During the past two decades, researchers have been accumulating a large body of evidence indicating that T1DMA, as well as other endocrine diseases associated with it, all have an autoimmune pathogeneses (5-7).

Since the inception of modern immunology, immune responses directed against self-structures have been considered potentially harmful. Initially, it was believed that autoimmune responses did not occur (the "horror autotoxicos" of Erhlich). This idea soon needed modification because autoimmune phenomena were subsequently found to be common, albeit often of insufficient intensity to induce clinical disease. In normal individuals, the destruction of autoreactive T-cells occurs in the thymus for self-antigens (autoantigens) that are expressed in this gland as restricted by human leukocyte antigen (HLA) molecules (central tolerance). Self-peptides presented in context of class-1 MHC (HLA-A, B and C) induces depletion of autoreactive cytotoxic (CD8+) T cells while those presented in context of class-II MHC (HLA-DR, DQ and DP) depletes self-reactive helper-suppressor (CD4+) T cells. Many self-antigens, however, are expressed only in peripheral tissues. At this level, an individual's T-lymphocytes will only recognize foreign antigens in conjunction with autologous HLA molecules presented on the surfaces of antigen-presenting cells (APC). In the absence of co-expression of accessory molecules necessary to create an affective response, anergy is usually induced (peripheral tolerance). Thus autoimmune diseases represent a breakdown of normal self-tolerance, often at several sites, disrupting the complex balance of immuno-regulation maintained by multiple mechanisms within the central and peripheral components of the immune system.

Immunologic recognition of self-structures is not by itself a pathologic event, but one that is essential for the normal function of the immune system. However, excessive anti-self reaction and production of high titers of autoantibodies or quantities of autoreactive T-lymphocytes with high affinities for self-molecules may ultimately lead to an autoimmune or immunological attack of self. The evidence that T1DMA has an autoimmune pathogenesis is considerable, e.g. disease associated autoantibodies specific for islet cells and islet cell constituents, cell-mediated immune abnormalities detectable in peripheral blood, the chronic lymphocytic infiltration of the pancreatic islets (insulitis lesions), and an immunogenetic susceptibility reflected mainly by HLA-DR/DQ gene associations and linkages (8, 9).

In 1849, Addison (10) first described the clinical and pathologic features of adreno-cortical failure in nine autopsied patients, some of whom also appeared to have pernicious anemia, a disease that also initially received his name. Ogle (11) reported the first instance of coexisting diabetes and adrenal insufficiency in 1866. In 1908, Claude and Gourgerot (12) suggested a common pathogenesis for the simultaneous expression of polyglandular insufficiencies involving pancreatic islets, thyroid, gonads, adrenals, and the anterior hypophysis, a fascinating and correct assertion. Parkinson (13) in 1910 noted an association between pernicious anemia and T1DMA. Mononuclear leukocyte infiltration of goitrous thyroid glands was observed by Hashimoto (14) in 1912, while a similar inflammatory lesion of pancreatic islets, termed insulitis, was described by von Meyenburg (15) in 1940. The association between adreno-cortical failure and thyroiditis was documented by Schmidt (16) in 1926, and the syndrome complex was extended by Carpenter et al. (17) in 1964 to include T1DMA. It was not until 1956 that the autoimmune pathogenesis of these disorders could be supported by laboratory evidence, beginning with the discovery of circulating precipitating autoantibodies to thyroglobulin in patients with Hashimoto's thyroiditis by Roitt and Doniach (18).

The ability to detect organ-specific humoral autoantibodies with methods developed by Anderson et al. (19) and Blizzard and Kyle (20) confirmed the clinical association between diabetes and idiopathic (autoimmune) adrenalitis. Solomon et al. (21) demonstrated the coexistence of adrenal atrophy in diabetics with thyroid and adrenal dysfunction. Irvine et al. (22) reported that both pernicious anemia and thyroid disorders occur with significant frequency in first-degree relatives of diabetic patients. Autoantibodies to specific thyroid and gastric antigens (23-27) as well as to adrenal and islet cell antigens (28-33) in diabetic patients have been studied extensively. Cellular autoimmunity to thyroid antigens in T1DMA has also been reported (34).

Neufeld, Blizzard and Maclaren (35, 36) distinguished the two major autoimmune polyglandular syndromes centered around Addison's disease (APS-1 and 2). APS-1 and APS-2 are present well-circumscribed entities; however, APS-2 without Addison's disease was originally classified as APS-3. In retrospect, the latter two APSs are sufficiently related that APS-2 can be divided into APS-2a with Addison's disease and APS-2b without Addison's disease. The APS types are summarized in Table 1. APS-1 became to be seen as a fully penetrant recessively inherited syndrome complex while APS-2a/2b appeared biased to female patients with components expressed in successive generations, suggesting a dominant mode of transmission. Recently, various studies have identified the gene called autoimmune regulator (AIRE) responsible for APS-1 (37, 38). The thymus is a predominant site of AIRE gene transcription suggesting that defective thymic functioning disrupting central tolerance could be responsible for the widespread autoimmunities of APS-1.

The evidence supporting the autoimmune nature of the component diseases of the APS is compelling: (1) affected organs demonstrate a chronic inflammatory infiltrate composed mainly of lymphocytes; (2) some of the component diseases are associated with immune-response genes encoded by class-II MHC (HLA-DR/DQ); and (3) the syndromes are replete with autoantibodies reacting to targeted tissue-specific antigens. The inductive events that lead to the initiation of autoimmune pathogenic processes remain poorly understood. It is clear, however, that both genetic and environmental factors are involved in the disease processes. In celiac disease, the ingestion of gliadin in wheat flour induces autoimmunity to transglutaminase in intestinal cells and the clinical disease. Removal of the inciting agent (gliadin) causes remission in celiac disease. However, for the remainder, no clear inciting agent has been identified albeit many have been suspected. While APS-1 is associated with recessive and often heterologous mutations of the AIRE gene, patients with same types of AIRE gene mutations can develop different component diseases of APS-1 at various ages of onsets, suggesting the involvement of environmental factors and/or other background genetic factors, notably the HLA genotype. When T1DMA occurs in APS-1 for example, the typical HLA-DR/DQ alleles associated with that disease out of context of APS-1 are usually present, though APS-1 itself is not an HLA linked disease (114). Although studies in the past demonstrated no association between disease components and HLA-DR/DQ alleles (191, 192), later studies reported associations were observed between some of the disease components and HLA-A alleles (193). A more recent study showed weak HLA associations in the pathogenesis of APS-1 (194). However, in the case of APS-2, at least one central predisposing gene has been identified, namely the cytotoxic T lymphocyte antigen-4 (CTLA-4) gene. Component disorders, however, are highly HLA-DR/DQ dependent. Linkage to CTLA-4 gene in APS-2 as well as in T1DMA, autoimmune thyroid disease, Addison's disease and other autoimmune disorders e.g. multiple sclerosis, and celiac disease suggests that CTLA-4 is a general autoimmune locus, and susceptibility polymorphysm(s) within the gene may lead to general defects in the immune regulation, while other tissue specific (e.g. insulin gene polymorphisms) or antigen specific (e.g. MHC) genetic factors and environmental factors determine the involvement of particular target organs (195). Autoantibodies may arise in a primary fashion through a breakdown in normal immunological tolerance, or secondarily by immunization with an environmental agent that is a molecular mimic of a self-antigen as discussed below. Other autoantibodies appear to arise during a secondary immune response which is stimulated by the release of intracellular antigens from damaged glands that are normally sequestered from the immune system. This "immunological spreading" of determinant reactivity often signals progressive autoimmune destruction and the approach of impending clinical disease onset.

