Review Article | Published:

Cell death and immunity

Negative selection — clearing out the bad apples from the T-cell repertoire

Nature Reviews Immunology volume 3, pages 383391 (2003) | Download Citation



Dead cells are a prominent feature of the thymic landscape as only 5% of developing thymocytes are exported as mature T cells. The remaining thymocytes die by one of two mechanisms; most thymocytes die because they are not positively selected and do not receive a survival signal, whereas a minority of thymocytes undergo T-cell receptor (TCR)-mediated apoptosis, a process known as negative selection. Negative selection is extremely important for establishing a functional immune system, as it provides an efficient mechanism for ridding the T-cell repertoire of self-reactive and potentially autoimmune lymphocytes. This review discusses several cellular and molecular aspects of negative selection.

Key points

  • CD4+CD8+ double-positive (DP) thymocytes undergo one of three fates in the thymus: positive selection, negative selection or death by neglect. Negative selection of thymocytes that express T-cell receptors (TCRs) with high affinity for self-peptide–MHC (5% of the total) deletes potentially self-reactive thymocytes, generating a largely self-tolerant peripheral T-cell repertoire.

  • Most negative selection is thought to occur in the thymic medulla as this contains two types of specialized antigen-presenting cell — dendritic cells and thymic epithelial cells (TECs). Medullary TECs transcribe genes that are normally expressed in peripheral tissues.

  • Negative selection can occur before or after positive selection and in thymocytes at all stages of development. So, positive and negative selection are probably independent, and not sequential, events.

  • Negative selection in response to high-affinity ligands might be due to increased TCR occupancy or a slower 'off-rate' (kinetic proofreading).

  • A second co-stimulatory signal, in addition to the TCR signal, might be required for negative selection. However, there are discrepancies between blocking experiments and genetically deficient mice in this regard.

  • The kinetics of mitogen-activated protein kinase (MAPK) signalling are thought to determine positive- versus negative-selection signals. Extracellular signal-regulated kinase (ERK) is induced more rapidly during negative selection, which might determine the particular transcription factors (such as NUR77 and NF-κB) that are triggered.

  • Linker for activation of T cells (LAT) and phosphatase and tensin homologue (PTEN) are also thought to be involved in negative selection.

  • Death-domain-containing proteins, such as CD95 (FAS) and tumour-necrosis factor receptors (TNFRs), that are involved in the apoptosis of peripheral T cells are not thought to be important for negative selection. Rather, negative selection might occur independently of death domain signalling.

Access optionsAccess options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.


  1. 1.

    Role of thymic organ structure and stromal composition in steady-state postnatal T-cell production. Immunol. Rev. 189, 8–20 (2002). A good review of thymic anatomy and the dynamics of T-cell production.

  2. 2.

    , , & Crucial role of the pre-T-cell receptor α gene in development of αβ but not γδ T cells. Nature 375, 795–798 (1995).

  3. 3.

    , & Thymocyte development in the absence of pre-T cell receptor extracellular immunoglobulin domains. Science 280, 905–908 (1998).

  4. 4.

    , & Positive and negative T cell selection. Annu. Rev. Immunol. 21, 139–176 (2003). This is a clear and comprehensive review of positive and negative selection.

  5. 5.

    & The thymus and central tolerance. Philos. Trans. R. Soc. Lond. B.Biol. Sci. 356, 609–616 (2001). This review of negative selection places many controversial issues into perspective.

  6. 6.

    et al. Immunologic tolerance maintained by CD25+ CD4+ regulatory T cells: their common role in controlling autoimmunity, tumor immunity, and transplantation tolerance. Immunol. Rev. 182, 18–32 (2001).

  7. 7.

    et al. Quantitative impact of thymic clonal deletion on the T cell repertoire. J. Exp. Med. 185, 377–383 (1997).

  8. 8.

    , & Using thymus anatomy to dissect T cell repertoire selection. Semin. Immunol. 11, 65–70 (1999). References 7 and 8 describe negative-selection-deficient mice and estimate the fraction of thymocytes that undergo negative selection.

  9. 9.

