Search

The present studies on the possible factors that can be implicated in the etiopathogenesis of diabetes are reviewed. The association between insulin-dependent diabetes and certain specifications of the HLA system are commented on, as well as the experimental studies and the clinical observations which indicate the possible relationship between this disease and some viruses. The role played by autoimmunity is thoroughly discussed, specially the significance of anti-pancreatic islet antibodies. Finally, new pathogenic theories and classifications of the disease are proposed.

The normally functioning immune system is subject to intricate networks of regulatory mechanisms: it is therefore not surprising to find that autoimmune diseases present a complex pathogenic picture in which the relative contributions of various factors probably determine the precise nature and course of disease. This is particularly evident in the effector mechanisms of organ-specific autoimmunity which are described in this chapter. These ultimately give rise to the disease symptoms, and can be directly cytotoxic, or may either stimulate or block functional activity or growth of the target cells. Their various contributions to human diseases are becoming more firmly established, as in Type I diabetes, or are only now being described, as in the case of EC-Ab in protracted diarrhea of infancy and as evidenced by the growing lists of receptor-stimulating or -blocking antibodies. The nature and precise location of relevant autoantigens is also coming under closer scrutiny. The answers to the question of why these diseases arise in the first place remain more elusive. However, it is again likely that a variety of factors can contribute. The attractive possibility of a role for idiotypic interactions is gaining ground, particularly within the context of antibodies to hormones and their receptors. Another potential mechanism which we believe may be of central importance, particularly in the development of organ-specific destructive autoimmunity, and which we have discussed here in detail, is the aberrant expression of HLA Class II molecules by target cells. Whether this is actually an initiating factor is presently not known, but its potential for promoting pathogenesis both early and late in the process is clear. Furthermore, the complex nature of the regulation of epithelial Class II expression may help to explain the heterogeneity of features and course of disease in different patients with the same underlying pathology. All these advances in our basic understanding of the disease processes should ultimately lead to more effective and specific means of therapeutic intervention.

By transfecting HLA class II-negative thyroid cells with the plasmid px-8 containing the SV-40 early region, we obtained a continuous line of thyrocytes which maintains most of the characteristics of the parental cells. After a period of 4 months, the transfected thyrocytes showed an increased surface expression of class I molecules and a de novo expression of class II products. Different clones derived from this thyroid cell line varied in their levels of constitutive class II basal expression and also in their sensitivity to interferon-gamma (IFN-β). This phenomenon constitutes a novel mechanism of HLA class II product induction and may be relevant to explain the in vivo inappropriate class II expression described in the autoimmune and neoplastic diseases of the thyroid.

When, in 1973, one of us (GFB) joined Deborah Doniach and the Autoimmunity Group in the Department of Immunology at The Middlesex Hospital in London, the main area of research was the pathogenesis of autoimmune liver diseases. Indeed, it is of interest to recall that another young Italian research fellow in the team at that time was Mario Rizzetto who, a few years later, discovered the Delta particle, and thus made a fundamental advance in the understanding of viral hepatitis. No doubt he had the right training! But to return to the story of the young fellow from Venice, he knew that the Department was famous all over the world for the investigation of endocrine autoimmunity and he was hesitant to become involved in the ‘diversion’ of liver disorders. There is nothing wrong with this subject but, after all, Ivan Roitt and Deborah Doniach were the discoverers of human thyroid autoimmunity 17 years previously and had persued this line of novel investigation for several years. ‘What about turning the emphasis again towards the first ‘love’ of endocrine autoimmunity?’ he asked. Deborah Doniach enthusiastically agreed. It was in 1974 when he then tested a few sera of patients with Addison’s disease on sections of human pancreas by the indirect immunofluorescence technique, that he observed the islets shining out against the dark background of the exocrine acini. With Alex Florin-Christensen, who actively collaborated in this new approach, the first dogma was thus challenged: a year previously during the first meeting on ‘Immunology of Diabetes’ held in Brussels, it was stated: ‘We are afraid that islet cell antibodies do not exist’.

