Previously, we discovered that influenza-generated CD4 effectors must recognize cognate Ag at a defined effector checkpoint to become memory cells. Ag recognition was also required for efficient protection against lethal influenza infection. To extend these findings, we investigated if vaccine-generated effectors would have the same requirement. We compared live infection with influenza to an inactivated whole influenza vaccine. Live infection provided strong, long-lasting Ag presentation that persisted through the effector phase. It stimulated effector generation, long-lived CD4 memory generation, and robust generation of Ab-producing B cells. In contrast, immunization with an inactivated virus vaccine, even when enhanced by additional Ag-pulsed APC, presented Ag for 3 d or less and generated few CD4 memory cells or long-lived Ab-producing B cells. To test if checkpoint Ag addition would enhance this vaccine response, we immunized mice with inactivated vaccine and injected Ag-pulsed activated APC at the predicted effector checkpoint to provide Ag presentation to the effector CD4 T cells. This enhanced generation of CD4 memory, especially tissue-resident memory in the lung, long-lived bone marrow Ab-secreting cells, and influenza-specific IgG Ab. All responses increased as we increased the density of peptide Ag on the APC to high levels. This suggests that CD4 effectors induced by inactivated vaccine require high levels of cognate Ag recognition at the effector checkpoint to most efficiently become memory cells. Thus, we suggest that nonlive vaccines will need to provide high levels of Ag recognition throughout the effector checkpoint to optimize CD4 memory generation.

Live infection with influenza and other viruses produces robust immune responses that generate long-lived CD4, CD8, and B cell memory that synergize to provide durable and long-lived protection (16). However, influenza and other ssRNA viruses (e.g., HIV and coronaviruses) rapidly accumulate mutations, leading to changes in viral Ags and consequential escape from pre-existing memory immunity. Memory B cells and neutralizing Abs predominantly recognize determinants on hemagglutinin (HA) and neuraminidase coat protein Ags that differ between heterosubtypic influenza strains (e.g., H1N1, H3N2, H5N1). In contrast, memory T cells, both CD4 and CD8, recognize epitopes from core proteins, many of which are conserved in the influenza strains in almost all outbreaks (7) and thus can provide heterosubtypic protection. New influenza strains are generated by genetic recombination (heterosubtypic strains) as well as by rapid mutation, including a lack of proofreading. Among other ssRNA viruses, coronaviruses accumulate mutations mostly through rapid replication (8), whereas HIV diversity is attributed in part to reverse transcription (9). Abs often drive selection of nonrecognized variants, but more T cell epitopes are conserved (10, 11), and so T cells provide key targets for vaccine design. Memory T cells are heterogeneous, functionally diverse, and able to provide protection through multiple synergizing mechanisms (5). The different subsets of CD4 memory T cells have varying lifespans (12) and distinct distribution among lymphoid and tissue sites. Central memory cells (TCM), defined by expression (or re-expression) of CD62L, are retained in the secondary lymphoid organs (spleen and draining lymph nodes [dLN]) (13), whereas effector memory cells (TEM) continuously recirculate, and tissue-resident memory cells (TRM) reside in the peripheral lymphoid and nonlymphoid tissues (3, 1315). Compared with TCM and TEM, CD4 TRM are more protective in part because they respond quickly at sites of entry to provide a first line of defense against pathogens (16). Influenza infection induces all these subsets, including a large population of lung-resident memory T cells (17). They clear virus and promote resolution and repair postinfection by multiple mechanisms (5, 18). However, currently used influenza vaccines generate only modest titers of neutralizing Ab that are often short lived and little T cell immunity, and thus, they do not provide robust long-term or heterosubtypic protection (1921), especially to new or mutated strains that develop each year or two.

Our previous studies (22, 23) showed that following live influenza infection CD4 effectors must recognize Ag again at a defined “checkpoint” to differentiate into memory cells. The checkpoint occurred during the time when effector accumulation was peaking at 6–8 d postinfection. Primary effectors needed to be induced to produce and respond to IL-2, which acted to block their default apoptosis and induced longevity (22). The same checkpoint is defined by the time when activated APCs (Ag/TLR agonist-activated APC [*APC]) and IL-2 are required (22, 23). The signals are delivered during the effector–Ag/*APC cognate interaction, and without these signals, effectors disappear, whereas with them, some can become memory cells. We now call this the “effector checkpoint” because during this time interval, the effectors can follow multiple fates that include 1) apoptosis, resulting in contraction when they do not see Ag; 2) further differentiation into late functional effectors such as T follicular helpers (TFH) and cytotoxic CD4 T cells; and 3) transition to resting memory cells in both the secondary lymphoid organs and infected lung, both when signals from Ag recognition. Using an in vivo sequential transfer model, we showed that without Ag presentation at the checkpoint, the CD4 effectors did not become memory cells. The provision of Ag to CD4 effectors at the checkpoint, even delivered by peptide-pulsed activated APC, which presented Ag for only 48 h, drove the effectors to differentiate into memory cells (19). With effector checkpoint cognate Ag recognition, memory generation increased 20–200 fold in spleen, LN, and lungs, indicating that the short cognate interaction was sufficient for effectors to avoid apoptosis and differentiate into memory (22). The requirement by CD4 effectors for Ag presentation through the checkpoint interval is consistent with other studies that suggested that the differentiation of naive CD4 T cells into memory cells requires prolonged Ag presentation (24).

