Title: IMMUNOLOGICAL UNRESPONSIVENESS "TOLERANCE" Identify the most important type of tolerance. Realize the importance of the tolerance mechanism.

Transcription

1 LECTURE: 05 Title: IMMUNOLOGICAL UNRESPONSIVENESS "TOLERANCE" LEARNING OBJECTIVES: The student should be able to: Define the term "immunological tolerance". Identify the most important type of tolerance. Realize the importance of the tolerance mechanism. Explain how the state of unresponsiveness is generated? Indicate how the state of tolerance is important in clinical medicine? Explain the autotolerance. List the four possible ways in which self reactive lymphocytes may be prevented from responding to self antigens such as: - Clonal deletion. - Clonal abortion. - Clonal anergy. - Suppression. Explain the central, and post thymic tolerance to selfantigens. Explain state of B-cell tolerance to self antigens. Discuss the acquired (immune) tolerance. List the characteristics of the acquired tolerance. List the factors (T and B cells) that influence the inductions of tolerance. Enumerate the methods of immune tolerance induction. Explain the light and dark zone.

2 LECTURE REFRENCE: 1. TEXTBOOK: ROITT, BROSTOFF, MALE IMMUNOLOGY. 6 th edition. Chapter 12. pp TEXTBOOK: ABUL K. ABBAS. ANDREW H. LICHTMAN. CELLULAR AND MOLECULAR IMMUNOLOGY. 5 TH EDITION. Chapter 10. pg HANDOUT. Immunological tolerance Tolerance mechanisms are needed because the immune system randomly generates a vast diversity of antigen-specific receptors and some of these will be self reactive; tolerance prevents harmful reactivity against the body's own features. Central thymic tolerance to self antigens (autoantigens) results from deletion of differentiating T cell that express antigen-specific receptors with high binding affinity for intrathymic self antigens. Low-affinity self-reactive T cells, and T cells with receptors specific for antigens that are not representative intrathymically, mature and join the peripheral T cell pool. Post-thymic tolerance to self antigens has five main mechanisms. Self-reactive T cells in the circulation may ignore self antigens, for example when the antigens are in tissues sequestered from the circulation. Their response to a self antigen may be suppressed if the antigen is present in a privileged site. Self-reactive cells may under certain conditions be deleted or rendered anergic and unable to respond. Finally a state of tolerance to self antigens can also be maintained by immune regulation. B-cell deletion takes place in both bone marrow and peripheral lymphoid organs. Differentiating B cells that express surface immunoglobulin receptors with high binding affinity for self-membrane-bound antigens will be deleted soon after their generation in the bone marrow. A high proportion of short-lived, low-avidity, autoreactive B cells appear in peripheral lymphoid organs. These cells may be recruited to fight against infection. Tolerance can be induced artificially by various regimens that may eventually be exploited clinically to prevent rejection of foreign transplants and to manipulate autoimmune and allergic diseases.

3 INTRODCTION Immunological tolerance is a state of unresponsiveness that is specific for a particular antigen; it is induced by prior exposure to that antigen. Active tolerance mechanisms are required to prevent inflammatory responses to the many innocuous air-borne and food antigens that are encountered at mucosal surfaces in the lung and gut. The most important aspect of tolerance, however, is self tolerance, which prevents the body from mounting an immune attack against its own tissues. There is potential for such attack because the immune system randomly generates a vast diversity of antigen-specific receptors therefore must be eliminated, either functionally or physically. Self reactivity is prevented by processes that occur during development, rather than being genetically preprogrammed. Thus while homozygous animals of histo-incompatible strains A and B reject each other's skin, and their F 1 hybrid offspring (which express the antigens of both the A and B parents) reject neither A skin nor B skin, the ability to reject such skin reappears in homozygotes of the F 2 progeny. Thus it is clear that self-non-self discrimination is learned during development: immunological 'self' must encompass all epitopes (antigenic determinants) encoded by the individual's DNA, all other epitopes being considered as non-self. However it is not the structure of a molecule per se that determines whether it will be distinguished as self or non-self. Factors other than the structural characteristics of an epitope are also important. Among these are: The stage of differentiation when lymphocytes first confront their epitopes. The site of the encounter. The nature of the cells presenting epitopes. The nature of lymphocytes responding to the epitopes. Historical background Soon after the existence of antibody specificity was established, it was realized that there must be some mechanism to prevent autoantibody formation. As early as the turn of the century, Ehrlich coined the term 'horror autotoxicus', implying the need for a 'regulating contrivance' to stop the production of antoantibodies. In 1938, Traub induced specific tolerance by inoculating mice in utero with lymphocytic choriomeningitis virus, producing an infection that was maintained throughout life. Unlike normal mice, these inoculated mice did not produce neutralizing antibodies when challenged with the virus in adult life. In 1945, Owen reported an 'experiment of nature' in non-identical cattle twins which showed that cells carrying self and non-self antigens could develop within a single host. These animals exchanged haemopoietic (stem) cells via their shared placental blood vessels and each animal carried the erythrocyte markers of both calves. They exhibited life-long tolerance to the otherwise foreign cells, in being unable to mount

4 antibody responses to the relevant erythrocyte antigens. Following this observation, Brnet and Fenner postulated that the age of the animal at the time of first encounter was the critical factor in determining responsiveness, and hence recognition, of non-self antigens. This hypothesis seemed logical, as the immune system is usually confronted with most self components before birth and only later with non-self antigens. Experimental support came in 1953, when Medawar and his colleagues induced immunological tolerance to skin allografts (grafts that are genetically non-identical, but are form the same species) in mice by neonatal injection of allogeneic cells (Figure-1). This phenomenon waqs easily accommodated in Burnet's clonal selection theory (1957), which states that a particular immunocyte (a particular B to T cell) is selected by antigen and then divides to give rise to a clone of daughter cells, all with the same specificity. According to this theory, antigens encountered after birth active specific clones of lymphocytes, whereas when antigens are encountered before birth the result is the deletion of the clones specific for them, which Burnet termed 'forbidden clones'. Implicit in the theory is the need for the entire immune repertoire to be generated before birth, but in fact lymphocytes differentiation continues long after birth. The key factor in determining responsiveness is thus not the development stage of the individual, but rather the state of maturity of the lymphocyte at the time it encounters antigen. This was suggested by Lederberg in 1959, in his modification of the clonal selection theory: immature lymphocytes contacting antigen would be subject to 'clonal abortion', whereas mature cells would be activated. It is now established that the neonate is in fact immunocompetent. The reason that one can induce tolerance to certain antigens in the neonate is simply that the type of immune response to antigen can be functionally different in the neonate compared with that in the adult. Past description of neonatal tolerance may therefore have been early examples of this type of 'immune deviation' (see below). Key discoveries in the 1960s established the immunological competence of the lymphocyte, the crucial role of the thyms in the development of the immune system, and the existence of two interacting subsets of lymphocytes: T and B cells. This set the scene for a though investigation of the cellular mechanism involved in tolerance. EXPERIMENTAL INDUCTION OF TOLERANCE Transgenic technology has allowed the study of tolerance to authentic self antigens Until recently, only artificially induced tolerance was amenable to experimental study: antigens or foreign cells were inoculated into an animal and the fate of responding T or B cells was investigated under a variety of circumstances. It was not clear, however, to what extent these experimental models resembled natural self tolerance.

