D. R. Kaufman, Division of Viral Pathogenesis, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, E/CLS 1024, Boston, MA 02115, USA. E-mail: email@example.com
Many pathogens use mucosal surfaces to enter and propagate within the host, making particularly desirable vaccines that target immune responses specifically to mucosal compartments. The majority of mucosal vaccine design strategies to date have been empirical in nature. However, an emerging body of basic immunological knowledge is providing new insights into the regulation of tissue-specific lymphocyte trafficking and differentiation. These insights afford the opportunity for the rational design of vaccines that focus immune responses at mucosal surfaces. Mucosal cellular immunity may prove critical for protection in the context of HIV infection, and thus there has been considerable interest in developing vaccines that target HIV-specific cellular immune responses to the gastrointestinal and vaginal mucosa. However, the optimal strategies for eliciting mucosal cellular immune responses through vaccination remain to be determined. Here, we review both recent vaccine studies and emerging paradigms from the basic immunological literature that are relevant to the elicitation of potent and protective mucosal cellular immune memory. Increasing the synergy between these avenues of research may afford new opportunities for mucosal vaccine design.
Vaccination strategies that elicit mucosal immunity will likely prove critical for protection against a wide variety of pathogens that use mucosal surfaces as portals of entry or sites of replication. Ongoing advances in our understanding of the basic biology of mucosal lymphocyte trafficking and differentiation are affording opportunities to optimize the design of mucosal vaccines. However, while a great deal is known about the biology of secretory immunoglobulin (Ig)A and other aspects of mucosal humoral immunity (for review, see [1,2]), the corresponding mechanisms underlying the induction and maintenance of mucosal cellular immune responses are only recently beginning to be elucidated in detail.
Significant progress has been made in characterizing the interactions between dendritic cells and T lymphocytes, and between T lymphocytes and vascular endothelial cells, that license T lymphocyte migration to mucosal and non-mucosal extralymphoid compartments. However, considerable uncertainty surrounds the fate of T lymphocytes that are licensed initially to migrate to mucosal surfaces. How does the activation phenotype of the T lymphocyte and the inflammatory state of the host influence the homing mechanisms that are required for mucosal entry and retention? Do mucosa-homing T lymphocytes undergo phenotypic changes in response to specific signals in the mucosal microenvironment that impact their ability to proliferate or elaborate various effector functions? How efficiently are T lymphocytes retained in the mucosa, and can memory T lymphocytes be reprogrammed dynamically following migration to distinct anatomical compartments? The answers to these questions are particularly germane to the design of vaccines that elicit optimal recall responses at mucosal surfaces.
Mucosal cellular immunity is likely to play a particularly critical role in vaccine-mediated protection in the context of HIV infection, where viral transmission typically occurs across mucosal surfaces, and viral replication within the intestinal mucosa contributes significantly to disease pathogenesis [3–6]. This is highlighted by the recent observation that a T cell-based vaccine capable of generating potent mucosal CD8+ T lymphocyte responses  can prevent the destruction of mucosal CD4+ T lymphocytes, lower viral load set-point and delay or abrogate disease progression in rhesus macaques challenged intravenously with simian immunodeficency virus (SIV) . However, the properties of a vaccine that will maximize protective mucosal cellular immunity in the context of HIV and other mucosal infections remain poorly defined. In particular, it is not clear whether a mucosal or systemic route of vaccine delivery is optimal for generating high-frequency, avid and functional mucosal cellular immune memory. This controversy, in turn, reflects limits in our knowledge of the underlying biology of CD4+ and CD8+ T lymphocyte trafficking and the influence of mucosal microenvironments on the phenotype and function of vaccine-elicited T lymphocytes. In this review, we will describe our current understanding of the mechanisms that control T lymphocyte homing and differentiation as they relate to vaccine-elicited mucosal cellular immune protection, and highlight future avenues of investigation that may increase our ability to optimize the protective mucosal immunity afforded by the next generation of T cell-based vaccines.
