MAVS Deficiency Is Associated With a Reduced T Cell Response Upon Secondary RSV Infection in Mice

Infections with respiratory syncytial virus (RSV) occurs repeatedly throughout life because sustained, protective memory responses fail to develop. Why this occurs is not known. During RSV infection the recognition of the virus via the cytosolic RIG-I like receptors and signaling via the adaptor protein MAVS is crucial for mounting an innate immune response. However, if this signaling pathway is important for T cell responses during primary infection and during re-infection is not fully elucidated. We describe a second peak of pro-inflammatory mediators during the primary immune response to RSV that coincides with the arrival of T cells into the lung. This second peak of cytokines/chemokines is regulated differently than the early peak and is largely independent of signaling via MAVS. This was concurrent with Mavs−/− mice mounting a strong T cell response to primary RSV infection, with robust IFN-γ; and Granzyme B production. However, after RSV re-infection, Mavs−/− mice showed fewer CD4+ and CD8+ short term memory T cells and their capacity to produce IFN-γ; and Granzyme B, was decreased. In sum, cytosolic recognition of RSV is important not only for initiating innate anti-viral responses but also for generating or maintaining efficient, short term T cell memory responses.


INTRODUCTION
Respiratory syncytial virus (RSV) is the most common cause of lower respiratory tract infections (LRTI) in children, resulting in bronchiolitis and viral pneumonia and often leading to hospitalization (1,2). Currently, no vaccines or effective treatments against RSV replication and induced disease are available with a pressing need to understand the underlying mechanisms for development of disease and how to obtain protection against the infection. Innate immune responses are important for an efficient, early anti-viral response (2,3). Type I interferons (IFNs) are essential mediators of the innate immune response and their protective role during RSV infection has been demonstrated in animal models (4)(5)(6)(7). Importantly, RSV disease severity has been linked to polymorphisms in genes that control the IFN system (8)(9)(10). Alveolar macrophages (AMs) are the primary source of type I IFNs during RSV infection and these type I IFNs lead to both cell intrinsic viral control and cell extrinsic anti-viral effects (5). One of the major cell extrinsic effect of type I IFNs is to trigger CCL2-induced recruitment of inflammatory monocytes, which contribute to the control of the virus (5). RSV infection of mice deficient in retinoic acid inducible gene I receptors (RLR) adaptor protein mitochondrial antiviral-signaling protein (MAVS; Mavs −/− ) revealed a severe deficiency in type I IFNs, pro-inflammatory cytokines and chemokines and inflammatory monocyte recruitment during RSV infection with associated increased illness and viral burden (5,6,11). Early anti-viral cytokines are not only important to contain virus replication and spread, they also play an essential role for the induction and regulation of protective adaptive immune responses (12). T cell responses are important for the final clearance of RSV. However, the T cells also contribute to illness and disease severity during RSV infection (2,13,14). In the absence of most innate immune responses several reports suggest that Mavs −/− mice are perfectly capable of inducing RSV-specific T cell responses (5,6,11). However, how viral sensing through the RLR-MAVS pathway influences short term T cell memory responses during RSV infection is not fully elucidated. Upon primary infection with a pathogen a population of long-lived memory T cells are maintained in the tissue and circulation in order to respond rapidly to potential secondary encounter with the same pathogen (15). In this study, we investigated the influence of innate immune responses on the initiation and effector function of T cells during the acute and memory phase of immune response to RSV infection. Following primary challenge with RSV, Mavs −/− mice displayed cytokine/chemokine production as well as infiltration of immune cells including an increased T cell recruitment into the lung tissue and airways with heightened T cell effector functions compared to wild type (wt) mice. However, after re-infection, Mavs −/− mice failed to mount a CD4 + or CD8 + T cell response to levels detected in wt mice. In addition, fewer of the memory T cells in the Mavs −/− mice produced Granzyme B and IFN-g. Our study demonstrates a dependency on the early innate immune responses for the presence of RSV-specific memory T cells during RSV re-infection.

