Basic and Clinical Endocrinology 7th International student edition Edition


Endocrine Autoimmunity

Juan Carlos Jaume MD

Epidemiologic analysis of a large population has reported that about one out of 30 people in the United States (more than 8.5 million individuals) are currently affected by autoimmune diseases. Graves' disease, type 1 diabetes, pernicious anemia, rheumatoid arthritis, chronic thyroiditis (Hashimoto's thyroiditis), and vitiligo are the most prevalent such conditions, accounting for 93% of affected individuals.

These autoimmune diseases have traditionally been looked upon as forming a spectrum. At one end are found organ-specific diseases with organ-specific autoantibodies. Hashimoto's thyroiditis is an example in which a specific lesion affects the thyroid (lymphocytic infiltration, destruction of follicular cells) and autoantibodies are produced with absolute specificity for thyroid proteins. At the other end of the spectrum are the systemic autoimmune diseases, broadly belonging to the class of rheumatologic disorders. Systemic lupus erythematosus is an example of a disease characterized by widespread pathologic changes and a collection of autoantibodies to DNA and other nuclear constituents of all cells. Many organ-specific autoimmune diseases are autoimmune endocrinopathies. Furthermore, autoimmune pathogenesis has been shown to be present in disorders affecting most endocrine glands such as the adrenals (autoimmune Addison's disease), the gonads (autoimmune oophoritis), the pancreas (type 1 diabetes), the pituitary (autoimmune hypophysitis), and the thyroid (autoimmune thyroid disease) (Table 4-1).

Table 4-1. Some autoimmune endocrinopathies, antigens, and autoantibodies.





Autoimmune (lymphocytic) hypophysitis Granulomatous hypophysitis


Pituitary cytosolic protein


Graves' disease




Hashimoto's thyroiditis



Anti-TPO, anti-Tg

Autoimmune (idiopathic) hypoparathyroidism




Type 1 diabetes mellitus

Pancreas (β cells)

GAD65, IA-2, insulin

Anti-GAD, anti-IA-2 (ICA), anti-insulin

Type B insulin resistance with acanthosis nigricans (rare)

Adipocytes, muscle cells

Insulin receptor

Insulin receptor blocking

Autoimmune Addison's disease (autoimmune adrenal failure)


21-Hydroxylase 17α-Hydroxylase P450scc

Anti-21-hydroxylase (ACA) Anti-17α-hydroxylase and anti-P450scc (SCA)

Autoimmune oophoritis (premature ovarian failure)


Not yet identified unequivocally, 17 α-hydroxylase, P450scc

Also SCA in association with adrenal insufficiency

Autoimmune orchitis Male
infertility (some forms)






By far the most common autoimmune endocrine diseases are autoimmune thyroid disease and type 1 diabetes. When the target is the thyroid gland and the clinical manifestation is hypothyroidism (Hashimoto's thyroiditis), the prevalence is about 1%. When the manifestation is hyperthyroidism (Graves' disease), the prevalence is about 0.4%. Both thyroid autoimmune disorders affect women preferentially. When the targets of the autoimmune response are the β cells of the pancreas, the clinical presentation is type 1 diabetes. The prevalence of type 1 diabetes is close to that of Graves' disease (0.2–0.5%); however, it has no gender bias.

Basic immunologic concepts as they apply to clinical autoimmune endocrine diseases as sole entities and as polyglandular failure syndromes are reviewed in this chapter.


The immune system is constantly confronted with a variety of molecules and recognizes them as either self or foreign. The adaptive immune system has evolved to recognize virtually any foreign molecule, either in existence or yet to come. The repertoire of immune recognition molecules randomly formed by gene rearrangements is not limited by the genetic information encoded in the genome (Figure 4-1). As a result, an enormously wide array of immune recognition molecules are acquired by the human immune system. By way of illustration, the theoretical diversity of T cell receptors (T cell recognition molecules) by random rearrangements reaches 1015. This mechanism of rearrangement also applies to B cell recognition molecules, ie, immunoglobulins. The random mechanism of gene rearrangement, however, produces immune recognition molecules that react with self components. Potentially dangerous immune cells carrying self-reactive recognition molecules are eliminated (negatively selected) during development of T lymphocytes in the thymus and of B lymphocytes in the bone marrow. It appears that only immune cells which react with foreign antigen strongly and with self antigen very weakly are positively selected and comprise the peripheral immune cell repertoire. This selection mechanism of immune cells is termed “central tolerance.” Self-reactive immune cells that skip central tolerance and reach the periphery are managed by other control mechanisms against autoimmunity and are either eliminated, rendered unresponsive, or suppressed (“peripheral tolerance”). Failures in these mechanisms of immunologic


regulation, as proposed by Mackay and Burnet in 1964, are central features of the concept of autoimmune disease.


Figure 4-1. Rearrangement of the T cell receptor (TCR) α and β genes to form a functionally diverse receptor. During T cell development, the TCR α and β gene segments rearrange by somatic recombination so that one of the Vα segments pairs with a single Jαsegment, and a Vβ segment pairs with a single Dβ and Jβ segment. The C (constant) segments are brought together with the rearranged segments by transcription and splicing to generate the functional mRNA that will be translated into the α and β protein chains that compose the TCR.


T and B lymphocytes are the fundamental and predominant immune cells. T lymphocyte precursors (pre-T cells) originate in the bone marrow and migrate to the thymus, where they undergo maturation and differentiation. At early stages, they express several T cell surface molecules but still have genomic (not rearranged) configuration of their T cell receptors (TCRs). These now pre-T cells, destined to become T cells with TCR α/β chains (T α/β cells), pass through a critical phase during which self-reactive T cells are deleted by negative selection (see T Cell Tolerance, below). Few pre-T cells will express other types of chains on their TCR (T γ/δ cells). T α/β cells differentiate into either mature CD4 or CD8 cells. These now mature lymphocytes migrate to T cell areas of peripheral lymphoid organs and exert their function as helper (TH) or cytotoxic (TC) cells when activated.

B lymphocytes mature and differentiate in the bone marrow and then migrate to the B cell areas of lymphoid organs. Influenced by factors derived from TH cells previously activated by professional antigen-presenting cells (APCs) such as macrophages, some B cells differentiate to become immunoglobulin M (IgM)-producing cells (plasma cells). Most of the other activated


B cells that do not differentiate into plasma cells revert to the resting state to become memory B cells. When memory B cells are further activated, two events occur: isotype switching (immunoglobulin class switching) and hypermutation of the immunoglobulin variable region to further increase diversity and specificity (affinity maturation).

Activation of B cells requires recognition of the antigen as a whole, while T cells require recognition of antigenic peptides bound to major histocompatibility complex (MHC) molecules on the surfaces of professional APCs. Therefore, T cell recognition is said to be MHC-restricted.

The human MHC (human leukocyte antigen; HLA) consists of a linked set of genes encoding major glycoproteins involved in antigen presentation (Figure 4-2). The complex locates to the short arm of chromosome 6 and divides into three separate regions: class I, class II, and class III genes. The class I “classic” region encodes HLA-A, HLA-B, and HLA-C loci; the nonclassic or class I-related region encodes HLA-E, HLA-F, and HLA-G loci and other immunity-related genes such as CD1. The class II region (HLA-D) encodes HLA-DP, HLA-DQ, and HLA-DR loci and other genes related to antigen processing, transport, and pres-entation such as transporter associated with antigen processing (TAP). The class III region encodes genes for tumor necrosis factors α and β (TNF-α and TNF-β); complement factors C2, C4, and B; and the steroidogenic enzyme 21-hydroxylase. MHC class I (classic) molecules are found on all somatic cells, whereas MHC class I nonclassic antigens are expressed only on some (eg, HLA-F on fetal liver, HLA-G on placental tissues). CD1 molecules are expressed on Langerhans cells, dendritic cells, macrophages, and B cells (all professional APCs). MHC class II molecules are exclusively expressed on these professional APCs. However, virtually all cells except mature erythrocytes can express MHC class II molecules under particular conditions (eg, stimulation with interferon-γ [INF-γ]). As a general rule, MHC class I molecules present peptides derived from endogenous antigens that have access to cytosolic cell compartments (eg, virus) to CD8 TC cells. On the other hand, MHC class II molecules present peptides derived from antigens internalized by endocytosis into vesicular compartments (eg, bacteria) to CD4 TH cells. MHC class II molecules also bind peptides derived from many membrane-bound self antigens.

