Research ArticleINFECTIOUS DISEASE

Resident memory CD8+ T cells in the upper respiratory tract prevent pulmonary influenza virus infection

See allHide authors and affiliations

Science Immunology  02 Jun 2017:
Vol. 2, Issue 12, eaam6970
DOI: 10.1126/sciimmunol.aam6970

T cells by a nose

The nose is the port of entry for inhaled pathogens such as influenza virus, yet little is known about nasal-specific border patrol. Pizzola et al. now report that resident memory T cells (Trm cells) in the nasal mucosa are critical for preventing pathogen spread to the lungs. In contrast to lung Trm cells, nasal CD8+ Trm cells were induced independently of local antigen and persisted in the long term. Moreover, upper respiratory tract immunization blocked influenza virus transmission to the lung, preventing severe pulmonary disease. Therefore, vaccines that induce nasal Trm cells may stop respiratory pathogens at the gates.

Abstract

Nasal epithelial tissue of the upper respiratory tract is the first site of contact by inhaled pathogens such as influenza virus. We show that this region is key to limiting viral spread to the lower respiratory tract and associated disease pathology. Immunization of the upper respiratory tract leads to the formation of local tissue-resident memory CD8+ T cells (Trm cells). Unlike Trm cells in the lung, these cells develop independently of local cognate antigen recognition and transforming growth factor–β signaling and persist with minimal decay, representing a long-term protective population. Repertoire characterization revealed unexpected differences between lung and nasal tissue Trm cells, the composition of which was shaped by the developmental need for lung, but not nasal, Trm cells to recognize antigen within their local tissue. We show that influenza-specific Trm cells in the nasal epithelia can block the transmission of influenza virus from the upper respiratory tract to the lung and, in doing so, prevent the development of severe pulmonary disease. Our findings reveal the protective capacity and longevity of upper respiratory tract Trm cells and highlight the potential of targeting these cells to augment protective responses induced to respiratory viral vaccines.

INTRODUCTION

Influenza is a highly contagious viral illness that continues to be a major public health burden, infecting 3 million to 5 million people globally per annum. Human infections with seasonal influenza A viruses tend to be initiated and localized to the upper respiratory tract (URT), whereas pandemic and zoonotic viruses often cause severe disease, possibly because of increased tropism for the lower respiratory tract (LRT) (1). However, although seasonal influenza A virus infections are generally established in the upper airways, progression to the lung is essential for the development of viral pneumonia and severe disease (2). Although the URT represents an important site for initiation of infection and/or transmission of virus via coughing, sneezing, or talking (2), little is currently known regarding the site-specific immunity present in the upper airways.

Much of what is known about the immune response after influenza virus infection of the respiratory tract has been deduced from studies of the lung (LRT), because infections in this site are associated with more severe disease and increased risk of death. Protective immunity against influenza virus is tightly correlated with the persistence of influenza-specific tissue-resident memory CD8+ T cells (Trm cells) in the lungs (3). Trm cells represent a memory T cell population that lodges in a variety of tissues (brain, intestine, female reproductive tract, salivary glands, lung, skin, and liver) (48) and are marked by the expression of CD69 and, in some cases, CD103 (9). Although Trm cells in most tissues are long-lived and self-sustaining, Trm cells within the lung have been shown to require continued replenishment from a circulating effector memory T cell pool and gradually wane over time (10). Trm cells provide superior protection against local secondary infections through direct effector functions and by promoting the recruitment of circulating memory T cells. Recent work has implicated interleukin-15 (IL-15) and transforming growth factor–β (TGFβ), as well as down-regulation of Krüppel-like factor 2 (KLF2) (11), T-bet and Eomes expression (12), and up-regulation of Hobit (13) in the formation and tissue retention of Trm cells. In addition to these universal Trm cell developmental requirements, local factors within the microenvironment in which these cells develop also influence their differentiation. Optimal development of influenza virus–specific CD8+ Trm cells in the lung requires local antigen recognition, exposure to TGFβ, and the presence of interferon-γ–secreting CD4+ T cells (1416). Currently, it is not known whether virus-specific CD8+ Trm cells also develop in the URT, whether their developmental requirements are similar to those of LRT Trm cells, or whether they contribute to the control of influenza virus infection.

Here, we investigated the deposition, longevity, and protective capacity of memory CD8+ T cells in the URT after influenza virus infection. We show that influenza virus–specific Trm cells persisted in the long term within the URT after intranasal influenza virus infection. We report that influenza-specific Trm cells in the nasal epithelia can block the transmission of influenza virus from the URT to the lung and, in doing so, prevent the development of severe pulmonary disease. These results have substantial implications for intranasal vaccines that aim to elicit protective CD8+ T cell–mediated immunity against respiratory pathogens and highlight the potential of URT Trm cells over lung Trm cells as key targets for respiratory viral vaccines.

RESULTS

Influenza virus–specific memory CD8+ T cells persist in the URT after virus infection

Although influenza virus infection induces virus-specific CD8+ Trm cells that persist in the lung parenchyma of mice (3, 17, 18), it was unclear whether CD8+ Trm cells were also embedded in the URT. To formally assess whether influenza infection induces Trm cells at this site, we adoptively transferred congenically marked (CD45.1+) ovalbumin (OVA)–specific OT-I T cell receptor (TCR) transgenic CD8+ T cells into C57BL/6 recipients (CD45.2+), which then received a total respiratory tract (TRT) infection with a recombinant influenza virus expressing the CD8 epitope from the model antigen OVA (X31-OVA). TRT infection [30 μl containing 104 plaque-forming units (PFU) of virus delivered intranasally to anesthetized mice] results in virus deposition and replication in both the lung (LRT) and nasal tissue (URT) (Fig. 1A). Immunofluorescence analyses of the nasal tissue at day 6 post-infection (p.i.) revealed the presence of CD3+ T cells in the nasal turbinates that congregated around cells that stained positive for influenza virus nucleoprotein (NP) (Fig. 1B). Most of the OT-I T cells recovered from the lung and nasal tissue on day 35 p.i. showed elevated expression of CD103 and CD69, a phenotype used to identify Trm cells (Fig. 1C). These CD103+CD69+ OT-I cells in the lung and nasal tissue appeared to emerge from a killer cell lectin-like receptor G1 (KLRG1) precursor population (fig. S1), which is consistent with previous reports showing that skin Trm cells developed from effector CD8+ T cells that lack KLRG1 expression (19). To validate that the CD103+CD69+ OT-I cells in the nasal tissue of the URT were “bona fide” Trm cells, we assessed whether they exhibited two key features of Trm cells, namely, whether they were located within the nasal tissue parenchyma and persisted with minimal replenishment from the circulating memory CD8+ T cell pool.

Fig. 1 Memory CD8+ T cells persist in the URT after influenza virus infection.

