REVIEWS

Tissue-tropic effector T cells: generation and targeting opportunities William W. Agace

Abstract | The localization of effector T cells to extralymphoid tissues is crucial for the generation of an effective immune response, but it also underlies many autoimmune and inflammatory disorders. Recent studies have highlighted a central role for draining lymph nodes and environmentally imprinted dendritic cells in the generation of tissue-tropic effector T cells. Here, I outline our current understanding of the mechanisms that regulate the generation and localization of tissue-tropic effector T cells, and the potential ways in which these pathways can be exploited for immunotherapeutic purposes.

Gut-associated lymphoid tissue Lymphoid structures and aggregates associated with the intestinal mucosa.

B-cell follicle An aggregate of B cells in lymphoid tissues. It contains naive B cells, as well as activated, proliferating and maturing B cells in germinal centres. B-cell follicles are contiguous with T-cell zones.

Efferent lymph vessels Lymphatic vessels that transport interstitial fluid and immune cells out of lymph nodes.

Thoracic duct The major lymphatic vessel that collects lymph from the lower limbs, abdomen, chest and pelvis and returns it to the circulation by the internal jugular and subclavian veins.

Immunology Section, Lund University, BMC I13, 22184 Lund, Sweden. e-mail: [email protected] doi:10.1038/nri1869

The initiation and maintenance of an effective immune response and the establishment of immunological memory are crucially dependent on the orchestrated migration of T-cell subsets to distinct tissue locations. After leaving the thymus, circulating antigen-inexperienced (that is, naive) T cells traffic continually through secondary lymphoid organs such as the spleen, peripheral lymph nodes and gut-associated lymphoid tissue (GALT) in search of their cognate peptide–MHC complex, and they are largely excluded from extralymphoid tissues1. Following their activation in secondary lymphoid tissues, naive T cells differentiate into effector T cells, the functions of which are determined, at least in part, by the environment in which they are initially activated. Some CD4+ effector T cells migrate towards the B-cell follicle in the secondary lymphoid organ in which they were activated, and provide help to B cells in a process that is crucial for germinal-centre formation. Others, together with CD8+ effector T cells, enter the circulation through the efferent lymph vessels and the thoracic duct, and localize to extralymphoid tissues, where they help to coordinate immune and inflammatory responses in the periphery (FIG. 1). Effector T-cell subsets show tropism for different extralymphoid tissues, such as the skin and intestinal tract. This tropism seems to be determined by the selective and rapid induction of tissue-homing receptor expression following priming, and the selective expression of the corresponding ligands in cutaneous and intestinal vascular beds. Which tissue-homing receptors are induced during T-cell priming depends on the lymphnode environment, which itself is influenced by tissuederived dendritic cells (DCs) and lymph. In this Review, I discuss our current understanding of the mechanisms that regulate the generation of tissue-tropic effector T-cell

682 | SEP T E M BER 20 0 6 | VOLU M E 6

subsets, and the possibility of using this knowledge for the improved development of immunotherapies and the rational design of vaccines.

Extralymphoid tropism of effector T-cell subsets T-cell migration to lymphoid and extralymphoid tissues is a multistep process that is regulated by the coordinated interaction of various cell-surface molecules on the T cell with their respective ligands on the surface of vascular endothelial cells (FIG. 2). The concept that antigen-experienced lymphocyte subsets show tropism for distinct extralymphoid tissues arose from early studies in which adoptively transferred lymphocytes were found to migrate preferentially to those tissues from which they were isolated2–4 . Subsequent studies defined two non-overlapping antigen-experienced T-cell populations in human blood: one subset expressing the α4β7-integrin, and the other expressing cutaneous leukocyte antigen (CLA), which is a carbohydrate epitope recognized by the antibody HECA-452 (for a review see REF. 5). These two T-cell subsets have preferential homing capacity for intestinal and cutaneous tissues, respectively 5 . The ligand for α4β7-integrin is MADCAM1 (mucosal vascular addressin cell-adhesion molecule 1) (REF. 6), which is expressed under steadystate conditions by endothelial cells in the intestinal tract and associated lymphoid tissues7. Studies using neutralizing antibodies specific for α4β7-integrin or MADCAM1, and studies using knockout mice, have shown a role for α 4β 7-integrin and MADCAM1 in mediating the entry of CD4+ and CD8+ effector T cells to these sites8–10 . The importance of α4β7-integrin as an intestine-specific ‘area code’ seems to be maintained during intestinal inflammation, as antibodies specific www.nature.com/reviews/immunol

© 2006 Nature Publishing Group

REVIEWS Ulcerative colitis An inflammatory bowel disease characterized by chronic inflammation of the colon.

SAMP1/yit mice A mutant mouse strain that spontaneously develops a chronic intestinal inflammation that is mainly localized in the terminal ileum.

for either α 4β 7-integrin or MADCAM1 attenuate inflammation in animal models of colonic inflammation 11–13 , and a humanized α 4β 7-integrin-specific antibody was found to induce clinical and endoscopic remission in patients with active ulcerative colitis14 . The α4β7-integrin–MADCAM1 interaction might also have a role in T-cell entry into the inflamed small intestine, as MADCAM1-specific antibody inhibited T-cell adhesion to the terminal ileal microvascular endothelium of senescence-accelerated mouse P1 (SAMP1)/yit mice, suppressed the development of spontaneous ileitis in these mice and attenuated established ileitis15 . Nevertheless, ileitis induced by the transfer of T cells from SAMP1/yit mice to severe combined immunodeficient (SCID) mice was unaffected by treatment with antibodies specific for either MADCAM1 or α4β7-integrin alone16 .

Spleen

Lymph node

Naive T cell

B-cell follicle

Circulation

Vein Artery

Blood in Blood out

HEV

Thoracic duct

T-cell zone

B

Afferent lymph vessel

Follicular B helper T cell

Efferent lymph vessel Medullary sinus Medulla

TH cell

Cytotoxic T cell Efferent lymph

B-cell follicle

Circulation

T-cell activation TCR Peptide–MHC

DC

Extralymphoid tissues: Inflamed skin Lung Intestine Liver Brain Others?

Figure 1 | Circulation of naive and effector T-cell populations. Naive T cells continually circulate through secondary lymphoid organs and the spleen in search of their cognate peptide–MHC complex on the surface of antigen-presenting cells. They enter lymph nodes across specialized high endothelial venules (HEVs) and return to the circulation through efferent lymphatics and the thoracic duct. Following activation in the lymph node, T cells differentiate into effector T cells (such as T helper (TH) cells and cytotoxic T cells), some of which migrate towards the B-cell follicle to provide help to B cells (follicular B helper T cells). A subset of CD4+ effector T cells together with CD8+ effector T cells leave the lymph node, return to the circulation through the lymph system, and might enter a wide range of extralymphoid tissues, where they help to coordinate immune responses in the periphery. DC, dendritic cell; TCR, T-cell receptor.

Furthermore, β7-integrin-deficient CD8+ effector T cells can enter the mouse small intestine and provide immunity to infection with rotavirus 9,17. So, α 4β 7integrin is not absolutely required for T-cell entry into the intestinal mucosa. CLA is a ligand for both endothelial-cell selectin (E-selectin) and platelet selectin (P-selectin) when displayed on P-selectin glycoprotein ligand 1 (PSGL1), and it was recently implicated as a ligand for E-selectin when displayed on CD43 (REFS 18,19); it is expressed by most skin-resident T cells20 . E-selectin and P-selectin are constitutively expressed by noninflamed dermal microvessels at sufficient levels to support T-cell rolling21. However, under homeostatic conditions, the entry of effector T cells into the skin is limited. During cutaneous inflammation, there is a marked increase in E-selectin and P-selectin expression, and mouse CD4+ and CD8+ effector T cells that express ligands for E-selectin and P-selectin enter the skin through an E-selectin- and P-selectin-dependent process22–26 . E-selectin and P-selectin are expressed on the vascular endothelium in several tissues during inflammation, and might also mediate effector T-cell migration to non-cutaneous sites of inflammation27–30 . Therefore, expression of ligands for E-selectin and P-selectin alone cannot be regarded as strict markers of skin-tropic T cells. Chemokine receptors also have an important role in regulating effector T-cell tropism for intestinal and cutaneous sites31. CC-chemokine receptor 9 (CCR9) is selectively expressed by a subset of circulating α4β7integrinhi T cells in human peripheral blood, by CD4+ and CD8+ T cells in the small intestine of humans and mice, and by a minority of T cells in the colon of humans and mice, but it is rarely expressed by T cells resident in other extralymphoid tissues32–35. The ligand for CCR9, CC-chemokine ligand 25 (CCL25), is constitutively and selectively expressed by epithelial cells of the human and mouse small intestine32,33,36 . Studies in mice using antibodies specific for CCL25, as well as competitive transfer experiments using CD4+ and CD8+ T-cell receptor (TCR)-transgenic Ccr9–/– and wild-type cells, have shown an important role for CCR9 in the recruitment of effector T cells to the small-intestinal lamina propria and epithelium34,37,38 but not to the colon, liver or lungs. Therefore, CCR9 seems to define a subset of effector T cells with selective tropism for the small intestine. Nevertheless, CD4+ effector T cells from Ccr9–/– mice can enter the small-intestinal lamina propria38 , and Ccr9–/– mice have relatively normal numbers of smallintestinal CD4+ and CD8+ T cells39,40 . Therefore, the requirement for CCR9 in T-cell localization to the small intestine is not absolute. Two other chemokine receptors, CCR4 and CCR10, have been implicated in regulating effector T-cell tropism to sites of cutaneous inflammation. CCR4 is expressed by most CLA+CD4+ T cells in human peripheral blood. One of its two ligands, CCL17, is expressed at low levels by skin venules and its expression is increased during inflammation41. CCR10 is also expressed by a subset of CLA+CD4+ T cells, and one of its ligands, CCL27, is VOLU M E 6 | SEP T E M BER 20 0 6 | 683

NATURE REVIEWS | IMMUNOLOGY © 2006 Nature Publishing Group

REVIEWS constitutively and selectively expressed by skin keratinocytes42–45. Studies in mice have indicated a role for both CCR4 and CCR10 in CD4+ effector T-cell localization to sites of delayed-type hypersensitivity (DTH)-induced skin inflammation42,46 . However, CCR10 is expressed Activation

Tethering and rolling

Adhesion Inactive LFA1

Naive T cell Inactive α4β7-integrin MADCAM1

CCR7

L-selectin

Inactive α4β7-integrin

Active LFA1

CCL21 ICAM2 and CCL19

PNAD

Active α4β7-integrin ICAM1

MADCAM1

HEV

Inactive α4β7-integrin

Effector T cell Inactive α4β7-integrin

MADCAM1

CCR9

PSGL1

Active α4β7-integrin CCL25 Others?

