Retinoic Acid

Immunity 458

autoreactive T cells that escaped thymic selection due to Id3 deficiency. Nonetheless, it is puzzling why Id3deficient mice develop autoimmune lesions only in exocrine glands, while Id3 is expressed in a variety of cell types. Redundancy between the various Id molecules could be an explanation for this. The Id3⫺/⫺ model for Sjo¨gren’s Syndrome will allow detailed investigation of the role of target organ abnormalities in autoimmunity and in particular, it may give insight into the initiation phase of the disease. A key question, of course is whether there are abnormalities in Id3 expression in patients with Sjo¨gren’s Syndrome and whether these abnormalities could serve as markers for Sjo¨gren’s Syndrome. In the end, it should be kept in mind that a single animal model may represent the disease process characteristic of only a subset of patient’s with this syndrome. If so, while this new model will provide new opportunities to study pathogenesis and potential therapies, the study of multiple animal models will be essential for a more complete understanding of this disease.

Selected Reading Cha, S., van Blokland, S.C., Versnel, M.A., Homo-Delarche, F., Nagashima, H., Brayer, J., Peck, A.B., and Humphreys-Beher, M.G. (2001). Exp. Clin. Immunogenet. 18, 143–160. de Heer, H.J., Hammad, H., Soullie, T., Hijdra, D., Vos, N., Willart, M.A., Hoogsteden, H.C., and Lambrecht, B.N. (2004). J. Exp. Med. 200, 89–98. Haneji, N., Hamano, H., Yanagi, K., and Hayashi, Y. (1994). J. Immunol. 153, 2769–2777. Heemskerk, M.H., Blom, B., Nolan, G., Stegmann, A.P., Bakker, A.Q., Weijer, K., Res, P.C., and Spits, H. (1997). J. Exp. Med. 186, 1597– 1602. Kowanetz, M., Valcourt, U., Bergstrom, R., Heldin, C.H., and Moustakas, A. (2004). Mol. Cell. Biol. 24, 4241–4254. Li, H., Dai, M., and Zhuang, Y. (2004). Immunity 21, this issue, 551–560. Pan, L., Sato, S., Frederick, J.P., Sun, X.H., and Zhuang, Y. (1999). Mol. Cell. Biol. 19, 5969–5980. Spits, H., Couwenberg, F., Bakker, A.Q., Weijer, K., and Uittenbogaart, C.H. (2000). J. Exp. Med. 192, 1775–1784. van Blokland, S.C., van Helden-Meeuwsen, C.G., Wierenga-Wolf, A.F., Drexhage, H.A., Hooijkaas, H., van de Merwe, J.P., and Versnel, M.A. (2000). Lab. Invest. 80, 575–585.

Marjan A. Versnel Department Immunology Erasmus MC Rotterdam The Netherlands

Retinoic Acid: An Educational “Vitamin Elixir” for Gut-Seeking T Cells

T cell priming by dendritic cells (DC) from gut-associated lymphoid tissues gives rise to effector cells with pronounced gut tropism. The mechanism for DC-dependent imprinting of gut specificity has remained unknown. New findings point to retinoic acid, which is uniquely produced by intestinal DC, but not by DC from other lymphoid organs (Iwata et al., 2004; this issue of Immunity). When naive T cells are activated by antigen they become effector/memory T cells, which acquire the capacity to migrate to extralymphoid tissues (von Andrian and Mackay, 2000). Effector cells that arise in response to antigens in the alimentary tract express mostly intestinal homing receptors, particularly the integrin ␣4␤7 and CCR9, the receptor for TECK/CCL25, a chemokine expressed in the small bowel. Consequently, these cells migrate preferentially to the small intestine, whereas T cell stimulation by cutaneous antigens induces effector cells expressing skin-homing receptors, such as carbohydrate ligands for endothelial P- and E-selectin and the chemokine receptors CCR4 and/or CCR10. In vivo studies have shown that the microenvironment in which T cells encounter antigen somehow instructs

