© 2001 Nature Publishing Group http://immunol.nature.com

R EVIEW

Intraepithelial lymphocytes: exploring the Third Way in immunology © 2001 Nature Publishing Group http://immunol.nature.com

Adrian Hayday1,2,*, Efstathios Theodoridis1, Elizabeth Ramsburg3 and John Shires1,2 Locally resident intraepithelial lymphocytes (IELs) are primarily T cells with potent cytolytic and immunoregulatory capacities, which they use to sustain epithelial integrity. Here, we consider that most IEL compartments comprise a variable mixture of two cell types: T cells primed to conventional antigen in the systemic compartment and T cells with ill-defined reactivities and origins, whose properties seem to place them midway between the adaptive and innate immune responses.We review the capacity of IELs to limit the dissemination of infectious pathogens and malignant cells and to control the infiltration of epithelial surfaces by systemic cells. An improved characterization of IELs would seem essential if we are to understand how immune responses and immunopathologies develop at body surfaces. Most immunologists have much in common with conservative (with a small “c”) politicians. Both focus on the pre-eminence of centralized decision-making in coordinating responses to a spectrum of challenges. As described by immunologists, the crucible of centralized control is the network of lymph nodes, where circulating T cells receive information from “the regions” conveyed by migrating antigen-presenting cells (APCs). A coordinated specific response is then developed and dispatched back to the regions. The form of the response varies according to the nature of the information received from the APCs, but it will invariably involve infiltration into the tissues of systemic lymphocytes and various accessory and effector cells. Although powerful, this centralized control paradigm conspicuously fails to mention intraepithelial lymphocytes (IELs) that are often already resident in the local tissues. The political form that centralized control should take has been the subject of doctrinaire debate for centuries. Recently, several mainstream politicians successfully shifted away from this debate, focusing instead on the “Third Way”, a philosophy that—among other things— places renewed emphasis on the power of local mechanisms to deal with local issues1. It seems appropriate for mainstream immunologists to do likewise and to reconsider the role played by local immune mechanisms in immune responses at body surfaces. Although there are many signatory features of the body surface immune system (see the Review by Neutra in this issue of Nature Immunology), emerging data on IELs

might be an appropriate departure point for an exploration of the Third Way in immunology. As with the Third Way in politics, exploring the Third Way in immunology is a voyage of rediscovery. Small round cells in the small intestinal epithelium were clearly described by Weber in 18472, and since then they have been extensively studied and reviewed3–7. Lymphocytes within the gut epithelium were noted to be an evolutionarily conserved trait by Fichtelius8, who also emphasized the primitive nature of antigen exposure at the gut and the skin. Soon afterwards intestinal IELs were identified as being T cells9–11, with a primarily cytolytic (CTL) phenotype12–17, and now IELs have been described in numerous tissues in a broad phylogenetic range of animals. Nonetheless, the critical benefit of IELs to their hosts has remained uncertain. Today, growing evidence portrays IELs as potent, rapidly activated cytolytic and immunoregulatory effectors that can protect their host tissues from infection, cell transformation and uncontrolled infiltration by systemic cells.

IELs are distinct from systemic T cells There are on average about 10–20 IELs per 100 villus enterocytes in the small bowel in humans3. Given the immense surface area of epithelia, resident IELs can comprise a substantial fraction of the body’s T cells. Nonetheless, the variable phylogenetic conservation of IEL compartments may have detracted from their importance. Thus, although rodent skin harbors an extensive network of IELs—known as dendritic epidermal T cells (DETCs)—T cells are less obvious in human skin and are enriched primarily in the dermis, as opposed to the epidermis18. A primary distinction of IELs from their systemic counterparts is their subset composition. Whereas spleen, peripheral blood and lymph node T cells readily subdivide into major histocompatibility complex (MHC) class II–restricted CD4+ αβ T cells and MHC class I–restricted CD8αβ+ αβ T cells, >70% of small intestinal IELs comprise CD8+ cells. A large fraction of these express a CD8αα homodimer, which is essentially absent from the circulation19. Likewise, CD4–CD8– “double negative” (DN) cells, which are rare in the systemic circulation, can account for >10% of murine small intestinal IELs and the majority of IELs in other compartments20–22. Conversely, CD4+ αβ T cells are under-represented in many IEL compartments and, of the few present in the small intestine, many also express CD8αα23. These T cell antigen receptor–positive (TCR+) “double positive” (DP) cells are also unprecedented in the systemic circulation. IELs include greater numbers of TCRγδ+ cells than are found in the murine or human circulation5,22,24–26. Between 35–65% of murine CD8+ small intestinal IELs and essentially all murine DETCs and vaginal IELs express TCRγδ with TCR repertoires specific to particular sites and distinct from those in the blood26. Further heterogeneity is evident among small intestinal TCRγδ+ cells. Thus, TCRγδlo cells are constitutively responsive to interleukin 2 (IL-2) and IL-7, whereas TCRγδhi cells are

1

Peter Gorer Department of Immunobiology, GKT School of Medicine, University of London,Third floor New Guy’s House, Guy’s Hospital, London SE1 9RT, UK. 2 Department of Molecular Cell and Developmental Biology and 3Section of Immunobiology,Yale University, New Haven, CT 06520, USA. Correspondence should be addressed to A. H. ([email protected]). http://immunol.nature.com

•

november 2001

•

volume 2 no 11

•

nature immunology

997

© 2001 Nature Publishing Group http://immunol.nature.com

R EVIEW

of Fas32 and also secrete T helper type 1 (TH1) cytokines33–36. In comparison to polyclonal small circulating T cells, the Properties Type a Type b References type a IEL repertoire comprises only ∼250 + + + + TCRαβ CD8αβ TCRαβ CD8αα clones37,38. Their composition varies in difTCRγδ+CD8αα+ ferent mice, which probably reflects stochastic responses to antigen challenges37,38. TCRγδ+DN Numerous TCR gene rearrangements Gene expression CD2 ++ – 86 in type a IELs are shared with gut lamina CD5 ++ +/– 23, 86 propria T cells and with thoracic duct CD28 ++ – 86, 116 CTLA4 ++ – 86 CD8+ T cell blasts38. This supports the Ly6C ++ +/– 86 hypothesis that type a IELs are primed to Tactile +/– ++ 86 antigen in gut-associated Peyer’s Patches FcεRIγ – ++ 122 (see the Review by Neutra in this issue of OCIL +/– ++ 86 Ly49E – ++ 86 Nature Immunology) or in extrafollicular TCRβ chain gene usage overlaps areas of the gut39, after which they drain with lamina propria and thoracic duct ++ – 38 via mesenteric lymph nodes and the lymConventional MHC restriction ++ – 31,61,83 phatics to the thoracic duct10.38. After –/– Representation in: TAP mice – +/– 31,61 entering the blood, they home back to the athymic mice – + 16,28,60,74 P-glycoprotein–/– mice Increased Reduced 123 lamina propria throughout the small and c-kit–/– mice ++ +/– 124 the large intestine. From the lamina proReconstitution from peripheral lymph nodes ++ +/– 42,44,45 pria they enter the epithelium with variImmunological memory of infection ++ – 47–50,70 able efficiency and may continue to exchange across the two compartments3,6,40. Past electron microscope not6. Other markers of IEL heterogeneity include morphology, size and images depicted IELs straddling the lamina propria and epithelium41, sedimentation density, with a substantial fraction of IELs appearing as and cell-labeling experiments suggested that ∼3 IELs per hour passed large granular lymphocytes, quite unlike systemic T cells3,11,27,28. back into the lamina propria40. Within the epithelium, particular type a Likewise, DETCs are, by definition, highly dendritic. Nonetheless, not clones will accumulate in response to repeated priming to an antigen all IEL compartments show such distinctive features. Thus, murine large that is expressed on enterocytes11,38,42. Consistent with this, type a IELs intestinal IELs are primarily αβ T cells expressing CD4 or CD8αβ6,29 are the predominant IEL subtype in the murine large bowel, which harbors the greatest microbial antigen load, and are heavily depleted in similar to those in the sytemic circulation. germ-free mice6,43.