P-selectin

MADCAM1

Small-intestinal venule

Inactive LFA1 Effector T cell CD43? Others?

CCR4

CCR10

PSGL1 CCL17

E-selectin

Inactive α4β1-integrin

P-selectin

Active LFA1

Active α4β1-integrin

CCL27? ICAM1

VCAM1

Cutaneous venule

Figure 2 | T-cell entry into lymphoid and extralymphoid tissues is a multistep process coordinated by selective expression of tissue-homing receptors. The mechanism by which naive T cells interact with high endothelial venules (HEVs) to gain entry into secondary lymphoid organs has been well characterized. Naive T-cell adhesion to HEVs is mediated initially by interactions between L-selectin on the T-cell surface and carbohydrate ligands known collectively as peripheral-node addressin (PNAD) on HEV surface glycoproteins, resulting in T-cell tethering and rolling. Mucosal addressin cell-adhesion molecule 1 (MADCAM1) is an L-selectin glycoprotein ligand on intestinal HEVs and can also interact with α4β7-integrin to mediate selectinindependent rolling. CC-chemokine ligand 21 (CCL21) presented on the HEV surface then signals through CC-chemokine receptor 7 (CCR7) on the rolling T cell to induce a rapid activation of integrins. Lymphocyte function-associated antigen 1 (LFA1) on the T cell, with a contribution from α4β7-integrin in intestinal lymph nodes, then binds to the immunoglobulin-superfamily ligands intercellular adhesion molecule 1 (ICAM1) (and ICAM2 in peripheral lymph nodes118) and MADCAM1, respectively, resulting in T-cell adhesion. T-cell adhesion or arrest is followed by transendothelial migration; however, the molecular mechanisms regulating this final step in the cascade are less well understood. Although a similar multistep adhesion cascade seems to be used for effector T-cell entry into extralymphoid tissues, such as the skin and small intestine, the use of different tissue-homing-receptor-family members promotes effector T-cellsubset localization to these tissues. PSGL1, P-selectin glycoprotein ligand 1, VCAM1, vascular cell-adhesion molecule 1.

684 | SEP T E M BER 20 0 6 | VOLU M E 6

by only a minority of CD4+ T cells in experimental Candida-extract-induced DTH and Haemophilus ducreyi-induced chancroid skin lesions in humans43. So, the contribution of CCR4 and CCR10 to T-cell homing to sites of cutaneous inflammation might vary depending on the type of inflammation. Of note, CCR4 is expressed by a subset of circulating human CLA– T cells, including some antigen-experienced α4β7-integrin+ T cells, and by subsets of T cells in the lungs and the synovial fluid of patients with rheumatoid arthritis43,47–49. Therefore, as with E-selectin ligands and P-selectin ligands, CCR4 expression alone cannot be regarded as a marker for skin-homing T cells. Together, these results highlight the importance of coordinated co-expression of tissuehoming receptors in regulating tissue tropism of effector T cells, and they indicate that there is a degree of redundancy in the mechanisms by which effector T cells can enter distinct extralymphoid tissues (FIG. 2).

Tissue-tropic effector T-cell generation in vivo Activation of naive CD4+ T cells in secondary lymphoid tissues leads to the concurrent generation of effector T cells expressing CXC-chemokine receptor 5 (CXCR5), which are specialized for B-cell help, and effector T cells expressing CXCR3 and ligands for P-selectin, which are specialized for tissue-inflammatory functions 50 . Expression of tissue-specific homing receptors is also induced following CD8+ and CD4+ T-cell priming in secondary lymphoid organs. However, which homing receptors are induced on T cells depends on the lymph node in which activation takes place. Original studies in humans indicated that CLA expression is preferentially induced on T cells during naive T-cell activation in skin-draining lymph nodes51, and these results have subsequently been confirmed and extended in mice using TCR-transgenic T-cell adoptive-transfer models. Two days after intraperitoneal administration of a model antigen, CD4+ T cells activated in skin-draining lymph nodes gain the ability to bind P-selectin52 , whereas CD4+ T cells activated in mesenteric lymph nodes upregulate their expression of α4β7-integrin and CCR9 (REFS 38,52). Similarly, CD8+ T cells activated in skin-draining lymph nodes gain the ability to bind E-selectin and P-selectin, whereas priming in mesenteric lymph nodes induces the expression of α4β7-integrin and CCR9 (REFS 37,38,53). Although expression of E-selectin ligands seems to be selectively induced on CD4+ and CD8+ T cells activated in skin-draining lymph nodes, the expression of P-selectin ligands has, in one study, been observed on CD4+ T cells primed in mesenteric lymph nodes10 . Together, these studies show that the expression of intestinal and cutaneous homing receptors are rapidly induced on T cells following their priming in gut- and skin-draining lymph nodes, respectively. Factors that regulate the efficiency of tissue-tropic T-cell generation in vivo include the route of antigen administration and the presence or absence of adjuvant34,37,54 . For example, in an ovalbumin-specific TCR-transgenic adoptive-transfer model, oral administration of ovalbumin alone led to an efficient induction of CCR9 and α4β7-integrin expression on responding www.nature.com/reviews/immunol

© 2006 Nature Publishing Group

REVIEWS Small-intestinal lymph node Small intestine

Circulation

Skin lymph node

Cervical lymph node

Efferent lymph

Other lymph nodes draining tissues such as the colon, lungs and liver

Skin

Brain

Effector T cells

Other tissues such as the colon, lungs and liver

or

Figure 3 | Generation of tissue-tropic effector T-cell subsets and their localization to extralymphoid tissues. Following their activation in secondary lymphoid tissues, effector T cells gain the ability to enter a wide range of extralymphoid tissues. Certain lymphoid tissues (small-intestinal, skin and cervical lymph nodes) seem to generate effector T-cell populations with enhanced tropism for the tissues that they drain. So, the small intestine and skin seem to selectively recruit intestinal- and skin-lymphnode-primed T cells, respectively, from the circulating effector T-cell pool. For simplicity, the skin, small intestine and brain are depicted as collecting only those effector populations generated in the relevant draining lymph node. However, there is likely to be redundancy in the system so that effector T-cell populations generated in other lymph nodes can enter these sites, albeit to a lesser extent. It is currently unclear whether T cells primed in lymph nodes that drain extralymphoid tissues such as the colon, lungs, liver and kidney display enhanced tropism for the tissues that they drain; however, effector T-cell populations generated in non-draining lymph nodes can readily enter the lungs and liver.

Lamina propria Connective tissue that underlies the epithelium of the mucosa and contains various myeloid and lymphoid cells, including macrophages, dendritic cells, T cells and B cells.

Delayed-type hypersensitivity A cell-mediated immune response that is evoked following antigen administration in the skin.

Chancroid Sexually transmitted disease that is caused by the Gram-negative bacterium Haemophilus ducreyi and is characterized by necrotizing genital ulceration.

Pertussis toxin Pertussis toxin blocks Gαicoupled receptor signalling (including chemokine-receptor signalling) by catalysing ADP ribosylation of Gαi.

CD8+ T cells in mesenteric lymph nodes, whereas intraperitoneal administration of antigen induced efficient expression of these receptors only after co-administration of adjuvant. These effects are probably due, at least in part, to differential targeting of DC populations, as is discussed later. Of note, recent studies in mice have shown that the initial T-cell-precursor frequency and the timing of T-cell activation during a primary immune response influence the expression of L-selectin (also known as CD62L) on responding T cells55,56 . Whether these parameters influence the induction of extralymphoid-tissue-homing receptors remains to be assessed.

