Autoimmune polyglandular syndrome Type 2: the tip of an iceberg?

Abstract

Autoimmune polyglandular syndromes (APS) are conditions characterized by the association of two or more organ-specific disorders. Type 2 APS is defined by the occurrence of Addison's disease with thyroid autoimmune disease and/or Type 1 diabetes mellitus. Clinically overt disorders are considered only the tip of the autoimmune iceberg, since latent forms are much more frequent. Historical, clinical, genetic, and immunological aspects of Type 2 APS are reviewed. Furthermore, data on 146 personal cases of Type 2 APS are also reported.

HISTORICAL BACKGROUND OF TYPE 2 AUTOIMMUNE POLYGLANDULAR SYNDROME

The association between Addison's disease and diabetes mellitus was first reported in 1886 by Oegle [1], but in the original description adrenocortical failure ensued from bilateral tuberculous destruction of the adrenal glands. During the subsequent 75 years, only 15 other cases were mentioned with this association [2]. Moreover, the combined occurrence of Addison's disease and chronic lymphocytic thyroiditis was first reported in two patients by Schmidt in 1926 [3], neither of whom had clinical signs of thyroid dysfunction. From that time, the coexistence of Addison's disease and autoimmune thyroid disease has come to be known as Schmidt's syndrome. To further investigate this clinical association, a few years later Wells studied 20 patients with Addison's disease revealing that a lymphocytic infiltration of the thyroid glands was far more common in patients with idiopathic Addison's disease than in those with tuberculous adrenal insufficiency [4]. The relationship between these three glands was revealed in 1931, when Rowntree and Snell [5] reported the first case with Addison's disease, hyperthyroidism and diabetes mellitus, and one year later Gowen [6] described a patient affected by Addison's disease, hypothyroidism and diabetes mellitus. This last patient died of diabetic ketoacidosis, and postmortem investigation showed that some islets of Langerhans had lymphocytic infiltrations similar to that seen in the adrenal and the thyroid glands, thus revealing for the first time that the pancreas can be involved in the same pathological process affecting both thyroid and adrenals.

The relationship between Addison's disease and diabetes mellitus was extensively reviewed in 1959 by Beaven et al. [7] on 66 cases, and some years later by Solomon et al. [8] on 113 cases. Diabetes mellitus was found to be the disease heralding the syndrome in 57–63% of the cases, while Addison's disease preceded diabetes mellitus in 23–35% and the two diseases appeared to be simultaneous in 8–10% of the cases; in 4% of the patients the sequence of the diseases was not specified. Many patients with this association died within one year by the time of the clinical diagnosis. Post-mortem investigation of these patients showed adrenals with atrophy and lymphocytic infiltration in 74%, tuberculous inflammation in 22%, and a neoplasia in 2% of the cases, denoting that the majority of them were affected by autoimmune Addison's disease. Unfortunately, no data were reported about the histopathological features of pancreatic islets in these cases. In the same years, Carpenter reviewed 142 cases with Schmidt's syndrome [9]. In the vast majority of cases, the thyroid autoimmune disease was Hashimoto's thyroiditis or idiopathic myxedema, and in the remaining cases Graves’ disease. The link between Schmidt's syndrome and diabetes mellitus was confirmed in this review, where 28 patients (20%) were found to suffer also from diabetes mellitus [9]. The complete triad of Addison's disease, thyroid autoimmune disease and Type 1 diabetes mellitus is also termed Carpenter's syndrome.

THE DISCOVERY OF AUTOIMMUNE DISEASES

During the 1950s and 1960s some discoveries greatly improved our knowledge about autoimmune diseases. In 1956, Roitt and Doniach [10] found that patients with Hashimoto's thyroiditis had circulating autoantibodies reacting to thyroid self antigens. In the same year, Adams and Purves [11] recognized that patients with Graves’ disease had a serum factor defined as long-acting thyroid stimulator (LATS), later found to be an immunoglobulin G binding to the TSH receptor [12–14]. Also in 1956, Rose and Witebsky [15] demonstrated that a lymphocytic thyroiditis similar to the spontaneous human disease can be induced in animals by immunization with autologous thyroid extracts in Freund adjuvant.