(A) Groups of five C57BL/6 mice receiving a TRT infection with 104 PFU of X31-OVA were sacrificed 3 days later, and titers of infectious virus were determined in clarified homogenates prepared from lung and nasal tissues by standard plaque assay. Symbols represent individual mice, and horizontal bars represent means ± SEM of viral titers. (B) Mice injected with 104 naïve CD45.1+CD8+ OT-I T cells and that received a TRT infection with 104 PFU X31-OVA were sacrificed 6 days later. Fluorescence microscopy (objective, 20×) was then performed to detect influenza virus NP and CD3+ T cells in the nasal tissues. (C) Mice injected with 104 naïve CD45.1+CD8+ OT-I T and infected with 104 PFU of X31-OVA were sacrificed at day 35. Representative flow cytometry plots of single-cell suspensions from the lung and nasal tissue were gated on OT-I (CD45.1+CD8+) cells and assessed for CD103 and CD69 expression. (D) Mice injected with 104 naïve CD45.1+CD8+ OT-I T cells before TRT infection with 104 PFU of X31-OVA received an intravenous injection of anti–CD8-phycoerythrin [CD8 intravenous (CD8-iv)] at day 30 p.i. and were sacrificed 5 min later for tissue harvest. For flow cytometry analysis, single-cell suspensions from the spleen, lung, blood, and nasal tissue were gated on OT-I (CD45.1+CD8+) cells and assessed for CD103 expression and location (CD8-ivlo, parenchyma associated cells; CD8-ivhi, circulating cells). (E and F) Mice seeded with 104 naïve CD45.1+CD8+ OT-I T cells and that received TRT infection with 104 PFU of X31-OVA were administered CFSE intranasally 20 days later into the URT, and the proportion of CFSE+CD45.1+CD8+ OT-I T cells in the nasal tissue and spleen was measured 3, 20, and 40 days after CFSE delivery. (E) Data are pooled from three independent experiments, with the symbols representing individual mice and the bars representing means ± SEM (n = 5 to 7 mice per group; one-way ANOVA with Tukey’s multiple comparison). ns, not significant. (F) Representative flow cytometry profiles showing expression of CFSE and CD103 on CD45.1+CD8+ OT-I T cells. (G) Microscopy of the nasal tissue highlighting the NALTs, nasal turbinates, and nasal septum [DAPI (4′,6-diamidino-2-phenylindole)]. (H) Microscopy of NALTs from mice staining of B220, CD3, and DAPI. (I) Mice injected with naïve CD45.1+CD8+ OT-I T cells received a TRT infection with 104 PFU X31-OVA and were sacrificed for analysis at day 30 p.i. Microscopy of the various areas of nasal tissue staining for CD45.1+ (OT-I cells) and DAPI. (J) Flow cytometry profiles depicting OT-I cells (CD45.1+CD8+) isolated from the NALTs and nasal tissue isolated from mice injected with naïve CD45.1+CD8+ OT-I T cells, which received a TRT infection with 104 PFU X31-OVA and were sacrificed for analysis at day 30 p.i. (K) Histograms represent the expression levels of CD103 on OT-I cells isolated from the NALTs or nasal tissue. Numbers represent MFI ± SEM. (L) Absolute number of CD103+CD69+ OT-I T cells in the NALTs and nasal tissue of mice described in (J). Data are pooled from three experiments, and bars represent means ± SEM (n = 8 mice per group; Mann-Whitney test).

To determine the anatomical location of the CD103+ OT-I cells in the nasal tissue, we used an in vivo antibody labeling method to discriminate between CD8+ T cells located within the circulation and the tissue. To do this, we intravenously injected mice seeded with naïve OT-I T cells and that received a TRT infection with X31-OVA 20 days earlier with a fluorescently coupled anti-CD8 antibody 5 min before tissue harvest. Most of the OT-I T cells in the blood were stained with the intravenously injected antibody, as did a proportion of OT-I cells within the spleen, confirming the efficiency of this in vivo labeling method (Fig. 1D). Few CD103+ OT-I cells in the nasal tissue and lung were labeled with the intravenously injected antibody, reflecting that they were likely to be located outside the blood stream and lodged within the parenchyma tissue (Fig. 1D).

We next assessed whether the CD103+ T cells in the nasal tissue persisted with minimal replenishment from the circulating population of memory CD8+ T cells. To assess this, we administered mice seeded with naïve OT-I.CD45.1+ T cells and that received TRT infection with X31-OVA 20 days earlier with a solution of carboxyfluorescein diacetate succinimidyl ester (CFSE) in a small volume via the intranasal route to limit labeling to cells of the URT. This technique labeled about 40% of CD103+ OT-I cells in the nasal tissue, as assessed 3 days after dye administration (Fig. 1, E and F). The proportion of CFSE+CD103+ OT-I T cells in the nasal tissue remained stable for a further 40 days after dye administration, indicating that nasal tissue CD103+ OT-I cells persist with minimal replenishment from the unlabeled circulating memory CD8+ T cell population and that turnover of the labeled cells was minimal.

We next assessed the distribution of CD8+ Trm cells that develop within the nasal tissue after influenza virus infection. The nasal tissue of rodents contains nasal-associated lymphoid tissues (NALTs), which are mucosal-associated lymphoid organs embedded in the submucosa of the nasal passage (Fig. 1G). NALT consists of paired lymphoid tissue located at the base of the nasal cavities at the entrance of the nasopharyngeal duct and can be identified by microscopy of nasal tissue sections as densely packed clusters of B and T cells (Fig. 1H). To determine the localization of influenza virus–specific Trm cells in the nasal tissue, mice injected with naïve OT-I.CD45.1+ T cells received a TRT infection with X31-OVA, and on day 20 p.i., mice were sacrificed and nasal tissue sections were prepared and stained to detect OT-I T cells (CD45.1+). In these sections, OT-I T cells were rare in NALTs but distributed throughout the nasal turbinate and septum (Fig. 1I). To confirm these findings, we teased NALTs away from the palate of immune mice (generated as described above), and cellular composition was assessed by flow cytometry. OT-I.CD45.1+ T cells were detected in both the NALTs and the nasal tissue (Fig. 1J). Most OT-I T cells isolated from the NALTs and nasal tissue expressed CD103, although its mean fluorescence intensity (MFI) was consistently lower on OT-I T cells isolated from the NALTs (Fig. 1K). Analysis of the distribution of CD103+ OT-I Trm cells in the NALTs and nasal tissue revealed that the vast majority of OT-I Trm cells after X31-OVA infection were localized outside the NALTs in the nasal tissue (Fig. 1L).

Trm cells isolated from nasal tissue express Trm cell signature genes

To further authenticate that the CD103+CD69+ OT-I T cells persisting in the nasal tissue were Trm cells, we assessed their expression of a set of Trm cell core signature genes (19, 20). Mice seeded with naïve OT-I cells received a TRT infection with X31-OVA, and >30 days later, subsets of OT-I memory T cells (identified on the basis of expression of CD103+ and CD69+) were purified from nasal and lung tissues (Fig. 2A). The expression of Trm cell signature genes in these cells was then compared with that in splenic central (CD62L+) and effector (CD62L) memory OT-I T cells isolated from the spleen of the same mice. Naïve OT-I T cells (CD44) were included as a control. As expected, genes defined as up-regulated in Trm cells (Ctla4, Cdh1, Hsp1a, Itgae, Rgsl1, Skil, and Xcl1) were elevated in CD103+CD69+ OT-I cells isolated from both the lung and nasal tissue (Fig. 2B and fig. S2). Likewise, genes down-regulated in Trm cells (Eomes, Fam65b, S1p1r, Sidt1, and Klf2) were lower in CD103+CD69+ OT-I cells isolated from the lung and nasal tissue compared to the splenic memory T cell populations (Fig. 2B and fig. S2). Collectively, these results demonstrate that a population of CD103+CD69+ Trm cells persists within the nasal tissue of the URT after influenza virus infection.