Additional tissue-tropic effector T-cell subsets Effector T-cell entry into inflamed extralymphoid tissues other than the intestine and skin is also an active process, involving — but not restricted to — members of the integrin, selectin-ligand and chemokine-receptor families, and it has been the subject of several excellent reviews57–59. In addition, mouse effector T cells can enter various non-inflamed extralymphoid tissues60–62 through, at least in some cases, an active process. For example, the non-inflamed liver seems to be a site of CD8+ effector T-cell localization that is mediated by intercellular adhesion molecule 1 (ICAM1) and vascular cell-adhesion molecule 1 (VCAM1) (REF. 63). In addition, CD8+ effector T-cell entry to the non-inflamed lung

parenchyma is lymphocyte function-associated antigen 1 (LFA1; CD11a–CD18) dependent and pertussis-toxin sensitive, in part because of the constitutive expression of CCL5 in the lungs64 . The entry of CD4+ T-cell blasts into the non-inflamed spinal-cord parenchyma is pertussis-toxin sensitive, and LFA1 and α4-integrin– VCAM1 dependent65,66 . CD8+ effector T cells generated in gut-draining lymph nodes in response to a model antigen that is expressed selectively by epithelial cells of the small intestine disseminate to several extralymphoid tissues, including the liver, lungs, brain and kidneys61. This indicates that effector T cells do not need to be primed in draining lymph nodes to enter these tissues. Consistent with these observations, intestinal-lymphnode-derived CCR9+α4β7-integrin+CD8+ effector T cells are readily detected in the liver and lungs 3 days after intraperitoneal administration of antigen37. A remaining question is whether T cells show increased tropism for extralymphoid tissues other than the intestine and skin, if they are primed in the relevant draining lymph node. Recently, mouse CD8+ T cells primed in cervical lymph nodes after intracerebral injection of tumour cells were found to express a partially overlapping, but distinct, adhesion-molecule expression profile compared with T cells primed in the mesenteric and inguinal lymph nodes. This expression profile included high levels of expression of ligands for P-selectin on a subpopulation of T cells and expression of α4β1-integrin53. Importantly, effector T cells primed in the cervical lymph nodes showed an increased ability to enter the central nervous system after transfer to brain-tumour-bearing mice compared with T cells primed in the inguinal lymph nodes. Of note, treatment of patients suffering from relapsing multiple sclerosis with natalizumab (Tysabri; Biogen Idec Inc. and Elan Corporation Plc), which is a humanized antibody specific for the α4-integrin chain, has shown therapeutic efficacy and provides proof of principle for the role of this integrin in the entry of encephalitogenic T cells into the central nervous system67,68 . Together, these results indicate the intriguing possibility that the acquisition of specific homing ‘area codes’ by T cells in draining lymph nodes might be a more generalized phenomenon (FIG. 3).

Role of dendritic cells DCs mediate the selective generation of gut- and skin-tropic effector T cells in intestinal and skindraining lymph nodes, respectively. Mouse CD4 + and CD8 + TCR-transgenic T cells stimulated with antigen-pulsed DCs isolated from intestinal lymph nodes, or with CD3-specific antibody in the presence of these DCs, express CCR9 and high levels of α4β7-integrin37,38,54,69,70 . The induction of α4β7-integrin expression on CD8+ T cells seems to be regulated at the level of α4-integrin mRNA expression, as α4-integrin, but not β7-integrin, mRNA levels are increased after CD8+ T-cell priming with Peyer’s patch DCs69. DCs are necessary and sufficient for the induction of CCR9 and α 4β 7-integrin expression, because CD8+ T cells stimulated with peptide-pulsed mesentericlymph-node cells that are depleted of DCs failed to VOLU M E 6 | SEP T E M BER 20 0 6 | 685

NATURE REVIEWS | IMMUNOLOGY © 2006 Nature Publishing Group

REVIEWS Retinoic acid Others?

MHC Intestinal lamina-propria DC or CD103+ mesentericlymph-node DC

TCR Naive T cell Competing soluble signal(s)

Retinoic acid Others? Spleen DC or CD103– mesentericlymph-node DC

Retinoic acid? Others? Peripherallymph-node DC or Langerhans cell

Competing soluble signal(s)

Competing soluble signal(s)

CCR4hi

α4β7-integrinlow

CCR9hi

E-selectin ligandhi

α4β7-integrin

hi

E-selectin ligandlow Gut-tropic effector T cell ↓ E-selectin ligands ↓ FucT-VII

Tissue tropism of effector T cell unclear ↓ CCR9

Skin-tropic effector T cell ↓ CCR9 ↓ α4β7-integrin ↑ FucT-VII

Figure 4 | Dendritic cells induce expression of tissue-homing receptors on T cells following their priming in regional lymph nodes. The tissue-homing-receptor repertoire induced on effector T cells depends on soluble signals derived from regionally imprinted dendritic cells (DCs). Expression of α4β7-integrin and the ligands for E-selectin are induced at low levels on T cells activated by a wide range of DCs. This ‘default’ expression of homing receptors can be regulated by soluble factors that are selectively released by gut- and skin-derived DCs. So, intestinal-derived DCs (mainly lamina-propria and CD103+ mesenteric-lymph-node DCs) produce soluble factors that enhance α4β7-integrin expression and induce CC-chemokine receptor 9 (CCR9) expression on responding T cells, while suppressing the expression of ligands for E-selectin and the expression of mRNA encoding fucosyltransferase-VII (FucT-VII). This seems to be mediated, at least in part, by the vitamin A metabolite retinoic acid. By contrast, skin-draining-lymph-node DCs generate soluble factors that enhance the expression of ligands for E-selectin and suppress the expression of α4β7-integrin and CCR9. Although three DC populations are depicted in the figure, DC variants along this axis are likely to exist. For example, Langerhans cells seem to be more potent in generating skin-tropic effector T cells than peripheral-lymph-node DCs. The ability of a particular DC subset to induce the expression of skin- or intestinal-homing receptors on responding T cells might therefore reflect the levels of retinoic acid and the levels of inducers of skin-homing receptors that they produce. TCR, T-cell receptor.

Inguinal lymph node The inguinal lymph nodes are situated in the upper thigh near the groin and receive draining lymph from cutaneous tissue of the lower extremities.

Peyer’s patches Specialized lymphoid follicles that are localized in the submucosa of the small intestine and appendix.

induce expression of these homing receptors37. By contrast, compared with intestinal-lymph-node DCs, DCs from skin-draining lymph nodes induce CD8 + T cells to express higher levels of fucosyltransferaseVII (FucT-VII) mRNA (FucT-VII is an enzyme that is required for the synthesis of ligands for E-selectin and P-selectin and for expression of the CLA epitope on human T cells), higher cell-surface levels of ligands for E-selectin and P-selectin, and more CCR4 mRNA and protein71–73. However, at least in vivo, CCR4 expression might be transiently induced on CD8+ T cells activated in non-cutaneous lymph nodes72 . Together, these results show that skin- and intestinal-lymph-node-derived DCs induce a similar expression pattern of tissue-homing receptors on responding T cells in vitro to that observed on T cells primed in the corresponding lymph node in vivo. Therefore, DCs seem to have a central role in the generation of tissue-tropic effector T-cell subsets. The mechanisms by which DCs from different secondary lymphoid tissues selectively generate tissuetropic effector T-cell subsets are currently the subject of intensive investigation. A recent breakthrough was made by Iwata et al.74 , who found that addition of retinoic acid (a metabolite of vitamin A) to mouse CD4+ or CD8+ T cells stimulated with CD3- and CD28-specific antibody or to mouse CD4 + T cells primed under T helper 1 (TH1)- or TH2-cell-polarizing conditions induced the expression of CCR9 and α4β7-integrin and

686 | SEP T E M BER 20 0 6 | VOLU M E 6

suppressed the expression of ligands for E-selectin and P-selectin74 . Synthesis of retinoic acid depends on the oxidative metabolism of retinol to retinal (which requires alchohol dehydrogenases) and then the conversion of retinal to retinoic acid (which requires retinal dehydrogenases (RALDHs)). Importantly, the ability of mesenteric-lymph-node and Peyer’s patch DCs to generate gut-tropic T cells was reduced in the presence of a RALDH inhibitor, and mesenteric-lymph-node and Peyer’s patch DCs, but not splenic DCs, could generate retinoic acid from all trans-retinol74 . Retinoic acid binds two families of nuclear receptors, the retinoic-acid receptors (RARs) and the retinoid X receptors (RXRs), and the ability of mesenteric-lymph-node and Peyer’s patch DCs to induce α 4β 7-integrin on responding T cells was blocked with an RAR antagonist74 . Together, these results suggest that the ability of intestinal DCs to generate CCR9+α4β7-integrin+ T cells lies in their selective ability to generate retinoic acid. Nevertheless, in vitro, T cells primed with antigen-pulsed splenic or peripheral-lymph-node DCs express α4β7-integrin, but not CCR9, after prolonged culture (REF. 71; M. Svensson and B. Johansson-Lindbom, unpublished observations). In my laboratory, splenic-DC-induced expression of α4β7-integrin on CD8+ T cells was also inhibited with a pan-RAR antagonist (M. Svensson and W.W.A., unpublished observations), indicating that splenic DCs can generate retinoic acid. The ability of non-intestinal www.nature.com/reviews/immunol

© 2006 Nature Publishing Group

REVIEWS

Langerhans cells Professional antigenpresenting dendritic cells that are localized in the skin epidermis.

Atopic dermatitis A common skin-associated chronic hypersensitivity disorder, the aetiology of which remains poorly understood, but in which responsible allergens can occasionally be identified.

Allergic contact dermatitis Cutaneous inflammatory condition caused by a T-cellmediated hypersensitivity to defined allergens.

