The human immune system is comprised of about 2 X 10 ¹²lymphocytes containing approximately equal ratios of T and B cells. B lymphocytes synthesize immunoglobulins that are first expressed on their membranes as clonally distributed antigen-specific receptors and then secreted as antibodies following antigenic stimulation. The ability of the immune system to recognize antigens is remarkable. A human being can produce more than 10 ₇antibodies with different specificities. The concentration of antibodies in human serum is 15 mg/ml, which represents about 3 x 10 ²⁰immunoglobulin molecules per person! Since each B cell has approximately 10 ₅antibody molecules of identical specificity on its surface, the human humoral immune system scans the antigenic universe with about 10 ¹⁷cell bound receptors. To maximize the chances of encountering antigen, lymphocytes recirculate from blood to lymphoid tissues and back to the blood. The 10 ¹⁰lymphocytes in human blood have a mean residence time of approximately 30 minutes, thus an exchange rate of almost 50 times per day.

T lymphocytes develop from precursor stem cells in fetal liver and bone marrow and differentiate into mature cell types during residence in the thymus. Mature T lymphocytes (antigen responding, response control, and response mediating cells) are present in thymus, spleen, lymph nodes, throughout skin and other lymphatic organs, and in the bloodstream. B lymphocytes (immunoglobulin producing cells) develop from precursor cells in fetal liver and bone marrow and are found in all lymphoid organs and in the bloodstream. The ontogeny and functions of these cells have been identified in a variety of ways, including morphologic and functional criteria, and by antibodies identifying surface proteins which correlate to a varying extent with specific functions. Lymphocytes develop through stages leading to pools of cells which can be operationally defined, and be recognized by acquisition of specific antigenic determinants (1) (Fig. 7-1, Table 7-1). Human B and T cells normally express class I (HLA-A, B, C) major histocompatibility complex (MHC) antigens on their surface, and B cells express class II antigens (HLA-DR, DP, DQ). "Activated" T cells also express class II antigens on their surface, and are then described as "DR+" (or sometimes as Ia+).

T cells have on their surface receptors (T cell antigen receptor -- TCR), which recognize an antigen/HLA complex, accessory molecules which recognize HLA determinants, and adhesion molecules which recognize their counterpart ligands on antigen presenting cells (APCs). After activation, T cells also have new receptors for cytokines, the hormone products mainly produced by macrophages, T cells and B cells, which control other T or B cells (2) (Table 7-2). The T cell antigen recognition complex, "receptor", consists of disulfide-linked heterodimers, usually the TCR-a and TCR-b chains, plus five or more associated peptides making up the CD3 complex (3). A small proportion of T cells have TCRg and TCRd chain instead of a and b chains. TCR-a and b peptides and gd peptides are derived from rearranged genes coding for proteins which are unique in each cell clone. The germline TCR genes are very large, containing 40 - 100 different V (variable) segments, D (diversity) segments (in genes), many J (junctional) segments, and one or two C (constant) segments (Fig. 7-2).

During development of each T cell, segments of the germline gene are rearranged so that one TCR gene V segment becomes associated with one D (in the case of TCR-b), one J, and one C segment to produce a unique gene sequence. This random combination of different V, D, and J and C segments, and additional variations in DNA sequence introduced in the J and D region during recombination, provides the enormous diversity of specific TCRs required to recognize the entire universe of T cell antigens. This process is described as "instructive", rather than "educational", in the sense that antigen specific TCRs develop because of intrinsic genetic instructions, rather than in response to exposure to antigens. This process, for example, means that all individuals can (before clonal deletion) have preformed TCRs able to recognize thyroid antigens as well as thousands of other antigens.

The set of V, D, and J segments present in one individual's inherited ("germ line") TCRa, b, g, and d chains differs from those comprising another individual's genes. This variation can be recognized by the process of "Southern blotting", in which DNA is digested by restriction enzymes which cut the DNA at specific infrequently-occurring sequences. However, the technique of reverse transcription-polymerase chain reaction (RT-PCR) is now much more widely used to analyse the V gene repertoire of T cells either ex vivo or after in vitro expansion (4), in the absence of suitable monoclonal antibodies to recognize the separate receptor families.