Adrenal autoantibodies (AA) detected by indirect immunofluorescent labeling have been reported in at least two of every three patients with non-tuberculous Addison's disease when tested at the time of their diagnoses (20). All layers of the adrenal cortex bind AA with striking sparing of the adrenal medulla. Fluorescence of the zona glomerulosa in particular gives a distinctive pattern. Occasionally, AA-positive individuals who do not yet have overt adrenocortical failure can be identified by screening patients with associated autoimmune endocrine diseases, as well as their high-risk family members. Up to 40% of such apparently healthy individuals were reported to subsequently develop Addison's disease over a follow-up period of 6 months to 10 years (48). This risk was observed to be especially high if the autoantibodies fixed complement in-vitro or were present in high titers. Another 20% of asymptomatic AA-positive relatives were reported to have elevated basal serum levels of ACTH or renin, or blunted adrenocortical responses to an intravenous infusion of ACTH, features that are indicative of subclinical glandular dysfunction.

Some 15% of AA-positive patients with Addison's disease also have an autoantibody that cross-reacts with other steroid-hormone producing cells, i.e. the placental syncytiotrophoblasts, ovarian theca/luteal cells, and/or testicular Leydig cells. These "steroidal cell" autoantibodies (SCA) are distinguished from AA by their ability to be adsorbed from serum by preincubation with adrenal, gonadal (ovarian or testicular), or placental homogenates, whereas AA are removed from positive sera by prior exposure to adrenal homogenates exclusively. When detected, SCA indicates a high risk for future gonadal failure, especially in females (49, 50). By contrast, patients with premature ovarian failure with positive test for AA would also have higher risk to develop Addison's disease than those who do not (48). As mentioned above, the major steroid cell autoantigens involved in the reactions of AA have now been identified as the p450 steroidogenic enzyme 21-hydroxylase (21-OH), alpha-hydroxylase (17-OH) and P450 side-chain cleavage enzyme (P450scc). The major antigen for SCA is 17-hydroxylase, a 55-kDa gonadal and adrenal steroid biosynthetic P450 microsomal enzyme (23). The frequency of antibodies to 21-OH varies depending upon the techniques used, and are generally at higher titers for patients with Addison's disease in association with APS than in patients with an isolated Addison's disease (30, 51-53). Using a 125I labeled recombinant human 21-OH as antigen, for example, antibodies to 21-OH protein were detectable in 72% (43/60) patients with isolated Addison's disease, 92% (11/12) patients with APS-1, and 100% (27/27) in patients with APS-2a (47). The frequency of antibodies to 17-OH and P450scc were also higher in patients with APS-1 than in patients with isolated Addison's disease (30). The dominant epitopes on 21-OH recognized commonly by autoantibodies from patients with Addison's' disease as an isolated disease, or in association with APS, are located in the C-terminal end and in a central region of 21-OH (54, 55). The recognized epitopes on 21-OH can either be conformational (56) or linear in nature (43). Whereas cellular immune mechanisms are thought to cause glandular destruction, a pathogenic role for humoral autoreactivity in autoimmune oophoritis has been suggested by studies showing complement-mediated cytotoxicity of cultured granulosa cells in the presence of sera from affected patients, though not in the presence of sera from control patients (57). Binding of SCA to granulosa cells by indirect immunofluorescence, however, can only be demonstrated when autoantibodies are present in high titers. Antibodies prepared from patients with Addison's disease had been shown to have inhibitory effects on recombinant 21-OH enzyme activity in-vitro (58), but such enzyme inhibitory effects are not so evident in-vivo (59), and conceptually are unlikely to account for the resultant disease.

Autoantibodies to parathyroid gland were firstly reported in 38% of 74 patients with idiopathic hypoparathyroidism by indirect immunofluorescent assay (60). Subsequent investigations, however, have found that anti-parathyroid serological immunoreactivity is rare in patients with failed glands (61) and usually is not parathyroid-specific (62). Antibodies considered to be against parathyroid antigens have been confused with mitochondrial autoantibodies in previous reports, and humoral sensitivity to parathyroid tissue may have delineated a tissue-specific response to antigens within the endothelial component of the gland (63). However, in a recent study by Western blotting, calcium sensing receptor was recognized in 32% (8/25) of patients with hypoparathyroidism associated with either APS-1 or hypothyroidism; the major epitope was located to the external domain of the receptor (64). In another study, however, calcium-sensing receptor autoantibodies were not detected in APS-1 patients despite that the majority of them had hypoparathyroidism (183). Two recent data one from Mayer et al., which demonstrated the presence of calcium-sensing receptor autoantibodies in the sera of APS-1 and APS-2 patients (184), and Kifor et al., showed similar results in patients with hypoparathyroidism (185) confirming the original study findings.

