Female, 6-8 week old BALB/c mice obtained from The Jackson Laboratory (Bar Harbor, ME) and housed under specific pathogen-free conditions with food and water provided ad libitum. All experiments described herein were approved by and performed according to the Icahn School of Medicine at Mount Sinai Institutional Animal Care and Use Committee. The following reagent was obtained through BEI Resources, NIAID, NIH: Fluzone® Influenza Virus Vaccine, 2005-2006 Formula, NR-10480. Mice were vaccinated intramuscularly in the hind legs with a seasonal trivalent inactivated influenza virus vaccine (TIV; Fluzone® Influenza Virus Vaccine) containing an influenza A H1N1 component (A/New Caledonia/20/1999/IVR-116), influenza A H3N2 component (A/New York/55/2004/X-157 [an A/California/7/2004-like strain]), and influenza B component (B/Jiangsu/10/2003 [a B/Shanghai/361/2002-like strain]).

For the first study assessing the impact of adjuvant on lung immune cell dynamics, TIV was adjuvanted with one of the following: no adjuvant, 2% alhydrogel (alum, Invivogen), or IMDQ-PEG-CHOL (abbreviated to IMDQ in this paper). IMDQ is a lymph node-targeting amphiphilic conjugate of an imidazoquinoline TLR7/8 agonist linked to poly(ethylene glycol) and cholesterol (77–79). A control group receiving alum adjuvant alone was included for comparison to the other TIV groups. An OVA-sensitized positive control group for lung eosinophilia and a PBS control were included in the study as well. 5 mice were used per treatment group.

For the second study assessing the impact of vaccine-mismatched heterosubtypic sublethal challenge on lung myeloid cell dynamics, TIV was adjuvanted with alum. A group receiving alum alone was included as a negative control. 3-4 mice were used per treatment group.

TIV and adjuvant mixtures were administered intramuscularly to both hind legs (50 µL/leg, 100 µL/mouse total).

Mice were anesthetized using an intraperitoneal (i.p.) injection of a ketamine and xylazine mixture, then challenged intranasally (i.n.) with 50 µL of mouse-adapted egg-grown influenza viruses or egg allantoic fluid as vehicle control. Body weights were monitored for 7-10 days post-infection to assess morbidity. 100% body weight was defined as the weight of the mouse immediately before the challenge. The viruses used were H1N1 A/New Caledonia/20/1999 (NC99) and H3N2 A/X-31 at a sublethal (0.2 LD₅₀) challenge dose.

Mice received a 100 µL i.p. injection containing 20 µg of Imject™ ovalbumin (OVA, Thermo Scientific) adsorbed to alum on days 0 and 7. On day 14, mice were anesthetized as described above and then challenged i.n. with 20 µg of Imject™ ovalbumin in PBS, diluted to a final volume of 50 µL/mouse.

Blood was collected via the submandibular route and then allowed to coagulate at 4°C overnight. Coagulated blood was then spun down at 400 x g for 5 minutes at 4°C and serum was collected and stored at -20°C until further analyses were performed.

