Naive T cells have the capacity to acquire any set of homing receptors on activation. However, the site of antigen entry exerts a strong influence on the traffic pattern that lymphocytes acquire. For example, pathogens entering through the skin, such as herpes simplex virus, preferentially generate lymphocytes with skin-homing receptors (Gonzalez et al. 2005; Kantele et al. 2003; Koelle et al. 2002, 2005). On the other hand, oral vaccination induces higher levels of a407 on effector/memory T cells (Kantele et al. 1999b; Lundin et al. 2002; Rojas et al. 2003; Rott et al. 1997) and B cells (Gonzalez et al. 2003; Kantele et al. 1997,1999a, 2005; Quiding-Jarbrink et al. 1997; Youngman et al. 2002) than parenteral administration of the same antigen. Importantly, among memory B and CD8 T cells, only a407+ (but not a4p7Neg) cells carried protection against intestinal rotavirus infection on adoptive transfer (Rose et al. 1998; Williams et al. 1998). In agreement with these observations, it has been shown that the homing potential of activated lymphocytes depends on the lymphoid tissue environment; T cells activated in MLN express higher levels of a407 and CCR9 as compared to those activated in skin-draining PLN (Campbell and Butcher 2002; Svensson et al. 2002). Conversely, skin-homing receptors (selectin ligands) are preferentially induced when T cells are activated in skin-draining PLN (Campbell and Butcher 2002).
In lymphoid tissues, DC are essential for efficient T cell activation (Jung et al. 2002; Probst and van den Broek 2005). DC modulate a number of T- and B cell properties in a tissue-specific fashion, such as Th1/Th2 bias (Everson et al. 1998; Iwasaki and Kelsall 1999; Rissoan et al. 1999) and antibody isotype switching (Sato et al. 2003; Spalding and Griffin 1986; Spalding et al. 1984). Recent work by several groups, including ours, has shown that DC are also responsible for the imprinting of tissue-specific homing potential. Intestinal DC from PP and MLN are sufficient to imprint activated T cells with the gut homing receptors a407 and CCR9 (Dudda et al., 2004, 2005; Johansson-Lindbom et al. 2003; Mora et al. 2003,2005; Stagget al. 2002), and a pronounced capacity to home to the small bowel mucosa (Mora et al. 2003) (Fig. 3A). On the other hand, DC from peripheral lymph nodes (PLN-DC) induce higher levels of E- and P-selectin ligands on T cells compared with intestinal DC (Dudda et al. 2004, 2005; Mora et al., 2005). Fucosyltransferase-VII (FucT-
Fig. 3A-D Imprinting mechanisms for gut- and skin-specific lymphocytes. Naive T and B cells access PLN, MLN, and PP through high endothelial venules (HEV) and become activated when they encounter a cognate antigen in the local T cell area or follicles, respectively. Once a lymphocyte has become activated in the lymphoid organs, it gives rise to tissue-seeking TEFF that leave the SLO via lymphatics or (from the spleen), enter directly into the blood, and home to peripheral organs through specialized postcapillary venules that express tissue-specific traffic molecules (vascular addressins). A PP-DC and MLN-DC express critical enzymes to metabolize vitamin-A (retinol) into RA, particularly retinaldehyde dehydrogenases (RALDH-1 and -2). T cell encounter of antigen in the presence of DC-derived RA induces gut-homing receptors and suppresses skin-homing molecules. This requires RA signaling through the RAR family of RA-receptors. It is likely that DC from lamina propria (LP-DC) can also make RA, but RA can be synthesized by other cells in the intestinal mucosa, such as enterocytes, which express RALDH-1. B When T cells are activated in skin-draining PLN, they are exposed to DC that cannot synthesize RA. The T cells then upregulate skin-homing molecules. The induction of this migratory preference appears to reflect the default differentiation pathway that T cells take on stimulation without gut-specific signals. However, it is still possible that skin-associated DC or other environmental factors are needed to upregulate CCR8 or CCR10 on cutaneous T cells. C B cells that become activated by oral antigens upregulate gut-homing molecules. Similar to the situation with T cells, we propose that PP-DC and MLN-DC, possibly by virtue of their ability to produce RA, may also trigger the acquisition of gut homing molecules by antigen-experienced B cells. In addition, B cells are exposed to other factors in the intestinal microenvironment, including TGFp and RA, which promote isotype switching to IgA. D Conversely, when B cells are activated in PLN or the spleen, they are not exposed to RA or other mucosal signals. It is not known whether there are specific imprinting signals for ASC or BMem targeting to the BM or other peripheral tissues
VII) is an essential enzyme for the generation of selectin ligands (Maly et al. 1996). Accordingly, PLN-DC induce higher levels of FucT-VII in CD8 T cells compared with DC from PP (PP-DC) (Mora et al. 2005). Interestingly, DC from spleen, glutaraldehyde-fixed DC (from any lymphoid tissue), or stimulation by anti-CD3 and anti-CD28 (without DC) also induced high levels of E- and P-selectin ligands on CD8 T cells (Mora et al. 2005). These results suggest that the acquisition of skin-homing molecules (at least E-/P-selectin ligands and probably also CCR4) maybe a default pathway whenever T cells are activated in the absence of gut-derived signals (Mora et al. 2005) (Fig. 3A). However, it must be cautioned that CCR10 and CCR8 (which can also be expressed on skin-homing T cells) are not induced by these in vitro conditions, suggesting that they might be controlled by skin-specific environmental cues.
