Vero E6 (CRL-1586), A549 (CCL-185), HEp-2 (CCL-23), and HEK-293 (CRL-1573) (all from American Type Culture Collection (ATCC), Manassas, VA) were maintained in 10% fetal bovine serum (FBS)/DMEM (Hyclone, Logan, UT). CX3CR1.293 cells (>90% CX3CR1⁺) were maintained in selection media (10% FBS/DMEM + 1.0 µg/mL puromycin) as previously described (51). RSV A2 and B1 were propagated in HEp-2 cells as described (55). RSV A2 expressing green fluorescent protein (GFP) was propagated in HEp-2 cells as described (56).

Nanoparticle (NP) vaccines were constructed using self-assembling Aquifex aeolicus lumazine synthase (57) fused to a next-generation SpyCatcher domain (54, 58, 59). To generate the NPs, a pET28a plasmid encoding an N-terminal 6-histidine tag, the Aquifex aeolicus lumazine synthase protein (UniProtKB entry O66529), and SpyCatcher003 (58) ( Table 1 ) was transformed into T7 Express E. coli and recombinant LumazineSynthase-SpyCatcher003 was expressed overnight at 18C. Cells were lysed by ultrasonication in wash buffer (10 mM Tris-Cl pH 8.0, 10 mM imidazole, 150 mM NaCl) with 1 mM MgCl₂, protease inhibitors, benzonase, and DTT. E. coli lysates were clarified by centrifugation and 0.22 um filtered. LumazineSynthase-SpyCatcher003 was purified from clarified lysates by affinity chromatography using a HisPur Nickel-NTA Resin and eluted using wash buffer with 250 mM imidazole. LumazineSynthase-SpyCatcher003 was dialyzed into PBS (pH 7.4) overnight at 4°C, resulting in empty vehicle control NPs. For negative stain imaging, LumazineSynthase-SpyCatcher003 protein was deposited onto glow-discharged, carbon-coated 400 mesh copper grids, stained with 2% (w/v) uranyl-formate, and viewed on a 200 kV FEI Glacios transmission electron microscope.

To generate CCD protein antigens, a synthetic gene encoding an N-terminal SpyTag003 (58) fused to RSV strain A2 G protein amino acids 157 to 197 (UniProtKB entry P03423) and a C-terminal 6-histidine tag was cloned into pRSFDuet-1 ( Table 1 ). Recombinant SpyTag003-RSV G CCD WT and SpyTag003-RSV G CCD S177Q proteins were expressed in T7 Express E. coli overnight at 18C. Cells were lysed by ultrasonication in wash buffer (20 mM Tris-cl pH 8.0, 25 mM imidazole, 150 mM NaCl) with 1 mM MgCl₂, protease inhibitors, and benzonase. E. coli lysates were clarified by centrifugation and 0.22 um filtered. SpyTag003-RSV G CCD WT and SpyTag003-RSV G CCD S177Q proteins were purified from clarified lysates by affinity chromatography using a HisTrap FF crude column and washed with wash buffer containing 6 M urea. Protein was eluted in wash buffer containing 500 mM imidazole. This CCD purification method has been used successfully to solve its structure bound to anti-G protein antibodies, confirming that the recombinant CCD protein adopts an antigenically-relevant conformation (52, 60, 61).

To generate CCD-coated NPs, LumazineSynthase-SpyCatcher003 protein was incubated with a 4 M excess of SpyTag003-RSV G CCD WT or SpyTag003-RSV G CCD S177Q overnight at 4°C. During this incubation, a covalent isopeptide bond is formed between the SpyTag003 and the SpyCatcher003 which is verified by SDS-PAGE and a shift in the molecular weight band. Aggregation was observed the following day and was confirmed to be the NP samples by SDS-PAGE. Insoluble pellets were resuspended in PBS (pH 7.4) to make a final concentration of 1 mg/ml. Protein concentrations for all NP samples were verified by Bradford assays.