Early after, Anderson described the presence of circulating autoantibodies to extracts of adrenal cortex in patients with idiopathic Addison's disease, suggesting an autoimmune pathogenesis of this form of adrenal insufficiency [16]. Based on these findings, Witebsky established some criteria that ideally should be fulfilled in order to define a disease as autoimmune in origin [17]: (1) direct demonstration of free circulating autoantibodies and/or of cell-mediated autoimmunity, (2) recognition of the specific antigen against which antibodies are directed, (3) production of antibodies against the same antigens in experimental animals, (4) appearance of pathological changes in the corresponding tissues of an actively sensitized experimental animal that are similar to those in the human disease. These postulates have been subsequently revised by Bona and Rose, who proposed the following lines of evidence: (1) direct (transfer of the disease by pathogenic antibody or pathogenic T cells), (2) indirect (reproduction of the disease in experimental animal models, isolation of autoantibodies or self reactive T cells), and (3) circumstantial (association with other autoimmune diseases in the same individual or in the same family, lymphocytic infiltration of the target organ, association with particular HLA-haplotypes or aberrant expression of HLA class II antigens on the affected organ, favourable response to immunosuppression) [18].

Besides Hashimoto's thyroiditis, Graves’ disease and Addison's disease, in the following years many other diseases initially designated as ‘idiopathic’ were included into the group of the autoimmune disorders, like chronic atrophic body gastritis, pernicious anaemia, chronic hypoparathyroidism, premature ovarian failure, vitiligo, alopecia, autoimmune hepatitis, myasthenia gravis, and so on. In fact, they revealed the presence of circulating autoantibodies to the relevant autoantigen(s) of the target organs (parietal cells, intrinsic factor, liver-kidney microsomes, steroid-producing cells, acetylcholine receptor, etc.), or met the above mentioned criteria [19].

It was only in 1974 that Type 1 diabetes mellitus, one of the three main diseases occurring in Type 2 APS, entered this group, when Bottazzo et al. [20] demonstrated that patients affected by Type 1 diabetes mellitus and other autoimmune endocrinopathies had circulating autoantibodies to the pancreatic islets.

Autoantibodies in organ-specific autoimmune diseases

During the past two decades, many enzymes, hormones and receptors have been identified as the targeted autoantigens in organ-specific autoimmune diseases, as reviewed by Song et al. [21]. The role of these autoantigens has been claimed to be crucial in initiating and perpetuating the autoimmune response, but their natural intracellular localization has raised many doubts on their unique responsibility in triggering autoimmunity. Nevertheless, their discoveries greatly improved the methods for the detection of organ-specific autoantibodies [22]. Moreover, International Committees coordinated standardization programmes in order to improve sensitivity, specificity and reproducibility of autoantibody determination among the various laboratories [23–25].

It was only in 1985 that thyroid microsomal antibodies were identified to react against the thyroid peroxydase, and subsequent investigations recognized numerous organ-related specific autoantigens involved in organ-specific autoimmunity (Table 1).

Table 1.

Main organ-specific autoimmune diseases and recognized relevant autoantigen targets

Target organ	Disease	Autoantigens	Reference
Thyroid	Graves’ disease	TSH-receptor†	[26]
	Hashimoto's thyroiditis	Thyroid peroxydase	[27,28]
	Idiopathic myxoedema	Thyroglobulin	[10]
Adrenal cortex	Addison's disease	21-hydroxylase	[29]
Gonads	Gonadal failure	P450 side-chain cleavage enzyme	[30]
		17α-hydroxylase
Parathyroid	Hypoparathyroidism	Calcium-sensing receptor	[31]
Endocrine pancreas	Type 1 diabetes mellitus	Glutamic acid decarboxylase	[32]
		Tyrosine-phosphatase like	[33]
		Insulin	[34]
Stomach	Body chronic atrophic gastritis	H+/K+ pump ATPase	[35]
	Pernicious anaemia	Intrinsic factor	[36]
Intestine	Celiac disease	Transglutaminase	[37]
	Idiopathic malabsorption	Tryptophan hydroxylase	[38]
Liver	Chronic autoimmune hepatitis	P450 (IID6, IA2)	[39]
Pituitary	Lymphocytic hypophysitis	68, 49, 43 kD from human	[40]
	Infundibuloneurohypophysitis	pituitary membrane (?)
Skin	Vitiligo	SOX9, SOX10	[41]
		Tyrosinase	[42]
	Alopecia	Tyrosine hydroxylase	[43]
Muscle	Myasthenia gravis	Acetylcholine receptor	[44]