Each individual T cell, and its progeny, have a "rearranged" gene with one unique combination of V, D, and J segments. Current technology makes it possible to clone individual T cells which respond to a specific antigen, to expand the progeny of a single cell many fold, and then to determine the DNA sequence of the V, D, and J regions which provide the unique recognition function of the TCR. Based on such studies, it is now evident that specific V segments are preferentially used in the response to certain antigens (4). We may thus infer that the availability of such a V segment in an individual's TCR repertoire must favor an immune response to a specific antigen, including an autoantigen.

Each TCR recognizes one specific antigenic peptide sequence (5), which may consist of 8 - 9 amino acids for class I restricted T cells, and 13 - 17 amino acids for class II restricted T cells. However, some T cells respond to various portions (epitopes) of one antigen; these may represent overlapping peptide segments of the epitope. Thus the response of each individual T (and B) cell is extremely specific, but the combined effect of many T (and B) cells acting together is observed in the typical final "polyclonal" response.

T cells recognize antigen in association with ("presented by") an MHC-molecule; CD4 ⁺T cells (often functioning as helper cells) recognize class II molecules + antigen, and CD8 ⁺T cells (often functioning as cytotoxic cells) recognize class I molecules plus antigen. The antigen (a small peptide, v.i.) fits within a cleft in the HLA-DR molecule (Fig. 7-3). The TCR functions to recognize the antigenic peptide on the MHC molecule (Fig. 7-3). The five associated peptides of the CD3 complex are believed to be signal-transducers and to initiate intracellular events following antigen recognition. The normal response proceeds via TCR antigen recognition, then "activation" of the T cell through the combined effect of antigen recognition and costimulatory signals, including interleukin (IL)-1 action, leading to T cell IL-2 secretion and IL-2 receptor expression, followed by proliferation of the T cell into an active clone.

Lymphocyte development is controlled by cytokines released by macrophages, lymphocytes, and other cells. Both T and B cells release a large array of cytokines which carry out their effector functions and alter the function of other cells (Fig. 7-1, Table 7-2). As lymphocytes mature in the thymus, and become activated on exposure to antigen, the types of cytokines to which they respond -- and produce -- become altered. In animals, and to a lesser extent in man, types of lymphocytes can be operationally defined by the cytokines produced. For example, "Th1" T cells produce IL-2, IFN-g and TNF and are predominant in "delayed hypersensitivity" type reactions, whereas "Th2" T cells produce IL-4 and IL-5, stimulate B cells, and are involved especially in antibody-mediated reactions. Cytokines produced by Th1 cells enhance the activity of this subset but inhibit Th2 cells, and vice versa. This type of regulation may be critical in determining an immune response and in suppressor phenomena.

As well as cytokines and their receptors, T cells express a number of receptors for chemokines, integrins and selectins which are involved in the sequential stages of cell adhesion which leads to T cell homing to tissues (7). A word of caution is necessary however in terms of translating these findings into the human situation where boundaries between the subsets are less clear. It is also increasingly recognized that the simple dichotomy of T cells into two types is over-simple (6a), with cytokines such as IL-12 being assigned to the Th1 subset although not being secreted by T cells, and production of this cytokine is stimulated by the Th2 cytokines IL-4 and IL-13, which will drive the immune response from Th2 towards Th1. The blurring of pattern that is seen in many autoimmune diseases challenges the dogma of an easy divide in the type of immune response.

Each B cell produces a unique immunoglobulin (Ig) programmed by an Ig gene which has also been "rearranged" from the germline V, D, J, and C segments (as for the TCR) (8). The TCR and Ig genes are, not surprisingly, members of one gene superfamily. Further diversity is provided by "antigen-driven" somatic mutations which occur during amplification of the progeny of a stimulated B cell, causing the production of a family of Igs with slightly different sequences. Some of these Igs may be better antibodies, and others might recognize other antigens including “self". B cells secrete their unique Igs into surrounding fluids, and some normally remain on the B cell surface, where they can bind the antigen which is recognized by (fits the structure of) the specific Ig (Fig. 7-4). The surface Ig is therefore a B cell receptor for antigen, having a specific face or "idiotype" which fits the conformation of the antigen molecule “epitope". The recognition process involves the shape of the epitope - i.e. it is "conformational" and for B cells probably normally involves unprocessed or "native" antigen. This implies that B cell and T cell epitopes for the same antigen are usually different segments or forms of the molecule.