Vaccine routing and the induction of mucosal cellular immunity
Significant controversy surrounds the question of how the route of antigen exposure or delivery influences the homing specificity and anatomical distribution of resultant T lymphocyte responses. Studies in humans demonstrated that circulating memory T lymphocytes elicited by intestinal rotavirus infection were enriched for expression of the mucosal homing integrin α4β7 [9,10]. Conversely, herpes simplex virus 2-specific memory T lymphocytes (but not those specific for non-dermatotropic herpesviruses such as cytomegalovirus and Epstein–Barr vius) were enriched for expression of P-selectin and E-selectin ligands, which are critical for skin homing . Oral immunization of human subjects with either an attenuated Salmonella vector or a soluble antigen similarly generated an α4β7 integrin-expressing population of circulating antigen-specific CD4+ and CD8+ T lymphocytes [12,13], while subcutaneous immunization with a recombinant DNA and modified vaccinia Ankara (MVA) prime-boost regimen failed to do so . However, the functional consequences of these patterns of homing marker expression are unclear, as memory T lymphocytes expressing tissue-specific homing receptors are not retained exclusively at mucosal or epithelial surfaces, but traffic readily through the blood and other anatomical compartments .
Animal models of localized infection have revealed a complex relationship between the initial site of antigen exposure and the anatomical distribution of the resultant cellular immune response. In an early study, subcutaneous or oral administration of rotavirus initially biased cytotoxic T lymphocyte (CTL) activity towards the site of inoculation. However, CTL activity equilibrated subsequently across systemic and mucosal anatomical compartments by day 21 post-infection despite limited dissemination of the virus . Consistent with this observation, antigen-specific memory CD8+ T lymphocytes were detected at high frequency in multiple systemic and mucosal anatomical compartments after systemic vaccinia virus infection  or localized infection with rotavirus (intestine) or Sendai virus (respiratory tract) . A similarly broad anatomical distribution of responding CD8+ T lymphocytes was observed when adoptively transferred T cell receptor-transgenic CD8+ T cells were activated by the tissue-restricted expression of cognate antigen in the intestine under control of a transgenic promoter . These studies suggest that an acute cellular immune response characterized by tissue-restricted patterns of antigen expression or T lymphocyte trafficking may subsequently give rise to a broad anatomical distribution of T lymphocytes in the memory phase.
More recently, a study of CD8+ T lymphocyte trafficking following cutaneous vaccinia virus infection has shed light on how the cellular immune system may maintain a balance between anatomically focused and broadly distributed cellular immune responses . Following intradermal vaccinia inoculation, CD8+ T lymphocytes were activated initially in draining lymph nodes to up-regulate skin-homing P-selectin and E-selectin ligands and home to the inoculation site. However, some activated T lymphocytes migrated subsequently to mesenteric lymph nodes, where they down-regluated P- and E-selectin ligands and up-regulated α4β7 integrin, resulting ultimately in a broad anatomical distribution of T lymphocyte responses. Despite this dynamic reprogramming of lymphocyte homing capacity, the site of initial antigen exposure had a long-term impact on patterns of T lymphocyte homing marker expression. Following intradermal vaccinia inoculation, memory CD8+ T lymphocytes in all anatomical compartments expressed P- and E-selectin ligands at a high frequency and α4β7 integrin at a low frequency, while the converse was true following intraperitoneal inoculation. In a separate study, when circulating rotavirus-specific memory CD8+ T lymphocytes were sorted into α4β7 integrin high and low populations and transferred adoptively to recombinant-activating gene-deficient mice infected chronically with rotavirus, only the α4β7 integrin high population was capable of clearing the infection , suggesting that the long-term maintenance of tissue-specific homing marker expression by circulating memory T lymphocyte populations may have functional relevance.
As in animal models of infection, animal models of vaccination have yielded complex findings regarding the impact of antigen delivery route on the anatomical distribution and functionality of the resultant cellular immune response. An early comparative routing study suggested that systemic and mucosal vaccination with a recombinant adenovirus (rAd) vector expressing herpesvirus glycoprotein B (gB) could generate equivalent acute systemic and mucosal CTL responses but divergent, anatomically biased recall responses . In a subsequent study, intrarectal rAd–gB administration proved superior to subcutaneous administration for protecting mice against rectal or vaginal HSV-2 challenge . In another series of reports, immunization with a rAd vector expressing the Mycobacterium tuberculosis (MTB) antigen Ag85A protected mice against respiratory MTB challenge when administered intranasally, but not intramuscularly [23,24]. Although intramuscular immunization generated higher frequencies of Ag85A-specific CT8+ T lymphocytes in the spleen and lung parenchyma, only intranasal immunization generated antigen-specific CD8+ T lymphocytes within the airway lumen, and these luminal T lymphocytes could mediate protection against respiratory MTB challenge in adoptive transfer studies . Similarly, another study demonstrated that intravaginal administration of a rAd vector expressing HIV Gag was superior to intramuscular administration for boosting Gag-specific vaginal CD8+ T lymphocyte responses primed by intravaginal administration of a recombinant, attenuated Listeria monocytogenes vector .