Evidence accumulated during the past decade has progressively consolidated the concept that type I (insulin-dependent) diabetes is a chronic, destructive, organ-specific autoimmune disorder [1, 2]. In general, the classical theories on the mechanisms of initiation and perpetuation of organ-specific autoimmunity have been considered in relation to the pathogenesis of beta-cell destruction, but none of them has so far given the final and decisive answer [3]. Nevertheless, we believe that all the pieces that will constitute the jigsaw puzzle of the pathogenesis of type I diabetes can now be put into place.

Physiologically, HLA Class II molecules are primarily expressed on immunoregulatory cells. In recent years, it has been shown that tissues affected by autoimmunity, including beta cells of 'diabetic' pancreases, 'inappropriately' express these glycoproteins. Cytokines are potent Class II inducers on a variety of cells but they exert a heterogenous effect when incubated with different human endocrine cells, i.e. thyroid cells are readily inducible, beta cells are much more resistant. The fine modulation of Class II expression by cytokines has been extensively studied and recent information on their action on endocrine cells is reported here. The possibility that distinct environmental factors from those postulated until now may be responsible for triggering the inappropriate expression of MHC molecules in epithelial cells in vivo is also emphasized. Despite several similarities between human Type I diabetes and spontaneous animal models of the disease, major differences still exist. Transgenic models have now been produced. However, even if they offer interesting new insights, they have not so far provided the decisive answer to explain the pathogenesis of human autoimmune diseases.

Islet cell antibodies (ICA) are present in the sera of most patients with Type 1 diabetes at diagnosis and in some of their genetically susceptible, but otherwise normal, first degree relatives. In this study we have investigated basal and stimulated insulin release by mouse islets following preincubation with human sera (with or without the addition of guinea pig complement) belonging to: 15 normal first degree relatives of diabetic probands; 7 patients with Type 1 diabetes; 7 control subjects with no history of diabetes. All sera had been previously screened for conventional (IgG), complement fixing (CF) and surface (S) ICA. Basal insulin release was not altered by any of the sera. The response to stimulus after incubation with ICA negative and IgG-ICA positive (but CF-ICA negative) sera was similar whether complement was present or not. Stimulated insulin release was significantly inhibited by complement and sera from 2 relatives and 3 diabetic patients. These sera were CF-ICA positive, the sera of the 2 relatives being also ICSA positive. One relative developed Type 1 diabetes 14 months later. This study demonstrates for the first time that sera containing CF-ICA and belonging to individuals susceptible to Type 1 diabetes, can impair insulin release in vitro. It is therefore likely that antibody-dependent, complement-mediated mechanisms are involved in the pathogenesis of Type 1 diabetes.

The existence of islet cell antibodies (ICA) strongly supports an autoimmune pathogenetic mechanism for insulin-dependent diabetes. ICA have been determined in a group of 205 diabetics with the following results: ICA were present in 76 percent of insulin-dependent diabetics with symptomatology lasting less than 8 weeks, in 20 percent of the remaining insulin-dependent diabetics, in 5 percent of non-insulin-dependent diabetics, and in 1 percent of the control subjects. Thyroid antimicrosome antibodies and antiparietal-cell antibodies were determined too, but no significant relationship was demonstrated in insulin-dependent diabetics with positive ICA. Under a clinical point of view the comparison between diabetics with and without ICA has been of no value for determining differential characteristics in relation to metabolic seriousness, epidemiological data, and complication rate.

HLA-DR antigens, the human equivalent of mouse I region-associated or Ia products, are polymorphic cell surface sialoglycoproteins involved in initiation of the immune response1. Their expression is normally restricted to B lymphocytes, macrophages, dendritic and other antigen-presenting cells and vascular endothelium2 and possibly some cells of the mucosa lining body cavities3,4. HLA-DR expression can be modified during cell differentiation; B lymphocytes become negative on maturing to plasma cells and human T lymphocytes acquire these antigens when activated in vitro or in vivo5. We report here that human thyroid follicular cells which are normally negative for HLA-DR molecules, can be induced to express these antigens when cultured with phytohaemagglutinin (PHA), concanavalin A (Con A) or pokeweed mitogen (PWM). These lectins exert their action directly on the thyroid cells with no concomitant mitogenic effect.