Female BALB/c mice (10-to-12-weeks old; Jackson Laboratories, Bar Harbor, ME) were housed in micro isolator cages with 12h light/dark cycle, and fed ad libitum. The mice received a priming dose of 10 µg NP-WT, NP-S177Q, or empty NPs. All vaccines were adjuvanted with 10 µg monophosphoryl Lipid A (MPLA; VacciGrade™ from S. Minnesota R595, InvivoGen, San Diego, CA), a TLR4 agonist, diluted in PBS. Similar to a related study that used using SpyCatcher multimerization of a SARS-CoV-2 spike vaccine candidate to induce a potent neutralizing antibody response at 21 days post-priming (62), vaccinated mice were boosted with either 10 µg or 25 µg of homologous vaccine or empty NPs and 10 µg MPLA diluted in PBS. Mice were i.m. vaccinated in the left and right and left quadriceps with 0.05 mL/quadriceps. Sera were collected on days 0, 14, 28, and 35 post-boosts. On day 21 post-boost, mice were i.n. anesthetized with Avertin (2, 2, 2-Tribromoethanol), and i.n. and challenged with 0.05 mL 10⁶ PFU RSV A2 diluted in PBS. Mice were monitored daily and euthanized on day 5 pi. Sera, BAL, lungs, and spleen were collected and stored on ice during organ processing for assays described below.

Sera were evaluated for anti-RSV IgG levels as described (51). Briefly, high-binding ELISA plates (Corning, Corning, NY) were coated with 5 µg/mL RSV A2 or B1 lysate overnight at 4°C. The next day, wells were washed 3x with KPL wash buffer (1x KPL in distilled water (diH₂O) (SeraCare, Milford, MA) and blocked with Blotto (5% non-fat dry milk) overnight at 4°C. Blotto was removed and sera (in 3-fold dilutions starting at 1:50) was diluted in Blotto and added to wells overnight at 4°C. Wells were washed 3x with KPL wash buffer and 2° goat-anti-mouse IgG-HRP (ThermoFisher, Waltham, MA), or secondary subtype IgG1 or IgG2a antibodies (Southern Biotech, Birmingham, AL) were added. Plates were incubated overnight at 4°C, washed 3x with KPL wash buffer, and developed with 1-Step™ Ultra 3,3’,5,5’-tetramethylbenzidine (TMB; ThermoFisher) for 20 min, and stopped with Stop Solution (ThermoFisher), then read immediately using a BioTek plate reader (BioTek, Winooski, VT) at OD₄₅₀.

To determine the level of RSV antibody neutralization in the mouse sera, a microneutralization assay was used as described with minor modifications (63). Briefly, sera were pooled and heat-inactivated at 55°C for 30 min. Diluted sera in 2% FBS/DMEM (1:40) were co-incubated with 200 FFU RSV A2-GFP +/- 10% guinea pig complement (C’) (Sigma-Aldrich, St. Louis, MO) for 1 h at 37°C. Following pre-incubation, the virus/sera mixture was added to 95% confluent A549 cells for 48 h. Fluorescent focus units (FFUs) were visualized using Cellomics ArrayScan (ThermoFisher), enumerated with HTS software, and mean FFUs of replicate wells were determined. Neutralization was determined as the percent reduction in mean FFUs compared to empty NP antisera.

