The lung is a respiratory organ specialized in gas exchange: within its alveoli, oxygen is extracted from the air and exchanged by carbon dioxide. A consequence of this fundamental role is the constant exposure to airborne antigens, innocuous or pathogenic. Many of these antigens elicit strong T cell responses, and understanding how those responses form is crucial to define how these immune responses can either eliminate pathogenic threats or promote unwanted responses to innocuous agents. CD4⁺ and CD8⁺ T cells are primed by antigen-presenting cells in secondary lymphoid organs, and effector cells migrate towards antigen-rich sites to perform their function (1, 2). After antigen clearance, a portion of T cells survive long-term, forming memory populations which are important to mount quick, efficient responses against secondary antigen exposure (3). Memory T cells can be divided by their migratory characteristics. Circulating memory T cells (T_CIRCM) recirculate between blood, secondary lymphoid organs, and tissues, without taking up residency; these cells can be further subdivided into central memory (T_CM), effector memory (T_EM) and, in the case of CD8⁺ T cells, long-lived effector cells (LLEC) (3, 4). In contrast, resident memory T cells (T_RM) establish long-term residency in tissues, mostly barrier tissues (5). The lung is, naturally, one of these tissues. Prior evidence strongly suggests that lung T_RM cells are pivotal in promoting local immune responses which can either be protective against pathogens (6, 7) or deleterious – for example, in response to allergens (8).

Since their discovery (9, 10), several studies aimed to define how T_RM cells form in the lung, as well as their specific function. From these studies, a few notions are relatively well-established. First, both CD4⁺ and CD8⁺ T cells can form lung T_RM or T_RM-like populations, and this is true in response to infections (11, 12) and to allergens (8). Second, while CD4⁺ lung T_RM cells are somewhat stable over time (7, 8), CD8⁺ lung T_RM cells are notoriously short-lived, with a faster rate of decay if compared to CD8⁺ T_RM cells in other tissues (13, 14). Finally, both CD8⁺ and CD4⁺ T cells form heterogeneous lung T_RM populations, with distinct transcriptional and functional characteristics (7, 14, 15). There are, however, several unanswered questions. One of the most important unsolved puzzles in the biology of lung T_RM cells lies on the nature of the signals that educate T cells to acquire a resident memory phenotype. While much evidence points out that the routes of infection (or sensitization) are paramount in defining the magnitude of the lung T_RM response, other works suggest that, at least partially, the type of antigen can dictate the homeostasis of lung T_RM cells. In this perspective article, we will briefly review previous research, provide preliminary data, and propose a hypothesis for this outstanding question.

T_RM cell development occurs through a series of processes where initial priming, T cell sensing of peripheral tissue-derived signals, and tissue microenvironmental factors play a role in the acquisition of a T_RM signature (16, 17). Because of their residency establishment inside the lung parenchyma, lung T_RM cells must acquire certain transcriptional and protein expression characteristics. Among these characteristics, T cells (a) downregulate molecules associated with tissue egress (e.g., CCR7, S1PR1 and S1PR5), as well as upregulate molecules associated with tissue retention (e.g., TGF-βRII, CD69 and/or CD103) (17–19), and (b) express chemokine receptors such as CXCR3, which sense CCL9 and/or CCL10 released in the lung parenchyma during local immune responses (12, 20). The need for sensing of lung-derived chemokines means that optimal alterations in the lung microenvironment are paramount for the formation of lung T_RM cells. These changes, such as production of CCL9 or CCL10 or of IL-33, a danger signal associated with heightened lung inflammation (21), are associated with local tissue antigen recognition. Indeed, airway infections or immunizations are very effective in the generation of lung T_RM cells, and persistent antigen in the lungs promote long-term survival of lung T_RM cells (22–26). The notion that lung initial antigen encounter is necessary for optimal lung-resident T cell responses is relatively well-established (27). In response to murine influenza, local antigen encounter is needed for CD8⁺ T_RM cell establishment (28), and the same is true for T_RM cells forming in response to Bacillus Calmette-Guerin (BCG) vaccination (23, 29).

Lung mucosal sites are often the first barrier encountered by pathogens or allergens. These sites are composed of a complex network of heterogeneous epithelial cells, peripheral nervous cells, innate and adaptive immune cells, covered by a mucous layer. Each one of these components can harness the tissue inflammation following local infections or allergen sensitization. IL-33 production by epithelial cells (21), nervous system regulation of immune responses (30), and the capture of antigens by local dendritic cells (31, 32) all play a role in the initiation and sustenance of lung T cell responses. The lung, due to its physiological role, must balance the induction of such responses with maintenance of its function of gas exchange. During viral infections, for example, the balance between pathogen clearance and immune modulation is tightly regulated by the epithelial cell-immune cell axis, and dysregulation of this balance can lead to severe tissue damage (33). Consequently, the production and release of effector molecules is likely regulated even within the lung tissue.

