/ml.

In response to tetanus, IL-2 was detected in 15/23 (65%) of diabetic compared to 1/9 (11%) non-diabetic twins (P = 0·01) (Fig. 3). However, compared to basal, IL-2 responses to tetanus were not altered in either group, indicating that the tetanus concentration used may have been suboptimal. In response to tetanus, IFN-γ was detected in similar proportions of diabetic (16/26, 62%) and non-diabetic (7/8, 88%) twins, and was not different between groups; similarly IL-10 was detected in 8/23 (35%) of diabetic compared to 1/9 (11%) non-diabetic twins and was not different between groups (Fig. 3).

Figure 3.

Figure 3

Open in a new tab

Cytokine production during 48 hr in whole blood from non-diabetic (ND) and diabetic (D) twins, in the presence of tetanus toxoid at 10 L.f.u./ml.

There were no apparent relationships between levels of cytokines in response to GAD or tetanus and the duration of diabetes, or between GAD antibodies and the levels of cytokines in response to GAD.

Cytokine responses in discordant twin pairs

Pair-wise comparisons between diabetic and non-diabetic twins for each cytokine failed to reach significance.

OAS

The activity of OAS was higher in diabetic (n = 22) than non-diabetic (n = 7) twins (6·8 ± 3·5 versus 4·8 ± 2·8, P = 0·02). In eight discordant twin pairs, OAS activity was higher in the diabetic (7·4 ± 5·3) than non-diabetic (4·8 ± 2·8) twin but in pair-wise comparison failed to reach statistical significance (P = 0·057). The intratwin pair correlation (r) of OAS activity was 0·7 (P = 0·03). There were significant correlations between OAS activity and basal IL-2 (r = 0·36; P = 0·02) and IFN-γ (r = 0·28; P = 0·04) and IL-10 (r = 0·34; P = 0·04) responses to GAD, consistent with a relationship between OAS production and T-cell hyperactivity.

Discussion

Diabetic twins had higher basal levels of IL-2 and higher levels of IL-2 and IFN-γ in response to GAD, than non-diabetic twins. Furthermore, basal IL-2, IL-2 responses to GAD and tetanus, and IFN-γ responses to GAD, were significantly more frequent in diabetic than non-diabetic twins. Although pairwise comparison between diabetic and non-diabetic twins for cytokine and OAS levels failed to just reach statistical significance, this may have reflected the small number of discordant twin pairs available for study. IL-10 responses to GAD or tetanus were not significantly different in magnitude or frequency between diabetic and non-diabetic twins. The possibility that stimulation by recombinant GAD was a reuslt of contamination by bacterial components cannot be totally excluded. However, this seems unlikely because the amount of endotoxin associated with GAD was negligible and non-specific stimulation should have occurred with all twins, not just those with diabetes.

Type 1 diabetes has previously been associated with cellular immune hyperactivity, reflected by increased expression of activation markers8–12 and a bias towards T helper 1 (Th1) cytokine secretion by peripheral blood T cells in response to non-specific stimuli.24 In a longitudinal study of individuals at risk for type 1 diabetes, T-cell proliferative responses to both GAD and tetanus increased in the 24 months prior to clinical diagnosis.13 The finding in the present study of higher T-cell cytokine responses in the diabetic twins supports the view that the cellular immune hyperactivity observed in type 1 diabetes is an acquired trait.

The only non-diabetic twin with substantial IL-2 and IFN-γ responses to GAD (#16.2) was ICA-positive and postitive for GAD and IA-2 antibodies, and had the high risk human leucocyte antigen (HLA) DR3,4 haplotype. Islet antibodies in non-diabetic twins denote high risk for diabetes19 and DR3,4 is more prevalent in concordant than discordant twin pairs.25 Therefore, this twin was at high risk for diabetes and his result is consistent with the cellular immune hyperactivity observed to precede clinical diagnosis.11,13 His results detracted from the statistical significance of group and discordant twin pair-wise comparisons but, ironically, are in keeping with our overall conclusion.

The finding of increased activity of OAS in diabetic compared to non-diabetic twins extends a previous report7 in which OAS activity was found to be increased in individuals with diabetes compared to controls. As OAS is induced directly by double-stranded RNA viruses such as rotavirus or by virus-induced α and β interferons, and correlated with T-cell cytokine production, we suggest that the acquisition of cellular immune hyperactivity in type 1 diabetes could reflect persisting virus infection.