DCs to induce the expression of α4β7-integrin, but not CCR9, on responding T cells in vitro suggests that α4β7integrin and CCR9 have different requirements for their expression. α4β7-Integrin expression by activated CD4+ T cells in vivo is reduced after treatment with a neutralizing antibody specific for the OX40 ligand75. So a possible scenario is that OX40 signalling might function in synergy with low levels of retinoic acid to trigger expression of α4β7-integrin on T cells, whereas the induction of CCR9 expression might require higher levels of retinoic acid and/or that the induction of CCR9 expression might be more sensitive to soluble inhibitors produced by DCs (FIG. 4; and see later sections). Of note, CD4+CD8+ double-positive thymocytes and intraepithelial CD8+ T cells in the small intestine express CCR9 but not α4β7-integrin34,76–78 , indicating that additional environmental factors have a significant effect on the selective expression of either receptor. For example, expression of α4β7-integrin, but not CCR9, is downregulated by effector T cells subsequent to their entry into the small-intestinal epithelium77. What then underlies the increased ability of peripherallymph-node DCs to induce expression of FucT-VII and selectin ligands? Mouse CD8+ T cells responding to fixed antigen-pulsed Peyer’s patch DCs failed to express CCR9 and α4β7-integrin but instead expressed increased levels of ligands for E-selectin and P-selectin. This indicates that expression of ligands for E-selectin and P-selectin is part of a default pathway that is actively suppressed by retinoic acid73 . However, even this default pathway must be differentially regulated by different DC populations, because Langerhans cells seem to be better than splenic and peripheral-lymph-node DCs at inducing the expression of ligands for E-selectin on responding CD8+ T cells. Indeed, supernatants from mouse peripheral-lymph-node DC–T-cell co-cultures increased the expression of ligands for E-selectin on T cells activated with CD3-specific antibody and suppressed the expression of α 4β 7-integrin71,72 . The soluble factor(s) is unlikely to be interleukin-12 (IL-12), which is an inducer of FucT-VII expression79, as IL-12-specific antibody had no effect on the ability of peripheral-lymph-node DCs to induce responding T cells to express ligands for E-selectin and P-selectin73. In summary, the increased ability of particular DCs to generate gut- or skin-tropic effector T cells might lie in their relative production of soluble mediators (including retinoic acid) that enhance the expression of CCR9 and α 4β 7-integrin while suppressing the expression of E-selectin ligand, and of soluble mediators that enhance E-selectin ligand expression while suppressing the expression of α4β7-integrin and CCR9 (FIG. 4).

Tissue-homing-receptor reprogramming Previous studies have shown that human CD4+ T-cell memory to rotavirus infection is found mainly in the circulating α4β7-integrin+ T-cell population80 . By contrast, circulating allergen-reactive T cells from patients with atopic dermatitis and allergic contact dermatitis , as well as skin-tropic virus-specific T cells, are found

mainly in the CLA+ T-cell population81,82 . Together, these results indicate that once imprinted, T cells maintain their gut- and skin-tropic phenotype. Nevertheless, two recent studies have shown that tissue-tropic memory T cells can be reprogrammed to express alternative homing receptors. That is, mouse CD8+ memory T cells that express markers of gut or skin tropism could be ‘re-educated’ in vitro to express the reciprocal homing receptors by restimulation with DCs from the reciprocal organ71,73. Such reprogramming might occur in vivo, as most circulating human α4β7-integrin+ and CLA+ T cells express both CCR7 and L-selectin83 and are therefore likely to circulate through all lymph nodes, and memory T cells might exit the peripheral tissues and re-enter the circulation through the lymph84,85 . This flexibility in homing commitment might have an important role in redirecting immune responses to less tissue-restricted pathogens and will be important to bear in mind when optimizing vaccination regimes.

Imprinting of dendritic cells Mouse DCs can be separated into discrete subsets based on their expression of cell-surface markers, such as CD4, CD8, CD11b and B220. CD8α+ and CD8α– DCs from mesenteric lymph nodes are both capable of generating α4β7-integrin+CCR9+CD8+ T cells37. Moreover, CD11c lowB220 +, CD8α hiCD11b low, CD8α –CD11b low and CD8α low/–CD11b low DC subpopulations from Peyer’s patches all induced higher levels of α4β7-integrin expression on responding CD8+ T cells, compared with their equivalent populations from peripheral lymph nodes, whereas the reverse was true for the induction of ligands for E-selectin and P-selectin73 . Therefore, based on conventional DC markers, there seem to be few differences in the ability of DC subsets from a given lymphoid tissue to generate tissue-tropic effector T cells. Together, these results indicate an important role for the environment of the lymph node in imprinting DCs with the ability to generate tissue-tropic T-cell subsets. However, DCs from the lamina propria of the small intestine and Langerhans cells are superior in their ability to induce responding CD8+ T cells to express CCR9 and ligands for E-selectin, respectively, compared with DCs from mesenteric lymph nodes and peripheral lymph nodes54,71. Therefore, DCs can be imprinted with the ability to generate tissue-tropic T-cell subsets before entering a draining lymph node. Moreover, Calzascia et al.53 , when examining the response of two different tumour-antigen-specific, TCR-transgenic T-cell populations to two distinct tumour cells (one injected intraperitoneally and the other subcutaneously) in the same mouse, found that multiple homing phenotypes could be generated simultaneously in the same lymph node53 . These results indicate that the identity of the lymph node is not crucial for determining the tissue-homing-receptor profile induced on responding T cells, but rather that it is the site where DCs acquire the antigen that controls this response. My laboratory recently provided direct evidence that the ability to generate gut-tropic T cells is not a property of all mesenteric-lymph-node DCs. Mesenteric-lymph-node VOLU M E 6 | SEP T E M BER 20 0 6 | 687

NATURE REVIEWS | IMMUNOLOGY © 2006 Nature Publishing Group

REVIEWS

Smallintestinal lumen

a ‘Imprinting’ signal

Lamina-propria lymphocyte Intraepithelial lymphocyte

CD103+ DC

CD103+ DC

b DC precursor

Afferent lymph vessel

Epithelial cell

Pre-imprinted DC precursor Blood vessel

Basement membrane CD103– DC?

Afferent lymph vessel

CD103+ DC

Naive T cell

CD103+ DC

Primed gut-tropic T cell CCR9 Circulation α4β7-integrinhi

B-cell follicle B

Medulla Mesenteric lymph node

CD103– DC

Naive T cell

Primed CCR9– T cell Function? α4β7-integrinlow

Figure 5 | The role of CD103+ dendritic cells in the generation of gut-tropic effector T cells. CD8+ and CD4+ T cells are induced to express CC-chemokine receptor 9 (CCR9) and high levels of α4β7-integrin after activation with CD103+, but not CD103–, mesenteric-lymph-node dendritic cells (DCs). Most small-intestinal lamina-propria DCs express CD103, and DCs derived from the small-intestinal lamina propria are more potent at generating gut-tropic CD8+ effector T cells than mesenteric-lymph-node DCs. This indicates that DCs are imprinted with the ability to generate gut-tropic T cells before they enter the mesenteric lymph nodes. a | One possibility is that CD103+ DCs or their precursors are imprinted with the ability to generate gut-tropic T cells after they enter the intestinal mucosa. b | Alternatively, precursors of CD103+ lamina-propria DCs might receive imprinting signals before they enter the intestinal mucosa and selectively migrate to this tissue. It is also currently unknown whether both CD103+ and CD103– lamina-propria DCs are imprinted with the ability to generate gut-tropic T cells and only CD103+ lamina-propria DCs drain to the mesenteric lymph nodes, or whether both populations drain to the mesenteric lymph nodes and only CD103+ lamina-propria DCs are imprinted (potentially owing to differences in their localization within the lamina propria and/or because they derive from different precursor populations).

Afferent lymph vessels Lymph vessels that drain interstitial fluid and immune cells (mainly dendritic cells and memory lymphocytes) from tissues to the draining lymph nodes.

DCs expressing CD103 (also known as αE-integrin), which make up about 40% of DCs in the mesenteric lymph node, and CD103 – mesenteric-lymph-node DCs were equally capable of priming CD4+ and CD8+ T cells in vitro, but only CD103+ DCs could induce the responding T cells to express CCR9. In addition, both DC populations induced the responding T cells to express α4β7-integrin, but CD103+ DCs did so to a greater extent54,86 . There is continual homeostatic migration of DCs to the mesenteric lymph nodes from the lamina propria that can be dramatically increased by administration of adjuvant87, and several lines of evidence indicate that a large proportion of CD103+ mesenteric-lymph-node DCs derive from this site. First, most mouse lamina-propria-derived DCs and a large proportion of DCs in the rat gut-draining afferent lymph vessels express CD103 (REFS 54,88). Second, DC

688 | SEP T E M BER 20 0 6 | VOLU M E 6

migration from the skin and intestinal lamina propria to draining lymph nodes is CCR7 dependent89–91, and T cells primed in the mesenteric lymph nodes of CCR7deficient mice following intraperitoneal antigen administration — in contrast to wild-type mice — failed to adopt a gut-tropic phenotype54 . Last, although total DC numbers are reduced in the mesenteric lymph nodes of CCR7-deficient mice, there is a greater reduction in the CD103+ DC population than other DC populations54,89,90 . Together, these results indicate that the tissue environment from which the DCs migrated, rather than the draining lymph node, has a role in imprinting DCs with the ability to generate tissue-tropic T cells. An additional possibility that is yet to be explored is that subsets of pre-imprinted DC precursors from the bone marrow selectively localize to distinct peripheral tissues (FIG. 5). www.nature.com/reviews/immunol

© 2006 Nature Publishing Group

REVIEWS Of note, antigen-pulsed bone-marrow-derived DCs (BMDCs) injected intraperitoneally can generate α4β7-integrin+ effector T cells in the mesenteric lymph nodes, although expression of CCR9 was not assessed in this study 72 . Although the mechanism by which intraperitoneally injected BMDCs gain entry to the mesenteric lymph nodes remains to be determined, one interpretation of these results is that BMDCs are imprinted in the peritoneum and/or the mesenteric lymph nodes with the ability to induce responding T cells to express α4β7-integrin. However, given that soluble factors seem to be crucial for the induction of CCR9 and α4β7-integrin expression on responding T cells, these mediators might be generated in sufficient quantities in the mesenteric lymph node to function in trans at the time of T-cell priming by injected BMDCs.