Mice
Wt mice and mice deficient in MAVS (Mavs −/− ), obtained from S. Akira [World Premier International Immunology Frontier Research Center, Osaka University, Osaka, Japan, (16)] were bred and maintained in specific pathogen-free conditions. Both strains were Ifna6 gfp/+ but since Ifna6 expression was not a primary readout the mice are designated as wildtype (wt) and Mavs −/− mice. The mice were gender-and age-matched (7-12 weeks) in each experiment. All animal experiments were reviewed and approved by Animal Welfare and Ethical Review Board (AWERB) within Imperial College London and approved by the UK Home Office in accordance with the Animals (Scientific Procedures) Act 1986 and the ARRIVE guidelines were followed.

Virus and Infection
Plaque-purified human RSV (hRSV herein referred to as RSV) A2 (originally from ATCC) was grown in HEp-2 cells with DMEM supplemented with 2% fetal calf serum (FCS) and 2mM L-glutamine. Mice were lightly anesthetized and infected intranasally (i.n.) with 1.8x10 6 FFU of RSV in 100 ml.

Isolation of Cells From Airway (BAL) and Lung
Mice were sacrificed at different time points after infection. For some mice as indicated in figure legends, an intravenous injection of 2 mg anti-CD45-BUV394 (Biolegend, Cambridge, UK) in 200 ml was performed 10 min before euthanasia. Bronchoalveolar lavage (BAL) was performed by flushing the lungs 3 times with 1 ml phosphate-buffered saline (PBS) containing 0.5 mM EDTA (Life Technology, Paisley, UK). The obtained lavage was spun at 3,500 x g for 5 min to separate the BAL fluid from the cells. Lungs were perfused with PBS. To obtain single cell suspension, lungs were collected into C-Tubes (Milteny Biotech, Surrey, UK) containing complete DMEM (cDMEM; supplemented with 10% FCS, 2mM L-glutamine, 100 U/ml penicillin and 100 mg/ml streptomycin), 1 mg/ml Collagenase D (Roche, Welwyn Garden City, UK) and 30 mg/ml DNase I (Sigma Aldridge, Dorset UK) and processed with gentleMACS dissociator according to manufacturer's protocol. After incubation for 30 min at 37°C the lungs were processed in the gentleMACS dissociator again. The red blood cells in BAL cells and lung single cell suspension were lysed using ACK buffer.
For intracellular staining the cells were re-stimulated with 5 mg/ml RSV M 187-195 -peptide for 1 h at 37°C. After addition of Golgi Plug (BD Biosciences) the samples were incubated for another 3 h, stained for surface marker expression as described above and fixed in fixation buffer (Biolegend). Cells were stained with fluorochrome-conjugated antibodies against granzyme B (GB11, PE-CF594) and IFN-g (XMG1.2, BV711) in the presence of purified rat IgG2b anti-mouse CD16/CD32 receptor antibody in permeabilization buffer (Biolgend) for 60 min. Samples were measured on a Becton Dickinson Fortessa LSR-SORP equipped with 20 mW 355 nm, 50 mW 405 nm, 50 mW 488 nm, 50 mW 561 nm, 20 mW 633 nm lasers and a ND1.0 filter in front of the FSC photodiode. Acquisition was set to 250,000 single, live cells. PMT voltages were adjusted after standardized CST checks minimizing the spectral overlap to increase data precision. All antibodies were purchased from BD, eBioscience or Biolegend. Data were analyzed using FlowJo software (Treestar, Ashland, OR, USA).
Chemokine, Cytokine, and RSV-Specific Ig Detection BAL fluid was assessed using a granzyme B and IFN-g ELISA kit (R&D Systems, Minneapolis, MN, USA) according to manufacturer's instructions. Detection limits were 31 pg/ml and 16 pg/ml respectively.
To quantify chemokines and cytokines in the lung tissue, lungs were homogenized using a tissue lyser (Qiagen, Crawley, UK) in PBS containing protease inhibitor according to manufacturer's instructions (Roche). After spinning for 10 min at 10,000 × g supernatants were analyzed using a Magnetic Cytokine 20-plex panel for Luminex (Life Technologies) according to manufacturer's instructions. The data were obtained from a Bioplex 200 system (Bio-Rad laboratories).
To quantify RSV-specific Ig, in serum and BAL, ELISA was used. Briefly, ELISA plates were coated with RSV or HEp-2 control antigens in the same concentration and incubated overnight at 4°C . After blocking with 1% BSA in PBS for 1 h at 37°C, serum or BAL fluid was added in serial dilutions. After a further incubation for 2 h at 37°C HRP-conjugated detection antibodies were added; anti-total Ig (0.25µg/ml, Bio-Rad, Hertfordshire, UK) or anti-IgA antibodies (0.5µg/ml, Bio-Rad). The plates were incubated for 2 h at 37°C before adding TMB substrate (Life-Tech) and after achieving proper color development, the reaction was stopped using 2 M H 2 SO 4 . The plates were analyzed using a spectrophotometry (SoftMax Pro v5.2; Molecular Devices) and optical density was read at 450 nm. The data are plotted as absorbance (Abs) 450 of RSV-specific Ig minus HEp-2 specific Ig.

