The study and all animal procedures were approved under ethical permits 1666/19, 38/23, and 3307/20 granted by the Gothenburg Regional Animal ethics committee.

C57BL/6 mice, all female, 6-8 weeks old, were purchased from Janvier Labs, France. B6. SJL-Ptprc^a Pepc^b/BoyJ (CD45.1) mice were maintained in-house. Mice were sacrificed at 0, 3, 7, 10, and 14 days post-infection or immunization. Blood was collected and dispensed into serum separation tubes (Microtube 1.1 mL Z-Gel, Sarstedt). Collected tissues were transferred into separate tubes containing PBS + 1% BSA for processing.

Peptides (0.5nM; NP 45: ₂₆₄LILRGSVAHKSCLPACV₂₈₀ ; NP 47: ₂₇₆LPACVYGPAVASGYDFE₂₉₂ ; NP 52: ₃₀₆LLQTSQVYSLIRPNENP₃₂₂ ; HA 13: ₇₃NIAGWLLGNPECDPLLP₈₉ ; HA 14 ₇₉LGNPECDPLLPVRSWSY₉₅ ; HA 15 ₈₅DPLLPVRSWSYIVETPN₁₀₁ ; HA 16 ₉₁RSWSYIVETPNSENGIC₁₀₇ ; HA 17 ₉₇VETPNSENGICYPGDFI₁₁₃ ; HA 18 ₁₀₃ENGICYPGDFIDYEELR₁₁₉) were mixed with LPS (0.6μg/mL) and Addavax (Invivogen). Mice were immunized subcutaneously in the footpad hock with 30μL/mouse. Depending on the experiments, three to six mice per group were infected, and their organs were pooled when necessary, for analysis. A measurement of 5 μg of each recombinant protein in PBS, mixed 1:1 with Addavax was injected i.p. in a volume of 140μl per mouse.

Influenza A/Puerto Rico/8/1934 H1N1 (PR8) was used to infect C57BL/6 mice at a dose of 500 TCID₅₀ per mouse in Hanks’ Balanced Salt Solution (HBSS) + 0.1% fetal calf serum (BSA) intranasally in a volume of 25μL/mouse.

C57BL/6 mice were primed with the immunodominant peptides (NP 45, 47, and 52, and OVA p323 control) as described under “immunization”. pLNs and spleens from five mice were pooled, and a single-cell suspension was prepared. CD4 T cells were purified with an EasySep Mouse CD4⁺ T Cell Isolation Kit (Stem Cell Technologies) according to the manufacturer´s instructions. CD4 T cells (2 x 10⁶ per mouse) were transferred i.v. and recipient mice were infected after 16 hours.

Mice were sacrificed, medLNs, pLNs, and spleens were harvested, and organs were processed. Red blood cells were lysed using ACK lysing buffer. For B cell characterization, 25μL of pre-mixed extracellular antibody mixture was added to each tube and incubated for 30 minutes at 4°C. The following antibodies were used for extracellular B cell staining: 1:700 anti-mouse CD3 BV510 (BD, catalog: 563024), 1:700 anti-mouse NK1.1 BV510 (BioLegend, catalog: 108737), 1:200 anti-mouse IgM FITC (BioLegend, catalog: 406506), 1:200 anti-mouse IgD Pacific Blue (BioLegend, catalog: 405712), 1:200 anti-mouse CD38 APC-Cy7 (BioLegend, catalog: 102728), 1:200 anti-mouse GL7 Pe-Cy7 (BioLegend, catalog: 144620), and 1:200 anti-mouse B220 BV786 (BioLegend, catalog: 103246). After a wash with 1mL of FACS-buffer (DPBS + 2% FBS + 0.5mM EDTA), the supernatant was discarded and 100μL of premixed HA and NP proteins at 6.6 nM were added to each tube for 1 hour at 4°C. Samples were washed and incubated for 30 minutes at 4°C with 50μL of 1:100 streptavidin-APC (Invitrogen, catalog: 17-4317-82). Live/Dead Aqua (Invitrogen, catalog: L34966) staining was performed, and samples were fixed with paraformaldehyde (1.5%). Samples were washed and resuspended in 200μl of FACS buffer and stored at 4°C until the day of analysis.

