T cell responses are elicited against Respiratory Syncytial Virus in the absence of signalling through TLRs, RLRs and IL-1R/IL-18R

Pattern recognition receptors (PRRs) and cytokine receptors are key players in the initiation of immune responses to infection. PRRs detecting viral RNA, such as toll like receptor (TLR)-3, -7/8, and RIG-I like receptors (RLRs; RIG-I and MDA-5), as well as cytokine receptors such as interleukin 1 receptor (IL-1R), have been implicated in responses to RNA viruses that infect the airways. The latter includes respiratory syncytial virus (RSV), a human pathogen that can cause severe lower respiratory tract infections, especially in infants. To evaluate the collective contribution of PRRs and IL-1R signalling to RSV immunity, we generated Myd88/Trif/Mavs−/− mice that are deficient in signalling by all TLRs, RLRs and IL-1R, as well as other cytokine receptors such as IL-18 receptor. Early production of pro-inflammatory mediators and lung infiltration by immune cells were completely abrogated in infected Myd88/Trif/Mavs−/− mice. However, RSV-specific CD8+ T cells were elicited and recruited into the lungs and airways. Consistent with these findings, Myd88/Trif/Mavs−/− mice survived RSV infection but displayed higher viral load and weight loss. These data highlight an unappreciated level of redundancy in pathways that couple innate virus sensing to adaptive immunity, providing the host with remarkable resilience to infection.

Scientific RepoRts | 5:18533 | DOI: 10.1038/srep18533 TLR2/6, TLR3, TLR4, TLR7, NOD2, RIG-I and MDA5 have all been implicated in RSV recognition 7,15 and loss of function of some of them (or of their adaptors) in mice can impact innate immunity to the virus 9,11,[16][17][18][19][20][21][22] . It is less clear to what extent these receptors are necessary for DC activation and priming of RSV-specific T cells. It has been reported that normal anti-RSV T cell responses can be elicited in Mavs −/− mice or mice doubly-deficient in MyD88 and MAVS 9,21,22 . This might suggest that TLR3 and TLR4 signalling through TRIF can compensate for loss of other TLRs and of RLRs, indicative of a high level of redundancy in the immune response to RSV. We set out to test this hypothesis by creating mice deficient in all three adaptor proteins; MyD88, TRIF and MAVS. Here, we show that after RSV infection these Myd88/Trif/Mavs −/− mice lose more weight and have higher viral load than control mice. The early inflammatory response to RSV is absent in Myd88/Trif/Mavs −/− mice but this does not impact greatly on the ability to prime and recruit RSV-specific T cells that secrete IFN-γ and granzyme B into the lungs and airways. In summary, RSV-infected Myd88/Trif/Mavs −/− mice are able to mount an RSV-specific T cell response and resist the virus despite a profound loss in innate immunity. These results indicate that mechanisms other than TLR, RLR and IL-1R signalling can induce adaptive immunity to RSV and, possibly, other RNA viruses.

Airway inflammation after RSV infection is abrogated in Myd88/Trif/Mavs −/− mice.
In order to delineate the role of TLRs, RLRs and IL-1/18R during RSV infection, Myd88/Trif/Mavs −/− triply deficient mice (henceforth, MTM −/− ) and control wild type (wt) mice (either C57BL/6 or littermate controls) were intranasally infected with RSV or mock infected with vehicle alone (PBS). At 18 h and 96 h post infection (p.i.), no IFN-α , IL-6, CXCL10 or CCL2 was detected in the bronchoalveolar lavage fluid (BAL) of infected MTM −/− mice, in contrast to BAL from wt animals (Fig. 1A). These results are in line with previously published data using RSVinfected Ifnar1 −/− , Mavs −/− or Myd88/Mavs −/− mice 9,22,23 . Expression of IL-1α and IL-1β in the lungs was also undetected in MTM −/− mice at any time point, in contrast to wt controls (Fig. 1B). Furthermore, very few neutrophils were recruited into the airways of MTM −/− mice (Fig. 1C) consistent with the lack of production of the neutrophil chemoattractant CXCL1, which was tested both at protein and mRNA level ( Fig. 1D and data not shown). Thus, signalling via TLRs and/or RLRs is necessary for the early pro-inflammatory cytokine and chemokine response to RSV infection and for ensuing infiltration of inflammatory cells. in the lower airways and are crucial to initiate early immune responses to RSV 9 . Primary wt AMs exposed ex vivo to RSV at different multiplicities of infection (MOI) produced IFN-α , IL-6 and TNF-α (Fig. 2). In contrast, no production was detected from AMs isolated from MTM −/− mice (Fig. 2). These findings indicate that AMs rely on TLRs and/or RLRs signalling to respond to RSV, confirming our previous observation that RLR signalling is a dominant pathway in the innate immune response of these cells to RSV 9 .