RNA Isolation and Quantitative RT-PCR
Lungs were homogenized using a TissueLyser LT (Qiagen), and total RNA was extracted using RNeasy Mini kit including DNA removal as described by the manufacturer (Qiagen). 1 mg RNA was reverse transcribed using High Capacity RNA-to-cDNA kit (Applied Biosystems, Paisley, UK) according to manufacturer's instructions. To quantify RNA levels in lung tissue, quantitative RT-PCR reaction for Tnfa, Ifng, and RSV L gene was performed using primers and probes as previously described (4) in QuantiTect Probe PCR Master Mix (Qiagen). For absolute quantification, the exact number of copies of the gene of interest was calculated using a plasmid standard, and the results were normalized to the housekeeping gene Gapdh (Applied Biosystems). The relative expression of Il1b, Il6, and Cxcl1 (all from Applied Biosystems) to the housekeeping gene Gapdh was determined. First, the DCt (Ct=cycle threshold) between the target gene and Gapdh for each sample was calculated, following the calculation of 2 -DCt . Analysis was performed using 7500 Fast System SDS Software (Applied Biosystems).

Statistical Analysis
Statistical analysis of data was performed using GraphPad Prism 6 (GraphPad Software Inc., La Jolla, CA, USA). Two-group comparisons were performed using unpaired, two-tailed Student's t test. One-way ANOVA with Tukey's post hoc test was used to compare multiple groups. For all tests a value of P<0.05 was considered as significant. *p<0.05; **p<0.01; ***p<0.001; ND, not detectable.

RESULTS
Inflammatory Cytokines and Recruitment of Innate Immune Cells During the Adaptive Response Is Independent of MAVS Signaling RSV infections are characterized by the rapid (18-24 h following infection) and transient induction of early innate immune mediators such as type I IFNs, IL-6, and TNF-a (3,5,6,17). The production of these cytokines are dependent on signaling via MAVS, as previously published (5,6). To monitor cytokine responses during the later course of primary RSV infection, cytokines and chemokines were measured in lung samples of RSV infected wild type (wt) and Mavs −/− mice, starting 4 days post infection (p.i.) up to day 9 p.i. The relative gene expression of Tnfa, Il6, and Il1b, and the protein expression of chemokine CCL2 were all detected on days 5 to 7 p.i. in both wt and Mavs −/− mice ( Figures 1A, B). Notably, this second peak of mediators (days 5-7 p.i.) was smaller than the first peak (18-24 h p.i.) identified in RSV infected wt mice [ Figure 1 and (5, 6)]. The neutrophil chemoattractant CXCL1, was induced more prominently in Mavs −/− mice than wt mice during the second peak (days 4-5; Figure 1B). This differential expression of the various cytokines and chemokines raised the question about the dynamics of pulmonary innate cells during RSV disease. Therefore, lung cell populations were assessed by flow cytometry from day 4 up to day 9 post RSV infection.
Gating was performed as previously described (5). Equivalent to the cytokine and chemokine responses the majority of the identified cell populations (CD103 + dendritic cells (DCs), plasmacytoid DCs (pDCs), inflammatory monocytes (CD64 + infMo) and neutrophils) showed recruitment into the lung of RSV infected wt and Mavs −/− mice from day 4 to day 9 p.i. (Figures 1C-G). CD11b + DCs were recruited slightly later in Mavs −/− mice and compared to wt mice ( Figure 1C). CD103 + DCs, pDCs and infMo were recruited in both Mavs −/− and wt mice but a slightly higher number of these cells was observed in Mavs −/− mice at the later time points compared to wt mice ( Figures 1E, F). Neutrophil numbers in Mavs −/− mice were increased over baseline for most time points while wt mice showed neutrophil recruitment on day 8 p.i. ( Figure 1G). In summary, late responses during RSV infection (days 4-9), including cytokine expression and immune cell recruitment depend on mechanisms other than MAVS signaling and are differently regulated than in the early response (days 0-3) where the cytokine and chemokine responses are mostly dependent on MAVS signaling (5).