© 2001 Nature Publishing Group http://immunol.nature.com

Table 1. Proposed simplified classification of IELs

IEL heterogeneity: a revised perspective To those unfamiliar with IELs, the heterogeneity within, and the variability between, IEL compartments may seem confusing. Improving methodologies will one day clarify the status and function of each cell type. However, until such time, we propose that matters might be simplified if IELs were classified into just two cell types: “a” and “b”. Type a includes TCRαβ+ cells that primarily recognize antigens presented by conventional MHC class I and II and are primed within the systemic circulation. Among the type b cells we include TCRαβ+CD8αα+ IELs and TCRγδ+ IELs that respond to antigens not restricted by conventional MHC. Although TCRαβ+CD8αα+ and TCRγδ+ cells are clearly different from one from another, type b cells share many “unconventional” features that distinguish them from type a cells (Table 1). Indeed, a recent genomics analysis identified the gene expression profiles of TCRαβ+CD8αα+ IELs and TCRγδ+ IELs as very similar and readily distinguishable from those of type a IELs (Tables 1 and 2). Interestingly, the distribution of type a and b cells varies greatly. Thus, murine skin and vagina are ∼100% type b; murine and human small intestine are ∼50% type a and ∼50% b; and the large intestine is ∼100% type a. Type b IELs are also reportedly over-represented in wild mice as opposed to laboratory mice30. Type a IELs of the small intestine are mostly CD8αβ+ αβ T cells, that, like splenic or lymph node CD8+ cells, are absent in MHC class Ia–deficient mice31. They are primarily cytolytic, killing via granzymes or by engagement nature immunology

Type a IELs can be experimentally reconstituted with donor inocula of peripheral lymph node cells, spleen cells or thoracic duct CD8+ blasts11,38,42,44,45. Reconstitution by peripheral lymph node cells requires antigen in the form of bacterial gut-associated flora44,46. Alternatively, ovalbumin (OVA)-specific peripheral lymph node T cells would reconstitute the IEL compartment of mice expressing an OVA transgene specifically in enterocytes42. The reconstituted antigen-specific type a IELs remained in stable juxtaposition with antigen-expressing epithelial cells until the antigen was presented by APCs in a pro-inflammatory context (via infection by a recombinant OVA-encoding virus). At this point, IEL numbers increased substantially, provoking severe pathology that resembled inflammatory bowel disease. The animals then recovered, as the IELs became functionally anergic for cytokine secretion (see later). These experiments clearly exposed the meta-stable coexistence of IELs and enterocytes, and the potential for immunopathology caused by IELs activated after infection.

Table 2. Subdivision of type b IELs

Type a IELs and the systemic circulation

998

Type a IEL responses to antigen

•

Properties Representation in β2M mice Lck–/– and Fyn–/– mice KGF production –/–

volume 2 no 11

•

november 2001

•

TCRαβ+CD8αα+

TCRγδ+

References

– – –

++ + ++

31, 82, 88 125 99

http://immunol.nature.com

© 2001 Nature Publishing Group http://immunol.nature.com

R EVIEW

© 2001 Nature Publishing Group http://immunol.nature.com

Protective functions of type a IELs

provide the diversity of encounters with antigen that is provided in the Antigen-specific type a IELs can transfer protection against lymphoid follicles, the IELs would require a greater probability that cytomegalovirus (CMV), rotavirus and Toxoplasma35,47–49. After intra- their TCR engages antigen. Thus, it was proposed that IELs are autorevenous (i.v.) transfer, type a IELs from Toxoplasma-primed mice pref- active, using invariant or semi-invariant TCRs to recognize self-moleerentially homed to the intestine, then shuttled between the gut and the cules induced by infection or cell transformation, and independent of systemic circulation50. Homing requires the α4β7 integrin, which binds specific foreign epitopes71. Consistent with this, enodogenous superto the mucosal addressin MadCAM, and, to a lesser extent, αEβ7, which antigen-reactive type b IELs are not deleted from the peripheral reperbinds to the epithelial adhesion molecule E-cadherin45,48,50. Homing of toire, as is the case for type a IELs and systemic T cells59. In addition, IELs is also guided by their expression of chemokine receptors CCR9 type b IELs expressing a transgenic TCR were positively selected, and CXCR3—which respond to mucosae-associated epithelial rather than negatively selected, by constitutive expression of the cogchemokine (MEC) and interferon (IFN)-inducible protein 10 (IP-10, nate antigen in the periphery72. also known as CXCL10) and monokine induced by γ IFN (MIG, also known as CXCL9), respectively—and are produced by gut epithelial Type b IEL development cells51–53. Likewise, nonintestinal epithelia recruit T cells with the use of Type b intestinal IELs are commonly detected in athymic mice, and other chemokines and a further adhesion molecule, lymphocyte- several groups have shown that fetal liver or bone marrow progenitors endothelial-epithelial cell adhesion molecule (LEEP-CAM)54. can reconstitute type b IELs in thymectomized recipients28,60,73–75. As a In addition to cells primed by result, intestinal type b cells were antigen entering across body surclassified as thymic-independent faces, type a IELs will include (TI) cells16,59,60,74 that may develop + + CD8αβ TCRαβ cytolytic cells in gut-associated cryptopatches76. that rapidly disseminate to several Indeed, intestinal γδ cell developanatomical sites, including the gut ment can be rescued in IL-7–/– epithelium, after i.v. antigen-primmice by the expression of IL-7 ing55,56. Although the retention of exclusively in the gut77. Nonesuch cells in the epithelium theless, not all type b IEL are TI: decreases with time, many are avian intestinal γδ+ IELs develop maintained in the apposite lamina in the thymus78 and murine propria, with the capacity to reDETCs and genital tract γδ+ IELs enter the epithelium38,55. Compared are primarily generated in the fetal thymus, where they are the to splenic memory cells, gut-assofirst mature T cells to arise79. In ciated T cells can be rapidly activated and may provide immediate addition, although some type b cytolytic and cytokine-mediated intestinal IELs develop extrathyresponses to local reinfection55,56. mically, their numbers are often substantially reduced in athymic Additionally, alloreactive type a IELs may contribute to graft-ver- Figure 1. A hypothesis for the development of an epithelium-associated mice. We therefore believe that sus-host disease (GvHD)—which repertoire of autoreactive type b IELs. These cells would be negatively selected functionally, the type a and b has clinically significant manifes- were they to develop in the thymus at a later timepoint or seed other peripheral sites80. classification may have some advantages over the TD and TI tations in the gut—and during classification. which IEL numbers increase57,58. Because type b IELs are not Given their similarities with conventional T cells, it is not surprising that type a IELs are thymus-depen- constrained by the recognition of polymorphic MHC, they may develop from DN thymic progenitors that escape to the periphery before dent (TD) and were previously classified as “TD-IELs”59,60. MHC-based selection of DP thymocytes. Possibly these DN cells acquire CD8αα in the gut under the influence of factors such as transType b IELs TCRγδ+ and TCRαβ+ type b IEL are present in mice lacking polymor- forming growth factor-β (TGF-β). It has been suggested that IELs phic MHC antigens31,61. They have not been detected circulating via the develop in the fetal or newborn thymus, before the establishment of lymphatics and the thoracic duct38 and, in humans and mice, the efficient central deletion (Fig. 1)80. After emigration from the thymus, TCRγδ+ IEL repertoires in different compartments show little overlap those cells that populate the spleen or peripheral lymph nodes are tolerwith those of peripheral blood, splenic or peripheral lymph node ized by autoantigen but those that invade the intestine are positively selected80. This is consistent both with the early development of type b TCRγδ+ cells26,62,63. Like type a IELs, type b IELs are cytolytic and secrete cytokines IELs and IELs having a higher threshold for activation than splenic T and chemokines16, 32–34,36,64. However, antigen-specific type b IELs are cells81, perhaps due to high local concentrations of TGF-β and with not readily evoked by CMV infection32 and, although type b IELs there being an unorthodox context of antigen presentation in the gut expand in parasite-infected animals, their development is less obvi- epithelium (see later). Once established in the IEL compartment, such ously dependent on either microbial or food antigens3,65–69. Of note, cells show poor responsiveness to stimulation81. Does this mean that there is little evidence for immunological memory in type b IELs70. they are effete cells, with little functional relevance? Drawing on funcBecause type b IELs do not recognize antigen presented by polymor- tional studies, outlined later, we would argue instead that type b IELs phic MHC on professional APCs, it is conceivable that they are primed are potent effector and immunoregulatory cells whose activities are directly by epithelial cells in situ. Because the epithelium cannot tightly held in check. http://immunol.nature.com