    , , , & The use of mammary tumor virus (Mtv)-negative and single-Mtv mice to evaluate the effects of endogenous viral superantigens on the T cell repertoire. J. Exp. Med. 182, 1493–1504 (1995).

  10. 10.

    Developmental biology of T cells in T cell-receptor transgenic mice. Annu. Rev. Immunol. 8, 531–556 (1990).

  11. 11.

    & T-cell tolerance. Curr. Opin. Immunol. 5, 873–879 (1993).

  12. 12.

    , , , & Unopposed positive selection and autoreactivity in mice expressing class II MHC only on thymic cortex. Nature 383, 81–85 (1996).

  13. 13.

    , , , & B7 expression on thymic medullary epithelium correlates with epithelium-mediated deletion of Vβ5+ thymocytes. J. Immunol. 152, 3254–3263 (1994).

  14. 14.

    & T-cell apoptosis detected in situ during positive and negative selection in the thymus. Nature 372, 100–103 (1994). This paper shows that most thymocytes that undergo apoptosis are located in the medulla.

  15. 15.

    Central tolerance of T cells. Int. Rev. Immunol. 13, 95–105 (1995).

  16. 16.

    & Does T-cell tolerance require a dedicated antigen-presenting cell? Nature 338, 74–76 (1989).

  17. 17.

    & Tolerogenicity of thymic epithelium. Eur. J. Immunol. 20, 2525–2528 (1990).

  18. 18.

    et al. Clonal deletion of Vβ5+ T cells by transgenic I-E restricted to thymic medullary epithelium. J. Immunol. 151, 3954–3960 (1993).

  19. 19.

    , , & CD4 T cell tolerance to human C-reactive protein, an inducible serum protein, is mediated by medullary thymic epithelium. J. Exp. Med. 188, 5–16 (1998).

  20. 20.

    , , & Intrathymic expression of genes involved in organ specific autoimmune disease. J. Autoimmun. 11, 309–318 (1998).

  21. 21.

    , & Aberrant expression of tissue-specific proteins in the thymus: a hypothesis for the development of central tolerance. Scand. J. Immunol. 47, 95–100 (1998).

  22. 22.

    , , , & Shaping of the autoreactive T-cell repertoire by a splice variant of self protein expressed in thymic epithelial cells. Nature Med. 6, 56–61 (2000).

  23. 23.

    , , & Promiscuous gene expression in medullary thymic epithelial cells mirrors the peripheral self. Nature Immunol. 2, 1032–1039 (2001).

  24. 24.

    & Self-antigen presentation by thymic stromal cells: a subtle division of labor. Curr. Opin. Immunol. 12, 179–186 (2000). References 19–24 describe the promiscuous expression of genes in the thymus. References 19 and 23 show that this promiscuous gene expression originates in thymic epithelial cells.

  25. 25.

    , & Self-reactive T cells selected on thymic cortical epithelium are polyclonal and are pathogenic in vivo. J. Immunol. 162, 5078–5084 (1999).

  26. 26.

    & In vivo T-lymphocyte tolerance in the absence of thymic clonal deletion mediated by hematopoietic cells. Blood 93, 3856–3862 (1999).

  27. 27.

    et al. APECED: a monogenic autoimmune disease providing new clues to self-tolerance. Immunol. Today 19, 384–386 (1998).

  28. 28.

    et al. Projection of an immunological self shadow within the thymus by the aire protein. Science 298, 1395–1401 (2002). This paper decribes autoimmune regulator (Aire)-deficient mice, which lack promiscuous gene expression in the thymus and consequently develop autoimmune polyendocrine syndrome.

  29. 29.

    , & T cell tolerance by clonal elimination in the thymus. Cell 49, 273–280 (1987). This was the first paper to directly show the deletion of self-reactive thymocytes.

  30. 30.

    et al. The Vβ-specific superantigen staphylococcal enterotoxin B: stimulation of mature T cells and clonal deletion in neonatal mice. Cell 56, 27–35 (1989).

  31. 31.

    & Negative selection in the thymus includes semimature T cells. J. Exp. Med. 185, 263–271 (1997).