For T cell characterization, 50μL of premixed tetramer mixture (I-Ab NP52_306-322 APC-LLQTSQVYSLIRPNENP, I-Ab HA16_91-107 BV412-RSWSYIVETPNSENGIC, I-Ab human CLIP_87-101 APC/BV421 - PVSKMRMATPPLMQA) was added to each tube and incubated for 3 hours at 4°C as the first staining step when indicated. After 2h 30 min, 25μL of premixed T cell-extracellular antibody mixture was added to each tube and incubated for 30 minutes at 4°C. The following antibodies were used for extracellular T-cell characterization: 1:200 anti-mouse CD3 Per-CP-Cy5.5 (BioLegend, catalog: 100218), 1:100 anti-mouse CD4 APC-Cy7 (BioLegend, catalog: 100526), 1:100 anti-mouse CD69 BV605 (BioLegend, catalog: 104529), 1:100 anti-mouse CD44 BV786 (BD, catalog: 563736), 1:100 anti-mouse CD19 BUV395 (BD, catalog: 563557), 1:100 anti-mouse NK1.1 BV510 (BioLegend, catalog: 108737), and 1:100 anti-mouse F4/80 BV510 (BioLegend, catalog: 123135).

The following antibodies were used for Tfh vs Tfr characterization: 1:100 anti-mouse B220 BUV395 (BD, catalog: 563793), 1:100 anti-mouse PD-1 BV711 (BioLegend, catalog: 135231), 1:100 anti-mouse CXCR5 BV650 (BD, catalog: 563981), 1:100 anti-mouse CD4 APC-Cy7 (BioLegend, catalog: 100526), 1:25 anti-mouse FOXp3 AlexaFluor 647 (BioLegend, catalog: 320014), and 1:100 anti-mouse CD45.1 PE (BioLegend, catalog: 553776). Permeabilization and intracellular staining were performed after the Live/Dead stain. In detail, 250μL of FOXP3 intracellular staining buffer (Invitrogen, catalog: 00-5523-00) was added for 1 hour at 37°C. The samples were then washed with a 2mL permeabilization buffer. A measurement of 25μL of 1:500 rat anti-mouse CD16/CD32 Fc block (BD, catalog: 553142) was added for 5 minutes at RT, followed by the addition of 25μL pre-mixed intracellular antibody for 45 minutes at RT. For intracellular staining, the following Abs were used: 1:100 anti-mouse FOXP3 PE (Invitrogen, catalog: 12577382), 1:50 anti-mouse Bcl6 AlexaFluor488 (BD, catalog: 561524), 1:50 anti-mouse Tbet Pe-Cy7 (Invitrogen, catalog: 25582582), and 1:50 anti-mouse RORγT Pe-CF-594 (BD, catalog: 562684). Following incubation, the samples were washed with perm buffer and resuspended in 200μL of FACS buffer per tube. The tubes were stored at 4°C until the day of analysis. Samples were analyzed using a BD LSR Fortessa X-20 instrument. Data analysis was performed using FlowJo V10 software (TreeStar), Microsoft Excel, and GraphPad Prism.

Ninety-six half-well plates were coated with recombinant protein at 1 μg/ml in PBS and incubated overnight at 4°C. Wells were blocked with 100μL 2% BSA in PBS for 1 hour at room temperature. Plates were washed three times with PBS + 0.05% Tween and incubated with twofold serially diluted sera (starting from 1:100) in PBS + 0.05% Tween for 1.5 hr at RT. After three washes, anti-mouse IgG H+L HRP (Aviva Systems Biology, catalog: OARA04973) at a 1:5000 dilution was added, and plates were incubated for 1 hr at RT. After three washes, plates were developed using TMB (Thermo Fisher, catalog: 34029) for 5 minutes at RT and subsequently blocked with H₂SO₄. The absorbance was measured with a TECAN Sunrise absorbance microplate reader (catalog: 16039400) at 450 nm.