Antigen is normally recognized by the TCR complex only when presented in association with a class I or class II MHC molecule. Typically CD8 ⁺T cells recognize antigen with class I (HLA-A, B, C) proteins, and CD4 ⁺T cells are "restricted" in their recognition to antigen presented with class II (HLA-DR, DP, DQ) proteins. Antigens which originate within the cell are preferentially presented by class I molecules to CD8 ⁺cells. This indicates an orientation of CD8 ⁺cells toward destroying cells invaded by viruses or producing abnormal antigens. Class II molecules are directed toward presentation of external or alloantigens to CD4 ⁺helper cells.

The genes for the HLA-A, B, C and HLA-DR, DP, DQ molecules are on chromosome 6, and comprise some of the genes in a large immune response control complex (Fig. 7-5). Each cell surface HLA molecule is made up of 2 peptide chains; an a and b2 microglobulin for class I molecules, and a and b chains for class II. Each individual inherits from each parent one HLA-A, B, and C, one DRa and 3 DRb genes, a pair each of DP and DQa and b genes, and other related genes which are not expressed, including DX and DO. b2 microglobulin polypeptides are the same for all individuals (Fig. 7-5). The genes are expressed in a co-dominant manner, and (in contrast to TCR and Ig molecules) are invariant in individuals. However, the genes are all highly "polymorphic", that is, many alleles may exist for each gene. The actual evolutionary drive for this diversity is unknown. We may note that different HLA molecules will presumably present different epitopes and thus lead to selection of a unique T cell repertoire for each HLA allele. While TCR gene rearrangement provides the T cell repertoire to respond to individual antigens, HLA diversity guarantees that different individuals will have different T cell repertoires. We may theorize that this is beneficial for survival of some members of the species when all are attacked by a pathogen - for example, the plague.

The HLA molecules play a central role in T cell clonal selection during fetal development, in normal immune responses, and in presentation of "self-antigens". In many instances -- including autoimmune thyroid disease (AITD) as detailed below -- inheritance of a specific HLA gene correlates with increased susceptibility to a specific disease. In some cases this can now be related to a gene coding for a specific amino acid in the HLA molecule which is believed to control epitope selection (often called determinant selection) and thus to be associated with disease susceptibility.

Antigen can be presented to CD4 ⁺T cells by "conventional" (or "professional") APCs, particularly dendritic cells (9), and also by B cells and activated T cells, and less effectively by a variety of other cells (fibroblasts, glial cells, thyrocytes), when these normally HLA-DR-negative cells are altered and express HLA class II molecules on their surface. This is because non-classical APCs cannot provide the necessary costimulatory signals, including the B7-1 (CD80) and B7-2 (CD86) molecules, which bind to CD28 on the T cell and are necessary for activation of certain T cells. Recent evidence in experimental type 1 diabetes mellitus suggests a simple two-stage model for these molecules in the development of autoimmunity against the islet, with B7-2 playing a primary role in T cell priming, while, if B7-1 has a role, it is dependent on expression in the local microenvironment and is most important in the efector phase of the autommune response (10).

If B7 molecules bind instead to CD152 (CTLA-4) on the T cell, the immune response is terminated. The individual roles of CD80 and CD86 are not clearly established, although some functions appear to be distinct (eg CD80 appears to stimulate CD152) and some overlapping (eg both stimulate CD28), and the tempo of their involvement at different times of the immune response is likely to be critical to the type of response produced (11). The maturation state of the dendritic cell is another determinant of immune homeostasis. Semi-maturation, induced by proinflammatory cytokines like TNF-a, allows the development of a tolerogenic stage for these cells. Full maturation, induced by signalling through toll-like receptors, complement receptors or antibody Fc receptors, induces proinflammatory cytokine production by the dendritic cell and allows them to generate T cell immunity (12).