  32. 32.

    et al. Cutting edge. Identification of the targets of clonal deletion in an unmanipulated thymus. J. Immunol. 170, 10–13 (2003).

  33. 33.

    , , , & Distinct sequence of negative or positive selection implied by thymocyte T-cell receptor densities. Nature 346, 861–863 (1990).

  34. 34.

    & Developmental regulation of thymocyte susceptibility to deletion by 'self'-peptide. J. Exp. Med. 176, 213–223 (1992).

  35. 35.

    , , & Negative selection of T cells occurs throughout thymic development. J. Immunol. 163, 689–698 (1999). References 33–35 show that positive and negative selection are not sequential events.

  36. 36.

    et al. Evidence for a differential avidity model of T cell selection in the thymus. Cell 76, 651–663 (1994).

  37. 37.

    et al. Positive and negative thymocyte selection induced by different concentrations of a single peptide. Science 263, 1615–1618 (1994). References 36 and 37 show that low doses of agonist peptide can mediate positive selection in fetal thymic organ cultures.

  38. 38.

    , & Strong agonist ligands for the T cell receptor do not mediate positive selection of functional CD8+ T cells. Immunity 3, 79–86 (1995). This paper shows that antagonist peptides mediate positive selection, whereas agonist peptides mediate negative selection. Low doses of agonist peptide did not result in positive selection in this study.

  39. 39.

    et al. T-cell-receptor affinity and thymocyte positive selection. Nature 381, 616–620 (1996). Reference 39 used surface plasmon resonance to show formally that positive-selecting peptide–MHC ligands have a lower affinity for the T-cell receptor (TCR) than do negative-selecting peptide–MHC ligands.

  40. 40.

    Kinetic proofreading in T-cell receptor signal transduction. Proc. Natl Acad. Sci. USA 92, 5042–5046 (1995).

  41. 41.

    , , & Two signals are required for negative selection of CD4+CD8+ thymocytes. J. Immunol. 151, 1868–1880 (1993).

  42. 42.

    , , , & Negative selection of CD4+CD8+ thymocytes by T cell receptor-induced apoptosis requires a costimulatory signal that can be provided by CD28. J. Exp. Med. 179, 709–713 (1994).

  43. 43.

    & Several different cell surface molecules control negative selection of medullary thymocytes. J. Exp. Med. 190, 65–73 (1999).

  44. 44.

    Cutting edge. Thymic selection and autoreactivity are regulated by multiple coreceptors involved in T cell activation. J. Immunol. 163, 3577–3581 (1999).

  45. 45.

    et al. Perinatal blockade of B7-1 and B7-2 inhibits clonal deletion of highly pathogenic autoreactive T cells. J. Exp. Med. 195, 959–971 (2002). This paper clearly shows that perinatal blockade of CD80 and CD86 inhibits negative selection and allows the development of autoreactive T cells. These results are in contrast to those in reference 48, which examines negative selection in CD80 and CD86 double-deficient mice.

  46. 46.

    , , , & CD28 expression is not essential for positive and negative selection of thymocytes or peripheral T cell tolerance. J. Immunol. 156, 1006–1013 (1996).

  47. 47.

    & Requirement for a complex array of costimulators in the negative selection of autoreactive thymocytes in vivo. J. Immunol. 166, 6050–6056 (2001).

  48. 48.

    , , & CD40 ligand functions non-cell autonomously to promote deletion of self-reactive thymocytes. J. Immunol. 168, 2759–2765 (2002). Reference 48 shows that CD40L does not have to be expressed by every thymocyte to induce efficient negative selection; that is, CD40L acts in a non-cell-autonomous manner. This paper also shows that deletion of superantigen-reactive thymocytes occurs in CD80 and CD86 double-deficient animals.

  49. 49.

    et al. An essential role for gp39, the ligand for CD40, in thymic selection. J. Exp. Med. 182, 1377–1388 (1995).

  50. 50.

    , & T cell receptor β chain lacking the large solvent-exposed Cβ FG loop supports normal α/β T cell development and function in transgenic mice. J. Exp. Med. 189, 1679–1684 (1999).

  51. 51.

    et al. Involvement of the TCR Cβ FG loop in thymic selection and T cell function. J. Exp. Med. 195, 1419–1431 (2002).