GraphPad Prism software was used for statistical analysis. We compared the frequency of NP tetramer⁺ cells between the NP- and OVA-immunized groups per organ using the Student’s t-test. To compare the counts of cells between the groups in the ELISpot screening analysis, we used one-way ANOVA with multiple comparisons using Dunnett correction for each individual pool and control peptide. For comparison between NP and OVA prime groups or PR8-OVA and PR8-WT groups, a two-sided unpaired Student’s t-test was performed. **** denotes a p-value of ≤0.0001, *** denotes a p-value of ≤0.001, ** denotes a p-value of ≤0.01, and * denotes a p-value of ≤0.05. To compare the avidity index between days 7 and 14 from the NP or OVA prime groups, two-way ANOVA and post hoc Tukey’s HSD at an alpha of 0.05 were performed. **** denotes a p-value of ≤0.0001 and *denotes a p-value of ≤0.05.

To determine the influence of CD4⁺ T cells on B cell responses and B cell immunodominance, we first conducted a screen to identify immunodominant T cell epitopes within the NP and HA proteins of A/Puerto Rico/8/34 (PR8) in C57BL/6 mice. We chosethese proteins because they have been identified as immunodominant CD4 targets, and NP is relatively conserved between distinct viral strains (3, 5).

We infected C57BL/6 mice intranasally with PR8, and 10 days after infection, we performed IFN-γ and IL-2 ELISpot assays with cells from individual spleens and pooled mediastinal lymph nodes (medLNs) stimulated with overlapping peptide pools spanning the whole protein ( Tables S1 , S2 ). OVA peptide p323 was used as a control. While IFN-γ is a prototypical Th1 cytokine, IL-2 is related to T-cell proliferative capacity, and importantly, IL-2-producing T cells have been shown to preferentially differentiate into Tfh cells (30, 31).

Upon stimulation with NP peptide pools, spot-forming cells (SFCs) were observed in the spleen samples, with NP 43-48 and 49-54 having the highest average SFCs and similar results for both cytokines ( Figure 1A ). For the medLN samples, while the IL-2 trend was very similar to that in the spleen, the IFN-γ response was much broader, with several peptides being recognized ( Figure 1B ). Subsequently, we deconvoluted the response by testing individual peptides within the positive pools and identified NP peptides 52 and 53 as sufficient to induce strong IL-2 cytokine secretion in both spleen ( Figure 1C ) and medLN ( Figure 1D ). In the medLN, peptides 45 and 47 were further identified as giving signals above the background ( Figure 1D ). For HA, we identified a single peptide pool (12–17) by ELISpot ( Figures S1A, B ); however, we were not able to deconvolute the response to a single peptide resolution as none of them gave a signal above background (data not shown). Given the similar results in the first screening, deconvolution was not carried out for IFN-γ. Overall, we identified immunodominant HA and NP PR8 CD4 T cell epitopes in BL6 mice. Our results were consistent with those previously reported from a similar screening (32).

Having identified immunodominant NP and HA peptides, we wanted to confirm their ability to induce a CD4⁺ T cell response. We immunized mice with 0.5 nM immunodominant peptide pools (NP 45, 47, and 52; HA 13-18) or OVA-p323 in the footpad hock. At 10 days post-immunization (d.p.i.), we collected draining pLNs and spleens and tested the ability of CD4 T cells to respond to peptide stimulation by ELISpot using OVA as a stimulation control. We detected that immunization with the NP peptide pool was able to prime T cells, as measured by IL-2 and IFN-γ ELISpot ( Figure 1E and Figures S1C–E ). However, HA-peptide immunization failed to generate CD4 T cells above the OVA control value ( Figures S1F–I ). Similarly, when we stained T cells by flow cytometry using NP52_306-322 and HA16_91-107-class II tetramers, we were able to detect a clear Tetramer⁺ T cell population in the pLN ( Figure 1F ) and spleen ( Figure S1J ) only after NP immunization. In line with the ELISPOT data, HA immunization failed to yield a significant frequency of peptide-specific T cells ( Figures S1K, L ).