Lung lobes were collected in 3 mL of RPMI-1640 media supplemented with 10% heat-inactivated fetal bovine serum (FBS) and 1X Penicillin-Streptomycin. Lung tissue was minced in 6-well plates using surgical scissors and then digested with Collagenase D at a working concentration of 2 mg/mL for 15 minutes at 37°C while shaking. After digestion, single-cell suspensions were generated by forcing digested tissue through a 70 µm cell strainer with the flat end of a 1 mL syringe plunger in a new 6-well plate. Wells were washed with an additional 3 mL of media then transferred to a 15 mL conical vial and centrifuged at 400 x g for 5 minutes at 4°C. Supernatants were discarded then pellets were resuspended in 5 mL of 1X ammonium chloride red blood cell lysis buffer and incubated at room temperature for 5 minutes. Cells were centrifuged and pellets were resuspended in 5 mL of 1X PBS to wash. After centrifugation, cell pellets were resuspended in 50 µL of Purified Rat Anti-Mouse CD16/CD32 Fc Block (clone 2.4G2, BD) diluted 1:100 in Staining Buffer (1% bovine serum albumin and 2 mM EDTA in 1X PBS), transferred to a 96-well V-bottom plate, and incubated at room temperature for 5 minutes. After incubation, 50 µL of surface stain antibody cocktail comprised of the following antibodies and dyes were added to cells: CD11c FITC (1:150, clone HL3, BD), CD125 PE (1:150, clone T21, BD), Siglec-F PE-CF594 (1:150, clone E50-2440, BD), Ly6G PerCP-Cy5.5 (1:150, clone 1A8, BD), CD101 PE-Cy7 (1:150, clone Moushi101, Invitrogen), CD11b APC (1:150, clone M1/70, BioLegend), MHC II Alexa Fluor 700 (1:150, clone M5/114.15.2, Invitrogen), CD62L APC-Cy7 (1:150, clone MEL-14, BD), Fixable Viability Dye eFluor 450 (1:200, eBioscience) for the eosinophil phenotyping study; or CD11b APC (1:100, clone M1/70, BioLegend), CD11c FITC (1:100, clone HL3, BD), MHC II eFluor™ 450 (1:200, clone M5/114.15.2, Invitrogen), Siglec-F (1:100, clone E50-2440, BD), CD64 PE (1:100, clone X54-5/7.1, BD), BD Horizon™ Fixable Viability Stain 700 (1:200, BD) for the vaccine-mismatched challenge study. After incubating in the dark for 20 minutes at room temperature, cells were washed by adding 120 µL of Staining Buffer on top of the cells, followed by centrifugation, discarding supernatants, resuspension of pellets in 200 µL Staining Buffer, and centrifugation. After discarding supernatants from the last wash, pellets were resuspended in 200 µL of Staining Buffer and 5 µL of CountBright Absolute Counting Beads (ThermoFisher) were added to all samples, excluding single-stained controls and the unstained lung cell control, to facilitate quantification of absolute cell numbers. UltraComp eBeads Plus Compensation Beads (ThermoFisher) were used to create single-stained compensation controls. Samples were acquired using a Beckman Coulter Gallios flow cytometer with Kaluza software. Data analysis was performed using FlowJo 10.8.1 (Treestar) and compensated using the built-in AutoSpill algorithm. Data were visualized using Graphpad Prism version 9.4.1. Unsupervised clustering analyses were conducted in the R computing language (ver. 4.05) using the CATALYST (Cytometry dATa anALYSis Tools) package version 1.20.1 in RStudio (ver. 1.4.1106) (80).

Lung left lobes were collected in 500 µL of 1X PBS in prefilled homogenizer bead tubes containing 3.0 mm high impact zirconium beads (Benchmark Scientific), snap-frozen on dry ice on the day of harvest, then stored at -80°C. On the day of the assay, samples were homogenized and then centrifuged at 10,000 x g for 5 minutes at 4°C after thawing. Clarified supernatants were serially diluted 10-fold, beginning at 1:10 for a total of 6 dilutions in 1X PBS. In a 12-well tissue culture plate seeded the day prior with Madin-Darby Canine Kidney (MDCK) cells, wells were washed twice with 1 mL of 1X PBS per well. After washing, 150 µL of each clarified lung supernatant dilution was added to one well and then incubated for 1 hr at room temperature, rocking plates every 10 minutes. After incubation, cells were washed once with 1 mL/well of 1X PBS then 1 mL/well of overlay (2% Oxoid agar in sodium bicarbonate buffered serum-free 2X minimum essential medium (MEM) supplemented with 1% diethylaminoethyl (DEAE)-dextran and 1 μg/mL tosylamide-2-phenylethyl chloromethyl ketone (TPCK)-treated trypsin) was added to each well. After allowing the overlay to solidify at room temperature for 15 minutes, plates were incubated at 37°C for 48 h and fixed with 1 mL of 4% formalin per well overnight at 4°C. After removal of the fixative and agar overlay, plates were washed with PBS-T (PBS with 0.05% Tween 20). Monolayers were immunostained with 150 μL/well hyperimmune rabbit polyclonal serum diluted in 5% milk in PBS-T overnight at 4°C on a plate rocker. Plates were washed 3 times with PBS-T then 150 μL/well of horseradish peroxidase (HRP)-conjugated anti-rabbit IgG Fc secondary antibody diluted in 5% milk in PBS-T for 1 hr at room temperature while rocking. Plates were washed, then 150 μL/well KPL TrueBlue was added and incubated at room temperature for 30 minutes on a plate rocker. Plates were washed and plaques were counted to determine lung viral titers.