Organ-specific autoimmunity has been recognised in virtually all the defined endocrine glands, where it is the hormone-producing epithelial cells themselves which are destroyed. The mechanisms leading to this breakdown of self-tolerance are still unknown, although a variety of hypotheses have been proposed (reviewed in 1 and 2). Two requirements which may well be central to these mechanisms are the accessibility of autoantigens to cells of the immune system, and the presentation of such antigens in an immunogenic fashion. In recent years we have performed a variety of experiments relevant to these points, mainly in thyroid autoimmunity, in which there is marked lymphocytic infiltration of the gland, including activated T cells (3). Our results have led to a general model for the generation of organ-specific endocrine autoimmunity. Central to this hypothesis is the ability of endocrine epithelial cells to present their own surface molecules as autoantigens.

Human islets from an adult subject with nesidioblastosis were isolated and used to perform in vitro studies. Isolated nesidioblastotic islets showed an increased basal rate of insulin secretion (nesidioblastotic islets, 81.3 +/- 6.4 vs. control islets, from cadaveric organ normal donors, 10.2 +/- 0.9 microU/islet/90 min) without any further release with increasing glucose concentration. In addition, islets isolated from the pancreas with nesidioblastosis contained more insulin than control islets, 1,547.0 +/- 128.7 vs. 935.0 +/- 51.7 microU/islet, and the levels of insulin mRNA were also higher than those measured in controls. Interestingly, the serum of the subject with nesidioblastosis contained autoantibodies that stained brightly and selectively a population of islet cells whose distribution coincided with that expected of alpha cells. In summary, pancreatic islets in nesidioblastosis display beta-cell functional abnormalities. Moreover, the finding of antibodies against islet cells is a common feature in the serum of nesidioblastotic subjects; nevertheless, their pathological significance warrants further investigation.

Celiac disease is a gluten-induced T-cell mediated autoimmune process that results in the destruction of the intestinal mucosa and is associated with an expansion of CD8 + CD103 + TCRαβ intraepithelial lymphocytes (IELs) in the damaged epithelium. The role of this IEL population in the pathology is unknown. The aim of this work was to compare the cytokine profile and the cytotoxicity pattern from CD8 + IEL clones isolated from celiac (CD) and non-celiac (NCD) biopsies. We report that the number of IL-10 producing CD clones was significantly lower (26%) than that obtained from the NCD sample (62%). Instead, IL-2 was produced by more CD (44%) than NCD clones (26%). Cytotoxicity patterns against intestinal epithelial cell lines suggest different functional subsets of CD8 + IELs. CD clones capable of high cytotoxicity produced IL-2 whereas most cytotoxic NCD IELs produced IL-10. This clonal analysis indicates that an impaired immune regulation in celiac mucosa may be partially attributed to the low generation of regulatory CD8 + IELs that produce IL-10.

The presence of antibodies directed against pancreatic islet cells was investigated in 18 insulin pendent diabetic patients with autoimmune disease (Group 1), in 8 insulin-independent diabetic patients with autoimmune disease (Group 2), in 42 non-diabetic patients with thyroid antibodies (Group 3), in 18 patients with antibodies against gastric parietal cells (Group 4) and in 100 healthy subjects (controls). Islet cell antibodies were found in 46% of group 1 patients and in only 1 patient of group 3 and 1 control. This would suggest the existence of a special form of autoimmune insulin-dependent diabetes characterized by a high incidence of islet cell antibodies and associated with other autoimmune diseases.