The intensive studies of humoral autoimmunities against antigens expressed by pancreatic cells (e.g., islet gangliosides, insulin, proinsulin, glutamic acid decarboxylase (mainly GAD65 and less often GAD67), and insulinoma antigen-2 (IA-2) highlight the complexity of disease-autoantibody relationships. The presence of islet cell autoantibodies (ICA) detected by immunofluorescence together with insulin autoantibodies and/or GAD and/or IA-2 has a high predictive value for developing T1DMA. However, autoantibodies against the cytoplasm of islet cells in frozen section (ICA) as well as autoantibodies against GAD65 and against aromatic L-amino acid decarboxylase (AADC)(196) also occur in many patients with APS-1 (65, 66), a syndrome with only a low likelihood of progression to clinically overt T1DMA (36, 66). Recent studies indicate that autoantibodies to islet cell autoantigens in patients with APS-1 have different reactive characteristics from those of patients with T1DMA as in APS-2a and 2b. For example, GAD65 autoantibodies from patients with APS-1 are readily detectable by Western blotting (66) indicating that these autoantibodies recognize linear epitopes on GAD65 protein. GAD65 autoantibodies present in patients with T1DMA, however, react with conformational epitopes of undenatured proteins (67). This suggests that different immune regulations could have been involved in driving the production of these two sets of GAD65 autoantibodies. In addition, the majority of patients with APS-1 were found to have positive autoantibodies to AADC (197,198); the presence of AADC autoantibodies in APS-1 patients did not correlate with the presence of T1DMA in the APS-1 patients (196,198). AADC autoantibodies were not detected in isolated T1DMA (197,198), suggesting that the presence of islet cell autoimmunity in APS-1 is not necessarily an indication of the destruction of islet cells, at least not very rapidly, since patients with APS-1 only unusually develop immune mediated T1DM (36, 66). Here the islet cell autoimmunity presumably lacks other components of the pathogenic process (e.g., antigen-specific cytotoxic T lymphocytes) that are necessary to produce b-cell damage and thus overt hyperglycemia. ICA seen in APS-1 may also be directed against additional antigens that are not targeted in T1DMA. Cystein sulfinic acid decarboxylase (CSAD) shows 50% amino acid sequence identity with GAD65. Recently, it has been shown that CSAD autoantibodies can cross-react with GAD65 (199). Despite close structural relation to GAD65, CSAD autoantibodies were negative in T1DMA serum samples, and did not appear to be associated with any of the known autoimmune manifestations of APS-1 except mucocuatenous candidiasis, which was the only common manifestation of APS-1 patients studied. The cross-reactivity with GAD65 and their presence only positive in APS-1 patient sera, CSAD autoantibodies may reflect the tendency of APS-1 patients to develop anti-GAD antibodies directed against different epitopes than T1DMA (200). Therefore, the islet cell non-destructive autoimmune response in patients with APS-1 may result from an impaired cellular immune regulatory mechanism, which differs from that in APS-2, albeit these speculations need to be systematically proven through specific studies.

Autoantibodies in patients with APS react with thyroid gland proteins, including thyroid peroxidase, thyroglobulin, and thyrotropin receptors. While immunoglobulins against the thyrotropin receptor may stimulate or inhibit both thyroid gland activity and growth, no consistently discernible effect on thyroid function has yet been attributed to autoantibodies that recognize thyroid peroxidase or thyroglobulin (Table 3). Nevertheless, immunization of susceptible strains of mice with thyroglobulin in complete Freund's adjuvant induces a thyroid-specific immune infiltrate in experimental allergic thyroiditis (EAT) (68).

Thyroid-associated ophthalmology (exopthalmos) is accepted as an autoimmune inflammatory disorder of the periorbital connective tissue (69). The presence of anti-TSH receptor antibodies, in particular TSAb (70), has been correlated with the presence of thyroid eye disease, even in the 10% of individuals who are biochemically euthyroid, although at levels lower than in hyperthyroid Graves' disease. Furthermore, anti-TSH receptor antibodies are not detected in 20% to 70% of euthyroid patients with Graves' ophthalmopathy (70) and any correlations with disease activity levels remain controversial (71, 72), making these assays of limited value to the clinician. The presence of anti-thyroglobulin (9%) and anti-microsomal antibodies (17%) is even lower in patients with euthyroid eye disease (70) and indeed, anti-thyroid peroxidase negativity has been suggested as a risk factor for ophthalmopathy, emphasizing the need for alternative markers. More recently, various groups have reported reactivity to a 64-kD protein in human and porcine eye muscle in the serum of patients with ophthalmology (73, 74) and this has been identified as a 67kD flavoprotein subunit of the mitochondrial enzyme, succinate dehyrogenase (75) besides three additional protein targets for anti-eye muscle antibodies (76-79). Antiflavoprotein and anti-G2 antibodies are most strongly, although not exclusively, associated with eye disease. Further studies using purified human antigens and confirmation of these results by other laboratories are required to verify these results. Currently, none of these antibodies can be recommended for routine clinical use.

Achlorhydria and pernicious anemia occurring as part of the APS are associated with the presence of circulating autoantibodies against gastric parietal cells (GPCA) and, less frequently, intrinsic factor (IFA). Approximately 10 percent of patients with T1DMA have co-existing-circulating GPCA, of whom many develop achlorhydria (6). The pathogenic importance of these immunoglobulins is suggested by their toxic effects on the gastric mucosas of frogs and rats (80, 81). The parietal cell proton pump (H, K+-ATPase) represents at least one target of GPCA (82). Thus GPCA appears to be primarily associated with achlorhydria which becomes complicated by iron deficiency, while IFA may arise secondarily as a consequence of gastric cell damage and are associated with increasing likelihood of clinical pernicious anemia as the patient ages. In APS-1, however, deficiency of vitamin B12 often arises during childhood years.

Melanocyte autoantibodies have been demonstrated in a small number of individuals with APS-1 and vitiligo, but a similar association has not been made in patients with isolated vitiligo (83). The antigens involved have been variously reported as tyrosinase (43) and tyrosinase associated proteins (84). Recently, Hedstrand et al, using sera from APS I patients for immunoscreening of a cDNA library from human scalp, identified the transcription factors SOX9 and SOX10 as novel autoantigens related to this syndrome (85). Immunoreactivity against SOX9 was found in 14 (15%) and against SOX10 in 20 (22%) of the 91 APS I sera studied. All patients reacting with SOX9 displayed reactivity against SOX10, suggesting shared epitopes. Among the 19 patients with vitiligo, 12 (63%) were positive for SOX10. Furthermore, Kemp et al, 2002 recently identified another pigment cell antigen that is a target in patients with vitiligo. Using IgG from vitiliginous patients to screen a melanocyte cDNA phage-display library, they identified the melanin-concentrating hormone receptor 1 (MCHR1) as a novel autoantigen (86).

