Environmental cues orchestrate regional immune surveillance and protection by pulmonary CTLs

Review on how T_RM cells are an important target for vaccination and explores how the mechanisms that control fate decisions of activated CTLs may be used to capitalize on this immunity.

Abstract

Tissue-resident memory CD8 T cells (T_RM) provide preemptive immunity against infections that begin in peripheral tissues by guarding the site of initial pathogen exposure. Their role in immunity to respiratory virus infection is particularly important because severe damage to the alveoli can be avoided when local populations of T_RM cells reduce viral burdens and dampen the responses of effector CD8 T cells in the lungs. Although a connection between rapid immune activation and early viral control is well established, the signals that keep T_RM cells poised for action in the local tissues remain poorly defined. Recent studies have shown that environmental cues influence the fate decisions of activated CTLs during memory formation. Manipulation of these signaling pathways could provide new ways to capitalize on protection from T_RM cells in mucosal tissues, while reducing collateral damage and pathology during vaccination.

Introduction

The lungs are comprised of 2 functionally distinct compartments with very different structural organization: one is the highly vascularized parenchymal tissue that is required for respiration and can be easily damaged during a robust immune response, and the other compartment includes the lumen and walls of the conducting airways, where the stratified epithelium is relatively resistant to immune damage. The organization of the tissues places constraints on the immune response because distinct populations of specialized memory CD8 T cells are required for local immune surveillance. Vascular tissues are perfused by memory CD8 T cells, which use the bloodstream to circulate around the body [1]. Additional populations of noncirculating T_RM cells colonize the airways and reduce the quantities of infectious virus that reach the alveoli during secondary challenge [2, 3]. This mechanism of protection is an underused resource for vaccine development, largely because reliable methods for establishing local populations of memory CD8 T cells in the mucosal tissues are not currently available. Efforts to develop safe vaccines for pulmonary immunization have been hampered by a limited understanding of the signaling pathways that influence the fate decisions of activated CTLs during memory formation. Some recent progress in the field indicates that new methods of immunization could soon be on the horizon.

MECHANISMS OF PROTECTION BY PULMONARY T_RM CELLS

Influenza viruses are genetically labile pathogens that constantly alter their surface proteins to evade the humoral immune response. CTLs overcome this problem by targeting conserved epitopes from internal viral proteins to provide cross-protection between serologically distinct strains [4–6]. Evidence that T_RM cells make an important contribution to immunity became known after sequential infections with different strains of IAV revealed a weak correlation between the numbers of antiviral memory CD8 T cells in the circulation of mice that had recovered from a prior infection with IAV and the severity of disease during reinfection with another strain [7]. Robust cells that mediate immunity developed only when viral antigens were present in the respiratory tract during priming [8, 9] and declined as an unusual subset of activated CTLs disappeared from the airways [10–12]. These lung-resident CTLs exhibited characteristics of T_RM cells in other barrier tissues and vital organs [3, 13, 14], including stable CD69 expression [10–12]. Collectively, T_RM cells are the largest subset of memory CD8 T cells in the body and can be found in most human tissues, including the lungs [15, 16].

The mechanisms by which T_RM cells enhance immunity may vary according to the site of infection. Early viral control in the female reproductive tract has been linked to rapid recruitment of immune cells to the site of inoculation [17] by a mechanism that can be mimicked using local delivery of inflammatory mediators [18]. Others found that T cell-derived IFN-γ enhanced immunity in the lungs by a similar mechanism after transfer [19]. Natural immunity to a respiratory infection is more complex because the lungs contain heterogeneous populations of memory CD8 T cells after pulmonary immunization, including some T_CM cells, which move through the vasculature and undergo robust proliferation in the local lymph nodes after secondary challenge [20]. Images of the lungs taken during reinfection showed that very few migrating CTLs arrived in the airways before the viral burdens declined. The immunity was not compromised when an agonist of the sphingosine S1P receptor was used to sequester circulating CTLs in the lymph nodes during reinfection, indicating that lung-resident CTLs control viral replication [2, 21].

Some pulmonary T_RM cells can be collected from the lumen of the airways by BAL [12, 22]. These CTLs exhibit weak lytic activity in vitro [23], which is caused by low CD11a expression after exposure to proteases in the lung surfactant (24, 25). It is not clear that the protease activity compromises immunity because many cells that enter the lumen of the airways are expelled from the lungs by the mucociliary escalator (26). Other T_RM cells remain embedded in the walls of the airways [2] and maintain CD11a expression, which supports robust lytic activity in vitro (27). The predominant populations of pulmonary T_RM cells also express αEβ7 integrin (CD103), which resists protease digestion and can replace CD11a during cytolysis of infected epithelial cells [28–31]. Together, these data indicate that pulmonary T_RM cells are equipped to have an active role in viral clearance, rather than simply sending a distress signal to cells in other tissues. CTLs with markers of T_RM cells had immediate lytic potential after recovery from human lungs [32].