5 Transgenic methods have now made possible the direct investigation of self tolerance. These methods allow one to introduce a specific gene into mice of defined genetic background and to analysis its effects upon the development of the immune system. Furthermore, if the introduced gene is linked to a tissue-specific promoter, its expression can be confined to specific cell types. The protein product encoded by a 'transgene' is treated by the immune system essentially as an authentic self antigen (autoantigen), and its effects can be studies in vivo without the trauma and inflammation associated with grafting foreign cells or tissues. In addition, the parent strain and the transgenic strain are ideal for control experiments and lymphocyte transfer studies because they are congenic that is they differ at only one locus. One can even create transgenic mice in which all of either their B or T lymphocytes express a single antigen receptor. By so increasing the frequency of antigen-specific precursor cells, one can readily dissect tolerance mechanisms. Finally, the use of targeted mutagenesis has allowed immunologists to 'knock out' specific genes in order to study the role of their gene products in the process of immunological tolerance. There are five possible ways in which self-reactive lymphocytes may be prevented from responding to self antigens: 1. Self-reactive T cells in the circulation may ignore self antigens, for example when the antigens are in tissues sequestered from the circulation. 2. Their response to a self antigen may be suppressed if the antigen is present in a privileged site. 3. Self-reactive cells may be deleted at certain stages of development, or 4. Self-reactive cells may be rendered anergic and unable to respond. 5. Finally, a state of tolerance to self-antigens can also be maintained by immune regulation. Which of these fates awaits the self-reactive lymphocyte depends on numerous factors, including: (i) the stage of maturity of the cell being silenced; (ii) the nature of this antigen; (iv) its concentration; (v) its tissue distribution; and (vi) its pattern of expression. CENTRAL THYMIC TOLERANCE TO SELF ANTGENS The process of generating new T cell receptors involves gene rearrangement in addition to N-region modifications. This allows the immune system to generate a vast array of T cell receptors. Such a broad repertoire is clearly necessary to provide protection against the multitude of different infections agents that any individual in the species is likely to encounter. T lymphocytes are not, however, simply effectors cells of the immune system. They also function as regulators of the system through provision of help for some and suppression of other responses. For effective control, lymphocytes must interact with other cells of the immune system and this is one reason why MHC restriction of T cell

6 recognition has evolved. Central tolerance among T lymphocytes evolves around a schooling process in which key cells are educated so that they become dependent on self MHC for survival, while at the same time potentially rebellious lymphocytes are identified and eliminated. This process of central tolerance among T lymphocytes takes place during their development within the thymus and depends on a number of check points through which cell have to pass in order to develop further. T-cell development involves positive and negative selection and lineage commitment T lymphocytes develop from precursors in the bone marrow and are derived from a common lymphoid progenitor cell that give rise to B cells, natural killer (NK) cells and both the αβ and γδ subsets of T lymphocytes (Figure-2). Here we will concentrate on the αβ population of T lymphocytes whose function in the maintenance of self tolerance is now well understood. When immature T cells enter the thymus they express neither CD4 nor CD8 co-receptor molecules (fig. 12.3). These so-called double negative (DN) cells constitute approximately 3% of total thymocytes. At this stage the T-cell receptor (TCR) β chain genes start their recombination. This involves sequential rearrangement of variable (V), diversity (D) and junctional (J) region genes from the multiple copies available in the genome. First diversity and junctional genes rearrange and this is followed by rearrangement of the DJ gene with a variable region gene (Figure-3). The VDJ then combines with a constant region by alternative splicing of RNA to give the complete β chain gene. At this point the α chain genes remain in their genomic configuration but the transcribed and translated β chain nevertheless appears at the cell surface. This is only possible because the β chain can pair with a 'surrogate' α chain and other component of the CD3 signaling complex, in order to migrate from the endo-plasmic reticulum to the cell surface. Surface expression of this complex allows double negative cells to switch off their RAG genes, begin to proliferate and mature into CD4 and CD8 double positive (DP) cells (Figure-3). There is no evidence that this 'checkpoint' (the β selection checkpoint) involves recognition of antigen. Newly formed DP cells reactivate RAG genes allowing rearrangement of the α chain. Like the immunoglobulin light chain, the T cell receptor α chain has no D segment and the first even is to direct rearrangement of Vα to Jα region genes. Suitable pairing of α and β chains at the α-selection checkpoint allows T cells to proceed to the next selection stage. Evidence shows, however, unlike the β chain that largely permits rearrangement of only one β gene through allelic exclusion, α chain rearrangement can continue to generate a second chain. In fact, up to 30% mature human T cells express more than one rearranged α chain. This implies that T cells like B cells undergo a degree of 'receptor editing' of the α chain in order to increase the likelihood of positive selection of cells selected to interact with self MHC. The potential for α-β pairing in combination with TCR gene rearrangement allows for a missive repertoire of TCR structure. Interestingly, however, some 95% of these structures fail to contribute to the T cell repertoire found in peripheral lymphoid tissues. This is