Due to the highly controlled immune environment in the lung, the presence of adjuvants to elucidate immune responses is widely used to enhance immunogenicity in the lung. Adjuvants (which are common components of vaccines) can increase the magnitude and durability of antiviral immunity, impacting the phenotype of recruited innate cells (34). In response to infections or airway allergen sensitization, natural adjuvants are pathogen-associated molecular patterns (PAMPs) present on viruses, bacteria, fungi, protozoans, recognized by pattern recognition receptors (PRRs) expressed by epithelial and resident immune cells. The establishment of an appropriate resistant or tolerant environment, the engagement of distinct PRR combinations, results in recruitment of immune cell types and cytokines produced (35). This, associated with a combination of cytokines, chemokines, and danger signals, offer evidence that the lung microenvironment is critical to promote lung T_RM cell establishment.

Studies on T_RM cells have focused on identification of tissue-derived signals, while understanding how priming of committed precursors in distinct secondary lymphoid organs has been less explored (36). Dendritic cells, for instance, are responsible for the imprinting of specific migratory patterns in the T cells during activation. DCs in skin-draining lymph nodes induce the preferential expression of homing molecules for entry into skin, whereas DCs in mesenteric lymph nodes elicit tropism for the small intestine (37). This indicates that T_RM cell preconditioning in an organ-dependent way already occurs during homeostasis, whereas imprinting for tissue-selective homing occurs during T cell priming. In addition, migratory DCs from different tissues might have varying capacities for TGF-β activation in draining lymph nodes, since preconditioning was less pronounced in mediastinal lymph nodes, even though these tissues had comparable induction of CD103 in naïve T cells at both sites (38). This adds another layer on how the route of antigen priming can regulate the quality and/or magnitude of lung T_RM cell establishment.

Other routes of antigen entry, such as intramuscular immunizations, can also induce lung T_RM cells, but the phenotype of these cells seems to be heterogenic, with lower proportion of cells located in the lung parenchyma (39). Thus, intramuscular immunizations have traditionally been considered poor inducers of mucosal T_RM cell responses (40, 41). Intranasal vaccination strategies can induce strong protection, as evidenced by past studies on RSV, Mtb and influenza (23, 26, 27, 42–44). However, this route has potential issues in antigen delivery to dendritic cells in the respiratory tract, perhaps due to physical barriers such as nasal clearing or mucus (45–47). A combination of intramuscular (i.e., distal antigen priming) and intranasal immunization approaches has been suggested as a candidate to enhance lung T_RM cell development in response to vaccines (41, 48, 49). Another evidence from vaccination also challenges the notion that lung antigen priming is the sole factor inducing optimal lung T_RM cells: the recent revolution in mRNA vaccines to combat SARS-CoV-2 (50) and, more recently, influenza (51). These immunizations, which are intramuscular, lead to robust lung CD4⁺ and CD8⁺ T_RM-like cell responses, as studies in mice suggest (48, 51). Future studies will be necessary to identify how mRNA vaccines promote lung T_RM cells even without intranasal priming, and whether long-lived lung T_RM cells are generated in humans. It is important to note that, although these studies suggest that intranasal immunization is not strictly necessary for lung T_RM cell responses, intranasal priming is still sufficient to improve lung T_RM cell establishment in these cases (23, 26, 27).

In contrast with evidence for route of antigen priming, other factors may also dictate the generation of lung T_RM cells, for example antigen (pathogen) load, pathogen life cycle characteristics, or the strength of TCR-MHC interaction. In mouse models of viral infection, CD8⁺ T_RM cells in brain and kidney express higher affinity to MHC class I tetramers (> 20x) than T_CIRCM cells (52). We observed a similar trend in preliminary experiments comparing influenza-specific lung T_RM cells with T_CIRCM cells ( Figure 1A ), suggesting that a selection of high-affinity clones may also happen for lung CD8⁺ T_RM cells.

A preferential selection of lung T_RM cells with high TCR affinity for antigen could either occur at the effector stage, after T cells migrated to the lung tissue, or at the priming and activation stage, in secondary lymphoid organs. Although these two hypothetical scenarios would suggest a major effect of the route of priming as a selector of lung T_RM cells, different lung viral infections induce distinct magnitudes of a lung T_RM cell response. In mice, while in response to influenza >50% of lung memory CD8⁺ T cells express the T_RM markers CD69 and CD103 (14), in response to RSV or BCG these numbers are much lower (23, 53). These differences may suggest that distinct antigen types or differences in TCR affinity regulate the establishment and phenotype of lung T_RM cells. Alternatively, they could also be explained by differences in how distinct pathogens interact with the lung immune system, for example differences in induction of cytokine production.