Antigen-presenting cells (APCs) process and present antigen in order to activate T cells utilizing MHC-peptide presentation (Figure 4-3). T cells require at least two signals to become activated. The interaction of a TCR expressed on antigen-specific T cells and the antigenic peptide-MHC complex expressed on APCs provides the first signal. The second signal is delivered by the interaction between costimulatory molecules CD80 (B7.1) and CD86 (B7.2) on APCs and CD28 on T cells. These two signals induce proliferation of T cells, production of interleukin-2 (IL-2), and expression of the anti-apoptotic protein Bcl-xL. TH cells and TC cells are effector cells that require both signals in order to become activated. However, TC cells also need the “help”provided by TH cells. Until very recently, it was thought that TH and TC cells needed to interact with the same APC simultaneously and that cytokines (such as IL-2) produced by the TH cell would then act on the TC cell to facilitate its response. New studies suggest that the interaction between another costimulatory molecule, CD40 ligand (CD154), present on T cells, and CD40, present on APCs, may provide an alternative explanation. It appears that TH cells recognizing antigenic peptides presented by APCs deliver a signal


through the CD154-CD40 complex that “licenses” APCs to directly stimulate TC cells (Figure 4-4). Thus, there is no need for simultaneous interactions of TH and TC cells while encountering the APC. CD154-CD40 interaction also enhances expression of CD80 and CD86 as well as secretion of cytokines (IL-1, -6, -8, -10, and -12 and TNF-α).


Figure 4-2. Gene organization of the human major histocompatibility complex or human leukocyte antigen (HLA) complex. Regions encoding the three classes of MHC proteins on top. APTP denotes a cluster of genes within the class II region encoding genes related to antigen processing, transport, and presentation. Class III region encodes genes unrelated to class I or class II not involved in antigen presentation (TNF-α and -β, complement factors C2, C4, B, and 21-hydroxylase and others).


Figure 4-3. Antigen recognition by T cells. From top to bottom: A. Antigen uptake: Incorporation of antigen (via phagocytosis, pinocytosis, or FcR-mediated endocytosis of immune complex). B. Antigen presentation: APCs deliver an antigen-specific signal through the MHC-peptide-TCR interaction on T cells (MHC I coupled to CD8 interacts with TC cells, MHC II coupled to CD4 interacts with TH cells). C. T-cell activation: The required second signal is provided via CD80/86 (B7.1; B7.2)-CD28 that induces the expression of CD154 (CD40 L) first and CD152 (CTLA-4) later on. Binding of CD154 on T cells with CD40 on APCs enhances expression of CD80/86. The APC-CD80/86 increased expression and consequent binding of CD28 on T cells perpetuates the activation and proliferation of these effector cells (downstream events). D. T-cell inactivation: CD152 (expressed 48–72 hours after T-cell activation) will preferentially bind to CD80/86 on APCs because of its higher affinity, displacing CD28 and in turn suppressing T-cell activity.

Yet another molecule on T cells, the CD28 homolog cytotoxic T lymphocyte antigen 4 (CTLA-4 or CD152), functions to suppress T cell responses (Figure 4-3). CD152 is expressed at low to undetectable levels on resting T cells. It is up-regulated by the ligation of CD28 on T cells with CD80/86 on APCs, or by IL-2. CD152 and CD28 on T cells share the same counterreceptors, namely, CD80/86 on APCs. However, CD152 has a 20-fold higher affinity than CD28 for their ligands.

The integration of all these interactions may be as follows (Figure 4-3): After processing antigen, APCs deliver an antigen-specific first signal through the MHC-peptide-TCR interaction on T cells. A second signal is provided by a costimulatory interaction of the CD80/86-CD28 complex that induces the expression of CD154 first and then CD152. Binding of CD154 on TH cells with CD40 on APCs enhances expression of CD80/86 and licenses APCs for direct activation of TC cells. Other inflammatory cytokines as well as lipopolysaccharides and viruses may do the same. The increased expression of APC-CD80/86 and consequent binding of CD28 on T cells then perpetuates the activation and proliferation of these effector cells. However, the expression of CD152 48–72 hours after T cell activation will lead to the preferential binding of this molecule to CD80/86 on APCs because of its higher affinity for CD80/86. This may displace CD28 from CD80/86 and, in turn, suppress T cell activity.

Activation and differentiation of B cells often require two signals also. Naive B cells are triggered by antigen but may also require accessory signals that come from activated TH cells. Some antigens can directly activate naive B cells without the need for TH cells (eg, lipopolysaccharides from gram-negative bacteria or polymeric protein structures). The former type of B cell activation (MHC class II-restricted T cell help) is called thymus-dependent (TD). The latter type is called thymus-independent (TI). TH cells also control isotype switching and initiate somatic hypermutation of antibody-variable region genes (see Tolerance, below).


Interaction between CD154 on TH cells and CD40 on B cells and the cytokines produced by TH cells are essential for isotype switching and formation of germinal centers in peripheral lymphoid organs. The immunoglobulin isotype switching is critical for the generation of functional diversity of a humoral immune response. Somatic hypermutation (point mutations of the variable region genes of immunoglobulins during the course of an immune response) is necessary for the affinity maturation of antibodies.


Figure 4-4. Licensed APCs directly activate TC cells. According to the traditional model, TH cells and TC cells recognize antigen on the same APC. The APC-activated TH cell produces IL-2, which contributes to the activation of TC cells while in simultaneous interaction with the same APC. According to the proposed new model (A), APCs are licensed to activate TC cells by TTH or other stimuli (lipopolysaccharides, INF-γ, viruses). APCs first interact with TH cells. The association of CD154 (CD40 L) on the TTH cell and CD40 on the APC allows (or licenses) the latter to activate TTC cells directly (B). Thus, once licensed, APCs are capable of activating cytotoxic T cells without the need of simultaneous interaction with TTH cells.

Overall, the immune response is a combination of effector mechanisms that function to eliminate pathogenic organisms. These effector mechanisms include, as innate immunity, phagocytosis (by macrophages, neutrophils, monocytes, and dendritic cells) and cytotoxicity (by natural killer [NK] cells); and as adaptive immunity, antibody-dependent complement-mediated cytotoxicity, antibody-dependent cell-mediated cytotoxicity (ADCC), cytotoxicity by T γ/δ cells that recognize heat shock proteins on target cells, and cytotoxicity by CD8 or CD4 TC cells. CD8 and CD4 TC cells are activated by the described recognition of specific antigenic peptides bound to class I (for CD8), class II (for CD4) MHC molecules on the APCs and classically by IL-2 from nearby activated CD4 TH cells. These cells kill the target by either secreting cytotoxins (perforin, granzyme) or by inducing apoptosis through the Fas-FasL interaction (see below).

The specificity of the immune response is crucial if self-reactivity is to be avoided. In order to ensure that lymphocyte responses and the downstream effector mechanisms they control are directed exclusively against foreign antigens and not against “self” components, a number of safety-check barriers must be negotiated before autoreactive lymphocytes can differentiate and proliferate.


T Cell Tolerance

T cells developing in the thymus (pre-T cells) are destined to become T α/β cells through rearrangement of the TCR β gene initially, followed by the TCR α gene (Figure 4-5). If unproductive rearrangements of TCR genes occur (nonfunctional TCR α or β proteins), apoptosis of these pre-T cells follows (Figure 4-5A). If functional rearrangements of TCR α and β proteins occur,


cells express TCR α/β dimer and CD3 molecules at low levels on the cell surface. TCR-rearranged cells proliferate 100-fold. Positive and negative selection occurs based on the ability of the rearranged TCR α/β to recognize antigenic peptides in association with self-MHC molecules on thymic epithelial and dendritic cells. Negative selection (clonal deletion) appears to take place in the thymus medulla, where pre-T cells bearing TCRs specific for self peptides bound to self-MHC molecules are deleted. At least 97% of developing T cells undergo apoptosis within the thymus (“central tolerance”). Positively selected pre-T cells increase expression of TCR α/β, express either CD4 or CD8, and become mature T cells. These mature T cells exit the thymus and go to the periphery. CD4 T cells are activated in the periphery in an MHC class II-restricted fashion, while CD8 T cells are activated in an MHC class I-restricted fashion.

A differential avidity model in which the fate of T cells is determined by the intrinsic affinity of TCRs for their ligands has been advanced to explain the paradox between positive and negative selection. According to this model, T cells with high avidity for MHC-self peptide complexes would be eliminated (negative selection), whereas T cells with low avidity to MHC-self peptide complexes would be positively selected. If the avidity is close to zero, T cells would not be selected (for lack of effective signal to survive). The biochemical factor or factors that signal survival (low avidity of TCR binding) versus apoptosis (triggered by high avidity interactions) have yet to be found.