Unlike viral infection, current inactivated, purified split-virion, or subunit vaccines focus on inducing Ab to seasonal HA variants and are mostly ineffective at inducing strong long-lived T or B cell memory (19, 25, 26). Current influenza vaccines are based on three basic formulations, including 1) whole inactivated influenza virus (WIV), which was the first influenza vaccine in 1936 and is still in use (27); 2) more purified split-virion or subunit vaccines, which are used most extensively; and 3) cold-adapted live vaccines, which were developed more recently (28) and which survive only a couple of days in the respiratory tract. In this study, we use WIV because it contains all viral components, including small amount of RNA, which provides pathogen-associated molecular patterns (PAMP), and a wide range of epitope peptides for T cell recognition (7) and immunogens for B cell recognition (e.g., HA protein). In contrast, the purified split-virion or subunit vaccines are composed of isolated and enriched components or recombinant proteins (e.g., HA and neuraminidase) from multiple viral strains to induce a broader Ab response, and T cell epitope peptides from internal viral proteins are minimized (29). These are given i.m. except for the cold-adapted FluMist, which is given as a nasal mist. In one case, an adjuvant is added. Given the short longevity of proteins, even in those in oil-based adjuvants, it is likely that vaccines do not present high amounts of Ag at the effector checkpoint, and thus, we postulate they are unlikely to efficiently drive CD4 memory differentiation or full effector differentiation. In this study, we ask if whole inactivated viral vaccine-induced CD4 effectors generated in mice do indeed require Ag recognition at an effector checkpoint to drive differentiation of effectors so that they reach the lung and therein become memory cells. If so, it would imply changes in the delivery of vaccines are called for to provide the additional signals needed.

Our studies in this article show that WIV “immunization” leads naive CD4 T cells to proliferate and accumulate, which peaks at 4–5 d, whereas the response induced by PR8 infection is more vigorous and persistent. We find WIV immunization presents Ag only during the first 3 d, suggesting that Ag presentation at the checkpoint interval does not occur. Even when we add *APC along with WIV during priming (WIV + Ag/*APC), which we showed earlier could strikingly enhance CD4 responses (30, 31), Ag presentation is still short lived and occurs effectively only during the first 3 d. In contrast, PR8 infection leads to a delayed but persistent Ag presentation. We find that the addition of Ag-pulsed APC following WIV vaccines, at 5 d posttreatment (dpt), increased the generation of donor memory cells, strongly enhanced lung donor T cells with a resident memory phenotype, and modestly enhanced the number of plasma cells secreting anti-PR8 Abs in the bone marrow (BM). Thus, our study shows that CD4 effectors generated by a vaccine (WIV + Ag/*APC) require Ag recognition at an effector checkpoint to drive differentiation into memory cells, especially lung TRM, supporting the concept that vaccine-induced effectors, like those generated by infection, have a checkpoint requiring Ag recognition that regulates memory generation. Our results also suggest that checkpoint Ag recognition by CD4 effectors also plays a role in supporting B cell responses. We discuss the implications for the vaccine and the possible reasons why the impact of adding Ag/*APC at the checkpoint is not as dramatic as it was with effectors generated by live virus.

Female and male 8–10-wk-old BALB/c, C57BL/6 and BALB/c.Thy1.1+/+ mice were purchased from The Jackson Laboratory. The transgenic (Tg) DO11.10 [BALB/cJ-Tg (DO11.10) Tg (CARΔ-1)1Jdgr] mice were purchased from Taconic Biosciences (Rensselaer, NY). HNT TCR Tg.Thy1.1+/− mice were generated by breeding BALB/c.Thy1.1+/+ mice with HNT TCR Tg mice. NP TCR Tg mice were generated in a collaborative effort by Dr. E. Huseby’s laboratory. FluNP TCR Tg mice express CD4 TCR that is specific for the MHC class II-IAb–restricted influenza NP311–325 epitope (QVYSLIRPNENPAHK). Thy1.1+/− mice were generated by breeding C57BL/6.Thy1.1+/+ mice with NP TCR Tg mice. Details of this strain will be described in an upcoming work (M. Jones, E. Huseby, L. Stern, and S.L. Swain, manuscript in preparation). All the animals were maintained in the animal facility at University of Massachusetts Medical School.