HLA Class II molecules are expressed by human thyroid epithelial cells (thyrocytes) in thyroid autoimmunity, although these cells are normally Class II-. gamma-Interferon (gamma-IFN) is probably involved in this expression, as suggested by its ability to induce Class II in cultured normal thyrocytes. We have now found that thyroid stimulating hormone (TSH) enhances Class II expression induced in cultured thyrocytes by gamma-IFN, and effects similar to those of TSH were obtained with dibutyryl cyclic AMP. A proportion of thyrocytes also expressed Class II following treatment with TSH or dibutyryl cyclic AMP in the absence of gamma-IFN, but the optimal activity of these mediators then appeared to be dependent upon the occurrence of some pre-existing Class II expression. These findings give insights into how a variety of mediators may influence Class II expression in thyroid autoimmunity.

Autoimmune thyroid diseases (AITD) and insulin-dependent diabetes mellitus (IDDM) are two autoimmune syndromes of unknown etiology with common immune features. One is that the target cells, thyrocytes and pancreatic islet beta cells respectively, hyperexpress several proteins encoded in the HLA region: HLA class I, HLA class II and transporter associated with antigen processing (TAP-1): the clinical course and many aspects of the immunopathology are, however, quite different. Low-molecular-mass polypeptides 2 and 7 (LMP2 and LMP7) are proteasome subunits that increase the efficiency of endogenous antigen processing and are encoded in close vicinity to the TAP genes. We investigated whether LMP2 and LMP7 are hyperexpressed in thyrocytes and islet cells in AITD and IDDM. Thyroid tissue from Graves' disease patients (GD, n = 8) and Hashimoto thyroiditis (HT, n = 1) and pancreatic tissue from IDDM patients (n = 4) as well as control tissues were examined by the two-color indirect immunofluorescence technique. The results demonstrate that, in normal glands, thyrocytes and pancreatic islet cells express comparable moderate to low levels of LMP2 and LMP7. In AITD and IDDM, expression of LMP2/7 in the endocrine cells was disparate: while in AITD glands there was hyperexpression of LMP2 and 7 parallel to that of HLA class I and TAP-1, in the islet cells of recent onset diabetic pancreases (n = 2) the level of LMP2 and 7 expression was totally normal, including islets that were infiltrated by lymphocytes and hyperexpressed HLA class I and TAP-1. These observations suggest different mechanisms of endogenous peptides generation at the target cells in AITD from IDDM. Since this is a key step for the maintenance of peripheral tolerance, it may help to understand some of the different clinical features of the two autoimmune diseases.

Objectives of this study were to examine the cytokine and perforin mRNA expression in minor salivary glands from patients with primary Sjogren's syndrome (pSS), searching for possible correlation with clinical parameters and to identify the dominant cytokine pattern in the different groups. Oral mucosa biopsy samples from 42 pSS patients were studied. Total RNA was analysed by normalized RT-PCR using oligo-dT as the RT primer and IL-2, IFN-gamma, IL-12, IL-18, IL-4, IL-10, TGF-beta, TNF-alpha and perforin-specific primers for amplification. Results were analysed taking into account: (1) biopsy grade I to IV (Chisholm's classification); (2) diagnosis of either definite pSS (n=30) or probable pSS (n=12), following the European classification criteria (ECC), and (3) length of disease evolution from the beginning of symptoms to the time of biopsy, using an arbitrary cut-off point of 12 months. This studied showed that Th1-related cytokines (IL-2, IFN-gamma, IL-12, IL-18, TNF-alpha) and perforin were present in most samples. IL-4 (Th2) was totally absent but other Th2 and regulatory cytokines (IL-10, TGF-beta) were detected in the majority of samples. No significant differences were found between definite and probable pSS nor between grades II, III, IV and fibrous tissue biopsies. A statistically significant increase of IL-2 (P=0.012) and IFN-gamma (P=0.019) was observed in samples from patients with longer disease evolution, whereas the two Th1-inducer cytokines IL-12 and IL-18 were equally and highly expressed in all samples. In conclusion, a predominant Th1 pattern of cytokines was observed in all pSS samples, irrespective of biopsy classification. In addition, a significant increase of Th1 cytokine expression frequency was associated with longer disease evolution.