Myd88/Trif/Mavs −/− mice display impaired viral control and enhanced weight loss.
To assess the impact of the lack of MyD88/TRIF/MAVS signalling on the course of infection, RSV replication in the lung was determined (Fig. 3A,B). MTM −/− mice displayed a much higher lung viral load than wt controls as measured by RSV L gene copies in lung tissue at day 4 and day 8 p.i.. As C57BL/6 is a relatively resistant mouse strain to RSV infection, isolating infectious virus from the lungs of these animals is not normally possible and they register as uninfected by immunoplaque assay where infected cells are stained positive for DAB (Fig. 3B). In contrast, lungs from MTM −/− mice contained detectable infectious virus at day 4 p.i. using the same assay (Fig. 3B). In addition, the MTM −/− mice lost significantly more weight than wt mice (17% versus 5% weight loss on day 7 p.i.; Fig. 3C). Nevertheless, from mice in a separate experiment, not sacrificed for analysis, it was clear that both wt and MTM −/− mice survived the infection equally (data not shown). We tested whether loss of signalling through MyD88/TRIF/MAVS would influence the expression of the epidermal growth factor, amphiregulin, which is associated with epithelial repair 24 . Interestingly, MyD88/TRIF/MAVS did not regulate the gene expression of amphiregulin as it increased comparably in both wt and MTM −/− mouse strains during the course of RSV infection (Fig. 3D). We conclude that mice deficient in MyD88/TRIF/MAVS display increased susceptibility to RSV infection, with higher virus titres and increased weight loss.  (Fig. 3E). Granzyme B was detected in wt mice but not MTM −/− mice at 18 h p.i. and reached similar levels in the two strains at day 8 p.i. (Fig. 3E). These data might suggest that an early source of granzyme B is missing in MTM −/− mice, consistent with the defect in innate immunity in that strain, but is compensated at later stages by another source, most likely T cells. Indeed, lung recruitment of T cells reaches its peak at day 8 post RSV infection 25 when they represent the main source of IFN-γ and granzyme B 26 . Consistent with the notion that lung T cell recruitment is functional in the absence of MyD88/TRIF/MAVS signalling, expression of the T cell chemoattractants CXCL9 and CXCL10 27 was strongly induced in MTM −/− mice at day 8 p.i., reaching even higher levels than in wt controls (Fig. 3F). Interestingly, MTM −/− mice lacked an earlier peak of chemokine expression seen at 18 h p.i. in wt animals (Fig. 3F) again an indication that MTM −/− mice might have a defect in recruitment of innate cells. Together, these data suggest that MyD88/TRIF/MAVS might be dispensable for a T cell response against RSV.

IFN-γ, granzyme B, Cxcl9 and
T cells are recruited to the airways and lungs of Myd88/Trif/Mavs −/− mice after RSV infection. To test the above hypothesis, airways and lungs of MTM −/− mice were analysed on day 8 p.i. for the composition of the immune infiltrate. The frequencies of lymphocytes (determined by differential counting of cytospin slides) and CD4 + and CD8 + T cells (determined by flow cytometry, gating strategy shown in Fig. S1A) in the BAL fluid were not different between MTM −/− and wt control mice (Fig. 4A,B). However, the total number of cells in BAL was lower in MTM −/− mice compared to wt mice (Fig. 4A) leading to a net reduction in CD4 + and CD8 + T cell numbers (Fig. 4A,B). Similar results were obtained when analysing lung tissue (Fig. 4C). Altogether, the defect in innate immune responses in the absence of MyD88/TRIF/MAVS did not hinder the T cell recruitment into the infected lungs and airways albeit MTM −/− mice showed lower levels of T cells than wt mice.  (Fig. 5A,B; gating strategy shown in Fig. S1B). Frequencies of tetramer + cells were similar between wt and MTM −/− mice, with around 30% of the CD8 + T cells staining positive for RSV M tetramers (Fig. 5A,B). However, the total number of RSV-specific CD8 + T cells was lower in the BAL but not the lungs of MTM −/− mice compared to wt mice (Fig. 5B), suggesting that MyD88, TRIF and/or MAVS signals might influence the extravasation of T cells into the airways. In sum, RSV-specific T cells can be elicited in mice deficient in MyD88/TRIF/MAVS.

