Tissue-resident memory lymphocytes at mucosal surfaces have been shown to be critical in long-term protection following mucosal infection. Tissue-resident memory T cells (Trm) have been well characterized in animal models and in humans, while knowledge about tissue-resident memory B cells (Brm) is currently more limited. Due to specific features of tissue-resident memory lymphocytes such as cross-reactivity and polyfunctionality, Trm and Brm have been demonstrated to be an ideal target for vaccine strategies aiming to induce protection at mucosal surfaces. The route of vaccine administration, the choice of antigenic epitopes and the impact of microenvironment appeared to be crucial parameters in the development of vaccine-induced mucosal tissue-resident memory responses in animal models. However, it remains significant gaps in understanding systemic and local signals needed to establish and maintain protective mucosal Trm subsets without inducing pathogenic populations. In addition, the development of Brm is currently not well understood. Discovery of innovative recombinant antigens and identification of safe mucosal adjuvants will be crucial in the development of vaccine formulations efficient to induce Trm and Brm at mucosal surfaces. This Opinion article will describe current knowledge about mucosal Trm and Brm and vaccine approaches already tested to induce tissue-resident memory lymphocytes at mucosal surfaces. It will pinpoint gaps in knowledge. It will suggest research avenues and highlight considerations to design vaccine strategies inducing mucosal tissue-resident memory lymphocytes and it will provide suggestions to improve methodology to quantify mucosal tissue-resident memory responses.

Three memory T cell subsets have been characterized: circulating effector memory T cells are abundant in non-lymphoid tissues, circulating central memory T cells are predominant in secondary lymphoid organs and non-circulating tissue-resident memory T cells (Trm) are able to persist in non-lymphoid tissues. Trm subset was discovered in parabiosis and tissue transplantation animal studies more than 15 years ago (1) and it is the most abundant memory T cell population. Trm have the ability to reside in mucosal tissues such as lung (2), nasal (3), gut (4), skin (5–8) and reproductive tract tissues (9), following infection (10). Trm were characterized in mice, non-human primates (NHP) (11) and humans (12–14). They were mostly defined by the high expression of the adhesion molecule CD69 which can be associated with an upregulation of the αE integrin CD103 (1), and by the downregulation of molecules preventing tissue exit such as CCR7 and CD62L (10). Other markers were associated with Trm such as CXCR3, CD49a (15) (16) or CD44 (1) in specific tissues. Regarding the functions of Trm, antigen-specific CD8 and CD4 Trm in the respiratory tract were shown to be associated with a better control of viral and bacterial infections (e.g. reduction of viral load (12), limitation of intracellular replication of bacteria (17)). In addition, several animal and human studies described the polyfunctionality of influenza- (18, 19) or respiratory syncytial virus (RSV)-specific CD8 Trm responses (20) as a key feature of protective Trm. The cross-reactivity of human Trm via recognition of conserved regions was also reported in the context of influenza (21, 22) and Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) (3) infections. Another interesting feature of Trm is their potential innate-like functions helping in mucosal protection. It was demonstrated that CD8 Trm located in lung parenchyma of mice intranasally infected with influenza virus engineered to express ovalbumin as an antigen, could reduce the severity of a subsequent pulmonary bacterial infection through neutrophil recruitment (23). A similar bystander activation of T cells was observed in a Herpes Simplex Virus (HSV) murine challenge model. Indeed, it was shown that a subset of vaginal CD8 T cells which were unspecific to the HSV antigen used for immunization, could play a partial role in genital protection. In this HSV challenge model, both peripheral CD8 T cells able to migrate to inflamed vaginal tissue and vaginal CD8 Trm of irrelevant antigen specificity were involved in this innate-like function (24).

The development of tissue-resident memory B cells (Brm) following infection has been discovered in murine parabiosis (25), adoptive transfer (26) and depletion studies (27). The presence of lung Brm early post-infection was reported in mouse models of influenza (25) and pneumococcal pneumonia (28). It was shown that the induction of Brm required local encounter with antigen (25) and that Brm contributed to early plasmablast responses leading to the secretion of cross-neutralizing antibodies against viruses and bacteria (27, 28). Compared to Trm, Brm have been less characterized and specific markers have not been fully defined. However, it was shown that Brm expressed CD69, the hallmark of tissue-resident lymphocytes, in mice (25), NHP and humans (14). Other markers were also associated with Brm in mice such as CXCR3 and CD44 (25, 26, 28). In addition, antigen-specific CD73 positive and negative pulmonary Brm subsets were described in a mouse influenza model (25). In animal studies, the analysis of Brm is based on the discrimination of resident and circulatory B cells by intravenous labelling. In animal models and humans, gating strategies based on characterized memory B cell markers and similarities with Trm transcriptional profiles are also used to describe Brm subsets (29). Interestingly, a study found a potential intestinal Brm subset. Indeed, most CD19⁺CD27⁺ B cells in human intestine were CD45RB⁺CD69⁺ B cells. In addition, sets of gene expressed in lung Trm were enriched in this gut B cell subset suggesting that it could be an intestinal-resident population (30). The presence of Brm in other mucosal tissues such as skin or reproductive tract is currently unclear.

For more than two centuries, vaccination has been a successful global strategy to reduce the burden of several infectious diseases (31). Historically, the immunogenicity induced by vaccines was associated with systemic humoral response which can be easily measured in blood using antibody assays (32). However, for several decades, efforts have been put into the understanding of vaccine-induced cellular and local immune responses. Could the protective capacities of mucosal Trm and Brm be harnessed to improve immunity at mucosal barriers?