Psoriasis is a common inflammatory skin disease with key immunological and genetic components. Recruitment of leukocytes into the skin is a central step in its pathogenesis, and it is mediated by cytokines. Among the cytokines expressed in psoriatic lesions, CCL4 and CCL4L chemokines appear to be pivotal elements for the skin recruitment of proinflammatory cells. The aim of this study is to evaluate the relationship between CCL4L polymorphisms [including Single Nucleotide Polymorphisms (SNPs) and Copy Number Variation (CNV)] and the course and prognosis of psoriasis. We analyzed the CNV and the rs4796195 SNP in 211 psoriatic patients and 234 controls; sera from both populations were also quantified for CCL4/CCL4L protein. Our results showed that a high CNV (≥3 copies) is associated with psoriasis severity, while moderate disease is more frequent in patients with lower CNV (≤2 copies); specifically CCL4L1 allele is more present in patients with severe psoriasis, while CCL4L2 correlates with a milder disease. In addition we found a positive correlation between the CNV and sera protein levels. Our results suggest that CCL4L genotyping could not only allow a better understanding of the psoriatic pathogenesis, but could also be used as a prognostic tool, even helping to modulate the efficacy of treatments.

We recently showed that human thyroid epithelial cells, which are normally negative for HLA-DR molecules, express HLA-DR in thyroid autoimmunity. Furthermore, induction of HLA-DR on normal thyroid cells can be achieved by culture with plant lectins. We have now found that recombinant human interferon-gamma (IFN-gamma) induces expression of HLA-DR molecules on cultured human thyroid cells, whereas Namalva IFN-alpha, recombinant IFN-beta or recombinant interleukin-2 (IL-2) do not. All three IFN, but not IL-2, enhanced thyroid cell HLA-A,B,C expression. The results strongly implicate T cells (which are the source of IFN-gamma) in the aberrant induction of DR on thyroid epithelial cells which is proposed to be a central feature of the immunopathological processes leading to autoimmunity.

SUMMARY One of the paradoxes of insulin-dependent diabetes mellitus is that the destruction of the pancreatic islets’ endocrine cells is restricted to the insulin-producing β cells, whereas the main autoantibodies, islet cell antibodies (ICA), are directed against all endocrine islet cells. GAD has recently been proposed as the main target of the humoral and cellular autoimmune attack to the islets, and since in rat pancreas this enzyme was expressed only in the β cells, this provided an explanation for the cell specificity of the destructive process. The finding of GAD-positive cells in the islets of two diabetic patients, one of whom had completely lost the β cells, led us to study in detail thedistribution of GAD in normal human islet ceils using a panel of GAD aniiscra and the double indirect immunofluorescence technique on cryostat sections, monolayer cultures and cytosmears. The results showed that GAD is present not only in the cytoplasm of β cells but also in 69% of the α and 27% of the δ cells. GAD was not present, however, on the surface of the islet cells. These results suggest that the cellular distribution of GADcan not by itself explain the selectivity of β cell destruction in insulin-dependent diabetes mellitus.

Cytokines are important modulators of immunological reactions, but it has been postulated that they might act on other unrelated epithelial cells. We studied the effects of recombinant interferon-gamma (rIFN gamma) and recombinant tumor necrosis factor-alpha (rTNF alpha) on normal human thyroid cells. We found that the combination of these two cytokines enhanced HLA class II molecule expression on these cells compared with the effect of rIFN gamma alone. This was proven by both immunofluorescence as well as a more sensitive and quantitative RIA. rTNF alpha alone had no effect on HLA class II molecule induction on the same thyrocytes, suggesting a synergistic rather than an additive action in combination with rIFN gamma. The addition of 600 U/ml rTNF alpha to low dose rIFN gamma (10 U/mL) enhanced class II expression by 50%, as quantified by RIA. We also demonstrated that normal thyrocytes possess distinct receptors for the two cytokines and that rTNF alpha probably augments IFN gamma binding, since it increased when the cells were first incubated with rTNF alpha. This increased binding provides an explanation for the synergistic action of rTNF alpha in enhancing class II molecule expression by rIFN gamma. We conclude that the presence of receptors for these cytokines on human thyroid cells gives a direct demonstration of their potential biological action on cells normally not involved in the immunological circuit. The phenomenon might also explain their direct or indirect involvement in vivo, such as in influencing inappropriate HLA class II molecule expression in epithelial cells affected by autoimmunity.