Lung left lobes were collected, homogenized, and clarified as described above. We used the Cytokine & Chemokine 26-Plex Mouse ProcartaPlex™ Panel 1 (ThermoFisher) to measure cytokine and chemokine concentrations in the lungs. The assay was conducted according to the manufacturer’s instructions and the plate was placed on an orbital shaker set to 300 rpm for all incubation steps. Briefly, 25 µL of lung homogenate was mixed with 25 µL of Universal Assay Buffer provided by the kit, then combined with beads in an optical bottom black 96-well plate and incubated for 30 minutes at room temperature protected from light. After 30 minutes, the plate was moved to 4°C for overnight incubation. The next day, the plate was equilibrated to room temperature for 30 minutes, then washed 3 times with 150 µL/well of 1X Wash Buffer diluted according to kit instructions. Following washing, 25 µL/well of 1X Detection Antibody mixture was added and incubated at room temperature for 30 minutes. The plate was washed 3 times then 50 µL/well of 1X Streptavidin-PE solution was added and incubated for 30 minutes at room temperature. After washing the plate 3 times, 120 µL/well of Reading Buffer was added and the plate was incubated for 5 minutes at room temperature. Subsequently, data were acquired on a Luminex 100/200 analyzer (Millipore) with xPONENT software (version 4.3). Data visualization and analysis were conducted using GraphPad Prism (version 9.4.1) and R computing language (ver. 4.05) in RStudio (ver. 1.4.1106).

To measure vaccine-specific total IgG, Nunc MaxiSorp 96-well plates (ThermoFisher) were coated with 100 µL/well TIV diluted 1:100 in carbonate-bicarbonate buffer and incubated overnight at 4°C. Plates were washed 3 times with PBS-T then blocked with 200 µL/well of blocking buffer (5% milk in PBS-T) and incubated for 1 h at room temperature. During the blocking step, sera were serially diluted 4-fold a total of 7 times starting from a 1:100 dilution for total IgG. After blocking, plates were washed 3 times then 100 µL/well of diluted sera was added to plates and incubated for 1 h at room temperature. Plates were washed then 100 µL/well of goat anti-mouse IgG-HRP (Cat. No. ab6823, Abcam) diluted 1:5000 in blocking buffer was added to each well and incubated for 1 h at room temperature. After washing plates 3 times, 100 µL/well of 1-Step Turbo TMB Substrate (ThermoFisher) was added and then incubated for 20 minutes at room temperature. The reaction was stopped by adding 100 µL/well of ELISA Stop Solution (Invitrogen) and plates were read at 450 nm and 650 nm. Background subtracted optical density values (OD_450-650nm) were used for downstream analyses.