CHALLENGES FOR VACCINE DEVELOPMENT

Pulmonary infection is the most effective method for inducing robust cell-mediated immunity in the lungs but raises important safety concerns when applied as a vaccine strategy, thus creating an impetus for the use of alternative approaches. Immunization in distal tissues can also have unexpected consequences, as shown by experiments with mice that were primed with IAV by peritoneal injection to induce the formation of circulating memory CD8 T cells without accompanying T_RM cells [2]. The mice were poorly protected against reinfection with another strain, even though large populations of antiviral memory CD8 T cells were found in the circulation [2, 7]. Ominously, the lungs were overwhelmed by massive numbers of reactivated CTLs, which proliferated in the lymph nodes before entering the respiratory tract a few days later (Fig. 1A). By comparison, relatively few migrating CTLs entered the lungs of intranasally primed mice (Fig. 1B), which had many T_RM cells in the airways [2]. These observations are not consistent with a proinflammatory role for T_RM cells in immunity to respiratory virus infection and suggest that T_RM cells can prevent pathology in the lungs. More important, the data show that a potentially serious complication could arise when vaccines target circulating memory CD8 T cells without mucosal immunization [33, 34] because a robust T_EFF response could severely damage the alveoli during infection [35, 36]. Vaccine-induced immunity has been linked to exacerbated pathology in the lungs during infection with respiratory syncytial virus [35, 37, 38]. Elderly patients are poorly protected by antibody-based vaccines [39] and may be particularly vulnerable to adverse outcomes from severe pathology during a robust CTL response.

Figure 1. Pulmonary T_RM cells blunt the T_EFF response in the lungs.

(A) Mice were primed with IAV by peritoneal injection and challenged with a different strain 1 mo later. Reactivated memory CD8 T cells entered the lungs after proliferating in the local lymph nodes but were too late for early viral clearance. (B) Mice were primed by intranasal inoculation and challenged with a different strain 1 mo later. Pulmonary T_RM cells controlled viral replication with minimal assistance from circulating CTLs.

The cell-mediated immunity that develops during respiratory virus infection declines over the course of approximately 6 mo [7], which corresponds to the gradual erosion of the lung T_RM population [40]. Eliciting prolonged cell-mediated immunity by vaccination is a significant hurdle for treatment of pulmonary infection. A recent study showed that extended immunity can be induced by vaccination with a replication-deficient AdV expressing IAV nucleoprotein, which was delivered by simultaneous intranasal and s.c. injections [41]. Although this approach shows great promise for vaccine-induced immunity, several questions remain to be addressed. The size of the T_RM population in the immunized lungs could not be accurately assessed without in situ imaging [14], but flow analysis indicated that early viral clearance correlated with many antiviral CTLs in the spleen [41]. It is important to note that replication-deficient AdVs persist for an extended time in some peripheral tissues and create a chronic source of antigen that maintains the high frequencies of activated CTLs [42]. Further work will show whether the T_RM population in the lungs of the AdV-NP–vaccinated mice was replenished with activated CTLs during chronic antigen stimulation.

Prolonged presentation of viral peptides is a characteristic of IAV infection [43, 44], which supports stable CD69 expression on antiviral CTLs in the draining lymph nodes when immunity is present [45]. The T_RM population in the airways is at least partially preserved by small numbers of migrating CTLs that enter the lungs after viral clearance [24, 43]. The replenishment process may be driven by chronic antigen stimulation [46, 47]. A vaccination study showed that antigen was required for NP-specific antibodies to provide cross-protection among different strains of IAV by a mechanism that involved FcRs and CD8 T cells [48]. The data from these combined studies support a model (Fig. 2) whereby persisting antigen [43, 44] or immune complexes [49] prolong cell-mediated immunity through chronic T cell activation [42], thereby producing activated CTLs, which enter peripheral tissues [25] and supplement the T_RM population in the lungs [24, 43]. It remains to be determined whether immune complexes persist after s.c. immunization with AdV-NP and support long-term immunity by a similar mechanism [41].

Figure 2. A hypothetical model for sustained cell mediated immunity.

(1) Local infection in the lungs is required for the formation of pulmonary T_RM cells after IAV and adenovirus infection. (2) Residual viral peptides remain in the mediastinal lymph node and support sustained CD8 T cell activation after IAV infection. (3) The adenovirus vector persists in peripheral tissues after s.c. vaccination. NP-specific antibodies form immune complexes, which maintain the supply of activated CTLs in spleen, and (4) replenish the T_RM population in the lungs.