Divergent effects of systemic and mucosal vaccine delivery have also been reported in a series of studies using peptide and recombinant MVA (rMVA)-based HIV vaccines. These data indicated that intrarectal but not subcutaneous immunization with a peptide-based vaccine could induce high-frequency CTL in the intestinal mucosa of rhesus macaques . Following intrarectal simian–HIV challenge, intrarectally but not subcutaneously vaccinated macaques mounted high-frequency mucosal CD8+ T lymphocyte responses, had lower viral load set-points and preserved systemic and mucosal CD4+ T lymphocytes [27,28]. The same investigators also reported that intrarectal vaccination with rMVA or recombinant plasmid DNA prime-rMVA boost regimens was superior to intramuscular or subcutaneous vaccination for generating mucosal CD8+ T lymphocyte responses in the intestinal mucosa [28,29]. They suggested that this was related to the superior ability of mucosal vaccination to elicit high-avidity, and therefore more functional, CD8+ T lymphocytes at mucosal inductive and effector sites [28,29]– a finding supported by an independent study of systemically and mucosally administered recombinant fowlpox vaccine vectors .
In contrast to these observations, a number of studies have demonstrated that systemic vaccination can generate potent, persistent and protective mucosal cellular immune responses, which can prove equivalent or superior to mucosal immunization in this regard. We and others have shown that intramuscular immunization of mice with rAd vectors, either singly or in heterologous prime-boost combinations, can induce high-frequency, functional and persistent CD8+ T lymphocyte responses in multiple mucosal compartments including the small and large intestines, vaginal tract and respiratory tract [7,31]. Recombinant Ad-elicited mucosal CD8+ T lymphocytes had an effector memory phenotype and expanded following mucosal challenge with recombinant vaccinia virus expressing cognate antigen. In contrast, mucosal routing of rAd vectors inefficiently induced CD8+ T lymphocyte responses in systemic and mucosal compartments, although mucosal rAd vector administration could prime effectively for intramuscular rAd boosting ([32,33]; Kaufman and Barouch, unpublished data). Using adoptive transfer studies, we also showed that intramuscular rAd vaccination licensed systemic CD8+ T lymphocytes to up-regulate mucosal homing receptors and to traffic to mucosal surfaces, where they accounted for the vast proportion of mucosal CD8+ T lymphocytes elicited by intramuscular immunization . Similarly, multiple preclinical HIV vaccine studies in rhesus macaques have also demonstrated that intramuscular vaccination with recombinant DNA, attenuated vaccinia viruses or replication-incompetent adenoviruses in various prime-boost combinations generates mucosal CD8+ T lymphocyte responses that are comparable in frequency to those seen in blood [7,34,35]. Intramuscular immunization of macaques with either recombinant DNA prime-rAd boost regimens or heterologous rAd prime-boost regimens utilizing two serotypically distinct rAd vectors protected animals from mucosal CD4+ T lymphocyte depletion [8,36], decreased plasma SIV viral load and prolonged survival [8,37] following high-dose SIV challenge. In particular, the data from the heterologous rAd prime-boost study are notable because no HIV envelope antigen was included in the vaccine, highlighting the role of T cell-mediated immunity in this context.
The reasons for the divergent results obtained in the studies detailed above are unclear. Variations in vectors, experimental models and immunological assays could account for some of these discrepancies. In addition, recombinant vaccine vectors and viruses may persist for long periods of time following inoculation, leading to ongoing local antigenic stimulation that can skew recall responses persistently [38–40]. However, it is likely that different vaccination regimens have the capacity to trigger unique T cell differentiation programmes that vary in their ability to overcome the anatomical compartmentalization of the cellular immune system. For this reason, the expanding body of data illuminating the basic biology of T lymphocyte trafficking and differentiation is likely to afford important insights that are critically relevant for mucosal vaccine design.
New paradigms in mucosal T lymphocyte homing: implications for mucosal vaccines
T lymphocytes migrate from the blood to both secondary lymphoid organs and peripheral tissues by extravasation through the endothelium of post-capillary venules. Lymphocyte transmigration is a highly regulated, multi-step process (for review, see [41–43]). Lymphocyte rolling and tethering are mediated through the interaction of lymphocyte selectin ligands and integrins with their cognate endothelial receptors. Subsequently, tissue-derived chemokines activate G-protein-coupled chemokine receptors on the lymphocyte surface, which signal reciprocally a conformational change in cell surface integrins. Integrin activation leads in turn to lymphocyte arrest and diapedesis. Following transmigration, extravascular lymphocyte homing is mediated by chemokine and cytokine gradients, and integrin-mediated interactions may influence tissue-specific lymphocyte retention. Cytokines, antigen and inflammatory mediators may affect lymphocyte trafficking patterns concomitantly by modifying the activation state of the lymphocyte, modulating endothelial permeability and altering patterns of endothelial adhesion marker expression. This multi-step paradigm allows for significant combinatorial diversity and complexity in the control of lymphocyte homing.