•

november 2001

•

volume 2 no 11

•

nature immunology

999

© 2001 Nature Publishing Group http://immunol.nature.com

R EVIEW

although we must again await biochemical evidence for the direct involvement of the TCR.

© 2001 Nature Publishing Group http://immunol.nature.com

Protection of epithelial integrity by type b IELs The unavailability of defined antigen-specific type b IEL clones makes it impossible to assign type b IEL function by conventional adoptive transfers. Likewise, no known mouse mutation exclusively depletes type b IELs. Therefore, much emphasis has been placed on immune responses in the skin of TCRδ–/– mice, as DETCs are ∼100% TCRγδ+. Although such mice lack all γδ+ cells, adoptive transfer of fetal thymocytes to perinatal recipients selectively repopulates DETCs79. Thus, DETCs can down-regulate the capacity of intradermally injected autoreactive CD4+ αβ T cells to cause GvHD in the skin95. The generality of such immunoregulation is suggested by TCRδ–/– mice, which show heightened cutaneous inflammation and ear swelling as a result of αβ T cell–mediated allergic responses (M. Girard, A. C. Hayday and R. E.Tigelaar, unpublished data). This immunopathology was specifically reduced by adoptive transfer of DETC progenitors. TCRδ–/– mice also respond to parasitic infection of the small intestine with a pathologically exaggerated pro-inflammatory αβ T cell response that could be suppressed by adoptive transfer of bulk IELs96. Down-regulation of systemic Figure 2. A two-step model for IEL activation and IEL responses. The IEL is depicted responses inevitably has an appearance of tolerance, and as receiving primary signals (1) through the TCR and secondary signals (2) through accessory some results have suggested that γδ+ IELs contribute to oral receptors (for example, 4-1BB) and cytokine receptors (for example, IL-15Rα).The primary contolerance97,98. Interestingly, almost 25 years ago it was hypothsequences of this activation (denoted I); include the release of heavily expressed cytolytic mediators and chemokines and of the cytokine IL-17B. The secondary consequences (denoted II) esized that IELs might deal with antigen in an anti-inflammainclude the expression of activation-induced genes such as high-affinity cytokine receptors, addi- tory manner, whereas antigens that gained immediate and tional costimulators (such as NKG2d), conventional cytokines and KGF. – and ↓ denote activa- unregulated access to the systemic circulation would be protion-induced down-regulation of some activities. inflammatory3. It is unclear whether the regulatory effects of IELs result directly from interactions with systemic cells or indirectly from IEL Unconventional antigen specificities of type b IELs Neither the identity nor the nature (proteinaceous or nonproteinaceous) interactions with epithelia. γδ+ IELs promote epithelial wound healing of the putative autoantigen(s) for type b IELs is known, although non- via the production of fibroblast growth factors, such as keratinocyte classical class IB MHC antigens have been strongly implicated. growth factors (KGF)99, and may more generally regulate the maturation Although CD8αα+TCRαβ+ IELs require neither class I nor CD1, they and homeostasis of the gut100. Indeed, IELs were assigned a primary role are depleted in β2 microglobulin–deficient (β2M–/–) and Qa-2–/– in regulating enterocyte turnover in monkeys and in guinea pigs6 and mice31,61,82–84. Interestingly, Qa-2–/– mice are susceptible to parasitic express enzymes regulating lipid and cholesterol metabolism101. infections85. However, the capacity of Qa-2 to function as a stressinduced epithelial cell antigen remains to be proven. In addition, the Type b IELs and coeliac disease interpretation is complicated because Qa-2 is heavily expressed by IEL A significant clue to defining the reactivities and functions of type b themselves86 and—although Qa-2 can bind a spectrum of self-pep- IELs must be their dramatic expansion (particularly of γδ+ cells) in tides—CD8αα+TCRαβ+ IELs develop, albeit to a lesser extent, in mice coeliac disease, a wheat protein allergy that provokes CD4+ αβ T deficient in the peptide transporter TAP31,61. cell–mediated erosion of the small intestinal villi3,102. Clearly, coeliac MHC class IB recognition is also implicated by the finding that disease is characterized by an intense infiltration of activated T cells, ∼0.6% of murine γδ+ intestinal IELs are stained by a peptide-indepen- but the prospect that type b IELs respond to these alone does not obvident tetramerized form of TL, a nonclassical MHC molecule expressed ously explain why type b IEL numbers are not amplified in other gut widely on the gut epithelium and on activated lymphoid cells87. immunopathologies. The gliadin peptides that provoke the allergy can However, because TCRγδ+ IELs develop in β2M–/– mice88, any wide- be directly enteropathic, and it is conceivable that type b IELs respond spread recognition of nonclassical MHC would require class IB pro- to a pathognomonic change in the enterocytes. By whatever means, the teins that do not bind to β2M. Increasingly such molecules are being end-result of type b IEL activation may again be to suppress tissue discovered. Two examples are human MICA and MICB, proteins that infiltration by systemic T cells, as peak type b cell representation activate Vγ1Vδ1+ IELs89. However, the full role of the TCR in any of seems to correlate inversely with disease symptoms103. these interactions remains uncertain. First, MICA and MICB, as well as their murine counterparts, can activate a spectrum of cytolytic cells, Type b IEL protection from infection and malignancy including γδ+ IELs, by binding to NKG2d90–92. Likewise, an additional DETCs respond to heat-shocked autologous keratinocytes by secreting or alternative receptor for TL is suggested by the fact that TL tetramers cytokines and killing the target cells104. Likewise, human Vδ1+ gut IELs will bind a large percentage of IELs, whether or not they express lyse MICA+ carcinoma cell lines89. Hence, it has been considered that TCRγδ93. Other human Vγ1Vδ1+ IELs reportedly recognize CD1c94, γδ+ IELs eradicate infected or transformed epithelial cells in vivo, 1000