them about their homing preference; upon systemic stimulation by antigen, activated T cells in mesenteric lymph nodes (MLN; receiving afferent lymph from the gut) upregulate gut-homing molecules, whereas those in peripheral lymph nodes (PLN, draining cutaneous tissues) express skin-homing receptors (Campbell and Butcher, 2002). This differential imprinting was reproduced in vitro by activating naive T cells with DC from different lymphoid organs. Thus, DC from Peyer’s patches (PP) or MLN, but not from PLN or spleen, induce effector T cells with high expression of ␣4␤7 and CCR9 and the capacity to migrate to the small bowel (Johansson-Lindbom et al., 2003; Mora et al., 2003; Stagg et al., 2002). However, until now, the mechanism(s) responsible for this imprinting of tissue specificity has remained a complete mystery. A groundbreaking article in this issue of Immunity identifies the first molecular mechanism for the imprinting of gut-homing T cells (Iwata et al., 2004). In an elegant series of experiments, Iwata and colleagues show that T cell exposure to subnanomolar concentrations of the vitamin A metabolite retinoic acid (RA) induced gut-homing receptors and the ability to migrate to the small intestine while simultaneously suppressing the expression of skin-homing molecules. These effects were reproduced with a synthetic agonist on RA receptors of the RAR isotype. Importantly, many DC from PP and MLN but few from PLN or the spleen expressed the prerequisite enzymes for oxidative conversion of vitamin A to RA, and inhibitors of these enzymes rendered intestinal DC incapable of inducing ␣4␤7high T cells. Consis-

Previews 459

Figure 1. Schematic Diagram of Putative Imprinting Signals that Direct Primed T Cells to the Skin or Gut When naive T cells are activated under the influence of retinoic acid (RA), they acquire a gut-homing phenotype with high expression levels of ␣4␤7 and CCR9. RA is produced by dendritic cells (DC) in gut-associated lymphoid tissues (Iwata et al., 2004) and possibly also by intestinal epithelium or other sources. Simultaneously, RA suppresses T cell acquisition of skin-homing molecules (E-/P-selectin ligands and CCR4). It has not yet been determined if DC from peripheral lymph nodes (PLN DC) generate specific imprinting signals that promote T cell homing to the skin or other nonintestinal organs.

tent with these in vitro experiments, vitamin A-deficient mice contained fewer ␣4␤7high effector/memory T cells than control animals. More importantly, T cells in the lamina propria and intraepithelial compartment in the small bowel of vitamin-A deficient mice were dramatically reduced in number. Pending independent confirmation, this work represents a major advance in our understanding of tissue-specific lymphocyte migration (Figure 1). Like all significant discoveries, these findings raise many new questions and intriguing possibilities. For example, although the current data show that vitamin A-derived RA is both necessary and sufficient to induce T cell homing to the small intestine under steady-state conditions, it remains to be determined if RA is required to generate gut-homing T cells in pathological settings. Indeed, viral infections generate effector cells that migrate to the intestine regardless of the site of activation or tissue of origin (Masopust et al., 2004). Perhaps inflammatory conditions induce RA-producing enzymes in DC that reside in nonintestinal compartments? Alternatively, it cannot be excluded that RA-independent mechanisms are evoked that promote gut tropism under these conditions. Of note in this regard, thymocytes and naive CD8 T cells express high levels of CCR9 but very low levels of ␣4␤7 (Carramolino et al., 2001; Mora et al., 2003). This dissociation of homing receptor expression supports the idea that there might be additional signals that may function separately or together with RA to regulate gut-homing receptors on T cells. While intestinal DC are clearly capable of producing RA, it is not clear if they are the only source of RA in vivo or if DC-derived RA is even necessary for T cell imprinting. The data by Iwata et al. (2004) indicate that intestinal epithelial cells express at least one relevant