Acknowledgments

We thank the Danish Twin Registry and Charlotte Fage-Larsen, Department of Endocrinology, Odense Hospital, Odense, for her assistance with these studies. Islet cell antibodies were measured by Dr Peter Colman, Royal Melbourne Hospital. N.P. and L.C.H. were supported by the National Health and Medical Research Council of Australia. Catherine O'Shea provided secretarial assistance.

References

  • 1.Barnett AH, Eff C, Leslie RD, Pyke DA. Diabetes in identical twins. A study of 200 pairs. Diabetologia. 1981;20:87–93. doi: 10.1007/BF00262007. [DOI] [PubMed] [Google Scholar]
  • 2.Kumar D, Gemayel NS, Deapen D, et al. North-American twins with IDDM. Genetic, etiological, and clinical significance of disease concordance according to age, zygosity, and the interval after diagnosis in first twin. Diabetes. 1993;42:1351–63. doi: 10.2337/diab.42.9.1351. [DOI] [PubMed] [Google Scholar]
  • 3.Kyvik KO, Green A, Beck-Nielsen H. Concordance rates of insulin dependent diabetes mellitus. A population based study of young Danish twins. Br Med J. 1995;311:913–7. doi: 10.1136/bmj.311.7010.913. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Hiltunen M, Hyoty H, Knip M, et al. Islet cell antibody seroconversion n children is temporally associated with enterovirus infections. Childhood Diabetes in Finland (DiMe) Study Group. J Infect Dis. 1997;175:554–60. doi: 10.1093/infdis/175.3.554. [DOI] [PubMed] [Google Scholar]
  • 5.Honeyman MC, Coulson BS, Stone NL, et al. Association between rotavirus infection and pancreatic islet autoimmunity in children at risk of developing type 1 diabetes. Diabetes. 2000;49:1319–24. doi: 10.2337/diabetes.49.8.1319. [DOI] [PubMed] [Google Scholar]
  • 6.Menser AM, Forrest JM, Honeyman MC, Burgess JA. Diabetes, HLA antigens and congenital rubella. Lancet. 1974;2:1508–9. doi: 10.1016/s0140-6736(74)90240-2. [DOI] [PubMed] [Google Scholar]
  • 7.Bonnevie-Nielsen V, Martensen PM, Justesen J, et al. The antiviral 2′5′-oligoadenylate synthetase is persistently activated in type 1 diabetes. Clin Immunol. 2000;96:11–8. doi: 10.1006/clim.2000.4874. [DOI] [PubMed] [Google Scholar]
  • 8.Alviggi L, Johnston C, Hoskins PJ, Tee DE, Pyke DA, Leslie RD, Vergani D. Pathogenesis of insulin-dependent diabetes: a role for activated T lymphocytes. Lancet. 1984;2:4–6. doi: 10.1016/s0140-6736(84)91994-9. [DOI] [PubMed] [Google Scholar]
  • 9.Hitchcock CL, Riley WJ, Alamo A, Pyka R, Maclaren NK. Lymphocyte subsets and activation in prediabetes. Diabetes. 1986;35:1416–22. doi: 10.2337/diab.35.12.1416. [DOI] [PubMed] [Google Scholar]
  • 10.Tun RY, Peakman M, Alviggi L, et al. Importance of persistent cellular and humoral immune changes before diabetes develops: prospective study of identical twins. Br Med J. 1994;308:1063–8. doi: 10.1136/bmj.308.6936.1063. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Peakman M, Leslie RD, Alviggi L, Hawa M, Vergani D. Persistent activation of CD8+ T-cells characterizes prediabetic twins. Diabetes Care. 1993;19:1177–84. doi: 10.2337/diacare.19.11.1177. [DOI] [PubMed] [Google Scholar]
  • 12.Petersen LD, Duinkerken G, Bruining GJ, van Lier RA, de Vries RR, Roep BO. Increased numbers of in vivo activated T cells in patients with recent onset insulin-dependent diabetes mellitus. J Autoimmunity. 1996;9:731–7. doi: 10.1006/jaut.1996.0095. [DOI] [PubMed] [Google Scholar]