Non-obese diabetic (NOD) mice An inbred strain of mice that spontaneously develop T-cell-mediated autoimmune diabetes.

Primary sclerosing cholangitis Chronic liver disease that is caused by progressive inflammation-mediated destruction of the hepatic bile ducts. The disease is of unknown aetiology and can develop as an extra-intestinal complication of inflammatory bowel disease.

Parabiotic mice Surgically joined mice that share a common blood circulation.

Aberrant tissue-tropic effector T-cell trafficking Although the constitutive expression of MADCAM1 and CCL25 in the small intestine normally functions to direct α4β7-integrin+CCR9+ effector T cells to this compartment, both MADCAM1 and CCL25 are expressed in non-intestinal vascular beds under certain inflammatory conditions. For example, the expression of MADCAM1 is strongly induced on pancreatic islet vessels in non-obese diabetic (NOD) mice and coincides with the appearance of α 4β 7-integrin + T cells 92,93 . Furthermore, treatment with MADCAM1-specific antibody reduced diabetes development in NOD mice and lymphocyte accumulation in the pancreas94 . Initial priming of diabetogenic T cells in NOD mice has been proposed to take place in intestinal lymph nodes, and antibodies to MADCAM1 reduced the ability of these cells to induce diabetes following transfer to NOD × SCID recipients95,96 . Similarly, MADCAM1 is expressed by hepatic endothelial cells in chronically inflamed human liver, particularly in primary sclerosing cholangitis (PSC) and liver pathologies associated with inflammatory bowel disease 97,98 , and expression of CCL25 is selectively induced in hepatic endothelial cells in individuals with PSC99. This aberrant expression of MADCAM1 and CCL25 seems to have a functional role, as CCR9+α4β7-integrin+ liver-infiltrating T cells accumulate in individuals with PSC but not other chronic liver diseases99. Together, these studies indicate that tissue-tropic effector T cells might be redirected to alternative compartments under certain inflammatory conditions. Whether these cells have an important role in disease pathogenesis, and the identity of the signals driving the aberrant expression of otherwise tissue-specific homing molecules, remains to be determined. Immune surveillance by memory T cells Although the discussion so far has mainly focused on the mechanisms by which effector T cells enter extralymphoid tissues, large numbers of memory T cells are present in extralymphoid tissues, where they can persist long after infection has been cleared. Furthermore, memory T cells can be detected in the gut, lung and peripheral lymph2,100–102 , indicating that extralymphoid

memory T cells can recirculate. However, the rate at which memory T cells enter different extralymphoid tissues, the origin of these memory T cells and whether memory T-cell subsets use similar ‘area codes’ as effector T cells to gain entry to these tissues is poorly understood56 . Recently, Klonowski and co-workers103 showed that there is considerable diversity in the degree of memory T-cell circulation through extralymphoid tissues, as well as in their mechanism of entry to these sites. Using parabiotic mice and MHC tetramers to track the dissemination of vesicular-stomatitis-virus-specific memory CD8+ T cells, they found a rapid turnover of memory cells in the lungs, liver and bone marrow. So, there seems to be continual surveillance of these tissues by blood-borne memory T cells, although the origin of these tissue-infiltrating memory T cells remains unclear. Increased percentages of influenza-virus- and respiratory-syncytial-virus-specific memory CD8 + T cells are found in normal human lungs compared with the circulating pool of memory CD8+ T cells104 . Furthermore, memory CD8+ T cells resident in normal lungs and in lung-draining afferent lymph have a distinct phenotype compared with those in intestinal and cutaneous tissues and lymph48,100,105,106 . These findings suggest that there might be some preferential memory CD8+ T-cell-subset migration to, and/or retention in, the lungs. In a recent study, long-term maintenance of influenza-virus-specific CD8+ T cells in the mouse lung parenchyma and airways was found to depend on persistent presentation of residual viral antigen in the lung-draining lymph nodes and on continual replenishment by migrating memory T cells from the circulation107. In contrast to the lungs, liver and bone marrow, homing of memory T cells to the brain and intestine seems to be extremely limited103 . This indicates that most of the T cells at these sites are long-term residents, probably generated from cells that entered during the effector stage of the immune response, and that only a small subset of blood-borne memory T cells exist with the correct ‘area code’ for these tissues. Indeed, memory T cells genetically deficient for β7-integrin show a reduced ability to enter the intestinal mucosa103, suggesting preferential entry of α4β7-integrin+ memory cells to this site. Although memory T-cell entry into the skin was not assessed in this study, a large proportion of non-cycling CLA+ memory T cells are resident in normal skin106,108. The chemokine receptor CCR8 is expressed by most of these cells and by a few circulating memory T cells, and its ligand, CCL1, is constitutively expressed by cells in the epidermis, including Langerhans cells and CD31+ dermal endothelial cells108 . These skin-resident T cells failed to migrate in response to the CCR10 ligand CCL27 and, although originally reported to be CCR4– (REF. 108), a subsequent study using different isolation techniques found CCR4 to be expressed by these cells106 . Therefore, immune surveillance of the skin in homeostatic conditions might be restricted to a small subset of blood-borne memory T cells, and is potentially mediated by CCR8 and/or CCR4, but not CCR10. VOLU M E 6 | SEP T E M BER 20 0 6 | 689

NATURE REVIEWS | IMMUNOLOGY © 2006 Nature Publishing Group

REVIEWS Immunotherapeutic aspects As our understanding of the mechanisms that regulate the generation of tissue-tropic effector T cells slowly unravels, the potential for using this knowledge for therapeutic purposes becomes apparent. Several possible scenarios can be envisioned in which manipulating the tissue-homing-receptor expression profile adopted by lymphocytes could have clinical application. First, vaccination: for example, most pathogens, including HIV, Vibrio cholerae and Salmonella enterica serovar Typhimurium, enter the body at mucosal surfaces and so a protective mucosal immune response to vaccination is often desired109. However, most current vaccines are administered intramuscularly or subcutaneously. Pharmacological agents that induce the expression of mucosal homing receptors on responding lymphocytes after vaccination at these sites might provide a means to increase vaccine efficacy for such pathogens. Alternatively, vaccines that target those DC subsets able to generate tissue-tropic lymphocytes could be used to increase lymphocyte localization to specific extralymphoid tissues. For example, targeting the CD103+ DC population in intestinal lymph nodes might function to increase protection against intestinal pathogens. Second, chronic inflammation: given the plasticity in tissue-homing-receptor expression by antigenexperienced T cells, pharmacological agents could be used to block DC-mediated induction of tissue-homingreceptor expression or modulate tissue-homing-receptor expression, redirecting pathogenic T-cell populations away from chronically inflamed tissues. A word of warning is warranted here, as such a strategy might potentially reduce access of T cells that are involved in protective immune surveillance, as recently highlighted by three cases of progressive multifocal leukoencephalopathy (a demyelinating disease of the central nervous system caused by the human polyomavirus JC virus) in patients with multiple sclerosis treated with natalizumab110–112 . Third, DC immunotherapy: DC immunization is a promising antitumour therapy113,114; however, recent data indicate a crucial role for the route of DC immunization in dictating the homing-receptor expression profile adopted by responding T cells71,72 . In a mouse melanoma model, intravenous and subcutaneous DC immunization induced immunity to metastatic lung tumours, whereas only subcutaneous DC immunization could control subcutaneous tumours115 , indicating that this has important relevance for therapeutic efficacy in vivo. Antigen-pulsed DCs, when injected intraperitoneally or subcutaneously, generate gut- and skin-tropic T cells, respectively, irrespective of the origin of the DCs71,72 .

1.

2.

3.

von Andrian, U. H. & Mempel, T. R. Homing and cellular traffic in lymph nodes. Nature Rev. Immunol. 3, 867–878 (2003). Cahill, R. N., Poskitt, D. C., Frost, D. C. & Trnka, Z. Two distinct pools of recirculating T lymphocytes: migratory characteristics of nodal and intestinal T lymphocytes. J. Exp. Med. 145, 420–428 (1977). Hall, J. G., Hopkins, J. & Orlans, E. Studies on the lymphocytes of sheep. III. Destination of lymph-borne immunoblasts in relation to their tissue of origin. Eur. J. Immunol. 7, 30–37 (1977).

4.

5.

6.