ENVIRONMENTAL CUES PROGRAM ANTIVIRAL CTL RESPONSES AND MAY YIELD NEW OPPORTUNITIES TO ENHANCE VACCINE PERFORMANCE

KLRG1 is an inhibitory molecule that has a poorly defined role in T cell biology [50]. High expression is a useful marker of polyfunctional CTLs with potent lytic activity and the capacity to produce large quantities of pathogenic cytokines [35, 36, 51]. Although KLRG1⁺ CTLs support immunity to some infections [52], most of the cells are deleted during the contraction of the T_EFF response without contributing to long-term memory [53]. Consequently, any KLRG1⁺ CTLs that develop in response to vaccine delivery have little long-term value and could damage the lungs. Discovery of new therapeutic approaches that inhibit formation of T_EFF cells during memory development could be a major advance for the treatment of pulmonary disease. Recent studies indicate that some progress has been made in this area.

Simplified experimental systems show that a brief period of antigen stimulation is sufficient for naïve CD8 T cells to embark on a self-perpetuating developmental program, which culminates in the formation of memory CD8 T cells [54]. Terminal differentiation of T_EFF cells that express KLRG1 requires additional signals from proinflammatory mediators that are released in infected tissues [55–57]. The cytokine milieu also contributes heterogeneity within the memory compartment and formation of T_RM cells. Defining the signals that control the fate decisions of activated CTLs would be a major advance toward production of safer vaccines for pulmonary infection.

Naïve CD8 T cells maintain early plasticity during clonal expansion because the progeny of individual precursors include mixed populations of T_EFF and memory CD8 T cells during infection [58, 59]. A branch in the developmental pathway leads to terminal differentiation of the T_EFF cells because the precursors of canonical memory CD8 T cells lack KLRG1 expression [56, 60]. Conversely, a few KLRG1⁺ CTLs survive the contraction of the T_EFF response after IAV infection but do not convert to a typical memory phenotype [61, 62]. Almost all the surviving KLRG1⁺ CTLs that were found in the lungs 30 d postinfection were located inside the blood vessels [61], indicating that KLRG1 marks a subset of memory CD8 T cells that is uniquely adapted for vascular surveillance. Although the relationships among different subsets of memory CD8 T cells remain obscure, environmental factors clearly have a role in the formation of T_RM cells [46, 63, 64], whereas conversion to a T_CM phenotype may be a default pathway.

The fate decisions of activated CTLs are controlled by a complex transcriptional network involving cooperative expression of T-bet and Zeb2 during a robust T_EFF response [65], whereas high eomesodermin expression favors memory formation [66, 67]. Evidence indicates that terminally differentiated T_EFF cells are a product of asymmetric cell division, which can be triggered by robust stimulation through the antigen receptor [68, 69]. Individual precursors produce daughter cells with different capacities for long-term survival after uneven inheritance of critical regulatory factors, such as T-bet, mTORC1, and c-MYC [70, 71]. Robust signaling via the antigen receptor cannot entirely account for the dichotomy in the CTL response because T_EFF cells do not up-regulate KLRG1 in vitro [72]. In vivo T_EFF generation is driven by proinflammatory molecules that are released during infection [58, 73]. Defining the cytokine milieu that supports optimal memory formation is an important priority for vaccine development. Eventually it may be possible to block the paracrine signals that drive the T_EFF response during vaccination (Fig. 3), whereas preserving heterogeneity within the memory compartment is essential for T_RM development and immunity.

Figure 3. An idealized response to vaccination.

The CTL responds to respiratory virus infection (red line). After vaccination, memory formation occurs without producing the pathogenic CTLs that damage the lungs (blue line). The shaded area shows a window when anti-inflammatory molecules might be used to block T_EFF generation.

This may be a complex task because activated CTLs receive costimulatory signals from a diverse array of proinflammatory cytokines during infection [74, 75], including IL-2 [76–78], IL-27 [79], IL-21 [80], IL-12, and type I IFN [55, 57, 81]. At least 1 additional signal is required for terminal differentiation of T_EFF cells because cultured cells do not express KLRG1 when costimulated with cytokines in vitro [72]. Although the ligand for the missing pathway remains elusive, some insightful clues about its influence on the CTL response can be found in the literature. Several studies have shown that regulation by TGF-β reduces the numbers of pathogen-specific CTLs that express KLRG1 during infection and autoimmunity, thus reducing pathology in mucosal tissues [61, 82–85]. In other experiments, circulating memory CD8 T cells converted to a KLRG1⁺ phenotype upon temporal ablation of the TGF-β receptor, without exposure to cognate antigen [86]. Revealingly, the numbers of memory CD8 T cells that expressed KLRG1 after ablation of the TGF-β receptor were reduced when the mice were pretreated with antibiotics [86], suggesting that the ligand for the unknown signaling pathway is expressed during infection or during exposure to microbial metabolites.