The mechanisms regulating T lymphocyte localization to mucosal surfaces are best understood in the context of homing to the intestinal lamina propria (Fig. 1). The interaction of α4β7 integrin with its cognate ligand mucosal vascular addressin cell adhesion molecule 1 (MAdCAM-1) on the high endothelial venules of Peyer's patches and the post-capillary venules of the small and large intestine is critical for steady state accumulation of T lymphocytes in the gut-associated lymphoid tissue (GALT) [44–46], and β7 integrin-deficient mice have a severly hypotrophic GALT . In the context of an active anti-viral immune response, α4β7 integrin is also important for the efficient migration of virus-specific CD8+ T lymphocytes to the intestinal lamina propria and epithelium, although a proportion of activated T lymphocytes can enter these anatomical compartments in an α4β7 integrin-independent manner . In the large intestine, α4β1 integrin may also synergize with α4β7 integrin to promote lymphocyte homing . Epithelial cells in the small intestine secrete the chemokine ligand 25 (CCL25) along a gradient that is highest at the proximal end (duodenum) and lowest at the distal end (ileum) , and CCL25 is bound and presented to T lymphocytes on the vascular endothelial surface . CCL25 binds to chemokine receptor 9 (CCR9) on T lymphocytes and promotes activation of α4β7 and αEβ7 integrins, thereby facilitating T lymphocyte homing to both the lamina propria and intestinal epithelium [51,52]. While steady-state accumulation of CD4+ and CD8+ T lymphocytes in the mucosa of the small intestine is, in large part, CCR9-independent , CCR9 plays an important role in the homing of recently activated T lymphocytes to the small bowel mucosa, particularly within the proximal small bowel [51,54,55]. However, there is also likely to be an important role for other G-protein-coupled receptors in this process . In contrast, CCL25 is not expressed in the large intestine , and CCR9 is likely to be dispensable for T lymphocyte homing to this compartment [50,57]. Following transmigration into the lamina propria, T lymphocytes down-regulate α4β7 integrin and up-regulate αEβ7 integrin , which may influence T lymphocyte retention within both the lamina propria and intestinal epithelium through interactions with its ligand, E-cadherin . CCR9- and β7 integrin-independent mechanisms for T lymphocyte homing to and retention within the intestinal mucosa remain incompletely defined.
Recent insights into the mechanisms underlying the imprinting of tissue-specific T lymphocyte homing specificity have highlighted the importance of microenvironment-specific signals in the priming phase of the immune response. CD4+ and CD8+ T lymphocytes that are primed in peripheral lymph nodes or activated in vitro by antibodies to CD3 and CD28 up-regulate P-selectin and E-selectin ligands and home to skin [19,60,61]. Conversely, T lymphocytes primed in mesenteric lymph nodes up-regulate β7 integrin and home to mucosal surfaces . In vitro priming of CD8+ T lymphocytes with peptide-pulsed dendritic cells derived from peripheral lymph nodes induces the up-regulation of P-selectin and E-selectin ligands and licenses skin migration following adoptive transfer to naive hosts . Conversely, in vitro priming with mesenteric lymph node or Peyer's patch-derived dendritic cells induces the up-regulation of β7 integrin and CCR9 on CD8+ T lymphocytes and licenses their migration to the small intestine [54,61,62]. The primacy of the dendritic cell–T lymphocyte interaction in imprinting tissue-specific lymphocyte homing was suggested in a murine model of CD8+ T lymphocyte-mediated anti-tumour immunity, where tumour cells expressing unique antigens were implanted simultaneously within different anatomical compartments. Tumours implanted at subcutaneous, intracerebral or intraperitoneal sites primed populations of CD8+ T lymphocytes simultaneously within a single lymph node to up-regulate distinct patterns of tissue-specific homing markers , presumably through the action of cross-presenting dendritic cells derived from each anatomical site. However, additional complexity is suggested by other studies, which have demonstrated that signals derived from the tissue parenchyma [64,65] or lymph node stroma  can influence the tissue-specific homing properties of T lymphocytes, either directly or by modulating the imprinting properties of tissue-derived dendritic cells.