A CX3C-CX3CR1 blocking assay was performed as described (24). Briefly, 500 nM RSV G ectodomain (G_ecto) was incubated +/- 5/mL heparin sulfate (HS) (Sigma) to prevent non-specific binding and +/- 20 µg/mL IgG (isolated from vaccinated mice by Protein G beads (Invitrogen) for 1 h on ice. CX3CR1.293 and HEK-293 cells were harvested, and 4 x 10⁶ cells/mL were blocked with 1 µg/mL Fc block diluted in FACS buffer (0.8% FBS/PBS) for 15 min on ice followed by incubation with 500 nM RSV G_ecto +/- 5 µg/mL HS +/- 10 µg/mL IgG for 1 h on ice. Cells were washed and resuspended in 20 µg/mL anti-G protein mAb (clone 130-5F) for 45 min on ice. Cells were washed and resuspended in goat-anti-mouse Alexa488 (1:200) (ThermoFisher) for 45 min on ice and protected from light. Cells were washed 3x with FACS buffer, resuspended in FACS buffer, and analyzed by flow cytometry. To determine FKN blocking, the assay was followed similarly except cells were incubated with 2 µg/mL biotinylated-FKN (Acro Biosystems, Newark, DE) +/- 5 µg/mL HS and +/- 10 µg/mL IgG. To detect CX3CR1-bound FKN, cells were incubated with Streptavidin-PE (1:200) (ThermoFisher). Identical times and temperatures were used for both ligands. Percent inhibition was determined as the difference of CX3CR1⁺ binding (G or FKN +HS + vehicle IgG) – (G or FKN +HS +NP IgG or mAb control bound to) x 100 as previously described (51). At least 20,000 events were collected using BD LSR II (BD Bioscience, Franklin Lakes, NJ).

Lungs were harvested at day 5 pi and homogenized in 1 mL DMEM using GentleMACS tissue homogenizer (Miltenyi Biotec, Gaithersburg, MD) as described (55). Homogenates were centrifuged at 500 xG at 4°C for 8 min, supernatant was 10-fold diluted in DMEM (Hyclone) and overlaid onto 90% confluent Vero E6 cells in 24-well plates. After 2h of absorption, cells were overlaid with 2% methylcellulose (Sigma Aldrich) and incubated at 37°C for 6 days. Following incubation, methylcellulose was aspirated, wells were washed with PBS, fixed with acetone: methanol (60:40, Sigma-Aldrich), and air-dried overnight. Wells were washed 3x with KPL wash buffer and blocked with blotto overnight at 4°C. The next day, Blotto was removed and a mAb cocktail against RSV F and G proteins (clones 131-2A, 131-2G) was diluted in blotto was added overnight at 4°C. Wells were washed 3x with KPL wash buffer and goat anti-mouse-AP (ThermoFisher) was added overnight at 4°C. Wells were washed 3x with KPL wash buffer and virus plaques were developed with 1-Step™ NBT/BCIP substrate solution (ThermoFisher) for 5 min, rinsed with diH2O, and enumerated using a dissection microscope.

Bronchioalveolar leucocytes (BAL) were collected by i.p. anesthetizing (Avertin) mice and terminally bleeding by severing the left axillary artery. The trachea was exposed and a small incision was made. The lungs were flushed 3x with 1 mL PBS and collected in 1.5 mL snap-cap tubes and BAL was centrifuged for 10 min at 500 xG at 4°C. The supernatant (BAL fluid) was separated and stored at -80°C until cytokine/chemokine analysis. BAL cells were resuspended in FACS buffer (0.8% FBS/PBS) and enumerated using a hemocytometer and Trypan blue. Cells were washed with FACS buffer and resuspended in Fc Block for 15 min on ice followed by the addition of anti-CD3, anti-CD8, and anti-CD11b, or isotype control Abs (all from BD Bioscience) for 1h on ice ( Supplementary Figure 2 ). Cells were washed, fixed with 2% PFA (Ted Pella, Redding, CA) for 20 min at room temperature, washed, and resuspended with FACS buffer. At least 10,000 events were collected with BD LSR II (BD).