nature immunology

•

volume 2 no 11

•

november 2001

•

http://immunol.nature.com

© 2001 Nature Publishing Group http://immunol.nature.com

© 2001 Nature Publishing Group http://immunol.nature.com

R EVIEW

IELs require additional activation in vivo to manifest full functional potential. For this reason, we have termed the constitutive state of the IEL compartment as “activated yet resting”86. Consistent with this is the high expression of the anti-proliferative genes Slfn2, Btg1 and Btg2; of Rgs1, which suppresses signaling from G protein–coupled receptors111; and of Junb, which at high JunB:c-Jun ratios suppresses cytokine gene transcription and maintains a resting state112. An “activated yet resting” state agrees with the scarcity of mitotic IELs in situ3,113, with the report that most extracted IELs are in G0 or G1114 and with the constitutive activation of mitogen-activating protein (MAP) kinase in nonproliferating IELs but not systemic Figure 3. Interactions between IELs and target cells. These may be influenced by a large number 115 cells . Interestingly, regulator of G protein signaling of proteins that appear to be expressed by IELs (top half). It is not yet known whether a single cell type would express all the cognate ligands or whether the interactions mediate the engagement by IELs of sev- 1 (RGS-1) is expressed to a far greater degree in IELs eral cell types (bottom half). + denotes those interactions that activate IEL; – denotes those that inhibit than in systemic T cells86. IELs; ? denotes unidentified ligand-receptor pairings. Provocatively, the disjunction between the high expression of cytolytic effectors and the rare expresthereby preventing systemic dissemination of infectious agents or sion of cytokines may relate to the split-anergy observed in OVA-spemalignant cells71. Consistent with this, TCRδ–/– mice are substantially cific IELs in OVA-transgenic mice recovering from gut immunopatholmore susceptible to mutagen-induced skin and colon carcinomas105,106. ogy42. Such IELs maintained cytolytic responses to stimulation but The expression of Rae-1 (a murine MICA-equivalent) is up-regulated failed to secrete cytokines, which emphasized the independent regulain the skin during chemical carcinogenesis, rendering transformed tion of the two pathways. Other studies have likewise shown that IELs cells targets of DETCs105. Interestingly, the pro-inflammatory systemic respond to CD3-TCR ligation with strong cytolytic, but weak proliferaαβ+ T cell response to cancer cells can sometimes promote tumor tive, responses114,115. For these reasons, we propose that IELs in the small development107. Therefore, the anti-inflammatory effects of DETCs intestine respond to activation in two steps (Fig. 2). The first step will involve the release of cytolytic mediators and immunoregulatory may also reduce tumor incidence, albeit indirectly. chemokines, whereas the second step will involve the up-regulation and release of cytokines, primarily TH1 cytokines. We suggest that the secMolecular characterization of IELs A more detailed description of IELs is necessary if one is to understand ond phase is facilitated by the down-regulation of JunB, Btg-1, their roles in down-regulating infection, transformation and systemic Schlafen-2 and RGS-1 and is accompanied by increased expression of infiltration of the local environment, as well as their involvement in high-affinity cytokine receptors, such as IL-15Rα, and costimulators, human disease. To this end, gene expression profiles were developed such as NKG2d, that collectively enhance afferent activation of the IEL. A fourth observation, also part-predicted by previous studies, is the for murine intestinal TCRαβ+ and TCRγδ+ IELs, which were sampled directly ex vivo. The expression of candidate genes was then validated under-representation of CD28, CD40 ligand and inducible costimulatory molecule (ICOS), which mediate costimulation. Instead, IELs may in type a and type b cells86. In aggregate, IELs were revealed as cytolytic TH1-skewed be costimulated through receptors such as 4-1BB, By55 and immunoregulatory cells, in keeping with earlier studies4,6,12,16,33–36. NKG2d90,116–119 (Figs. 2 and 3). In aggregate, IELs encode a broad range Specific results, including the expression of cytolytic mediators (such of surface receptors, with some, like CD7, expressed highly (Fig. 3). as, granzymes and Fas ligand), chemokines (such as, lymphotactin, Understanding their functional relevance will be important in clarifying macrophage inflammatory protein 1α (MIP-1α) and MIP-1β), the nature and breadth of IEL interactions with epithelial cells (Fig. 3) chemokine receptors (such as, CCR5 and CXCR3) and adhesion mole- and/or with other cells of the immune system. Indeed, IELs also cules (such as, αE and β7) were likewise predictable32,52,53,108,109. But even express several genes conventionally associated with APCs86, fuelling in these areas, the analysis has provided additional insight. For exam- speculation that they may interact directly with other T cells. ple, there had been little appreciation of the very high amounts in which A fifth observation was the similarity of gene expression between IELs express a restricted subset of effector molecules: thus, ∼3% of IEL TCRγδ+ DN IELs, TCRγδ+CD8αα+ IELs and TCRαβ+CD8αα+ IELs, mRNA is devoted to granzymes A and B and RANTES86. By this crite- which forms the basis for their collective type b classification (Table rion, IELs appear as activated effector cells, akin to plasma cells in the 1a). Of note, Ly6C—a T cell memory marker—was expressed highly in type a, but not type b, IELs86. B lineage. A second observation was the selectivity of effector molecule expression. Although the expression of chemokines that recruit CCR5+ T cells IELs and adaptive and innate immunity and APCs was very high, the expression of chemokines that recruit neu- Type a IELs bear the hallmarks of cells in the adaptive response. In trophils was negligible86. It is tempting to relate this to the anti-inflam- contrast, type b IELs could be considered “revertants” from the adapmatory effects of type b IELs. Indeed, there was high expression of pro- tive to the innate response, using gene rearrangement (a hallmark of the adaptive response) to generate receptors for conserved autoantithymosin β4, which can suppress neutrophil migration110. Other than the constitutive expression of lymphotoxin-β, TGF-β1 and gens that specify infection, cell transformation or other forms of dysthe gut-associated cytokine IL-17B, there was negligible expression of regulation within local tissues120. It is intriguing that type b IELs are conventional cytokines such as IL-2 and IFN-γ. The under-representa- expressed to varying degrees in different anatomical compartments; tion of IFN-γ and KGF, which are produced by IELs33,34,99, suggests that possibly site-specific selective pressures have operated to retain the http://immunol.nature.com