To measure total IgE, Nunc MaxiSorp 96-well plates (ThermoFisher) were coated with 100 µL/well of anti-mouse IgE antibody (Clone R35-72, BD) diluted to a concentration of 2 µg/mL in carbonate-bicarbonate buffer and incubated overnight at 4°C. Similar to the IgG ELISA described above, plates were washed the next day 3 times with PBS-T and then blocked with 200 µL/well of blocking buffer for 1 h at room temperature. As plates were incubated with blocking buffer, sera were diluted 4-fold a total of 7 times, starting from a 1:50 dilution. Unlabelled purified mouse IgE (SouthernBiotech) diluted 4-fold a total of 11 times from a starting concentration of 400 ng/mL to generate a standard curve. Plates were washed after blocking then 50 µL/well of diluted sera was added to plates and incubated for 1 h at room temperature. After incubation with sera, plates were washed and then incubated with 100 µL/well of goat anti-mouse IgE-HRP (SouthernBiotech) for 1 h at room temperature. Plates were then washed, developed, and read as described for the IgG ELISA above after incubation. Subtracted optical density (OD) values of the 1:50 dilution were used to determine serum concentrations of total IgE.

All data were analyzed using GraphPad Prism (version 9.4.1).

6 µL of serum was pretreated with 18 µL of receptor destroying enzyme II (Seiken) and incubated at 37°C overnight. After incubation, 18 µL of 2.5% sodium citrate solution was added and incubated at 56°C for 30 minutes followed by an addition of 18 µL of 1X PBS. Pretreated sera were stored at 4°C until use in the assay. Hemagglutination (HA) units were calculated via HA assay by combining a serial 2-fold dilution of stock virus in PBS with an equal volume of 0.5% turkey red blood cells (RBC) and incubated at 4C for 45 minutes. For the HA inhibition (HAI) assay, 50 µL of pretreated sera were serially diluted 2-fold a total of 11 times in PBS. Stock virus was diluted to a concentration of 8 HA units per 50 µL, then 25 µL/well of diluted virus was added to all serum-containing wells. Plates were incubated at room temperature for 30 minutes, then 50 µL/well of 0.5% turkey RBC was added to all wells. After incubation at 4°C for 45 minutes, HAI titers were recorded.

The following metrics from each individual mouse were integrated into one data frame for principal component analysis (PCA): flow cytometry absolute numbers for 18 cell populations of interest; body weight percentages from 7 timepoints; 7 DPC lung viral titers; IgG and HAI titers from 14 days post-vaccination and post-challenge sera; total IgE concentrations post-vaccination and post-challenge sera; concentrations of 26 cytokines/chemokines in clarified lung homogenate supernatants. Data visualization and analysis were conducted using prcomp in the R computing language (ver. 4.05) in RStudio (ver. 1.4.1106).

We used the 2005-2006 Fluzone trivalent inactivated influenza vaccine (TIV) as our model vaccine. Given that the previous study by Choi et al. was conducted using TIV adjuvanted with alum (TIV+alum), a known Th2-skewing adjuvant, we included the following groups in the study to disentangle any potential adjuvant effects: TIV with no adjuvant, TIV adjuvanted with alum to match the original study (TIV+alum), TIV adjuvanted with a lymph node targeting amphiphilic conjugate of the imidazoquinoline TLR7/8 agonist connected to poly(ethylene glycol) and cholesterol to potently elicit Th1-skewed responses (TIV+IMDQ), and alum alone with no TIV to assess if any effects previously observed were inherently due to alum itself rather than the vaccine ( Figure 1A ) (75, 77, 78). Ovalbumin (OVA)-sensitized allergic mice were included positive control for lung eosinophilia, alongside a PBS negative control ( Figure 1A ). Three weeks following intramuscular (i.m.) vaccination with the described regiments, mice were challenged with a sublethal (0.2 LD₅₀), vaccine-matched challenge with A/New Caledonia/20/1999 (NC99) or mock-challenged with egg allantoic fluid since all viruses in the study were propagated in embryonated chicken eggs. At 7 days post-challenge (DPC) for the vaccination groups and 1 DPC for the control groups, we quantified key lung immune cell populations of interest: eosinophils, neutrophils, and alveolar macrophages ( Figure S1 ).

