key: cord-0306325-qdgxg88j
authors: Zhao, Bali; Yang, Jingyi; He, Bing; Li, Xian; Yan, Hu; Liu, Shuning; Yang, Yi; Zhou, Dihan; Liu, Bowen; Fan, Xuxu; Zhong, Maohua; Zhang, Ejuan; Zhang, Fan; Zhang, Yue; Chen, Yao-Qing; Jiang, Shibo; Yan, Huimin
title: A Safe and Effective Mucosal RSV Vaccine Consisting of RSV Phosphoprotein and Flagellin Variant
date: 2021-02-10
journal: bioRxiv
DOI: 10.1101/2021.02.09.430425
sha: 632bcfe0b8b1ee5a93220b679051f37b34e09d62
doc_id: 306325
cord_uid: qdgxg88j

Respiratory syncytial virus (RSV) is a major cause of serious acute lower respiratory tract infection in infants and the elderly. No licensed RSV vaccine available thus far calls for the development of vaccines with new target(s) and vaccination strategies. Here, we constructed a recombinant protein, designated P-KFD1, comprised of RSV phosphoprotein (P) and E. coli K12 strain-derived flagellin variant KFD1. Intranasal (i.n.) immunization with P-KFD1 inhibits RSV replication in both upper and lower respiratory tract, and protects mice against lung disease without vaccine-enhanced disease (VED). The P-specific CD4+ T cells provoked by P-KFD1 i.n. immunization, either reside in or migrate to respiratory tract, mediate protection against RSV infection. Sc-RNA seq and carboxyfluorescein succinimidyl ester (CFSE) labeled cell transfer further characterized the Th1 and Th17 responses induced by P-KFD1. Finally, we found the anti-viral protection depends on either IFN-γ or IL-17A. Collectively, P-KFD1 is promising as a safe and effective mucosal vaccine candidate to prevent RSV infection. HIGHLIGHTS A new subunit RSV vaccine candidate with new target and vaccination strategy, P-KFD1, is designed and generated Intranasal immunization with P-KFD1 protects mice against RSV infection and averts vaccine-enhanced disease Sc-RNA seq and CFSE-labelled cell transfer identified characteristics of the protective CD4+ T cells Local and peripheral CD4+ T cells provide protection against RSV infection dependent on either IFN-γ or IL-17A

Human respiratory syncytial virus (RSV) was first isolated and identified as an 6 important cause of bronchiolitis in infants as early as 1956 (Blount et al., 1956; 7 Chanock and Finberg, 1957). It remains a major cause of severe acute lower 8 respiratory tract infection (ALRI) not only in infants, but also in young children, the complications, including bronchiolitis or pneumonia, often with acute respiratory 12 distress, requiring hospitalization and creating a heavy medical cost and burden for 13 treatment and prevention all over the world. In 2015 alone, RSV infection resulted in 14 about 3.2 million episodes of hospitalization across the globe and 59,600 deaths in 15 children younger than 5 years (Shi et al., 2017) . A recent multi-site case-control study 16 found that RSV had the greatest etiological fraction as high as 31.1% of all bacteria 17 and virus pathogens that caused severe pneumonia requiring hospital admission in 18 children from Africa and Asia (Group, 2019). Even worse, natural RSV infection does 19 not elicit long-lasting immunity, and repeated infections occur throughout life (Glezen 20 et al., 1986; Henderson et al., 1979) . Only one licensed monoclonal antibody product 21 (Palivizumab) is currently used to reduce the frequency of severe disease in the 22 3 high-risk neonates (Group, 1998) . The development of a safe and effective RSV 1 vaccine has been elusive thus far. 2 The intramuscularly administered formalin-inactivated RSV (FI-RSV) vaccine 3 primed for enhanced illness in infants on natural RSV infection, this phenomenon was 4 replicable in animal models and named as vaccine-enhanced disease (VED) ( ). Intensive investigation showed that VED is usually involved in Th2-biased 7 immune response and substantial binding antibodies (Ruckwardt et al., 2019) . 8 Subsequent studies then adopted subunit-based vaccines, mostly by targeting cell 9 surface fusion (F) protein or attachment glycoprotein (G), owing to their 18 have failed (Ruckwardt et al., 2019) . Consequently, new RSV vaccine targets and 19 vaccination strategies warrant intensive investigation. 20 Because RSV infection mostly initiates from the upper respiratory tract and is 21 restricted to the lung, an RSV vaccine could prophylactically prevent infection at 22 upper and lower respiratory tracts if mucosal immunity could be elicited effectively in 1 the respiratory mucosa (Yang and Varga, 2014) . Moreover, appropriate immunization 2 routes, such as intranasal (i.n.) administration, and specific adjuvant formulations to 3 induce potent and broad mucosal immune responses are needed. Therefore, we first 4 chose the cholera toxin B subunit (CTB) as a mucosal adjuvant to test whether the 5 RSV internal phosphoprotein (P), nucleoprotein (N), non-structural protein 1 (NS-1) 6 and M2-1 could be potential vaccine targets in BALB/c mice. We found that i.n. 7 administration of P, together with CTB (P+CTB), significantly reduced viral loads in 8 both noses and lungs of mice upon RSV challenge compared with mock immunized 9 mice. This encouraging result urged us to further evaluate P as a vaccine target by 10 engaging more mucosal adjuvants and immunization strategies. 11 However, certain degree of increased inflammatory cell infiltration and mucus 12 production were observed in the P+CTB immunized mice compared to saline 13 immunized mice post RSV challenge, suggesting potential vaccine-induced 14 immunopathogy. Thus, we further tried another mucosal adjuvant flagellin (Mizel and 15 Bates, 2010), and integrated flagellin with the newly identified vaccine target P, as we , where the P and KFD1 were covalently coupled. Here, we report that i.n. 21 5 immunization with P-KFD1 protects mice against RSV infection, as well as lung 1 disease. 2 