Spleens from mice were collected at day 5 pi. Single-cell suspensions of spleen cells were made by dissociation through a 100 uM cell strainer (Corning), washed with Hanks Balanced Salt Solution (HBSS) (HyClone), and red blood cells were lysed with Gey’s solution (Sigma-Aldrich) for 5 min. Splenocytes were washed 2x with HBSS, resuspended in media containing 10% FBS + RPMI-1640, and enumerated using a hemocytometer, and 2 x 10⁷ cells/mL were plated in a round bottom 96-well plate (Corning). Splenocytes were stimulated with 10 µg/mL RSV G (₁₈₃WAICKRIPNKKPGKK₁₉₇) and M2 peptides (₈₂SYIGSINNI₉₀) (42) or control (GFP, aa 200-208), phorbol 12-myristate 13-acetate (PMA)/ionomycin (Sigma), or left unstimulated and were treated with GolgiPlug (Brefeldin A) (BD) to retain cytokines and incubated at 37°C for 6h. After 6 h, cells were washed 3x with FACS buffer, blocked with 1 µg/mL Fc block (BD), and stained with anti-CD3 and anti-CD4 or isotype controls (all from BD Bioscience) for 1 h on ice. Cells were fixed with 2% PFA for 20 min at room temperature, washed with permeabilization buffer (BD Bioscience), and incubated with anti-IFNγ and anti-IL-4 or isotype controls diluted in permeabilization buffer for 1 h at 4°C ( Supplementary Figure 2 ). Cells were washed 3X with permeabilization buffer, resuspended in FACS buffer, and analyzed with BD LSR II (BD Bioscience) with at least 10,000 events collected.

Data were analyzed by one-way ANOVA with Dunnett’s multiple comparison test. p<0.05 was considered significant. Data are represented as mean +/- SEM. A vaccination study was performed once. Experiments were performed at least in duplicate with representative data shown.

NP immunogens were constructed using SpyTag/SpyCatcher technology (54, 58) ( Figure 1 ). Briefly, a construct of lumazine synthase, which self-assembles into 60-mer spherical particles, was fused to a SpyCatcher domain. Recombinant lumazine synthase – SpyCatcher protein was purified and confirmed by negative stain electron microscopy to self-assemble into NPs ( Figure 1D ). To generate CCD-coated NPs, the lumazine synthase – SpyCatcher NPs were incubated with recombinant RSV G CCD protein fused to a SpyTag, allowing for the formation of the covalent isopeptide bond between the SpyCatcher and SpyTag and display of the CCD antigen on the surface of the NPs ( Figure 1B ). Covalent linking of the SpyTagged CCD to the lumazine synthase – SpyCatcher was verified by SDS-PAGE and a change in molecular weight of the bands ( Figure 1C ). We previously identified that a point mutation at site 177 (serine to glutamine) improved immunogenicity in a G glycoprotein vaccine compared to wild-type G protein adjuvanted with MPLA (51). Thus, in addition to wild-type CCD antigen loaded onto NPs (NP-WT), the S177Q CCD antigen was also generated and loaded onto NPs (NP-S177Q). Notably, upon overnight incubation of CCD antigens with NPs, precipitation was observed. Pelleting of the precipitate by centrifugation and evaluation by SDS-PAGE revealed that the precipitate is the NP-WT and NP-S177Q nanoparticle samples ( Figure 1C ). No precipitation is observed by incubation of empty NPs or CCD alone, suggesting that the loading of the CCD, which contains many hydrophobic amino acids, promoted insolubility of the NPs. To generate samples for immunization, pellets were resuspended in PBS.