^†TSH, thyroid stimulator hormone. Autoantibodies may exert both stimulating (Graves’ diseases) or blocking (idiopathic myxedema, atrophic variant of Hashimoto's thyroiditis) activity.

All the studies aimed at recognizing the antigen-antibody reactions contributed to improve the diagnosis of autoimmune diseases, as well as permit the early detection of individuals at risk for the future development of organ-specific autoimmune diseases.

As a consequence of these discoveries, many diagnostic tests that employed the immunofluorescence technique on cryostat sections of human or animal tissues were progressively substituted by RIA or ELISA tests using cloned autoantigens [30,32,33,45–48].

CLASSIFICATION OF AUTOIMMUNE POLYGLANDULAR SYNDROMES

In general, organ-specific autoimmune diseases do not cluster casually, but reveal preferential associations. In 1980, Neufeld and Blizzard [49] published a classification of the autoimmune polyglandular syndromes (APS) on clinical grounds indicating the existence of four main distinct types (Table 2). Type 2 APS, or Schimidt's syndrome, is characterized by the obligatory occurrence of autoimmune Addison's disease (which represents the pivotal disease) in combination with thyroid autoimmune diseases and/or with Type 1 diabetes mellitus.

Table 2.

Classification of autoimmune polyglandular syndromes (APS). Adapted from [49]

Type 1	Chronic candidiasis, Chronic hypoparathyroidism, Addison's disease (at least two present)
Type 2	Addison's disease (always present) and Thyroid autoimmune diseases, and/or Type 1 diabetes mellitus
Type 3	Thyroid autoimmune diseases associated with other autoimmune diseases (excluding Addison's disease and/or hypoparathyroidism)
Type 4	Combination of organ-specific autoimmune diseases not included in the previous groups

TYPE 2 AUTOIMMUNE POLYGLANDULAR SYNDROME

Epidemiology

The frequency of Type 2 APS in humans is rare, being described in about 1·4–4·5 per 100 000 inhabitants [50,51]. However, recent observations revealed that the disease is much more frequent if one considers also the cases with subclinical forms (see below).

Animal models

Animal models exist for several autoimmune diseases and may serve to focus on autoimmune targets and the immune mechanisms involved in the pathogenesis of self-aggression. Spontaneous models, such as the NOD mouse for Type 1 diabetes mellitus, are believed to reflect their human counterparts. Moreover, animal manipulation by means of drugs, infectious agents, or genetic engineering may be helpful to unveil possible ethiopathological factors in triggering autoimmunity. Unfortunately, spontaneous animal models of Type 2 APS are rare. The White Leghorn Chicken is an obese inbred strain which develops a lymphocytic thyroiditis, but can also develop adrenal cortex autoantibodies. However, no clear impairment of the corresponding target organs usually occur, so the syndrome remains solely at subclinical level [52]. A spontaneous form of Type 2 APS has also been described in a boxer dog affected by primary hypothyroidism and partial adrenocortical deficiency: the autopsy study has demonstrated thyroid atrophy and lymphocytic adrenalitis with complete destruction of the zona fasciculata and reticularis [53]. Apart from these spontaneous models, there are animal models of experimentally induced autoimmune endocrine diseases obtained after environmental perturbation (e.g. viruses, toxic substances), thymectomy procedures, or genetic manipulation. For example, some strains of mice infected with cytomegalovirus may give rise to a Type 2 APS with lymphocytic infiltration of the adrenals, pancreatic islets, liver, myocardium, salivary glands, and circulating organ-specific autoantibodies are also detectable [54]. An APS affecting thyroid, adrenals, ovary, pancreatic islets, and stomach in various combination has been shown in mice treated with cyclosporin A at birth followed by removal of the thymus [55]. This last observation suggests that polyglandular autoimmunity ensues from a more profound T-cell disturbance than that required for the induction of a single organ disease.