Fig. 2 Trm cells develop in the URT after influenza infection independently of local antigen presentation, and establishment of Trm cells is only partially dependent on TGFβ signaling.

(A) Expression of CD69 and CD103 on OT-I T cells in the lung and nasal tissue of mice seeded with 104 naïve CD45.1+CD8+ OT-I T cells, followed by TRT infection with 104 PFU of X31-OVA and analyzed on day 30 p.i. (B) Heat map depicts relative expression of Trm cell signature genes in subsets of OT-I cells sorted from the nose and lung [as indicated in (A)] and memory (CD44+) CD62L+ and CD62L OT-I cells sorted from the spleen of mice that received 104 naïve CD45.1+CD8+ OT-I cells, followed by TRT infection with 104 PFU of X31-OVA and assessed at 30 days p.i. Results are presented as expression relative to housekeeping, and data are pooled from five experiments. (C and D) The absolute number of CD45.1+CD8+ OT-I T cells in the nose (C) and lung (D) of mice seeded with either WT or Tgfbr2f/fdLck-Cre (TGFβRII KO) naïve CD45.1+CD8+ OT-I T cells, followed by TRT infection with 104 PFU of X31-OVA and analyzed on days 6, 10, and 20 p.i. Data are pooled from three experiments, and bars represent means ± SEM (n = 6 to 16 mice per group; two-way ANOVA with Sidak’s multiple comparison). (E and F) Absolute numbers of WT and TGFβRII KO CD103+CD69+ Trm OT-I in the lung (E) and nasal tissues (F) of mice at days 10, 20, and day 60 p.i. Data are pooled from three experiments, and bars represent means ± SEM (n = 6 to 19 mice per group; two-way ANOVA with Sidak’s multiple comparison). (G) Quantitative polymerase chain reaction analysis of the expression of various genes in CD103+CD69+ TGFβRII KO CD45.1+CD8+ OT-I (KO) or CD45.1+CD8+ OT-I (WT) T cells sorted from the nasal tissue of mice that received naïve TGFβRII KO CD45.1+CD8+ OT-I or CD45.1+CD8+ OT-I cells, followed by TRT infection with 104 PFU of X31-OVA and assessed at day 20 p.i. Results are presented as expression relative to housekeeping. Data are pooled from three experiments, and bars represent means ± SEM (*P < 0.01, Mann-Whitney test). (H and I) Analysis of CD103 and CD69 expression on memory CD45.1+CD8+ OT-I T cells in the lung and nasal tissue isolated from naïve mice (Nil) or mice receiving TRT infection with 104 PFU of X31 (Flu) or X31-OVA (Flu-OVA), “seeded” with in vitro–activated OT-I T cells 2 days later and analyzed on day 20 p.i. The percentage (H) and absolute number (I) of CD45.1+CD8+ OT-I cells expressing CD103 and CD69 in the lung and nasal tissue. Data are pooled from three experiments, and bars represent means ± SEM (n = 6 to 8 mice per group; one-way ANOVA with Tukey’s multiple comparison).

Developmental requirements of Trm cells in the nasal tissue

Recent reports indicate that the optimal conversion of effector cytotoxic T lymphocytes into Trm cells within the lung requires both local antigen recognition and exposure to TGFβ (3, 15, 21). Therefore, we next assessed whether Trm cells that develop in the nasal tissue of the URT were bound by the same differentiation requirements.

We assessed whether exposure to TGFβ facilitated Trm cell development in the nasal tissue of the URT. To do this, we seeded mice with either wild-type (WT) naïve OT-I T cells or OT-I cells that lacked expression of the TGFβ receptor (TGFβR) [TGFβRII KO (knockout) OT-I], and then they received a TRT infection with X31-OVA; the infiltration of WT and TGFβRII KO OT-I cells into the lung and nasal tissue and their conversion into CD103+CD69+ Trm cells were assessed. There was little difference between the total number of WT and TGFβRII KO OT-I cells in the nose and lung days 6 to 20 p.i. (Fig. 2, C and D); however, the proportion of these cells that converted into Trm cells differed at each site (Fig. 2, E and F). Consistent with previous reports, assessment of the number of CD103+CD69+ OT-I cells that had elevated expression of the Trm cell markers CD103 and CD69 on day 60 revealed fewer TGFβRII KO OT-I cells compared with WT OT-I T cells that expressed these markers in the lung up to day 60 p.i. (Fig. 2E). However, in the nasal tissue, there was only a minor reduction at best (2.5-fold) in the number of TGFβRII KO OT-I cells expressing CD103 and CD69 compared with WT OT-I T cells (Fig. 2F). To gain further insight into the identity of the CD103+CD69+ T cells that develop in the nasal tissue independently of TGFβ signaling, we assessed these cells for the expression of key Trm cell signature genes. Critically, most of the genes in our Trm cell signature panel (Chd1, Itgae, Skil, Xcl1, Eomes, Fam65b, S1pr1, Sidt1, and Klf2) were all expressed in these CD103+CD69+ TGFβR KO OT-I cells at levels similar to those observed in TGFβ-sensitive CD103+CD69+ WT OT-I cells isolated from the nasal tissue (Fig. 2G). The exception, Ctla4, was significantly elevated in the TGFβRII KO CD103+CD69+ OT-I cells compared with their WT equivalents, implicating a role for TGFβ as a regulator of expression of this gene. Although TGFβ exposure boosts Trm cell differentiation in both the lung and nasal tissue after influenza virus infection, here we show that cells with a phenotypic and genetic Trm cell signature profile can develop in both of these sites, albeit with varying efficiencies, independently of TGFβ signaling.

We next examined the requirement for local cognate antigen recognition in Trm cell differentiation in the URT. To do this, mice received a TRT infection with a parental strain of influenza virus (X31) or the strain that had been engineered to express the model antigen OVA (X31-OVA) 2 days before receiving 106 in vitro–activated effector OT-I CD8+ T cells. Although effector OT-I T cells will be recruited to the lung and nasal tissue in both cohorts of mice, only in the group infected with the X31-OVA virus will interactions with local cognate antigen occur. The presence of CD103+CD69+ OT-I Trm cells in the respiratory tract was then assessed by flow cytometry at day 20 p.i. in both cohorts of mice. As previously demonstrated, OT-I Trm cells only developed in the lung of mice that received X31-OVA infection, validating the requirement for local antigen recognition in lung Trm cell development (Fig. 2, H and I). In contrast, a similar proportion of OT-I cells converted into Trm cells in the nasal tissue of mice infected with either X31-OVA or X31 (Fig. 2H). Hence, the requirement for antigen in driving lodgment of Trm cells differed between the URT and the lung, with Trm cells being deposited in the nasal tissues via an “antigen-independent” mechanism (Fig. 2, H and I). Nevertheless, the presence of local antigen within the nasal tissue did increase the number of OT-I cells that converted into Trm cells, which is consistent with recent reports showing that local antigen also boosts the number of Trm cells that develop in the skin (22). Furthermore, local inflammation within the nose in response to influenza infection did increase the proportion of effector OT-I cells that converted into Trm cells, and in the absence of inflammation, the conversion of the OT-I cells into nasal Trm cells was less efficient (Fig. 2H), which is consistent with recent reports showing that inflammatory environments favor Trm cell development (10).