Furthermore, for NP-specific CD4 T cells in pLNs, we were able to classify the response obtained from immunizing with the positive peptide pools. Among NP⁺ CD44⁺ T cells, we detected a significant enrichment of Bcl6⁺ and Tbet⁺ CD4⁺ T cells compared to OVA-immunized controls ( Figure 1G ). Importantly, our results show the ability of a single NP peptide immunization to induce peptide-specific CD4 T cells that can assume Th1 and Tfh phenotypes. Given the HA-peptide prime results, we decided to focus on NP-peptide for the rest of the study.

The results from previous studies have been partially contradictory regarding the ability of peptide-primed CD4 T cells to help GC response and to provide intermolecular help (16, 17, 25, 29, 33). Hence, we wanted to test whether CD4 T cells induced by NP-peptide immunization could influence the humoral response to other viral proteins.

First, we used a simple immunization model in which NP peptide-primed (NP 45, 47, and 52) animals were subsequently vaccinated with an equal mixture of recombinant HA and NP proteins (5μg per protein) intraperitoneally (i.p.) in Addavax. Here, we did not expect any intermolecular help as the proteins are not physically attached to each other. Importantly, previous studies have shown similarities between immunodominant CD4 epitopes after infection and immunization (34). Mice were sacrificed 7 days post-immunization for B cell ELISpot and 14 days after protein immunization for flow cytometry ( Figure 2A ). Serum was collected at both time points to measure Abs. NP peptide priming slightly increased the amount of spleen antigen-specific ASCs after vaccination, as measured by ELISpot; however, no difference was detected in the response to specific proteins ( Figure 2B ). As a control, we measured ASC in pLNs, as these cells should not be affected by i.p. immunization, and confirmed that, in the majority of mice, NP-peptide immunization by itself was not sufficient to elicit detectable ASC responses ( Figure S2A ). Remarkably, in primed mice, NP-specific Abs showed accelerated kinetics, with responses already at high levels at 7 d.p.i. ( Figures 2C, D ). At 14 d.p.i., NP-Abs were still slightly higher compared to the control, but they almost plateaued from day 7 ( Figures 2C, D ). Importantly, responses to HA were not impacted ( Figures 2C, D ), as expected. Overall, when proteins are mixed, we demonstrated that CD4 T cell help is specific for the cognate protein and provides accelerated Ab kinetics.

However, when we measured total GC and antigen-specific responses by flow cytometry ( Figure S3 ), we did not detect a higher frequency of total GC B cells in the spleen ( Figure 2E ). Moreover, mixed immunization was able to potently induce NP-specific GC B cells, regardless of priming, but the HA-specific response was modest ( Figure 2G ). As i.p. immunization drains to the medLN (35), we also assayed the B cell response in this organ but similarly, we were unable to detect any differences in total or antigen-specific GC B cell frequency ( Figures S2B, D ). When we analyzed memory B cells defined as CD38⁺ GL7^- IgD^- IgM^- B cells, we found that the presence of memory CD4 T cells did not influence the frequency of total or antigen-specific memory B cells ( Figures 2F, H and Figures S2C, E ).

In summary, we demonstrated that memory CD4 T cells induced by peptide immunization were able to specifically accelerate the vaccine-induced serum Ab response but did not affect GC dynamics.

Prior to testing in an infection model, we decided to verify the ability of primed T cells to differentiate into Tfh cells within the medLN upon challenge. Therefore, we primed animals either with NP or control peptide and thereafter transferred total CD4 T cells to congenic mice. Recipients were challenged by i.n. infection and at 14 d.p.i. the T cells within the medLN were analyzed ( Figures 3A, B ). Overall, animals that received NP-primed CD4 T cells showed preferential expansion and differentiation of NP-specific Tfh cells compared to control-primed, transferred CD4 T cells ( Figure 3C ). In addition, animals receiving NP-primed CD4 T cells showed an overall increased frequency of Tfh cells but not Tfr cells ( Figures 3D, E ). Furthermore, we analyzed the Tfr : Tfh ratio or proportion of CXCR5⁺ cells, which are also Foxp3⁺; this measure has previously been associated with suppression of the Ab response (36). Indeed, animals that received NP-CD4 had a much lower Tfr : Tfh ratio ( Figure 3F ).