T Cells Are Increased in the Lung of Mavs −/− Mice During Primary RSV Infection
The induction of appropriate adaptive immune responses and the recruitment of RSV-specific T cells play an essential role in viral clearance, regulating inflammatory responses and disease pathology (2,13). In previous reports we and others have shown that wt and Mavs −/− mice recruit CD4 + and CD8 + T cells into the lung (5,6,11). A quantitative and qualitative analysis of differences in airway and lung T cells during experimental RSV infection and the possible impact of innate immune responses remained elusive. Time course analysis (days [4][5][6][7][8][9] in the RSV infection model revealed an induction of IFN-g mRNA from day 5 p.i. in lungs from both wt and Mavs −/− mice albeit a more prominent increase was seen in wt mice at day 5 p.i. 2A). T cells are recruited into inflamed tissue by chemokines such as CXCL9 or CXCL10, which are induced by IFNs (18). Both CXCL9 and CXCL10 protein was detected in the lungs of wt and Mavs −/− mice and the levels of these chemokines were slightly higher at days 7 to 9 in the Mavs −/− mice compared to wt mice ( Figure 2B). For analysis of T cell responses during RSV infection, wt and Mavs −/− mice were sacrificed on day 8 p.i., representing the peak of the T cells response (19,20). Cells from the lungs and the airways (bronchoalveolar lavage; BAL) were isolated and stained for cell surface markers to phenotypically distinguish different T cell subsets (gating shown in Supplemental Figure S1A). Numbers of total CD4 + and CD8 + T cells were significantly increased in the lung tissue of Mavs −/− mice compared to wt mice ( Figure 2C). However, the recruitment of T cells into the airways showed no differences between Mavs −/− and wt mice ( Figure 2D). RSVspecific CD8 + T cells, determined by tetramer staining (gating shown in Supplemental Figure S1B) were represented in equal percentages and numbers ( Figures 2E, F) in the lung tissue of wt and Mavs −/− mice, consistent with previous reports from our group and others (5,6). However, viral clearance was delayed in Mavs −/− mice with more RSV L gene copies detected at day 8 p.i.
[ Figure 2G (5, 11)]. This is most likely a result of the absence of a complete innate immune response (5,6,11) and an increased