Trm cells that develop in the URT persist with minimal decay and provide long-term protection against influenza virus rechallenge

Trm cells that deposit within the murine lung after influenza virus infection undergo attrition, limiting their potential to provide long-term protection against future infections (3). We assessed whether Trm cells lodged within the nasal tissue after influenza infection were also susceptible to the same numerical erosion. To compare the persistence of influenza virus–specific Trm cells in the nasal tissue and lung, mice injected with naïve OT-I T cells received a TRT infection with X31-OVA, and the number of CD103+CD69+ OT-I Trm cells in these tissues 20, 60, 100, and 120 days later was measured. Although the number of CD103+CD69+ OT-I Trm cells in the lung declined ~100-fold between days 20 and 120 p.i., this population of cells remained relatively stable in the nasal tissue, declining less than 2-fold over the same time period (Fig. 3A).

Fig. 3 Trm cells that develop in the URT persist with minimal decay and provide long-term protection against influenza virus rechallenge.

(A) Mice injected with 104 naïve CD45.1+CD8+ OT-I T cells and that received a TRT infection with 104 PFU X31-OVA were sacrificed at various time points p.i., and the absolute numbers of CD103+CD69+ OT-I T cells (Trm cells) in the lung and nasal tissue were determined. Symbols represent individual mice, and bars represent means ± SEM (n = 6 to 29 mice per group; data are pooled from five independent experiments; two-way ANOVA with Sidak’s multiple comparison). (B and C) Mice injected with 104 naïve CD45.1+CD8+ OT-I T cells and that received a TRT infection with 104 PFU X31-OVA were rechallenged 20 or 120 days later via a TRT infection with PR8-OVA, and the viral titers in the lung (B) and nasal tissue (C) were determined 3 days later. Symbols represent individual mice. Data are pooled from two experiments, and bars represent means ± SEM (n = 5 to 8 mice per group; one-way ANOVA with Tukey’s multiple comparison).

We next determined whether the persistence of Trm cells in the nasal tissue conferred prolonged local protection against secondary influenza virus infection. To assess this, we rechallenged mice seeded with naïve OT-I T cells and infected with a TRT X31-OVA infection 20 or 120 days later with a heterologous influenza virus, which also expresses the CD8 epitope from the model antigen OVA (PR8-OVA). The use of a heterologous virus that expresses different hemagglutinin and neuraminidase proteins to the primary strain allows the contribution of the humoral response to protection to be excluded. Viral growth in the nasal tissue and lung was determined 3 days later.

Consistent with previous reports (3), we observed that the capacity of influenza virus–infected mice to control a secondary heterologous influenza virus infection of the lung waned over time. Whereas mice rechallenged 20 days after the primary influenza infection exhibited a 20-fold reduction in influenza viral titers in the lung (Fig. 3B), mice rechallenged 120 days after the primary infection displayed minimal protection with viral titers in the lung tissue being comparable with that seen in a naïve cohort that only received the challenge virus. The protection against secondary heterologous influenza infection within the nasal tissue did not decay over the time course of this experiment. We observed significant reduction in viral titers in the nasal tissue of mice that were rechallenged either 20 or 120 days after the primary infection (Fig. 3C), which correlated with the stable persistence of influenza-specific Trm cells within the nasal tissue. The protective capacity and longevity of URT Trm cells mark these cells as ideal targets for respiratory viral vaccines.

Different immunodominance hierarchies among influenza-specific Trm cells in the lung and nasal tissue

We next assessed whether we could achieve the same pattern of Trm cell lodgment along the respiratory tract when evoking an endogenous CD8+ T cell response directed against influenza viral proteins. Whereas viruses express many proteins, the virus-specific CD8+ T cell responses generally segregate into reproducible hierarchies with certain epitopes eliciting either immunodominant or subdominant responses after virus infection. We tracked the influenza virus–specific CD8+ T cell response against two classically defined immunodominant epitopes derived from the viral nucleoprotein (NP366) and acid polymerase (PA224) and a subdominate epitope derived from the viral polymerase (PB1703) using H-2Db tetramers loaded with the PA224 and NP366 epitopes or H-2Kb tetramers loaded with the PB1703 epitope. After a TRT infection with PR8, sizable populations of NP366-, PA224-, and PB1703-specific CD8+ T cells could be detected in the spleen, lung, and nasal tissue of mice at day 10 p.i., and although numbers contracted thereafter, long-term memory cells of all specificities were detectable in each of these sites at day 60 p.i. (Fig. 4, A and B). Consistent with previous reports (21, 23, 24), splenic NP366-specific CD8+ T cells numerically dominated over PA224- and PB1703-specific CD8+ T cells at day 10 p.i.; however, this effect was lost as cells progressed into long-term memory. Similarly, we observed the same pattern of immunodominance in the nasal tissue (Fig. 4, A to C), as previously reported (24). We observed a different pattern of immunodominance in the lung where equivalent numbers of each of the three CD8+ T cell specificities were recovered at the acute (day 10 p.i.) and late memory (day 60 p.i.) time points, although the numbers of each had dropped markedly by day 60 p.i. (Fig. 4, A and B).

Fig. 4 Different immunodominance hierarchies among influenza-specific Trm cells in the lung and nasal tissue.

Absolute numbers of NP366-, PA224-, and PB1703-specific CD8+ T cells in the spleen, lung, and nasal tissue of mice that received a TRT infection with 50 PFU of PR8 were determined at days 10 (A) and 60 p.i. (B). Data are pooled from two independent experiments, and symbols represent means ± SEM (n = 5 to 11 mice per group; two-way ANOVA with Tukey’s multiple comparison). (C to F) Expression of CD69 and CD103 on NP366-, PA224-, and PB1703-specific CD8+ T cells in the lung and nasal tissue of mice that received a TRT infection with 50 PFU of PR8 and analyzed on day 60 p.i. The percentage (D) and absolute number (E and F) of CD103+CD69+ NP366-, PA224-, and PB1703-specific CD8+ T cells in the nasal tissue (E) and lung (F) on day 60 p.i. are shown. Data are pooled from two experiments, and bars represent means ± SEM (n = 11 mice per group; two-way ANOVA with Sidak’s multiple comparison).