A recent report has shown that mice depleted of vitamin A have dramatically reduced numbers of effector/memory T cells in the gut mucosa, but not elsewhere (Iwata et al. 2004). Concomitant in vitro experiments showed that the presence of the vitamin A metabolite retinoic acid (RA) during T cell activation induces a4ß7 and CCR9 on T cells, even in the absence of DC (Iwata et al. 2004). RA also blocked the activation-induced default upregulation of E- and P-selectin ligands on T cells (Iwata et al. 2004), a phenomenon that was also observed when T cells were activated in the presence of PP-DC, even when the PP-DC did not present the activating antigen (Mora et al. 2005). Importantly, gut DC (from PP and MLN) express higher levels of retinaldehyde dehydrogenases (RALDHs), which are essential enzymes for RA biosynthesis (Iwata et al. 2004). These enzymes were not detected in DC from PLN or spleen. Blocking RALDHs in DCs or RA receptors in T cells significantly decreased the induction of a4ß7 by PP-DC and MLN-DC (Iwata et al. 2004). Therefore, the ability to produce RA is an important mechanism by which gut DC induce gut-homing molecules and suppress skin-homing molecules on T cells (Fig. 3A).
It should be mentioned that even though intestinal DC are sufficient to induce gut homing molecules in vitro, it is formally possible that other sources of RA may contribute in vivo. For example, enterocytes in the small intestine express high levels of RALDH-1 (Iwata et al. 2004), and it has been shown that RA can be produced by a small intestinal epithelial cell line in vitro (Lampen et al., 2000) (Fig. 3A). On the other hand, intraperitoneal immunization efficiently generates gut-homing T cells (Campbell and Butcher 2002; Johansson-Lindbom et al. 2003; Svensson et al. 2002), and some nonintestinal viral infections also induce the generation of a4ß7+ T cells (Masopust et al. 2004). Thus it is possible that gut-homing T cells can be imprinted in regions other than gut-associated lymphoid tissues. It will be interesting to determine whether RA synthesis can be induced in nonintestinal DC under inflammatory or infectious conditions and whether there are additional, RA-independent mechanisms of intestinal imprinting (Mora and von Andrian 2004). Moreover, the expression of a4ß7 and CCR9 is not always linked. For example, naive CD8 T cells express high levels of CCR9 (Carramolino et al. 2001) but low levels of a4ß7 (Mora et al. 2003), and T cells infiltrating the colon mucosa are a4ß7High, but CCR9- (Zabel et al. 1999). Because RA induces both gut-homing molecules simultaneously (Iwata et al. 2004), these findings suggest that there are additional mechanisms that may regulate the expression of a4ß7 and/or CCR9 (Mora and von Andrian 2004).
It should also be emphasized that although PP-DC generate T cells that potently migrate to the small intestine, PP-DC-induced effector T cells do not migrate efficiently to the colon (Mora et al. 2003). Moreover, T cells activated with appendix-derived DC did not migrate better to the colon mucosa than those activated with PP-DC (Mora and von Andrian, unpublished obser vations), suggesting that there additional, still unknown factors involved in the imprinting of colon-homing T cells (assuming that specific colon-tropic T cells exist).
As discussed above, many IgA-ASC express a407 and CCR10, and some also express CCR9 (Cyster 2003; Kunkel and Butcher 2003). PP-DC can affect B cell function in a tissue-specific fashion by promoting class-switching to IgA (Sato et al. 2003; Spalding and Griffin 1986; Spalding et al. 1984). Similarly, B cell exposure to RA induces class-switching to IgA (Tokuyama and Tokuyama 1999). Therefore, it will be interesting to determine whether gut DC and/or RA can also imprint gut tropism in ASC and to understand how is this linked to IgA switching (Fig. 3B).
Was this article helpful?