Antibodies detected by indirect immunofluorescent labeling of hypothalamic vasopressin-producing cells also have been reported in a small number of patients with central diabetes insipidus who had other autoimmune endocrinopathies (87). In a report of 19 patients with a variety of endocrine autoimmunities, autoantibodies against anterior pituitary lactotrophs were detected (88), and scattered reports of humoral responses against somatotrophs and perhaps even gonadotrophs also have been published, but not independently confirmed. However rarely, if ever, have these patients had symptomatic disease of the hypothalamic-pituitary axis. In contrast, among 30 reported patients with proven or presumed symptomatic lymphocytic hypophysitis, autoantibodies directed against the pituitary gland have been described (89). Using transformed rodent pituitary cell lines as a substrate in an indirect immunofluorescence assay, immunoglobulins that specifically bound the hypophyseal cells in culture were observed in the serum of humans with the empty sella syndrome (1). Clearly, pituitary autoimmunity is an area of potential research that needs more attention.

In addition to the common features of APS, malabsorption, alopecia, chronic active hepatitis, and hypogonadism may develop in some APS patients. Therefore screening for marker autoantibodies for these associated diseases facilitates early diagnosis and treatment of the corresponding diseases. For example, antibodies to tyrosine hydroxylase are found in patients with alopecia areata (184); autoantibodies to tryptophan hydroxylase are associated with malabsorption (185); autoantibodies to mitochondria, smooth muscle and/or liver kidney microsomes are associated with autoimmune liver disease and chronic active hepatitis (186), while hypogonadism is associated with steroid producing cell autoantibodies (187). In their 35 patients with APS-1, Betterle et al., reported that hypergonadotropic hypogonadism (61%) was the most frequently observed additional disease followed by alopecia (38%), vitiligo (22%), chronic hepatitis (19%), and malabsorption (15%) (188). Very recently, a report indicates that 100% of APS-1 patients have high titer antibodies to (IFN-α and ώ). Whereas the finding may have diagnostic importance, such antibodies may provoke an immunological dysfunction and thus be a target for immunological intervention.

The autoantibodies described above that react with cellular enzymes or hormones generally have unknown pathogenic significance. The functional role of autoantibodies that alter organ function by binding to their hormone receptors is, however, readily understandable. Whereas autoantibodies that bind thyrotropin and acetylcholine receptors have been long recognized to have pathogenic importance in Graves' disease and myasthenia gravis respectively, a similar mechanism has only recently emerged as a potential pathogenic process in other endocrine autoimmunities. Thus immunoglobulins that recognize receptors for gonadotropins, insulin, or ACTH receptors may inhibit the action of their respective hormone ligands (3).

Among APS-1 patients with chronic active hepatitis (CAH), autoantibodies against mitochondrial, nuclear, or smooth-muscle antigens are frequently found, albeit their clinical significances are also unclear. Since the liver is one of the organs with the highest levels of AADC, AADC, autoantibodies may also be involved in the pathogenesis of CAH (196, 201). Abnormal B-lymphocyte function has been variably described in patients with APS. Deficiency of IgA is one such feature while high levels of IgG and IgE also have been observed in some patients (2, 4). Age-specific expansions of activated CD5+ B lymphocytes were observed in patients with T1DMA (90) while our group has reported deficiencies in regulatory CD4+/CD25+ and natural killer T cells (NKT cells) (91).

The presence of circulating tissue-specific autoreactive leukocytes in APS patients was first demonstrated by the elaboration of migration inhibitory factors (MIF's) after incubation of target-organ homogenates with peripheral blood mononuclear cells (PBMC's) from affected individuals. Subsequently, increased levels of PBMC's expressing activation markers such as HLA-D have been observed in patients with early but not end-stage T1DMA, Graves' disease, thyroiditis, Addison's disease, and oophoritis (93, 99). Since surface antigen phenotyping does not reliably distinguish lymphocytes with different functions, cytokine production profiles of PBMC's are coming under scrutiny. Diminished production of interferon- (INF-) and interleukin 2 (IL-2) in response to mitogens has been observed in patients at high risk for T1DMA (100). In contrast, autologous thyroid cells elicited interferon- production by PBMC's harvested from patients with autoimmune thyroid disease, but not from those with nontoxic goiters or thyroid cancer (101). Examinations of affected end organs obtained early in the disease process will ultimately be more informative than those of circulating lymphocytes. Unfortunately, tissue specimens obtained at or after the time disease becomes clinically apparent contain infiltrates that represent a complex response against a multitude of antigens. They are thus unsuitable for studies of autoimmune initiators. As programs of disease prediction expand, improved availability of more appropriate organ specimens with early autoimmune infiltrates is expected. Animal models of disease are now being used to follow the kinetics of leukocyte infiltration of autoimmune targets. In NOD mice, descriptions of early insulitis have described initial infiltration by macrophages and CD8+ T lymphocytes, followed by CD4+ T lymphocytes and B lymphocytes (102).

It appears unlikely that the action of a single T-lymphocyte clone can result in clinically important organ failure since adoptive transfer of either T1DMA or thyroiditis requires transfusion of both CD4+ and CD8+ lymphocytes. Nonetheless, it is likely that autoimmunity against a single antigen initiates disease, albeit it may be that several antigens may be involved in this initial response. Target-organ invasion by restricted T-cell families, identified by their expression of T-cell receptor genes that contain uniquely rearranged variable (V) or complementarity-determining regions (e.g., CDR3), or monoclonal expansion of B lymphocytes has not been demonstrated convincingly in either APS. Preferential use of certain TCR families may occur, however, in an antigen-specific fashion during the inductive events (103).

Despite the multitude of investigations, the sequence of effector events leading to eventual cell destruction has not been resolved. It has been difficult to determine how the local effects of cytokines (released from either leukocytes, damaged endothelium, or possibly endocrine epithelium (92, 104)), aberrant targeted cell expression of class I and/or class II MHC (9, 93), and adhesion molecules (ICAM-1) (105, 106) on the endocrine epithelium surface contribute to the pathological process. In diabetic rodents, b-cell expression of class I MHC is observed early in the pathogenic sequence, enhancing the ability of CD8+ T cells to lyse these cellular targets. Later, there may be enhanced class-II MHC reactivity due to the invasion of macrophages and perhaps some patchy aberrant expression of these antigens on pancreatic -cells. Some have suggested that T-cell elaboration of INF- may have an important role in cell destruction, since it can be cytotoxic either alone, or in combination with tumor necrosis factor-alpha (TNF-) (107). These two cytokines also have been reported to increase aberrant class-II MHC expression on thyroid (108), islet (109), and ovarian granulosa cells (110). The pancreatic b-cells are especially sensitive to cytotoxic effects of IL-1, an effect that may operate through cytokine-mediated accumulation of local nitric oxide (NO) (111). Recently, it has been suggested that cellular lysis through inflammation is often perpetuated by generation of NO (111, 112), but it remains to be shown whether this mechanism will prove to be the principal mode of glandular cell destruction in the APS.