This indicates that DCs are re-imprinted at the site of immunization, or that dominant soluble factors in the draining lymph node might operate in trans at the time of T-cell priming. Identification of these imprinting and soluble factors might allow for conditioning of the immunization site before DC immunization. Last, T-cell immunotherapy: although adoptive transfer of tumour-specific T cells is also used as an antitumour therapy, a problem with this approach is the inefficiency of in vitro expanded T cells to localize to the relevant tumour site and draining lymph nodes116,117. Identification of the factors that regulate the induction of tissue-homing receptors on effector T cells might allow for the in vitro generation of tumour-specific T cells with an increased ability to target distinct tumours. The in vitro generation of gut-tropic effector T cells by addition of retinoic acid to antigen-primed T-cell cultures provides an example of this possibility.

Concluding remarks Although many homing receptors that are involved in effector T-cell localization to extralymphoid tissues have been defined, our understanding of the mechanisms that regulate their in vivo expression remains limited. Studies in the past few years have provided remarkable new insights into the cellular and molecular mechanisms that regulate the generation of tissue-tropic effector T cells, and have highlighted a central role for DCs and the tissue environment in this process. Nevertheless, several outstanding issues remain. First, and most importantly, our current understanding of the mechanisms that generate tissue-tropic T cells is based on mouse models, and confirmation that similar processes occur in humans is urgently required. Second, it will be important to address whether the ability to generate distinct tissue-tropic effector T-cell subsets is unique to the intestine and skin or a property of all tissue-draining lymph nodes. Third, it is necessary to identify the sites where DCs and/or their precursors are imprinted with the ability to generate tissue-tropic effector T-cell subsets, as well as the molecular mechanisms underlying this process. Fourth, we still require a greater understanding of the molecular signals generated by DC subsets that regulate homingreceptor expression by responding lymphocytes. Last, it will be important to determine the effect of inflammation and infection on the site-specific induction of tissuehoming receptor expression. Solving these key issues should increase our understanding of tissue-specific immune and inflammatory processes, and will allow for more focused strategies to modulate these processes in clinical settings.

McDermott, M. R. & Bienenstock, J. Evidence for a common mucosal immunologic system. I. Migration of B immunoblasts into intestinal, respiratory, and genital tissues. J. Immunol. 122, 1892–1898 (1979). Butcher, E. C., Williams, M., Youngman, K., Rott, L. & Briskin, M. Lymphocyte trafficking and regional immunity. Adv. Immunol. 72, 209–253 (1999). Berlin, C. et al. α4β7 Integrin mediates lymphocyte binding to the mucosal vascular addressin MAdCAM-1. Cell 74, 185–195 (1993).

690 | SEP T E M BER 20 0 6 | VOLU M E 6

7.

8.

9.

Briskin, M. et al. Human mucosal addressin cell adhesion molecule-1 is preferentially expressed in intestinal tract and associated lymphoid tissue. Am. J. Pathol. 151, 97–110 (1997). Hamann, A., Andrew, D. P., Jablonski-Westrich, D., Holzmann, B. & Butcher, E. C. Role of α4-integrins in lymphocyte homing to mucosal tissues in vivo. J. Immunol. 152, 3282–3293 (1994). 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).

www.nature.com/reviews/immunol © 2006 Nature Publishing Group

REVIEWS 10. Haddad, W. et al. P-selectin and P-selectin glycoprotein ligand 1 are major determinants for TH1 cell recruitment to nonlymphoid effector sites in the intestinal lamina propria. J. Exp. Med. 198, 369–377 (2003). 11. Picarella, D. et al. Monoclonal antibodies specific for β7 integrin and mucosal addressin cell adhesion molecule-1 (MAdCAM-1) reduce inflammation in the colon of scid mice reconstituted with CD45RBhigh CD4+ T cells. J. Immunol. 158, 2099–2106 (1997). 12. Kato, S. et al. Amelioration of murine experimental colitis by inhibition of mucosal addressin cell adhesion molecule-1. J. Pharmacol. Exp. Ther. 295, 183–189 (2000). 13. Hesterberg, P. E. et al. Rapid resolution of chronic colitis in the cotton-top tamarin with an antibody to a gut-homing integrin α4β7. Gastroenterology 111, 1373–1380 (1996). 14. Feagan, B. G. et al. Treatment of ulcerative colitis with a humanized antibody to the α4β7 integrin. N. Engl. J. Med. 352, 2499–2507 (2005). 15. Matsuzaki, K. et al. In vivo demonstration of T lymphocyte migration and amelioration of ileitis in intestinal mucosa of SAMP1/Yit mice by the inhibition of MAdCAM-1. Clin. Exp. Immunol. 140, 22–31 (2005). 16. Rivera-Nieves, J. et al. L-selectin, α4β1, and α4β7 integrins participate in CD4+ T cell recruitment to chronically inflamed small intestine. J. Immunol. 174, 2343–2352 (2005). 17. Kuklin, N. A. et al. α4β7 Independent pathway for CD8+ T cell-mediated intestinal immunity to rotavirus. J. Clin. Invest. 106, 1541–1552 (2000). 18. Matsumoto, M. et al. CD43 functions as a ligand for E-selectin on activated T cells. J. Immunol. 175, 8042–8050 (2005). 19. Fuhlbrigge, R. C., King, S. L., Sackstein, R. & Kupper, T. S. CD43 is a ligand for E-selectin on CLA+ human T cells. Blood 107, 1421–1426 (2005). 20. Picker, L. J., Michie, S. A., Rott, L. S. & Butcher, E. C. A unique phenotype of skin-associated lymphocytes in humans. Preferential expression of the HECA-452 epitope by benign and malignant T cells at cutaneous sites. Am. J. Pathol. 136, 1053–1068 (1990). 21. Weninger, W. et al. Specialized contributions by α(1,3)-fucosyltransferase-IV and FucT-VII during leukocyte rolling in dermal microvessels. Immunity 12, 665–676 (2000). 22. Austrup, F. et al. P- and E-selectin mediate recruitment of T-helper-1 but not T-helper-2 cells into inflammed tissues. Nature 385, 81–83 (1997). 23. Picker, L. J., Kishimoto, T. K., Smith, C. W., Warnock, R. A. & Butcher, E. C. ELAM-1 is an adhesion molecule for skin-homing T cells. Nature 349, 796–799 (1991). 24. Reinhardt, R. L., Bullard, D. C., Weaver, C. T. & Jenkins, M. K. Preferential accumulation of antigenspecific effector CD4 T cells at an antigen injection site involves CD62E-dependent migration but not local proliferation. J. Exp. Med. 197, 751–762 (2003). 25. Tietz, W. et al. CD4+ T cells migrate into inflamed skin only if they express ligands for E- and P-selectin. J. Immunol. 161, 963–970 (1998). 26. Hirata, T., Furie, B. C. & Furie, B. P-, E-, and L-selectin mediate migration of activated CD8+ T lymphocytes into inflamed skin. J. Immunol. 169, 4307–4313 (2002). 27. Symon, F. A., McNulty, C. A. & Wardlaw, A. J. P- and L-selectin mediate binding of T cells to chronically inflamed human airway endothelium. Eur. J. Immunol. 29, 1324–1333 (1999). 28. Bonder, C. S. et al. P-selectin can support both TH1 and TH2 lymphocyte rolling in the intestinal microvasculature. Am. J . Pathol. 167, 1647–1660 (2005). 29. Ley, K. & Kansas, G. S. Selectins in T-cell recruitment to non-lymphoid tissues and sites of inflammation. Nature Rev. Immunol. 4, 325–335 (2004). 30. Piccio, L. et al. Efficient recruitment of lymphocytes in inflamed brain venules requires expression of cutaneous lymphocyte antigen and fucosyltransferase-VII. J. Immunol. 174, 5805–5813 (2005). 31. Kunkel, E. J. & Butcher, E. C. Chemokines and the tissue-specific migration of lymphocytes. Immunity 16, 1–4 (2002). 32. 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