7 because thymocytes undergo a rigorous education before they exit the thymus. Education requires preliminary selection of cells for survival and their subsequent commitment to a particular lineage (positive selection). This is then followed by death of those cells that interact strongly with MHC (negative selection). In other words, T cells are positively selected for 'usefulness' (MHC restriction) and negatively selected against 'dangerous' autoreactivity. The controlling element in thymic education is the MHC expressed antigenpresenting cells (APC) in the thymus (Figure-4). This is such that T cell development is blocked at the DP stage in a thymus that does not express MHC. In fact, cell at this stage of development need to be nurtured by cells expressing MHC. Cells whose TCR fails to engage either a class I or class II MHC molecules undergo programmed cell death (death by neglect), while cells that recognize MHC with moderate affinity on cortical epithelial cell survive. Cells mature from DP cells to the single positive (SP) cells where they express either CD4 or CD8. It is clear that MHC plays a role in this selection process. Thus mice lacking class I MHC protein in the thymus have few CD8 single positive cells, while mice lacking class II have few CD4 single positive cells, it is likely that evolution has shape complementarily-determining regions (CDR) 1 and 2 of the TCR, so that the TCR preferentially matches MHC molecules. Signaling via CD4 and CD8 drives lineage commitment Why has the immune system evolved two separate types of T cell? Would it not be more economical just to have on double positive cell that could interact with either class I or class II expressing cell? Selection of cells expressing either CD4 or CD8 has evolved just as the need for two pathways of antigen processing has been driven by encounter of vertebrates with increasingly sophisticated pathogens. Class I and class II pathways has evolved to allow the immune system to recognize either cytoplasmic or extracellular/intravacuolar infectious agents respectively. This has then driven evolution of two subsets of T cell equipped with the means to help eradicate these infectious agents. The remaining question is how CD4 and CD8 cells develop form one common precursor? Until recently there were two theories to explain this process. The instructive model predicts that the co-receptor molecules CD4 and CD8 are involved in signaling. Thus recognition and simultaneous binding of class II MHC by suitable TCR and associated CD4 molecules directs inactivation of CD8 expression. Accordingly, a cell with a class II binding TCR develops into a CD4 cell and a cell bearing a class I binding TCR into a CD8 cell. The alternative hypothesis is that this whole process occurs in a random or stochastic fashion. Recent evidence strongly favours the instructional model. Logic dictates that a newly rearranged TCR should not necessarily distinguish between class-i or II MHC prior to thymic selection. It is therefore reasonable to assume that a signal for lineage commitment might be delivered by co-receptors, CD8 and CD4. These molecules are, after all, known to have an inherent affinity for class I and class II molecules respectively. This belief is supported by the fact that transgenic mice expressing a class I restricted TCR aberrantly generate predominantly CD4 cells when their developing T cells express a hybrid CD8 (extracellular)/cd4 (Intracellular) co-

8 receptor molecule. This suggests that intracellular signaling via the CD4 intracellular domain leads to selective commitment along the CD4 lineage with associated inactivation of CD8 expression. Recent experiments have revealed the nature of a molecular switch that associates with co-receptor molecules and controls lineage commitment. The src-family kinase Lck plays an important role in this process and is known to associate better with CD4 when compared with CD8 molecules. In DP thymocytes some 25-50% of surface CD4 associate with Lck, compared with only 2% for CD8. Consistent with a role in lineage commitment, it turns out that a constitutively active Lck molecule drives over selection of CD4 cells, whereas a constitutively inactive Lck drives CD8 selection (Figure-5). These results show that the intensity of Lck signaling is crucial for the development of lineage commitment. Further experiments indicate the strength of signaling via the TCR can also influence lineage commitment. Artificially increasing the signal delivered via the TCR during thymocyte development has been shown to encourage the generation of CD4 SP cells even among class I restricted cells. DP cells therefore dictate their own fate depending on the strength of integrated signal delivered via both their cell surface TCR and coreceptor-linked Lck molecules. At low signal strength CD8 selection takes place, while at a higher strength CD4 commitment is observed (Figure-6). Co-receptor molecules dictate lineage commitment by virtue of the amount of the src-kinase Lck that they bring to the TCR-MHC synapse during positive selection (see Figure-21, and 22). Antigen recognition is important for development of the T-cell repertoire Do T cells need to see antigen for positive selection and if so does this have to be a specific MHC-bound peptide? Mice deficient in the proteins required to transport peptides into the endoplasmic reticulum (TAP proteins) do not allow selection of CD8 cells. This proves that peptide antigen in conjunction with MHC class I is required for CD8 cell differentiation. But how many peptides are required for a completely functional T-cell repertoire. This question has been addressed in the class Ii system through the creation of transgenic mice in which the vast majority of class II MHC molecules are occupied by a single peptide. These mice produce CD4 sp cells are able to respond to a number of different antigens. Careful analysis of these mice, however, reveals their T-cell repertoire is far from normal. Although the number of cells present in these mice was only reduced by about 50%, the resulting repertoire was stunned and the single MHCpeptide complex was unable to select a number of known TCRs when these were introduced as transgenes. T-cell selection is compartmentalized in the thymus The thymus is made up of lobes, each of which is organized into outer cortical and inner medullary regions (Figure-7). Immature lymphocytes are found in the cortical region associated with cortical epithelial cells. Cells in the outer cortex are rapidly proliferating immature cells. Cells in the inner cortex are more mature DP cells probably undergoing

9 positive selection. The medulla contains mature SP lymphocytes, medullary epithelial cells and bone marrow-derived macrophages and dendritic cells. It is a matter of hot debate as to whether spatial separation of MHC on different APCs and their isolation in different thymic region influence positive and negative selection. Clearly haematopoietic cells are restricted to the medulla. Two important considerations are the fact that the thymic cortex is relatively inaccessible to large circulating proteins because of its vascular supply and the observation that cortical epithelial cells are inefficient at presenting exogenous proteins. These cells would thus be predicted to present endogenous antigens only and not antigens carried to the thymus in the blood supply. Cortical epithelial cells definitely play a role in positive selection because mice expressing class II MHC only on these cells show normal levels of positive selection but impaired negative selection. By contrast, bone marrow-derived macrophage and dendritic cells account for the removal of at least 50% of all positively selected cells. A further question relates to how the thymus could ever possibly express all of the antigens that a T cell might encounter outside of the thymus. There seems little doubt that the thymus does not express all potential self antigens. Nevertheless, there is increasing evidence suggesting that medullary epithelial cells can express antigens such as insulin from the pancreas and proteolipid protein from brain, previously thought to be expressed only in peripheral tissues. Medullary epithelial cells may, therefore, contribute to negative selection either by direct presentation of antigen or possibly by transfer of antigens to myeloid APCs such as dendritic cells (Figure-8). T-cell development includes a series of checkpoints In conclusion, the architecture of the thymus appears to be designed to compartmentalize thymic selection. Cortical epithelial cells present a wide range of endogenous antigens and contribute to positive selection. Interestingly, it is estimated that a developing thymocyte might only ever interact with a single cortical epithelial cell. The imprint that this leaves on the T cell clearly has a profound effect on the resulting T cell repertoire. Medullary APCs have access to circulating antigens and are largely responsible for negative selection. It is now also clear that antigens from a wide variety of tissues are expressed at low levels in the thymus, most probably by medullary epithelial cells. These cells may well contribute to negative selection either alone or in partnership with myeloid cells. Furthermore, there is evidence that antigen recognition in the thymus may contribute to the generation of regulatory T lymphocytes that play such an important role in peripheral tolerance (see below). Checkpoints in central T cell tolerance include: β-selection checkpoint: only cells with a rearranged β-chain mature from double negative to double positive cells. This process is not dependent on MHC proteins. Α-selection checkpoint: cells expressing an αβ complex must interact with MHC to survive.