•

november 2001

•

volume 2 no 11

•

nature immunology

1001

© 2001 Nature Publishing Group http://immunol.nature.com

© 2001 Nature Publishing Group http://immunol.nature.com

R EVIEW

pattern recognition and rapid primary responses that type b cells can provide. The identities of these selective pressures are unknown, but we have previously considered that they may be age-related26. Thus, in all animals examined, γδ+ cells within the type b subset are some of the first T cell subsets to develop. Hence, they would be well placed to respond to the onslaught of environmental challenges to which the newborn is exposed. Intriguingly, antigen nonspecific cytotoxicity of rat IELs reportedly peaked at about 3 weeks of age and declined thereafter121. Despite our emphasis on features common among IELs, the IELs in one site (for example, the lungs) may exert different functions to IELs in other sites. In addition, there may be other cell types that likewise are “revertants to innate immunity”, cells with limited autoreactive specificities that may have been strongly selected for in particular anatomical sites; for example, because they could include natural killer T cells in the liver and gut and B1 cells in the mouse peritoneum and gut. Although the selective pressures for such cells may have operated locally, it is clear that such cells can have profound effects (directly or indirectly) on systemic cells. Thus, there seems no doubt that a complete understanding of immune responses and pathologies that develop within particular anatomical locations requires a substantially greater understanding of local T cells and their relatives. As one of America’s more durable politicians, the late Tip O’Neill, famously advised: “All good politics is local”. Acknowledgements We thank S. Creighton and J. Cridland for expert assistance and R.Tigelaar, J. Lewis, M. Girardi, A.Turner, P. Kilshaw and D. Oppenheim for critical discussions. Supported the Wellcome Trust, the NIH and the Dunhill Medical Trust. 1. Giddens, A. The Third Way:The renewal of social democracy (Polity Press, Cambridge, 1998). 2. Weber, E. Uber den Mechanismus der Einsaugung des Speisesaftes beim Menschen und bei einigen tieren. Physiol.Wissenschaftliche Med. Archiv Anat. 400–402 (1847). 3. Ferguson, A. Intraepithelial lymphocytes of the small intestine. Gut 18, 921–937 (1977). 4. Ernst, P., Befus, A. & Bienenstock, J. Leukocytes in the intestinal epithelium: an unusual immunological compartment. Immunol.Today 6, 50–55 (1985). 5. Klein, J. & Moseley, R. in Mucosal Immunology: Intraepithelial lymphocytes, advances in host defense mechanisms (eds Kiyono, H. & McGhee, J.) 33–60 (Raven Press, New York, 1993). 6. Beagley, K. & Husband, A. Intraepithelial lymphocytes: origins, distribution, and function. Crit. Rev. Immunol. 18, 237–254 (1998). 7. Yoshikai,Y.The interaction of intestinal epithelial cells and intraepithelial lymphocytes in host defense. Immunol. Res. 20, 219–235 (1999). 8. Fichtelius, K.-E.The mammalian equivalent to bursa Fabricii of birds. Exp. Cell Res. 46, 231–234 (1967). 9. Guy-Grand, D., Griscelli, C. & Vassalli, P.The gut-associated lymphoid system: nature and properties of the large dividing cells. Eur. J. Immunol. 4, 435–443 (1974). 10. Meuwissen, S. et al. Analysis of the lympho-plasmacytic infiltrate in Crohn’s disease with special reference to identification of lymphocyte subpopulations. Gut 17, 770–780 (1976). 11. Guy-Grand, D., Griscelli, C. & Vassalli, P.The mouse gut T lymphocyte, a novel type of T cell. Nature, origin, and traffic in mice in normal and graft-versus-host conditions. J. Exp. Med. 148, 1661–1677 (1978). 12. Davies, M. & Parrott, D.The early appearance of specific cytoxic T cells in murine gut mucosa. Clin. Exp. Immunol. 42, 273–279 (1980). 13. Tagliabue, A., Befus, A., Clark, D. & Bienenstock, J. Characteristics of natural killer cells in the murine intestinal epithelium and lamina propria. J. Exp. Med. 155, 1785–1796 (1982). 14. Klein, J. & Kagnoff, M. Nonspecific recruitment of cytoxic effector cells in the intestinal mucosa of antigen primed mice. J. Exp. Med. 160, 1931–1936 (1984). 15. Ernst, P., Clark, D., Rosenthal, K., Befus, A. & Bienenstock, J. Detection and characterization of cytotoxic T lymphocyte precursors in the murine intestinal intraepithelial leukocyte populations. J. Immunol. 136, 2121–2126 (1986). 16. Klein, J. R. Ontogeny of the Thy-1-, Lyt-2+ murine intestinal intraepithelial lymphocyte. Characterization of a unique population of thymus-independent cytotoxic effector cells in the intestinal mucosa. J. Exp. Med. 164, 309–314 (1986). 17. Viney, J., Kilshaw, P. & MacDonald,T. Cytotoxic αβ+ and γ/δ + T cells in murine intestinal epithelium. Eur. J. Immunol. 20, 1623–1626 (1990). 18. Holtmeier,W. et al. The TCRδ repertoire in normal human skin is restricted and distinct from the TCRδ repertoire in the peripheral blood. J. Invest. Dermatol. 116, 275–280 (2001). 19. Jarry, A., Cerf-Bensussan, N., Brousse, N., Seiz, F. & Guy-Grand, D. Subsets of CD3+ and CD3- lymphocytes isolated from normal human gut epithelium display phenotypic features different from their counterparts in peripheral blood. Eur. J. Immunol. 20, 1097–1103 (1990). 20. Stingl, G. et al. Thy-1+ dendritic epidermal cells express T3 antigen and the T-cell receptor γ chain. Proc. Natl Acad. Sci. USA 84, 4586–4590 (1987). 21. Asarnow, D. et al. Limited diversity of γ δ antigen receptor genes of Thy-1+ dendritic epidermal cells. Cell 55, 837–847 (1988). 22. Itohara, S., Nakanish, N., Kanagawa, O., Kube, R. & Tonegawa, S. Monoclonal antibodies specific to native murine T-cell receptor γδ: analysis of γ/δ T cells during thymic ontogeny and in peripheral lymphoid organs. Proc. Natl Acad. Sci. USA 86, 5094–5098 (1989). 23. Lefrancois, L. Phenotypic complexity of intraepithelial lymphocytes of the small intestine. J. Immunol. 147, 1746–1751 (1991).

