Molecular

REVIEW Molecular and genetic parameters defining T-cell clonal selection Nathalie Labrecque1,2,3, Troy Baldwin4 and Sylvie Lesage1,3 Clonal selection of T cells occurs in the thymus and is responsible for generating a useful and functional repertoire of T cells. Aberrations in clonal selection result in altered T-cell homeostasis in the secondary lymphoid organs ranging from an absence of T cells to an overabundance of autoreactive T cells. The advent of new technologies facilitating the manipulation of the mouse genome has helped refine our understanding of the molecular and genetic pathways involved in clonal selection and has also revealed a high degree of complexity. Herein, we attempt to review recent advances in thymic selection processes, achieved mostly through genetic manipulations. Immunology and Cell Biology (2011) 89, 16–26; doi:10.1038/icb.2010.119; published online 19 October 2010 Keywords: negative selection; positive selection; thymus The clonal selection theory, described by Sir Frank Macfarlane Burnet in 1957, pertained to the specificity of antibodies to different antigens. Not only does this theory still holds true today, it has been broadened to include specificity of T cells to antigen. Indeed, the response of the adaptative immune system is mediated by two cell types, namely B cells and T cells, each of which respond to antigen through their antigen receptor; the B-cell receptor on B cells and T-cell receptor (TCR) on T cells. The B-cell receptor and TCR are produced through somatic DNA recombination events allowing sufficient diversity to recognize pathogens. All the while, selection events must occur to prevent B cells and T cells from reacting towards healthy self-tissue. This review aims to highlight genetic variations in T-cell clonal selection processes in the thymus, focusing on recent advances pertaining to specific molecular pathways in positive and negative thymic selection processes. We will first briefly introduce T-cell differentiation events in the thymus as well as common tools and experimental approaches used to dissect clonal selection. OVERVIEW OF T-CELL DIFFERENTIATION IN THE THYMUS Hematopoietic progenitors of myeloid, lymphoid or mixed myeloid/ lymphoid potential enter the thymus and, although a fraction of these will differentiate into dendritic cells, NK cells, B cells or gd T cells, the majority will yield ab T cells.1 Commitment to the T-cell lineage is imposed by the Notch signalling pathway at early stages of develop- ment.2 The earliest stages of ab-T-cell differentiation are characterized by a lack of CD4 and CD8 co-receptor expression (double-negative, DN) on immature T cells, thymocytes. DN thymocytes can be further subdivided into four differentiation steps (DN1 to DN4 subsets) based on CD44, CD25 and CD117 expression (Figure 1). Extensive proli- feration at the DN1 and DN2 stages allows for the expansion of the few precursors that seed the thymus. At the DN3 stage, thymocyte proliferation is halted, while somatic recombination events proceed at the TCRb locus. Only when somatic recombination events yield a productive TCRb chain do DN3 thymocytes express a functional pre- TCR, composed of pre-Ta and TCRb chains as well as the CD3 molecules. In the absence of ligand and possibly through oligomer- ization, the pre-TCR complex provides a signal allowing survival, proliferation and further differentiation of thymocytes to the DN4 stage.3 Concomitantly, pre-TCR signals result in the cessation of TCRb chain rearrangement, thus ensuring allelic exclusion. DN3 thymocytes that fail to generate a productive TCRb chain cannot express the pre-TCR and are, therefore, eliminated by apoptosis. DN4 thymocytes proliferate extensively and differentiate into double-positive (DP) thymocytes expressing both CD4 and CD8 co-receptors. At the DP stage, thymocyte proliferation is again halted to allow somatic recombination events, this time, at the TCRa locus. Following a successful rearrangement, DP thymocytes will express the abTCR at low levels relative to mature T cells. As the abTCR is generated from random and flexible juxtaposition of TCR gene segments, positive and negative selection events, respectively, test whether the TCR is useful (MHC restricted) and not autoreactive. Similar to Goldilocks and her porridge, to continue their maturation process, DP thymocytes must receive neither too little (death by neglect), nor too much signal (negative selection); the signal must be just right (positive selection). Indeed, if DP thymocytes do not recognize self-peptide–MHC complexes (spMHC) with sufficient affinity and fail positive selection, Received 9 September 2010; accepted 12 September 2010; published online 19 October 2010 1Maisonneuve-Rosemont Hospital Research Center, Montre´al, Que´bec, Canada; 2Department of Medicine, University of Montreal, Montre´al, Que´bec, Canada; 3Department of Microbiology and Immunology, University of Montreal, Montre´al, Que´bec, Canada