10 Lineage commitment checkpoint: cells are instructed to repress expression of either CD4 or CD8 and to develop into single positive cells. Negative selection checkpoint: cells that interact strongly with MHC and antigen in the thymus are deleted (Figure-9). It is important to consider how a cell with one TCR can receive signals instructing it to survive and undergo lineage commitment without coincidently receiving a signal to undergo negative selection. The most likely explanation is that this relies on the avidity of the cells' interaction with MHC and peptide. Experiments have shown that the decision to undergo positive negative selection is directly related to the half-life of TCR binding to the MHC-peptide complex (Figure-10). Selection also depends on the architecture of the thyumus, the nature of APCs in the cortex versus the medulla of the thymus and the types of antigen that these cells are able to present. PERIPHERAL OR POST-THYMIC TOLERANCE TO SELF ANTIGENS There is no doubt that many potentially autoreactive T cells escape central tolerance. This reflects the fact that many antigens are either not present or are present at insufficiently high levels to induce tolerance in the thymus. Thus, for example, peripheral blood lymphocytes from healthy individuals respond vigorously to purified myelin basic protein, a major constituent of myelin in the brain, following their culture in vitro (Figure-11). So, how are these cells kept at bay in healthy individuals and why are autoimmune disease directed to such proteins so incredibly rare? This is because various mechanisms have evolved to maintain tolerance in peripheral lymphoid organs (Figure- 12). Sequestration of antigen occurs in some tissues Both developing and mature lymphocytes may never encounter self antigens. Many of these are sequestered away from the immune system y physical or immunological barriers. In this way, tissue antigens may never be available to T lymphocytes, either because of their location or the fact that they may never be processed by functional APCs. Privileged sites are protected by regulatory mechanisms Cells that have escaped tolerance in the thymus can also ignore self antigens of they are expressed in a privileged site. Within these sites pro-inflammatory lymphocytes are controlled either by apoptosis (Fas-ligand expression) or cytokine (transforming growth factor-β/interleukin 10, TGFβ /IL-10) secretion. Well-characterized immunologically privileged sites include the brain, anterior chamber of the eye and testes. These are defined as privileged sites because transplanted tissues have an enhanced chance of survival within them. Immune privilege in the eye is known to result from an active down-regulation of systemic and local immunity rather then 'ignorance'. Antigens introduced into the anterior chamber of the eye are collected by APCs and subsequently

11 carried to the spleen. In this case, antigen-specific regulatory T cells are generated in the spleen. Regulatory T cells generated by this process of anterior chamber-associated immune deviation (ACAID) can be adoptively transferred to naïve animals where they confer antigen-specific protection against inflammatory responses. A similar process of immune privilege is found in the brain. Here, however, the process depends on presentation of antigen in cervical lymph node rather than the spleen. Does immune privilege contribute to self tolerance or is it a phenomenon only seen by the introduction of foreign antigen to the privileged site? Immune privilege is clearly designed to dampen down inflammatory responses in certain vital organs. The same suppressive mechanisms would equally apply to inflammatory caused by an immune response to either an infections or self antigen. Immune privilege does not discriminate and therefore contributes to peripheral tolerance, at least in these particular organs. T-cell death can be induced by persistent activation or neglect Apoptotic death of lymphocytes is an extremely important mechanism of immune control and is essential for the maintenance of immune homeostasis in healthy individuals. It contributes both to the deletion of cells with high avidity for antigen and death of lymphocytes when the immune response is no longer required. These functions are fulfilled by two distinct mechanisms, activation-induced cell death (AICD) and passive cell death (PCD). Cells repeatedly stimulated with antigen undergo AICD by mechanisms involving so-called 'death receptors' of the tumor necrosis factor-receptor family. Among these, the most important molecule is Fas that on cross-linking by its ligand (FasL) leads to activation of the caspase cascade via caspase 8 and subsequent apoptopic death of the cell (Figure-13). This can occur by cell-cell interactions and there is, in addition, evidence that T lymphocytes can kill themselves through 'fratricidal cell death' following the secretion of soluble FasL (Figure-14). In addition, many activated cells die by PCD because their antigen is simply eliminated, as happens, for example, following clearance of an infection. Removal of the antigen then deprives cells of essential survival stimuli including growth factors. Under these conditions mitochondria in the cell respond by releasing cytochrome c. This, in combination with apoptosis activating factor 1, leads activation of the caspase cascade following cleavage and activation of caspase 9 (Figure- 13). Presumably, the survival rate of T cells that cross-react with self antigens, but which are generated during the immune response to infection, will be in the absence of AICD. The importance of the Fas pathway for AICD has been revealed by genetic defects in both mouse and man. For example, the lpr mouse has a mutation in Fas while the gld mouse has a mutation in Fas L. Both mutations lead to lymphadenopathy (expanded secondary lymphoid tissue). Importantly this lack of regulation also leads to the generation of autoimmunity, autoantibody production and nephritis with similarities to systemic lupus erythematous in humans. Note that thymus selection in the lpr mouse is normal, showing that the Fas pathway is not essential for central tolerance but is clearly required for peripheral tolerance. Recent studies have shown that analogous mutations lead to a similar form of disease known as human autoimmune lymphoproliferative syndrome (ALPS) characterized by