1002

nature immunology

•

volume 2 no 11

•

24. Bucy, P., Chen, C. L., Cihak, J., Losch, U. & Cooper, M. Avian T cells expressing γ/δ receptors localize in the splenic sinusoids and the intestinal epithelium. J. Immunol. 141, 2200–2205 (1988). 25. Goodman,T. & Lefrancois, L. Expression of the γ/δ T cell receptor on intestinal CD8(+) intraepithelial lymphocytes. Nature 333, 855–858 (1998). 26. Hayday, A. C. γδ cells: a right time and a right place for a conserved third way of protection. A. Rev. Immunol. 18, 975–1026 (2000). 27. Rudzik, O. & Bienenstock, J. Isolation and characteristics of gut mucosal lymphocytes. Lab. Invest. 30, 260–266 (1974). 28. Mayrhofer, G.Thymus-dependent and thymus-independent subpopulations of intestinal intraepithelial lymphocytes: a granulated subpopulation of probable bone marrow origin. Blood 55, 532–535 (1980). 29. Camerini,V., Panwala, C. & Kronenberg, M. Regional specialization of the mucosal immune system: intraepithelial lymphocytes of the large intestine have a different phenotype and function than those of the small intestine. J. Immunol. 151, 1756–1776 (1993). 30. Mosely, R. et al. Phenotype and TCR γ/δ variable gene repertoire of intestinal intraepithelial lymphocytes in wild mice (Mus musculus domesticus): abundance of Vγ1 transcripts and extensive δ gene diversity. Int. Immunol. 2, 231–238 (1994). 31. Fujiura,Y. et al. Development of CD8aa+ intestinal intraepithelial T cell in β2-microglobulin- and/or TAP-deficient mice. J. Immunol. 156, 2710–2715 (1996). 32. Corazza, N., Muller, S., Brunner,T., Kagi, D. & Mueller, C. Differential contribution of Fas- and perforin-mediated mechanisms to the cell-mediated cytotoxic activity of naïve and in vivo-primed intestinal intraepithelial lymphocytes. J. Immunol. 164, 398–403 (2000). 33. Taguchi,T. et al. Novel function for intestinal intraepithelial lymphocytes: murine CD3+, γ/δ TCR+ T cells produce IFN-γ and IL-5. J. Immunol. 147, 3736–3744 (1991). 34. Barrett,T., Gajewski,T., Danielpour, D. & Chang, E. Differential function of intestinal intraepithelial lymphocyte subsets. J. Immunol. 149, 1124–1130 (1992). 35. Chardes,T., Buzoni-Gatel, D., Lepage, A., Bernard, F. & Bout, D. Toxoplasma gondii oral infection induces specific cytoxic CD8α/β+ Thy-1+ gut intraepithelial lymphocytes, lytic for parasite-infected enterocytes. J. Immunol. 153, 4596–4603 (1994). 36. Lundqvist, C., Melgar, S.,Yeung, M. M., Hammarstrom, S. & Hammarstrom, M. L. Intraepithelial lymphocytes in human gut have lytic potential and a cytokine profile that suggest T helper 1 and cytotoxic functions. J. Immunol. 157, 1926–1934 (1996). 37. Regnault, H., Cunano, H.,Vassalli, P., Guy-Grand, D. & Kourilsky, P. Oligoclonal repertoire of the CD8α/α TCRα/β murine intestinal intraepithelial T lymphocytes evidence for random emergence of T cells. J. Exp. Med. 180, 1345–1358 (1994). 38. Arstila,T. et al. Identical T cell clones are located within the mouse gut epithelium and lamina propia and circulate in the thoracic duct lymph. J. Exp. Med. 191, 823–834 (2000). 39. Nanno, M. et al. Development of intestinal intraepithelial T lymphocytes is independent of Peyer’s patches and lymph nodes in aly mutant mice. J. Immunol. 153, 2014–2020 (1994). 40. Marsh, M. Studies of intestinal lymphoid tissue. II. Aspects of proliferation and migration of epithelial lymphocytes in the small intestine of mice. Gut 16, 674–682 (1975). 41. Meader, R. & Landers, D. Electron and light microscopic observations on relationships between lymphocytes and intestinal epithelium. Am. J. Anat. 121, 763–774 (1967). 42. Vezys,V., Olson, S. & Lefrancois, L. Expression of intestine-specific antigen reveals novel pathways of CD8 T cell tolerance induction. Immunity 12, 505–514 (2000). 43. Guy-Grand, D., DiSanto, J., Henchez, P., Malassis-Seris, M. & Vassalli, P. Small bowel enteropathy: role of intraepithelial lymphocytes and of cytokines (IL12, IFNγ,TNF) in the induction of epithelial cell death and renewal. J. Immunol. 28, 730–744 (1998). 44. Camerini,V. et al. Generation of intestinal mucosal lymphocytes in scid mice reconstituted with mature, thymus-derived T cells. J. Immunol. 160, 2608–2618 (1998). 45. Lefrancois, L. et al. The role of β7 integrins in CD8 T cell trafficking during an antiviral immune response. J. Exp. Med. 189, 1631–1638 (1999). 46. Umesaki,Y., Setoyama, H., Matsumoto, S., Imaoka, A. & Itoh, K. Differential roles of segmented filamentous bacteria and clostridia in development of the intestinal immune system. Infect. Immun. 67, 3504–3511 (1999). 47. Dharakul,T., Rott, L. & Greeenberg, H. B. Recovery from chronic rotavirus infection with mice with severe combined immunodeficiency: virus clearance mediated by adoptive transfer of immune CD8+ T lymphocytes. J.Virol. 64, 4375–4382 (1990). 48. Lepage, A. C., Buzoni-Gatel, D., Bout, D.T. & Kasper, L. H. Gut-derived intraepithelial lymphocytes induce long term immunity against Toxoplasma gondii. J. Immunol. 161, 4902–4908 (1998). 49. Muller, S., Buhler-Jungo, M. & Mueller, C. Intestinal intraepithelial lymphocytes exert potent protective cytotoxic activity during and acute virus infection. J. Immunol. 164, 1986–1994 (2000). 50. Buzoni-Gatel, D. et al. Intraepithelial lymphocytes traffic to the intestine and enhance resistance to Toxoplasma gondii oral infection. J. Immunol. 162, 5846–5852 (1999). 51. Agace,W.W. et al. Human intestinal lamina propria and intraepithelial lymphocytes express receptors specific for chemokines induced by inflammation. Eur. J. Immunol. 30, 819–826 (2000). 52. Kunkel, E. J. et al. Lymphocyte CC chemokine receptor 9 and epithelial thymus-expressed chemokine (TECK) expression distinguish the small intestinal immune compartment: epithelial expression of tissue-specific chemokines as an organizing principle in regional immunity. J. Exp. Med. 192, 761–767 (2000). 53. Shibahara,T.,Wilcox, J. N., Couse,T. & Madara, J. L. Characterization of epithelial chemoattractants for human intestinal intraepithelial lymphocytes. Gastroenterology 120, 60–70 (2001). 54. Agace,W.W., Higgins, J. M., Sadasiven, B., Brenner, M. B. & Parker, C. M.T-lymphocyte-epithelial-cell interactions: integrin αΕ(CD103) β7, LEEP-CAM and chemokines. Curr. Opin. Cell Biol. 12, 563–568 (2000). 55. Masopust, D.,Vezys,V., Marzo, A. L. & Lefrancois, L. Preferential localization of effector memory cells in nonlymphoid tissue. Science 291, 2413–2417 (2001). 56. Masopust, D., Jiang, J., Shen, H. & Lefrancois, L. Direct analysis of the dynamics of the intestinal mucosa CD8 T cell response to systemic virus Infection. J. Immunol. 166, 2348–2356 (2001). 57. Slavin, R. & Santos, G.The graft versus host reaction in man after bone marrow transplantation: pathology, pathogenesis, clinical features, and implication. Clin. Immunol. Immunopathol. 1, 472–498 (1973). 58. Mowat, A. & Ferguson, A. Intraepithelial lymphocyte count and crypt hyperplasia measure the mucosal component of the graft-versus-host reaction in mouse small intestine. Gastroenterology 83, 417–423 (1982). 59. Rocha, B.,Vassalli, P. & Guy-Grand, D.The Vβ repertoire of mouse gut homodimeric α CD8+ intraepithelial T cell receptor α/β+ lymphocytes reveals a major extrathymic pathway of T cell differentiation. J. Exp. Med. 173, 483–486 (1991). 60. Rocha, B.,Vassalli, P. & Guy-Grand, D.Thymic and extrathymic origins of gut intraepithelial lymphocyte populations in mice. J. Exp. Med. 180, 681–686 (1994). 61. Sydora, B., Brossay, I., Hagenbaugh, A., Kronenberg, M. & Cheroutre, H.TAP-independent selection of CD8+ intestinal intraepithelial lymphocytes. J. Immunol. 156, 4209–4216 (1996). 62. Triebel, F. & Hercend,T. Subpopulations of human peripheral T γδ lymphocytes. Immunol.Today 10, 186–188 (1989).