Influenza A virus (IAV), A/PR8/34 (PR8, H1N1), was derived from a stock in A. Harmsen’s laboratory originally from D. Morgan at Scripps Research Institute. It was passaged through BALB/c mice and grown in chicken eggs at the Trudeau Institute. This same batch of virus has been used in our previous studies (32). Formalin-inactivated influenza vaccine (WIV, purified and inactivated influenza A/PR/8/34 [H1N1]) was purchased from Charles River Laboratories (material no. 10100782). PR8 infection was carried out by intranasal (i.n.) administration of virus at 0.3 LD50. Immunization with WIV was by i.v. injection of 2.5 μg WIV. To enhance the CD4 T cell response, 3 × 105 HNT/TLR agonist-activated BM-derived dendritic cells (BMDC) (*BMDC), *BMDC pulsed with 12.5 μM HA126–138 for 1 h (see below), was added with the WIV as in our previous study (30). The immunization strategy is referred to as WIV + Ag/*APC.

For BMDC generation, BM from BALB/c mice was harvested from femurs and tibias, and a single-cell suspension was generated. Then, 8 × 106 BM cells were seeded in each petri dish (no. FB0875712; Thermo Fisher Scientific, Hampton, NH) and cultured in 10 ml T cell medium with GM-CSF (no. 576306; BioLegend, San Diego, CA) (at 10 ng/ml). On day 3, 10 ml medium was added. On day 5, the medium was refreshed by removing 10 ml culture and replacing with 10 ml fresh medium. On day 6, poly(I:C) (low m.w.) (InvivoGen, San Diego, CA) and CpG ODN 1826 (5′-TCCATGACGTTCCTGACGTT-3′) (Integrated DNA Technologies, San Diego, CA) were added to activate the cultured cells. On day 7, the cells were harvested, and CD11c+ cells were isolated using the CD11c MicroBeads UltraPure for mouse (no. 130-108-338; Miltenyi Biotec, Bergisch Gladbach, Germany) and LS column (no. 130-042-401; Miltenyi Biotec). These activated CD11c+ cells (*BMDC) were pulsed with the HNT epitope peptide, HA126–138 (HNTNGVTAACSHE; New England Peptide, Gardner, MA), at the indicated concentrations (1.25, 12.5, and 125 μM) for 1 h at 37°C. The HNT peptide-pulsed, activated BMDCs are referred to as Ag/*APC.

Spleen and prominent LN were harvested from HNT Thy1.1+ BALB/c TCR Tg mice or NP Thy1.1+ TCR C57BL/6 Tg mice. Single-cell suspensions were generated by processing the tissues. Small resting cells were isolated by 40, 53, 62, and 80% Percoll (no. 17-0891-01; GE Healthcare Life Sciences, Pittsburgh, PA) density gradient centrifugation and collecting the cells between the 40 and 53% inner layers. Naive CD4 T cells were isolated from those small resting cells by mouse CD4 (L3T4) MicroBeads (no. 130-117-043; Miltenyi Biotec) and LS column (no. 130-042-401; Miltenyi Biotec).

For transfer experiments (Figs. 36), 105 naive Thy1.1+ HNT TCR Tg CD4 T cells were transferred into DO11.10 hosts, treated either with WIV + HNT/*BMDC (Ag/*APC) immunization or by infection with 0.3 LD50 PR8 1 d prior to treatment and −1 dpt. To study the impact of checkpoint Ag recognition on memory generation and T helper function, 3 × 105 Ag/*APC were injected i.v. to the WIV + Ag/*APC-immunized mice at 5 dpt. For the CD4 T cell kinetics study in Fig. 1, naive Thy1.1+ HNT TCR Tg CD4 T cells were labeled with CFSE by incubating the naive CD4 T cells with 5 μM CFSE (no. C34554; Invitrogen, Carlsbad, CA) solution at 107 cells/ml at 37°C for 20 min. Then, 105 CFSE-labeled naive CD4 T cells were injected i.v. on −1 dpt into the BALB/c mice given WIV + Ag/*APC or PR8 live virus. For the CD4 T cell kinetics study in Supplemental Figs. 1 and 4, 105 naive Thy1.1+ NP TCR Tg CD4 T cells were i.v. injected on −1 dpt into C57BL/6 mice given WIV or PR8 live virus. For Supplemental Fig. 4, *BMDC pulsed with NP311–325 were used as Ag/*APC to assess the impact of checkpoint Ag, and 3 × 105 of these Ag/*APC were i.v. injected to hosts at 5 dpt. For the Ag presentation kinetics studies (Fig. 2, Supplemental Figs. 2, 3), 106 CFSE-labeled naive Thy1.1+ HNT TCR Tg CD4 T cells were i.v. injected into treated BALB/c mice on 0, 3, 5, 7, and 11 dpt. After 2.5 d, spleen, lung dLN/mediastinal LN (mLN), and lung cells were harvested from the host mice. We gated on donor Thy1.1+ cells and determined the loss of CFSE dye, indicating division induced by Ag presentation.