4 or P+CTB protect mice against RSV infection 5 We generated a recombinant protein P-KFD1 (Figures S1A and S1B) , and 6 investigated the immunogenicity and anti-viral efficacy of P-KFD1 as a potential 7 mucosal subunit RSV vaccine targeting on P. P-KFD1 retains Toll like receptor 5 8 (TLR5) activity at a marginal lower level than full length KF (Figure S1C). We 9 firstly evaluated the immunogenicity of P by i.n. immunization with P-KFD1 or 10 cholera toxin B subunit (CTB) adjuvanted P protein in BALB/c mice. Briefly, 20 μg 11 of P-KFD1 in saline or 10 μg of P protein mixed with 2 μg of CTB (P+CTB) were 12 prepared as one dose of immunogen. BALB/c mice were immunized with P-KFD1, 13 P+CTB or saline only for sham control, respectively, at weeks (w) 0, 4 and 8 ( Figure   14 1A, upper panel). P-specific IFN-γ secreting T cell response could be detected in 15 both lungs and spleens of P-KFD1 or P+CTB immunized mice ( Figure 1B) . 16 Meanwhile, high levels of P-specific saliva IgA, serum IgG and serum IgA antibody 17 responses could be induced by i.n. immunization with either P-KFD1 or P+CTB 18 compared to sham immunization ( Figure 1C) . These results indicated that 19 recombinant P is immunogenic in mice by i.n. immunization in either P-KFD1 or 20 P+CTB formulation, although the levels of P-specific immune responses induced by 21 P-KFD1 were lower than P+CTB (Figures 1B and 1C) . Next, the protective efficacy 22 6 against RSV infection by i.n. immunization with P-KFD1 or P+CTB was evaluated 1 and compared with that by intramuscular (i.m.) immunization of formalin-inactivated 2 RSV (FI-RSV) in mice. Thus, one more group of mice was immunized with FI-RSV 3 at weeks 8 and 10 with Alum as adjuvant ( Figure 1A, lower panel) . All immunized 4 mice were challenged with RSV strain A2 at week 12, and sacrificed at 4 days post 5 challenge for evaluating viral loads in noses and lungs ( Figure 1A) . Compared with 6 control group, P-KFD1, P+CTB or FI-RSV immunized mice significantly reduced 7 viral loads in both noses and lungs ( Figure 1D ). In addition, viral loads decrease 8 could not be observed in either CTB only or KFD1 only intranasally immunized mice 9 ( Figures 1E and 1F) , suggesting that i.n. immunization of P-KFD1 or P+CTB 10 elicited P-specific immune responses conferred protection against RSV infection.