november 2001

•

http://immunol.nature.com

© 2001 Nature Publishing Group http://immunol.nature.com

© 2001 Nature Publishing Group http://immunol.nature.com

R EVIEW

63. Deusch, K. et al. A major fraction of human intra-epithelial lymphocytes simultaneously expresses the γδ T cell receptor, the CD8 accessory molecule, and preferentially uses the Vδ1 gene segment. Eur. J. Immunol. 21, 1053–1059 (1991). 64. Mosley, R.,Wang, J., Hamad, M. & Klein, J. Functional heterogeneity of murine intestinal intraepithelial lymphocytes: studies using TCR-αβ+ IEL lines and fresh IEL isolates reveal multiple cytotoxic subsets differentiated by CD5, CD8α/α, and CD8α/β expression. Dev. Comp. Immunol. 18, 155–164 (1994). 65. Ferguson, A. & Parrott, D.The effect of antigen deprivation on thymus-dependent and thymus-independent lymphocytes in the small intestine of the mouse. Clin. Exp. Immunol. 12, 477–488 (1972). 66. Ropke, C. & Everett, N. Kinetics of intraepithelial lymphocytes in the small intestine of thymusdeprived mice and antigen-deprived mice. Anat. Rec. 185, 101–108 (1976). 67. MacDonald,T. & Ferguson, A. Small intestinal architecture and protozoal infection in mice. Gastroenterology 74, 496–500 (1978). 68. Bandeira, A. et al. Localization of γδ T cells to the intestinal epithelium is independent of normal microbial colonization. J. Exp. Med. 172, 239–244 (1990). 69. Findly, R. C., Roberts, S. J. & Hayday, A. C. Dynamic response of murine gut intraepithelial T cells after infection by the coccidian parasite Eimeria. Eur. J. Immunol. 23, 2557–2564 (1993). 70. Smith, A. L. & Hayday, A. C. An αβ T-cell-independent immunoprotective response towards gut coccidia is supported by γδ cells. Immunology 101, 325–332 (2000). 71. Janeway, C. A., Jones, B. & Hayday, A. Specificity and function of T cells bearing γ δ receptors. Immunol. Today 9, 73–76 (1988). 72. Rocha, B., von Boehmer, H. & Guy-Grand, D. Selection of intraepithelial lymphocytes with CD8α/α co-receptors by self-antigen in the murine gut. Proc. Natl Acad. Sci. USA 89, 5336–5340 (1992). 73. Poussier, P., Edouard, P., Lee, C., Binnie, M. & Julius, M.Thymus-independent development and negative selection of T cells expressing T cell receptor α/β in the intestinal epithelium: evidence for distinct circulation patterns of gut- and thymus-derived T lymphocytes. J. Exp. Med. 176, 187–199 (1992). 74. Poussier, P. & Julius, M.Thymus-independent T cell development and selection in the intestinal epithelium. A. Rev. Immunol. 145, 521–553 (1994). 75. Lefrancois, L. & Olson, S. Cutting edge: reconsitution of the extrathymic intestinal T cell compartment in the absence of irradiation. J. Immunol. 159, 538–541 (1997). 76. Kanamori,Y. et al. Identification of novel lymphoid tissues in murine intestinal mucosa where clusters of c-kit(+) IL-7R(+) Thy1(+) lympho-hempoietic progenitors develop. J. Exp. Med. 184, 1449–1459 (1996). 77. Laky, K. et al. Enterocyte expression of interleukin 7 induces development of γδ T cells and Peyer’s patches. J. Exp. Med. 191, 1569–1580 (2000). 78. Dunon, D., Cooper, M. & Imhof, B.Thymic origin of embryonic intestinal γ/δ T cells. J. Exp. Med. 177, 257–263 (1993). 79. Allison, J. & Havran,W.The immunobiology of T cells with invariant γ/δ antigen receptors. A. Rev. Immunol. 9, 679–705 (1991). 80. Lin,T. et al. Autospecific γ/δ thymocytes that escape negative selection find sanctuary in the intestine. J. Clin. Invest. 104, 1297–1305 (1999). 81. Guehler, S. R., Finch, R. J., Bluestone, J. A. & Barrett,T. A. Increased threshold for TCR-mediated signaling controls self reactivity of intraepithelial lymphocytes. J. Immunol. 160, 5341–5346 (1998). 82. Park, S. H. et al. Selection and expansion of CD8α/α(+) T cell receptor α/β(+) intestinal intraepithelial lymphocytes in the absence of both classical major histocompatibility complex class I and nonclassical CD1 molecules. J. Exp. Med. 190, 885–890 (1999). 83. Das, G. & Janeway, C. A. Jr Development of CD8α/α and CD8α/β T cells in major histocompatibility complex class I-deficient mice. J. Exp. Med. 190, 881–884 (1999). 84. Das, G. et al. Qa-2-dependent selection of CD8 α/α T cell receptor α/β(+) cells in murine intestinal intraepithelial lymphocytes. J. Exp. Med. 192, 1521–1528 (2000). 85. Fragoso, G. et al. Increased resistance to Taenia crassiceps murine cysticercosis in Qa-2 transgenic mice. Infect. Immun. 66, 760–764 (1997). 86. Shires, J.,Theodoridis, E. & Hayday, A. Biological insights into TCRγδ+ and TCRαβ+ intraepithelial lymphocytes provided by serial analysis of gene expression (SAGE). Immunity 15, 419–434 (2001). 87. Crowley, M. P. et al. A population of murine γδ T cells that recognize an inducible MHC class IB molecule. Science 287, 314–316 (2000). 88. Correa, I. et al. Most γδ T cells develop normally in β2-microglobulin-deficient mice. Proc. Natl Acad. Sci. USA 89, 653–657 (1992). 89. Groh, B., Steinle, A., Bauer, S. & Spies,T. Recognition of stress-induced MHC molecules by intestinal epithelial γδ T cells. Science 279, 1737–1740 (1998). 90. Bauer, S. et al. Activation of NK cells and T cells by NKG2d, a receptor for stress-inducible MICA. Science 285, 727–729 (1999). 91. Diefenbach, A., Jamieson, A., Liu, S., Shastri, N. & Raulet, D. Ligands for the murine NKG2D receptor: expression by tumor cells and activation of NK cells and macrophages. Nature Immunol. 1, 119–126 (2000). 92. Cerwenka, A. et al. Retinoic acid early inducible genes define a ligand family for the activating NKG2D receptor in mice. Immunity 12, 721–727 (2000). 93. Tsujimura, K. et al. The binding of thymus leukemia (TL) antigen tetramers to normal intestinal intraepithelial lymphocytes and thymocytes. J. Immunol. 167, 759–764 (2001). 94. Spada, F. M. et al. Self-recognition of CD1 by γ/δ T cells: implications for innate immunity. J. Exp. Med.