Corning 96 Well EIA/RIA Assay Microplates (MilliporeSigma, Burlington, MA) were coated with WIV (Charles River Laboratories) in ELISA coating buffer (0.05 M carbonate–bicarbonate [pH 9.6]). Plates were blocked with BSA in PBS. Serial dilutions of serum were plated and incubated at 4°C overnight. Plates were washed with ELISA wash buffer (0.05% Tween 20 in PBS) followed by incubation with HPR-conjugated goat anti-mouse IgG (Southern Biotech, Birmingham, AL). Plates were developed with the peroxidase substrate, o-phenylenediamine (MilliporeSigma). OD readings at 490 nm obtained from serial dilution were analyzed by nonlinear regression. The end point titer was defined as the serum dilution, resulting in OD readings at 490 nm value equivalent to the cutoff value.

MultiScreen-HA Filter Plates (MilliporeSigma) were coated with WIV (Charles River Laboratories) in ELISA coating buffer (0.05 M carbonate–bicarbonate [pH 9.6]). Plates were blocked with BSA in PBS. Serial dilutions of splenocytes and BM cells were plated and incubated at 37°C overnight. Plates were washed with ELISA wash buffer (0.05% Tween 20 in PBS) followed by incubation with HPR-conjugated goat anti-mouse IgG (Southern Biotech). Plates were developed with AEC Substrate Set (BD Biosciences, San Jose, CA). PR8-specific IgG Ab-secreting cell (AbSC) spots were quantified using the ImmunoSpot reader, and the data were analyzed with ImmunoSpot Software (Cellular Technology, Cleveland, OH).

Spleens, lungs, and LNs (dLN/mLN) were harvested from the infected, immunized, or naive mice. Single-cell suspensions were prepared, blocked with anti-Fc RII/III mAb (Bio X Cell, West Lebanon, NH), and LIVE/DEAD stained using LIVE/DEAD Fixable Yellow Dead Cell Stain Kit (Invitrogen). Cells were then surface stained with Abs: FITC–anti-mouse CD69 (clone H1.2F3), PE-Cy7–anti-mouse CD44 (clone IM7), PE-Cy–anti-mouse Thy1.1 (clone OX-7), APC–anti-mouse CXCR6 (clone SA051D1), and Alexa Fluor 488 anti-mouse ICOS (clone C398.4A) were purchased from BioLegend; APC–anti-mouse Thy1.1 (clone HISS1), APC–eFluor 780–anti-mouse CD4 (clone RM4-5), PerCP–eFluor 710–anti-mouse CXCR5 (clone SPRCL5), and PerCP-Cyanine5.5–anti-mouse CD62L (clone MEL-14) were purchased from eBioscience (San Diego, CA); and PerCP-Cy rat anti-mouse CD44 (clone IM7) and PE–anti-mouse Bcl-6 (clone K112-91) were purchased from BD Biosciences. For the in vivo and ex vivo Ab labeling, Pacific Blue anti-mouse CD4 (clone RM4-4; BioLegend) was injected into the immunized mice. At 5 min after injection, mice were euthanized with isoflurane (Patterson Veterinary Supply, Devens, MA) followed by cervical dislocation. Blood, spleens, and lungs of the mice were harvested, and the single-cell suspensions were prepared. Cells from different tissues were stained as indicated above. A successful in vivo Ab labeling was assured by over 99% blood CD4 T cells accessible to in vivo Abs as indicated in the published protocol (33). Stained samples were fixed with 1% formaldehyde (Thermo Fisher Scientific, Waltham, MA). Data were acquired on a BD LSRII Fortessa instrument (BD Biosciences) and analyzed using the FlowJo software program (Tree Star, Ashland, OR).

Experimental animal procedures were conducted in accordance with guidelines outlined by the Office of Laboratory Animal Welfare of the National Institutes of Health. Protocols were approved by the Institutional Animal Care and Use Committee of the University of Massachusetts Medical School (Worcester, MA).

For statistical analysis, significance comparing the mean of two normally distributed groups was determined by the two-tailed unpaired Student t test. Significance comparing the mean of over three normally distributed groups was determined by one-way ANOVA with Bonferroni multiple comparison posttest. The p values <0.05 were considered significant. Error bars in figures represent the SEM: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. The results were analyzed by GraphPad Prism software (GraphPad Software, San Diego, CA).