Intranasal immunization with P-KFD1 protects mice against lung disease caused 12 by RSV infection 13 A primary concern for development of an RSV vaccine is the potential for VED that 14 occurred in FI-RSV vaccinated infants upon natural RSV infection. We thus evaluated 15 the pathological changes in P-KFD1 or P+CTB immunized mice post RSV challenge. 16 In contrast to the quick and massive body weight loss in FI-RSV immunized mice, 17 body weight changes in P-KFD1 or P+CTB immunized mice were not significantly 18 different from those in saline group after RSV challenge ( Figure S2A) . We further 19 evaluated the respiratory function of mice by testing their airway 20 hyper-responsiveness (AHR). As shown in Figure 2A , P-KFD1 immunized mice had 21 normal inhalation resistance (Ri), exhalation resistance (Re) and lung dynamic 22 compliance (Cdyn) after challenge, which were similar to those measured parameters 1 in saline immunized mice and uninfected naï ve mice. The FI-RSV immunized mice 2 showed dramatic increases of both Ri and Re, and significant decrease of Cdyn by 3 excitation with increasing concentrations of methacholine (MCH). The P+CTB 4 immunized mice showed lower values of Ri and Re than the FI-RSV group, but 5 higher values than those of either the P-KFD1 or saline groups. These results 6 indicated that P-KFD1 intranasally immunized mice retained normal respiratory 7 function after RSV challenge, in contrast to FI-RSV immunized mice which showed 8 severe airway obstruction as reported in many studies (Knudson et al., 2015) . In 9 parallel with the respiratory function, we observed lung histopathology changes by 10 H&E and PAS staining at 8 days post RSV infection (dpi) in naï ve or immunized 11 mice. As shown in Figure 2B , consistent with typical VED, overt inflammation and 12 mucus production (black arrows) were observed in lungs of FI-RSV immunized mice. 13 In contrast, neither inflammatory cell infiltration nor mucus secretion was observed in 14 lungs of P-KFD1 immunized mice ( Figure 2B ). It should be noted that more 15 inflammatory cell infiltration and mucus production were observed in the P+CTB 16 immunized mice compared to the P-KFD1 group, but less compared to the FI-RSV 17 group. In detail, the highest pathological scores of immunocyte aggregation around 18 both bronchioles and pulmonary vessels, and the highest score of interstitial 19 pneumonia were observed in FI-RSV immunized mice, while significant lower scores 20 were observed in the P-KFD1 or saline immunized mice ( Figure 2C ). The P+CTB 21 immunized mice showed lower scores of inflammations than the FI-RSV immunized 22 mice, but higher scores than the P-KFD1 immunized mice. In addition, a significant 1 lower score of mucus production was observed in the P-KFD1 immunized mice 2 compared to that in either FI-RSV or P+CTB immunized mice ( Figure 2D ). 3 Furthermore, analysis of differential infiltrated immunocytes in lungs by flow 4 cytometry (FCM) showed an increase of eosinophils and T cells in the FI-RSV 5 immunized mice, and an increase of T cells in the P+CTB immunized mice, compared 6 to those of P-KFD1 or saline immunized mice upon RSV challenge ( Figure 2E ). We 7 also analyzed the Th bias in lungs of the immunized mice at 8 days after challenge by 8 ex-vivo PMA and ionomycin stimulation ( Figure 2F ). Compared to saline immunized 9 mice, percentage of IL-4 secreting cells in CD4 + T cells increased significantly in 10 FI-RSV immunized mice as many studies previously reported (Knudson et al., 2015), 11 while no increase was observed in either the P-KFD1 or P+CTB immunized mice. 12 Moreover, there was no difference in frequency of CD25 + Foxp3 + Treg ( Figure S2B ) 13 or IFN-γ + CD4 + T cells among all groups of mice. It should be noted that a significant 14 elevated percentage of IL-17A + CD4 + T cells was detected in P+CTB immunized 15 mice. In summary, in contrast to FI-RSV immunization, i.n. immunization with 16 P-KFD1 rather than P+CTB completely avoided occurrence of VED. Taken the 17 anti-RSV efficacy in account ( Figure 1D) , P-KFD1 might be a potential safe and 18 effective RSV vaccine candidate. 20 The encouraging protective effects conferred by P-KFD1 i.n. immunization urged us 21 to further investigate the relationship between protective efficacy and immunization 22 strategy. We found that inoculation with 20 μg of P-KFD1 provided better protection 1 against infection than with a dosage of 5 μg or 1.25 μg (Figure S3A) , and boosting 2 twice provided better protection against infection than once ( Figure S3B) . Hence, the 3 protective efficacy of P-KFD1 i.n. immunization is dependent on both dosage of 4 immunogen and times of immunization. These results suggested that the degree of 5 immune responses induced by P-KFD1 i.n. immunization is correlated with protective 6 efficacy. Then we tried to elucidate, whether humoral immune response, or cellular 7 immune response, or both, mediated protection in P-KFD1 immunized mice. 8 Based on previous studies on the intracellular neutralization activities of 9 anti-viral IgA in vitro and in vivo through pIgR mediated transcytosis (Burns et al., Figure S3C ). In parallel, P-KFD1 i.n. immunization induced lower 16 level of P-specific serum IgG, but higher level of serum IgA and undetectable 17 P-specific saliva IgA in pIgR -/mice compared to those in wild type mice ( Figure 3A) . 18 Despite the striking difference in P-specific secreting antibody response, the 19 protection provided by P-KFD1 i.n. immunization against RSV infection in both 20 noses and lungs of pIgR -/mice was as similar as that of wild type mice ( Figure 3B) . 21 This result suggested that P-specific mucosal IgA antibodies did not mediate the 22 protection against RSV by intraepithelial neutralization related to pIgR. We then 1 performed anti-serum transfer experiment and found no anti-viral efficacy could be 2 conferred to naï ve recipient mice by transfer of serum from the P-KFD1 immunized 3 mice ( Figure 3C ). It suggested that P-specific antibodies in serum were not the main 4 protective factors. Collectively, P-specific humoral immune response induced by 5 P-KFD1 i.n. immunization does not appear to be the main mediator for protection 6 against RSV infection. 7 Thus, we speculated that the T cell response induced by P-KFD1 i.n. 8 immunization do mediate protection. As showed in Figures 1B and S3D , P-KFD1 i.n. 9 immunization induced P-specific IFN-γ, IL-17A but not IL-4 secreting T cell 10 responses. P-specific IFN-γ secreting T cells could be detected upon stimulation with 11 P in both lungs and spleens at day 7 post second boost immunization (8w+7d), 12 decreasing to nearly undetectable levels thereafter (8w+28d), but starting to rebound 13 after challenge at 1 dpi (12w+1d) and increasing to a still higher level at 4 dpi 14 (12w+4d) ( Figure 3D ). Accordingly, after challenge, the virus replicated only from 15 1×10 3 at 1 dpi to 1.5×10 3 at 4 dpi in lungs of P-KFD1 immunized mice, but replicated 16 dramatically from 0.5×10 3 at 1 dpi to 1×10 5 at 4 dpi in lungs of saline immunized 17 mice ( Figure S3E ). This delayed anti-virus effect implied that T cells might take part 18 in the protection conferred by P-KFD1 i.n. immunization. With regard to IFN-γ 19 production by purified CD4 + or CD8 + T cells, P-KFD1 i.n. immunization induced 20 much higher fraction of P-specific IFN-γ producing CD4 + T cells than the fraction of 21 P-specific IFN-γ producing CD8 + T cells ( Figure 3E ). This indicated that the 22 P-specific CD4 + T cells elicited by P-KFD1 i.n. immunization should be the main 1 cellular source of IFN-γ. 2 To confirm the role of T cells played in the protection against infection, CD4 + 3 and/or CD8 + T cells were depleted in vivo by using anti-CD4-specific and/or 4 anti-CD8-specific antibodies ( Figure S3F ). When T cells were depleted 5 simultaneously with anti-CD4 and anti-CD8 antibody together (α-CD4 + α-CD8), the 6 anti-viral activity in P-KFD1 immunized mice was abrogated in both noses and lungs 7 ( Figure 3F ). When only CD4 + T cells were depleted, the anti-viral efficacy in 8 P-KFD1 immunized mice declined significantly to the same level as that in saline 9 immunized mice. However, if only CD8 + T cells were depleted, anti-viral efficacy 10 remained intact in both noses and lungs ( Figure 3F) . Hence, the protection against 11 RSV conferred by i.n. immunization with P-KFD1 was mainly CD4 + T cell 12 dependent. 13 Figure 4A , FTY720 treatment did not affect the number of P-specific IFN-γ 22 producing T cells in spleens of P-KFD1 immunized mice, but rather decreased those 1 in lungs significantly from about 200 to 70 SFC/10 6 lymphocytes. This result 2 indicated that in lung post infection, more than half of the P-specific IFN-γ producing 3 T cells migrated from lymphoid organs, and could be blocked by FTY720 treatment. 4 Consistent with the decreased migration of T cells, the anti-viral efficacy of P-KFD1 5 immunized mice was mitigated by the FTY720 treatment, while no distinguishable 6 difference was observed in saline immunized mice, irrespective of FTY720 treatment 7 ( Figure 4B ). This suggested circulating T cells took part in protection afforded by 8 P-KFD1 i.n. immunization. Meanwhile, FTY720 treatment just partially abrogated the 9 anti-viral efficacy of the P-KFD1 immunized mice, in which viral titers in both noses 10 and lungs were still significantly lower than those of the saline immunized mice, also 11 treated with FTY720 ( Figure 4B ), further demonstrating that resident T cells in the 12 lungs also played a protective role in P-KFD1 i.n. immunization. On the other hand, 13 compared with i.n. immunization, intraperitoneal (i.p.) immunization with P-KFD1 14 induced comparable levels of P-specific IFN-γ + T cell response and production of 15 IFN-γ in the spleen, but lower levels in lung ( Figure S4 ). In addition, i.p. 16 immunization with P-KFD1 resulted in significantly reduced viral load in lung, but 17 not in nose ( Figure 4C ). Taken together, both the migrated T cells from lymphoid 18 organ and the resident T cells in the respiratory tract contributed to local in situ 19 anti-viral immunity in P-KFD1 immunized mice. 20 To further verify that it was CD4 + T cells evoked by P-KFD1 immunization that 21 directly contributed to protection against viral infection, we performed adoptive 22 13 transfer experiments. Briefly, CD4 + T cells from the spleens of intraperitoneally or 1 intranasally immunized mice were sorted by antibody-conjugated microbeads and 2 then transferred into naï ve mice prior to RSV challenge. Mice that received CD4 + T 3 cells from P-KFD1 intraperitoneally immunized donors had lower viral loads in both 4 noses and lungs post challenge ( Figure 4D ). Mice that received CD4 + T cells from 5 P-KFD1 intranasally immunized mice also exhibited lower viral load in lungs, but 6 only marginally lower viral titer in the noses ( Figure 4E ). Collectively, these data 7 suggested that the P-specific CD4 + T cells induced by P-KFD1 i.n. immunization, 8 either resided in, or migrated to the respiratory tract, played key protective roles 9 against RSV infection. CD4 + T cell subset and antigen-unexperienced CD11a low CD49d -CD4 + T cell subset 18 from lungs of saline or P-KFD1 intranasally immunized mice on 4 dpi by FACS 19 ( Figures 5A and S5A ). As expected, P-specific IFN-γ + CD4 + and IL-17A + CD4 + T 20 cell responses could only be detected in the CD11a high CD4 + T cell subsets of P-KFD1 21 immunized mice ( Figure S5B ). To focus on these responsive CD11a high cells in the 22 CD4 + T cells, we mixed the sorted CD11a high subset with hashtag antibody labeled 1 sorted CD11a low CD49dcell subset at a ratio of approximately 5:3, before performing 2 sc-RNA seq of the transcriptome profile and TCR repertoire of CD4 + T cells ( Figure   3 5B). Unsupervised hierarchical clustering and t-distributed stochastic neighbor 4 embedding (t-SNE) dimensional reduction analysis based on the transcriptomes of 5 pooled samples identified fifteen CD4 + T cell clusters (Figures 5C and S5C ). The 6 cells in clusters 1, 3 and 7 were naï ve-like CD4 + T cells, since they expressed 7 transcripts of Ccr7, Lef1 and Igfbp4 (Figures 5D and S5E ). Cells in cluster 0 highly 8 expressed transcripts of Itgb1, Cd40lg, Il2, and Ifng, which resembled effector 9 memory or tissue resident memory CD4 + T cells (Tem/Trm) ( Figure 5D ). Whereas, 10 cluster 4 cells highly expressed transcripts of Il2ra, Foxp3 and Ikzf2 which seemed 11 like Treg cells. Notably, cluster 6 cells were enriched in Il17a, Il17f, Ccr6 and Rorc, 12 indicative of Th17 subset ( Figure 5D ). Other clusters were also designated based on 13 the signature genes expression ( Figure 5D ). Moreover, sc-RNA seq revealed a 14 significantly higher proportion of Th17 cells in P-KFD1 induced CD4 + T cells 15 compared to saline group ( Figure 5C ). 