In most of the cases, once a given endocrine gland is involved, the corresponding specific autoantibody may be detected; for example, anti-islet cell antibodies are produced within the first few years, after onset of symptoms in insulin-dependent diabetes (DID). Accompanying autoantibodies are quite frequently found in the patient himself. In Schmidt's syndrome (thyroid and adrenal glands are involved and associated to IDD in 30% of the cases) thyroid microsomal antibodies are found in 38% of the cases, thyroglobulin antibodies in 11% of the cases, islet-cell antibodies in 7% of the cases and steroid cell antibodies in 17% of the cases. Associations are also possible in patient family members. Aberrant expression of HLA-DR molecules at the membrane of follicular thyroid cells (as of any other endocrine gland), following a viral aggression, could well account for the endocrinopathy combinations, but alternative mechanisms should be discussed.

SUMMARY A new bioassay is described for detecting the growth stimulating immunoglobulins (TGI)* that contribute to goitre formation in human thyroid autoimmune diseases. It measures the incorporation of tritiated thymidine into intact rat thyroid follicles grown in tissue culture. This radiometric assay demands much less technical skill than the cytochemical bioassays (CBA) previously employed. It has good reproducibility and the techniques and apparatus are available in many clinical laboratories. Immunoglobulins (Igs) from 68% of patients with goitrous GRAVES' disease were positive, in proportion with goitre size, and this showed no correlation with T3 levels, or three accepted methods for conventional thyroid stimulating antibodies. Non-toxic nodular goitre cases gave positive results in 3/9 who had recurrences after one or more thyroidectomies and in 1/10 cases of familial simple goitre. All normal subjects and all endemic goitre cases were negative as well as 21 cases of sporadic non-toxic nodular goitre. Although it is less sensitive than the ‘growth CBA’ it clearly emphasizes the essential difference between the intensity of growth stimulus which leads to the regular hyperplasia of thyroid epithelium seen in GRAVES' thyrotoxicosis and the disorganized and metabolically uncoordinated hyperplasia typical of non-toxic nodular goitre.

By definition, autoimmunity is a circumvention of self-tolerance. Considering the complex network of control mechanisms which initiate and regulate immune responses to foreign agents and maintain the natural state of self-tolerance, autoimmunity may result from a weakening of this delicate equilibrium at various levels. Several hypotheses have therefore been put forward to explain the 'active' immunization process against self-constituents which are normally tolerated.

Immune responses are initiated by HLA-DR+ cells, which present antigen to T cells. Observations that HLA-DR may be experimentally induced on thyroid epithelium and that HLA-DR occurs on thyrocytes in autoimmune thyroid diseases suggest a mechanism of autoimmunity with special relevance to organ-specific diseases. This involves the local aberrant expression of HLA-DR antigens by epithelial cells and their subsequent capacity to present autoantigens occurring on their surfaces to T lymphocytes. For autoantigens which T cells recognise infrequently because of their restricted tissue location and low concentration in the circulation, T-cell tolerance is unlikely, and so induction of autoreactive T cells would occur. Because interferon is the best known inducer of DR antigen expression and viral infections may predate endocrine autoimmunity, the following sequence seems likely: local viral infection which causes interferon production, or other local environmental factors which would induce DR expression, presentation of autoantigens, and subsequent autoimmune T-cell induction. These T cells would activate effector B and T cells. Whether the initial induction of autoimmune T cells leads to autoimmune disease would depend on factors such as abnormalities of the suppressor T-cell pathway, reported to coexist with autoimmunity and necessary to induce autoimmune disease in mice. This mechanism of autoimmune disease induction explains vague associations with viral infections and long latency periods before disease becomes manifest and gives a simple explanation for the well-documented association between HLA-DR and autoimmune diseases in man.