12 defective lymphocyte apoptosis, lymphocyte accumulation and humoral autoimmunity. The ALPS phenotype is associated with inherited mutations in the Fas gene (ALPS type la) or the Fas ligand (ALPS type 1b). Both activation-induced and programmed cell death are tightly regulated and the two apoptopic pathways are under independent regulation. Bel family members, for example, block PCD, by inhibiting the release of cytochrome c, but do not affect AAICD. AICD, on the other hand, is inhibited by proteins binding to the death receptor complex. Of these most important is FLIP (FLICE inhibitory protein where FLICE is the FADD-like IL-1β converting enzyme). FLIP binds to the adaptor protein FADD or a precursor form of caspase-8 and blocks generation of the Fas-associated death receptor complex. AICD is also regulated by IL-2. This cytokine stimulates Fas-mediated AICD by enhancing transcription and expression of Fas L while inhibiting transcription of FLIP. Note that disruption of the genes for IL-2, IL-2Rα or IL-2Rβ leads to lymphadenopathy and autoimmunity in these knockout mice apoptotic death (propriocidal death) in homeostasis and peripheral tolerance. The balance of co-stimulatory signals affects immune homeostasis and self tolerance Naïve T lymphocytes require tow signals to proliferate and differentiate. The first signal is triggered by TCR recognition of the appropriate peptide-mhc complex. The second signal is delivered by CD80 (B7.1) and CD86 (B7.2) co-stimulatory molecules expressed by APCs. How a T cell interprets co-stimulation depends on which co-stimulation receptor it uses. The CD28 mlecule is constitutively expressed on T cells and signaling via CD28 enhances cell survival, prevents anergy induction and enhances CD40L expression. Ligation of CTLA-4 (CD152), however, inhibits T cell responses. At the molecular level CTLA-4 ligation inhibits early T cell activation including expression of the IL-2 receptor α chain and secretion of IL-2. The CTLA-4 pathway also inhibits IL-2 messenger RNA accumulation and inhibits upregulation of cyclindependent kinases 4 and 6, hence inhibiting progression though the cell cycle. CTLA-4 has a higher avidity (100x) for CD80 and CD86 but is normally isolated to the peri-nuclear Golgi apparatus. On T cell contact with an APC,CTLA-4 traffics to the plasma membrane at the TCR-APC interface (Figure-15). The role of CTLA-4 in normal homeostasis is revealed in the CTLA-4 knockout mouse. These mice show normal thymus selection but suffer from polyclonal T cell expansion and die from a fatal lymphoproliferative disease. Interestingly, CTLA-4 influences CD4 cells more than CD8 cells, since depletion of CD4 cells from the CTLA-4 knockout mouse prevents the lymphoproliferative disease in this mouse. Lymphoid dendritic cells contribute to peripheral tolerance Lymphocytes do not possess an inherent capacity to distinguish between foreign and self antigens. One selected, autoreactive cells are controlled by regulatory mechanisms but their activation is not differently controlled. How then are responses to foreign antigens induced while avoiding autoimmune responses? Foreign antigens encountered by the immune system are predominantly components of infectious agents and the immune

13 system has evolved ways to recognize the inherent adjuvant properties of these infectious agents. It is known that these elements upregulate co-stimulatory molecules but they also control migration of APC into the T cell zones of secondary lymphoid tissues. Most dendritic cells in the T-cell zones of resting lymph nodes are lymphoid dendritic cells that arise from the same progenitor cells as T and B lymphocytes. Myeloid dendritic cells are normally found outside the T cell zones but migrate into the zone when they first encounter the types of adjuvants contained in infectious agents (Figure-16). In addition, myeloid dendritic cells are functional APCs whereas lymphoid dendritic cells are unable to internalize exogenous antigens. Lymphoid dendritic cells do, however, present endogenous antigens and recognition of these 'self' antigens presented by lymphoid dendritic cells leads to apoptopic cell death among potential autocreative T cells. Only when antigen-bearing myeloid dendritic cells migrate into the T cell zones does the balance swing in favour of immunity to antigen presented by these cells. In summary, homeostatic balance in the immune system is required to prevent lymphoproliferative responses. Lymphoproliferative disorders are associated with responses to both foreign and self antigen. Lack of homeostatic control leads to autoimmunity and hence the molecules involved in homeostatic control are important regulators of peripheral tolerance. These molecules include: CTLA-4 that acts as a break on the normal immune response to both foreign and self antigens. Members of the TNF-R family, particularly Fas and Fas L. Components of the caspase cascade. IL-2 and the IL-2 receptor that together regulate sensitivity to Fas-mediated apoptosis. The role of regulatory lymphocytes in peripheral tolerance Peripheral tolerance to antigens can be 'infectious'. The experimentally induced tolerance to one antigen can thus maintain tolerance or suppress the immune response to a second antigen as long as the two antigens are structurally or physically associated (e.g. within the same tissue). This implies that mechanisms other than ignorance and cell death must be involved in tolerance. One explanation for such phenomena depends on the existence of two populations of T lymphocytes that produce distinct cytokines. Many inflammatory autoimmune disease are caused by TH1 cells that produce cytokines such as interferon-γ (IFNγ) and tumor necrosis factor-α (TNFα). Cytokines derived from TH2 calls (IL-4, IL- 5, IL-6, IL-10) support antibody production. A major additional effect of TH2-derived cytokines such as IL-10, however, is down-regulation of macrophage effector functions, including antigen presentation to TH1 and naïve T cells. TH2 cells are thus able to suppress inflammatory (derived-type hypersensitivity, DTH) responses. TH1-cell-derived IFNγ can prevent the differentiation of TH0 to TH2 cells. This type of immune deviation was defined more than 30 years ago to describe how an individual animal could respond to the same antigen in two completely different ways. Guinea pigs primed with antigen in

14 alum produced high levels of IgG1 antibody, but did not support a DTH response, whereas animals primed with the same antigen in complete Freund's adjuvant developed strong DTH responses. It was subsequently suggested that the ability of the same antigen to induce either 'humoral' or 'cellular' immunity could reflect the distinct activation of two mutually antagonistic arms of the immune system. The results of these experiments were undoubtedly a form of immune deviation resulting from the selective induction of TH2 rather than TH1 cells. Immune deviation can influence hypersensitivity conditions. Diabetes in the NOD mouse is known to be caused by TH1 cells and can be prevented by antigen-primed TH2 cells, whereas allergic disorders can be treated by induction of TH1 cells. Peripheral T-cell tolerance depends on the genetic make-up of the individual. A transgenic model has been created in which a viral antigen (influenza haemagglutinin, HA) is expressed in the islet of the pancreas and a TCR specific for this antigen is expressed on the T cells. These double-transgenic mice are then bred with mice that differ in their non-mhc genes that is, the mice have different background genes. In one mousse strain (BALB/c background), the HA-reactive T cells produce large amounts of both IL-4 and INFγ and show no signs of inflammatory disease in the pancreas. In another mouse strain (B10.D2 background), the HA-reactive cells produce only TH1 cytokines and the T cells are able to infiltrate the pancreatic islets and cause diabetes. Immune deviation is clearly controlled b background gene, many of which combine to control the susceptibility of an individual to autoimmune disease. T-cell-mediated diseases (such as insulin-dependent diabetes, thyroiditis and gastritis) can be produced in otherwise normal mice b simply eliminating a subpopulation of CD4 + T cells expressing CD5, CD25 or a particular isoform of CD45. This is best illustrated by the transfer of subsets of kurine cells isolated from healthy mice into Rag/ mice that otherwise do not contain T cells of their own. Naïve and activated murine T cells can be distinguish according to the level of cell surface expression of CD45 RB activated cells have low level of this isoform and naïve cells have high levels. Activated or naïve populations are then transferred into recipient Rag/ mice. Transfer of naïve cells leads to inflammatory bowel disease (IBD) in the Rag/ recipients, but co-transfer of relatively few activated cells prevents disease (Figure-17). The activated cells either produce or induce production of the immune suppressive cytokine TGFβ. Another cytokine, IL-10, undoubtedly plays a role in the function of the activated, regulatory cells because transfer of these regulatory cells frokm an IL-10 knocknout mouse fails to suppress the induction of IBD by naïve cells. A further subset of regulatory lymphocytes is distinguished by expression of the α chain of the IL-2 receptor (CD25). Elimination of these cells from normal mice leads to the generation of various organ-specific autoimmune conditions. These natural regulatory T cell are anergic to TCR-mediated activation but potently suppress the activation of other T cells. In contrast to cells such as the CD45RB low regulators, however, the CD25 + CD4 T cells are probably cytokine independent but suppress other cells by an APC-dependent mechanism involving cognate cellular interactions.