16 As expected, the hashtag labeled antigen-unexperienced CD11a low CD49d -CD4 + 17 T cells were mainly located in the naï ve-like clusters 1, 3 and 7 cells (Figures 5C and   18 S5D), and showed genes expression pattern similar as naï ve CD4 + T cells ( Figure   19 S5E). Moreover, similar as that in the naï ve-like cells, differential expression genes 20 (DEGs, log 2 |FC|＞1, P<0.05) were rarely found in hashtag + cells between P-KFD1 21 group versus saline group ( Figure S5F ). In contrary, large amount of DEGs (139 22 genes) could be identified in the non-naï ve cells that excluded clusters 1, 3 and 7 in 1 P-KFD1 group compared to saline group ( Figure 5E ). David analysis revealed that 2 the upregulated DEGs in P-KFD1 group were enriched in two KEGG pathways 3 "cytokine-cytokine receptor interaction" and "inflammatory bowel disease", and in 4 two Gene Ontology (GO) biology processes "inflammatory response" and "immune 5 response" in P-KFD1 group (Figure 5F ). The representative DEGs included Th17 6 marker genes Il17a, Il17f, Rorc, Ccr6, activation related marker genes Il2ra, Cxcr6, 7 and cytotoxicity-associated genes Lta etc. (Figures 5G and S5G ). 8 Then we aimed at the set of T cells with the same TCRs, which were the 9 sequencing identified clonal expanded CD4 + T cells. TCR sequencing totally 10 identified 5695 and 5769 barcodes in saline group and P-KFD1 group, respectively. It 11 is obviously that the frequency and the number of T cells with repeatedly used TCRs 12 in P-KFD1 group (625) were much higher than that in saline group (330) (Figures 6A, 13 S6A and S6B). Moreover, the clone sizes of CD4 + T cells in P-KFD1 group were 14 much larger than that in saline group ( Figure 6B ). In addition, in P-KFD1 group, a 15 large portion of the CD4 + T cells with repeatedly used TCRs were located in cluster 6 16 Th17 cells. These suggested epitope-specific CD4 + T cell responses especially the 17 Th17 cells were elicited by P-KFD1 i.n. immunization. Consistently, between saline 18 group and P-KFD1 group, TCR usage in the non-naï ve CD4 + T cells that excluded 19 clusters 1,3,7 or cells with repeatedly used TCRs were significantly different ( Figures   20   S6C and S6D) . In CD4 + T cells with repeatedly used TCRs, large amount of DEGs 21 existed between P-KFD1 group versus saline group ( Figure 6C) . These upregulated 22 genes in P-KFD1 group also included Th17 related transcripts of genes (Ccr6, Il17a, 1 Il17f and Il23r), activation related genes such as Il2ra and Cxcr6, as well as some 2 immune response-associated genes such as Ccr1, Csf1, Lta, Il1r1 and Tnfsf11 3 (Figures 6D and S6E) . The results above all suggested that P-KFD1 i.n. 4 immunization induced expansion and activation of Th17 cells on 4 dpi at 5 transcriptional level. 6 ELISpot assay of pulmonary immunocytes showed that much higher levels of 7 P-specific IL-17A + T cell response were elicited in P-KFD1 intranasally immunized 8 mice on 4 days after RSV challenge ( Figure 6E ). Generally, P-KFD1 i.n. 9 immunization boosted P-specific Th17 immune responses. (Figure 7A) . The CFSE-labeled cells derived from donor mice could be 18 detected and differentiated as un-proliferated CFSE high cells and proliferated CFSE low 19 cells ( Figure 7B ). As depicted in Figure 7C , in the CFSE + cells, percentages of 20 CFSE low CD4 + T cells from P-KFD1 immunized mice were significantly higher than 21 those from KFD1 or saline immunized donors, in both lungs and spleens of the 22 recipient mice, but CFSE low CD8 + T cells were not. This result indicated that 1 P-specific CD4 + T cells induced by P-KFD1 i.n. immunization could expand upon 2 RSV challenge. 3 In order to detect potential atopy of Th responses, lymphocytes of the recipient 4 mice were stimulated by PMA and ionomycin ex vivo and analyzed. The percentages 5 of IFN-γ + cells, IL-17A + cells or IL-4 + cells in CFSE -CD4 + T cells did not differ 6 among the three different recipient groups in lungs and spleens, respectively, 7 suggesting that adoptive transfer did not change Th bias of recipient mice after RSV 8 challenge (Figures 7D and 7E, hollow column) . In CFSE + CD4 + T cells, both the 9 percentages of IFN-γ + cells and IL-17A + cells from P-KFD1 immunized donors were 10 significantly higher than those from saline immunized donors, in lungs and spleens of 11 recipients respectively. Whereas, no differences were found in percentages of CFSE + 12 IL-4 + CD4 + T cells among all groups (Figures 7D and 7E, solid column) . These 13 results indicated P-KFD1 i.n. immunization induced both Th1 and Th17 responses at 14 4 dpi. We also found that the percentages of CFSE + IFN-γ + CD8 + T cells from 15 P-KFD1 intranasally immunized donors were slightly higher than those from saline 16 immunized donors ( Figure S7A) . However, the percentages of CFSE + IL-17A + CD8 + 17 T cells were all at low levels and showed no difference among the recipients ( Figure   18 S7A). 19 Furthermore, the percentages of CFSE + IFN-γ + CD4 + T cells, CFSE + IL-17A + 20 CD4 + T cells ( Figure 7D ) and CFSE + IFN-γ + CD8 + T cells ( Figure S7A ) from 21 P-KFD1 immunized donors were higher than those from KFD1 immunized mice. 22 These results implied that P-KFD1 i.n. immunization induced P-specific Th1 and 1 Th17 responses, as well as IFN-γ + CD8 + T cell response, especially in RSV infected 2 tissue. When the same CFSE labeling and adoptive transfer experiments were 3 performed for the FI-RSV and P+CTB immunized mice, a dramatically higher level 4 of CD4 + T cell proliferation at 4 dpi could be observed compared to that of P-KFD1 5 immunized mice ( Figure S7B) . Besides, a robust Th2-biased response in the FI-RSV 6 immunized mice and a robust Th17-biased immune response in the P+CTB 7 immunized mice could be observed in both lungs and spleens, respectively ( Figure   8 S7C), which were also consistent with the results in Figure 2F . 9 To determine whether the Th1 and/or Th17 responses mediated anti-viral 10 efficacy in P-KFD1 intranasally immunized mice, antibodies against IFN-γ (α-IFN-γ) 11 and/or IL-17A (α-IL-17A) were injected into the P-KFD1 immunized mice during 12 RSV infection, respectively. As shown in Figure 7F , simultaneous administration of 13 both anti-IFN-γ and anti-IL-17A antibodies could abolish anti-viral efficacy in both 14 noses and lungs of P-KFD1 intranasally immunized mice. In contrast, administration 15 of only anti-IFN-γ antibody or anti-IL-17A antibody could not abrogate anti-viral 16 efficacy; instead, significantly lower viral loads were retained in both noses and lungs 17 of P-KFD1 intranasally immunized mice compared to those of saline immunized mice, 18 respectively. These data suggested that T cell-derived IFN-γ or IL-17A alone could 19 mediate the inhibition of RSV replication in vivo. Taken together, P-KFD1 i.n. 20 immunization induced P-specific IFN-γ + CD4 + and IL-17A + CD4 + T cell responses 21 without changing Th bias, and provided protection against RSV infection depended 1 on IFN-γ or IL-17A. 12 Here, we reported how we used new strategies to develop a safe and effective 13 mucosal RSV vaccine targeting an internal RSV antigen P that has been seldom 14 investigated. And we examined the immune responses in relation to virological and 15 pathological features of the fusion protein P-KFD1 intranasally immunized animals 16 before and after RSV challenge by comparison with those immunized with P+CTB or 17 FI-RSV. Our data showed that P-KFD1 i.n. immunization induces P-specific immune 18 responses and inhibits RSV replication in both the upper and lower respiratory tract of 19 mice. It is noteworthy that P-KFD1 i.n. immunization does not result in VED, 20 whereas P+CTB i.n. immunization do cause some enhanced respiratory disease (ERD) 21 20 which is different from the typical VED featured by the infiltration of eosinophils and 1 highly Th2-biased immune response in FI-RSV immunized mice. 2 As part of our initial hypothesis, P-specific secreted IgA response in P-KFD1 3 immunized mice would play a role in prevention against viral infection. However, 4 neither P-specific secreted IgA nor P-specific serum was found to be relevant to 5 protection against RSV infection in our immunization regimen, though our 6 experiments performed with pIgR -/mice, or by serum transfer, could not completely 7 exclude the possible roles of antibodies or B cells through other unknown 8 mechanisms. 9 We thus analyzed the kinetics of T cell response associated with P-KFD1 i.n. 10 immunization and RSV challenge. We demonstrated that P-specific CD4 + T cells, Th17 representative genes (such as Rorc, Ccr6, Il17a and Il17f) and activation related 15 genes (such as Il2ra and Cxcr6), some cytotoxicity-associated genes such as Lta, and 16 Tnfsf11 were also significantly upregulated. Future investigation to determine whether 17 these cytotoxicity-associated genes contribute to the protection against RSV infection 18 is warranted. In our adoptive transfer experiments, both Th1 and Th17 immune 19 responses were detected in the P-KFD1 and P+CTB groups, but the magnitude of 20 Th17 immune response in the P+CTB group was much higher than that in the 21 P-KFD1 group. A kind of aberrant Th17 response might explain why impaired 22 respiratory function, inflammatory cell infiltration and mucus production occurred in 1 the P+CTB immunized and RSV challenged mice. The underlying mechanism also 2 needs further investigation. 15 Our animal model is a limiting factor in this study. Mice are semi-permissive for 16 RSV replication. BALB/c mice infected with commonly used laboratory RSV strains, 17 including the A2 strain we used, do not exhibit high viral loads or pulmonary mucus 18 (Meng et al., 2014) . Nevertheless, either viral load or pulmonary mucus production 19 was measurable for comparatively evaluating both anti-viral efficacy and 20 immunopathology in the mice subjected to different vaccinations in our study. 21 Another caveat is that although the Th17 immune response in P-KFD1 intranasally 22 immunized mice were characterized, the detail protective mechanism remains elusive. 1 Further investigation on the mechanism and function of Th17 response during RSV 2 infection is warranted to strike an optimal balance between anti-viral activity and 3 immunopathology. Finally, we asked if P-specific T cell responses could be boosted 4 to afford greater protection against RSV infection in human beings, even though 5 researchers (Guvenel et al., 2020) have already detected P-specific T cell responses in 6 peripheral blood of healthy adults under experimental RSV inoculation. 7 In summary, our study provides a new concept for RSV vaccine development, 8 taking into account a balanced Th1/Th17 CD4 + T cell response for evaluation of an 9 RSV vaccine. This concept is exemplified by P-KFD1 i.n. immunization that 10 generated RSV P-specific CD4 + T cell responses and conferred protection against 11 RSV infection and disease in mice. Overall, the E. coli-produced recombinant 12 flagellin-phosphoprotein can prevent RSV infection without raising any VED 13 concerns. 14 ACKNOWLEDGMENTS 15 We thank Professor Zishu Pan at Wuhan University for providing RSV strain A2, and 16 Professor Xu Yang at Central China Normal University for technical assistance in 17 detecting airway responsiveness of mice. We thank Xuefang An, Li Li, He Zhao, Dr. 15 independent experiments. In Figure 1B , two columns were compared using unpaired t 16 test. In Figures 1C to 1F , groups were compared using one-way ANOVA. * p < 0.05, 17 ** p < 0.01, *** p < 0.001, ns means non-significant. See also Figure S1 . in saline or P-KFD1 intranasally immunized mice that were treated with depletion 12 antibodies against CD4 + and/or CD8 + T cells were detected on 4 dpi, n = 5~6 mice per 13 group. Data are represented as mean ± SEM of two independent experiments. In 14 Figures 3A and 3B , two groups were compared using un-paired t test. In Figure 3C , 15 groups were compared using one-way ANOVA. In Figures 3D to 3F , groups were 16 compared using regular two-way ANOVA. * p < 0.05, ** p < 0.01, *** p < 0.001, ns 17 means non-significant. See also Figure S3 . intravascularly injected into naï ve mice, respectively, followed by RSV challenge. 10 Viral loads in noses and lungs of recipients that received CD4 + T cells from 11 intraperitoneally immunized donors (D) or from intranasally immunized donors (E) 12 were detected at 4 dpi, n = 6 mice per group. Data are represented as mean ± SEM of 13 two pooled experiments for Figures 4A and 4C . Data are represented as mean ± 14 SEM of two independent experiments for Figures 4B, 4D and 4E . In Figure 4A , 15 groups were compared using regular two-way ANOVA. In Figures 4B to 4E , groups 16 were compared using one-way ANOVA. * p < 0.05, ** p < 0.01, *** p < 0.001, ns 17 means non-significant. See also Figure S4 . spleens (E) of recipients were assessed at 4 dpi, n = 4~5 mice per group. (F) Saline or 10 P-KFD1 intranasally immunized mice were treated with antibodies to block IFN-γ, 11 IL-17A or both IFN-γ and IL-17A during RSV infection, respectively, viral loads 12 were assessed on 4 dpi, n = 4~5 mice per group. Data are represented as mean ± SEM 13 of at least two independent experiments. In Figures 7C and 7F , groups were 14 compared using regular two-way ANOVA. In Figures 7D and 7E , groups were 15 compared using one-way ANOVA. * p < 0.05, ** p < 0.01, *** p < 0.001, ns means 16 non-significant. See also Figure S7 . 17 