and 4Department of Medical Microbiology and Immunology, University of Alberta, Edmonton, Alberta, Canada Correspondence: Dr S Lesage, Maisonneuve-Rosemont Hospital, Research Center, 5415, boulevard de l¢Assomption, Montre´al, Que´bec, Canada H1T 2M4. E-mail: Sylvie.lesage@gmail.com Immunology and Cell Biology (2011) 89, 16–26 & 2011 Australasian Society for Immunology Inc. All rights reserved 0818-9641/11 www.nature.com/icb they will continue to rearrange their TCRa locus until a productive and useful abTCR is generated or they will die by neglect. Alterna- tively, autoreactive DP thymocytes expressing a TCR with high affinity for spMHC undergo negative selection, which includes cellular apoptosis. Only when DP thymocytes express a useful TCR, restricted to self-MHC, are they allowed to mature, a process defined as positive selection (Figure 1). Importantly, the CD4 and CD8 co-receptors, by sequestering the tyrosine kinase lck, have a crucial function in this process. A recent study took advantage of elaborate genetic manipula- tions to support this view.4 Van Laethem et al.4 observed that although positive selection of thymocytes is completely deficient in the absence of MHC class I and class II molecules, positive selection could proceed in mice genetically deficient for CD4 and CD8 co-receptors as well as MHC class I and class II molecules. Moreover, using CD4 chimeric molecules unable to bind lck, their data support the view that co-receptors relocate lck near the TCR complex, a key event to ensure positive selection of thymocytes.4 Positively selected thymocytes proceed through further differentia- tion steps to ultimately become either CD4 or CD8 single-positive (SP) thymocytes, which is determined by their restriction to MHC class II and class I, respectively. SP thymocytes migrate to the thymic medulla in which they continue to verify the reactivity of their TCRs; autoreactive SP thymocytes are again eliminated by apoptosis. Even- tually, useful non-autoreactive SP thymocytes exit the thymus to populate the peripheral organs and compose part of the adaptative immune system. EXPERIMENTAL APPROACHES FOR TESTING THE THYMIC CLONAL SELECTION THEORY The selection processes described above are an extension of the clonal selection theory proposed by Sir Macfarlane Burnet. Together, death by neglect, positive and negative selection ensures the maturation of useful non-autoreactive T cells to form a broad repertoire of T cells recognizing a variety of antigens. Since the clonal selection theory was first proposed, there have been tremendous technical advances that have allowed it to be experimentally tested and, nevertheless, this theory withstands the test of time. Perhaps the most important of these advances has been the ability to genetically manipulate the mouse genome allowing the creation of transgenic and knock-out mice. Furthermore, in the post-genomic era, rapid advances in sequencing technologies have provided opportunities for forward genetic approaches and genetic linkage analyses to uncover molecular pathways involved in clonal selection. Studies of clonal selection in non-TCR transgenic mice have been incredibly difficult for a number of reasons. Perhaps the greatest limiting factor for the study of clonal selection in a wild-type repertoire is the exceeding low frequency of antigen-specific T-cell clones. Recently, the Jenkins laboratory pio- neered a tetramer-based protocol to identify and quantify the antigen- specific T-cell clones present within the wild-type repertoire.5 Using this protocol, they were able to examine both positive and negative selection of 2W-peptide-specific thymocytes.6 While the sensitivity of this protocol does not currently permit the evaluation of the mole- cular events of selection at the DP stage of development because of low levels of TCR expression, it is likely to be important for examining selection in situations of gene deficiency in otherwise wild-type T cells. Furthermore, following the identification of natural autoantigens targeted during autoimmunity, this tetramer-based protocol can be utilized to examine clonal selection of disease causing, autoreactive T cells in humans or mouse model systems. TCR transgenic mice have proven to be an invaluable tool for the study of positive and negative selection and recently new approaches have been utilized to generate TCR transgenic mice that more Figure 1 Overview of thymic TCRab + T-cell differentiation. Early thymic progenitors (ETP), derived from bone marrow hematopoietic stem cells, colonize the thymus where they commit to the T-lineage under the influence of the Notch signalling pathway. The different stages of T-cell differentiation can be followed using CD4 and CD8 expression. The more immature double-negative (DN; CD4�CD8�) thymocytes can be further divided into DN1 to DN4 based on their expression of CD44, CD25 and CD117 (c-kit) as depicted. A successful rearrangement of the TCRb