Expression of GzmB in CD8 + T Cells Is Enhanced in Mavs −/− Mice During Primary Infection With RSV
During acute viral infection, effector CD8 + T cells engage several mechanisms that mediate killing of infected cells and influence antiviral immune responses. For example, the production of cytotoxins, such as granzyme B (GzmB) and perforin, or the expression of surface proteins, like FasL induces programmedcell death in target cells (14). Additionally, the release of IFN-g and TNF−a have pleiotropic and immune stimulatory effects (21,22). We assessed the effector function of lung and airway CD8 + T cells after primary encounter with RSV by stimulating them ex vivo with RSV M 187-195 peptide and evaluated their capacity to produce GzmB and IFN-g, or both, by intracellular staining and flow cytometry analysis (gating shown in Supplemental Figure  S1B). In the airway and lungs of Mavs −/− mice, 40% to 50% of CD8 + T cells expressed GzmB, compared to wt mice where 30% to 40% of CD8 + T cells stained positive for GzmB ( Figures 3A, B). Also, total numbers of CD8 + GzmB + T cells in the lung tissue was significantly higher in Mavs −/− mice at day 8 during primary infection ( Figure 3A). In terms of the capacity of CD8 + T cells to produce IFN-g, wt and Mavs −/− mice displayed no significant differences ( Figures 3A, B). The same accounted for multifunctional CD8 + T cells capable of producing both GzmB and IFN-g (Supplemental Figures S2A, B). This suggested that Mavs −/− mice were able to recruit highly activated effector CD8 + T cells into the infection site and a large fraction of them were able to secrete effector mediators. Of note, the peptide stimulation did not allow for stimulation of CD4 + T cells, which could also be potent producers of IFN-g. Furthermore, to detect GzmB and IFN-g; levels in vivo, BAL fluids were analyzed. The overall GzmB Each symbol represents one individual mouse. Statistical significances of differences between groups were determined by one-way ANOVA with Tukey's post hoc test. *P < 0.05; ***P < 0.001. and IFN-g; responses were significantly increased in Mavs −/− mice at day 8 p.i. compared to wt mice ( Figures 3C, D). Altogether, our data show that CD4 + and CD8 + T cell numbers and CD8 + T cell effector functions are elevated in the absence of MAVS signaling during primary RSV infection.

Fewer T Cell in the Lungs After RSV Re-Infection of Mavs −/− Mice
During secondary encounter with a pathogen memory T cells within the tissue or recruited from circulation are rapidly activated for defense of the host tissue (15). We hypothesized that the loss of innate immune activation in Mavs −/− mice (5,6,11) impacts on the quality of the short-term memory responses to RSV. The virus is cleared in both wt and Mavs −/− mice after 8 to 9 days post infection (6,23) so to study short-term memory responses in Mavs −/− mice, we re-infected wt and Mavs −/− mice 21 days after primary infection with RSV and determined the T cell responses 4 days post re-challenge by flow cytometry (gating strategy see Supplemental Figure S1A). To note, in vivo labelling of CD45 + cells were performed in wt mice at day 4 post re-infection to distinguish T cells residing in the vasculature versus in the lung parenchyma and a majority (70-80%) of the CD4 + and CD8 + T cells were shown to reside in the lung parenchyma (Supplemental Figure 3).
Overall, lung CD4 + and CD8 + T cells were increased in RSV re-challenged mice compared to mice 25 days post RSV primary infection ( Figure 4A). Between the re-challenged groups, the absolute numbers of CD4 + and CD8 + T cell in the lung tissue were higher in wt mice compared to Mavs −/− mice ( Figure 4A). Although, the expansion/recruitment of T cells was similar in the airways of both re-infected strains ( Figure 4B) and the T cellattracting chemokine CXCL9 was detectable in the BAL fluid of both, wt and Mavs −/− mice. Albeit, Mavs −/− mice had a two-to three-fold higher level of CXCL9 than wt mice ( Figure 4C).
To elucidate if this was due to an altered viral load, the number of RSV L gene copies was determined on day 4 after reinfection with RSV. The amount of detectable RSV L gene was low but no differences in lungs from wt and Mavs −/− mice were detected ( Figure 4D). This is probably due to effective neutralization of RSV virions by RSV-specific antibodies (24) as both wt and Mavs −/− mice induce neutralizing RSV-specific antibodies. In our study, Mavs −/− mice showed higher levels of RSV-specific Ig compared to wt mice ( Figures 4E, F) while a previous study showed an impaired RSV-specific Ig response in Mavs −/− mice (6). Overall, these data suggest a defect in the accumulation of lung memory T cells in the Mavs −/− mice.