To assess whether these influenza virus–specific CD8+ T cells converted into Trm cells along the respiratory tract, we analyzed mice receiving a TRT infection with PR8 on day 60 p.i. for the expression of the Trm cell markers CD103 and CD69 by NP366-, PA224-, and PB1703-specific T cells recovered from the lung and nasal tissue. We observed that NP366-, PA224-, and PB1703-specific T cells differentiated into CD103+CD69+ Trm cells with different efficiencies depending on whether they developed into Trm cells within the lung or nasal tissue (Fig. 4, C to F). In the nasal tissue, similar proportions (~55 to 65%) of NP366-, PA224-, and PB1703-specific T cells had up-regulated Trm cell markers at day 60 p.i. (Fig. 4D), and all three CD8+ T cell specificities represented equivalently sized Trm cell populations at this site (Fig. 4E). In the lung, ~60 to 70% of PA224- and PB1703-specific T cell pools had differentiated into CD103+CD69+ Trm cells by day 60 p.i., whereas the proportion of NP366-specific T cells up-regulating these markers (~20%) was much lower (Fig. 4D). The poor conversion of NP366-specific T cells into Trm cells resulted in a significant reduction in the absolute number of NP366-specific Trm cells in the lung compared to the other specificities (Fig. 4F). Using the in vivo anti-CD8 labeling approach described earlier, we confirmed that the reduction of CD103+CD69+ Trm NP366-specific cells in the lung did correlate with the reduction of NP366-specific cells within the lung parenchyma tissue as coexpression of these markers identify cells localized to this compartment (fig. S3). Overall, we observed distinct immunodominance hierarchies within influenza virus–specific Trm cell populations in the lung (PB1703 > PA224 > NP366) compared to the nasal tissue (PB1703 = PA224 = NP366) (Fig. 4, D to F). Collectively, these data show that the selection of different CD8+ T cells specificities into the Trm cell pool is influenced by local factors within the microenvironment where they develop.

Relative epitope abundance modulates the immunodominance hierarchy within the lung Trm cell pool

The development of Trm cells within the lung is heavily dependent on recognition of local cognate antigen. We speculated that changes in viral protein production and, thus, epitope availability within the lung during virus infection might influence the immunodominance hierarchy and result in the preferential selection of different specificities into the Trm cell pool. To address this, mice primed by intraperitoneal infection with PR8 received bone marrow–derived dendritic cells (BMDCs) pulsed with equivalent doses of NP366 and PA224 peptides or unpulsed BMDCs as control [and adjuvant; lipopolysaccharide (LPS)] 6 days later via the intranasal route. The development of NP366- and PA224-specific Trm cells in the lung was then determined 21 days after BMDC immunization (Fig. 5A). In the spleens of control mice (infected with PR8 via the intraperitoneal route and given unpulsed BMDCs), the numbers of NP366-specific CD8+ T cells exceeded the number of PA224-specific CD8+ T cells (Fig. 5B), whereas in animals intranasally immunized with peptide-pulsed BMDCs, the numbers of NP366- and PA224-specific CD8+ T cells in the spleen were equivalent (Fig. 5B). As expected, the proportions (Fig. 5C) and numbers (Fig. 5D) of NP366- or PA224-specific CD8+ Trm cells were negligible in the lungs of control animals, consistent with the requirement for local antigen recognition in lung Trm cell development (Fig. 5, C and D). In contrast, mice boosted by intranasal immunization with BMDCs pulsed with NP366- and PA224 peptide or BMDCs pulsed with the individual peptides (fig. S4) developed CD103+CD69+ Trm cells in the lung, and delivery of equivalent doses of influenza virus epitopes resulted in equivalent proportions and numbers of NP366- and PA224-specific CD8+ T cells developing into lung Trm cells (Fig. 5, C and D). These data indicate that the immunodominance hierarchy and selection of CD8+ T cell specificities into the Trm cell pool within the lung may be a consequence of relative epitope abundance within the tissue over the course of the virus infection. In contrast, in the nasal tissue where there is no requirement for local cognate antigen recognition in Trm cell development, all CD8+ T cell specificities developed into Trm cells with equivalent efficiencies (Fig. 4, D and E).

Fig. 5 Relative epitope abundance modulates the immunodominance hierarchy within the lung Trm cell pool.

(A) Mice were injected via the intraperitoneal route with 104 PFU of PR8 and, 6 days later, were inoculated via the intranasal route with BMDCs pulsed with NP366 and PA224 peptide or unpulsed BMDCs in the presence of LPS. Then, 21 days after BMDC immunization, mice were sacrificed, and the absolute number of NP366- and PA224-specific CD8+ T cells in the spleen (B) and the proportion (C) and absolute number (D) of CD103+CD69+ NP366- and PA224-specific CD8+ T cells in the lung were determined. Data are pooled from two experiments, and bars represent means ± SEM (n = 6 mice per group; two-way ANOVA with Sidak’s multiple comparison).

Lodging influenza virus–specific Trm cells in the URT prevents pulmonary influenza virus disease

In humans, influenza A virus infection is initiated and largely confined to the URT, although migration of the virus into the lungs can occur in high-risk groups or with highly virulent influenza strains. Because the transmission of influenza virus from the URT to the lung is often associated with more severe disease, the capacity of Trm cells in the URT to prevent this transition was assessed. To do this, we established infection protocols that allow the selective deposition of Trm cells in either the URT (nose) or TRT (nose and lung) by exploiting the differential requirement for local antigen recognition in Trm cell development in these regions. The mouse-adapted PR8 strain of influenza virus does not disseminate efficiently from the URT (nasal tissue) to the URT (lung) (25). Therefore, when PR8 is administered as a URT infection (10 μl to the nares; no anesthetic), virus replication is largely localized to the upper airways with limited replication in the lung (Fig. 6, A and B). In contrast, after TRT infection (30 μl to anesthetized mice), PR8 replicates in the nose and to high titers in the lung (Fig. 6, A and B).

Fig. 6 Influenza-specific Trm cells in the URT blocks pulmonary influenza virus infection.

Viral titers in nasal tissue (A) and lungs (B) of mice 5 days after URT infection or TRT infection with PR8. Data are pooled from two independent experiments; symbols represent titers from individual mice, and bars represent means ± SEM (n = 5 to 6 mice per group; Student’s t test). (C) Flow cytometry analysis of CD103 and CD69 expression on NP- or PA-tetramer+ cells in the lung and nasal tissue of mice receiving a URT or TRT infection with PR8 and analyzed on day 30 p.i. (D to F) Mice injected intraperitoneally with PR8 [intraperitoneal (ip) PR8] or that received a URT or TRT infection with PR8 infection were sacrificed and analyzed on day 30 p.i. (D) The absolute number of CD8+ NP- and PA-tetramer+ cells in the spleen and the absolute number of CD8+ NP- and PA-tetramer+ CD103+CD69+ Trm cells in the lung (E) and nasal tissue (F) were determined. Data are pooled from two independent experiments, and bars represent means ± SEM (n = 9 to 18 per group; one-way ANOVA followed by Sidak’s multiple comparison test). (G to J) Mice generated as described in (D) to (F) were challenged on day 30 p.i. with X31 via a URT infection, and the viral titers in the nasal tissues (G and I) and lung (H and J) were determined 3 days (G and H) or 5 days (I and J) later. Data are pooled from two independent experiments; symbols represent individual mice, and bars represent means ± SEM (n = 8 to 10 per group; one-way ANOVA followed by Sidak’s multiple comparison).