Despite the stimulating information provided by animal models, data in animals do not necessarily reflect human disease in vivo. No infectious agents or noticeable immunodeficiency states have been indeed demonstrated in human Type 2 APS.

Clinical features

In 1981, Neufeld [56] reviewed from the literature and from personal cases 224 patients affected by Type 2 APS. This author reported that the syndrome began after 20 years of age in 84% of the cases with an increased prevalence in middle-aged women. Thyroid autoimmune diseases (encompassing Hashimoto's thyroiditis, primary myxedema, symptomless autoimmune thyroiditis, Graves’ disease, isolated ophthalmopathy) were present in 69% and Type 1 diabetes in 52% of the cases. Clinical aspects and laboratory findings of the components of the syndrome are similar when the diseases occur isolated or in the context of APS. However, these aspects are beyond the aims of the present paper and have been extensively treated in previous excellent reviews [57–60]. Minor autoimmune diseases might also occur, like vitiligo, chronic atrophic gastritis, or hypergonadotropic hypogonadism (Table 3), but they were less frequent if compared to Type 1 APS. In the following years, other investigators studied further groups of patients from Italy [61], Sweden [62], Norway [63], and Germany [64]. Since 1970, we have studied 146 patients with Type 2 APS, and their main clinical features are summarized in Table 3. Thyroid autoimmune disease was diagnosed with a greater frequency and Type 1 diabetes with a lower frequency in comparison to the reports of Neufeld, indicating that the clinical combinations may vary according to the different populations examined. Other minor autoimmune diseases have been diagnosed in patients with Type 2 APS, with a frequency ranging from 1 to 12% [51] (Table 3).

Table 3.

Clinical features of patients with Type 2 APS

	From Neufeld et al. [56]	Personal data
Patients (no.)	224	146
Female/Male ratio	1·8	>4
Family history of Type 2 APS	n.r.	0
Adults/Children	n.r.	133/13
Main diseases
Addison's disease	100%	100%
Thyroid autoimmune diseases	69%	88%
Type 1 diabetes mellitus	52%	23%
Minor diseases
Vitiligo	4·5%	12%
Hypergonadotropic hypogonadism	3·6%	10%
Chronic autoimmune hepatitis	n.r.	3%
Alopecia	0·5%	4%
Pernicious anaemia	<1%	2%
Seronegative arthritis	n.r.	2%
Myasthenia gravis	n.r.	0
Adenohypophysitis	n.r.	0

n.r., not reported.

As far as clinical combinations concern, among our patients 129 (88·4%) had two main diseases, while only 17 (11·6%) had the complete tri-glandular syndrome (Carpenter's syndrome). The most frequent association was Addison's disease and Hashimoto's thyroiditis, while the least was Addison's disease, Graves’ disease and Type 1 diabetes mellitus (Table 4).

Table 4.

Prevalence of the main autoimmune diseases in Type 2 APS patients (personal data)

Endocrine diseases	No. of cases	Prevalence (%)
Addison's disease + chronic thyroiditis	82	56·1
Addison's disease + Graves’ disease	31	21·2
Addison's disease + Type 1 diabetes mellitus	16	10·9
Addison's disease + chronic thyroiditis + Type 1 diabetes mellitus	14	9·6
Addison's disease + Graves’ disease + Type 1 diabetes mellitus	3	2·0

The mean ages at the clinical onset of the different diseases in our patients with Type 2 APS are summarized in Table 5. The diseases more frequently diagnosed in the youngest patients were vitiligo, followed by Type 1 diabetes and hypergonadotropic hypogonadism. Among thyroid autoimmune disorders, Graves’ disease usually preceded, whereas Hashimoto's thyroiditis followed Addison's disease.

Table 5.