Mice received a priming dose of 10 µg NP-WT, NP-S177Q, or empty NPs adjuvanted with 10 µg MPLA. On day 21 post-prime, mice were boosted with either 10 µg or 25 µg of homologous vaccine or 10µg empty NP, all adjuvanted with 10 µg MPLA. At day 7 post-boost, the mice were bled, and anti-RSV Abs were detected by ELISA ( Figure 2 , Supplementary Figure 1 ). NP-WT and NP-S177Q vaccination induced anti-RSV Ab responses. Abs generated by NP-S177Q were significantly increased (p <0.05), and NP-WT 25µg and NP-WT 10µg Abs were increased (p = 0.28, p = 0.06, respectively) compared to empty NP vaccination. NP-S177Q vaccination induced moderately higher serum IgG titers than NP-WT ( Figure 2 ), although the IgG responses did not statistically differ between vaccine doses. At day 21 post-boost, the NP-vaccinated mice were challenged with RSV A2, and on day 5 the serum Ab responses were determined. Similar to pre-challenge IgG titers, all vaccinated mice had greater anti-RSV A2 IgG compared to vehicle control ( Figure 3A ). Mice boosted with 25 µg of NP-S177Q vaccine had significantly (p<0.05) increased Ab titers compared to vehicle control, however, NP-S177Q vaccination was not statistically improved over NP-WT boosted mice. Contrary to our previous study demonstrating improved Ab recall responses (51), these data show a less robust recall response as sera Ab levels were roughly 1 log₃ lower in each group on days 7 post boost and 5 post-challenge. Previous constructs utilized full-length G protein as opposed to restricting antibody responses to the CCD, which may partially explain this phenomenon. It is also possible Abs were present in the lung during infection and would not be detected in sera. It is also notable that serum Ab titers against RSV B1 were markedly lower than RSV A2 ( Figure 3B ). This finding was similar to a previous report suggesting anti-G Abs generated against A2 G protein bind with lower affinity to RSV B compared to RSV A2, likely due to variable residues encompassing the CCD between subtypes, despite conservation of the CX3C motif (40). NP-S177Q vaccination induced greater Abs compared to vehicle control. NP-WT also induced anti-B1 Abs although the titers were lower compared to NP-S177Q.

To determine if the serum Ab response were Th1- or Th2-like, ELISAs were performed to determine the specific IgG subclass ( Figure 3C ). It is established that IgG2a corresponds to a Th1-type response, while IgG1 corresponds to a Th2-type in response (64) and determines Fc effector function (e.g., complement-dependent cytotoxicity) (65, 66). There were no significant changes in Th2-type Ab responses between vehicle and NP-WT or NP-S177Q vaccinated mice. Further, Th1-type responses were only significantly (p<0.05) increased in the 25 µg NP-WT vaccinated mice, while there were no significant IgG2a responses in NP-S177Q vaccinated mice. These findings do not recapitulate the increased Th1-type responses which were previously observed with G protein immunogen (51). Conformationally designed epitopes such as those in the NP vaccines may require adjuvants that do not denature or emulsify the antigens, and or the insolubility of NP-WT and NP-S177Q vaccines may have contributed to these differences (67).

Blocking CX3C-CX3CR1 interaction or ablating the CX3C motif is correlated with protection against RSV disease in mice and cotton rats (43, 49, 50, 68). To evaluate the efficacy of G protein CX3C-CX3CR1 blocking Abs generated in response to NP-WT or NP-S177Q vaccination, serum IgG from NP-vaccinated mice was isolated and tested. Similar to the G protein vaccinated mice (51), vaccination with NP-WT or NP-S177Q candidates induced significant (p<0.05) CX3C-CX3CR1 blocking Abs compared to vehicle IgG ( Figure 4A ), and Ab from NP-S177Q vaccination induced slightly higher blocking Abs (35%) than NP-WT vaccination (20%). As expected, mAb 131-2G which binds to a conserved epitope in the G protein blocked up to 90% G protein binding to CX3CR1. Contrary to our previous report showing that G protein induced greater CX3C-CX3CR1 blocking Abs compared to vaccination with an S177Q G protein mutant, in this study, we observed a slight improvement in G protein CX3C-CX3CR1 blocking, and in agreement with previous reports, 131-2G blocked G protein binding more effectively than polyclonal IgG from vaccinated mice. These Abs did not cross-react and block FKN binding to CX3CR1 (data not shown). This is not unexpected as there are structural differences that may preclude anti-G protein binding (69). Thus, this NP-S177Q vaccine platform induces G protein CX3C-CX3CR1 blocking Abs which have been shown to protect against RSV disease and are not implicated in modifying endogenous FKN signaling.