We addressed the impact of restricting influenza virus to the URT or TRT on Trm cell development within these compartments. Mice were infected via either the URT or TRT with PR8, and at day 30 p.i., the numbers of NP366- and PA224-specific Trm cells in the lung and nasal tissue were determined, as well as the proportion of influenza tetramer+ cells in the spleen (as a reflection of the circulating memory T cell pool). Additionally, a control group of animals received an intraperitoneal inoculation with PR8. As seen in Fig. 6D, NP366- and PA224-specific cells were observed in the spleens of all cohorts of mice. Animals that received PR8 by the intraperitoneal route alone did not develop influenza-specific Trm cells in the lungs or nasal tissue (Fig. 6, E and F). After TRT infection, CD103+CD69+ NP366- and PA224-specific Trm cells were detected in both the lung (Fig. 6, C and E) and nasal tissue (Fig. 6, C and F); however, after URT infection, tetramer+ Trm cells were only detected in the nasal tissue (Fig. 6, C, E, and F). Thus, we have developed an experimental model whereby influenza virus–specific Trm cells are lodged within the upper airways but not in the lung.

We next determined whether lodging influenza virus–specific Trm cells in the URT alone could prevent the development of pulmonary influenza virus infection. Mice infected as described above (URT, TRT, or intraperitoneal PR8) were rested for 30 days before a URT rechallenge with X31 (10 μl to the nares; no anesthetic), a heterologous virus strain capable of migrating into the LRT after URT infection. At day 3 p.i. with X31, mice were sacrificed, and viral growth in the respiratory tract was determined. Compared to naïve mice that received a primary infection with X31, the presence of circulating memory T cells alone (intraperitoneal PR8 cohort) did not result in any significant reduction in viral titers in the nasal tissue (Fig. 6G) or the lung (Fig. 6H). In contrast, when the circulating memory T cell pool was supplemented by depositing influenza virus–specific CD8+ Trm cells in the nasal tissue and lung (TRT PR8 cohort), virus growth in both the lung and nose was greatly reduced (Fig. 6, G and H). Depositing Trm cells in the nasal tissue alone (URT PR8) was just as effective at reducing X31 growth in both the nose and the lung (Fig. 6, G and H). To confirm that deposition of Trm cells in the URT did not just delay the migration of the virus from the nasal tissue to the lung, we repeated the abovementioned experiment but assayed the nasal tissue and lung 5 days after secondary X31 challenge. Again, we observed that deposition of Trm cells in the nasal tissue alone (URT PR8) significantly reduced X31 growth in both the nose and the lung (Fig. 6, I and J). The depletion of all memory CD8+ T cells from these PR8-primed mice (circulating and resident) before heterologous rechallenge with X31 abrogated the reduction in viral titers in the nasal tissue, validating that the control of influenza virus within this tissue was CD8+ T cell–mediated (fig. S5). These data highlight the effectiveness of lodging Trm cells in the URT as an approach to limit the development of viral pneumonia in the lung.

DISCUSSION

The respiratory tract is divided anatomically into upper (including the nose, mouth, and pharynx) and lower (trachea, bronchi, and lungs) compartments and is a major portal through which viruses enter the body. The URT represents an important site for initiation of infection and/or transmission of many human respiratory pathogens, including influenza virus. Despite this, little is known of the immunity that can be evoked within the URT. Therefore, we investigated the deposition, protection, and recall response of influenza virus–specific memory CD8+ T cells in the URT. These studies showed that influenza virus–specific Trm cells developed within the URT after intranasal influenza virus infection. Moreover, URT Trm cells rapidly cleared a secondary heterosubtypic influenza infection from the nasal mucosa and, in doing so, prevented virus spread into the lung. The URT is a key region to limiting viral spread to the LRT and, therefore, in preventing the development of viral pneumonia and severe disease.

Whereas virus-specific Trm cells lodge both within the URT and within LRT after influenza virus infection, the differentiation requirements of Trm cells that develop in these compartments varied markedly. Nasal tissue (URT) Trm cells developed independently of local antigen recognition and were not heavily dependent on exposure to TGFβ for their differentiation. In contrast, optimal differentiation of lung Trm cells required both local antigen recognition (3, 18) and TGFβ signaling (21). The mechanism underlying the differences in Trm cell differentiation in the nasal tissue and lung remains unknown, and although Trm cell development is most likely regulated by the tissue microenvironment and the local microbiota, we have yet to identify the factors that are essential for nasal tissue Trm cell development. Defining these conditions will be important to facilitate the development of vaccines that lodge Trm cells in the upper airways.

Although all the molecular factors that drive the differentiation and retention of Trm cells in their tissue of residency are not completely understood, TGFβ has been highlighted as one factor that facilitates Trm cell development in a variety of tissues. TGFβ signaling has been shown to induce the up-regulation of CD103 on Trm cell precursors and drives the down-regulation of t-box transcription factors Eomes and T-bet (12)—this series of events supports the deposition and retention of Trm cells in skin, gut, and lung (19, 21, 26, 27). Although TGFβ exposure improves Trm cell differentiation in both the lung and nasal tissue after influenza virus infection, here we show that cells with a phenotypic and genetic Trm cell signature can develop in both of these sites, albeit with varying efficiencies, independently of TGFβ signaling. In particular, the absence of TGFβ had minimal impact on the development and long-term persistence of Trm cells within the nasal tissue. Whereas identifying the other environmental cues evoking Trm cell differentiation in the nasal tissue requires further investigation, potential candidates may include exposure to IL-15, tumor necrosis factor–α, and IL-33 (11), because all these cytokines have been implicated in Trm cell differentiation. This work further highlights the diversity among Trm cells that reside in different anatomical sites and emphasizes the local microenvironment as a key factor influencing the differentiation and maintenance of these cells.

Whereas viruses express many proteins, virus-specific CD8+ T cell responses are often directed at only a fraction of all the potential antigenic epitopes. The relative sizes of viral epitope–specific CD8+ T cell responses generally segregate into reproducible hierarchies with certain epitopes eliciting either immunodominant or subdominant responses after virus infection. Earlier studies report that the immunodominance hierarchy of memory CD8+ T cells recovered from the lung and nasal tissue after influenza infection mirrored the pattern observed within the circulating memory CD8+ T cell compartment (24, 28). However, these studies profiled bulk influenza-specific memory CD8+ T cells within the upper and lower airways, which comprise a mixture of both Trm cells and circulating memory T cells. Here, we show that the pattern of immunodominance observed among purified influenza virus–specific Trm cells isolated from the lung does not match the hierarchy observed within the circulating influenza virus–specific memory CD8+ T cell compartment. We observed diversity within the immunodominance hierarchies among influenza virus–specific Trm cell pools located in the URT and LRT.