A range of vaccine approaches have especially been tested in mouse models to improve mucosal Trm development (33). The route of vaccine administration has been shown to play a crucial role. Growing evidences suggest that mucosal vaccines on their own or combined with systemic vaccines could be a promising strategy to enhance the development of mucosal Trm. For example, a study found that intranasal administration of live-attenuated influenza virus induced the development of CD4 and CD8 Trm in lungs, whereas systemic immunisation with live-attenuated influenza virus did not generate similar Trm response in mice (34). Similarly, intranasal immunization of mice with a chimpanzee adenoviral-based SARS-CoV-2 vaccine was shown to induce CD103⁺CD69⁺CD8 T cells in lungs, while vaccination by intramuscular route failed to generate pulmonary Trm cells (35). Another strategy named ‘prime and pull’ was tested to generate Trm in vaginal tract. Mice were subcutaneously immunized with an attenuated strain of HSV-2 (prime). Then, pro-inflammatory chemokines were applied to the vagina of mice in order to recruit HSV-specific CD8 T cells to this mucosal site (pull). Compared to the other experimental groups which were primed and boosted by intravaginal or subcutaneous routes only, the prime and pull strategy was the only one leading to the establishment of CD8 Trm in the genital mucosa. This study suggested that inflammation on its own could lead to the recruitment of Trm to genital mucosa and that a persistent antigen stimulation was not needed for the establishment of Trm (36). Similar findings were reported in nasal, upper respiratory tract (2) and skin (6). However, it was demonstrated that a local antigen encounter was needed to establish CD8 Trm in lungs (37–39). For instance, mice immunized by intraperitoneal route with influenza virus (prime) could exclusively generate pulmonary Trm following an intranasal immunization (pull) with CpG oligodeoxynucleotides combined with the antigen. The authors of this study suggested that circulating antigen-specific CD8 T cells could cause local tissue damage, which could play a role in Trm conversion (39). Interestingly, the importance of vaccine epitopes was pinpointed in a study describing the sequence design and immunogenicity of a CD8 T cell peptide Coronavirus Disease 19 (COVID-19) vaccine. This COVID-19 vaccine candidate was based on a range of 11 structural and non-structural SARS-CoV-2 proteins including conserved regions. The sequence was designed using SARS-CoV-2 immuno-dominant epitopes determined by screening and SARS-CoV-2 neoepitopes selected using a computational multi-neoepitopes based peptide vaccine approach, which had been shown to be safe and efficient in clinical trials evaluating a vaccine candidate against lung cancer. Following one subcutaneous vaccination in a mouse model, the COVID-19 vaccine candidate could induce a significant number of peripheral viral-specific CD8 T cells expressing Trm markers such as CD103 and CD49a, in spleen and draining lymph nodes. Even though the authors did not analyse Trm in lungs, this study highlights the importance of epitope selection to induce lymphocytes with a tissue-residency signature (40). Another study compared the functionality of T cells and Trm in lung biopsies collected for cancer suspicion in SARS-CoV-2 infected patients or individuals vaccinated with COVID-19 mRNA vaccines. The current spike-based COVID-19 mRNA vaccines were shown to induce similar SARS-CoV-2 spike-specific IFNγ CD4 T cell responses in lungs of vaccinees and convalescents, while antigen-specific CD8 T cell responses were not induced, neither in convalescents, nor in vaccinated individuals. Regarding tissue-resident memory responses in lungs, polyfunctional CD4 and CD8 Trm induction was shown to be limited post-vaccination compared to post-infection. A selection of SARS-CoV-2 epitopes, as previously described, could help to improve tissue-resident memory responses generated by the current COVID-19 mRNA vaccines (41). Determining the role of local antigen stimulation, mucosal inflammation and antigenic epitopes in specific vaccine strategies to induce tissue-resident memory lymphocytes is fundamental. These parameters may impact on the choice of the antigen, the administration route and the vector used to deliver the vaccine (10).

Regarding Brm, their induction has especially been evaluated in mouse pulmonary tissues (25). As it was determined that the establishment of pulmonary Brm post-infection required a local antigen encounter (25), it could be hypothesized that mucosal vaccination might also be beneficial to generate Brm in lung tissues at least. However, this hypothesis remains to be demonstrated. Interestingly, evidences revealed that the establishment of Brm pool did not correlate with the presence of Trm in the reproductive tract of female mice immunized with HSV (42). These data suggest that Trm and Brm responses could be induced in an independent manner at least in the reproductive tract. Trm and Brm development might have different kinetics and/or might require different local microenvironments. This correlation needs to be studied in other tissues given a lack of correlation might significantly impact on the development of vaccines targeting mucosal tissue-resident memory responses. Do specific vaccine approaches need to be developed to induce either Trm or Brm? It is currently unknown.

Vaccines able to induce long-term mucosal responses are needed to improve protection against infection at mucosal surfaces. Expanding polyfunctional and cross-reactive tissue-resident memory responses using mucosal vaccination on its own or combined to systemic vaccination looks a promising way to reach this goal. An ideal vaccine would induce controlled and balanced Trm and/or Brm responses at mucosal sites ( Figure 1 ). For this purpose, a better knowledge is needed to understand the formation of effective tissue-resident memory responses at mucosal surfaces and to determine the specific environment needed in each mucosal tissue to generate protective Trm and Brm subsets. Identifying effective mucosal adjuvants able to induce mucosal tissue-resident memory responses is a key parameter to optimize vaccine formulations ( Figure 2 ). However, it will be challenging to move vaccine candidates into clinical trials if there are not any standard procedures to quantify Trm and Brm responses in human mucosal tissues. Identification of peripheral markers correlating with Trm and/or Brm responses could be an easy way to evaluate mucosal tissue-resident memory responses post-vaccination in humans.

SP and SL have written the manuscript. All authors contributedto the article and approved the submitted version.