Costimulatory interactions between CD28 and CD80/86 and between CD154, CD40, and adhesion molecules, such as lymphocyte function-associated antigen-1 (LFA-1), are also involved in preferential deletion of self-reactive T cells in the medullary region of the thymus. It is known that negative selection is not 100% effective and that some potentially autoreactive T cells do escape to the periphery. Not all self peptides, including those derived from proteins expressed in highly tissue-specific organs such as endocrine glands, would be presented to pre-T cells during their development in the thymus. Therefore, the peripheral immune system must maintain tolerance through complementary control mechanisms.

“Peripheral tolerance” (Figure 4-5B) may be maintained by the induction of unresponsiveness to self antigen (anergy) or by the induction of regulatory cells, such as suppressor T cells (active suppression). Peripheral clonal deletion (apoptosis) of autoreactive T cells that have escaped from the thymus may play an important role in limiting rapidly expanding responses, but there are many examples where autoreactive T cells persist. Some autoreactive T cells may never encounter the self antigen because it may be sequestered from the immune system (ignorance). Lastly, immune deviation, whereby noninflammatory TH2 responses suppress an autoreactive inflammatory TH1 response, inducing peripheral tolerance, deserves further discussion (see also Autoimmune Aspects of Thyroid Disease, below). TH1 cells, which regulate cell-mediated responses, secrete interferon-γ (INF-γ) and small amounts of IL-4. In contrast, TH2 cells, which provide help for antibody production, secrete abundant IL-4 and little INF-γ. A prevailing concept in human autoimmunity is that TH1 responses are believed to dominate. It has been shown in animal models that induction of TH2 responses ameliorates TH1 responses. Hence, unbalanced TH1 immune deviation may lead to a breakage of peripheral tolerance. However, evidence to the contrary exists in some endocrinopathies. (See Autoimmune Response in the section on Autoimmune Aspects of Thyroid Disease, below.)

Clonal deletion and anergy occur through apoptosis at the site of activation or after passage through the liver. High antigen dose and chronic stimulation induce peripheral elimination of both CD4 and CD8 T cells. Activated T cells express Fas molecules on their surfaces but are resistant to Fas ligand-mediated apoptosis because of the simultaneous expression of Bcl-xL, induced by CD28 ligation during activation (see Immune Recognition and Response, above). Several days after activation, when Bcl-xL has declined, CD4 cells become susceptible to Fas-mediated apoptosis (activation-induced cell death; AICD). A similar mechanism via p75 tumor necrosis factor (TNF) receptor has been shown for CD8 cells. Therefore, autoreactive T cells might be deleted by apoptosis induced by chronic stimulation with self antigens, present abundantly in the periphery. However, autoreactive T cells specific for very rare self antigens may be difficult to eliminate.

Anergy also results from the lack of a second costimulatory signal. When nonhematopoietic cells stimulated by INF-γ present antigen in an MHC class II-restricted fashion (as thyrocytes do in AITD), autoreactive T cells may be rendered unresponsive because of the absence of a CD28-CD80/86-mediated signal (nonhematopoietic cells do not express CD80/86 as professional APCs do). However, even if the two signals are provided, anergy may result from the lack of TH cell-originated cytokines (IL-2, -4, -7, etc). Recently, it has also been shown that in vivo T cell anergy may be induced by CD80/86-CD152 interaction (see also Immune Recognition and Response, above).

T cell active suppression is considered to be a major regulatory mechanism of peripheral tolerance; however, its mode of action is still under study. As mentioned above, nonhematopoietic cells stimulated by INF-γ present antigen in an MHC class II-restricted fashion to T cells and render them anergic. These nonhematopoietic cells (nonprofessional APCs) may also present to CD4 T




suppressor (TS) cells. Before becoming unresponsive, these cells may induce specific CD8 T suppressor (TS) cells. In turn, these CD8 TS cells may regulate (via T cell suppressor factors or cytotoxicity) antigen-specific autoreactive T cells (see also Figure 4-5B).


Figure 4-5. A: Central T cell tolerance. Mechanisms of central tolerance (at the thymus level) are depicted. From top to bottom, pre-T cells first rearrange their TCR. Unproductive (nonfunctional) rearrangements lead to apoptosis, while productive ones engage pre-T cells in self antigen recognition. Clonal deletion indicates elimination of cells based on their high or no avidity for self antigen (apoptosis). Surviving low-avidity cells reach the periphery as mature CD4 and CD8 cells. B: Peripheral T cell tolerance. May be accomplished through any of the five depicted mechanisms. 1. Clonal deletion: After encountering self antigen in the context of self MHC molecules and simultaneous delivery of a second signal (CD80/86-CD28) by APCs (top left), autoreactive T cells become activated. These activated T cells express Fas molecules on their surface but are resistant to Fas ligand-mediated apoptosis because of the simultaneous expression of Bcl-xL (not shown) induced by CD28 ligation during activation. Several days after activation, when Bcl-xL presence has declined, CD4 cells become susceptible to Fas ligand-mediated apoptosis. Natural killer cells (NK-T) may then accomplish the task of eliminating these autoreactive T cells. 2. Anergy: Anergy may be induced via CD80/86-CD152 interaction 48–72 hours following activation or may result from the lack of a second costimulatory signal from APCs presenting self antigen (nonprofessional APCs). 3. Active suppression: Active suppression is thought to occur when nonhematopoietic cells (stimulated by INF-γ) present antigen in an MHC class II restricted fashion to CD4 T suppressor (Ts) cells. Before becoming unresponsive, these cells may induce specific CD8 Ts cells. In turn, these CD8 Ts cells may suppress antigen-specific autoreactive T cells. 4. Ignorance (top right): Some autoreactive T cells may never encounter self antigen because it may be sequestered from the immune system. Although they may persist in the circulation, they never become activated. 5. Immune deviation: Under specific circumstances, noninflammatory TH2 responses could suppress inflammatory (autoreactive) TH1 responses (see text).

B Cell Tolerance

Instead of the thymus, the bone marrow provides the setting for central B cell tolerance. Pre-B cells rearrange their B cell receptor (BCR or membrane-bound immunoglobulin) early in development. The immunoglobulin heavy (H) chain genes rearrange first, followed by light (L) chain gene rearrangement. Unproductive rearrangements and pairings leading to formation of nonfunctional immunoglobulin drive pre-B cells to apoptosis (Figure 4-6A). Functional rearrangements (functional BCRs) allow immature B cell expansion and expression of IgM and CD21 (a marker of functionality). Only one-third of the precursor cells reach this stage. The random rearrangement of the V, D, and J segments of immunoglobulin genes during this period inevitably generates self-recognizing immunoglobulins. Negative selection of autoreactive B cells occurs at the immature B cell stage on the basis of the avidity of the BCR for self antigens. Similar to the T cell clonal deletion, immature B cells that strongly bind antigens in the bone marrow are eliminated by apoptosis. Some autoreactive immature B cells, instead of undergoing apoptosis, resume rearrangements of their L chain genes in an attempt to reassemble new κ or λ genes. This procedure, called BCR editing, permanently inactivates the autoreactive immunoglobulin genes. Soluble antigens, presumably because they generate weaker signals through the BCR of immature B cells, do not cause apoptosis but render cells unresponsive to stimuli (anergy).These anergic B cells migrate to the periphery, where they express IgD. They may be activated under special circumstances, making anergy less than sufficient as a mechanism of enforcing tolerance. Only immature B cells in the bone marrow with no avidity for antigens (membrane-bound or soluble) become mature B cells with the capacity to express both IgM and IgD. As with T cells, 97% of developing B cells undergo apoptosis within the bone marrow. Also, and as with T cells, central clonal deletion, anergy, and BCR editing eliminates autoreactive B cells, recognizing bone marrow-derived self antigens.

Peripheral B cell tolerance (Figure 4-6B) is also crucial for protection against autoimmunity. It appears that in the absence of antigen, mature B cells are actively eliminated in the periphery by activated T cells via Fas-FasL and CD40-CD154 interactions. In the presence of specific antigen but without T cell help, antigen recognition by BCRs induces apoptosis or anergy of mature B cells. If antigen and specific T cell help are provided—ie, if antigen bound to the BCR is internalized, processed, and presented in an MHC class II-restricted fashion to a previously activated TH cell specific for the same antigen—two events occur. One, the B cell becomes an IgM-secreting plasma cell, and—in the presence of the appropriate cytokines and after expression of




CD40 (for TH cell CD154 interaction)—class switching occurs. Two, further somatic hypermutation of the immunoglobulin variable region genes of such mature B cells, which changes affinity of BCRs for antigens, also occurs in germinal centers (see also Immune Recognition and Response, above). Mutants with low-affinity receptors undergo apoptosis, while enhanced-affinity BCRs are positively selected. In the presence of CD40 ligation of CD154, antigen-stimulated B cells become memory B cells (Figure 4-6B).