We based our experimental design on the cell transfer model that we previously developed to evaluate the requirements for naive CD4 T cell response to WIV immunization. To evaluate the development of CD4 effectors and memory, we transferred HNT TCR Tg CD4 T cells (BALB/c. HNT. Thy1.1) into BALB/c mice. We previously showed the naive HNT CD4 T cells, whose TCR specifically recognizes the HA126–138 peptide in the HA of PR8 (H1) influenza, respond to PR8 infection in vivo (23, 32) and in vitro to HNT peptide-pulsed APC (31). We treated the host mice with WIV vaccine (i.v.) and, for comparison, infected with influenza (PR8) live virus. By following the proliferation and differentiation of donor cells, this model allows us to study both the Ag presentation of the HNT epitope peptide following infection versus vaccination and the fate of naive CD4 T cells postinfection versus vaccine. We used our previously published strategy, WIV + Ag/*APC, as our standard immunization because adding exogenous Ag/*APC induced a more robust naive CD4 T cell effector response of both young and aged CD4 T cells (30), and we want to ensure efficient initial naive CD4 effector generation. HNT peptide-pulsed BMDC were activated with poly(I:C) and CpG (Ag/*APC) because activated APCs express high levels of MHC and costimulatory ligands (CD80/86) that maximize naive CD4 T cell response (30). The coinjection WIV/Ag-*APC strategy induced greater donor CD4 T cell proliferation and differentiation and greater PR8-specific IgG in an IL-6–dependent manner (30). In key experiments, we compared responses to WIV alone. We know the Ag/*APC only present Ag for 48 h or less (22), so we can readily evaluate the subsequent requirement for Ag recognition by adding Ag/*APC after 4–5 d.

To determine the kinetics of the vaccine-induced CD4 T cell response, we followed donor naive BALB/c.HNT.Thy1.1 CD4 T cell recovery and division following WIV + Ag/*APC–immunized mice and compared it to that in PR8-infected mice. WIV + Ag/*APC immunization induced a rapid and vigorous expansion of CD4 T cells from 2 to 5 dpt in the spleen (Fig. 1B), LN (Fig. 1D), and lung (Fig. 1F). CD4 donor cell number peaked on 4 or 5 dpt and then rapidly decreased in spleen (Fig. 1B) and mLN (Fig. 1D). We found fewer CD4 T effectors in the lung, and their numbers were maintained at similar levels during 4–7 dpt (Fig. 1F). In contrast to the vaccination, PR8 infection induced a delayed CD4 T cell response. Postinfection, donor CD4 T cells did not increase detectably until 3–4 dpt in dLN (Fig. 1D) and even later (5 dpt) in spleen (Fig. 1B). In earlier studies, we found PR8-induced donor HNT CD4 effector and polyclonal CD4 effector response peaked at 6–7 dpt in spleen, lung, and dLN (15, 34) and stayed elevated through 10 dpt. Thus, WIV + Ag/*APC induced an earlier but transient CD4 T cell response, whereas live PR8 infection induced a later response that was sustained longer, consistent with our earlier studies (15, 34).

To determine if the addition of Ag/*APC to WIV would alter the kinetics of CD4 response to WIV immunization, we compared the response kinetics of WIV alone (without the added Ag/*APC) and to live PR8 virus. We found the same pattern of the early and transient response in spleen, LN, and lung in WIV-treated mice as with WIV + Ag/*APC–treated mice (Supplemental Fig. 1), indicating that the addition of Ag*/APC did not alter the kinetics of CD4 response to WIV immunization.

To further evaluate when naive CD4 T cells actually started proliferating following WIV immunization, we assessed donor CD4 T cell division by analyzing the loss of CFSE (Fig. 1C, 1E, 1G). After immunization with WIV + Ag/*APC, donor cells were dividing by 2 dpt in the spleen and by 3 dpt in the LN and lung (Fig. 1C, 1E, 1G). We compared this to WIV alone without Ag/*APC. After both WIV and WIV with Ag/*APC, detector cells had undergone >6–8 divisions in all organs by 3 dpt. In contrast, after PR8 infection, divided donor CD4 HNT cells were not detectable in spleen until 5 dpt (Fig. 1C), in LN until 4 dpt (Fig. 1E), and in lung until 5 dpt (Fig. 1G). By 7 dpt, transferred cells in all three organs were fully divided. Thus, compared with PR8 infection, vaccine immunization, with or without extra Ag/*APC at 0 dpt, induced earlier but more transient donor CD4 T cell division that led to an earlier peak of effector accumulation and an early contraction.