To test the potential contributions of route of priming versus antigen type, we infected mice with LCMV, Armstrong strain (a systemic virus with no tropism for the mouse lung) or influenza, using intraperitoneal versus intranasal infection routes. Intranasal infection with LCMV or influenza led to significantly increased numbers of lung parenchymal antigen-specific CD8⁺ T cell accumulation ( Figure 1B ). Per se, these results are indicative of the importance of airway antigen entry in the formation of lung T_RM cells. However, the magnitude of lung T_RM cell accumulation is higher in response to intranasal influenza if compared to intranasal LCMV ( Figure 1B ). A more detailed characterization of these lung T_RM cells also show that intranasal influenza is unique in promoting upregulation of CD69 and CD103, in comparison to intranasal LCMV ( Figures 1C, D ). Confirming previous findings (27, 54), intraperitoneal influenza, despite failing to promote the numerical accumulation of lung T_RM cells ( Figure 1B ), was sufficient to induce a consistent upregulation of CD69 and CD103 in a small proportion of lung T_RM cells ( Figures 1C, D ). These preliminary findings suggest that, despite an important role for the intranasal route of immunization, the acquisition of a classic lung T_RM phenotype may strongly rely on the antigen type, more specifically their natural lung tropism.

Most past studies strongly suggest that airway exposure to antigens is an important factor in the establishment of lung T_RM cells, but additional evidence from us and others also point to the antigen type, more specifically its natural lung tropism, as another regulating factor. We believe that optimal lung T_RM cell generation will take advantage of these two variables, and the magnitude of lung T_RM cell responses obeys a continuum ( Figure 2 ). In response to airway exposure to lung allergens or to respiratory infections, both lung antigen tropism and airway route of exposure are present, and consequently a strong lung T_RM cell response is mounted. On the other end of the spectrum, systemic infections with pathogens lacking lung tropism do not elicit lung T_RM cell responses. In “hybrid” scenarios, such as intraperitoneal exposure to influenza or intranasal exposure to LCMV, lung T_RM cell generation will be partial, with the magnitude of the response relying on other factors induced by either lung inflammatory responses or antigen persistence.

Some questions, however, are still unanswered. First, what is the exact influence of lung inflammatory responses (such as the ones induced by innate immune cells or epithelial cells) in this “T_RM continuum”? When considering our intranasal LCMV system, for example, the lack of a CD69/CD103 phenotype can be due to changes during T cell priming, but the lack of lung tropism of LCMV possibly also translates in decreased infectivity in the lungs. Consequently, inflammatory responses during the acute phase are expected to be lower in lungs, which could influence the local release of signals such as TGF-β, which are necessary to educate nascent T_RM cells for CD103 expression (55). CD8⁺ T_RM cells can be generated in tissues without antigen if sterile inflammation is administered to such tissues simultaneously to systemic antigen immunization. This is true for skin T_RM cells (56, 57) and female reproductive tract T_RM cells (58). Future systematic investigations on whether such “prime and pull” strategies are sufficient for lung T_RM cell generation will be important.

Another important point to consider is the fact that CD4⁺ and CD8⁺ T_RM cells, despite sharing common pathways and molecular requirements, are not the same. CD4⁺ T_RM cells typically locate outside of epithelial sites, partly due to their inability to respond to TGF-β – which is controlled by their downregulation of Runx3 (59). This is also true in the lungs, where CD4⁺ T_RM cells are mostly concentrated within the lung parenchyma, with some of them in close contact with B cells and other immune cells (7, 15). Our model heavily takes into consideration our findings with lung CD8⁺ T_RM cells, as well as the abundant literature on these cells (6, 13, 14, 28). It will be interesting to assess the relative contributions of the route of priming versus antigen tropism (versus local inflammation) for lung CD4⁺ T_RM cell establishment. In conclusion, in this perspective we provided a short review of the known literature on how lung T_RM cells form, and how lung tissue versus antigen type can influence their formation. Understanding the relative roles of each one of these variables will lead, in our opinion, to the discovery of more efficient approaches to boost the generation of lung T_RM cells that can provide protection against infections – or to block undesirable lung T_RM cell formation in response to lung allergens.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was approved by IACUC, project number A00005542-20. The study was conducted in accordance with the local legislation and institutional requirements.

BM: Conceptualization, Writing – original draft, Writing – review & editing. MM: Data curation, Formal analysis, Investigation, Writing – original draft, Writing – review & editing. HB: Conceptualization, Data curation, Formal analysis, Funding acquisition, Project administration, Resources, Supervision, Writing – original draft, Writing – review & editing.