A model for the possible pathogenesis of insulin-dependent diabetes mellitus (IDDM) is presented. It is partly based on studies of autoimmune thyroiditis and founded on the idea that inappropriate class II molecule expression in insulin-producing beta-cells of pancreatic islets could cause an organ-specific autoimmune disorder. The model is supported by results of several IDDM case studies as well as by the critical evaluation of the known effects of various lymphokines and lymphotoxins on endocrine target tissues. The possible roles of islet capillary endothelial cells and islet cell antibodies in the pathogenesis of IDDM are discussed.

Human thyroid epithelial cells (thyrocytes) express HLA Class II molecules in autoimmune thyroid diseases (ATD). Normal thyrocytes do not express Class II, but can be induced to do so by culture with interferon-gamma (gamma-IFN). We have examined HLA-D subregion expression in sections and monolayers of thyroid by indirect immunofluorescence using appropriate monoclonal antibodies. The results indicate that, in ATD, the incidence and intensity of Class II subregion expression by thyrocytes varies between patients, and follows the pattern DR greater than DP greater than DQ. The same hierarchy is observed in cultured normal thyrocytes treated with gamma-IFN: strong induction of Class II, and of DP and DQ in particular, requires relatively high concentrations of gamma-IFN or additional factors such as thyroid stimulating hormone. These findings suggest that HLA-D subregion expression by thyrocytes in on-going ATD is determined by the levels of disease related factors in the affected tissue.

On Southern blotting of DNA extracted from thyroid glands of five patients with Graves' disease, two probes (720 bp and 942 bp) for gag human immunodeficiency virus type 1 (HIV-1) gave a positive hybridisation signal in all samples tested. DNAs from peripheral blood mononuclear cells hybridised with the 720 bp gag HIV-1 probe in three of the five patients, none of whom had antibodies to HIV-1. Negative results were obtained with DNA from normal thyroid glands, thyroid neoplasms, various unrelated normal tissues, and virus-infected human cell lines. The intensity of the signal and the pattern of bands observed with the DNA of Graves' patients were heterogeneous and, in general, were not the same in the thyroid glands and peripheral blood mononuclear cells of individual patients. Similarly, no correlation was found between the positive hybridisation signals and other genetic and immunological indices or the duration of anti-thyroid drug treatment at the time the patients were investigated. The findings suggest the presence of a novel retrovirus, and the retrovirus-like sequences seem to be closely associated with thyroid autoimmunity.

We investigated the expression of intercellular adhesion molecule-1 (ICAM-1) and lymphocyte function-associated antigen-3 (LFA-3) by thyrocytes and their regulation by cytokines. Immunofluorescence studies on cryostat sections and on freshly dispersed cell preparations showed that ICAM-1 and LFA-3 are barely detectable in non-autoimmune thyrocytes. However, thyrocytes acquired ICAM-1 expression in culture. IFN-gamma, IL-1 beta and TNF-alpha produced a clear enhancement of ICAM-1 expression. When tested in combination, IL-1 beta and TNF-alpha were additive to the IFN-gamma effect. LFA-3 expression was not modulated by these cytokines. In the HT93 thyroid cell line generated by transfection with SV40, ICAM-1 and LFA-3 were both constitutively expressed at high levels. Cytokines modulated ICAM-1 expression similarly, but to a greater extent than in normal thyrocytes. LFA-3 remained unmodified. These results support the notion that normal thyrocytes are immunologically silent cells. The capability of cytokines to induce ICAM-1 together with HLA class I and class II-expression on thyrocytes suggests that under their influence, these cells may express all the surface molecules required for antigen presentation and/or for being recognized as target cells in the context of thyroid autoimmune disease.