Extracellular (usually foreign protein) antigen is endocytosed by macrophages and dendritic cells, facilitated by a variety of antigen uptake receptors (9). B cells collect the antigen by binding it to cognate surface Ig, and internalizing the Ig-antigen complex. Inside these cells the antigen molecule is broken down to peptides which are 13 - 17 amino acids long. Many of these peptides are destroyed, but some are retained to reappear on the cell surface as T cell antigen "epitopes". These peptides, possibly in a Golgi-like organelle, become associated with HLA-DR, DP, or DQ molecules, are transported to the cell surface, and there can be "presented" to a T cell (as in Figure 7-3). The initial response to antigen proceeds via the antigen-APC-T cell route. In autoimmunity this route may be favored by the particular differentiation pathway of dendritic cells, initiated by certain cytokines, or because of intrinsic defects in genes controlling dendritic cell function (14). Secondary T cell responses may follow the same route, or may more frequently follow the antigen-B cell-T cell route. This is because the surface Ig of the B cell allows it to collect specific antigen circulating at a very low level and concentrate it for presentation to a T cell. In addition it is increasingly recognized that 50-90% of proteins which might form autoantigens are post-translationally modified, and some of these modifications may be important in creating new self antigens. These modified proteins have unknown effects on positive and negative selection of T cells, and of course complicate the analysis of epitopes based on native protein sequences that do not encompass the modified proteins.

Proteins produced within the cell (including normal proteins or products encoded by invading viruses) are processed by a separate pathway and appear on the cell surface as peptides of 8 - 9 amino acids associated with HLA class I molecules.

Antigen presentation to T cells leads to a variety of responses which include proliferative or suppressive functions, development of cell cytotoxic responses, control of Ig secretion, and many more. In addition, under specific circumstances, antigen presentation may cause the T cell to become non-responsive or "anergized" (10, 15).

Presumably the APC, with its surface covered by HLA-DR-antigen complexes, is met by T cells having the correct receptor (idiotype) matching the processed antigen exposed on the surface of the HLA molecule. T cell and APC adhere by segments of HLA-DR and CD4 molecules which are, effectively, "sticky" (see Figure 7-3), fostering close contact of APC and T cells. Co-stimulation is provided by other "adhesion molecules" on the APCs which also pair with their counterparts on T cells. Specifically, LFA-1 on T cells binds to ICAM-1 on the APC, and CD2 binds to ICAM-2 and these increase intercellular adhesion as a preliminary step. Binding of B7 to T cell receptor CD28 or CTLA-4 is especially important in costimulation but other costimulators exist (16). A recent development has been the recognition that the B7/CD28 pathway is more complex than previously realized, and includes a molecule, inducible costimulator (ICOS), that binds to a B7-like protein, B7RP-1; in mice both ICOS and C28 regulate T cell expansion and deliver complementary signals necessary for optimal T cell activation (17). These reactions dramatically increase the bond between APC and T cell, ensure close approximation, and provide the additional, or "second signal" needed to activate the T cell.

Presentation of antigen and the accompanying second signal are required to activate a naive T cell and initiate an immune response; previously activated T cells are much less dependent on B7-mediated costimulation. Antigen recognition, and APC-produced IL-1 (Table 7-2), cause T cell stimulation. This activates the T cell to express IL-2 receptors and to secrete IL-2 itself. Increased T-cell secreted IL-2 induces the responding T cell, and nearby ("bystander") T cells to proliferate. T-cell secreted IL-2, IL-6 and other cytokines and IL-4 cause B cells to be stimulated and proliferate and cell surface receptors such as CD40 on B cells and its ligand on T cells are also involved in B cell activation (18). B cells themselves secrete distinct profiles of cytokines, in response to the engagement on CD40, and these cytokines can upregulate or downgregulate an immune response in a manner which depends on whether the B cell is simultaneously stimulated by antigen (19). Intimate T-cell to B-cell contact may account for "antigen-specific help" for T cell and B cell responses, whereas the effect of T cell-secreted lymphokines on "bystander" T or B cells may account for stimulation of "non-antigen-specific" responses by these lymphocytes (20). T cell - B cell interaction usually occurs as T cells percolate through lymph nodes. The T cell receptors must in some way "scan" the exposed DR-antigen combinations on B cells, until by chance the T cell finds a B cell presenting the antigen epitope recognized by the TCR. The beneficial effects of rituximab, a CD20 specific, B depleting monoclonal antibody, in autoimmune conditions including Graves’ disease (20a) is related to its effects on inhibiting this interaction between T and B cells.