Recurrent mucocutaneous candidal infections in APS-1 that are often resistant to treatment most certainly reflect an abnormality of T-lymphocyte function. No specific T-cell defect to account for these findings has been consistently identified, albeit one must exist (99). The possible role of the elusive transfer factor remains unclear (113).

APS-1 is an autosomal recessive disease with a pattern of inheritance initially observed by analysis of patients with idiopathic Addison's disease and hypoparathyroidism (131), and later reported by others in different racial groups (35, 39, 40, 132). By allelic association and linkage analyses a candidate gene named autoimmune regulator (AIRE) was initially mapped to the long arm of chromosome 21 (21q22.3) by two individual groups (37, 38). AIRE gene consists of 14 exons and encodes a protein of approximately 545 amino acids with two zinc finger domains typical of a nuclear transcription factor (Figure 1). AIRE gene was later narrowed down to within 500 kb of a gene encoding phospho-fructokinase of liver type (PFKL), by linkage analyses and physical mapping in a relatively homogenous Finnish and European patients with APS-1 (133, 134), and later in a heterogeneous US patients with APS-1 (135). At least 45 different mutations have been detected in patients with APS-1 with different racial backgrounds indicating that this gene is the disease gene responsible for APS-1 (38, 115-126). The major AIRE gene mutations with different frequencies in APS-1 patients of various ethnic backgrounds are listed in Table 2 (see below).

In addition, 15 polymorphisms have been reported, six of which were found in the coding region, but only one resulted in amino acid substitution (123, 125). The mutations are spread throughout the coding region of the gene, although there are two mutation hotspots, as shown by the presence of the same mutations present on many different haplotypes from several populations (123, 125, 127, 128). Other mutations common in isolated populations include a single nucleotide substitution A374G, which leads to a substitution of a tyrosine residue in cysteine in the HSR domain in the Iranian Jewish population (129) and the predominant Sardinian APS-1 mutation R139X, responsible for 90% of Sardinian APS-1 alleles (121). Because the functions of the AIRE protein are still unresolved, the mechanisms by which different mutations disturb the physiological functions of the protein are unknown. However, recent data indicates that mutations in different regions of the gene have different effects on the intracellular targeting and transcriptional regulation functions of the AIRE protein (129, 130). The predicted outcomes for most of the AIRE mutations are truncated conceptual protein of AIRE due either to the introduction of a stop codon, or a frame-shift of the coding gene. However, there are missense mutations e.g. R15L, L28P, Y90C and K83E, which result in the substitution of a single amino acid in exon 1 or 2. It remains to be confirmed whether such missense mutations could disrupt the function of AIRE protein. Two of the most frequently detected mutations in various racial groups, one R257X is located at exon 6, and the other 1094del13 is located at exon 8 of AIRE gene. Other mutations are much less frequent, and some of them have only been detected in a single allele. R257X is due to a transition of C to T at amino-acid position 257. This results in the change of an Arg codon (CCA) to a stop codon (TGA) and would produce only a protein with about 256 amino acids (37, 38). R257X is a dominant mutation for Finnish patients with APS-1 and is also frequently present in patients with other ethnic backgrounds, such as in north Italians, Swiss, British, Germans, New Zealanders and American Caucasians (123, 128). 1094del13 is a 13 bp deletion at nucleotide position of 1094-1106, and result in a frame shift to produce a truncated 372 amino acid residue. 1094del13 occurs in APS-1 patients with various ethnic backgrounds (38, 120, 121, 123, 128). In addition, 1094del13 is a dominant AIRE gene lesion for British patients with APS-1, since 74% (17/23) of mutated AIRE gene alleles from British patients with APS-1 contains this deletion (120).

Figure 1. Chromosome localization of AIRE gene, AIRE gene mutations and AIRE protein. AIRE gene is located on chromosome 21 q22.3, close to gene encoding phosphor-fructokinase of liver type (PFKL). To date some 45 different mutations have been detected with R257X and 1094del13 to be the dominant mutations detectable in patients with different ethnic backgrounds. Shown in the figure are the mutations in various exons. AIRE protein contains two PHD zinc-finger motifs (PHD), three LXXLL motifs (L) and a proline-rich region (PRR), suggestive of its putative role as nuclear transcriptional regulator.

Despite its rarity, the prevalence of APS-1 is higher in certain ethnic groups, e.g., Iranian Jews (1/9,000) (40), Finns (1/25,000) (39), and Sardinians (1/14,000) (182). Since it is more prevalent in certain populations, it could be related to a founder gene effect. Founder effects exist for some genetically isolated populations according to the analyses on mutations and haplotype of polymorphic markers closely associated with AIRE gene locus. Recombination events become less for two genomic markers that are located closely than those remotely located in the same chromosome, i.e. linkage disequilibrium happens to the closely located genes or polymorphic markers. Thus, individuals are likely to have common ancestors if they share same haplotype for polymorphic markers, which are in linkage disequilibrium, especially for those from genetically isolated populations. Haplotype analyses on polymorphic markers located closely to AIRE gene on Finnish patients have suggested that more than 85% of cases of APS-1 in Finnish patients are due to one major mutation that is commonly present in the ancestors of the Finnish population (134). This is in concordance with the finding that the mutation R257X was present in up to 82% of Finnish patients with APS-1 and accompanied with one haplotype of closely linked polymorphic markers, D21S1912, and PFKL (37, 38). D21S1912 is located at approximately 130 kb upstream of AIRE gene, and PFKL located at 1.5 kb downstream of AIRE gene (37, 38). This evidence suggests that R257X occurred as a single mutation event in the Finnish population. Studies of 12 British families with APS-1 for AIRE gene mutations found that 17 of the 24 possible mutant AIRE alleles tested had 1094del13 with a common haplotype spanning the AIRE gene locus, suggesting the presence of a founder effect in the British population (120). Mutation R139X has been found in 90% (18/20) of independent alleles with identical haplotypes for D21S1912 - PFKL in that ethnic group (121).