  52. 52.

    et al. Imaging synapse formation during thymocyte selection: inability of CD3ζ to form a stable central accumulation during negative selection. Immunity 16, 595–606 (2002).

  53. 53.

    , , , & Immature CD4+CD8+ thymocytes form a multifocal immunological synapse with sustained tyrosine phosphorylation. Immunity 16, 839–848 (2002). References 52 and 53 describe the immunological synapse formed by double-positive (DP) thymocytes.

  54. 54.

    et al. CD8 is needed for positive selection but differentially required for negative selection of T cells during thymic ontogeny. Eur. J. Immunol. 23, 212–216 (1993). This paper shows that negative selection is not completely dependent on co-receptor expression.

  55. 55.

    , & A role for the α-chain connecting peptide motif in mediating TCR–CD8 cooperation. J. Immunol. 169, 2964–2970 (2002).

  56. 56.

    et al. CD3δ establishes a functional link between the T cell receptor and CD8. J. Biol. Chem. 278, 3257–3264 (2003).

  57. 57.

    , , & Positive selection through a motif in the αβ T cell receptor. Science 281, 835–838 (1998).

  58. 58.

    , & A motif in the αβ T-cell receptor controls positive selection by modulating ERK activity. Nature 406, 422–426 (2000).

  59. 59.

    , , & Signaling life and death in the thymus: timing is everything. Science 299, 1859–1863 (2003). This is a recent review on TCR signalling in thymocytes.

  60. 60.

    et al. Duration and strength of extracellular signal-regulated kinase signals are altered during positive versus negative thymocyte selection. J. Immunol. 167, 4966–4973 (2001).

  61. 61.

    et al. The JNK pathway regulates the in vivo deletion of immature CD4+CD8+ thymocytes. J. Exp. Med. 188, 1817–1830 (1998).

  62. 62.

    et al. c-Jun NH2-terminal kinase (JNK)1 and JNK2 have similar and stage-dependent roles in regulating T cell apoptosis and proliferation. J. Exp. Med. 193, 317–328 (2001).

  63. 63.

    , , & Differential roles of ERK and p38 MAP kinase pathways in positive and negative selection of T lymphocytes. Immunity 9, 565–574 (1998).

  64. 64.

    et al. Induction of T helper type 2 immunity by a point mutation in the LAT adaptor. Science 296, 2036–2040 (2002).

  65. 65.

    et al. A LAT mutation that inhibits T cell development yet induces lymphoproliferation. Science 296, 2040–2043 (2002). References 64 and 65 describe an autoimmune syndrome that is generated by a point mutation in the linker for activation of T cells (LAT) adaptor.

  66. 66.

    et al. Association of Grb2, Gads, and phospholipase C-γ 1 with phosphorylated LAT tyrosine residues. Effect of LAT tyrosine mutations on T cell antigen receptor-mediated signaling. J. Biol. Chem. 275, 23355–23361 (2000).

  67. 67.

    & Identification of the minimal tyrosine residues required for linker for activation of T cell function. J. Biol. Chem. 276, 29588–29595 (2001).

  68. 68.

    et al. Disruption of T cell signaling networks and development by Grb2 haploid insufficiency. Nature Immunol. 2, 29–36 (2001). This paper shows that mice carrying one functional allele of growth-factor receptor-bound protein 2 (Grb2) are defective in negative but not positive selection.

  69. 69.

    et al. T cell-specific loss of Pten leads to defects in central and peripheral tolerance. Immunity 14, 523–534 (2001).

  70. 70.

    , & A gene inducible by serum growth factors encodes a member of the steroid and thyroid hormone receptor superfamily. Proc. Natl Acad. Sci. USA 85, 8444–8448 (1988).

  71. 71.

    Nerve growth factor induces a gene homologous to the glucocorticoid receptor gene. Neuron 1, 183–188 (1988).

  72. 72.

    , , , & Apoptotic signals delivered through the T-cell receptor of a T-cell hybrid require the immediate-early gene nur77. Nature 367, 281–284 (1994).