Responding lymphocytes can be segregated into groups based on whether they are naive or memory cells, CD4 ⁺"helper" cells, or CD8 ⁺"cytotoxic/suppressor" cells. CD8 ⁺T cell activation requires additional costimulatory pathways to the B7-dependent pathway largely used by CD4 ⁺T cells, and ICAM-1 is particularly important (21, 22). Although not providing as a clear separation as in the murine system, a functional separation based on lymphokine secretion seems to provide an important categorization for CD4 ⁺cells. Th1 cells function as "inflammatory" cells, typical of a delayed hypersensitivity type reaction, while Th2 cells are more specifically helper cells for B cell immunoglobulin synthesis. A number of factors including TCR affinity and ligand density, and non-T cell-derived cytokines such as IL-4 and IL-12, determine whether the outcome of an immune response is predominantly by Th1 or Th2 cells (23). Some CD4 ⁺and some CD8 ⁺cells appear to provide suppressor signals. The nature of both the recognition process in the suppressor function and the putative effector molecules is not clear. It may involve TCR recognition of T cell or B cell idiotype, secreted molecules representing part of the TCR receptor, or lymphokines but none of these has conclusive support (24). Most recent attention has focused on alternates of local secretion of TGF-β and direct cell contact which involves the binding of CTLA-4 on the T regulatory T cell (25).

A third population of T helper cells has been defined recently, based on their secretion of the pleiotropic proinflammatory cytokine IL-17, and are so called Th17 cells. The differentiation and expansion of these cells depends on the coordinate effects of IL-6, transforming growth factor beta (TGFb) and IL-23 (25a). These Th17 cells are responsible for defense against certain micro-organisms such as Klebsiella, Borrelia and fungi. Of relevance to this discussion, they also have important roles in tissue inflammation and organ-specific autoimmunity.

Although the concept of suppressor cells fell into disrepute during the late 1980s, there has been resurgence in interest with the recognition that CD4 ⁺cells expressing high levels of the IL-2 receptor, CD25, act in a way entirely in keeping with the previously defined suppressor population. These CD4 ⁺, CD25 ⁺T cells have been termined regulatory or Treg cells. Such cells can prevent autoimmunity when transferred from healthy, naïve animals and their depletion results in autoimmune disease. Such cells express Foxp3 which encodes a critical transcription factor for their function: mutation of this gene in man results in the lethal immunological disorder IPEX syndrome that includes autoimmune hypothyroidism anmongst its manifestations (26). The exact mechanism by which suppression is induced is unclear but IL-2 appears to be critical to maintaining peripheral tolerance by supporting the function of these cells, whcih in turn explains some of the apparently contradictory effects of this classically agonistic cytokine (27). Similar cells have been found in man and it is likely that the next few years will see a renewed attempt to identify defects in this population in thyroid autoimmunity, especially as the fundamental questions of antigen specificity and recognition, and mechanisms of action, become clear (28).

Recently it has been established that APCs have a central role in controlling Treg cells, with resting APCs (including thymic epithelial cells) promoting their development through the induction of the transcription factor Foxp3 (29). Activation of APCs, for instance through their T cell-like receptors, has the opposite effect, and at least one component responsible for the suppression of Tregs then is the cytokine IL-6; this pathway allows effector T cells to predominate over Tregs, thereby shifting the dynamic equilibrium in favor of an immune (or autoimmune) response. Another critical molecule in the Treg cell pathway is the costimulatory signal receptor, CD28, which is required for both development and maintenance of Treg function. CD4 ⁺, CD25 ⁺Treg cells may be particularly suited to respond to their selecting self-peptide, which may direct their accumulation to where the self-peptide is expressed, in turn allowing them to be effective in maintaining non-self-responsiveness (30).

TGFb is yet another critical factor and recent work suggests that exposure to this cytokine induces Tregs, but when combined with IL-6, Th17 effector cells are generated (25a). The absence or presence of IL-6 is thus critical to determining whether there is a regulatory milieu or a proinflammtory response mounted by Th17 cells. It seems that both Th1 and Th17 cells are potent inducers of organ-specific autoimmunity, but their relative roles in each type of disease remain to be clarified.