Figure 4-6. A: Central B cell tolerance. As T cells do in the thymus, B cells rearrange their B cell receptor in the bone marrow. Unproductive rearrangements drive pre-B cells to apoptosis. Functional rearrangements allow expansion and expression of IgM. Next, similar to T cell clonal deletion, immature B cells that strongly bind self antigens in the bone marrow are eliminated by apoptosis. Some autoreactive immature B cells, instead of becoming apoptotic, however, resume rearrangements of their L chain genes, attempting to reassemble new allelic κ or λ genes (BCR editing). Soluble self antigens presumably generate weaker signals through the BCR of immature B cells; they do not cause apoptosis but make cells unresponsive to stimuli (anergy). These anergic B cells migrate to the periphery, expressing IgD, and may be activated under special circumstances. Only immature B cells with no avidity for antigens become mature B cells, expressing both IgM and IgD. These are the predominant cells that make it to the periphery. B: Peripheral B cell tolerance. In the “absence” of antigen (top right), mature B cells are actively eliminated by activated T cells via Fas-FasL and CD40-CD154 interactions. In the “presence” of specific self antigen but“without T cell help,” antigen recognition by BCRs induces apoptosis or anergy on mature B cells. If self antigen and specific autoreactive T cell help are provided, two events develop (center): (1) The B cell becomes an IgM-secreting plasma cell (top left), and, in the presence of the appropriate cytokines after expression of CD40 (for TH cell CD154 interaction), class switching occurs (bottom left). (2) Further somatic hypermutation of the Ig-variable region genes, which changes affinity of BCRs, occurs. Mutants with low-affinity receptors undergo apoptosis, while improved-affinity BCRs are positively selected. In the presence of CD40 ligation of CD154, antigen-stimulated B cells become memory B cells. These two events are the same as in foreign antigen recognition.

The ability of mature B cells to capture very low quantities of antigen via high-affinity BCRs allows them to amplify their antigen-presenting capacity to more than 1000 times that of other professional APCs. This particular property may become critical when developing chronic organ-specific autoimmune diseases in which the source of antigen is limited. Thus, autoreactive B cells that happen to escape the control mechanisms described could amplify and perpetuate autoimmune responses in patients with failing endocrine organs when tissue destruction has left only minute amounts of residual antigen.


Although the breakage of self-tolerance seems to be a central pathogenic step in the development of autoimmune diseases, autoimmunity is a multifactorial event. Specifically, defects in apoptosis-related molecules (Fas-FasL) of thymic dendritic cells have been shown to impair central clonal deletion. Also, in the periphery, similar defects (Fas-FasL, CD152) on T cell-APC molecules may prevent apoptosis of autoreactive T cells. However, it is difficult to consider these general defects as causative of organ-specific disorders. Furthermore, clonal ignorance of T cells cannot be maintained if antigens sequestered from the immune system are released in blood or if cryptic epitopes of antigens that have never been recognized by the immune system are presented to T cells for recognition (after tissue destruction, for example). Defects of active suppression, immune deviation (TH1/TH2 imbalance), and defects in B cell tolerance may all be involved in the pathogenesis of autoimmune diseases. How and why loss of immune self-tolerance occurs is not completely understood. Both genetic and environmental factors appear to be necessary.


Epidemiologic studies demonstrate that susceptibility to most autoimmune diseases has a significant genetic component. In type 1 diabetes, for example, there is a clear association between race and susceptibility to disease—the incidence is approximately 40 times higher in Finland than in Japan. Family studies also demonstrate a strong underlying genetic component. The lifetime risk of developing type 1 diabetes in the United States general population is 0.4%, whereas in relatives of diabetics the risk is substantially higher (4% for parents, 5–7% for siblings, 20% for HLA-identical sibs, 25–40% for monozygotic twins).

The inheritance pattern of autoimmune disorders is complex. These disorders are polygenic, arising from


several independently segregating genes. The most consistent genetic marker for autoimmune diseases to date is the MHC genotype. Considering again genetic susceptibility to type 1 diabetes, up to 95% of Caucasians developing diabetes express the HLA alleles DR3 or DR4—compared with about 40% of normal individuals. Individuals heterozygous for both DR3 and DR4 have the highest risk. It has been shown that the DQ rather than DR genotype is a more specific marker of susceptibility and that the association of both markers is due to the fact that they are products of closely linked genes. But what is more important than the fact that HLA genes are linked to diabetes is that HLA haplotypes are no longer simply undefined genetic markers. It has been shown that the polymorphisms of the DQ molecules are critical for high-affinity recognition of autoantigens (eg, islet cell antigens) by TCRs. Although the crystal structure of HLA-DQ has not yet been defined, comparisons with HLA-DR structure suggest that the lack of aspartic acid at position 57 (Asp57) on the DQ β chain allows the autoantigen (processed peptide) to fit better in the antigen-binding groove formed by this molecule. To the contrary, the presence of Asp57allows the formation of a salt bridge with a conserved arginine at position 76 on the DQ α chain, preventing the accommodation of the immunogenic peptide recognized by the TCR. Several autoimmune diseases have been linked to HLA-DQβ1 genes, including type 1 diabetes, celiac disease, bullous pemphigoid, autoimmune hepatitis, and premature ovarian failure, and the structure of the DQβ1 molecule may be the reason for the increased susceptibility.

Other candidate genes associated with autoimmune endocrinopathies are discussed further under the single and polyglandular syndromes.


Environmental factors also play a critical role in the pathogenesis of autoimmune disease. The strongest evidence for this statement comes from studies of monozygotic twins, which show that concordance rates for autoimmune disorders are imperfect (never 100%). As mentioned above, in type 1 diabetes, identical twins show less than 50% concordance.

The environmental factors thought to have greatest influence on disease development include infectious agents, diet, and toxins. In type 1 diabetes, viruses have been strongly suspect. Up to 20% of children prenatally infected with rubella develop type 1 diabetes. Children with congenital rubella also have an increased incidence of other autoimmune disorders, including thyroiditis and dysgammaglobulinemia. The mechanisms by which these pathogens may induce autoimmune responses include molecular mimicry and direct tissue injury. The hypothesis of molecular mimicry suggests that immune responses directed at infectious agents can cross-react with self antigens, causing tissue or organ destruction. Support for this concept is found in well-known clinical syndromes, eg, rheumatic fever (immune responses directed against streptococcal M protein seem to cross-react with cardiac myosin, inducing clinical myocarditis). In autoimmune diabetes, the best-studied example of molecular mimicry is the B4 coxsackievirus protein P2-C. Coxsackie B4 virus has also been epidemiologically implicated in the development of type 1 diabetes. There is a striking amino acid sequence similarity between P2-C viral protein and the enzyme glutamic acid decarboxylase (GAD), found in pancreatic β cells (see Autoimmune Aspects of Type 1 Diabetes, below).

The importance of diet in the development of autoimmune diseases remains controversial. An association between early exposure to cow's milk proteins and the risk of type 1 diabetes has been observed in several epidemiologic studies. For example, one study demonstrated that primary immunity to insulin is induced in infancy by oral exposure to cow's milk insulin, but the relevance of this observation is still unknown. On the other hand, selected antigens (from bovine serum albumin to porcine insulin) have been administered orally to mice with a broad spectrum of autoimmune disorders, including nonobese diabetic (NOD) mice, with favorable outcomes. Those data in mice were so compelling that oral tolerance trials in humans have been conducted or are ongoing. Unfortunately, the results of already completed trials in other autoimmune diseases have been disappointing. As examples of ongoing trials in diabetology, the Diabetes Prevention Trial 1 (DPT-1) in the United States and the DIOR study (in recently diagnosed diabetic children and adults) in France are currently attempting to prevent type 1 diabetes in high-risk patients with oral or parenteral insulin. However, another study has shown that oral administration of autoantigen to mice may induce a cytotoxic T lymphocyte response leading to the onset of autoimmune diabetes. This suggests that caution should be used when applying this approach to the treatment of human autoimmune diseases.


Organ-specific autoimmune endocrine disorders may present as single entities or may cluster in polyglandular syndromes. Most endocrine glands are susceptible to


autoimmune attacks. Some are affected more frequently than others (Table 4-1).


Autoimmune thyroid disease can present in a polarized fashion with Graves' disease (thyroid hyperfunction) at one end and Hashimoto's thyroiditis (thyroid failure) at the other. This functional subdivision is clinically useful. However, both diseases have a common autoimmune origin.