Defect in Effector Function of RSV-Specific Memory CD8 + T Cells in Mavs −/− Mice During Secondary RSV Infection
To evaluate the RSV-specific CD8 + T cell response three strategies were employed. First, tetramer staining was Statistical significances of differences between groups were determined by unpaired Student's t test. In (B) only difference between RSV re-infected groups are shown. *P < 0.05; **P < 0.01; ***P < 0.001.
performed to quantify the frequency and total number of RSVspecific CD8 + T cells. Mavs −/− mice had fewer lung M-tetramer positive cells compared to wt mice ( Figure 5A). To confirm these results and quantify total CD8 + T cell responses, lung single-cell suspensions were re-stimulated with RSV M 187-195 peptide at different concentrations (5, 0.5, and 0.05 ng/ml), control peptide RSV M2 82-90 (5 ng/ml) and RSV (MOI 1) for 3 days and the supernatants assessed for IFN-g ( Figure 5B). These data confirmed that Mavs −/− mice have fewer RSV-specific CD8 + T cells as their lung cells secreted less IFN-g upon CD8 + T cell peptide stimulation than wt lung cells ( Figure 5B).
We also speculated that, in addition to fewer RSV-specific CD8 + T cells, Mavs −/− mice additionally exhibit functional defects in the RSV-specific CD8 + T cell responses. After a short peptide restimulation, frequency and number of CD8 + GzmB + and multifunctional CD8 + T cells (IFN-g and GzmB) in both lung tissue and airways, were significantly reduced in Mavs −/− mice compared to wt mice ( Figures 5C-F). In absolute numbers, a decrease in GzmB + and GzmB + IFN-g; + cells were noted in the lung (Supplemental Figure 4). Taken together, our data show that numbers and full functionality of short-term memory CD8 + T cells are dependent on signaling via MAVS. Type I IFN secretion comprises the imminent response triggered in infected cells with robust antiviral and immunomodulatory response to contain virus replication and spread (25). During RSV infection, type I IFNs derived from alveolar macrophages (AMs) regulate the early inflammatory and antiviral activity of different lung cell populations (4,5,17). This response is completely dependent on MAVS signaling since type I IFNs and many other early mediators are completely absent in Mavs −/− mice after RSV infection (5,6). Also, deficiency in MAVS signaling and/or deficiency in type I IFNs receptor signaling promote susceptibility to RSV-induced disease (4,5). However, the cytokine and chemokine expression in the lung of wt mice after infection with RSV is characterized by a biphasic induction resulting in two peaks of leukocyte infiltration and cytokines and chemokine secretion in the lungs (previously described (26-29) and confirmed in our study). A second peak (days 4-9 p.i.) was detected for the chemokines CCL2, CXCL1, CXCL9 and CXCL10, the pro-inflammatory cytokines TNF-a, IL-6, IL-1b, and IFN-g; and DCs, CD64 + inflammatory monocytes and neutrophils. These observations resembles experimental RSV challenge studies in humans, where a transient induction of the neutrophil chemokine CXCL8 (mouse homologue CXCL1) shortly after inoculation was followed by an additional CXCL8 peak during RSV-induced illness and viral shedding (30). Additionally, IL-8, CCL2, CCL3, and CCL5 were elevated in nasal secretions during symptomatic illness (30). Importantly, this second wave of cytokine and chemokine gene upregulation with cellular infiltration was initiated by mechanism/s independent of the RLR-MAVS pathway. In the later stages of primary RSV infection, apoptotic epithelial cells and immune cells are markedly increased in the lungs (31). It is therefore possible that signaling of endosomal TLR3 and TLR7 in phagocytic cells that clear apoptotic cells containing viral antigen or cells expressing DAMP receptors become increasingly important.