To complement findings from our manual gating analyses, we then used unsupervised clustering of our flow cytometry data using CATALYST to provide an unbiased insight into the phenotypic dynamics of the lung across all treatment groups and cell types.

Within Siglec-F⁺ cells, a large majority of the cells were AMs (72.34%), followed by eosinophils (25.17%) and neutrophils (2.49%) ( Figures 2A , S4 ). A small subset (3.77% of Siglec-F⁺) of AMs also expressed CD11b in addition to CD11c, which might suggest infiltrating monocyte origin (83). Neutrophils were predominantly negative for CD62L expression, however a minor population expressed CD62L (0.32% of Siglec-F⁺). We noticed that there was heterogeneity within the eosinophil subset, clustering on the basis of MHC II expression levels and CD62L. A majority of eosinophils were CD62L⁺, likely corresponding to a more rEos-like phenotype while the CD62L^- eosinophils may correspond to iEos. Both CD62L⁺ and CD62L^- eosinophils were further bifurcated on the basis of MHC II expression, with low MHC II levels on a majority of cells in either subset. In line with previous manual gating results, a greater proportion of eosinophils were observed in TIV-vaccinated NC99-challenged mice compared to the corresponding mock-challenged control groups. AM depletion was also readily observed in unsupervised clustering results in the alum-only, NC99-challenged group and to a lesser extent in OVA-challenged mice.

We then used the unsupervised clustering analyses on the total eosinophil population (Ly6G^- CD11b⁺ CD11c^- CD125^int Siglec-F⁺) to dissect differences within our population of interest at a higher resolution ( Figures 2B , S5 ). In the absence of other cell types, differential CD101 expression levels could be demarcated more clearly for binning of cells based on iEos and rEos phenotypic designations. Confirming results from manual gating analyses, a large majority of eosinophils were rEos with intermediate expression levels of Siglec-F (73.73%). Approximately a fifth (21.19%) of rEos had elevated expression of MHC II. iEos also expressed predominantly low levels of MHC II, although 28.55% of iEos were MHC II^hi. 5.02% of eosinophils were double negative for both CD62L and CD101 expression. There was an enrichment in iEos in the TIV-vaccinated then NC99-challenged mice and OVA-sensitized mice, similar to observations for manual gating. Interestingly, the proportion of MHC II^hi versus MHC II^low within each eosinophil subset appeared consistent across treatment groups and inflammatory conditions.

Given the critical role of environmental cytokines in shaping cellular dynamics within tissues during inflammatory responses, we next characterized 26 chemokines and cytokines using a bead-based assay from lung homogenate supernatants collected at 7 DPC for the vaccinated and challenged mice and 1 DPC for the OVA-sensitized and PBS controls.

From our analyses, no strong Type 1, 2, or 3 inflammatory signal was observed at 7 DPC in the lungs of mice that received both TIV-vaccination and NC99 challenge ( Figure 3 ). In contrast, alum-only mice had potent induction of acute inflammation-associated cytokines (IL-1β, IL-6, and IL-18) alongside a clear and distinct Th1 module (IFN-γ, TNF-α). This is in line with expectations given that this is a primary infection for the alum-only mice since they were not primed with any influenza antigens. OVA-sensitized mice exhibited both a pro-inflammatory module (IL-1β, IL-6, TNF-α), albeit to a lesser extent compared to the alum group, alongside a Th2 cytokine module (IL-4, IL-5, IL-13, IL-9, IL-27, CCL11) at 1 DPC. IL-22 concentrations were modestly elevated in all groups receiving intranasal NC99 challenge ( Figure S6 ). Interestingly, GM-CSF concentrations were highest in the alum-only NC99-challenged group as well as in the OVA-sensitized mice, but were undetectable in the majority of the vaccinated mice barring one mouse in the TIV-vaccinated group despite observing an influx of eosinophils into the lung ( Figure S6 ). Concentrations of the chemokines CCL2, CCL3, CCL4, CCL5, CCL7, CXCL1, and CXCL10 were significantly higher in the alum-only NC99-challenged mice compared to all other groups ( Figure S6 ). OVA-sensitized mice exhibited higher concentrations of CCL2, CCL3, CCL7, and CXCL2 ( Figure S6 ). CCL5 and CCL7 concentrations in the lungs of TIV-vaccinated and challenged mice were both significantly elevated above mock-challenged counterparts, other TIV-vaccinated groups, and the PBS control group but below that of the alum-only NC99-challenged group ( Figure S6 ). Of note, only the alum-only NC99-challenged group had significantly elevated concentrations of IL-10 out of all groups in the study ( Figures 3 , S6 ).