MOLECULAR CONTROL OF T_RM DEVELOPMENT AND TISSUE LOCALIZATION

Eliciting protective CTL populations in the lungs by vaccination is a challenging endeavor because antigens must reach the lungs during T_RM formation [19, 43, 46, 87]. Peptide antigens have a controversial role in tissue localization and may be dispensable for formation and/or maintenance of T_RM cells at some sites [88, 89]. Tonic signaling through the TcR is required for CTLs to maintain canonical markers of T_RM cells (CD69 and CD103) in tissues that lack an epithelial layer, such as the secondary lymphoid organs and brain [43, 90]. At other sites, CD69 expression may be maintained independent of the antigen receptor, including mucosal tissues, such as lungs [88, 89]. The CTLs that express CD69 do not enter the circulation when the receptor for S1P is modulated from the cell surface [91–93]. A related mechanism is required for T_RM cells to settle in peripheral tissues, such as the lungs, where cytokines promote the formation of T_RM cells and suppress expression of Kruppel-like factor 2 and the S1P1 receptor [94].

Most pulmonary T_RM cells also express αeβ7 integrin (CD103), which mediates interactions with epithelial cells that express E-cadherin and supports retention at the mucosal surface [31, 95]. Naive CD8 T cells constitutively express CD103, which may facilitate interactions with specialized DCs that express E-cadherin [96]. Expression disappears from the surface of activated CTLs soon after TCR stimulation and returns when memory CD8 T cells enter tissues containing TGF-β [46, 64]. The concentrations of TGF-β reach high levels during infection with mildly pathogenic strains of IAV and other pathogens that encode enzymes, which cleave the latent cytokine and release the active peptide that is responsible for receptor binding [97, 98]. Fewer T_RM cells were found in the lungs of IAV-infected mice after targeted deletion of CD103 or TGF-β receptor II from T cells [46, 61], indicating that interactions with E-cadherin promote retention of CD103⁺ T_RM cells at the mucosal surface. Studies suggest that CD103 has a similar role in the gastrointestinal tract [99] and in positioning of tumor-infiltrating lymphocytes within the tumor microenvironment [100, 101].

Expression of the Itgae gene (CD103) is controlled by integrated signals from the antigen receptor, and TGF-β and CD4 help [61, 63, 67, 102]. The TGF-β receptor employs multiple signaling pathways for immune regulation [103], including a cascade Smads that are involved in regulation of the Itgae gene [102, 104, 105]. Smad4 is a central adaptor protein for multiple overlapping pathways and chaperones complexes of receptor-activated Smads into the nucleus during signal transduction [106]. A recent study showed that CTLs, which lack either the TGF-β receptor or Smad4, adopt reciprocal phenotypes during IAV infection [61]. Although the TGF-β receptor-deficient CTLs did not express CD103, targeted deletion of Smad4 was sufficient for widespread CD103 expression on antiviral CTLs in peripheral and lymphoid tissues [61]. Conversely, very few Smad4-deficient CTLs expressed KLRG1 or CD62L, which were highly expressed in the absence of the TGF-β receptor. Although some Smad4-deficient CTLs survived the contraction of the T_EFF response, very few cells were found in the resting lymph nodes without CD62L expression [61]. Similar data from other models indicate that Smad4 deletion has a global effect on CD8 T cell differentiation [107, 108]. Surprisingly, the phenotype of the Smad4-deficient CTLs was not altered in the absence TGF-β receptor expression [108] [Y. Hu and L. Cauley, personal communication, December 10, 2015], indicating that an alternative signaling pathway is involved in regulation of the Itgae gene. Because Smad4 has a critical role in terminal differentiation of KLRG1⁺ CTLs, this pathway may provide new avenues for therapeutic intervention by blocking a robust T_EFF response in the lungs during vaccination.

CONCLUDING REMARKS

Understanding the signaling pathways that modify the fate decisions of activated T cells is necessary for understanding immune responses to viral pathogens and may provide the key to optimal vaccine design. T_EFF generation via external signaling is appealing to vaccinologists aiming to induce memory formation without the pathogenic CTLs that combat infection. Whether this concept will hold true for T_RM cells in mucosal tissues remains to be determined, but an ability to manipulate the signaling pathways that drive memory formation would constitute an important step toward rational design of vaccines for pulmonary infections.

AUTHORSHIP

L.S.C. wrote the article.

Glossary

AdV

adenovirus vector

BAL

bronco alveolar lavage

IAV

influenza A virus

KLRG1

killer cell lectin-like receptor G1

NP

nuclear protein

S1P1

sphingosine 1 phosphate receptor 1

Smad

Sma- and Mad-related protein

T_CM cell

central memory CD8 T cell

T_EFF cell

effector T cell

T_RM cell

resident memory CD8 T cell

DISCLOSURES

The author declares no conflicts of interest.