15 It is now clear that natural regulations play a central role in maintaining self tolerance. How and where are these cells generated? Interestingly not all CD25 + cells serve as regulators. The CD25 marker is upregulated on naïve T cells in response to antigen and yet this is normally associated with an active immune response rather than regulation It now appears that the thymus plays a crucial role in generation of CD25 + regulators (Figure-18). Thymectomy of mice before thymus-derived CD25 + cells have had time to populate peripheral lymphoid organs leads to generation of autoimmune diseases similar to those caused by elimination of CD25 + cells from lymphoid organs of the adult mouse. How than does the immune system ever make a response to foreign antigens and infectious if CD25 + cell are such potent regulators? Note that these cells are normally anergic. Both their anergic and reulatory phenotypes can, however, be temporarily reversed by high local levels of cytokines such as IL-2. Thus a strong immune response to an infectious agent would temporarily overcome regulatory activity of the CD25 + cells, thus permitting the localized immune response to the infectious agent to proceed. It would appear that the active regulatory T cells identified in mice are antigen specific and regulate peripheral tolerance by the production of immune suppressive cytokines such as TGFβ and IL-10. The thymus-derived CD25 + population, on the other hand, appears not to depend on such cytokines and may suppress neighbouring T cells in a contact-dependent fashion. Other subsets of lymphocytes may also contribute to the regulation of immune responses. In certain case immune suppressivecd8 + T cells have been described. These could in fact be cells of TCR-γδ type, since CD8 + γδ T cells have been shown to both suppress allergic responses and control diabetes in a mouse model of the disease. Diabetes has also been controlled in mice by transfer of theymocytes bearing the NKI marker that appears to produce the TH2 cytokines IL-4 and IL-10. In conclusion, regulatory lymphocytes play a crucial role in the control of autoimmune responses. Scientists are only just beginning to reveal the mechanisms by which these cells mediate their suppressive activity. Furthermore, we know very little about how these cells are generated in the normal immune repertoire of healthy individuals. There is little doubt that clarification of these questions will help in the control of many hypersensitivity conditions including allergic and autoimmune diseases. B-CELL TOLERANCE TO SELF ANTIGENS High-affinity IgG production is T cell dependent. For this reason, and because the threshold of tolerance for T cells is lower than that for B cells, the simplest explanation for non-self reactivity by B-cells is a lack of T-cell help. Until quite recently it was therefore felt that immunological tolerance should be the responsibility of T lymphocytes alone. The logical argument being: why evolve a complex process of immunological tolerance amongst cells such as B cells when they are then allowed to hypermutate? All nature would have to do would be to evolve a system in which T cells never react with self antigens; a 'perfect' system would delete all selfreactive T cells. The immune system would then allow the B cell repertoire to expand widely and generate as many autocreative B cells as possible by random rearrangement.

16 Without help from T cells these B cells would remain harmless. Now we appreciate that the immune system is not allowed to be 'perfect'. If it were, it would almost inevitably develop 'holes' in the immunological repertoire through which faster evolving microorganisms would inevitably break. We now appreciate that autoreative T cell exist in all of us and that the balance between health and autoimmunity is a fine one, dependent on numerous polymorphisms in regulatory mechanism. It is now clear that the B lymphocyte pool is subject to analogous but subtly different mechanisms of immunological tolerance to those that apply to the T lymphocyte pool. There are clearly circumstances in which B cell must be tolerized directly. For example, some microorganisms have cross-reactive antigens that have both foreign T cell-reactive epitopes and other epitopes that resemble self epitopes and are capable of stimulating B cells. Such antigens could provoke a vigorous antibody response to self antigens (Figure- 19). Furthermore, in contrast to TCRs, the immunoglobulin receptors on mature, antigenically stimulated B cells can undergo hypermutation and may acquire anti-self reactivities at this late stage. Tolerance thus needs to be imposed on B cells both during their development and after antigenic stimulation in secondary lymphoid tissues. Self reactive B cell may be deleted or anergized, depending on the affinity of the B cell antigen receptor and the nature of the antigen. Tolerance induction by self antigen can lead to one of several results, such as deletion or anergy. The outcome depends on the affinity of the B cell antigen receptor and on the nature of the antigen it encounters, whether this is an integral membrane protein or a solube and largely monometric protein in the circulation. The fate of self-reactive B cells has been determined using transgenic technology (Figure-20, and 21). B-cells pass through several development checkpoints B-cell development shows similar features to T cell development but takes place largely in the bone marrow. Development is marked by ordered expression of surrogate receptor molecules. The first notable event takes place at the late pro-b stage when the CD79a and CD79b (Igα and Igβ) molecules appear at the cell surface (see Figure-1, and 3). Progression to the pre-b stage is accompanied by VDJ recombination at the heavy chain locus. The rearranged heavy chain can then appear at the cell surface in combination with both CD79a and b and the VpreB and γ5 molecules that act as surrogate light chains (see Figure-2). There are clear analogies with T-cell development here. T cells rearrangement the heavy chain genes first and also employ a surrogate light chain. In addition, it appears that expression of the B cell receptor is necessary for successful allelic exclusion at the heavy chain locus. As pre-b develop, recombination at the light chain locus proceeds and the cells become immature B cells. At this point the cells are IgD. B cells then go through a 'transitional' stage in development becoming IgD low as well as IgM +. Mature B cells express IgD at higher levels than IgM. Checkpoints in B-cell development include: Successful expression of CD79a and b in late pro-b clls.