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

principle in regional immunity. J. Exp. Med. 192, 761–768 (2000). This paper was the first to highlight potential mechanistic differences in T-cell trafficking to the small intestine and colon, and to suggest a role for CCR9 specifically in small-intestinal immunity. Papadakis, K. A. et al. CCR9-positive lymphocytes and thymus-expressed chemokine distinguish small bowel from colonic Crohn’s disease. Gastroenterology 121, 246–254 (2001). Svensson, M. et al. CCL25 mediates the localization of recently activated CD8αβ+ lymphocytes to the smallintestinal mucosa. J. Clin. Invest. 110, 1113–1121 (2002). This study was the first to show that CCR9 expression is selectively induced on CD8+ T cells activated in intestinal lymph nodes, and to show an in vivo role for CCL25 in mediating CD8+ T-cell localization to the small-intestinal epithelium. Zabel, B. A. et al. Human G protein-coupled receptor GPR-9–6/CC chemokine receptor 9 is selectively expressed on intestinal homing T lymphocytes, mucosal lymphocytes, and thymocytes and is required for thymus-expressed chemokine-mediated chemotaxis. J. Exp. Med. 190, 1241–1256 (1999). Wurbel, M. A. et al. The chemokine TECK is expressed by thymic and intestinal epithelial cells and attracts double- and single-positive thymocytes expressing the TECK receptor CCR9. Eur. J. Immunol. 30, 262–271 (2000). Johansson-Lindbom, B. et al. Selective generation of gut tropic T cells in gut-associated lymphoid tissue (GALT): requirement for GALT dendritic cells and adjuvant. J. Exp. Med. 198, 963–969 (2003). Stenstad, H. et al. Gut associated lymphoid tissue primed CD4+ T cells display CCR9 dependent and independent homing to the small intestine. Blood 107, 3447–3454 (2006). Uehara, S., Grinberg, A., Farber, J. M. & Love, P. E. A role for CCR9 in T lymphocyte development and migration. J. Immunol. 168, 2811–2819 (2002). Wurbel, M. A. et al. Mice lacking the CCR9 CCchemokine receptor show a mild impairment of early T- and B-cell development and a reduction in T-cell receptor γδ+ gut intraepithelial lymphocytes. Blood 98, 2626–2632 (2001). Campbell, J. J. et al. The chemokine receptor CCR4 in vascular recognition by cutaneous but not intestinal memory T cells. Nature 400, 776–780 (1999). Homey, B. et al. CCL27–CCR10 interactions regulate T cell-mediated skin inflammation. Nature Med. 8, 157–165 (2002). This study was the first to implicate a role for CCR10 in effector T-cell localization to inflamed cutaneous sites. Soler, D., Humphreys, T. L., Spinola, S. M. & Campbell, J. J. CCR4 versus CCR10 in human cutaneous TH lymphocyte trafficking. Blood 101, 1677–1682 (2003). Morales, J. et al. CTACK, a skin-associated chemokine that preferentially attracts skin-homing memory T cells. Proc. Natl Acad. Sci. USA 96, 14470–14475 (1999). Hudak, S. et al. Immune surveillance and effector functions of CCR10+ skin homing T cells. J. Immunol. 169, 1189–1196 (2002). Reiss, Y., Proudfoot, A. E., Power, C. A., Campbell, J. J. & Butcher, E. C. CC chemokine receptor (CCR)4 and the CCR10 ligand cutaneous T cell-attracting chemokine (CTACK) in lymphocyte trafficking to inflamed skin. J. Exp. Med. 194, 1541–1547 (2001). 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). Kunkel, E. J. et al. Expression of the chemokine receptors CCR4, CCR5, and CXCR3 by human tissue-infiltrating lymphocytes. Am. J. Pathol. 160, 347–355 (2002). Panina-Bordignon, P. et al. The C-C chemokine receptors CCR4 and CCR8 identify airway T cells of allergen-challenged atopic asthmatics. J. Clin. Invest. 107, 1357–1364 (2001). Campbell, D. J., Kim, C. H. & Butcher, E. C. Separable effector T cell populations specialized for B cell help or tissue inflammation. Nature Immunol. 2, 876–881 (2001). Picker, L. J. et al. Control of lymphocyte recirculation in man. II. Differential regulation of the cutaneous lymphocyte-associated antigen, a tissue-selective homing receptor for skin-homing T cells. J. Immunol. 150, 1122–1136 (1993).

52. Campbell, D. J. & Butcher, E. C. Rapid acquisition of tissue-specific homing phenotypes by CD4+ T cells activated in cutaneous or mucosal lymphoid tissues. J. Exp. Med. 195, 135–141 (2002). This paper, together with reference 34, shows that T cells rapidly acquire intestinal- and skin-homing receptors in vivo following their activation in intestinal- and skin-draining lymph nodes, respectively. 53. Calzascia, T. et al. Homing phenotypes of tumorspecific CD8 T cells are predetermined at the tumor site by crosspresenting APCs. Immunity 22, 175–184 (2005). This article shows that T cells activated in cervical lymph nodes have an increased ability to migrate to the central nervous system. The study also indicates that the homing phenotype adopted by the responding T cells depends on the site where the antigen-presenting cell acquires antigen. 54. Johansson-Lindbom, B. et al. Functional specialization of gut CD103+ dendritic cells in the regulation of tissue-selective T cell homing. J. Exp. Med. 202, 1063–1073 (2005). 55. Catron, D. M., Rusch, L. K., Hataye, J., Itano, A. A. & Jenkins, M. K. CD4+ T cells that enter the draining lymph nodes after antigen injection participate in the primary response and become central-memory cells. J. Exp. Med. 203, 1045–1054 (2006). 56. Marzo, A. L. et al. Initial T cell frequency dictates memory CD8+ T cell lineage commitment. Nature Immunol. 6, 793–799 (2005). 57. Engelhardt, B. & Ransohoff, R. M. The ins and outs of T-lymphocyte trafficking to the CNS: anatomical sites and molecular mechanisms. Trends Immunol. 26, 485–495 (2005). 58. Luster, A. D., Alon, R. & von Andrian, U. H. Immune cell migration in inflammation: present and future therapeutic targets. Nature Immunol. 6, 1182–1190 (2005). 59. Luster, A. D. & Tager, A. M. T-cell trafficking in asthma: lipid mediators grease the way. Nature Rev. Immunol. 4, 711–724 (2004). 60. Reinhardt, R. L., Khoruts, A., Merica, R., Zell, T. & Jenkins, M. K. Visualizing the generation of memory CD4 T cells in the whole body. Nature 410, 101–105 (2001). 61. Masopust, D. et al. Activated primary and memory CD8 T cells migrate to nonlymphoid tissues regardless of site of activation or tissue of origin. J. Immunol. 172, 4875–4882 (2004). 62. Masopust, D., Vezys, V., Marzo, A. L. & Lefrancois, L. Preferential localization of effector memory cells in nonlymphoid tissue. Science 291, 2413–2417 (2001). 63. John, B. & Crispe, I. N. Passive and active mechanisms trap activated CD8+ T cells in the liver. J. Immunol. 172, 5222–5229 (2004). 64. Galkina, E. et al. Preferential migration of effector CD8+ T cells into the interstitium of the normal lung. J. Clin. Invest. 115, 3473–3483 (2005). 65. Laschinger, M., Vajkoczy, P. & Engelhardt, B. Encephalitogenic T cells use LFA-1 for transendothelial migration but not during capture and initial adhesion strengthening in healthy spinal cord microvessels in vivo. Eur. J. Immunol. 32, 3598–3606 (2002). 66. Vajkoczy, P., Laschinger, M. & Engelhardt, B. α4-integrin–VCAM-1 binding mediates G proteinindependent capture of encephalitogenic T cell blasts to CNS white matter microvessels. J. Clin. Invest. 108, 557–565 (2001). 67. Miller, D. H. et al. A controlled trial of natalizumab for relapsing multiple sclerosis. N. Engl. J. Med. 348, 15–23 (2003). 68. Polman, C. H. et al. A randomized, placebo-controlled trial of natalizumab for relapsing multiple sclerosis. N. Engl. J. Med. 354, 899–910 (2006). 69. Mora, J. R. et al. Selective imprinting of gut-homing T cells by Peyer’s patch dendritic cells. Nature 424, 88–93 (2003). 70. Stagg, A. J., Kamm, M. A. & Knight, S. C. Intestinal dendritic cells increase T cell expression of α4β7 integrin. Eur. J. Immunol. 32, 1445–1454 (2002). This paper, together with references 37 and 69, demonstrates a crucial role for intestinal DCs in the generation of gut-tropic effector T-cell subsets. 71. Dudda, J. C. et al. Dendritic cells govern induction and reprogramming of polarized tissue-selective homing receptor patterns of T cells: important roles for soluble factors and tissue microenvironments. Eur. J. Immunol. 35, 1056–1065 (2005).

VOLU M E 6 | SEP T E M BER 20 0 6 | 691

NATURE REVIEWS | IMMUNOLOGY © 2006 Nature Publishing Group

REVIEWS 72. Dudda, J. C., Simon, J. C. & Martin, S. Dendritic cell immunization route determines CD8+ T cell trafficking to inflamed skin: role for tissue microenvironment and dendritic cells in establishment of T cell-homing subsets. J. Immunol. 172, 857–863 (2004). This paper shows that antigen-pulsed DCs isolated from the skin or skin-draining lymph nodes induce T cells to express skin-homing receptors. 73. Mora, J. R. et al. Reciprocal and dynamic control of CD8 T cell homing by dendritic cells from skin- and gut-associated lymphoid tissues. J. Exp. Med. 201, 303–316 (2005). 74. Iwata, M. et al. Retinoic acid imprints gut-homing specificity on T cells. Immunity 21, 527–538 (2004). This paper shows an important role for retinoic acid in the generation of gut-tropic T cells, and thereby provides the first mechanistic insights as to how intestinal-lymph-node DCs induce the expression of gut-homing receptors on responding T cells. 75. Malmstrom, V. et al. CD134L expression on dendritic cells in the mesenteric lymph nodes drives colitis in T cell-restored SCID mice. J. Immunol. 166, 6972–6981 (2001). 76. Carramolino, L. et al. Expression of CCR9 β-chemokine receptor is modulated in thymocyte differentiation and is selectively maintained in CD8+ T cells from secondary lymphoid organs. Blood 97, 850–857 (2001). 77. Ericsson, A., Svensson, M., Arya, A. & Agace, W. W. CCL25/CCR9 promotes the induction and function of CD103 on intestinal intraepithelial lymphocytes. Eur. J. Immunol. 34, 2720–2729 (2004). 78. Andrew, D. P., Rott, L. S., Kilshaw, P. J. & Butcher, E. C. Distribution of α4β7 and αEβ7 integrins on thymocytes, intestinal epithelial lymphocytes and peripheral lymphocytes. Eur. J. Immunol. 26, 897–905 (1996). 79. Wagers, A. J., Waters, C. M., Stoolman, L. M. & Kansas, G. S. Interleukin 12 and interleukin 4 control T cell adhesion to endothelial selectins through opposite effects on α1,3-fucosyltransferase VII gene expression. J. Exp. Med. 188, 2225–2231 (1998). 80. Rott, L. S. et al. Expression of mucosal homing receptor α4β7 by circulating CD4+ cells with memory for intestinal rotavirus. J. Clin. Invest. 100, 1204–1208 (1997). 81. Koelle, D. M. et al. Expression of cutaneous lymphocyte-associated antigen by CD8+ T cells specific for a skin-tropic virus. J. Clin. Invest. 110, 537–548 (2002). 82. Santamaria Babi, L. F. et al. Circulating allergenreactive T cells from patients with atopic dermatitis and allergic contact dermatitis express the skinselective homing receptor, the cutaneous lymphocyteassociated antigen. J. Exp. Med. 181, 1935–1940 (1995). 83. Campbell, J. J. et al. CCR7 expression and memory T cell diversity in humans. J. Immunol. 166, 877–884 (2001). 84. Bromley, S. K., Thomas, S. Y. & Luster, A. D. Chemokine receptor CCR7 guides T cell exit from peripheral tissues and entry into afferent lymphatics. Nature Immunol. 6, 895–901 (2005). 85. Debes, G. F. et al. Chemokine receptor CCR7 required for T lymphocyte exit from peripheral tissues. Nature Immunol. 6, 889–894 (2005). 86. Annacker, O. et al. Essential role for CD103 in the T cell-mediated regulation of experimental colitis. J. Exp. Med. 202, 1051–1061 (2005). 87. MacPherson, G. G., Jenkins, C. D., Stein, M. J. & Edwards, C. Endotoxin-mediated dendritic cell release from the intestine. Characterization of released dendritic cells and TNF dependence. J. Immunol. 154, 1317–1322 (1995). 88. Brenan, M. & Puklavec, M. The MRC OX-62 antigen: a useful marker in the purification of rat veiled cells with the biochemical properties of an integrin. J. Exp. Med. 175, 1457–1465 (1992).