Ages of onset of the different autoimmune diseases in patients with Type 2 APS and frequency of the relevant antibodies

Autoimmune disease	Mean age at disease onset (years) (range)	Frequency of the relevant autoantibody at disease onset† (%)
Vitiligo	27·7 (9–43)	None
Type 1 diabetes mellitus	28·4 (2–63)	70
Hypergonadotropic hypogonadism	29·0 (18–40)	100
Graves’ disease	33·4 (7–58)	80
Addison's disease	34·6 (1–85)	91
Pernicious anaemia	35·5 (34–37)	100
Alopecia	38·6 (32–52)	None
Chronic thyroiditis	40·2 (12–80)	97
Chronic atrophic gastritis	45·4 (16–65)	70
51·6 (42–61)	100

^†i.e. less than 1 years from the clinical diagnosis.

Humoral immunity

The frequencies of the relevant autoantibodies detectable at the clinical onset of the diseases constituting Type 2 APS are summarized in Table 5. As a rule, the frequency of antibodies against adrenal cortex detected by immunofluorescence on normal human adrenal glands and/or to 21-hydroxylase measured by radioimmunoassay [65] was relatively stable over time. Similar trends were found for antibodies to thyroid and gonadal antigens. By contrast, antibodies to endocrine pancreas (classical islet cell antibodies, glutamic acid decarboxylase or tyrosine phosphatase-like antibodies) revealed a rapid decline in the course of time, being around 50% in Type 1 diabetes patients with long-standing disease. In addition, vitiligo and alopecia in Type 2 APS were not associated with any autoantibody specificity, thus different from patients with the same diseases in the context of Type 1 APS [41,43,66].

Cellular immunity in the target organs

The pathological hallmark of autoimmune thyroiditis is lymphoplasmacytic infiltration. Frequently, lymphocytes are organized into well-developed germinal centres. Thyroid follicles are of reduced size and variable degrees of fibrosis are present. The infiltrating T cells are mainly CD8⁺, but also CD4⁺ T cells are present, many of which are activated as they express HLA class II molecules [67,68].

The pattern of insulitis in newly diagnosed Type 1 diabetes is characterized by an infiltration of lymphocytes, which are primarily T lymphocytes [69]. The majority of the infiltrating lymphocytes are of T cytotoxic/suppressor phenotype, with a few B cells and macrophages. Some of the T lymphocytes show the markers of cell activation [70]. In the advanced phases acinar cell atrophy is usually found.

The pattern of infiltration of adrenals in Addison's disease at the onset is characterized by a widespread mononuclear cell infiltrate consisting of lymphocytes, plasma cell and macrophages. Residual cortical nodules of regenerating cells secondary to high levels of corticotropin (ACTH) may be seen, but in the advanced stages of the disease, fibrosis and atrophy greatly predominate. In contrast to autoimmune thyroiditis and Type 1 diabetes, there has been no phenotypic characterization of infiltrating lymphocytes [71].

Studies on histopathology of the target organs involved in Type 2 APS have given results similar to those observed in isolated autoimmune forms. Adjacent organs or tissues non directly targeted by the autoimmune reaction are typically spared by the autoimmune attack.

Imaging

Regarding the imaging of the involved glands, cross-sectional imaging techniques, such as computed tomography (CT) and nuclear magnetic resonance (NMR) are able to show the adrenals with a resolution and clarity unimagined even 20 years ago. This has brought about a remarkable improvement in the diagnosis and characterization of adrenal insufficiency. As a rule, corticoadrenal failure due to autoimmune adrenalitis, both as isolated form or as component of APS syndromes, shows normal or minuscule adrenal glands bilaterally [72]. We evaluated the adrenal glands in 57 patients Type 2 APS at the onset of Addison's disease by CT or NMR, and in 50 of them we found normal adrenal patterns, while in the remaining cases the adrenals were reduced in volume consistent with gland atrophy.

Ultrasound technique has also greatly enhanced the diagnosis of thyroid autoimmune diseases in the recent years, and a diffuse or multifocal hypoechoic pattern has claimed to be typical of autoimmune thyropathy, both in goitrous or in chronic atrophic thyroditis and in Graves’ thyrotoxicosis [73].