This experiment demonstrates the ability of primed T cells to differentiate into Tfh cells after infection, with the potential to influence the antibody response.

Having verified that NP-peptide prime was able to induce a Tfh CD4 T cell response and boost Ab responses to vaccination, we wanted to test their influence on the host humoral response after IAV infection. Here, we were particularly interested in verifying whether pre-existing NP-specific CD4 T cells would be able to help the humoral response to HA. As NP is more conserved, it would be desirable if such a response would favor neutralizing immunity to HA. Therefore, we immunized mice with the NP peptide pool and challenged them after 28 days by i.n. infection with PR8 virus ( Figure 4A ). At 0, 3, and 7 d.p.i, medLNs and spleens were harvested and processed for B cell ELISpot to monitor the emergence of antigen-specific ASCs. Day 3 ELISpot captures early extrafollicular responses, while day 7 captures a combination of extrafollicular and early GC-derived ASCs (7). Again, we analyzed both the medLN and the spleen. Previous studies have demonstrated the presence of antigen-specific B cells in both lymphoid organs after influenza infection, albeit with different dynamics (7, 8). Similar to the vaccination results, we did not detect a significant increase in NP-specific ASCs upon NP priming in any of the organs at any of the time points analyzed ( Figures 4B, C ; Figures S4A, B ). Despite this, and in line with the data from the immunization, analysis of serum Abs demonstrated a significantly higher response in primed animals, which was specific to NP ( Figures 4D, E ). In NP-primed animals, a trend toward a higher response to NP was already observed at 7 d.p.i. and was solidified at 14 d.p.i., thus demonstrating intramolecular enhancement of the Ab response when CD4 T cell help was present ( Figures 4D, E ). HA Ab levels were not influenced by the primed CD4 T cells. As our results again showed a disconnect between ASC numbers and serum Abs, we reasoned that the total number of plasmablasts could be higher in primed animals and therefore counted total cells within the medLN. While there was a trend toward higher cellularity in the medLNs of day 7 primed animals, the number was not significantly higher ( Figure S4E ). Furthermore, we hypothesized that primed CD4 T cells could promote Abs of higher affinity. Therefore, we performed an avidity ELISA to determine the avidity index of antigen-specific Abs. We did not detect any difference in the avidity of NP-specific Abs in primed vs non-primed animals ( Figure 4F ), and while we cannot offer a definitive explanation, the results suggest that the higher Abs measured by ELISA in primed animals may be due to a slightly higher cell number possibly combined with a higher secretion rate of Abs. Interestingly, regardless of prime, the avidity of NP Abs was generally higher at 7 d.p.i, compared to HA-Abs, which showed more canonical affinity maturation ( Figure 4F ).

We also analyzed GC and memory B cell formation by flow cytometry ( Figure S3 ). Similar to the vaccination results, we failed to detect an enhancement in total or antigen-specific GC B cells at 14 d.p.i in both the medLN ( Figures 5A, B, D ) and the spleen ( Figures S4C,F ). However, in this experiment, we also sampled 7 d.p.i. and demonstrated the role of pre-existing CD4 T cells in accelerating GC formation after IAV infection, similar to what was previously described (16). Indeed, we detected an overall increase in GC B cell frequency at 7 d.p.i. in both medLNs ( Figure 5B ) and spleens ( Figure S4C ). The frequency of switched memory B cells was unchanged ( Figure 5C and Figure S4D ). Importantly, when looking at antigen specificity, the overall higher GC frequency did not translate to a higher proportion of antigen-specific cells ( Figure 5D and Figure S4F ), an aspect that had not been investigated previously. In addition, we identified a significant increase in NP-specific memory B cells at 14 d.p.i. in the medLN ( Figure 5E ) but not in the spleen ( Figure S4G ).

Overall, our results suggest that CD4 T cell help is efficiently recalled after infection; however, while it can accelerate GC formation and size, it might not be crucial to enhance the frequency of antigen-specific B cells within the GC. Rather, it has clear effects on overall serum Ab levels and could be important for early fate decisions and terminal differentiation. Furthermore, in agreement with previous results, we show that after IAV infection, CD4 T cells do not provide intermolecular help.