Anti-G protein Abs are neutralizing in human airway epithelial cells infected with RSV and in vivo (27). The addition of complement aids the neutralization of some anti-G protein Abs including the highly potent 3D3 and 3G12 anti-G protein mAbs that can be detected in immortalized cell lines (69, 70). To determine if serum from NP-WT or NP-S177Q vaccinated mice was neutralizing, heat-inactivated sera +/- 10% guinea pig complement (C’) were co-incubated with RSV-GFP (56) and added to RSV-infected human A549 cells for 48 h ( Figure 4B ). In the absence of complement, there was no significant (p>0.05) neutralization for any vaccine groups, however, serum plus complement from NP-WT and NP-S177Q vaccinated mice significantly (p<0.05) neutralized infected A549 cells compared to empty NP vaccination. Moreover, serum from 25 µg S177Q induced significantly (p<0.05) greater neutralization compared to NP-WT at the same dose. These data suggest neutralization is complement-dependent and not CX3C:CX3CR1-mediated neutralization.

Lung viral titers showed that NP-WT and NP-S177Q vaccination reduces lung titers in vivo ( Figure 5 ). On day 5 pi, corresponding to peak lung viral titers (71), 10 µg NP-WT or 25 µg NP-S177Q vaccination resulted in significantly (p<0.05) reduced viral titers in the lungs of RSV A2 challenged mice. 25 µg NP-WT vaccinated mice and 10 µg NP-S177Q vaccinated mice also reduced lung titers compared to empty NP vaccination (p=0.16, p=0.15, respectively). These findings are consistent with other G protein vaccines that reduce lung titers and induce anti-G protein neutralizing Abs (nAbs), however, others have reported that Abs to G protein are non-neutralizing but this was determined in the absence of complement, an effect which has caused misunderstanding (40, 41, 72, 73). There may be mechanisms aside from nAbs that result in reduced viral titers after NP vaccination including a cytotoxic T lymphocyte (CTL) response or improved macrophage activity, however these were not examined here. Moreover, while in vivo and in vitro neutralization data suggest 25 µg NP-S177Q vaccination resulted in the greatest levels of nAbs and reduced lung titers, the lack of significant in vivo reduction for 25 µg NP-WT and 10 µg NP-S177Q does not correlate with our findings in vitro. The findings from this study show that G protein immunogens are capable of inducing nAbs that are detectable in vitro with additional complement and vaccination may reduce lung viral titers in mice.

Aspects of RSV disease are connected with the expression of the G protein CX3C motif (50). Blocking G protein CX3C-CX3CR1 interaction with mAbs specific to this motif or the CCD domain is correlated with reduced RSV disease in vivo (40, 41, 73). To determine if NP-WT or NP-S177Q vaccination is protective against G protein-mediated disease, vaccinated mice were challenged with RSV A2 and BAL leukocytes were evaluated ( Table 2 ). A significant (p<0.05) reduction in BAL cell numbers (3.9 x 10⁴ cells) in RSV-challenged mice that were vaccinated with 10 µg of NP-S177Q vaccine was evident compared to the empty NP control vaccinated mice (7.5 x 10⁴ cells). Interestingly, no other vaccination group including mice vaccinated with 25 µg of NP-S177Q vaccine had substantially reduced BAL cells following RSV challenge. Consistent with an overall reduction in BAL cell infiltration, RSV-challenged 10 µg NP-S177Q mice vaccinated had reduced CD11b+ cell numbers (2.3 x 10³) and a trend toward lower in CD8+ T cell numbers (2.0 x 10³) compared to RSV-challenged empty NP vaccinated mice (5.4 x 10³ and 3.3 x 10³ cells, respectively). Taken together, these support lung disease protection in mice vaccinated with NP-S177Q vaccine compared to vehicle control vaccinated mice. We also examined intracellular cytokine production by splenocytes stimulated with RSV G peptide encompassing the CCD and M2 as previously described (42, 68), however, there were no statistical differences detected in the production of IFNγ+ or IL-4+ by CD3+/CD4+ T cells between groups (data not shown).