It is increasingly clear that Treg are more complex a group of cells than originally clear. Currently T regulatory cells can be classified as those which arise within the thymus and express Foxp3, and a Th3-like population which probably does not express this molecule and which develops in the periphery. The glucocorticoid inducible tumor necrosis factor receptor (GITR) is expressed by both populations but CD25 –bright expression is not a requirement for regulatory T cell function. In addition, there is a CD4 ⁺, CD69+ population of regulatory T cells and a CD8 ⁺suppressor population which recognizes self-peptides expressed by an MHC class Ib (Qa1 in mice) molecule (31). The latter seem particularly critical in the remissions of multiple sclerosis that characterize many cases of this disease. Since T cells stimulate or suppress other T or B cells, they may develop feedback circuits to limit responses to antigens.

The reciprocal relationship between Th1 and Th2 cells, exerted through secretion of cytokines, serves as another model of suppressor function and this could be an important future immunotherapeutic area. This paradigm is conceptually useful but is almost certainly too simplistic, not least because there may exist within the Th2 compartment different types of T cells, some with pathological effector function and others which act as physiological regulators of Th1 responses. Endeavors to manipulate the entire Th2 population, to deviate an immune response away from Th1 cells, may therefore lead to exacerbation of the immune response, and may explain the reciprocal relation between the prevalence of infectious disease and autoimmunity (32).

It must be noted that the complexity of T cell function is much greater than suggested by this brief outline. For example, CD4 ⁺T cells can function as suppressor or cytotoxic cells and may do so at different times in their life cycle. In fact an important generalization is that the T cell surface molecules so far defined only rarely identify unique functional sets of cells. On the other hand, functions of T cells to help T or B responses, to suppress such responses, and to kill target cells under certain conditions, can be operationally defined, and these functions often correlate with expression of specific T cell antigens. While recognizing the inherent simplifications we have employed, it remains useful to discuss T cell subsets in relation to surface antigens and function, since so much work has been published using these concepts.

In addition to the standard T cell function described above, other cells participate in immune responses. Macrophages may destroy cells having immune complexes on their surface through recognition of the Fc portion of bound Ig. Other cells which do not bear the CD3 marker of T cell lineage exist (K and NK cells) and have the ability to spontaneously kill other cells (especially those expressing HLA antigens). NK cells can be detected by specific monoclonal antibodies such as anti-CD16, and are recognized phenotypically as large granular lymphocytes. Like T cells, NK cells can have a type 1 or 2 pattern of cytokines release (33). Macrophages, T, K, NK, or other cells also kill cells coated with immune complexes in the process of antibody-dependent-cell-cytotoxicity (ADCC) (Fig. 7-6).

The immune system, which evolved to defend us from invading foreign proteins, normally "tolerates" (does not develop recognizable responses to) self-antigens. The level of this control is variable. For example, self-reactivity to serum albumin is not seen. However, antibodies to thyroid antigens exist in up to 36% of adult women, and their presence must be considered effectively normal. The development of tolerance is closely associated with the restriction of TCRs to recognizing an antigen only when presented by an HLA molecule. The process, which for T cells occurs in the fetal thymus, leads to elimination of some T cells, and retention of others with TCRs having desirable features. Self-antigens are believed to be presented on HLA molecules to T cells developing in the thymus. This implies that antigen must be in the thymus or in the circulation for tolerance to develop. T cells bearing TCRs which react strongly to HLA molecules not bearing antigen ("autoreactive cells") are largely inactivated or destroyed. T cells which have the capacity to react with foreign antigens presented by self MHC molecules are somehow retained. Most T cells with TCRs which bind strongly to class I or class II molecules bearing self-antigen are also "clonally deleted" (34) (Fig. 7-7). Presumably some T cells which react with MHC molecules plus self antigen are not deleted, since otherwise an excessive number of T cells would be lost, leaving a "hole" in the available TCR repertoire which would compromise its ability to mount a future immune defense.