We also monitored weight loss as a measure of morbidity and mortality up to 7 days post-challenge. No mortality was observed, in line with the sublethal nature of the influenza challenge. Additionally, NC99-challenged mice lost minimal body weight irrespective of vaccination status, indicating little morbidity in this sublethal challenge model ( Figure 4A ). Minimal morbidity (<5% body weight loss) was also observed at 1 DPC for the control mice following OVA sensitization compared to un-challenged PBS controls ( Figure 4B ). Furthermore, all groups that received TIV vaccination did not have detectable lung virus titers, barring the alum-only group, indicating that all vaccinated animals were able to control viral replication by 7 days post-challenge ( Figure 4C ).

Using hemagglutination inhibition (HAI) titers as a readout of serum neutralization capacity for NC99, we demonstrated that all TIV-vaccinated mice generated modest HAI titers, while alum-only mice had predominantly undetectable HAI titers consistent with their antigen exposure history ( Figure 4D ). The low titers are in line with the suboptimal dose of vaccine administered, which is enough to induce seroconversion and vaccine-specific antibodies, but not enough to confer completely sterilizing immunity in the lungs (75, 76). Although HAI activity was low to modest for TIV-vaccinated mice, all mice that were vaccinated mounted robust vaccine-specific IgG responses ( Figure 4E ). Here we observe a clear adjuvant effect, where mice vaccinated with TIV+IMDQ exhibit significantly higher endpoint titers of TIV-binding IgG compared to mice that received TIV+Alum or TIV alone. The inclusion of alum as an adjuvant did not appear to elicit significantly higher IgG titers compared to TIV alone. No detectable TIV-specific IgG titers were measured in the alum-only groups, as expected. Given the eosinophil influx in the lungs in the absence of a strong Th2 cytokine module, we wanted to investigate if an allergy-like response was elicited in the serum by measuring total IgE ( Figure 4F ). All vaccinated mice generated a baseline level of total IgE by 2 weeks post-vaccination, which increased in concentration by 7 DPC. There appears to be no adjuvant effect on total IgE induction after TIV-vaccination. In line with previous literature, OVA-sensitized mice generated the highest titers of IgE compared to all other groups.

Furthermore, we used serum titers of IgG1 and IgG2a to indirectly discern host Th2 and Th1 skewing, respectively ( Figure S7 ) (84–86). We observed robust induction of vaccine-specific IgG1 and IgG2a, in line with findings from total IgG analysis. TIV+Alum vaccination elicited the highest titers of IgG1, both at 2 weeks post-prime and 7 DPC, while TIV+IMDQ vaccination elicited the highest titers of IgG2a at both time points ( Figures S7A, B ). Unadjuvanted TIV-vaccinated mice had comparable IgG1 titers to the TIV+IMDQ vaccinated mice, and IgG2a titers that were significantly elevated above that of the TIV+Alum-vaccinated mice but substantially lower than that of the TIV+IMDQ-vaccinated mice. Using the IgG2a/IgG1 ratio as a metric to assess Th1/Th2 skewing, we observed that mice vaccinated with TIV+IMDQ had the highest ratio, TIV-only mice had the second highest ratio, followed by TIV+Alum mice ( Figure S7C ). This suggests that TIV vaccination in the absence of any adjuvant elicits a balanced Th1/Th2 response, while TIV+IMDQ skews towards Th1 and TIV+Alum skews towards Th2 (78, 79).