17 Successful rearrangement at the heavy chain locus in pre-b cells. Successful rearrangement at the light chain locus and receptor editing. Self tolerance begins when IgM first appears at the surface of the developing cell. Immature cells are resistant to apoptosis, although the development of immature B cells can still be blocked by interaction with self antigen. However immature B cells can edit their receptor, thus allowing development to progress. Receptor editing allows potentially self-reactive B cells to continue development When the IgM receptor on an immature, bone marrow B cell reacts with self antigen further cell differentiation is blocked but light chain rearrangement can continue. If the new IgM receptor does not react with s self antigen in bone marrow, B-cell development can proceed. Interestingly the 'immature' B cell is relatively resistant to apoptotic cell death, whereas the late 'transitional' cell is sensitive. Allowing light chain rearrangement to continue among immature cells permits the B cell to edit its receptor and thus rescue potentially autocreative cells from inevitable death. This mechanism of altering offending receptors before the B cell becomes sensitive to antigen-mediated cell death clearly allows the immune system to optimize the generation of its repertoire. In fact it looks like mechanisms have evolved to promote receptor editing in B cells. The K light chain locus can be inactivated by recombination of a recombining sequence (RS). Recombination of the RS element results in deletion of CK and other sequences required for transcription of the K allele. It is important to appreciate that 40-60% of IgM + λ + B cells carry a VK JK rearrangement inactivated by this RS recombination. Receptor editing therefore plays an important role in generation of the normal B cell repertoire. Self reactive B cells are usually deleted As B cells mature through the transitional stage they become poor at reactivating the recombinase activating genes. So, if receptor editing has failed to eliminate autoreactive B cells they are likely to be eliminated by apoptotic death, since they can no longer select a new receptor. Interestingly these transitional IgM hi IgD lo cells can immigrate to the periphery where the high IgM level ensures that apoptotic cell death can still take place. Peripheral B cell tolerance Recent studies have shown that the bone marrow exports a higher proportion of new B cells than the thymus does new T cells. Also, unlike T cells leaving the thymus, B cells leaving the bone marrow are relatively immature. These cells express the heat stable antigen (HAS) and migrate from the bone marrow to the outer T cell zone of the spleen. It is among these cells that the most significant amount of negative selection takes place. This splenic environment eliminates unwanted B cells as effectively as the thymus removes unwanted T cells (Figure-22). Self reactive B cells are purged by a process

18 that (i) induces anergy; (ii) prevents migration into B cell follicles, and (iii) rapidly leads to cell death. Self-reactive B cells in the outer T cell zone are short-lived (1-3 days), whilst cells selected to enter B cell follicles becomes long-lived and recirculate from 1 to 4 weeks. Short-lived, autoreactive B cells may, however, contribute to immune responses to foreign antigens. Anergy in such an autoreactive B cell can be overcome by highavidity antigen. Self-reactive B cells be overcome by high-avidity antigen. Self-reactive B cells may then be recruited into the functional immune repertoire because their B cell receptor cross-reacts sufficiently strongly with foreign antigen. At first sight, allowing B cell tolerance to self antigen to be overwhelmed by foreign antigen would seem a risky business. Why promote a pluripotent B-cell repertoire at the risk of autoimmunity? This question must be considered a balance between the risk of autoimmunity and protection of the species from infectious pandemics. The immune system has evolved to ensure that among the population there will be individuals equipped to fight almost any infection. Diversity among T cells is generated by rearrangement of their receptors in contact with MHC polymorphism. MHC polymorphism and T cell receptor selection together provide sufficient diversity across the population to protect against almost any infection. MHC polymorphism has evolved to fill holes in the protective repertoire of the species. Likewise the fact that the B-cell pool permits a proportion of short-lived and potentially autoreactive B cells to emigrate to the periphery provides an additional level of protection from infection. In the absence of an infection these cells die rapidly. This short-lived pool of cells can, however, contribute to broadening the potential B cell repertoire. Such repertoire diversity and further differences between the repertoires of individuals within the population will ensure that there will always be some individuals ready to mourn an effective immune response to infection. A second window of susceptibitly to tolerance occurs transiently during the generation of B cell memory. Secondly B cells (derived from memory B cells produced by T celldependent stimulation) are highly susceptible to tolerance by epitopes presented multivalently in the absence of T cells help. Such a tolerance-susceptible stage probably ensures that newly derived memory B cells that have acquired self reactivity (as a result of accumulated) somatic mutations) are purged from the repertoire. ARTIFICIALLY INDUCED TOLERANCE IN VIVO Chimerism is associated with tolerance Tolerance can be induced by the inoculation of allogeneic cells into hosts that lack immunocompetence, for example neonatal host, or adult hosts after immunosuppressive regiments such as total body irradiation, durg (e.g. cyclosporine) or anti-lymphocyte antibodies (anti-lymphocyte globulin, anti-cd4 antibodies, etc.). For tolerance to be maintained, a certain degree of chimerism the coexistence of cells from genetically different individuals must be maintained. This is best achieved if the inoculum contains cells capable of self renewal (e.g. hone-marrow cells).

19 If mature T cells are present in the injected cell population, they may react against the histocompatibility antigens of their host and induce a severe and often fatal disease known as graft-versus-host-disease. Antibodies to co-receptor and co-stimulatory molecules induce tolerance to transplants Tolerance of transplanted tissues can be achieved in adult animals by monoclonal antibodies directed against the T cell molecules, CD4 and CD8 (the antibodies can be either the T cell-depleting or the non-depleting type). In this situation, tolerance of skin allografts is obtained even in the absence of cellular chimerism. Another highly promising approach to transplantation tolerance has arisen through the use of agents designed to blockade co-stimulatory molecules. As mentioned above, T cells require co-stimulatory signal for effective priming. CD28 and CD154 both play important roles in co-stimulation. The CD28 pathway of activated can be inhibited by blocking both B7 molecules with a soluble form of CTLA-4 (CT-4-IG). In combination with an antibody to the ligand for CD40 (CD154), CTLA-4-Ig has been shown to block recognition of allografts and allow long-term skin allograft survival in mice. Antibodies to CD154 alone will prolong renal allograft survival in non-human primates. It is thought that anti-cd154 prevents the three cell interplay between CD4,CD8 and dendritic cells that is required for the maturation of CD8 cells (Figure-23). Soluble antigens readily induce tolerance Tolerance is inducible in both neonatal and adult animals by administering soluble protein antigens in disaggregated form. T and B cells differ in their susceptibility to tolerization by these antigens. Thus tolerance is achieved in T cells from spleen and thymus after very low antigen doses and within a few hours. Tolerance of spleen B cells requires much more time and higher doses of antigen (Figure-24). The antigen levels which will produce B cell tolerance in neonates are about one-hundredth of those required in adults. Oral administration of antigens induces tolerance Orally administered antigens induce tolerance to themselves by a variety of mechanisms. High doses of antigen can cause anergy or deletion; this could be one situation in which the term anergy is being used to describe a state of paralysis preceding cell death. Lower doses of antigen can, however, induce the priming of T cells in the gut. As an effective antibody response in the gut requires class-switching to the IgA isotype, it comes as no surprise that antigen feeding induces T cells that support IgA production. These 'TH3- like' cells produce cytokines including IL-10 and TGFβ and serve to inhibit inflammatory responses mediated by TH1 cells. TGFβ inhibits the proliferation and function of B cells, cytotoxic T cells and NK cells, inhibits cytokine production in lymphocytes and