Discussion
DCs are activated in tissues by PRR and cytokine receptor signals and migrate to draining lymph nodes where they prime T cells. Many PRRs have been shown to recognise RSV but which are necessary to initiate virus-specific T cell responses is unclear. Here, we show that mice lacking the ability to signal via TLRs, RLRs, IL-1R or IL-18R (Myd88/Trif/Mavs −/− mice) can still elicit RSV-specific T cells and their recruitment to the lung and airways. Our data indicate the existence of mechanisms for coupling detection of RSV infection to adaptive immunity that are not predicted from current knowledge in the field. It is not clear how DCs get activated in the lungs of MTM −/− mice during RSV infection in the absence of PRR signals. It has been suggested that, during influenza virus infection, CD8 + T cell priming depends on IL-1R rather than PRR signalling in DCs 13 . As MyD88 is required for IL-1R and IL-18R signaling 14 , this mean of activating DCs is not operative in our model. Consistent with that notion, we were also unable to detect any IL-1α or IL-1β in the lungs of MTM −/− mice during RSV infection. Another possibility is that DCs get activated by products derived from dying cells 28 . DNGR-1 is a receptor that can couple detection of dead cells to CD8 + T cell priming 29 , even though it is not, technically, a DC-activating receptor 30,31 . However, we have recently shown that DNGR-1 is dispensable for T cell responses to RSV in mice 32 although it could become non-redundant in the MTM −/− mice, when activation of DCs via PAMPs is absent. Other dead cell-derived mediators that could promote DC activation include heat shock proteins, HMGB1, uric acid, and ATP 28 . An interesting candidate is DNA released from dead cells, which has been implicated in the adjuvant effects of alum 33 and would trigger the cGAS/STING pathway normally involved in immunity to DNA viruses 34 . To what extent DNA sensing might impact the response to cytopathic RNA viruses is an interesting area for future research.
Underscoring the importance of PRRs in RSV infection, RSV bronchiolitis has been associated with polymorphism in genes encoding PRRs or their signalling pathways [35][36][37][38][39] . Here, we describe for the first time the impact of global loss of TLR and RLR signalling in experimental RSV infection. Consistent with a key role for PRRs in antiviral immunity, MTM −/− mice showed markedly reduced innate responses with a deficit in production of pro-inflammatory cytokines or chemokines early post-infection. No infiltration by neutrophils was detected in the lungs of MTM −/− mice and these mice were defective in restricting viral replication. NK cell responses were not directly evaluated in this study but our previous data showed that Mavs −/− mice fail to recruit NK cells into the lungs during RSV infection 9 . Despite this early failure to contain RSV infection, MTM −/− mice did not succumb to the virus as previously shown in STAT1 −/− mice 40 . This is most probably due to their capacity to mount RSV-specific T cell responses, as demonstrated here. Such responses have previously been reported following RSV infection of Mavs −/− and Myd88/Mavs −/− mice 9,21,22 although the impact of additional TRIF deficiency has not been evaluated until now. We suggest that T cells recruited to the lung and airways in the MTM −/− mice are more actively restimulated due to the higher viral burden, resulting in enhanced IFN-γ production that ultimately contributes to viral clearance.
Furthermore, we found that amphiregulin is upregulated in the lungs of RSV infected wt and MTM −/− mice. Amphiregulin is a member of the epidermal growth factor family important for maintaining lung epithelial cell barrier function 24,41 . This and other growth factors might help to maintain the epithelial barrier during the innate phase of the response until T cells can be recruited. Interestingly, like T cell responses and the T cell recruiting chemokines CXCL9 and CXCL10 on day 8 p.i., amphiregulin was induced independently of MyD88, TRIF and MAVS signalling. Thus, signs of tissue repair and adaptive immune responses to RSV can occur independently of known classical PRR pathways implicated in the recognition of RNA viruses. It will be important to elucidate the mechanisms underlying this phenomenon as they might have an impact on future vaccine designs and therapeutics.