  • 13.Honeyman MC, Stone N, DeAizpurua H, Rowley MJ, Harrison LC. High T-cell responses to the glutamic acid decarboxylase (GAD) isoform 67 reflect a hyperimmune state that precedes the onset of insulin-dependent diabetes. J Autoimmunity. 1997;10:165–73. doi: 10.1006/jaut.1996.0124. [DOI] [PubMed] [Google Scholar]
  • 14.Atkinson MA, Bowman MA, Campbell L, Darrow BL, Kaufman DL, Maclaren NK. Cellular immunity to a determinant common to glutamate decarboxylase and coxsackie virus in insulin-dependent diabetes. J Clin Invest. 1994;94:2125–9. doi: 10.1172/JCI117567. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Jones DB, Crosby I. Proliferative lymphocyte responses to virus antigens homologous to GAD65 in IDDM. Diabetologia. 1996;39:318–24. doi: 10.1007/s001250050576. [DOI] [PubMed] [Google Scholar]
  • 16.Schloot NC, Roep BO, Wegmann DR, Yu L, Wang TB, Eisenbarth GS. T-cell reactivity to GAD65 peptide sequences shared with coxsackie virus protein in recent-onset IDDM, post-onset IDDM patients and control subjects. Diabetologia. 1997;40:332–8. doi: 10.1007/s001250050683. [DOI] [PubMed] [Google Scholar]
  • 17.Honeyman MC, Stone NL, Harrison LC. T-cell epitopes in the type 1 diabetes autoantigen tyrosine phosphatase IA-2: potential for mimicry with rotavirus and other environmental agents. Mol Med. 1998;4:231–9. [PMC free article] [PubMed] [Google Scholar]
  • 18.Honeyman MC, Brusic V, Stone N, Harrison LC. Neural network-based prediction of candidate T-cell epitopes. Nature Biotech. 1998;16:966–70. doi: 10.1038/nbt1098-966. [DOI] [PubMed] [Google Scholar]
  • 19.Johnston C, Millward BA, Hoskins P, Leslie RD, Bottazzo GF, Pyke DA. Islet-cell antibodies as predictors of the later development of type 1 (insulin-dependent) diabetes. A study in identical twins. Diabetologia. 1989;32:382–6. doi: 10.1007/BF00277263. [DOI] [PubMed] [Google Scholar]
  • 20.Heaton DA, Millward BA, Gray P, Tun Y, Hales CN, Pyke DA, Leslie RD. Evidence of beta cell dysfunction which does not lead on to diabetes: a study of identical twins of insulin dependent diabetics. Br Med J. 1987;294:145–6. doi: 10.1136/bmj.294.6565.145. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Millward BA, Alviggi L, Hoskins PJ, et al. Immune changes associated with insulin dependent diabetes may remit without causing the disease: a study in identical twins. Br Med J. 1986;292:793–6. doi: 10.1136/bmj.292.6523.793. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Malhotra U, Spielman R, Concannon P. Variability in T cell receptor V beta gene usage in human peripheral blood lymphocytes. Studies of identical twins, siblings, and insulin-dependent diabetes mellitus patients. J Immunol. 1992;149:1802–8. [PubMed] [Google Scholar]
  • 23.Colman PG, McNair P, Margetts H, et al. The Melbourne Pre-Diabetes Study: prediction of type 1 diabetes mellitus using antibody and metabolic testing. Med J Aust. 1998;169:81–4. doi: 10.5694/j.1326-5377.1998.tb140188.x. [DOI] [PubMed] [Google Scholar]
  • 24.Kallmann BA, Huther M, Tubes M, Feldkamp J, Bertrams J, Gries FA, Lampeter EF, Kolb H. Systemic bias of cytokine production toward cell-mediated immune regulation in IDDM and toward humoral immunity in Graves' disease. Diabetes. 1997;46:237–43. doi: 10.2337/diab.46.2.237. [DOI] [PubMed] [Google Scholar]
  • 25.Johnston C, Pyke DA, Cudworth AG, Wolf E. HLA-DR typing in identical twins with insulin-dependent diabetes: difference between concordant and discordant pairs. Br Med J. 1983;286:253–5. doi: 10.1136/bmj.286.6361.253. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Immunology are provided here courtesy of British Society for Immunology

RESOURCES