Endocrine epithelial cells do not normally express human leukocyte antigen (HLA) class II molecules, but do so in a variety of autoimmune diseases. This finding suggests the hypothesis that such inappropriate class II-positive expression may enable these cells to present autoantigens and thus contribute to autoimmune pathogenesis. Indeed, class II-positive thyrocytes can present both exogenous antigenic peptides and intrinsic autoantigens to the appropriate T cells. Class II expression by thyrocytes can be induced by interferon-gamma, and is positively and negatively regulated by thyroid-stimulating hormone and epidermal growth factor, respectively. Furthermore, heterogeneity of thyrocyte class II subregion expression appears to be related to the nature of the inducing stimulus. The complexity of regulatory signals is underlined by findings in type I diabetes: islet beta cells aberrantly express class II in this disease, but class II cannot be induced in normal beta cells by interferon-gamma.

SUMMARY According to the ‘aberrant HLA expression’ hypothesis, endocrine autoimmunity is driven by presentation of self antigens by target cells over-expressing HLA molecules. In autoimmune thyroid diseases (AITD), thyroid follicular cells (thyrocytes) over-express HLA class I and HLA class II molecules. Since efficient presentation of endogenous peptides via class I requires transporters that translocate endogenous peptides from the cytoplasm to the endoplasmic reticulum, i.e. transporters associated with antigen processing (TAP) -1 and -2, the capability of thyrocytes to express TAP and whether TAP is hyperexpressed in AITD glands are issues relevant to the above hypothesis. Results from immunofluorescence and Northern blotting studies on primary thyrocyte cultures and on a thyroid cell line demonstrate that thyrocytes express constitutively TAP-1 at a low level, and that this expression is readily induced by interferon-gamma (IFN-γ) and to a lesser extent by IFN-α. In AITD, but not in non-autoimmune glands, thyrocytes hyperexpress TAP-1, as demonstrated by both immunohistopathology and flow cytometry. The cytokine pattern does not bear, as assessed by reverse transcriptase-polymerase chain reaction (RT-PCR), a clear relationship with TAP-1 expression. These results have broad implications and suggest that the core concept of the ‘aberrant HLA expression’ hypothesis of endocrine autoimmunity could be incorporated in the currently prevailing view of ‘autoimmunity by breach of peripheral tolerance’.

Endocrine autoimmune disorders are mediated by T cell-dependent responses to organ-specific antigens, but the mechanisms initiating the process remain unknown. Lymphocytes which use the gamma delta heterodimer as T cell receptor (TCR) for antigen constitute a distinct subset of T cells whose function remains elusive. In order to investigate their possible involvement in endocrine autoimmunity we have determined the proportion of gamma delta T cells in the peripheral blood of 23 patients with type 1 (insulin-dependent) diabetes mellitus (type-1 DM) and 30 patients with autoimmune thyrotoxicosis (Graves' disease). T lymphocyte TCR expression was assessed by fluorescence-activated flow cytometry on peripheral blood mononuclear cells using MoAbs UCHT1 (CD3), TCR delta 1 (gamma delta TCR), WT31 and beta F1 (alpha beta TCR) and both the percentage of T cells expressing gamma delta and the ratio gamma delta/alpha beta were calculated. In the diabetic patients gamma delta cells were not significantly different from the control group (7.7 +/- 54% versus 8.0 +/- 5.5% of T cells, P NS). There was no relation between the proportion of gamma delta lymphocytes and the presence of islet cell antibodies (ICA) in the sera. The Graves' patients showed a tendency towards a higher proportion of gamma delta T lymphocytes than the controls (gamma delta/alpha beta ratios: 0.095 +/- 0.047 versus 0.063 +/- 0.022, P = 0.03). In 14 Graves' patients the number of gamma delta were measured in paired samples of peripheral and intrathyroidal lymphocytes, demonstrating an expansion of gamma delta within the thyroid glands (0.21 +/- 0.3 versus 0.095 +/- 0.047, P = 0.032). Immunohistochemical studies showed that gamma delta cells were scattered among the predominant alpha beta lymphocytes infiltrating the thyroid gland and that they account for 10% of intraepithelial lymphocytes. No relation was found between the increase of gamma delta lymphocytes and any clinical features.