  73. 73.

    et al. Regulation of the Nur77 orphan steroid receptor in activation-induced apoptosis. Mol. Cell. Biol. 15, 6364–6376 (1995).

  74. 74.

    , , , & A role for the orphan steroid receptor Nur77 in apoptosis accompanying antigen-induced negative selection. Immunity 3, 273–282 (1995).

  75. 75.

    et al. Unimpaired thymic and peripheral T cell death in mice lacking the nuclear receptor NGFI-B (Nur77). Science 269, 532–535 (1995).

  76. 76.

    , , & Functional redundancy of the Nur77 and Nor-1 orphan steroid receptors in T-cell apoptosis. EMBO J. 16, 1865–1875 (1997). Reference 74 shows that a dominant-negative version of NUR77 impairs negative selection, whereas reference 75 shows that negative selection is unimpaired in Nur77-deficient mice. This discrepancy is explained in reference 76, which shows that NOR1 and NUR77 are structurally and functionally redundant.

  77. 77.

    , , & Apoptosis of T cells mediated by Ca2+-induced release of the transcription factor MEF2. Science 286, 790–793 (1999).

  78. 78.

    et al. The lack of NF-κB transactivation and PKCε expression in CD4+CD8+ thymocytes correlates with negative selection. Cell Death Differ. 7, 1253–1262 (2000).

  79. 79.

    et al. Peptide-induced negative selection of thymocytes activates transcription of an NF-κB inhibitor. Mol. Cell 9, 637–648 (2002).

  80. 80.

    , & Deletion of potentially self-reactive T cell receptor specificities in L3T4-, Lyt-2- T cells of lpr mice. J. Exp. Med. 168, 2221–2229 (1988).

  81. 81.

    , & A role for Fas in negative selection of thymocytes in vivo. J. Exp. Med. 187, 1427–1438 (1998).

  82. 82.

    & The thymus and negative selection. Immunol. Rev. 185, 126–135 (2002).

  83. 83.

    , , , & A dominant interfering mutant of FADD/MORT1 enhances deletion of autoreactive thymocytes and inhibits proliferation of mature T lymphocytes. EMBO J. 17, 706–718 (1998). This paper shows that negative selection is intact in mice that express a dominant interfering mutant of FAS-associated death domain (FADD). This makes it unlikely that FAS (CD95), tumour-necrosis factor receptor 1 (TNFR1), DR3 or tumour-necrosis factor-related apoptosis-inducing ligand receptor 2 (TRAILR2) contribute to negative selection through their death domains.

  84. 84.

    , , , & Defective thymocyte apoptosis and accelerated autoimmune diseases in TRAIL−/− mice. Nature Immunol. 4, 255–260 (2003). Reference 84 shows that Trail−/− mice are defective in negative selection. These results could be contrary to those in reference 83.

  85. 85.

    The suicide in the thymus, a twisted trail. Nature Immunol. 4, 207–208 (2003). This review puts forward a hypothesis to reconcile the results of references 83 and 84. TRAIL receptors might support thymocyte apoptosis by activating JUN N-terminal kinase (JNK) without using their death domain.

  86. 86.

    et al. The apoptotic protease-activating factor 1-mediated pathway of apoptosis is dispensable for negative selection of thymocytes. J. Immunol. 168, 2288–2295 (2002).

  87. 87.

    et al. Blocked negative selection of developing T cells in mice expressing the baculovirus p35 caspase inhibitor. EMBO J. 18, 156–166 (1999).

  88. 88.

    , & Caspase enzyme activity is not essential for apoptosis during thymocyte development. J. Immunol. 164, 4071–4079 (2000).

  89. 89.

    et al. T-cell receptor ligation by peptide–MHC induces activation of a caspase in immature thymocytes: the molecular basis of negative selection. EMBO J. 16, 2282–2293 (1997).

  90. 90.

    , , , & Bcl-2 inhibits multiple forms of apoptosis but not negative selection in thymocytes. Cell 67, 879–888 (1991).

  91. 91.

    , , & Positive and negative selection of T cells in T-cell receptor transgenic mice expressing a Bcl-2 transgene. Proc. Natl Acad. Sci. USA 91, 1376–1380 (1994).