Unfortunately, the imaging of endocrine pancreas in patients with Type 1 diabetes has proven to be frustrating, and no reliable imaging procedures of the insulitis process are routinely available so far.

Genetic susceptibility

Type 2 APS is a combination of three of autoimmune diseases and for many years it was believed that the genetic profile of patients with Type 2 APS could be influenced by the disease accompanying Addison's disease. In 1986, Addison's disease was reported to be linked with DR3 and/or DR4 aplotypes and this association was found to be independent from the presence or the absence of Type 1 diabetes [74]. Subsequently, it was found that in patients with Type 2 APS there was an association with HLA-DR3/DQB1*0201 haplotype when Addison's disease was not combined to pancreatic autoimmunity, and with HLA-DR4/DQB1*0302 when Addison's disease was combined to pancreatic autoimmunity [74]. In this study, however, patients with pancreatic autoimmunity included also nondiabetic subjects with islet cell antibodies, many of whom did not necessarily develop diabetes.

Subsequent investigation evaluating a large group of Norwegian patients with Addison's disease demonstrated that those with Type 2 APS were significantly associated with DRB1*04; DQA1*03; DQB1*0302 and DRB1*03; DQA1*0501; DQB1*02, independently from the presence of Type 1 diabetes. Furthermore, it was demonstrated that HLADRB1*01; DQA1*01; DQB1*0501 haplotype conferred protection against Addison's disease [63]. Other genes have been studied in Type 2 APS patients. A polymorphism of the cytotoxic T lymphocyte antigen-4 (CTLA-4) gene was found to be associated with Addison's disease in the context of Type 2 APS in English, but not in Norwegian, Finnish or Estonian patients [76]. Recently, mutations in the AIRE gene (typically found in Type 1 APS) have been investigated in Addisonian patients with Type 2 APS, but this gene was not found to be implicated [77,78].

We studied for the class II HLA haplotype 54 patients with Type 2 APS affected by Type 1 diabetes and/or thyroid autoimmune diseases. Our investigation confirmed that DRB1*03; DQB1*02, DRB1*04; DQB1*03 and DRB1*03,*04 were significantly associated with Addison's disease (P < 0·0001, P < 0·01 and P < 0·0005, respectively), independent from the presence of diabetes mellitus or the form of thyroid autoimmune disease [79]. Furthermore, we demonstrated a negative association with DRB1*01; DQB1*05 (P < 0·0001) and DRB1*13 (P < 0·02). Similar data have been reported in Type 2 APS patients from Norway [63].