Investigations on the efficiency, in which different influenza-specific CD8+ T cell specificities within a polyclonal response were recruited into the lung and nasal tissue Trm cell pool, revealed notable differences in the immunodominance hierarchy at these sites. In the nasal tissue, NP366-, PA224-, and PB1703-specific CD8+ T cells develop into Trm cells with equivalent efficiencies. As Trm cell development in the nasal tissue does not require local cognate antigen recognition, all CD8+ T cell specificities recruited to this site have an equal opportunity to develop into Trm cells because there is no local selection process fine-tuning the repertoire. In contrast, in the lung, we observed that NP366-, PA224-, and PB1703-specific CD8+ T cells adopted Trm cell status with varying efficiencies ranging from 20 to 70% Trm cell conversion. Because Trm cell development in the lung is heavily dependent on local cognate antigen recognition, changes in viral protein production and epitope availability within the lung over the course of virus infection are likely to govern the selection of different specificities into the Trm cell pool. It is known that different influenza virus epitopes are presented by different cell types in the lung, and this would likely result in a difference in the kinetics of antigen presentation of these epitopes over the course of virus infection (29). The tailoring of lung Trm cell diversity through antigen recognition is likely driven by a need for strict numeric limitations, given the sensitivity of this site to immunopathology. Our data support recent findings that show antigen-dependent cross-competition as a mechanism that shapes the repertoire of polyclonal antiviral Trm cells (30). Thus, antigen presentation within the tissue over the course of an infection is a factor that influences the recruitment of CD8+ T cell specificities into the Trm cell pool, especially within microenvironments where Trm cell development is antigen-dependent. Understanding what shapes the immunodominance hierarchy within these different memory CD8+ T cell compartments and at these different anatomical locations is important, particularly with regard to developing vaccines that elicit effective memory CD8+ T cell responses.

Although mouse models and more recently clinical studies (31) confirm that Trm cells protect the lung from respiratory pathogens, this protection is only transient because Trm cells in the lung, unlike populations in the skin and intestinal mucosae, undergo attrition and decay in number, leaving the lung susceptible to reinfection (3). Slütter et al. (10) recently reported that lung Trm cells are not self-sustaining and required continued replenishment from the circulating effector memory CD8+ T cell pool to persist. The time-dependent loss of effector memory CD8+ T cells from the circulating memory CD8+ T cell compartment depleted the system of lung Trm precursor cells, and this was proposed to underlie the decay of lung Trm cells. The gradual loss of lung Trm cells may represent a mechanism to remove cells capable of evoking excessive inflammation and pathology from the lung, an organ noted for its intolerance for inflammation (32). Whereas vaccination strategies that deposit influenza virus–specific Trm cells in the lung provide exquisite protection against heterosubtypic influenza challenge (21), the decay of lung Trm cells diminishes their potential for long-term protection against future infections (3). Here, we show that Trm cells in the nasal tissue of the URT are not susceptible to the same erosion and can provide long-term protection against secondary influenza virus infections. Trm cells present within the URT could also limit transmission of influenza virus from the upper to the lower airways. Hence, vaccination strategies that deposit Trm cells within the URT could safeguard the TRT and provide longer-lasting protection against respiratory pathogens.

FluMist, a live attenuated influenza vaccine (LAIV) currently on the market, is administered in the form of a nasal spray. Intranasal immunization with LAIV delivers the weakened influenza virus strains into the upper airways, and because of modifications that restrict replication to lower temperatures of the URT (<33°C), this is where the virus remains (3335). Although recent reports by Zens et al. (36) demonstrate that immunization of mice with LAIV results in highly protective humoral and cellular immune responses, as well as the development of lung Trm cells, the immunological outcomes of LAIV immunization in humans are ambiguous and likely affected by differing levels of preexisting influenza immunity between recipients (37). Immunization with LAIV does elicit robust influenza-specific neutralizing antibodies in humans, and this response is initiated in the mucosal-associated lymphoid tissue that services the upper airways (palatine tonsils) (38). However, the influenza-specific CD8+ T cell response evoked by the LAIV is highly variable (3941).

Currently, it is not known why the LAIV fails to elicit robust influenza-specific CD8+ T cell responses. One possibility is that preexisting humoral immunity in humans limits the replication of LAIV in the upper airways, providing sufficient antigen to boost B cell responses but insufficient for effective CD8+ T cell priming, expansion, and/or activation. Moreover, the requirements for effective B cell priming or expansion may be anatomically distinct in the human airways to those for CD8+ T cells, such that only effective humoral responses are elicited. Defining CD8+ T cell priming, expansion, activation, and decay in humans in response to LAIV is critical to uncover the immunological mechanisms underlying the apparent inability to induce effective CD8+ T cell (including Trm cell) immunity. Although necessary, the design and execution of such studies in humans are extremely challenging. Greater understanding of the mechanisms involved in inducing effective cellular immunity and URT Trm cell development after intranasal immunization will drive the development of effective new strategies that safeguard the upper airways from respiratory pathogens.

Our results demonstrate that heterosubtypic immunity to respiratory influenza challenge is tightly correlated with the presence of influenza virus–specific Trm cells within the respiratory tract. Our findings reveal the protective capacity and longevity of URT Trm cells and highlight the potential of these cells over lung T cells as key targets for respiratory viral vaccines.

MATERIALS AND METHODS

Study design

The main aim of the study was to characterize the development of influenza virus–specific Trm cells in the nasal tissue and assess the capacity of these cells to block influenza infection in the upper airways and stop the development of severe pulmonary disease. For this purpose, we adopted a mouse model of influenza virus infection and assessed nasal Trm cells development using a model antigen system and transgenic OT-I CD8 T cells and complemented this work by tracking the endogenous CD8 T cell response.

All experiments were performed at least twice. The study involved sublethal infections with influenza A/Puerto Rico/8 (H1N1) PR8 or Hong Kong/31 (H3N2) X31 influenza viruses strains or these parental strains that were engineered by reverse genetics to express the model antigen OVA. No outliers were excluded from the data analyses.

Mice and viruses

C57BL/6, OT-I.CD45.1, and Tgfbr2f/fdLck-Cre OT-I.CD45.1 mice were bred in-house and housed under specific pathogen–free conditions in the animal facility at the Peter Doherty Institute of Infection and Immunity, University of Melbourne, Melbourne, Australia. All experiments were done in accordance with the Institutional Animal Care and Use Committee guidelines of the University of Melbourne. Mice were infected intranasally with X31-OVA or PR8-OVA (42) (which encodes the OVA257–264 epitope within the neuraminidase stalk) provided by S. Turner, Monash University, Melbourne, Australia. For TRT infection, mice were anesthetized with inhalation isoflurane anesthetic and infected with 50 PFU of PR8 or PR8-OVA or 104 PFU of X31 or X31-OVA in a volume of 30 μl. For URT infection, 10 μl of 104.5 PFU of PR8, PR8-OVA, or X31 virus was placed onto the nares of unanesthetized mice.

In vivo labeling of T cells with carboxyfluorescein diacetate succinimidyl ester (CFSE) or anti-CD8 and CD8 depletion

CFSE

A total of 10 μl of 5 mM CFSE diluted in RPMI was placed onto the nares of unanesthetized mice.

Anti-CD8 antibody labeling

Mice were injected intravenously with 3 μg of phycoerythrin-conjugated antibody to CD8 (clone YTS-169) 5 min before they were sacrificed. Mice were perfused with phosphate-buffered saline (PBS), and tissues were collected, processed, and stained with allophycocyanin-conjugated antibody to CD8 (anti-CD8; clone 53-6.7, eBioscience).