  92. 92.

    et al. BH3-only Bcl-2 family member Bim is required for apoptosis of autoreactive thymocytes. Nature 415, 922–926 (2002). Reference 92 establishes the importance of BCL-2-interacting mediator of cell death (BIM) in negative selection.

  93. 93.

    et al. Activated T cell death in vivo mediated by proapoptotic Bcl-2 family member Bim. Immunity 16, 759–767 (2002). This paper shows that BIM deficiency compromises the death of activated peripheral T cells.

  94. 94.

    , , , & BH3-only proteins that bind pro-survival Bcl-2 family members fail to induce apoptosis in the absence of Bax and Bak. Genes Dev. 15, 1481–1486 (2001).

  95. 95.

    , , , & Activation of mitochondrial voltage-dependent anion channel by a pro-apoptotic BH3–only protein Bim. Oncogene 21, 4944–4956 (2002).

Download references


I am indebted to the efforts of many laboratories that have contributed to this field and apologize to those whose work was not cited. Thanks to J. Sprent and G. Werlen for interesting discussions, and D. Gil-Pagès, B. Hausmann, D. Naeher, A. Schrum, G. Werlen and E. Teixeiro-Pernas for reading the manuscript. My laboratory is supported by grants from the Swiss National Science Foundation and Hoffmann La Roche, Ltd., and by a generous gift from Novartis, AG.

Author information


  1. Laboratory of Transplantation Immunology and Nephrology, University Hospital Basel, Hebelstrasse 20, CH-4031 Basel, Switzerland.

    • Ed Palmer


  1. Search for Ed Palmer in:



Animals are lethally irradiated to destroy most of their haematopoietic-lineage cells then given bone-marrow cells from a donor animal to reconstitute their haematopoietic cells. In the context of thymic development, positive selection is mediated by radiation-resistant cells in the cortex of the host's thymus. Negative selection is mediated by residual epithelial cells from the host's thymus and, by antigen-presenting cells that are derived from the transferred bone marrow.


These are type II cell-surface proteins, which are encoded by mouse mammary tumour viruses (MMTVs) that have integrated into the mouse germline. These viral proteins are presented by MHC class II molecules and bind to the T-cell receptor through its Vβ domain. Because these endogenous proteins are self-antigens and are expressed by antigen-presenting cells in the thymus, they delete large numbers of thymocytes that express a particular Vβ domain.


Recombinant MHC molecules can be engineered to contain an antigenic peptide in their peptide-binding groove. These peptide–MHC monomers are biotinylated and then used to form tetrameric complexes by binding to fluorescently labelled streptavidin (streptavidin has four biotin-binding sites). Tetramers can be used to stain T cells expressing T-cell receptors that are specific for that peptide–MHC complex.


The detection of alterations in plasmon waves generated at a metal–liquid interface. Changes in surface plasmon resonance are a function of the mass of molecules bound to the interface, so this technique allows sensitive detection of ligand binding in real time without requiring the chemical modification of ligands to enable their detection.


(FTOC). Removal of day-16 fetal thymi allows the analysis of antigen-driven positive- and negative-selection events during in vitro culture.


A technique that is used to enrich messenger RNA (complementary DNA) sequences that are specifically expressed by a particular cell line or under particular physiological conditions.


Family of cytosolic proteases that contain a cysteine residue in the active site and that cleave their substrate after an aspartic-acid residue. They can be divided into inflammatory caspases (1, 4, 5 and 11), which cleave and activate pro-inflammatory cytokines, and pro-apoptotic caspases, which cleave and activate pro-apoptotic substrates. Pro-apoptotic caspases consist of initiator caspases (2, 8 and 9), which, in turn, cleave and activate effector caspases (3, 6 and 7).


Mice transgenic for an αβ T-cell receptor specific for a peptide derived from the male antigen HY.


An apoptotic protein complex formed from the union of apoptotic protease activating factor 1 (APAF1), cytochrome c and dATP with pro-caspase-9. Complex formation leads to the cleavage and activation of caspase-9, which activates caspase-3 and other effector caspases, leading to cell death.

About this article

Publication history



Further reading