DISCUSSION
Interestingly, Mavs −/− mice showed significant increased influx of CD103 + DCs, pDCs, infMo and neutrophils in the respiratory tract [our data and (11)]. Notably, the viral load in the lung of Mavs −/− mice was higher than in wt mice (5,6,11) and this could drive the increased recruitment of these cell types to the lung. Similarly, influx of monocytes and neutrophils was abundant in autopsy material from infants who died from RSVinduced LRTI (31).
The second peak of cytokines and cellular infiltration in the lung coincides with T cells infiltrating the lungs and also RSV-induced illness characterized by increased airway hyperresponsiveness, airway resistance and weight loss in mice (17,26). In most infections the activation of the adaptive immune responses and their potent effector functions is required for the final elimination of invasive pathogens. It is commonly believed that the initiation of the adaptive immune system requires innate cytokines, such as type I IFNs (12). Accordingly, it was very surprising to discover similar or slightly increased T cell responses in Mavs −/− mice during primary infection [day 8 p.i.; data presented here and (5, 6, 11)]. There are several possible explanations for this. Firstly, the viral load in the lung of Mavs −/− mice is higher than in wt mice (5,6,11) and this could be driving stronger T cell responses. It is important to note that similar levels of RSV specific CD8 + T cells were detected in wt and Mavs −/− mice, which could suggest a heightened recruitment of CD8 + T cells with other TCR specificities. Second, a MAVS independent mechanism could drive the maturation and antigen presentation capacity of DCs. We have previously shown that mice deficient in MyD88/TRIF and MAVS can mount T cell responses albeit to a reduced level compared to wt mice (32). In addition, Bhoj et al. described a migratory DC subset in RSV-infected Mavs −/− mice suggested to be involved in cross-presentation of exogenous antigen to CD8 + T cells in the draining lymph nodes (6). A specific characterization of this particular DC subset and the specific molecular signals that lead to priming of T cells by DCs remain unclear.
After two to three weeks the majority of activated T cells undergo apoptosis and only a small pool of virus-specific memory T cells are maintained (15,33). Upon secondary encounter with a pathogen, activation and clonal expansion of the memory T cell pool is crucial for a rapid host defense and survival. We show that activation and/or expansion of shortterm memory cells was compromised in the lungs of mice lacking MAVS signaling, which led to a significant impairment in the production of the T cell effector molecules GzmB and IFNg. To note, this was observed in the lung but not in the airways. This was not due to a difference in viral load or RSV-specific antibody titers. Interestingly, the Mavs −/− mice showed higher RSV-specific Ig responses. The deficiency of short term memory T cells could be the result of a defect in the priming phase of the T cell response that leads to a problem in establishing or retaining a memory T cell pool. Alternatively, MAVS signaling could be necessary during the re-activation of memory T cell in the lungs during RSV secondary infection. To fully re-activate memory T cells, cognate and non-cognate signals (such as type I IFNs, IL-15, or IL-18) are required to rapidly acquire effector functions (34)(35)(36). Kohlmeier et al. showed that signaling of type I IFNs through the type I IFN receptor (IFNAR) on memory CD8 + T cells enhances GzmB production and cytolytic activity in recall responses to influenza virus (34). Therefore, it is possible that the memory T cells formed during primary RSV infection in Mavs −/− mice cannot be re-activated due to a lack of essential inflammatory cytokines. Also, the type of memory cells in the lung will be important to evaluate. Resident memory T cells have been shown to be important during RSV infection (37,38) and the formation and function of these cells in MAVS deficient mice will be important to elucidate in future studies.
In conclusion, our work emphasizes a role for MAVS signaling beyond the initiation of early anti-viral responses that protect from RSV infection. In particular, we identified a critical role for the RLR-MAVS pathway in the numbers and effector functions of the short term memory T cell pool during secondary RSV infection. RSV elicits a short-lived memory response in humans and our data suggest that a weak innate response results in an impaired short-term T cell memory response, especially in the CD8 + T cells, which highlights the importance of a balanced immune response to achieve an efficient and long-lasting antiviral immune response important both during infection and vaccination.

DATA AVAILABILITY STATEMENT
All datasets presented in this study are included in the article/ Supplementary Material.

ETHICS STATEMENT
The animal study was reviewed and approved by Imperial College AWERB.

AUTHOR CONTRIBUTIONS
MiP designed, performed and analyzed the experiments and wrote the paper. NP, AV, and FK performed specific experiments and reviewed the paper. MaP helped in designing the flow cytometry panel. CJ supervised the project, designed and analyzed the experiments, and wrote the paper. All authors contributed to the article and approved the submitted version.