enzyme, RALDH1, to metabolize vitamin A, but it was not determined if epithelial cells produce RA (Iwata et al., 2004). Conditional deletion of key enzymes for RA production in DC or epithelium could clarify this point. Another intriguing possibility is that RA produced by DC and/or epithelial cells in the gut may become lymph borne and contribute remotely to intestinal imprinting in draining MLN. This could explain why the in vivo induction of gut-homing T cells is highly efficient in MLN even though in our hands (and consistent with data shown by Iwata et al.), purified DC from MLN are less efficient at inducing ␣4␤7 than those from PP (our unpublished data). An analysis of mRNA levels for different isoenzymes involved in RA production further suggests that DC residing in MLN are surprisingly distinct from those in PP (Iwata et al., 2004). Since DC are thought to travel from PP and the lamina propria to MLN via the lymph, it will be interesting to assess if the relocating DC adjust their enzyme repertoire for RA production. Moreover, it remains to be determined if the differential expression of these isoenzymes has functional consequences for T cell imprinting. Experiments along these lines might also provide clues to the question how DC are “educated” to acquire a gut-imprinting potential. Although it is conceivable that DC or their precursors arrive in the gut already committed to a “gut-imprinting program,” it seems more likely that they acquire this property in the intestine. It will be interesting to elucidate the role of gutassociated cytokines, exposure to specific commensal microorganisms, toll-like receptor signals, and other environmental cues in this regard. What are the implications of the observations by Iwata et al. for our understanding of lymphocyte homing specificity in other tissues? For example, the colon, unlike

Immunity 460

the small intestine, is still poorly understood with regard to T cell recruitment. While ␣4␤7 is likely to play a role, it is clear that at least the chemokine signal(s) for this part of the intestine must be distinct from the CCL25/ CCR9 pathway, which predominates in the small bowel (Kunkel and Butcher, 2002; Mora et al., 2003). It will be interesting to investigate if RA or other vitamin A metabolites are involved in the generation of colon-homing T cells. RA not only upregulated gut-homing molecules on activated T cells but also simultaneously inhibited the induction of skin homing receptors (Iwata et al., 2004). This raises the possibility that T cells are biased toward a skin-homing phenotype whenever they are activated in the absence of RA. In other words, the acquisition of skin-homing molecules might be the default pathway in the absence of modulating signals like RA and possibly others that redirect T cells to the gut or elsewhere. However, it should be cautioned that one chemokine receptor implicated in skin homing, CCR10, is not induced under standard in vitro conditions for T cell activation, suggesting that the acquisition of at least some skinspecific homing receptors might require specific imprinting conditions possibly from PLN DC. On the other hand, although RA suppressed the upregulation of skinhoming molecules during activation of naive T cells, its effect on effector/memory T cells has not been determined. Could exposure to RA reprogram memory cells that have already committed to a skin homing phenotype? The answer to this question may have important functional and therapeutic implications. In fact, such a mechanism could explain, at least in part, the therapeu-

tic effect of retinoids in T cell-mediated cutaneous autoimmune diseases like psoriasis (Kuenzli and Saurat, 2001). J. Rodrigo Mora and Ulrich H. von Andrian The CBR Institute for Biomedical Research and Department of Pathology Harvard Medical School 200 Longwood Avenue Boston, Massachusetts 02115 Selected Reading Campbell, D.J., and Butcher, E.C. (2002). J. Exp. Med. 195, 135–141. Carramolino, L., Zaballos, A., Kremer, L., Villares, R., Martin, P., Ardavin, C., Martinez, A.C., and Marquez, G. (2001). Blood 97, 850–857. Iwata, M., Hirakiyama, A., Eshima, Y., Kagechika, H., Kato, C., and Song, S.-Y. (2004). Immunity 21, this issue, 527–538. Johansson-Lindbom, B., Svensson, M., Wurbel, M.A., Malissen, B., Marquez, G., and Agace, W. (2003). J. Exp. Med. 198, 963–969. Kuenzli, S., and Saurat, J.H. (2001). Curr. Opin. Investig. Drugs 2, 625–630. Kunkel, E.J., and Butcher, E.C. (2002). Immunity 16, 1–4. Masopust, D., Vezys, V., Usherwood, E.J., Cauley, L.S., Olson, S., Marzo, A.L., Ward, R.L., Woodland, D.L., and Lefrancois, L. (2004). J. Immunol. 172, 4875–4882. Mora, J.R., Bono, M.R., Manjunath, N., Weninger, W., Cavanagh, L.L., Rosemblatt, M., and von Andrian, U.H. (2003). Nature 424, 88–93. Stagg, A.J., Kamm, M.A., and Knight, S.C. (2002). Eur. J. Immunol. 32, 1445–1454. von Andrian, U.H., and Mackay, C.R. (2000). N. Engl. J. Med. 343, 1020–1034.