Lead contact 21 32 Further information and requests for resources and reagents should be directed to and 1 will be fulfilled by the Lead Contact, Huimin Yan (yanhuimin@shphc.org.cn) 2 Materials Availability 3 This study did not generate new unique reagents. Mice were intranasally (i.n.) immunized three times with P-KFD1 or P+CTB at 8 four-week intervals after anesthesia with pentobarbital sodium (50 mg/kg), and were 9 challenged with 2 ⅹ 10 6 PFU of RSV A2 on 28 days post last immunization. 10 Mice were intramuscularly (i.m.) immunized with FI-RSV in the lower hind limb 11 twice at a two-week interval and were challenged with RSV A2 14 days post boost 12 immunization. 13 E.coli BL21 DE3, purified by affinity chromatography on a Ni-NTA column (Qiagen), 7 and removed contaminating lipopolysaccharide (LPS), respectively, as previously 8 described (Yang et al., 2013) . The residual LPS content was determined using the 9 Limulus assay (Associates of Cape Cod) to be less than 0.01 EU/μg protein. (Waris et al., 1996) . Virus stocks or tissue homogenates from infected mice were 10 titrated by immuno-plaque assay as described elsewhere (Quan et al., 2011) . Titers 11 were recorded as PFU/nose or PFU/g lung. 12 Enzyme-Linked Immunosorbent Assay (ELISA) 13 Antibody responses were assessed by enzyme-linked immunosorbent assay (ELISA Airway responses to methacholine challenge 3 Airway responsiveness was assessed in mice using an AniRes2005 lung function Healthcare) and layered carefully onto 70% Percoll to generate discontinuous Percoll 10 gradients and centrifuged at 2000 rpm for 25 minutes at 22℃. Cells were aspirated in 11 the interface between 40% and 70% Percoll gradients. 12 For isolation of splenic lymphocytes in mice, spleens were minced and pressed 13 through the 70-μm nylon mesh screen. Lymphocytes were obtained by density In vivo antibody treatment 10 For depletion of CD4 + , CD8 + , or both CD4 + and CD8 + T cells, mice were (HRPN), mouse IgG 1 (MOPC-21), or both rat IgG 1 and mouse IgG 1 were assigned 1 as isotype control groups, respectively. 2 All antibodies used above were purchased from Bio X Cell.