Genes & Environment

Major susceptibility genes in autoimmune thyroid disease have yet to be identified. Although certain HLA alleles (mainly HLA-DR3 and DQA1*0501) have been shown to be present more frequently in Graves' disease than in the general population, this association has frequently been challenged. In fact, no consistent association has been found between Graves' disease and any known HLA polymorphism. Furthermore, the risk of developing Graves' disease in HLA-identical siblings (7%) is not significantly different from that in control populations. HLA-DR5, -DR3, -DQw7 in Caucasian, -DRw53 in Japanese, and -DR9 in Chinese patients were found to be associated with Hashimoto's thyroiditis. However, genetic linkage between Hashimoto's thyroiditis and a specific HLA locus has not been demonstrated consistently. Overall, the HLA loci are likely to provide less than 5% of the genetic contribution to autoimmune thyroid disease, confirming the relative importance of non-HLA related genes in susceptibility. For example, it has been shown that the inheritance pattern of autoantibodies to thyroperoxidase (TPO) is genetically transmitted. Other candidates are currently under study. However, autoimmune thyroid disease linkage to CTLA-4, HLA, IgH chain, TCR, thyroglobulin (Tg), TPO, and thyrotropin receptor (TSH-R) genes has been excluded.

An important environmental factor influencing the natural history of autoimmune thyroid disease is that of iodine intake (dietary, or present in drugs such as amiodarone, x-ray contrast media). There is considerable evidence that iodine adversely affects both thyroid function and antibody production in those with occult or overt autoimmune thyroid disease.

Autoimmune Response

In Graves' disease, thyrocytes are the differentiated carriers of TSH-Rs and the target cells of autoantibodies and probably the autoimmune response. The development of autoantibodies that functionally stimulate the TSH-R mimicking the action of TSH was the first example of antibody-mediated activities of a hormone receptor in humans. Autoantibodies that may stimulate the calcium-sensing receptor (another G-protein couple receptor) and signal the inhibition of PTH production have been described in autoimmune hypoparathyroidism. Similarly, stimulating antibodies that bind to the adrenocorticotropin (ACTH) receptor may be involved in the pathogenesis of primary pigmented nodular adrenocortical disease (also referred as nodular adrenal dysplasia).

In Graves' disease, antibodies to the TSH-R present with different types of activity. Thyroid-stimulating immunoglobulins (TSIs), the cause of the hyperthyroidism, are most frequently detected by a bioassay that measures cAMP production in a rat thyroid cell line (FRTL5). TSH-blocking autoantibodies which may produce hypothyroidism are identified by their ability to prevent TSH binding to TSH-R. This has allowed development of the TSH-binding inhibition (TBI) assay. No direct immunoassay for the measurement of TSH-R autoantibodies is available yet, and its development may be difficult because TSH-R autoantibodies are present at very low concentrations.

A particular feature of Graves' disease is its early clinical presentation. Unlike other autoimmune endocrinopathies (type 1 diabetes, Hashimoto's thyroiditis, autoimmune Addison's disease), in which much of the target organ has to be destroyed before the disease is manifested, Graves' hyperthyroidism presents with an enlarged and active gland. Minimal lymphocytic infiltration is present when hyperthyroidism (due to the presence of TSH-R-stimulating antibodies) develops. This unique feature may ultimately allow early immune intervention in preference to current ablative therapeutic options.

Another peculiar feature of Graves' disease is the helper T cell response observed in this disease. The activation of antibody-producing B cells by T helper (TH) lymphocytes in Graves' disease is well recognized. At present, a prevailing concept of human autoimmunity suggests that as in acute allograft rejection, “deviation” toward a TH1 response dominates its pathogenesis. Counterdeviation toward TH2 is thought to be a consequence of tolerance induction and has been postulated as a potential therapeutic approach. Graves' disease seems to challenge that concept. Analysis of TSH-R-specific T cell clones from patients with Graves' disease has provided direct evidence for polarization of THresponses; however, instead of TH1 deviation, TH0 and TH2 responses have been observed. As mentioned before, TH1 cells, which regulate cell-mediated responses, secrete mainly IFN-γ and small amounts of IL-4. In contrast, TH2 cells that regulate antibody production


(such as TSH-R autoantibodies in Graves' disease) preferentially produce IL-4 and little IFN-γ. T cells expressing both IL-4 and IFN-γ are known as TH0 cells. These experimental results suggest that in Graves' disease TH0-TH2 cell responses appear to be dominant. Hence, in human autoimmunity, Graves' disease appears to be an exception to the usual TH cell pattern.

In Hashimoto's thyroiditis, the hallmark of the humoral immune response is the presence of autoantibodies to TPO. Although the effector mechanism for TPO (or thyroglobulin) autoantibodies is still controversial, under special circumstances (at least in vitro) the autoantibodies are themselves cytotoxic agents or activators of cytotoxic T lymphocytes. Furthermore, in secondary T cell responses, antibodies may play a critical role in antigen processing or presentation to T cells. In short, macrophages internalize (and subsequently process) antigen by phagocytosis and antigen-antibody complex uptake via Fc receptors. B cells have membrane-bound antibodies (B cell receptors; BCRs) which provide a much more powerful system for antigen capture. Indeed, recombinant TPO-specific membrane-bound autoantibody captures antigen and allows presentation efficiently. Antibody binding also modulates antigen processing of immune complexes, enhancing or suppressing the presentation of different T cell peptides. Hence, APCs (internalizing immune complexes through Fc receptors) and B cells (capturing antigen through BCRs) can influence the secondary T cell response that perpetuates autoimmune disease. The potential role of autoantibodies in modulating presentation of T cell determinants in thyroid (and diabetes) autoimmunity is being explored.

Animal Models of Autoimmune Thyroid Disease

The classic immunologic approach to development of an animal model of an autoimmune disease is to immunize the animal with soluble antigen in adjuvant. For autoimmune thyroid disease, the induction of thyroiditis in rabbits using human thyroglobulin (Tg) was one of the earliest attempts to do this by Rose and coworkers in 1956. In subsequent studies, mice immunized with human or murine Tg developed experimental autoimmune thyroiditis. Immunization with TPO (human or porcine) induces thyroid autoantibodies and, as in the case of Tg, causes thyroiditis to develop in particular MHC strains of mice. However, unlike spontaneous thyroiditis in chickens, none of the immunized mouse models of thyroiditis develop hypothyroidism.

In 1996, Shimojo and coworkers developed a mouse model that clearly mimics some of the major features of Graves' disease. This was achieved by the ingenious approach of immunizing mice with fibroblasts stably transfected with the cDNA for the human TSH-R and syngeneic MHC class II (Figure 4-7). Most of the animals had moderately high TBI activity in their sera, and about 25% were clearly thyrotoxic, with elevated T4 and T3 values, detectable TSI activity, and thyroid hypertrophy. For the first time, therefore, an animal model has been established in which a significant number of affected subjects have the immunologic and en-docrinologic features of Graves' hyperthyroidism. However, it is not clear that the immune response is confined to the thyroid gland. Nevertheless, this model opens the way to investigation of the pathogenesis of Graves' disease.


Most type 1 diabetes results from autoimmune destruction of pancreatic β cells in a process that can span several years. This results in glucose intolerance and clinical disease when the majority of β cells have been destroyed. The destruction is marked by circulating antibodies to pancreatic β cells and by massive infiltration of mononuclear lymphocytes into the islets of Langerhans


where pancreatic β cells remain. The lymphocytes disappear when the β cells are gone. Although insulin is available for replacement therapy, type 1 diabetes remains a chronic disorder of major socioeconomic impact that mainly afflicts the young. Elucidation of the molecular mechanisms underlying this destruction—and the development of methods to prevent autoimmunity—may ultimately lead to effective treatment. Such developments, however, require animal models of type 1 diabetes that closely resemble the disease in humans.


Figure 4-7. Animal model for Graves' disease. This animal model was achieved by injecting AKR/N mice with syngeneic (MHC-identical) fibroblasts dually transfected with mouse MHC class II (H2-k) and human TSH-R cDNA. About 25% of the animals developed endocrinologic (hyperthyroidism) and immunologic (TSI, TBII) features of Graves' disease.

Genes & Environment

The susceptibility to develop type 1 diabetes is associated with certain alleles of the MHC class II locus that have been statistically linked to a variety of autoimmune disorders. The most recent analyses indicate that in Caucasians, HLA-DR3 DQ2 (DQB1*0201) and HLA-DR4 (DRB1*0401), DQ8 (DQB1*0302) haplotypes are most strongly associated with type 1 diabetes. In Asian populations, DRB1*0405 is the major susceptibility haplotype. In contrast, the DR2, DQ6 (DQB1*0602) haplotype is negatively associated with type 1 diabetes. More importantly, susceptibility requires both HLA-DQ β chain alleles to be negative for aspartic acid at position 57 (Asp57) on the amino acid sequence. Studies of different populations have shown a linear relationship between the incidence of type 1 diabetes and the estimated frequency of homozygous absence of Asp57.