Incomplete forms of Type 2 APS

In his original review Neufeld had established that a patient could be classified as having Type 2 APS if affected by thyroid autoimmune disease or Type 1 diabetes mellitus, and if at least one member in the family had Type 2 APS [49]. However, a positive family history for Type 2 APS is in truth exceptional (Table 3). Moreover, from the clinical point of view, hardly ever the syndrome blows up simultaneously with two or three main autoimmune diseases in one individual, while it usually initiates with a single disease (i.e. Type 1 diabetes mellitus, Graves’ disease, Hashimoto's thyroiditis, or Addison's disease), and sometimes even with a minor disease (e.g. vitiligo, pernicious anaemia, premature ovarian failure, alopecia, chronic atrophic gastritis) (Table 5). After a variable period of latency, a proportion of these subjects may develop the other components of the syndrome. For this reason, it is of considerable importance to identify among patients with a single disease those at risk for future development of the ‘fully expressed’ Type 2 APS. The only way to become acquainted with this is to screen patients with one organ-specific autoimmune disease for the circulating autoantibodies relevant to the other main diseases at the clinical diagnosis and every two or three years. Large population studies have defined their frequency. For example, in patients with Type 1 diabetes but no clinical adrenal failure, autoantibodies to the adrenal cortex (markers of potential Addison's disease) are found with a frequency of 0·4–1·6% [80–83]. In patients with thyroid autoimmune diseases with no clinical adrenal failure, autoantibodies to the adrenal cortex are present in about 1% of the cases [81,82], while in those with Addison's disease without overt thyroid disease, autoantibodies to the thyroid are present in 40–58% of the cases [84–86]. Moreover, patients with Addison's disease but no clinical manifestations of Type 1 diabetes, pancreatic autoantibodies are present in about 6–20% of them [51,85–87]. We designated these conditions as ‘incomplete’ APS, therefore extending the concept of Type 2 APS to the forms not yet fully expressed at clinical level. Antibody positive individuals are considered at high risk of developing clinical dysfunctions and will require to be further studied by using specific functional/morphological tests (i.e. determination of thyroid hormones, TSH and thyroid examination by ultrasound in patients with antibodies to thyroid, oral or intravenous glucose tolerance test in those with antibodies to endocrine pancreas, or ACTH test in those with antibodies to adrenal cortex). This approach will identify the patients with potential Type 2 APS (if functional tests are normal) or with subclinical Type 2 APS (if functional tests are abnormal). In the case of a positive antibody test despite normal function of the relevant organ, appropriate follow ups are advisable. In this way, we’ll be able not only to initiate early treatment of the incoming autoimmune disease in patients already affected by one endocrine disorder, but possibly to prevent the clinical outbreak of the ongoing autoimmune disease. Early recognition and treatment of a second or third complicating autoimmune endocrine disease, like acute adrenal failure in one patient with Type 1 diabetes mellitus or hypothyroidism in one with Addison's disease, may be crucial and life-saving in some cases.

Table 6 summarizes the frequency of the organ-specific autoantibodies detected in patients with one of the diseases composing Type 2 APS and the cumulative risks of developing clinical diseases in positive cases.

Table 6.

Incomplete forms of Type 2 APS

Clinical manifestation	Prevalence per 100 000†	Autoantibodies to:	Frequency (%)‡	Risk of future disease (%)‡
Thyroid autoimmune diseases	2375	Adrenal cortex or 21-hydroxylase	0·5–1	∼30
Type 1 diabetes mellitus	192	Adrenal cortex or 21-hydroxylase	0·5–1·6	∼30
Addison's disease	14	TPO and/or Tg	40–48	∼50
Addison's disease	14	GAD, IA2, insulin	6–20	From 5 to 90*
Minor autoimmune diseases (e.g. alopecia, vitiligo, pernicious anaemia)	150–400	Adrenal cortex and Endocrine pancreas and/or thyroid	0·5–1	∼10

TPO, thyroid peroxydase; Tg, thyroglobulin; GAD, glutamic acid decarboxylase; IA2, second antigen of islet cells.

^‡From [80–86].

^†From [88].

^*On the basis of the number of positive antibodies [87].

Bearing in mind that the estimated prevalence of the autoimmune diseases constituting Type 2 APS ranges from 14 to 2375 cases per 100 000 individuals [88] and that the prevalence of the other relevant autoantibodies in each autoimmune disease varies from 0·5 to 48% (Table 6), it may be inferred that the prevalence of the incomplete forms of Type 2 APS is much more frequent (up to 150 cases per 100 000 individuals) than the complete form (1·4–4·5 cases per 100 000 individuals) [50,51]. However, only a proportion of the patients with the incomplete forms will develop a complete clinically overt syndrome (Table 6).

Pathogenesis

The paradox of autoimmunity and the consequent diseases represent one of the mysteries that still bewilder immunologists, and several questions remain unanswered. Organ-specific autoimmune diseases develop in genetic susceptible individuals under the stimulation of environmental factors. As a consequence, these individuals produce specific humoral and cell-mediated immune responses against the constituents of the body's own tissues, and may involve one or more organs. A persistent defective capacity has recently been described in CD4⁺ CD25⁺ regulatory T cells, a subset of specialized T lymphocytes involved in suppression of autoreactivity, in patients with Type 2 APS but not in patients with single autoimmune endocrinopathy or in normal healthy controls [89]. However, these data have been observed in a small number of patients and require further investigation.