We then integrated individual-level data from all mice using principal component analysis (PCA) of 58 total parameters from flow cytometry, body weights, lung viral titers, TIV-specific IgG titers, HAI titers, total IgE concentrations, and cytokine/chemokine concentrations from lung homogenate supernatants ( Figure 5 ). Distinct vaccination and challenge response profiles were observed after dimensionality reduction. Alum-vaccinated, NC99-challenged mice cluster further and separately from TIV-vaccinated, OVA-sensitized, and PBS control mice, driven by factors such as lung viral titers, IL-10, IL-18, IFN-γ, and loss of AMs ( Figure S8 ). OVA-sensitized mice also clustered together away from the TIV-vaccinated mice, on the basis of Th2 cytokine concentrations and eosinophil absolute numbers. TIV-vaccinated, NC99-challenged mice clustered closer to the OVA-sensitized mice. TIV+Alum- and TIV+IMDQ-vaccinated mice clustered close to all mock-challenged groups and the PBS negative control.

We next expanded on the previously described breakthrough infection model by testing if pulmonary eosinophilia was also induced upon virus challenge with a vaccine-mismatched H3N2 influenza virus ( Figure 6A ). As an alternative to intramuscular TIV as the antigenic priming, we also tested if prior challenge with NC99, similar to TIV vaccination, could result in pulmonary eosinophilia upon rechallenge with a heterosubtypic H3N2 virus. After priming mice with TIV+Alum or alum alone, mice were sublethally (0.2 LD₅₀) challenged with either the vaccine-matched NC99 or mock-challenged with egg allantoic fluid at 3 weeks post-vaccination. 4 weeks following homologous challenge, mice were then challenged with a sublethal dosage (0.2 LD₅₀) of H3N2 A/X-31 (X31), which is a laboratory-adapted reassortant virus containing the HA and NA gene segments from H3N2 A/Hong Kong/1/1968 in the backbone of H1N1 (A/Puerto Rico/8/1934) and is mismatched from the TIV we use in our studies (87). Contrary to more recent H3N2 influenza viruses, this virus is able to efficiently infect mice and thereby cause severe morbidity for the chosen virus dose. At 7 DPC following heterosubtypic, vaccine-mismatched X31 challenge, we used flow cytometry to discern myeloid populations of interest ( Figure S9 ).

Absolute numbers of eosinophils were significantly elevated in mice that were vaccinated and challenged, irrespective of challenge virus ( Figure 6B ). This elevation was not observed in alum-only mice even after a second intranasal influenza challenge with X31 after priming with NC99, suggesting that the nature of the primary antigenic exposure (intramuscular priming versus intranasal challenge) may be linked to distal influx of eosinophils in the lungs upon subsequent challenge. AM absolute numbers were highest in alum-vaccinated, NC99- and X31-challenged mice, at 1.8- to 2.6-fold higher than other experimental groups. It is likely that protection from vaccine alone or vaccine and NC99 challenge did not confer a sufficient amount of protection against loss of AMs ( Figure 6C ). Dendritic cells (DCs) were defined as Siglec-F^- MHC II^hi CD11c⁺ CD64⁺ cells, likely of monocyte origin (88). DCs were elevated in animals that were mock-challenged then X31-challenged, irrespective of initial vaccination regimen ( Figure 6D ). An influx of CD11b⁺ macrophages was only observed in TIV+Alum-vaccinated mice challenged with X31 7 weeks after initial vaccination and 4 weeks after mock intranasal challenge with egg allantoic fluid, and no other group ( Figure 6E ). These cells were Siglec-F⁺ CD11c⁺ similar to canonical AMs, but also exhibited surface expression of CD11b potentially indicating monocytic origin.