locus allows the thymocytes to achieve their first development checkpoint, b-selection. This leads to survival, proliferation and differentiation into double-positive thymocytes (DP; CD4+CD8+). DP thymocytes can be further divided into three subsets based on their level of cell surface expression of the ab TCR. TCR-negative DP thymocytes undergo TCRa gene rearrangement; if the rearrangement is successful, DP thymocytes will express low levels of the TCR and will verify the specificity of their TCR. If the DP thymocytes’ TCR has a low affinity for self-peptide MHC molecules, the DP thymocytes undergo positive selection and continue their differentiation into single-positive (SP) thymocytes (SP4:CD4+CD8� if restricted to MHC class II molecules; SP8: CD4�CD8+ if MHC class I restricted). As a result of positive selection, DP thymocytes will increase their TCR expression level and will up-regulate CD69 and CD5. If the DP thymocyte has no detectable affinity for self- peptide MHC molecules, they will try successive TCRa gene rearrangements. If they fail to express an MHC-restricted TCR, DP thymocytes will undergo death by neglect. Thymocytes will undergo negative selection, upon expression of a TCR bearing a too high affinity for self-peptide–MHC complexes. This can occur either at the DP or SP stage of differentiation. T-cell clonal selection N Labrecque et al 17 Immunology and Cell Biology accurately reflect endogenous TCR expression patterns. Historically, TCR transgenic mice have suffered from artificial premature expres- sion of the TCR, in which expression was initiated within the DN thymocyte compartment when, under normal circumstances, the mature abTCR is not expressed until the DP stage of development. This results in a situation where DN thymocytes that are not normally able to recognize pMHC are subject to self-antigen-dependent selec- tion pressures.7 Recently, two groups have utilized different strategies to drive expression of the H-Y TCR, which is specific for the ubiquitously expressed male antigen, at the DN48 or DP9 stage of development. The Sant’Angelo group placed a DNA ‘stuffer’ sequence flanked by RSS sequences between the initiation codon and the rearranged VJa sequence of the H-Y TCRa. Following Rag-mediated recombination, this stuffer sequence is excised and expression of the H-Y TCRa chain can proceed. This strategy resulted in mature H-Y TCR expression beginning at the DN4 stage, mimicking endogenous Va2 expression in non-trangenic mice.8 The Hogquist group utilized a Cre-LoxP conditional expression approach in which a floxed tran- scriptional and translational ‘STOP’ cassette was placed upstream of the rearranged H-Y TCRa cDNA. Breeding of these conditional mice to CD4-Cre transgenic mice resulted in mature H-Y TCR expression at the DP stage.9 While positive selection appears to proceed normally when the mature TCR is expressed prematurely in conventional H-Y TCR transgenic female mice, lineage commitment, negative selection and agonist selection are dramatically altered. Therefore, in terms of understanding the genetics underlying clonal selection, and negative selection in particular, the newer TCR transgenic models that restrict TCR expression to post-b-selection thymocytes are likely to be extremely useful. Other advances in our understanding of tolerance induction to self-antigens expressed in a tissue-restricted manner, so-called tissue- restricted antigens, have also improved our understanding of clonal selection. The identification of the autoimmune regulator10,11 and the realization that it can direct tissue-restricted antigen expression in the medullary region of the thymus12 has allowed for the study of negative selection to tissue-restricted antigen using TCR transgenic mice.13–15 As TCR transgenic thymocytes will only encounter their high-affinity self-antigen following positive selection and CCR7-mediated traffick- ing to the medulla,16,17 in these model systems early TCR expression in classical TCR transgenic mice is not problematic.18 For example, the Goodnow group recently utilized the HEL-specific 3A9 TCR trans- genic mouse in conjunction with RIP-HEL mice to query the gene expression changes associated with clonal selection (both positive and negative selection).19,20 Collectively, we now have physiological TCR transgenic model systems in place that allow for the examination of the genetics underlying clonal selection. POSITIVE AND NEGATIVE SELECTION: ROLE OF TCR AFFINITY FOR PEPTIDE–MHC COMPLEXES Decoding TCR affinity As presented above, key events during T-cell differentiation are required to generate a repertoire of T cells expressing a useful self- MHC-restricted TCR with a low affinity for spMHC, while eliminat- ing T cells bearing an autoreactive TCR with high affinity for spMHC. The ability of thymocytes to discriminate a low- versus a high-affinity interaction has puzzled immunologists for many decades now. A realistic assumption is that different