http://immunol.nature.com

•

november 2001

191, 937–948 (2000). 95. Shiohara,T., Moriya, N., Hayakawa, J., Itohara, S. & Ishikawa, H. Resistance to cutaneous graft-vs.-host disease is not induced in T cell receptor δ gene–mutant mice. J. Exp. Med. 183, 1483–1489 (1996). 96. Roberts, S. J. et al. T-cell αβ+ and γδ+ deficient mice display abnormal but distinct phenotypes toward a natural, widespread infection of the intestinal epithelium. Proc. Natl Acad. Sci. USA 93, 11774–1179 (1996). 97. Fujihashi, K. et al. Regulatory function for murine intraepithelial lymphocytes.Two subsets of CD3+,T cell receptor-1+ intraepithelial lymphocyte T cells abrogate oral tolerance. J. Immunol. 145, 2010–2019 (1990). 98. Ke,Y., Pearce, K., Lake, J. P., Ziegler, K. H. & Kapp, J. A. γδ T lymphocytes regulate the induction and maintenance of oral tolerance. J. Immunol. 158, 3610–3618 (1997). 99. Boismenu, R. & Havran,W. Modulation of epithelial cell growth by intraepithelial γδ T cells. Science 266, 1253–1255 (1994). 100. Komano, H. et al. Homeostatic regulation of intestinal epithelia by intraepithelial γδ T cells. Proc. Natl Acad. Sci. USA 92, 6147–6151 (1995). 101. Fahrer, A. et al. Attributes of γδ intraepithelial lymphocytes as suggested by their transcriptional profile. Proc Natl Acad. Sci. USA 98, 10261–10266 (2001). 102. Ferguson, A. & Murray, D. Quantitation of intraepithelial lymphocytes in human jejunum. Gut 12, 988–994 (1971). 103. Hayday, A. et al. Intraepithelial γ/δ + T cells in natural infection and in coeliac disease: protectos of epithelial integrity and mediators of immune regulation - a hypothesis. 6th Int. Proc. Coeliac Disease 46–57 (Oak Tree Press, Dublin, 1994). 104. Havran,W., Chien,Y. & Allison, J. Recognition of self antigens by skin derived T cells with invariant γδ antigen receptors. Science 252, 1430–1432 (1991). 105. Girardi, M. et al. Regulation of cutaneous malignancy by γδ T cells. Science 10. 1126/1063916 (Science Express, 2001). 106. Matsuda, S., Kudoh, S. & Katayama, S. Enhanced formation of Azoxymethane-induced colorectal adenocarcinoma in γ/δ T lymphocyte-deficient mice. J. Cancer Res. 92, 880–885 (2001). 107. Coussens, L. & Werb, Z. Inflammatory cells and cancer.Think different! J. Exp. Med. 193, 23–26 (2001). 108. Kilshaw, P. & Murant, S. Expression and regulation of β7 integrins on muse lymphocytes: relevance to the mucosal immune system. Eur. J. Immunol. 21, 2591–2597 (1991). 109. Boismenu, R., Feng, L., Xia,Y.Y., Chang, J. C. & Havran,W. L. Chemokine expression by intraepithelial γδ T cells. Implications for the recruitment of inflammatory cells to damaged epithelia. J. Immunol. 157, 985–992 (1996). 110. Young, J. D. et al. Thymosin β4 sulfoxide is an anti-inflammatory agent generated by monocytes in the presence of glucocorticoids. Nature Med. 5, 1424–1427 (1999). 111. Rief, K. & Cyster, J. RGS molecule expression in murine B lymphocytes and ability to down-regulate chemotaxis to lymphoid chemokines. J. Immunol. 164, 4720–4729 (2000). 112. Szabowski, A. et al. c-Jun and JunB antagonistically control cytokine-regulated mesenchymal-epidermal interaction in skin. Cell 103, 745–755 (2000). 113. Darlington, D. & Rogers,W. Epithelial lymphocytes in the small intestine of the mouse. J. Anat. 100, 813–830 (1966). 114. Yamamoto, M. et al. Cytokine synthesis and apoptosis by intestinal intraepithelial lymphocytes: signaling of high-density αβ and γδ T cells via T cell receptor-CD3 complex results in interferon-γ and interleukin-5 production, while low-density T cells undergo DNA fragmentation. Eur. J. Immunol. 24, 1301–1306 (1994). 115. Sydora, B. et al. Intestinal intraepithelial lymphocytes are activated and cytolytic but do not proliferate as well as other T cells in response to mitogenic signals. J. Immunol. 150, 2179–2191 (1993). 116. Ohteki,T. & MacDonald, H. Expression of the CD28 costimulatory molecule on subsets of murine intestinal intraepithelial lymphocytes correlates with lineage and responsiveness. Eur. J. Immunol. 23, 1251–1255 (1993). 117. Vinay, D.S. & Kwon, B.S. Role of 4–1BB in immune responses. Semin. Immunol. 10, 481–489 (1998). 118. Anumanthan, A. et al. Cloning of BY55, a novel Ig superfamily member expressed on NK cells, CTL, and intestinal intraepithelial lymphocytes. J. Immunol. 161, 2780–2790 (1998). 119. Takahashi, C., Mittler, R. S. & Vella, A.T. Cutting edge: 4-1BB is a bona fide CD8 T cell survival signal. J. Immunol. 162, 5037–5040 (1999). 120. Boismenu, R. & Havran,W. An innate view of γ δ T cells. Curr. Opin. Immunol. 9, 57–63 (1997). 121. Tice, D. Ontogeny of NK activity in rat small bowel. Transplant. Proc. 22, 2458–2495 (1990). 122. Guy-Grand, D. et al. Different use of T cell receptor transducing modules in two populations of gut intraepithelial lymphocytes are related to distinct pathways of T cell differentiation. J. Exp. Med. 180, 673–679 (1994). 123. Eisenbraum, M. D., Mosley, R. L., Teitelbaum, D. H. & Miller, R. A. Altered development of intestinal intraepithelial lymphocytes in P-glycoprotein deficient mice. Dev. Comp. Immun. 24, 783–795 (2000). 124. Puddington, L., Olson, S. & Lefrancois, L. Interations between stem cell factor and c-kit are required for intestinal immune system homeostasis. Immunity 1, 733–739 (1994). 125. Page, S.T., van Oers, N. S. C., Perlmutter, R. M.,Weiss, A. & Pullen, A. M. Differential contribution of lck and fyn protein tyrosine kinases to intraepithelial lymphocyte development. Eur J. Immunol 27, 554–562 (1997).

•

volume 2 no 11

•

nature immunology

1003