Patients with same AIRE gene mutations often had different closely linked haplotypes, suggesting that either ancient mutational events or multiple independent events occurred to account for the AIRE mutations and haplotypes observed. For example, R257X is the major mutation present in patients from European countries other than Finland. Those patients who had R257X tend to have diversified haplotypes of D21S1912 - PFKL (123). This is also the case for the other major mutation, 1094del13, since different haplotypes were present in patients with different ethnic origins. For example, 9 out of 15 alleles of 1094del13 detected in 13 patients from a group of American Caucasian patients with APS-1 had different haplotypes of D21S1912 - PFKL (128), suggesting multiple independent events led to the 1094del13 mutation. This should be expected since the patients were of heterogeneous origins typical of the North American population.

An understanding of the biological role of the AIRE protein should provide needed insights into the mechanism of autoimmunity and to APS-1 in particular. AIRE gene is expressed as in mRNA prevalently in the thymus, but is also expressed in other tissues such as lymph nodes, pancreas, adrenal cortex and PBMCs (37, 38). Recent studies in AIRE gene knockout mice indicate an important role of AIRE protein in eliminating autoreactive T cells through transcriptional control of tissue specific antigens at the levels of the thymus (202). Attention has already been drawn to the putative nuclear localization and its role in transcriptional regulation of the encoded protein based on the analyses of the predicted amino-acid sequence (37, 38). The AIRE protein contains two PHD zinc-finger motifs, three LXXLL motifs and a proline-rich region, suggestive of its putative role as nuclear transcriptional regulator (Fig. 1). The pattern of the two zinc-finger motifs in AIRE-1 is similar to the Mi-2 and TIF1 autoantigens (136). Mi-2 autoantigen is a 240-kDa human nuclear protein recognized by sera from patients with autoimmune dermatomyositis (137). The Mi-2 autoantigen is actually a partial fragment of a chromo-helicase-DNA (CHD) binding protein, CHD3, identified recently (138). The families of CHD proteins are known to play roles in gene expressions and regulations (138). Also, TIF1 is actively involved in the transcriptional control of the estrogen receptor (136, 139). Accordingly, AIRE gene is most likely participating in the regulation of the expression of another gene(s). There are no obvious correlations between the mutant genotypes of AIRE gene and the clinical phenotypes of APS-1, suggesting that the outcome of the syndrome may be influenced by environmental factors. The core phenotype of APS-1 includes the 3 diseases, mucocutaneous candidiasis, hypoparathyroidism and Addison's disease. Patients with APS-1 are also frequently accompanied with one or more other autoimmune diseases, such as chronic active hepatitis, alopecia, vitiligo, as well as evidence for an immunodeficiency state, represented by chronic diarrhea/malabsorption, chronic mucocutaneous candidiasis and our observation of oro-pharyngeal carcinomas. Not all patients with APS-1 express all of the three core component diseases or the frequently accompanied diseases. Even patients of same ethnic origin with the same AIRE mutation often present with different component diseases of APS-1 or the orders of the appearance of component diseases (120, 121, 123). Different phenotypic expressions are also present in affected siblings. Such variations in the clinical phenotype may be the result of the involved gene interacting with the immune system when exposed to some unknown environment factors. However, some ethnic groups of patients with APS-1 may develop specific component diseases, suggesting that the outcome of the syndrome is influenced by background genes within population. For example, there is a relative rarity of candidiasis among the Iranian Jewish patients with APS-1 (40). Also, immune mediated diabetes is rarely seen in patients with APS-1 in the US; however it does occur in some Finnish patients with APS-1, especially with increasing age. Finnish patients with APS-1 often have ectodermal and enamel hypoplasia (39). Calcium deficiency due to hypoparathyroidism should not be the primary cause for enamel dystrophy, since ectodermal and enamel hypoplasia occurs in APS-1 patients with or without hypoparathyroidism in Finnish patients (39, 140). In addition, those non-Finnish patients with APS-1 who had hypoparathyroidism are seldom seen with enamel hypoplasia (40, 135). Presumably, different mutations in the responsible gene could be involved in these various phenotypes as shown in the presence of different haplotypes for genetic markers among patients with different ethnic backgrounds (134). In other words, background genes or genes with epistatic effects may be responsible for variations in the expressed phenotype. Alternatively, more than one gene may be responsible for the development of APS-1, albeit, this is becoming increasingly unlikely. Thus, studies for the function of the identified AIRE gene could shed more light into the pathogenic mechanism of APS-1, especially, autoimmune Addison's disease, which also occurs either as an isolated disease or in association with APS-2. Recent studies have shown that the individual HLA class II alleles may modify the APS-1 phenotype rather than AIRE mutation genotype (194). In the same study, component diseases in APS-1 such as Addison's disease was associated with HLA-DRB1*03, alopecia with HLA-DRB1*04-DQB1*0302, whereas T1DMA was negatively correlated with HLA-DRB1*15-DQB1*0602.

APS-2a/2b remains to be genetically defined. In contrast to the autosomal recessive pattern of APS-1, APS-2a and AP-2b express an autosomal dominance pattern with incomplete penetrance, since APS-2a/2b often affect individuals in many generations of the same family (4, 114). As the major component diseases of APS-2a, Addison's disease is associated with HLA and whether non-HLA related gene(s) are involved in the disease remains to be demonstrated. Patients with APS-2a often share same susceptible HLA alleles with those patients with only an individual component disease (141, 142). For example, DR3-DQB1*0201/DQB1*0302 are associated with APS-2a when there is T1DM (143). Such HLA associations suggests that particular molecules of HLA are required in the development of component autoimmune diseases, and the expressions of a particular autoimmune phenotype depends on the involvement of other gene products, especially in a multicomponent autoimmune syndrome like APS-2a. Unlike APS-1, the order of the appearance of the component diseases of APS-2 varies greatly. Individual patients can present with Addison's disease and then develop T1DMA and/or autoimmune thyroid disease, or any other sequence. This could indicate the presence of different pathogenic pathways during the development of APS-2a. Thus, it might be expected that other gene products could influence the susceptibility of individual patients to particular alleles of HLA. For example, the association of DQB1*0302 with APS-2a was abolished in one study, when those patients with APS-2a, and overt clinical T1DMA or positive for autoantibodies to islet antigens, were excluded from the analyses (144). Also, the development of the same disease with different susceptible HLA alleles has been observed in inter-racial studies for HLA susceptibility (145). In addition, multi-genetic involvement in the development of individual component diseases of APS-2a has been proven, such as T1DMA that is linked to more than 10 loci in non-HLA genomic regions (146), or autoimmune thyroid disease which appears to be polygenic as well (147, 148).