3 FTY720 treatment 4 Mice were intraperitoneally injected with 250 μL of FTY720 (Sigma) at a dose of 1 5 mg/kg daily beginning 3 days before RSV challenge, until they were sacrificed.

Purification of splenic CD4 + T cells and adoptive transfer 7 Splenocytes were separated from saline or P-KFD1 immunized mice at 7 days post 8 last immunization. Then CD4 + T cells were positively selected using anti-CD4 (L3T4) 9 microbeads following the manufacturer's instructions (Miltenyi Biotec, Germany). 10 The purity of sorted CD4 + T cells was at least 95%. Naï ve recipient mice were 11 intravenously injected with 3×10 6 purified CD4 + T cells 12 hours prior to RSV 12 challenge and sacrificed at 4 days post infection. 13 FACS sorting, co-culture and sc-RNA seq 14 Firstly, CD4 + T cells were positively isolated by magnetic sorting (Miltenyi Biotec, 15 Germany) from pulmonary immunocytes of saline or P-KFD1 intranasally immunized 16 mice on 4 dpi. Then the purified CD4 + T cells were stained with monoclonal 17 antibodies specific to CD3 (clone 145-2C11), CD4 (clone RM4-5), CD11a (clone 18 M17/4), CD49d (clone 9C10) and 7-AAD viability staining solution (BioLegend) for 19 FACS sorting on a FACSAriaIII (BD, Heidelberg, Germany). 20 For detection of P-specific immune responses, the FACS sorted CD11a high CD4 +