To explain multiple organ involvement in APS, Tadmor et al. [90] have hypothesized that organs derived from the same embryonal germ layer share common specific antigens. However, if this hypothesis may explain the pathogenesis of Type 3 APS (e.g. thyro–gastric autoimmune syndrome where both tissues are derived from the endodermal layer), it does not clarify Type 2 APS, in that the adrenal cortex is of mesodermal origin, while the thyroid and the pancreas are of endodermal origin. It still remains unclear why autoimmunity is focused on proteins typically present in endocrine tissues and not in other organs of the same germ layer, how the immune system selectively recognizes these autoantigens, and why multiple organs may be involved in the same individual on different occasions.

Another crucial question is if the break of the immune tolerance in (Type 2) APS induces the simultaneous activation of multiple autoreactive clones of lymphocytes, or if this occurs at different times in the course of the life. The observation that at the onset of the heralding autoimmune disease further autoantibodies to other endocrine organs are frequently detectable is in agreement with the first hypothesis. However, the fact that other autoantibodies may be absent at the onset of the heralding disease but appear at a future time is in favour of the second hypothesis.

Furthermore, it should be kept in mind that even if autoimmunity starts at the same time, the target organs may be destroyed with different latency periods owing to the various size of the implicated glands and/or the ability of the different endocrine cells to regenerate after the autoimmune injury. Obviously, the immunological mechanisms involved are also crucial in the development of the autoimmune disease and the intervention of activated self-reacting T cell is considered to be necessary in the majority of the cases to achieve full destruction of the target organ [91]. In fact, in Addison's disease, chronic lymphocytic thyroiditis and in Type 1 diabetes the relevant circulating autoantibodies that precede, accompany, and follow the clinical diagnosis of these endocrine disorders are not pathogenic in vivo. Nevertheless, some circulating autoantibodies, such as those stimulating the TSH-receptor in Graves’ disease or those blocking the TSH-receptor in the atrophic variant of chronic thyroiditis, as well as those to the intrinsic factor in pernicious anaemia, are pathogenic in vivo.

It is conceivable that the coexistence of blocking antibodies to the TSH receptor (which prevent the regenerating effects of its natural ligand TSH) and infiltrating autoreactive T cells in the thyroid may give rise to an accelerated form of hypothyroidism with atrophy. On the other hand, the presence of stimulating antibodies to the TSH receptor in a chronic lymphocytic thyroiditis may result in a slow-onset hypothyroidism, in ‘mixed’ forms of thyroid dysfunction as the case of Hashitoxicosis [58,68], or in thyroid ‘yo–yo’ syndrome [92].

Therapy

The therapies regarding the different components of Type 2 APS are similar whether they occur as single or in multiple association with other autoimmune diseases. However, it is worth remembering that the thyroid hormone replacement therapy in patients with autoimmune hypothyroidism and misdiagnosed adrenal insufficiency can precipitate an adrenal failure owing to the action of thyroxine in enhancing hepatic corticosteroid metabolism. In addition, some patients with Addison's disease show a reversible increase in thyrotropin levels, regardless of the presence of thyroid autoantibodies, that is related to the loss of the inhibitory effects of glucocorticoids on thyrotropin secretion [93]. Moreover, a reduction in insulin requirement may be the first sign of Addison's disease in a patients with Type 1 diabetes mellitus. Thus, before initiating the therapy with thyroxine or simply modifying insulin dosage, it is prudent to investigate the possible coexistence of an underlying adrenal insufficiency [94].

CONCLUSIONS

It is now well established that organ-specific diseases frequently occur in privileged clusters of association and Type 2 APS is considered one of the most typical. Clinically overt syndrome is considered only the tip of the iceberg, since latent forms are much more frequent. Organ-specific autoantibody screening in patients with monoglandular autoimmune endocrinopathies undoubtedly facilitates the identification of those at risk of developing a future APS. Early identification and treatment of another autoimmune endocrine disease may be critical and even life-saving. Currently, management of these disease is restricted to the pharmacological replacement therapy. However, progress in understanding the inner immunological mechanisms implicated in these conditions, should allow common treatments aimed to prevent or at least dampen the progression to irreversible multiple organ damage.