Retinoic acid (RA) plays a critical role in the acquisition of intestinal mucosal homing capacity by T lymphocytes. CD8+ T lymphocytes primed in vitro in the presence of RA up-regulate mucosal homing markers and can traffic to the lamina propria of the small intestine . Moreover, dendritic cells derived from the intestinal mucosa, but not peripheral tissues, express high levels of RA-producing enzymes and can synthesize RA from retinol (vitamin A). Inhibition of RA synthesis by mucosal dendritic cells abrogates their ability to imprint gut homing on CD8+ T lymphocytes . Highlighting the central role of RA in the generation of gut-tropic T lymphocytes, CD4+ and CD8+ T lymphocytes are nearly absent from the small intestine lamina propria of mice with a dietary deficiency of retinol . Moreover, RA also plays an important role in the induction of gut-homing, IgA-secreting B lymphocytes and forkhead box P3-positive (FoxP3+) regulatory CD4+ T lymphocytes [68–71]. A unique CD103+ subset of dendritic cells found in the intestinal lamina propria and mucosal lymphoid inductive sites expresses high levels of the RA synthesizing enzyme retinaldehyde dehydrogenase 2 , and these dendritic cells are superior to their CD103- counterparts in their ability to imprint mucosal capacity on CD8+ T lymphocytes .
Taken together, these data suggest that a complex and highly regulated series of tissue-specific molecular signals dictates the imprinting of homing specificity on naive T lymphocytes during the initial acute phase of the immune response. However, significant questions remain regarding the dynamic changes in T lymphocyte homing capacity and effector function that occur later in the acute phase or in the memory phase. The phenotypic plasticity of memory T lymphocytes in this regard may have important implications for the long-term capacity of these cells to maintain their initial tissue-restricted patterns of migration or mount subsequent recall responses at locations distinct from the initial site of antigen exposure. These concerns are particularly pressing in the context of mucosal vaccine design.
Phenotypic plasticity and dynamic migration of memory T lymphocytes: new areas of investigation for mucosal vaccine design
A number of studies suggest that patterns of T lymphocyte homing receptor expression do not represent the fixed features of terminally differentiated cells, but rather can be reprogrammed following initial imprinting. Reprogramming of T lymphocyte tissue-specific homing marker expression and homing capacity has been observed both during the primary immune response  and upon subsequent reactivation of memory T lymphocytes under different priming conditions . Moreover, memory T lymphocytes can recirculate between anatomical compartments, although the rate of turnover differs significantly between anatomical sites. Parabiosis experiments in mice demonstrated a rapid equilibration of resting antigen-specific memory CD8+ T lymphocytes in the liver, lung, spleen, lymph nodes and bone marrow of conjoined animals . In contrast, equilibration in the peritoneal cavity, intestinal lamina propria and central nervous system was delayed, and turnover within the intestinal epithelium was highly restricted . In these studies, memory CD8+ T lymphocytes trafficking to the intestinal mucosa were able to up-regulate β7 integrin and the activation marker CD69, highlighting their phenotypic plasticity. The phenotypic and migratory plasticity of anatomically distinct memory T lymphocyte populations was demonstrated further in a study where memory CD8+ T lymphocytes were isolated from the spleen or intestinal epithelium of mice and adoptively transferred intravenously to naive recipients . Upon antigenic rechallenge, both spleen- and intestine-derived memory T lymphocytes migrated broadly to different anatomical compartments. Regardless of their anatomical origin, responding T lymphocytes migrating to the spleen acquired phenotypic markers characteristic of systemic T lymphocytes, while the converse was true for those migrating to the intestinal mucosa. These studies highlight the influence of T lymphocyte recirculation patterns and phenotypic plasticity on the geography and efficacy of subsequent recall responses. From a vaccinologist's perspective, understanding the extent and limits of memory T lymphocyte recirculation and reprogramming may illuminate the underlying biology of mucosal cellular immune protection and may also help resolve the discrepancies between the previous comparative vaccine routing studies highlighted above.