The earlier but more transient CD4 T cell response to vaccine immunization suggests that the Ag presentation was short lived. Thus, we determined when the donor CD4 T cells saw the Ag, as in earlier studies (32). We used CFSE-labeled naive donor HNT CD4 T cells as detectors that lose the dye as they divide to visualize ongoing Ag presentation. We transferred these detector cells into the immunized or infected hosts at 0, 3, 5, 7, and 11 dpt (Fig. 2A), harvested them 2.5 d later, and assessed CFSE dilution of the donor CD4 T cells in spleen, LN, and lung (Fig. 2A). We quantitated the extent of Ag presentation at the indicated intervals (0–2.5, 3–5.5, 5–7.5, 7–9.5, and 11–13.5 dpt). In mice immunized with WIV + Ag/*APC, donor cells divided most extensively in the spleen when added at 0 dpt, whereas those added at 3 dpt or later showed much less division (loss of CFSE) (Fig. 2B). Immunization with WIV alone showed a similar pattern of results, confirming that both the WIV alone and mixed with Ag/*APC present little Ag after 3 dpt. In contrast, in PR8-infected mice, donor cells showed no evidence of Ag recognition during the 0–2.5 dpt interval, with no division of detector cells in spleen (Fig. 2A), dLN, or lung (Supplemental Figs. 2A, 3A) when added at 0 dpt. Instead, PR8 infection caused donor division that was sustained at a higher level at intervals from day 3 on, continuing to days 5–7, with less but still detectable division at 11 dpt in spleen (Fig. 2B, 2C), LN, and lung (Supplemental Figs. 2B, 2C, 3B, 3C). The sustained division lead to greater expansion by 7 dpt (Fig. 1, Supplemental Fig. 1).

We calculated the total donor cell recovery and the divided donor cells. In vaccine-immunized hosts, the highest recovery of both total and divided donor cells occurred when the detectors were added at 0 dpt, with the numbers decreasing when added at 3 dpt and thereafter (Fig. 2C, 2D). Similar results were seen in the LN and lung (Supplemental Figs. 2B, 2C, 3B, 3C). Thus, in the WIV + Ag/*APC– or WIV alone–immunized mice, the magnitude of Ag presentation, shown as both division and recovery of detectors, peaked before 2.5 dpt (Fig. 2B) and declined abruptly after 3 dpt. In contrast, following PR8 infection, CFSE loss was delayed for at least 2–3 d, peaking at day 5, and then gradually declining through days 7 and by 11 dpt in the spleen, dLN (Supplemental Fig. 2), and lung (Supplemental Fig. 3). Donor cells recovered in the lung at later times might be effector cells that recognized Ag and may have divided in spleen and LN and thereafter migrated to the lung; this set of results may not accurately reflect in situ Ag presentation in lung. Ag presentation induced by WIV alone was of a somewhat lower magnitude, with significantly fewer divisions and lower recovery of detectors, but it followed similar kinetics as that induced by our standard WIV + Ag/*APC vaccine (Fig. 2, Supplemental Figs. 2, 3).

Together, these data indicate that effective Ag presentation by either WIV + Ag/*APC or WIV immunization is early but transient, leading to early but robust CD4 T cell proliferation and generation of highly divided effectors peaking at 4–5 dpt, which then contract (Figs. 1, 2), whereas PR8 infection leads to a longer period of Ag presentation with effectors peaking at 5–7 dpt before the onset of contraction. This is consistent with the concept that even if WIV-generated effectors get extra help from Ag/*APC during priming, the CD4 donor effectors are not exposed to sufficiently strong Ag presentation at 4–5 dpt when the response peaks because high amounts of Ag are present for only 3 d. In contrast, PR8-induced effectors peak at 7 dpt and are exposed to Ag presentation from 3 to 11 dpt. Thus PR8-generated but not WIV-generated CD4 effectors can and do recognize Ag during the peak of their effector response, which we earlier identified as a checkpoint, coincident with the effector peak.

Because Ag presentation signals induced by WIV + Ag/*APC vaccine had nearly disappeared by 3–5 dpt (Figs. 1, 2), we tested whether providing Ag/*APC exogenously at 5 dpt, at the time when effector numbers peak, would augment memory generation. We reasoned that at their peak, effectors are most committed to default apoptosis and require Ag/*APC to stimulate IL-2 that promotes their rescue, as suggested in our earlier studies (22, 23). As above, mice were immunized by WIV + Ag/*APC on 0 dpt, and Ag/*APC, known to effectively provide Ag presentation signals that drive PR8-induced effectors to become CD4 memory generation effectors (22), were added at 5 dpt. To study the effects of checkpoint Ag on both memory CD4 T cell generation and CD4 helper function, indicated by Ab production, we used our previously established reductionist model. We transferred donor HNT.Thy1.1 naive CD4 T cells into the DO11.10 TCR hosts and followed the donor CD4 T cell response after WIV immunization. DO11.10 TCR hosts bear a Tg TCR specific for an influenza-irrelevant OVA peptide, and host CD4 T cells do not respond to WIV. The transfer of HNT CD4 T cells into DO11.10 hosts allows us to evaluate not only the generation of donor CD4 memory cells but also their functional contribution (e.g., helper function) independent of host CD4 responses (23, 30). The advantage of this model is the ability to follow the response and impact of the naive cohort of cells unambiguously and assess the impact of providing the peptide Ag to these cells without introduction of Ag seen by B cells or host T cells.