The CTLA-4 gene located on chromosome 2q33 encodes a co-stimulatory molecule that inhibits T-cell activation (202, 203) and is linked to T1DMA, autoimmune thyroid disease and Addison's disease (204, 205, 206). Association between Addison's disease appears to be stronger in the presence of APS-2a than in those isolated Addison's disease (206).

Patients with APS-2b lack definitive genetic features except for their associations with its component diseases and their associated HLA alleles. For example, DQB1*0301 is increased in Hashimoto's thyroiditis, DRB1*03 in Addison's disease, DRB3 genes are increased in Graves' disease, and DRB1*13 in vitiligo. Non-HLA genes are however expected to be involved in the development of APS-2b also, and attempts have been made to map for non-HLA genes responsible for component disease of APS-2b. For example, Tomer et al., (149) linked a susceptible locus for Graves' disease to within a 6 cM at the chromosome 20q11.2. Although this susceptible locus is waiting to be confirmed, the finding of a gene for the component disease of APS-2b would help to look into the pathogenesis of the syndrome. The susceptible locus at the chromosome 20q11.2 was linked to Graves' disease but not to Hashimoto's thyroiditis. This again suggests that the cause of pathogenic autoimmunity may involve different pathogenic processes, and that alternative pathways exist for the break down of self-tolerance of the immune system. Thus, the genetic studies for autoimmune endocrine syndromes should bring insights to understand the pathogenic process of these diseases, which may be applicable to the component autoimmunities. APS-2 is commonly inherited as an autosomal dominant trait with incomplete penetrance and is related to the HLA-DR3 and DR4 phenotypes (114). However, thyroid autoimmunity in pedigrees with type 1A diabetes may segregate independently from the HLA complex (6). The molecular genetics of the alpha and chains of the HLA-DQ protein so intensively investigated in type 1A diabetes have not yet been closely examined in APS patients. As discussed elsewhere, unique amino acid sequences in the DQ and DQ chains that probably relate to the antigen-binding cleft of the class-II molecule are important to the inherited susceptibility to T1DMA. Multiple genes are likely to be involved in the APS's, and those outside the HLA complex have yet to be identified.

APS-1 is a rare childhood disease affecting males and females equally, but more is prominent in certain races such as in Finns (39), Sardinians (183) and Iranian Jews (40). APS-1 is diagnosed when a patient presents with at least two of its' three cardinal clinical features: hypoparathyroidism, chronic mucocutaneous candidiasis and hypoadrenocorticism (Addison's disease). Any young person afflicted by troublesome moniliasis without the systemic infection which is generally associated with severe immune deficiency, should be assessed for a possible T-lymphocyte deficiency state as well as for APS-1. In one study, nearly 45% of pediatric patients with refractory monilial infections but no overt underlying T-cell defect had an autoimmune endocrinopathy (150). Of the 50 to 100% of APS-I patients who develop recurrent monilial infection, most have lesions that are restricted to the skin, nails, and oral and perianal mucosa. Whereas remissions of varying length occur, progressive courses are common, and gastrointestinal involvement can become severe, especially when complicated by bacterial overgrowth, chronic diarrhea, or gastrointestinal hemorrhages. More than 75% of APS-1 patients develop hypoparathyroidism, usually presenting before age 10 years. Severe hypocalcemia manifested by carpopedal spasms, seizures, or laryngospasm can be the presenting feature of APS-1, especially in young children. Patients with chronic mucocutaneous candidiasis and hypoparathyroidism have a high chance of developing Addison's disease. The latter rarely appears before the development of hypoparathyroidism (6). The symptoms of hypoparathyroidism, however, can be masked in the presence of untreated Addison's disease where it may manifest upon steroid replacement therapy (41). Adrenocortical failure typically develops between the ages of 10 and 30 years. Deficiencies of mineralocorticoids and glucocorticoids usually arise simultaneously, but their onsets can be dissociated by up to 5 years (39). Females suffer from gonadal insufficiency more often than males and usually present with maturational arrest after the onset of a normal pubarche and menarche. Autoimmune oophoritis also may present with failed pubertal development or with menstrual irregularities (151).

Fat malabsorption, which may be episodic, has been linked by some to hypoparathyroidism. More likely causes include IgA deficiency, gluten sensitivity, and bacterial overgrowth of the upper small bowel. Malabsorption/chronic diarrhea has been associated with autoantibodies to tryptophan hydroxylase, Deficiencies of iron or vitamin B12 result from parietal cell autoimmunity with subsequent early appearance of achlorhydria followed by intrinsic factor deficiency and pernicious anemia. Typical atrophic gastritis arises in 15% of APS-1 cases with a mean age at onset of 16 years. Studies of Finnish patients have particularly emphasized manifestations of APS-1 in the teeth and integument. In decreasing order of frequency, enamel hypoplasia, ungual dystrophy (pitting), keratopathy, and tympanic membrane sclerosis have all been reported at rates from 33 to 77% (39). Vitiligo may be missed if not specifically sought using ultraviolet light (Woods lamp examination). Alopecia totalis or universalis is frequent, but all types occur. It has been suggested that hair loss may diminish after treatment of hypoparathyroidism is started (152), but this does not reflect the authors' experience. The appearance of hepatomegaly or jaundice with dark urine and clay-colored stools often heralds the onset of chronic active hepatitis. It occurs in up to 10% of patients and is not associated with persistent immunological hypersensitivity to hepatitis viruses. Sjőgren's syndrome (parotitis, arthritis, and sicca syndrome) is not infrequent, while T1DMA, chronic thyroiditis, and hypophysitis are distinctly uncommon.

APS-2a is far more common than APS-1 and is diagnosed when a patient has adrenocortical deficiency with T1DMA, chronic lymphocytic thyroiditis, or Graves' disease. Unlike APS-1, this syndrome can be more difficult to recognize before the onset of clinically significant multigland disease. The disease commonly manifests in the third or fourth decade, but it is not uncommon before or after these ages. It is heralded by adrenocortical failure in almost half the cases, although this estimate may be skewed by a selection bias in the literature. As many as 20 years can elapse before polyglandular involvement becomes evident. Furthermore, isolated thyroiditis and T1DMA are common enough in these age groups that routine adrenal autoantibody screening of such affected patients is not justified unless adrenocortical insufficiency is clinically suspected. APS-2b, on the other hand is characterized by the presence of autoimmune thyroid disease in association with one of the other organ-specific autoimmune diseases such as atrophic gastritis/pernicious anemia, vitiligo and/or T1DMA, but in the absence of Addison's disease (6). Frequent associations common to APS-2b as centered around T1DMA, myasthenia gravis (42), and vitiligo (43) have been well documented. Graves' disease and Hashimoto's thyroiditis are both frequent in APS-2, as are vitiligo and pernicious anemia. The authors recommend routine thyroid autoantibody screening of all T1DMA diabetes patients and full endocrine autoantibody testing in those found to be positive. However, physicians should routinely elicit historical and physical features relevant to the diagnostic triad in all patients with T1DMA and/or autoimmune thyroiditis. A family history of poly-glandular failure is often present in past generations that can serve as a flag for those patients who need extra monitoring. The presence of extra endocrinological manifestations such as alopecia or vitiligo is less common than in APS-1, but when such manifestations are present, they are important clinical indicators, especially if they are profound. The mortality risk of untreated adrenocortical failure in the 2% of patients with myasthenia gravis who develop associated endocrinopathies requires that all these patients under 40 years should be assessed closely for endocrinological disorders during their initial investigations.

