Armed and ready: How effector T cells deploy in reactive lymph nodes to modulate immunity

New evidence demonstrates how two different effector T cell subsets traffic in reactive lymph nodes to modulate T cell and B cell responses.

T cells home to specific locations in secondary lymphoid tissue by means of their differential expression of chemokine receptors. For example, naive T cells and central memory T cells (T_CM cells) home to T cell areas of lymphoid tissue through their expression of CCR7, whereas effector T cells (T_E cells) and effector memory T cells (T_EM cells) lose expression of CCR7 and traffic in peripheral tissues (T_CM cells and T_EM cells are subsets of memory T cells defined by the absence or presence, respectively, of immediate effector function and the ability to home to peripheral rather than lymphoid tissue)¹. In this issue of Nature Immunology, two papers provide new evidence that T cell subsets can traffic in reactive lymph nodes through expression of the chemokine receptors CXCR3 and CXCR5 to inhibit or enhance immune responses. Findings from both papers have implications for vaccine development, administration and monitoring.

In the first paper, Guarda et al.² show how CD8⁺ T_E and T_EM cells, which express CXCR3, home to high endothelial venules (HEVs) in reactive but not resting lymphoid tissue through interactions with inducible CXCR3 ligands in the HEV lumen. These cells then gain entry to T cell areas of the reactive node, where they can target and kill cognate antigen-bearing dendritic cells (DCs) as a means to limit secondary, and perhaps even primary, immune responses. In the second paper, Fazilleau et al.³ focus on follicular B helper T cells (T_FH cells), a subset of antigen-experienced CD4⁺ T cells known to traffic to B cell follicles through expression of CXCR5 to promote antibody responses⁴. The authors demonstrate that 'effector' T_FH cells accumulate 'preferentially' in lymphoid tissue draining sites of vaccination, locally promoting germinal center formation and the generation of plasma cells. As the antibody response contracts, T_FH cells diminish in number and function but persist at the site. They retain the ability to mount effector function after local re-exposure to antigen, thus demonstrating a 'memory' T_FH phenotype.

In the steady state, naive T cells and T_CM cells home to HEVs and T cell areas in secondary lymphoid tissue through their expression of L-selectin (CD62L), the integrin LFA-1 and the receptor CCR7, which interact with L-selectin ligands (peripheral node addressin), intercellular adhesion molecules and the chemokines CCL19 and CCL21 (Fig. 1a). In contrast, CD8⁺ T_E and T_EM cells lack L-selectin and CCR7, have higher expression of CXCR3 and traffic in tissues, blood and spleen. They also acquire the ability to bind P- and E-selectin, which along with CXCR3 expression permits homing to inflamed tissues where the effector cells lyse targets bearing cognate antigen. Guarda et al.² present the new observation that CD8⁺ T_E and T_EM cells, which are generally believed to avoid lymph nodes (although this has been described for infection with human immunodeficiency virus⁵), can in fact migrate to lymph nodes but 'preferentially' traffic to reactive rather than resting nodes. They obtained this result in reactive nodes conditioned by a variety of stimuli, including mature DCs, tumor necrosis factor, vaccine adjuvant and local listeria infection.

Figure 1: Effector T cell subsets traffic in T cell and B cell areas of lymph nodes to modulate immune responses.

Kim Caesar

(a) Naive CD8⁺ T cells and T_CM cells home to HEVs and T cell areas in secondary lymphoid tissue through their expression of L-selectin and CCR7 (left). In contrast, CD8⁺ T_E and T_EM cells lack L-selectin and CCR7 and do not migrate toward resting nodes. However, T_E and T_EM cells have higher expression of CXCR3 and migrate to HEVs in reactive lymph nodes (right), which upregulate expression of CXCR3 ligands (CXCL9 and CXCL10). T_E and T_EM cells then gain access to T cell areas of the node through interactions with unknown adhesion molecules and chemokines ('?'), where they can kill cognate DCs to limit secondary immune responses. PNAd, peripheral node addressin. (b) T_FH cells are specialized CD4⁺ T cells that express CXCR5 and home to B cell follicles through interactions with CXCL13. In the effector phase (left), T_FH cells accumulate in lymphoid tissue draining sites of inoculation and express molecules that promote the generation of germinal center (GC) B cells and plasma cells. With contraction of the B cell response (right), the number of T_FH cells decreases, although 'reservoirs' of antigen-specific T_FH cells persist in conjunction with antigen–MHC class II. The persistent T_FH cells are ICOS^loOX40^lo and have less expression of effector cytokines, but they express CD69 and have recall response ability. Persistent antigen–MHC class II may be provided by dendritic cells (which can be found in B cell zones¹²), although the precise cell type is not known ('?'). IL-, interleukin; IFN- γ, interferon-γ; MHC II, MHC class II; APC, antigen-presenting cell.

Guarda et al.² show that CD8⁺ T_E and T_EM cells rapidly migrate to and re-enter reactive lymph nodes through HEVs by a mechanism that depends mainly on expression of CXCR3, which interacts with HEV-lumenal CXCR3 ligands (CXCL9 and CXCL10; Fig. 1a). Notably, they show that these ligands are rapidly induced in the HEV lumen in reactive nodes by inflammatory stimuli but are not expressed in resting lymph node HEVs. Migration of T_E and T_EM cells to reactive lymph nodes may depend partly on low expression of CCR7 on the T cells, as T cells deficient in both CCR7 and CXCR3 were less able to home to reactive lymph nodes than were T cells deficient in CXCR3 alone. The authors report that after gaining entry to the node, CD8⁺ T_E and T_EM cells limit the induction of secondary responses by killing DCs bearing cognate antigen². This is the first time, to our knowledge, that T cell–mediated killing of cognate DCs has been demonstrated in lymphoid tissue (it has been reported in peripheral tissues before)⁶.