T cells and CD11a low CD49d -CD4 + T cells were stimulated with RSV P protein or 22 41 irrelevant protein or medium alone in 96-well plates for 48 hours, respectively. In 1 detail, 3ⅹ10 4 sorted cells were co-cultured with 1ⅹ10 4 naï ve mice derived splenocytes 2 in a well. And the collected cell supernatants were used to detect the amount of IFN-γ, 3 IL-4 and IL-17A by ELISA. were not compatible with the standard cell surface protein library procedure were 20 used in this study. Therefore, the Hashtag cell surface libraries were constructed as 21 following steps. Partial of the amplified cDNAs were ligated with illumina read 2 22 adaptor, then with the indexing PCR to get the Hashtag libraries. Libraries were 1 sequenced on Illumina NovaSeq6000 sequencing platform with the following read 2 lengths: read 1-150 cycles; read 2-150 cycles; and i7 index-8 cycles. 3 Single cell transcriptome analysis 4 The Cell Ranger Single-Cell Software Suite (versions 3.0.2) were used to perform 5 barcode processing and single-cell gene counting (Paulson et al., 2018) 6 (http://10xgenomics.com/). First, "cellranger mkfastq" was carried to generated fastq 7 files. Second, feature-barcode matrices were generated by "cellranger count" function 8 using GRCm38 mouse as reference genome (Ensembl). Then the cell ranger 9 aggregation function (aggr) was used to combine the two libraries. A correction for 10 sequencing depth was also performed during the aggregation. IntegrateData function 11 in R package Seurat V3.6.3 was used to merge Saline and P-KFD1 fastq data. 12 Low-quality genes and cells were filtered by removing cells with 1) expressed genes 13 fewer than 200, 2) expressed genes more than 5,000, 3) percentages of mitochondrial 14 genes >20% and 4) genes expressed in less than 3 cells. The filtered gene-barcode 15 matrix was first normalized using "LogNormalize" method. The top 2,000 variable 16 genes were then identified using the "vst" method in Seurat FindVariableFeatures 17 function. Principal component analysis (PCA) was performed using variable genes, 18 and the top 20 principal components (PCs) were used to perform t-distributed 19 Stochastic Neighbor Embedding (t-SNE) to visualize the cells. The resolution was set 20 to 0.5 for clustering. T-SNE coordinate points and cell clusters were exported into 21 Loupe Cell Brower V3.1.1 (combine data by cell ranger aggregation function) to 22 analyze data. For the TCR data, the Cell Ranger Single-Cell Software Suite (versions 1 3.0.2) were used to perform barcode processing, assembly contig, cell calling, 2 annotation the contigs and CDR3 region for each clonotype. KEGG (Kyoto 3 Encyclopedia of Genes and Genomes) pathways and GO (Gene ontology) terms 4 analysis were performed as follows: genes with Benjamini-Hochberg-adjusted P value 5 < 0.05 and log 2 |FC| between two groups larger than 1 were used for DAVID analysis 6 (Guo et al., 2018) (https://david-d.ncifcrf.gov/). 7 8 Statistical parameters including the exact value of n, the definition of center, 9 dispersion, and precision measures (geometric mean ± SEM) and statistical 10 significance are displayed in Figures and Figure Legends . Data were considered to be 11 statistically significant if p < 0.05. All statistical analyses were performed using 12 GraphPad Prism version 5.0 (GraphPad Software, San Diego, CA). Data were 13 compared using unpaired, two-tailed t test between two groups or one-way ANOVA 14 with Tukey's multiple comparison for more than two groups. Data were also analyzed 15 by regular two-way ANOVA if two independent variables existed in one experiment. 