20 antagonizes the effects of TNF. Although the induction of mucosal TH3 cells of antigen specific, the suppressive activity of cytokines such as TGFβ is not. Hence, the induction of oral tolerance to one antigen is able to suppress the immune response to a second, associated antigen. The effect of 'bystander suppression' by mucosal TH3 cells allows the suppression of complex organ-specific autoimmune diseases by feeding with a single antigen derived from the affected tissue. Other mucosal surfaces are providing equally effective as routes for the induction of antigen-specific tolerance. Nasal deposition of calss II-restricted peptides has been used to control both 'humoral' and 'cellular' immune responses. In addition, administration of aerosolized antigen to the lung can be used to control either allergic responses to foreign antigens or autoimmune responses to self antigens. The fact that one can inhibit autoimmune (TH1 type) and allergic (TH2 type) responses argues against immune deviation as a mechanisms for tolerance induction in this case. Recent evidence suggests that cell producing IL-10 may be responsible for nasal peptide-induced tolerance. It could be that other cell types, including CD8 cells, are responsible for suppressing the immune response that follows aerosol administration of protein antigens. Extensive clonal proliferation can lead to exhaustion and tolerance Tolerance in T cells, and to a lesser extent in B cells, can be due to clonal exhaustion, the end result of a powerful immune response. Repeated antigenic challenge may stimulate all the antigen-responding cells to differentiate into short-lived end cells, leaving no cells that can respond to a subsequent challenge with antigen. Anti-idiotypic responses can be associated with tolerance An antibody's combining site may act as an antigen and induce the formation of 'antiidiotypic antibodies'. By cross-linking immunoglobulin on B cells, these antibodies can block B-cell responsiveness of the cell. Because, in some animals, most of the antibodies produced in response to particular antigens bear a particular idiotype, suppression of this idiotype by anti-idiotype antibody can significantly alter the response. This type of tolerance will be partial, however, because it affects only those B cells carrying the idiotype. The idiotype of a T cell is represented by the polymorphic regions of the TCR α and β chains. TCR-specific regulatory cells, induced after vaccination with self-reactive T cells, TCR peptides or even DNA-encoding TCR molecules, have been shown to prevent autoimmunity in animal models. Tolerance in vivo relates to persistence of antigen Persistence of antigen plays a major part in maintaining a state of tolerance to it in vivo when the antigen concentration decreased below a certain threshold, responsiveness is restored. If the tolerance results from clonal deletion, recovery of responsiveness is

21 related to the time required to generate new lymphocytes from their precursors. This tolerance can be prevented by measures such as thymectomy. If, on the other hand, the tolerance is maintained by a suppressive mechanism resulting, for example, from the induction of regulatory T cells, then the state of tolerance can be relatively long lived. POTENTIAL THERAPEUTIC APPLICATIONS OF TOLERANCE A better understanding of tolerogenesis could be valuable in many ways. It could be used to promote tolerance of foreign tissue grafts to control the damaging immune responses in hypersensitivity states and autoimmune diseases. The various ways of establishing artificial tolerance in adult animals are being investigated for their potential clinical applications. Some success has been obtained in the case of transplants associated with chimerism and performed under the umbrella of immunosuppressive agents. Treatment with monoclonal non-depleting anti-cd4 and anti-cd8 or anti-cd154 antibodies has also been used successfully with foreign tissue or organ transplants. The possibility that tolerance can be induced either by mucosal administration of the target antigen or through the use of peptide drugs awaits extensive clinical trials in humans. It is also important to learn how to activate T cells that ignore certain antigens, so as to enable the immune system to mount an appropriate active response. This could be exploited to limit the growth of tumors that may express their own unique tumor-specific genes. Figure-1 The experiment demonstrates the induction of specific tolerance to grafted skin, induced by neonatal injection of spleen cells from a different strain. Mice of strain A normally reject grafts from strain B. However, if newborn mice of strain-a receive cells

22 from a strain B, they show tolerance to skin grafts from this donor at 6 weeks of age, but reject grafts from other strains (C). This phenomenon is due to immune deviation. Figure-2 Both B and T lymphocytes develop from a common precursor found in fetal liver or adult bone marrow. T cell progenitors give rise to both the αβ and γδ lineages as well as NK T cells. Recent studies have that early T-cell progenitor cells also give rise to lymphoid dendritic cells in the mouse.

23 Figure-3 Precursor thymocytes develop into 'double positive' cells expressing low levels of the αβ TCR. These undergo positive selection for interaction with self MHC class I or class II molecules on cortical epithelium. Unselected cells (the majority) undergo programmed cell death by apoptosis. Cells undergoing positive selection lose one or the other of their co-receptor molecules (CD4 or XD8). Finally, self-reactive cells are eliminated by their interaction with self peptides presented on cells at the corticomedullary junction and in the thymic medulla.

24 Figure-4 Host mice (F1[H-2 b xh-2 k ]) were thymectomized, then engrafted with 14-day fetal thymuses of various genotypes. They were subsequently irradiated to remove their resident T-cell populations, and then reconstituted with F1 bone marrow to provide stem cells. After proming with antigen (keyhole limpet Hemocyanin, KLH, the proliferateive response of lymph node T cells to KLH on APCs from each parental strain was evaluated. In some experiments, thymus lobes were incubated before grafting with deoxyguanosine (dg), which destroys intrathymic cells of macrophage/dendritic cell lineage. The results show (i) that the thymic environment is necessary for T cells to learn to recognize MHC and (ii) that bone marrow-derived cells (removed by dg treatment) are not required for this process to occur.