Additional avenues of investigation are likely to prove particularly fruitful for the design of the next generation of mucosal T cell-based vaccines (Table 1). First, little is known about how systemic and mucosal microenvironments influence the phenotype and activation state of memory T lymphocytes that populate distinct anatomical locations. In particular, an intricate balance of activating and toleragenic signalling mechanisms exists within the intestinal mucosa to allow for mucosal immune protection while promoting symbiosis with gut microflora. Little is known about the susceptibility of memory T lymphocytes to environmental signals in this context. Secondly, the overlapping geography of effector and regulatory T lymphocyte activation may influence the balance of these populations within particular anatomical compartments, with significant consequences for the functionality of the cellular immune response at each site. In this regard, it is intriguing to note that RA plays an important role in the generation of both gut homing CD8+ T lymphocytes and CD4+ FoxP3+ regulatory T lymphocytes [67–69,71]. Finally, memory T lymphocytes are divided into central and effector memory subsets with distinct functional properties [75–77], and the relative contribution of each subset to mucosal cellular recall responses has not been determined. The majority of central memory cells are found in lymphoid inductive sites and the bone marrow, while mucosal T lymphocytes exhibit primarily an effector memory phenotype. However, it is not known how rapidly central memory T lymphocytes can influx into mucosal tissues following antigenic challenge, nor how these cells might synergize with mucosa-resident T lymphocytes to clear mucosal pathogens.
Table 1. Avenues of research into basic T lymphocyte biology with the potential to impact mucosal vaccine design.
Naive T lymphocyte priming phase
Primary immune response
DC, dendritic cell.
DC-derived molecular signals that influence T lymphocyte homing marker expression
Role of inflammation in modulating T lymphocyte extravasation at mucosal surfaces
Recirculation of resting memory T lymphocytes between anatomical compartments
Relative contribution of central memory and effector memory T lymphocytes to mucosal recall responses
Role of anatomically distinct DC subsets in imprinting tissue-specific T lymphocyte homing specificity
Multiplicity and redundancy of tissue-specific homing receptors and pathways
Influence of microenvironment-specific signals on the phenotype of memory T lymphocytes
Effector functions of mucosa-resident T lymphocytes
Role of DC-independent signals in imprinting T lymphocyte homing specificity
Dynamic reprogramming of T lymphocyte homing specificity and phenotype
Dynamic reprogramming of T lymphocyte homing specificity and phenotype
Dynamic reprogramming of T lymphocyte homing specificity and phenotype
Correlation between T lymphocyte activation state and receptiveness to reprogramming of homing specificity
Comparative anatomical distribution of effector and regulatory T lymphocyte subsets
Comparative anatomical distribution of effector and regulatory T lymphocyte subsets
Contribution of mucosal innate immune signalling to mucosal DC and T lymphocyte activation
Role of antigen persistence in maintaining tissue-specific T lymphocyte populations
Molecular mechanisms governing T lymphocyte retention at mucosal surfaces
Important advances in mucosal vaccine design are also likely to come from the convergence of two additional areas of research – empirical studies on the utility of mucosal adjuvants and emerging insights into the role of innate immune signalling in mucosal homeostasis. A wide variety of molecules and macromolecular structures, including liposomes and other lipid formulations, immunostimulating complexes, Toll-like receptor (TLR) ligands, and cholera toxin and Escherichia coli heat-labile toxin conjugates have been utilized as adjuvants for mucosal vaccine formulation. These have shown promise for enhancing vaccine-elicited mucosal humoral immunity and, to a lesser extent, cellular immunity (for review, see ). Many of these adjuvants are thought to work by improving antigen targeting to mucosal DC and by triggering innate immune receptors at mucosal surfaces. Recent studies have demonstrated that innate immune signalling occurs constitutively at the gut mucosa, driven by interactions between gut commensual bacteria, mucosal epithelial cells and mucosal immune cells [79–81]. Importantly, alterations in TLR-mediated mucosal signalling have been shown to impact directly mucosal DC activation states and to alter the balance between regulatory and effector T lymphocytes elicited by mucosal pathogens . These data suggest that mucosal innate immune signalling regulates mucosal DC and T lymphocyte function continuously and actively. It will therefore be important to evaluate whether mucosal vaccine adjuvants can modulate these homeostatic mechanisms directly to overcome mucosal immune tolerance and improve vaccine-elicited mucosal cellular immunity.
While significant progress has been made in developing vaccine candidates that target cellular immune responses to mucosal surfaces, uncertainty remains about how to optimize persistent and protective mucosal T lymphocyte responses against HIV and a wide variety of other mucosal pathogens. In this context, a deeper understanding of the underlying biology of T lymphocyte activation, trafficking and differentiation will be needed to inform the design of the next generation of T cell-based mucosal vaccines (Table 1). We suggest that increased synergy between basic and applied disciplines may be fruitful for accelerating the development of vaccines against mucosal pathogens.