To test the impact of Ag recognition at the effector checkpoint on memory CD4 T cell generation, we transferred Ag/*APC to the vaccine-immunized hosts on 5 dpt and compared the results with hosts that did not receive Ag/*APC. We enumerated donor cells in spleen and lung on 35 dpt. The addition of Ag/*APC at 5 dpt led to a slight 1.7-fold increase in total donor memory cell number in spleen (Fig. 3B) but a larger 4.5-fold donor cell increase in the lung (Fig. 3C). The selective increase of donor memory cells in the lung led us to analyze whether Ag/*APC at 5 dpt favors generation of CD62Llo memory cells, which include both tissue-restricted memory TRM and recirculating TEM. Following vaccine immunization with no Ag/*APC at 5 d postinfection, donor cells were mostly CD62lo in both spleen (80%) and lung (78%) (Fig. 3D). When Ag/*APCs were injected on 5 dpt, there was a minor increase in percentage, and a 1.7-fold increase in the number of CD62Llo donor CD4 memory cells in the spleen and a higher increase in percentage (>90%) and a 4.2-fold increase in lung (Fig. 3D–F). Thus, the addition of checkpoint Ag/*APC on 5 dpt led to a more than 4-fold higher generation of both total and CD62lo memory and CD4 T cells in the lung with a smaller increase in the spleen.

We also evaluated if memory augmentation by checkpoint Ag presentation occurs after immunization with WIV alone in a different but comparable model into which we transfer FluNP TCR CD4 T cells (as in Supplemental Fig. 1A) as the donor cells whose fate we follow. The mice were immunized with WIV alone and injected with Ag/*APC or not at 5 d postvaccination as in Supplemental Fig. 1B. Similar to WIV + Ag/*APC immunization, donor memory cells in both spleen (Supplemental Fig. 4B) and lung (Supplemental Fig. 4C) were significantly increased in both the frequency (3.9-fold for spleen and 9.4-fold for lung) and number (∼2.0-fold for spleen and lung). To confirm that Ag is essential for the impact, we compared the impact of Ag/*APC with that of *APC not pulsed with peptide Ag. In the HNT model, we added *APC alone without peptide at 5 dpt. We compared the generation of memory donor cells to those in immunized mice without 5 dpt *APC addition. We found adding*APC without Ag at 5 dpt had no impact on donor memory cell recovery, which was at a similar frequency and number as in untreated immunized mice (Supplemental Fig. 4E, 4F) and the same general magnitude as in Fig. 4, without Ag/*APC. Thus, as expected, the impact of Ag/*APC is peptide Ag dependent.

Memory CD4 T cells detected in peripheral tissues include both recirculating TEM and retained TRM in the tissue. TRM play critical roles in providing long-lived immunity locally (35). Specifically, in the lung, TRM provide strong protection against influenza challenge (36). Both TEM and TRM are CD62Llo cells. Therefore, we further analyzed what fraction of the augmented CD62Llo memory CD4 T cells are TRM rather than TEM.

We examined expression of CD69, a TRM signature marker of all TRM at 35 dpt (36, 37). The addition of Ag/*APC at 5 dpt increased both CD69+ donor lung memory cell frequency (2.1-fold) and number (6.6-fold) (Fig. 4A, 4B). Moreover, expression of CXCR6, shown to be upregulated on TRM (38), was also significantly upregulated by the addition of Ag/*APC at 5 dpt. The addition of Ag/*APC at 5 dpt also increased the frequency (2.0-fold) and number (3.3-fold) of CXCR6+ donor lung memory cells (Fig. 4C, 4D). To confirm that donor memory cells were bona fide TRM, we used in vivo Ab labeling in the same experiments to identify more Ab-inaccessible donor TRM from those circulating that are Ab accessible. Consistent with expectation (33), over 99% blood CD4 T cells were labeled by i.v.-injected Ab, whereas most of the donor memory CD4 T cells at 35 dpt in the spleen remained unlabeled, confirming that the spleen TCM share tissue-resident properties (Fig. 4C). We found that the addition of Ag/*APC at 5 dpt led to a significantly increased fraction of unlabeled donor CD4 memory in lung (Fig. 4E) but not spleen (Fig. 4D), strongly suggesting many are TRM. The total number of donor TRM was increased 4.7-fold in the lung, with little impact on donor TRM in spleen. Taken together, Ag/*APC addition on 5 dpt promoted immunization-generated CD4 effectors to migrate and accumulate in the lung in larger numbers and to further differentiate into CD69+/CXCR6+ TRM. Thus, recognition of checkpoint Ag selectively promoted CD4 effectors to become memory lung TRM but not TEM.