Non-HLA candidate genes consistently associated with type 1 diabetes include the “variable number of tandem repeats” (VNTR) polymorphisms in the insulin gene and the CTLA4 gene (CD152). The VNTR polymorphisms are located adjacent to defined regulatory sequences that influence insulin gene expression. Of immunologic importance, CTLA4 gene (see Immune Recognition and Response, above) is the other non-HLA candidate gene consistently found to be associated with type 1 diabetes.

Although environmental factors definitely play a role in the development of type 1 diabetes (eg, Coxsackie B4 virus, mumps virus, rubella virus, Kilham rat virus in the BB-rat; or cow's milk formula exposure), more studies are needed to establish a definite etiologic link.

Autoimmune Response

The autoantibodies associated with β cell destruction can be present up to several years before the clinical onset of disease and are thus excellent markers of disease risk. Furthermore, they have served as important tools to identify human pancreatic β cell autoantigens. In 1990, Baekkeskov and coworkers identified a 64-kDa islet cell protein as the smaller isoform of the GABA-synthesizing enzyme glutamic acid decarboxy-lase (GAD65). This autoantigen was shown to be recognized by 70–80% of prediabetic and newly diagnosed type 1 diabetic patients' sera. A second component of the 64-kDa antigen was shown to be a putative tyrosine phosphatase, termed IA-2. IA-2 is recognized by 60–70% of prediabetic and newly diagnosed type 1 diabetic patients. Together, GAD65 and IA-2 autoantibodies detect over 90% of individuals who develop type 1 diabetes and can be used to detect individuals at risk several years before the clinical onset of disease.

Although autoantibody responses to GAD65 are not easily detected, there is strong evidence to suggest that GAD65 is an important T cell autoantigen in the nonobese diabetic (NOD) mouse. Thus, GAD65 is the earliest known target of the autoimmune T cell response in the NOD mouse. Administration of the protein in a tolerogenic form prevents disease. In contrast, induction of tolerance to other potential autoantigens in this model (such as carboxypeptidase H and hsp60) does not prevent disease. The NOD mouse does not develop autoimmunity to the IA-2 molecule and thus distinguishes itself from the human disease with regard to this target antigen (see Models, below).

Insulin is a third well-characterized autoantigen in type 1 diabetes. Insulin autoantibodies can be detected in about 50% of newly diagnosed children with type 1 diabetes. Insulin-specific T cell clones can transfer disease in the NOD mouse. Furthermore, administration of whole insulin, insulin B chain, or an insulin peptide epitope in a tolerogenic form can protect against disease in NOD mice. Because animals receiving insulin or insulin B chain continue to have intra-islet insulitis—in contrast to young NOD mice treated with GAD65 in a tolerogenic way—it has been suggested that insulin reactivity is more distal in disease progression. Additional but less well characterized proteins have been implicated as targets of autoantibodies in type 1 diabetes in humans.

Autoantibodies, although they are good markers of disease, do not seem to be directly involved in destruction of pancreatic β cells. Adoptive transfer of diabetes to NOD mice with spontaneous combined immune deficiency lacking T cells (NOD-SCID) can be mediated by T cells alone. Because β cell-deficient NOD mice do not develop disease, it is possible that B lymphocytes function as important APCs in the islet to perpetuate an ongoing autoimmune response and thus are essential for presentation of rare antigens such as GAD65 and I-A2. (See alsoAutoimmune Response in the section on Autoimmune Aspects of Thyroid Disease, above.)

An important question is whether GAD65, IA-2, and insulin are major target antigens of T cell mediated


β cell destruction that results in type 1 diabetes in humans. Proliferative and cytotoxic T cell responses to GAD65 are detected in peripheral blood of newly diagnosed type 1 diabetes patients, but their pathogenicity has not been addressed. Induction of neonatal tolerance to GAD65 specifically prevents diabetes in the NOD mouse model. The role of IA-2 in destructive autoimmunity to the pancreatic β cell in humans is suggested by the high predictive value of IA-2 antibodies for clinical onset of diabetes.

Both GAD65 and IA-2 are neuroendocrine proteins, which are expressed at significant levels in brain and in the β cell. Stiff-man syndrome—a very rare neurologic disorder in humans with a high coincidence of diabetes—is characterized by a strong autoantibody response to GAD65, the titer of which is several orders of magnitude higher than in diabetes. It has been suggested that impairment of γ-aminobutyrate-secreting neurons in stiff-man syndrome is mediated by GAD65 autoantibodies, whereas development of type 1 diabetes is associated with a cellular immune response to GAD65. The low incidence of stiff-man syndrome compared with type 1 diabetes (only one in 104 type 1 diabetic patients develops stiff-man syndrome, whereas 40% of stiff-man syndrome patients develop type 1 diabetes) probably reflects in part the protection of GABAergic neurons by the blood-brain barrier and the absence of MHC class II antigen expression in normal neurons. The cellular localization of IA-2 expression in brain is not known, and there are no known disorders of the central nervous system that involve autoimmunity to IA-2.

In the NOD mouse, the destruction of pancreatic β cells requires both the CD4 T helper (TH) cells and CD8 cytotoxic (TC) cells. Whereas THcells seem to be required for the development of an autoimmune response to the islets and generation of intrainsulitis, TC cells are probably the effector cells of β cell destruction. Furthermore, there is evidence that in the CD4 lineage, the TH1 subset is important for development of disease in the NOD mouse. TH1 cells are induced by IL-12 and are biased toward secreting IFN-γ and IL-2. In contrast, there is evidence that the TH2 cytokine IL-4 exerts a dominant-negative effect on diabetes progression in the NOD mouse. In humans, low autoantibody titers associated with type 1 diabetes and high titers associated with a protective haplotype (DR2) suggest that a strong TH2 response can be inhibitory for β cell destruction. A role for TH1 cells in human disease is also suggested by results of cytokine profiles of peripheral human NK cells in identical twins which are discordant for the development of diabetes. This is different from the observed TH responses in Graves' disease. (See Autoimmune Aspects of Thyroid Disease, above.)

Animal Models of Autoimmune Diabetes Mellitus

The nonobese diabetic (NOD) mouse has been invaluable for studies of molecular mechanisms of autoimmunity directed toward the pancreatic β cells and the development of diabetes. It has several features, however, which distinguish it from the human disease. The incidence of diabetes is two to three times higher in female than in male NODs, whereas in humans there is a slight preponderance of type 1 diabetes in males. Furthermore, while the induction of organ-specific autoimmunity and inflammation in humans may be caused by human pathogens or toxins, autoimmunity seems to be the default mechanism in the NOD mouse. Thus, mice in clean, pathogen-free environments have a high incidence of disease, whereas a variety of regimens that stimulate the immune system of the mouse, such as viral infection or injection of complete Freund's adjuvant, prevent disease. To date, more than 125 treatments for successful prevention or delay of diabetes in the NOD mouse (Table 4-2) have been identified, but none have been identified for humans.

The Bio Breeding (BB) rat develops spontaneous T cell-mediated diabetes. The BB-rat disease is significantly distinct from the human disease in that it is accompanied by autoantibodies to lymphocytes and a severe lymphocytopenia, which is essential for development of β cell autoimmunity and diabetes in this model.

While these models of type 1 diabetes have been invaluable for studies of basic immunologic mechanisms associated with pancreatic β cell autoimmunity, other models of type 1 diabetes closer to what occurs in humans are needed if immunoprevention and immunomodulatory techniques are to be tested.