Methods
Mice. Myd88/Trif −/− and Mavs −/− mice (obtained from S. Akira, Japan [42][43][44] were screened to ensure the genotype was maintained and were crossed in-house to generate Myd88/Trif − /Mavs −/− mice. Littermates or C57BL/6 mice (purchased from Charles River or Harlan, UK) were used as controls. No difference between littermates and C57BL/6 controls were detected for any of the assays performed. All mice were bred and maintained in pathogen-free conditions and gender and age-matched mice aged 8-14 weeks were used for each experiment. All animal experiments were performed in accordance with the Animal Welfare and Ethical Review Board (AWERB) within Imperial College London and the ARRIVE guidelines and were approved by the UK Home Office in accordance with the Animals (Scientific Procedures) Act 1986. Cell isolation and processing. Mice were sacrificed using a fatal dose of pentobarbital injected intraperitonially (i.p.) according to UK home office guidelines. Bronchoalveolar lavage (BAL) and lungs were obtained as described previously 32 . Briefly, BAL was collected by flushing the lungs three times with 1 ml of PBS/0.5 mM EDTA (Life Technologies). Lung lobes were collected and digested with Collagenase D (1 mg/ml, Roche) and DNase I (30 μ g/ml; Sigma-Aldrich) using a gentleMACs cell dissociator (Miltenyi Biotech) according to the manufacturer's protocol and incubated at 37 °C for 30 min. Red blood cells were lysed by treating lung cells with ACK lysing buffer and lung cells were passed through 100 μ M cell strainers to create single-cell suspensions. Total cell counts were determined using hemocytometer slides and dead cells were excluded by Trypan blue staining (Sigma-Aldrich). To determine the cellular composition in the BAL, cells were transferred onto a microscope slide (Thermo Scientific, UK) using a Shandon Cytospin 3 Centrifuge and slides were stained with hematoxylin and eosin (H&E; Reagena, Gamidor, UK). Cells were categorised as macrophages, lymphocytes or neutrophils based on their morphology and size under a light microscope.
For ex vivo stimulation of AMs, AMs were collected by bronchoalveolar lavage through flushing the lungs as above repeated twice and AMs from several mice were pooled. The purity of the AMs was > 98%. Collected cells were incubated in a flat-bottom 96-well plate (1.25 × 10 5 cells/well) in complete DMEM (cDMEM; supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 100 U/ml penicillin and 100 μ g/ml streptomycin) for 3 h. After washing, the adherent cells were exposed to cDMEM (medium) or various MOIs of RSV or UV-inactivated RSV for 20 h. Inactivation was performed by exposing virus to UV light for 2 min (UV-RSV) in a CX-2000 UV cross-linker (UVP).

Virus and infection.
Plaque-purified human RSV (originally A2 strain from the ATCC, US) was grown in HEp-2 cells 45 . For infection, mice were lightly anesthetised and installed intranasally (i.n.) with 2 × 10 6 focus-forming units (FFU) of RSV in 100 μ l. As control, intranasal administration of 100 μ l PBS was given to naïve mice. Intranasal inoculation with PBS or HEp-2 cell supernatant does not change BAL cell numbers or cell composition compared to naïve mice (data not shown).
RSV titre was assessed in lungs 4 days post RSV infection using an immunoplaque assay (optimized from 26 ). Briefly, lung homogenate was titrated on HEp-2 cell monolayers in 96-well, flat-bottom plates and incubated for 2 h on a shaker. Then, pyruvate-free DMEM containing 2% FCS was added and the plate was further incubated for 20 h before fixed with methanol and incubated with biotin-conjugated goat anti-RSV antibody (Biogenesis, United Kingdom). Infected cells were detected using Streptavidin (Sigma) and DAB substrate (Diaminobezindine tetrahydrochloride), enumerated by light microscopy and used to calculate titre as focus forming units (FFU).