Anti-CD8 depletion

Mice were injected with 200 μg of anti-CD8 (clone 2.43, Walter and Eliza Hall Institute Antibody Facility) for three consecutive days or normal rat serum and then every second day for the duration of the experiment.

Influenza virus plaque assay

Plaque assay for influenza virus was performed as described previously (43).

In vitro activation of OT-I cells

OT-I T cells were activated in vitro with 10−6 M SIINFEKL peptide–pulsed splenocytes, as previously described (4). Mice were injected with 106 in vitro–activated effectors (CD103CD69).

Adoptive transfer and isolation of naïve T cells

Naïve OT-I CD8 T cells isolated from OT-I TCR transgenic mice were purified from single-cell suspensions prepared from the lymph node (LN) and spleen. Cells were purified after a depletion step using antibodies against CD11b (M1/70), F4/80, Ter-119, Gr-1 (RB6), major histocompatibility complex class II (M5/114), and CD4 (GK 1.5), followed by incubation with anti-rat immunoglobulin G–coupled magnetic beads (Dynal Biotech) following the manufacturer’s protocols. Naïve OT-I T cell preparations were 90 to 95% pure as determined by flow cytometry.

BMDC generation and immunization

Bone marrow flushed from tibias and femurs of C57BL/6 mice was resuspended at 1 × 106/ml in RPMI 1640 supplemented with 2.5 mM Hepes, 5.5 × 10−5 M mercaptoethanol, penicillin (100 U/ml), streptomycin (100 μg/ml), 5 mM glutamine, 10% fetal bovine serum (FBS), and granulocyte-macrophage colony-stimulating factor (10 ng/ml). Cells were incubated at 37°C with 10% CO2 and cultured for 6 days with a medium change on day 3. Dendritic cells were loaded with 10−6 M NP336 and PA224 peptide at 37°C for 45 min. Dendritic cells were washed and resuspended in 30 μl of PBS, and 2 × 106 cells with 1 μg of LPS were delivered intranasally into mice.

Isolation of nasal tissue and NALTs

The nasal tissue, including the nasal cavity, nasal turbinates, and NALTs, was obtained by cutting down the vertical plane of the skull and scraping out the tissues and small bones from both sides of the nasal passages. The NALTs were extracted by removing the head from the body, dissecting away the lower jaw, tongue, and connective tissue to expose the soft palate of the upper jaw. The front incisors were then cut away to reveal the anterior end of the soft palate. The palate was then peeled back from the anterior end, revealing the paired NALT structures at the posterior of the hard palate.

Flow cytometry and cell sorting

Single-cell suspensions were prepared from spleens, LNs, and NALTs by mechanical disruption. Mice were perfused before the harvest of the lung tissue and nasal tissue, which were enzymatically digested for 1 hour at 37°C in 3 ml of collagenase type 3 (3 mg/ml in RPMI 1640 supplemented with 2% FBS). Cells were stained for 25 min on ice with the appropriate mixture of monoclonal antibodies and washed with PBS with 1% bovine serum albumin. The conjugated monoclonal antibodies were obtained from BD Pharmingen or eBioscience. H2-Db-NP366, H2-Db-PA244, and H2-Kb-PB1701 tetramers were made in-house. Flow cytometry gating strategies are described in figs. S6 and S7.

Real-time polymerase chain reaction

Cells were washed twice with PBS before RNA extraction by RNeasy Plus Mini Kit (QIAGEN, Venlo, Netherlands). RNA template was prepared, at equal concentrations, and treated to in-solution deoxyribonuclease I digestion (Sigma-Aldrich, St. Louis, USA) to remove trace genomic DNA. Equal volumes of RNA template were used for each SensiFAST cDNA (complementary DNA) synthesis reaction (Bioline, London, UK) before resuspension of template in equal volumes of ultrapure high-performance liquid chromatography water. Real-time polymerase chain reaction was performed with SensiFAST Lo-ROX SYBR Green (Bioline, London, UK) on Mx3005P (Stratagene, La Jolla, USA) with a standardized cDNA template (2 ng) used per well. Fold change was calculated relative to the geometric mean of three housekeeping genes (RPL13a, HPRT, and IPO8).

Immunofluorescence microscopy

Euthanized mice were decapitated, their heads were skinned, and the lower jaws, including tongue, were removed. Tissue was fixed in 4% paraformaldehyde for 6 hours on ice, washed twice with PBS, and incubated with agitation for 48 hours at room temperature in 20% EDTA. Tissue was embedded in optimum cutting temperature, and frozen sections (14 μm) were cut using a cryostat. Tissue sections were acetone-fixed, blocked in serum-free protein block, and stained with the specified antibodies from eBioscience, BioLegend, or Abcam (CD3-Alexa660, CD45.1-Alexa647, B220-Alexa594, and anti-influenza virus NP–fluorescein isothiocyanate).

Statistical analysis

Comparison between two study groups was statistically evaluated by unpaired two-tailed t test or Mann-Whitney test. Comparison between more than two groups (single factor) were evaluated using one-way analysis of variance (ANOVA) with Tukey’s multiple comparison on log10-transformed values. Two-way ANOVA with Sidak’s multiple comparison on log10-transformed values was used to evaluate more than two groups at different time points. In all tests, statistical significance was quantified as *P < 0.5, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Statistical analysis was performed using GraphPad Prism 6 software.

SUPPLEMENTARY MATERIALS

immunology.sciencemag.org/cgi/content/full/2/12/eaam6970/DC1

Fig. S1. Lung and nasal tissue Trm cells develop from KLRG1 precursor cells.

Fig. S2. Molecular signature of nasal tissue Trm cells.

Fig. S3. Localization of NP, PA, and PB1 CD103+CD69+ Trm cell in the lung after influenza virus infection.

Fig. S4. Trm cell development after dendritic cell immunization.

Fig. S5. Depletion of CD8 T cells eliminates local protection in the nasal tissue after influenza virus rechallenge.

Fig. S6. Gating strategy for identifying transgenic OT-I in spleen, lung, and nasal tissue of influenza virus–infected mice.

Fig. S7. Gating strategy for identifying endogenous influenza virus–specific CD8+ T cells in spleen, lung, and nasal tissue of influenza virus–infected mice.

Table S1. Source data file.

REFERENCES AND NOTES

Acknowledgments: We thank S. Turner (Monash University, Australia) for the influenza viruses X31-OVA and PR8-OVA and N. Zhang and M. Bevan for TGFβRII KO OT-I mice. Funding: This work was supported by National Health and Medical Research Council (NHMRC) of Australia and the Australian Research Council (ARC). The Melbourne World Health Organization Collaborating Centre for Reference and Research on Influenza is supported by the Australian Government Department of Health. Author contributions: A.P., P.C.R., and L.M.W. designed the experiments. A.P., J.M.S., and L.M.W. performed the experiments and data analysis. T.H.O.N. and K.K. provided crucial experimental reagents. A.P., P.C.R., A.G.B., W.R.H., and L.M.W. contributed to the writing and editing of the manuscript. A.P. and L.M.W. performed statistical analysis. Competing interests: The authors declare that they have no competing interests.
View Abstract

Stay Connected to Science Immunology

Navigate This Article