Autoimmune Adrenal Failure

Autoimmune Addison's disease seldom develops as a single gland syndrome. In about 50% of cases, the disease is associated with other glands and organ failures. Anderson and coworkers described the existence of adrenal-specific autoantibodies for the first time in 1963. Using immunofluorescence techniques based on sections of human, bovine, or monkey adrenals, antibodies specifically recognizing the adrenal cortex (ACA) were described. Steroid-producing cell autoantibodies (SCA) reactive with cells of the adrenals, gonads, and placenta were described by Anderson and coworkers in 1968. SCAs are detected predominantly in ACA-positive


patients with Addison's disease who have premature ovarian failure in the context of autoimmune polyglandular syndrome type I (APS-I; seePolyglandular Syndromes, below). Steroid 21-hydroxylase has been identified as a major adrenal autoantigen in ACA-positive patients with Addison's disease. Using a sensitive assay based on the immune precipitation of radiolabeled recombinant 21-hydroxylase, workers in one study reported positive testing for 72% of sera from patients with isolated Addison's disease, 92% of patients with APS-I, 100% of patients with APS-II, and 80% of patients who were positive for ACA by immunofluorescence but did not have clinically overt Addison's disease (apparently healthy blood donors showed 2.5% positivity) (Table 4-3). Another study measured ACA in 808 children with organ-specific autoimmune diseases without adrenal insufficiency. ACAs were detectable in 14. Ten of these ACA-positive children (also positive for 21-hydroxylase antibodies) and 12 ACA-negative children were prospectively followed with adrenocortical function testing and antibodies. Overt Addison's disease developed in 9 (90%) ACA/21-hydroxylase antibody-positive children within 3–121 months, and the one remaining child had subclinical hypoadrenalism throughout a 24 month-observation period. The progression to adrenal failure was not related to ACA titer, sex, adrenal function, type of associated autoimmune disease, or HLA status. Although ACA 21-hydroxylase antibodies appear to be highly predictive in children, in adults the cumulative risk of developing Addison's in patients with organ-specific autoimmune diseases is about 32%.

Table 4-2. Some reported therapies for prevention of type I diabetes in the NOD mouse.

   1,25-Dihydroxyvitamin D therapy
   Insulin or β chain, orally or parenterally or by nasal insufflation
   Partial pancreatectomy
   Vitamin E

   BCG immunization
   Anti-β 7 integrin
   Anti-CD3, -4, -8, -28
   Anti-complement R
   GAD65 or its peptides, orally or parenterally or by intrathymic injection
   IL-1, -2, -3, -4, -10, -12
   Anti-MHC class I, II

   Monosodium glutamate

   Cold exposure
   Elevated temperature
   Lactate dehydrogenase-elevating virus infection
   Lymphocytic choriomeningitis virus infection
   Mouse hepatitis virus infection
   Mycobacterium infection
   Non-pathogen-free state
   Saline (repeated injections)

Steroid 17α-hydroxylase is another adrenal autoantigen. 17α-Hydroxylase antibodies were found in


5% of patients with isolated Addison's disease, 55% of patients with APS-I, 33% of patients with APS-II, and 20% of sera from patients who were positive for ACA but did not have clinically overt Addison's disease (Table 4-3). Antibodies against another recently described adrenal autoantigen, cytochrome P450 side-chain cleavage enzyme (P450scc), were found to be present in 9% of patients with isolated Addison's disease, 45% of patients with APS-I, 36% of patients with APS-II, and 20% of sera from patients who were positive for ACA but did not have clinically overt Addison's disease (Table 4-3). The prevalence of P450scc antibodies in these groups of patients was always lower than that of 21-hydroxylase antibodies but similar to that of 17α-hydroxylase antibodies. Furthermore, almost all sera that were positive for 17α-hydroxylase or P450scc antibodies were also positive for 21-hydroxylase antibodies. In addition, a comparison of SCAs measured by immunofluorescence with 17α-hydroxylase and P450scc antibody measurements suggested that 17α-hydroxylase and P450scc are the major components of the SCA antigen just as 21-hydroxylase is the major component of ACA antigen.

Table 4-3. Adrenal autoantibodies in different syndromes.


Addison's Disease (%)

APS-I1 (%)

APS-II2 (%)

ACA(+)3 (%)
















1APS-I: Autoimmune polyglandular syndrome type I: Autoimmune polyendocrinopathy, candidiasis, and ectodermal dystrophy.
2APS-II: Autoimmune polyglandular syndrome type II: Adrenal insufficiency, thyroid disease, and diabetes mellitus.
3ACA(+): Adrenal cortex antibody-positive without clinically overt Addison's disease.

Overall, immune responses in autoimmune adrenal disease may involve other antigens, but reactivity to the three described, particularly 21-hydroxylase, appears to predominate. Although inhibition of enzymatic activity by these antibodies has been shown in vitro, no clear relationship to the pathogenesis of the clinical syndrome has yet been established.

Autoimmune Oophoritis & Orchitis

An autoimmune origin for premature ovarian failure with concomitant Addison's disease or oophoritis can be based on the following: (1) the presence of autoantibodies to SCA in most cases, (2) the characterization of shared autoantigens between the adrenals and the ovaries (ie, 17α-hydroxylase and P450scc), (3) the histologic features of the ovaries (lymphocyte and plasma cell infiltrate involving steroid-producing cells), and (4) animal models of the syndrome. There is some evidence of autoimmunity in idiopathic premature ovarian failure not associated with Addison's disease (cellular immune abnormalities, presence of various ovarian antibodies in some patients, and associations with type 1 diabetes and myasthenia gravis); however, the absence of histologic confirmation (lack of oophoritis) makes the autoimmune pathogenesis less credible.

Less is known about the autoimmune pathogenesis of human orchitis. Animal models, however, have shown that infectious or traumatic injury to the testes can induce autoimmune responses in this immune-privileged tissue (conferred by ignorance; see Tolerance, above).

Autoimmune Hypophysitis

Autoimmune hypophysitis (also called lymphocytic hypophysitis) should be considered in the differential diagnosis of pituitary abnormalities in women (8:1 female:male ratio) during the latter half of pregnancy and in the first 6 months postpartum, as well as in patients with coexisting autoimmune disorders, eg, thyroiditis, adrenalitis, autoimmune hypoparathyroidism, or atrophic gastritis. More than 100 cases have been described since the original report in 1962. Antipituitary antibodies have been detected in a minority of patients. Owing to the lack of markers for the disease, the diagnosis can only be confirmed with histologic examination. Nevertheless, because of the usually transient endocrine and compressive features of this condition, conservative management based on clinical suspicion may prevent the consequences of unnecessary pituitary surgery. Granulomatous hypophysitis—another form of autoimmune hypophysitis—appears to have a similar autoimmune pathogenesis but more commonly affects postmenopausal women.

Autoimmune Hypoparathyroidism

Autoimmune hypoparathyroidism, also called idiopathic hypoparathyroidism, is one of the major components of autoimmune polyglandular failure syndrome type I (APS-I; see next section). It also presents as a sporadic disease, sometimes associated with Hashi-moto's thyroiditis in women. The fact that autoimmune hypoparathyroidism presents in association with other autoimmune diseases and the presence of autoantibodies reactive with parathyroid tissue in many affected patients suggest an autoimmune pathogenesis. Parathyroid autoantibodies have been reported to show a complement-dependent cytotoxic effect on cultured bovine parathyroid cells. At least one major parathyroid autoantigen has been identified as the calcium-sensing receptor (Ca-SR). Ca-SR is of great importance in the regulation of parathyroid hormone secretion and renal tubular calcium absorption. This receptor is a member of the seven-membrane spanning domain G protein-coupled receptor family. It is also expressed in thyroid C cells, the pituitary, the hypothalamus, and in other regions of the brain. The relationship of the autoimmune response directed against the receptor to the pathogenesis of the disease is not completely clear. However, stimulation of the Ca-SR with consequent inhibition of PTH synthesis and secretion has been suggested. The prevalence of these antibodies in clinically diagnosed idiopathic hypoparathyroidism was found to be 56% in one study. Measurement of these antibodies may have value in predicting development of autoimmune hypoparathyroidism in patients with autoimmune endocrinopathies who are at risk.




Associations of multiple autoimmune endocrine disorders have been classified in two different syndromes. Autoimmune polyglandular syndrome (APS) type I and type II can be clearly separated clinically (Table 4-4). Some authors have attempted to subdivide APS-II (ie, APS-II and APS-III) on the basis of the association of some autoimmune disorders but not others. However little information is gained by making this subdivision in terms of understanding pathogenesis or prevention of future endocrine failure in patients or their relatives.