Protection against influenza virus infection in polymeric Ig 21 receptor knockout mice immunized intranasally with adjuvant-combined vaccines

Human Anti-fungal Th17 Immunity and 25

Reduces Hospitalization From Respiratory Syncytial Virus Infection in High-risk Infants

Pulmonary toxicity and adjuvant effect of di-(2-exylhexyl) phthalate in ovalbumin

Global characterization of T cells in non-small-cell lung cancer by single-cell sequencing

Epitope-specific airway-resident CD4+ T cell dynamics during 20 experimental human RSV infection

Alteration of lymphocyte trafficking by sphingosine-1-phosphate 2 receptor agonists

Induction of Potent Neutralizing Antibody Responses 5 by a Designed Protein Nanoparticle Vaccine for Respiratory Syncytial Virus

Quantifying antigen-specific CD4 T cells during a viral 8 infection: CD4 T cell responses are larger than we think

Memory CD4+ T cells protect against influenza through multiple synergizing mechanisms

Structure-based design of a fusion glycoprotein vaccine for 14 respiratory syncytial virus

Structure of RSV fusion glycoprotein trimer bound to a prefusion-specific 17 neutralizing antibody

An overview of respiratory syncytial 19 virus

Flagellin as an adjuvant: cellular mechanisms and potential

IL-17-induced pulmonary pathogenesis during respiratory viral infection and 24 exacerbation of allergic disease

Global burden of acute lower respiratory infections due to 27 respiratory syncytial virus in young children: a systematic review and meta-analysis

Single-Cell RNA Sequencing of Tumor-Infiltrating NK Cells 31

Reveals that Inhibition of Transcription Factor HIF-1alpha Unleashes NK Cell Activity

Pulmonary eosinophilic response to respiratory 34 syncytial virus infection in mice sensitized to the major surface glycoprotein G

Acquired cancer resistance to combination 38 immunotherapy from transcriptional loss of class I HLA

Viruslike particle vaccine induces protection against respiratory syncytial virus infection in mice

Immunological Lessons from Respiratory

Syncytial Virus Vaccine Development

Commensal Candida albicans Positively Calibrates Systemic Th17

Global, regional, and national disease burden 5 estimates of acute lower respiratory infections due to respiratory syncytial virus in young children in 6 2015: a systematic review and modelling study

Generation of polymeric immunoglobulin 9 receptor-deficient mouse with marked reduction of secretory IgA

Flagellin-PAc fusion protein is a high-efficacy anti-caries mucosal vaccine

Respiratory syncytial virus infection in 14 mice

Atypical pulmonary eosinophilia is mediated by 16 a specific amino acid sequence of the attachment (G) protein of respiratory syncytial virus

Respiratory synctial virus 19 infection in BALB/c mice previously immunized with formalin-inactivated virus induces enhanced 20 pulmonary inflammatory response with a predominant Th2-like cytokine pattern

Preexisting influenza-specific CD4+ T cells correlate 24 with disease protection against influenza challenge in humans

Characterization of the T cell response to 26 human rhinovirus in children: implications for understanding the immunopathology of the common 27 cold

Shows High Protective Efficacy 30 against Dental Caries with Low Potential Side Effects

Antigen replacement of domains D2 and D3 in flagellin promotes mucosal IgA production and 33 attenuates flagellin-induced inflammatory response after intranasal immunization

Mucosal vaccines against respiratory syncytial virus

Immunoglobulin A Targeting on the N-Terminal Moiety of Viral Phosphoprotein Prevents

Measles Virus from Evading Interferon-beta Signaling

Airway Memory CD4(+) T Cells Mediate Protective 42 Immunity against Emerging Respiratory Coronaviruses