89. Worbs, T. et al. Oral tolerance originates in the intestinal immune system and relies on antigen carriage by dendritic cells. J. Exp. Med. 203, 519–527 (2006). 90. Jang, M. H. et al. CCR7 is critically important for migration of dendritic cells in intestinal lamina propria to mesenteric lymph nodes. J. Immunol. 176, 803–810 (2006). 91. Ohl, L. et al. CCR7 governs skin dendritic cell migration under inflammatory and steady-state conditions. Immunity 21, 279–288 (2004). 92. Hanninen, A. et al. Vascular addressins are induced on islet vessels during insulitis in nonobese diabetic mice and are involved in lymphoid cell binding to islet endothelium. J. Clin. Invest. 92, 2509–2515 (1993). 93. Hanninen, A., Salmi, M., Simell, O. & Jalkanen, S. Mucosa-associated (β7-integrinhigh) lymphocytes accumulate early in the pancreas of NOD mice and show aberrant recirculation behavior. Diabetes 45, 1173–1180 (1996). 94. Yang, X. D., Sytwu, H. K., McDevitt, H. O. & Michie, S. A. Involvement of β7 integrin and mucosal addressin cell adhesion molecule-1 (MAdCAM-1) in the development of diabetes in obese diabetic mice. Diabetes 46, 1542–1547 (1997). 95. Jaakkola, I., Jalkanen, S. & Hanninen, A. Diabetogenic T cells are primed both in pancreatic and gut-associated lymph nodes in NOD mice. Eur. J. Immunol. 33, 3255–3264 (2003). 96. Hanninen, A., Jaakkola, I. & Jalkanen, S. Mucosal addressin is required for the development of diabetes in nonobese diabetic mice. J. Immunol. 160, 6018–6025 (1998). 97. Grant, A. J., Lalor, P. F., Hubscher, S. G., Briskin, M. & Adams, D. H. MAdCAM-1 expressed in chronic inflammatory liver disease supports mucosal lymphocyte adhesion to hepatic endothelium (MAdCAM-1 in chronic inflammatory liver disease). Hepatology 33, 1065–1072 (2001). 98. Hillan, K. J. et al. Expression of the mucosal vascular addressin, MAdCAM-1, in inflammatory liver disease. Liver 19, 509–518 (1999). 99. Eksteen, B. et al. Hepatic endothelial CCL25 mediates the recruitment of CCR9+ gut-homing lymphocytes to the liver in primary sclerosing cholangitis. J. Exp. Med. 200, 1511–1517 (2004). This paper provides an intriguing example of how tissue-tropic T-cell populations can be redirected under certain inflammatory conditions by the aberrant expression of tissue-homing ligands. 100. Abitorabi, M. A. et al. Differential expression of homing molecules on recirculating lymphocytes from sheep gut, peripheral, and lung lymph. J. Immunol. 156, 3111–3117 (1996). 101. Mackay, C. R., Marston, W. L. & Dudler, L. Naive and memory T cells show distinct pathways of lymphocyte recirculation. J. Exp. Med. 171, 801–817 (1990). 102. Mackay, C. R. et al. Tissue-specific migration pathways by phenotypically distinct subpopulations of memory T cells. Eur. J. Immunol. 22, 887–895 (1992). 103. Klonowski, K. D. et al. Dynamics of blood-borne CD8 memory T cell migration in vivo. Immunity 20, 551–562 (2004). This paper examines, for the first time, the dynamics of memory T-cell entry into extralymphoid tissues and highlights important differences in both the kinetics and the mechanisms by which this process occurs. 104. de Bree, G. J. et al. Selective accumulation of differentiated CD8+ T cells specific for respiratory viruses in the human lung. J. Exp. Med. 202, 1433–1442 (2005). 105. Campbell, J. J. et al. Expression of chemokine receptors by lung T cells from normal and asthmatic subjects. J. Immunol. 166, 2842–2848 (2001). 106. Clark, R. A. et al. The vast majority of CLA+ T cells are resident in normal skin. J. Immunol. 176, 4431–4439 (2006).

692 | SEP T E M BER 20 0 6 | VOLU M E 6

107. Zammit, D. J., Turner, D. L., Klonowski, K. D., Lefrancois, L. & Cauley, L. S. Residual antigen presentation after influenza virus infection affects CD8 T cell activation and migration. Immunity 24, 439–449 (2006). 108. Schaerli, P. et al. A skin-selective homing mechanism for human immune surveillance T cells. J. Exp. Med. 199, 1265–1275 (2004). This paper highlights a potential role for CCR8 in the immunosurveillance of cutaneous tissues by memory T cells. 109. Neutra, M. R. & Kozlowski, P. A. Mucosal vaccines: the promise and the challenge. Nature Rev. Immunol. 6, 148–158 (2006). 110. Van Assche, G. et al. Progressive multifocal leukoencephalopathy after natalizumab therapy for Crohn’s disease. N. Engl. J. Med. 353, 362–368 (2005). 111. Kleinschmidt-DeMasters, B. K. & Tyler, K. L. Progressive multifocal leukoencephalopathy complicating treatment with natalizumab and interferon β-1a for multiple sclerosis. N. Engl. J. Med. 353, 369–374 (2005). 112. Langer-Gould, A., Atlas, S. W., Green, A. J., Bollen, A. W. & Pelletier, D. Progressive multifocal leukoencephalopathy in a patient treated with natalizumab. N. Engl. J. Med. 353, 375–381 (2005). 113. Banchereau, J. & Palucka, A. K. Dendritic cells as therapeutic vaccines against cancer. Nature Rev. Immunol. 5, 296–306 (2005). 114. O’Neill, D. W., Adams, S. & Bhardwaj, N. Manipulating dendritic cell biology for the active immunotherapy of cancer. Blood 104, 2235–2246 (2004). 115. Mullins, D. W. et al. Route of immunization with peptide-pulsed dendritic cells controls the distribution of memory and effector T cells in lymphoid tissues and determines the pattern of regional tumor control. J. Exp. Med. 198, 1023–1034 (2003). 116. Dudley, M. E. & Rosenberg, S. A. Adoptive-celltransfer therapy for the treatment of patients with cancer. Nature Rev. Cancer 3, 666–675 (2003). 117. Kershaw, M. H., Teng, M. W., Smyth, M. J. & Darcy, P. K. Supernatural T cells: genetic modification of T cells for cancer therapy. Nature Rev. Immunol. 5, 928–940 (2005). 118. Lehmann, J. C. et al. Overlapping and selective roles of endothelial intercellular adhesion molecule-1 (ICAM-1) and ICAM-2 in lymphocyte trafficking. J. Immunol. 171, 2588–2593 (2003).

Acknowledgements I thank all present and former colleagues and collaborators for their contributions to this work and in particular B. Johansson-Lindbom, F. Ivars and C. Johansson for their valuable comments during the preparation of this manuscript. I also apologize to those researchers in the field whose important contributions I have been unable to cite because of space constraints. This work was supported by grants from the Swedish Medical Research Council, the Wellcome Trust, the Crafoordska, Österlund, Åke Wiberg, Richard and Ruth Julins, Nanna Svartz and Kocks foundations, the Royal Physiographic Society, and the Swedish foundation for Strategic Research ‘Microbes and Man’ and INGVAR II programmes.

Competing financial interests The author declares no competing financial interests.

DATABASES The following terms in this article are linked online to: Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?db=gene CCL25 | CCR4 | CCR8 | CCR9 | CCR10 | CD103 | CXCR5 | ICAM1 | LFA1 | MADCAM1 | VCAM1 Access to this links box is available online.

www.nature.com/reviews/immunol © 2006 Nature Publishing Group