affinities of TCR–spMHC inter- actions will lead either to the activation of different TCR signalling pathways or to different intensities of activation of the TCR signalling pathway. These non-mutually exclusive possibilities have been tested by many different groups over the years. A first hint suggesting that TCR signalling pathways differ in the context of positive and negative selection was obtained by comparing the TCR signalling cascades in response to either positively or negatively selecting spMHC ligands. Both types of ligands activated the three classical mitogen-activated protein kinase (MAPK) pathways: extracellular signal-regulated kinase (ERK), c-jun NH2-terminal kinase (JNK) and p38.21 However, the kinetics of ERK activation, but not that of p38 or JNK, was different between positive and negative selection.21 Indeed, Werlen et al. have nicely shown that ligands inducing positive selection induce a slow but sustained activation of ERK, while negatively selecting ligands activate ERK strongly, but only transiently (Figure 2). These findings were later corroborated using foetal thymic organ cultures.22 The key role of ERK activation during positive selection was also highlighted using genetic manipulations. For one, a mutation of the TCRa connecting peptide domain selectively affects positive selection and ERK activa- tion.21 In addition, CD3d chain-deficiency blocks thymocyte positive selection and is necessary to properly couple TCR signalling to ERK activation.23 Finally, as expected from these experiments, T-cell abla- tion of ERK1/2 confirmed the selective role of this MAPK pathway in positive selection.24,25 These observations then opened the question as to how ERK activation is differentially regulated in thymocytes following different affinity signals. Examination of Ras activation, upstream of ERK, provided a possible explanation. Indeed, Ras activation can occur via two families of Ras guanine exchange factors, RasGRP or Grb2-SOS. Interestingly, it was shown that RasGRP is necessary for positive, but not negative selection,26,27 while Grb2-SOS haploinsufficiency selectively influences negative selection28 (Figure 2). A very elegant study by Daniels et al. in 200629 further refined the concept that differences in affinity lead to opposing cell fate decisions (thymocytes survival versus death) by modulating the strength and kinetics of ERK activation. They have shown that ligands of increasing affinity within the positive selection threshold led to incremental increases in phospho-p23-z and phospho-ZAP-70 while higher-affi- nity ligands that led to negative selection induce more phospho-ZAP- 70 at the membrane. A high concentration of phospho-ZAP-70 correlates with the activation and translocation of RasGRP1 from the cytoplasm to the membrane and recruitment of Grb2-SOS to phospho-LAT. This results in the transient and strong activation of the Ras/Raf-1/ERK signalling cascade at the membrane. On the other hand, ligands inducing positive selection are unable to recruit Grb2- SOS at the cytoplasmic membrane; rather they recruit RasGRP1 to the Golgi, where a slow and sustained activation of ERK ensues.29 Therefore, the intracellular site of ERK activation allows thymocytes to distinguish positively and negatively selecting ligands. The conclusion of Daniels experiments were further confirmed using computer simulations.30 Very interestingly, their simulation helps to explain the sharp boundary separating ligands that mediate positive and negative selection. Indeed, there is no feedback regulation of RasGRP activity and, therefore, TCR signalling increases in a linear manner with ligand potency. At the sharp boundary defining negative selec- tors, the activation of Ras by SOS leads to a large increase in Ras activity because of positive feedback regulation of this pathway.30 Their model is supported by previously described results showing that deletion of RasGRP only influences positive selection,26,27 while Grb2 haploinsufficiency affects negative selection.28 However, more recent experiments using Grb2�/� thymocytes indicate that Grb2 is necessary for both positive and negative selection.31 This discrepancy might be explained by weaker tyrosine phosphorylation of multiple proteins controlling TCR signalling (lck, ZAP-70, TCR-z, LAT, PLC-g-1) and by a reduction of JNK and p38 activation in the absence of Grb2,31 which will certainly interfere with positive selection. Future T-cell clonal selection N Labrecque et al 18 Immunology and Cell Biology experiments are needed to understand how Grb2 can affect proximal TCR signalling. Alternatively, the possible interaction between Grb2 and Thymus-expressed molecule involved in selection (Themis) may also explain the defect in positive selection in the absence of Grb232–34 (see below for more details on Themis). RasGRP and Grb2-SOS thus appear central in defining the fate of the thymocyte. Moving upstream in the signalling cascade, we find