APS-I is an inherited autosomal recessive disorder with 25% incidence among siblings of affected individuals. Also known as APECED, or autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy, APS-I is characterized by the triad of chronic mucocutaneous candidiasis, autoimmune hypoparathyroidism, and adrenal insufficiency (only two are required in the index case for the diagnosis, and only one in the siblings). Chronic mucocutaneous candidiasis (involving oral mucosa and nails or less frequently the esophagus) is usually first manifested as the initial problem early in life. In most individuals, the development of autoimmune hypoparathyroidism and Addison's disease follows. However, lifelong surveillance is important since decades may elapse between the development of one feature of the disorder and the onset of another. There is no female preponderance in this syndrome, and it is not HLA-associated. APS-I may occur sporadically or in families. The genetic locus responsible for the disease has been mapped to the long arm of chromosome 21 (21q22.3). The haplotype analysis of this region in different populations has shown that APS-I is linked to different mutations of a gene identified as the autoimmune regulator (AIRE). AIRE encodes a putative nuclear protein with transcription factor motifs (including two zinc finger motifs). It is expressed in different tissues but particularly in the thymus. The mechanism by which mutations of this putative transcription factor lead to the diverse manifestations of APS-I is still unknown. Other immune response-related genes as well as environmental factors probably play a role in development of the syndrome. Several studies of large cohorts of patients from different ethnic backgrounds have reported the appearance of chronic candidiasis at different sites in all patients. Hypoparathyroidism and Ad-dison's disease present with similar high frequency (Table 4-4). The occurrence of the diagnostic triad reportedly presents in 57% of patients. Female hypogonadism, presenting as total or partial failure of pubertal development or as adult premature ovarian failure, has been reported in up to 60% of patients. Male hypogonadism is less frequent (14%). Type 1 diabetes is not as frequent as in APS-II but if present usually develops early (under 21 years of age). Autoimmune hypothyroidism (atrophic thyroiditis) is also less frequent than in APS-II; however, thyroid autoantibodies may be present in many euthyroid patients. Other manifestations are described in Table 4-4. Acute autoimmune hepatitis is reportedly less common than chronic hepatitis, which appears to be present in most individuals. Autoantibodies to aromatic l-amino acid decarboxylase (AADC) are associated with chronic active autoimmune hepatitis and vitiligo which are found in APS-I. These antibodies, if present, can be helpful in making the diagnosis. Autoantibodies against tryptophan hydroxylase have been associated with gastrointestinal dysfunction in APS-I. Autoantibodies to the H+-K+ ATPase and to intrinsic factor are associated with pernicious anemia, and autoantibodies to tyrosinase are associated with vitiligo. Other autoantibodies associated with the single gland disorders that make up this polyglandular syndrome have been discussed above.


APS-II is the most common of the polyglandular failure syndromes. It affects women in a 3:1 ratio to men. APS-II is diagnosed when at least two of the following are present: adrenal insufficiency, autoimmune thyroid disease (thyroiditis with hypothyroidism or Graves' disease with hyperthyroidism) and type 1 diabetes. Historically, Schmidt (1926) first described the association of Addison's disease and thyroiditis. Carpenter and coworkers in 1964 included type 1 diabetes in the syndrome. Other components of APS-II include the following (Table 4-4): primary hypogonadism, myasthenia gravis, celiac disease, pernicious anemia, alopecia, vitiligo, and serositis. The most frequent association appears to be type 1 diabetes (over 50%) and autoimmune thyroid disease (70% in some series). Adrenal insufficiency may be concurrent, may be delayed in onset for up to 2 decades, or may never be manifested. Some diabetic patients (2–3%) develop celiac disease. Gluten-free diet is usually effective. If the celiac disease is untreated, hypocalcemia (not due to hypoparathyroidism), osteopenia, and occasionally gastrointestinal lymphoma may occur.

Although this syndrome and its components aggregate in families, there is no identifiable pattern of inheritance.


Susceptibility is probably determined by multiple gene loci (HLA being the strongest) that interact with environmental factors. Many of the disorders of APS-II are associated (some genetically linked) with the HLA haplotype identified in single disorders. HLA-A1, B8, DR3 and DR4, DQA1*0501, and DQB1*0201 have all been described as associated with APS II.

Table 4-4. Comparison of the different components of autoimmune polyglandular syndromes (APS).


Type I

Type II


Autosomal recessive


Genetic association or linkage

Linked to AIRE

Some HLA association


Equal distribution

Female preponderance

Age at onset


Age 20–40

Endocrine disorders

 Addison's disease




Common (79–96%)

Rare (late onset)

 Autoimmune thyroid disease

Less frequent (about 5%)

More frequent (about 70%)

 Type 1 diabetes

14% (lifetime)


 Primary hypogonadism

60% female, 14% male

About 5%


Not reported



 Chronic mucocutaneous candidiasis

Often at onset (about 100%)

Not reported


Common (about 29%)



About 13%

About 5%

 Dermatitis herpetiformis

Not reported



 Celiac disease

None (only steatorrhea)

Present in 2–3%

 Autoimmune hepatitis

About 12%

Not reported


 Pernicious anemia

About 13%

As common as in APS-I

 Pure red cell hypoplasia


Not reported

 Idiopathic thrombocytopenic purpura

Not reported



 Enamel hypoplasia

All reported

Not reported

 Nail dystrophy



 Tympanic membrane calcification




 Myasthenia gravis

None reported

All reported

 Stiff-man syndrome



 Parkinson's disease






Not reported



Not reported

 Progressive myopathy


Not reported

 IgA deficiency

Not reported



Not reported


 Idiopathic heart block

Not reported


 Goodpasture's syndrome

Not reported



Hormonal replacement therapy remains the only form of treatment of the autoimmune polyglandular syndromes. The clinical management of these disorders mandates early diagnosis of associated components.


Since the age at onset of associated disorders is clinically unpredictable, long-term follow-up is needed. Endocrine disorders are treated as they develop and are diagnosed. Hormonal treatments for the specific gland failures are described elsewhere in this book. However, specific combinations of endocrine organ failure require specific management. For example, thyroxine replacement can precipitate life-threatening adrenal failure in a patient with untreated Addison's disease. Furthermore, hypoglycemia or decreasing insulin requirements in a patient with type 1 diabetes may be the earliest symptom or sign of adrenal insufficiency. Hypocalcemia, seen in APS-II, is more commonly due to celiac disease than to hypoparathyroidism. Treatment of mucocutaneous candidiasis with ketoconazole in patients with APS-type I may induce adrenal insufficiency in a failing gland (this antifungal medication is a global P450 cytochrome inhibitor). These drugs may also elevate liver enzymes, making the diagnosis of autoimmune hepatitis—requiring treatment with immunosuppressants—more difficult in these patients.

Screening of affected individuals as well as their relatives is the only way of preventing morbidity and mortality. Annual measurement of TSH is recommended as cost-effective in first-degree relatives of patients with type 1 diabetes. Autoantibody measurements may help in the preclinical assessment of several disorders. Electrolytes, calcium and phosphorus levels, thyroid and liver functions, blood smears, and vitamin B12 measurements are all recommended in the follow-up of APS-I. For APS-II patients with type 1 diabetes, thyroid disease and celiac disease coexist with sufficient frequency to justify not only TSH measurement but also screening for endomysial antibodies containing transglutaminase antibodies, which are prevalent in celiac disease.


Blanas E et al: Induction of autoimmune diabetes by oral administration of autoantigen. Science 1996;274:1707.

Jacobson DL et al: Epidemiology and estimated population burden of selected autoimmune diseases in the United States. Clin Immunol Immunopathol 1997;84:223.

Janeway CA et al (editors): Immunobiology: The Immune System in Health and Disease, 4th ed. Garland, 1999.

Jaume JC et al: Cellular thyroid peroxidase (TPO), unlike purified TPO and adjuvant, induces antibodies in mice that resemble autoantibodies in human autoimmune thyroid disease. J Clin Endocrinol Metab 1999;84:1651.

Jaume JC et al: Evidence for genetic transmission of thyroid peroxidase autoantibody epitopic “fingerprints.” J Clin Endocrinol Metab 1999;84:1424.

Jaume JC et al: Thyrotropin receptor autoantibodies in serum are present at much lower levels than thyroid peroxidase autoantibodies: analysis by flow cytometry. J Clin Endocrinol Metab 1997;82:500.

Jaume JC, Rapoport B, McLachlan SM: Lack of female bias in a mouse model of autoimmune hyperthyroidism (Graves' disease). Autoimmunity 1999;29:269.

Kita M et al: Regulation and transfer of a murine model of thyrotropin receptor antibody mediated Graves' disease. Endocrinology 1999;140:1392.

Kwok WW et al: HLA-DQB1 codon 57 is critical for peptide binding and recognition. J Exp Med 1996;1831253.

Lanzavecchia A: Immunology. Licence to kill. Nature 1998; 393:413.

Rapoport B et al: The thyrotropin (TSH) receptor: interaction with TSH and autoantibodies. Endocr Rev 1998;19:673.

Roitt IM, Delves PJ: Roitt's Essential Immunology, 10th ed. Blackwell, 2001.

Shimojo N et al: Induction of Graves-like disease in mice by immunization with fibroblasts transfected with the thyrotropin receptor and a class II molecule. Proc Natl Acad Sci USA 1996;93:11074.

Volpe R (editor): Autoimmune Endocrinopathies. Humana, 1999.