T_FH cells are a specialized type of antigen-experienced (CD45RO⁺) CD4⁺ helper T cell that orchestrates B cell responses in secondary lymphoid tissues. These cells home to B cell follicles through interactions between CXCR5 and CXCL13 (CXCL13 is produced by stromal cells in B cell follicles, primed T_FH cells, and myeloid and plasmacytoid DCs)⁴. T_FH cells promote B cell responses, including B cell memory development and plasma cell differentiation, through their expression of costimulatory molecules and cytokines such as ICOS, OX40, CD40L and interleukins 10 and 21 (ref. 4). How these cells develop and whether they have the capacity for memory are issues now being investigated.

The paper by Fazilleau et al.³ sheds light on both issues. The authors report local accumulation of T_FH cells after vaccination with a foreign protein (pigeon cytochrome c) using two different adjuvants. Subcutaneous vaccination promotes high-affinity helper T cell responses in draining lymph nodes but promotes them to a much lesser extent in spleen. A substantial fraction of these helper T cells are CXCR5⁺ICOS⁺OX40⁺ effector T_FH cells (Fig. 1b). The development of these cells is associated with the production of germinal center B cells and plasma cells. Notably, the reverse is true with intraperitoneal vaccine administration: high-affinity T_FH cells, germinal center B cells and plasma cells develop in the spleen but develop to a much lesser extent in lymph nodes.

After contraction of the B cell response, the authors report fewer T_FH cells in the draining node³. However, 'reservoirs' of antigen-specific T_FH cells persist in the node for over 28 days. These persistent T_FH cells are ICOS^loOX40^lo and have less effector capacity and lower expression of genes encoding effector cytokines (Fig. 1b). However, they have recall response ability, regaining expression of effector cytokine mRNA after antigen re-exposure. In association with this 'memory' T_FH cell compartment, the authors note the development of antigen-specific memory B cells and, by indirect evidence, the presence of persistent complexes of peptide and major histocompatibility complex (MHC) class II that they detected over 70 days after priming (although they do not identify a specific cell responsible for this). In addition, they detected continued expression of CD69 in a subset of these memory T_FH cells as late as 56 days after priming. They suggest that CD69, which may locally tether the lymphocytes through ligation of an as-yet-undescribed ligand, as well as the persistence of specific antigen, help to maintain the memory T_FH cell population in the draining node^7,8.

The results of these studies raise many issues. Although CXCR3 ligands promote the homing of CD8⁺ T_E and T_EM cells to HEVs in reactive nodes², it is still not apparent what molecular interactions are involved in the sticking and rolling of these cells on the HEV lumen and their migration into T cell area of the node (Fig. 1b). In addition, it is not known how long the CD8⁺ effector T cells persist in reactive lymph nodes. Also, the homing of effector CD4⁺ T cells, which apparently traffic by different mechanisms, is not addressed in these studies.

One chief unresolved issue is the relative importance of the phenomenon of DC killing in the node by cognate CD8⁺ effector T cells. Is this an important means of negative feedback? In addition, what is the main site of DC elimination by effector CD8⁺ T cells; is this reactive lymphoid tissue or the periphery? A published study has reported that DC elimination by this mechanism occurs in the periphery but not in lymph nodes⁶.

Another important issue is how to reconcile the observed killing of cognate DCs in reactive nodes with the established effectiveness of many booster vaccination regimens⁹. What makes some DCs stimulatory, whereas some DCs apparently become targets? One possible explanation is that DCs can be rendered resistant to killing by cytolytic T lymphocytes depending on the maturation stimulus. A published study has shown that DCs activated by lipopolysaccharide, CD40L or T helper type 1 CD4⁺ T cells (but not those activated by T helper type 2 CD4⁺ T cells) are rendered resistant to killing by cytolytic T lymphocytes¹⁰. This effect is mediated by upregulation in the DC of a specific serum protease inhibitor (Spi-6) that inhibits the function of granzyme B.

There are also many unresolved issues about the development and function of memory T_FH cells. CD4⁺CXCR5⁺ T cells are also found in the circulation^4,11. So what leads to the retention of some T_FH cells in lymphoid tissue and the release of other T_FH cells to the blood? In particular, the precise MHC class II–positive antigen-presenting cell associated with and perhaps maintaining T_FH cell memory is not known. It is not a follicular DC (which do not process antigen or express MHC class II molecules); perhaps myeloid DCs or a subset of B cells are responsible (Fig. 1b). The function of CD69 in the retention of memory T_FH cells and the location of these memory T_FH cells in the node also must be elucidated.

Many issues raised by both reports are relevant to vaccination studies. Can adjuvants be selected for that will enhance beneficial effects such as the development of T_FH cells but avoid unwanted effects such as DC killing? How should booster vaccinations be administered without inadvertent promotion of negative feedback and downregulation of specific immunity? In addition, in vaccination studies, peripheral blood is typically monitored for the presence of high-affinity T cells after vaccination. But if a large proportion of these cells remain localized or are recruited to lymph nodes draining inoculation sites, how will this affect the assessment of vaccine efficacy? These issues must be considered in the design of future vaccine studies in animal models and in human trials.