Rare diagnoses have on occasion been reported in association with an APS. The authors and others have followed patients with APS-1 who developed severe, idiopathic, noninflammatory myopathy with eventual respiratory failure. Separate reports suggested that pure red cell hypoplasia and male infertility in patients with APS-1 responded well to glucocorticoid therapy (153, 154). In rare cases, neo-osseous porosis and sarcoidosis have been linked to APS-1 and APS-2 respectively (155, 156). Asplenia/hyposplenism may not be uncommon and it can be suspected when Howell-Jolly bodies are found in a peripheral blood smear (189). We recommend that all patients with APS-1 be vaccinated against pneumococcus since they are prone to septicemias with the organism and even sudden death from such infection.

Given the interrelationship between T1DMA and other endocrinopathies in the polyglandular syndromes, how should the clinician monitor the T1DMA patient? The consequences of undiagnosed endocrinopathies can involve significant morbidity or mortality, particularly Addison's disease. One could argue that the alert clinician would recognize the onset of a new endocrinopathy in a T1DMA patient or the symptoms of endocrine insufficiency in a relative of a patient; however, disease onset may be insidious and easily missed unless specifically and routinely thought of and tested for.

Ideally, all patients with T1DMA should be screened for the presence of adrenal, steroidal, gastric parietal, thyroid microsomal, and thyroglobulin antibodies at the time of their diagnosis, which itself should be confirmed with ICA, GAD65A, IA-2A and IAA testing when clinical presentations are not classic. Family members of a proband with T1DMA should also undergo these autoantibody studies and those who are positive or who have family members with other endocrinopathies should have full autoantibody profiles performed. The necessity of further workup of T1DMA patients or their family members depends on their panel results. ICA (or GAD plus IA-2 autoantibody) screening should be repeated annually in children younger than 10 years and every 2 years in relatives younger than 20 years because these age groups are at higher risk for developing -cell autoimmunity; the risk of conversion from negative to positivepancreatic markedly falls with increasing age.

About 25% of T1DMA female patients (half this number of male patients) develop autoimmune lymphocytic thyroiditis (e.g., Hashimoto's thyroiditis, Grave's disease), most commonly recognized after the diagnosis of T1DMA. The consequences of untreated hypothyroidism include congestive heart failure, dyslipidemia, infertility, growth retardation, and rarely, slipped capital femoral epiphysis. Graves' disease can cause significant weight loss, weakness, and congestive heart failure and exopthalmus/ptosis. We would suggest screening for thyroid microsomal and thyroglobulin autoantibodies in all patients with T1DMA. Patients with thyroid autoantibodies or goiter should have TSH screening annually and careful monitoring of vital signs at each examination. Those found positive should be also screened for adrenal autoimmunity. If a patient is found to have thyroiditis, family members should be screened for thyroid autoantibodies and goiter also.

All patients with T1DMA plus thyroid autoimmunity or has family history of Addison's disease should be screened for adrenocortical, steroidal cell or best 21-hydroxylase autoantibodies. Possibly all patients with T1DMA should be screened to minimize the risks of sudden, unexplained death due to unrecognized Addison's disease. Addison's disease is readily diagnosed and readily treated when diagnosed. However, the disease can be fatal and has been occasionally diagnosed at autopsy. Patients with adrenal 21-hydroxylase or steroidal antibodies should be carefully evaluated for clinical signs and symptoms of adrenal insufficiency; they should have annual basal plasma adrenocorticotrophic hormone (ACTH), 8:00 AM serum cortisol and supine renin levels studied because these will become elevated well before the onset of electrolyte abnormalities and most clinical symptoms. A history of weight loss, hyperpigmentation (often described as ``dirty'' pigmentation of the neck or elbows), easy fatigability, muscular weakness, dehydration, emesis, or malaise should be sought at each clinical visit. Patients should be educated to recognize and report these symptoms immediately. Electrolyte abnormalities are a late finding, and normal electrolytes do not exclude disease.

Addison's disease may be present in as many as one in every 250 patients with T1DMA. It should be recognized and treated as it occurs; left unrecognized, this disease poses an absolute risk to patients. Tumor, tuberculosis or other granulomatous disorders should be eliminated from the differential diagnosis.

T1DMA patients with (1) strong family histories of thyroiditis or pernicious anemia, (2) megaloblastic anemia, or (3) unexplained fatigue should be screened for gastric parietal cell antibodies. Any patient with these antibodies should have yearly determinations of HgB, ferritin and vitamin B12 levels. If abnormalities are found, chronic oral iron or monthly B12 injections should be administered. Most patients reach mid life before this becomes necessary. Any T1DMA patient with new onset of anemia or peripheral paraesthesia should notify his or her endocrinologist promptly.

Impaired fertility may result from inadequate glycemic control in T1DMA. Gonadal autoimmunity, however, particularly in patients with steroidal cell antibodies, has been well-documented in patients with T1DMA. Patients with steroidal antibodies may develop both Addison's disease and hypogonadism. Thus, antibody-positive patients should be admonished to recognize the symptoms of adrenal insufficiency described above. They also should be monitored for secondary amenorrhea, oligomenorrhea, or infertility. Any of these signs should alert the clinician to the possibility of autoimmune ovarian or testicular failure. Determination of serum steroidal cell antibodies, LH, FSH, estrogen, and testosterone levels will facilitate diagnosis. Annual determinations of follicle-stimulating hormone (FSH) and luteinizing hormone (LH) may permit prediction of gonadal failure. Patients also should be examined for evidence of mucocutaneous candidiasis or for hypoparathyroidism, if these autoantibodies are detected.