The best evidence that thymic deletion prevents autoimmunity in man comes from auoimmune polyglandular syndrome (APS) type 1, which is the result of a muatation in the AIRE (AutoImmune REgulator) gene. Such patients have multiple autoimmune disorders, principally Addison’s disease and hypoparathyroidism but including thyroid autoimmunity. The AIRE protein is expressed in the thyrmus by medullary epithelial cells and regulates the rather surprising expression of an array of self proteins (normally confined to extrathymic tissues) by these cells during fetal development. When such self antigens cannot be expressed to allow clonal deletion, autoimmunity ensues and this accounts for the multiple autoimmunity and early onset found in this syndrome (reviewed in 34). It is also now clear that additional factors must be involved in determinig the target –organ specificity of the consequences of AIRE deficiency, and these may include other proteins that regulate the expression of autoantigens for utoimmune disease that are rare or absent in APS type 1 (35).

This fine discrimination between perfection in clonal deletion, and repertoire maintenance, allows a limited number of autoreactive and self-antigen-reactive T cells to survive, and thus sets the stage for autoimmune disease. It is unclear how this mechanism produces tolerance to antigens not present in the fetal thymus or circulation - especially antigens which may be expressed only in the mature individual. It is presumed that a mechanism working outside the thymus -- "peripheral tolerance” -- completes T cell selection. Not all suppressor cells are of the CD8 ⁺lineage. In rodents an important regulatory T cell population is CD4 ⁺CD25 ⁺which mediate their effects by both cell-cell contact and cytokine secretion, as discussed in Section4.

Factors which control the production of Treg include the way antigen is presented. CD28 and CTLA-4 are both expressed by these cells and can influence their function, while stimulation of T reg via B7-2 inhibits their suppressive function, whereas B7-1 enhances this function. Since the maturation state of dendritic cells influences the expression of B7-1 and B7-2, this helps to explain how the activation state of the antigen presenting cell may influence the immune response. IL-2 is critical for the development and expansion of Treg (36) and may represent an important point at which this immune system could be regulated therapeutically, as any way to enhance Treg function is clearly likely to be of benefit in autoimmune disorders. Perhaps the clearest evidence so far that these cells play an important role in maintaining self-tolerance is the demonstration that their removal in vitro allows the derivation of CD4 ⁺cells from peripheral blood that respond to islet cell, melanocyte and testis autoantigens, while adding CD4 ⁺, CD25 ⁺T cells back suppresses the expansion of these autoractive helper cells. (37).

During maturation in the thymus, probably 95% or more of the lymphocytes produced are negatively selected, and die through a process described as "programmed cell death" -- or apoptosis. This process involves several genes including those required for apoptosis, such as Fas. A similar process is thought to ensue whenever a T cell is stimulated by its cognate antigen but does not receive a "second signal", and during induction of anergy by other mechanisms. Defects in Fas lead to preservation of autoreactive T cells in some models of animal autoimmune disease (38).

B cells undergo a similar selection process in fetal bone marrow or liver, except for the participation of MHC molecules. If exposed to antigen during this early stage of development, B cells are somehow permanently inactivated. As for T cells, the selection process is not perfect, and leaves some B cells having the ability to make antibodies directed to self-antigens in the adult. However, B cells require T cell "help" in order to proliferate and differentiate into mature Ig secreting cells. In the effective absence of self-reactive T cells to "help" such B cells, these B cells remain dormant, and expanding clones do not develop. Although such clonal ignorance may be an important pathway in preventing B cell autoreactivity, it is not the only mechanism, and physiological concentrations of autoantigen may induce anergy of B cells, even when their affinity for autoantigen is low.

Tolerance to self-antigen can be overcome ("broken") in animals by injecting the antigen in an unusual site on the body, especially in the presence of adjuvant compounds such as tubercule bacillus fragments and oil, or alum, or by slightly altering the antigen structure, or by altering the responding immune system (for example, by whole body irradiation, or depletion of suppressor T cells). It is clear that some degree of recognition of self-antigens is normal, but is also normally well controlled by suppressor circuits in the immune system. Based on current knowledge, a number of therapeutic strategies already exist through which memory T and B cells can be targeted, with increasingly impressive results in autoimmune disease (39). Factors which may lead to amplification of the normal small number of self reactive T cells, and thus produce "autoimmune" disease, are described subsequently.