Given the increased donor memory generation, we asked if adding checkpoint Ag/*APC would lead to more helper cells and to better helper activity. We assessed expression of helper phenotypes on CD4 donor effectors, the recovery of B AbSC responses (35 dpt), and influenza-specific Ab production. As above, HNT naive CD4 donor cells were transferred to DO11.10 hosts, which were immunized with WIV + Ag/*APC, and Ag/*APC were added or not on 5 dpt. To assess the induction of more differentiated effectors, we measured the expression of ICOS, one signature marker of functional CD4 Th cells (39), 2 d later (Fig. 5A–C). Adding Ag/*APC on 5 dpt enhanced the expression of ICOS significantly (Fig. 5A), with higher expression per effector (2.3-fold) (Fig. 5B) and increased numbers of ICOS+ donor effectors (2.4-fold) (Fig. 5C). Thus, Ag presentation at the effector checkpoint enhanced effectors to upregulate ICOS expression, which might correlate with enhanced helper activity. To analyze whether the effectors developed other indications of a more TFH-like phenotype, we evaluated the coexpression of Bcl-6 and CXCR5 on donor cells with or without Ag/*APC treatment at 5 dpt. We found adding the checkpoint source of Ag resulted in the generation of a modest 2-fold increase in Bcl-6+CXCR5+ and Bcl-6+ICOS+ expression on a relatively small population of cells (Fig. 5E, 5F), which were fewer and with less intensity of staining than those generated by PR8 infection (Fig. 5D). This moderate impact may be because this is a time 1–2 d before the peak of TFH, and/or it may indicate that the effectors require additional signals to progress to full TFH status.

To test if checkpoint Ag presentation resulted in the generation of a prolonged AbSC response, we analyzed the B cell Ab response to PR8 IAV after 35 dpt using the same experimental design. We measured the serum titer of anti-PR8 IgG by ELISA and determined the number of anti-PR8 IgG-secreting cells (AbSC) by ELISpot. Anti-PR8 IgG titers were significantly increased (2.8-fold) in the vaccine-immunized mice treated with Ag/*APC at 5 dpt (Fig. 5C). Consistent with serum titers, PR8-specific IgG AbSC in spleen (Fig. 5D) were modestly increased in number (1.7-fold) with the Ag/*APC treatment at 5 dpt. Anti-PR8 IgG AbSCs in BM (Fig. 5E) were significantly increased in both frequency (2.1-fold) and number (3.4-fold). In summary, the addition of the checkpoint Ag/*APCs following WIV + Ag/*APC vaccine immunization increased anti-influenza long-term (35 dpt) AbSC and the level of IgG1 in the serum, presumably indicating better helper CD4 effector activity.

Although it was clear that adding checkpoint Ag/*APC was effective in enhancing lung CD4 memory and inducing TRM generation, we were still puzzled to see only a modest increase in CD4 memory in the spleen because in the live virus studies, almost all spleen memory generation depended on the checkpoint Ag recognition, which increased memory 20–200-fold (22, 23). We reasoned that live influenza virus infection provides a very high level of Ag throughout the checkpoint interval. Thus, we analyzed whether increasing Ag density at checkpoint (5 dpt) would result in an additional increase in CD4 memory generation. For the studies above, we used a dose of HA126–138 peptide (12.5 μM) to pulse *APC at the checkpoint (5 dpt), which is higher than doses of 1–2.5 μM per mouse routinely used in studies (40). In this study, we pulsed *APC with three peptide doses, 1.25, 12.5, and 125 μM (Fig. 6A), and added these at 5 dpt. There was a clear impact of increased peptide density, which led to an increase in memory generation. At low density of peptide (1.25 μM), Ag/*APC on 5 dpt only slightly increased the frequency of donor cells in spleen compared with no Ag/*APC addition, and there was no enhancement in the numbers of donor memory (Fig. 6B). As the Ag concentration increased, donor memory cells in spleen and lung both increased (Fig. 6B, 6C). The highest dose (125 μM) increased spleen donor memory cells 2–4-fold and lung donor memory over 6-fold (Fig. 6B). The quantity of CD69+ TRM memory donor cells in the lung at 35 dpt in the immunized mice were also highly dependent on the checkpoint Ag dose, with no significant increase at low dose (1.25 μM), 5.9-fold increase at medium dose (12.5 μM), and a striking 11.6-fold increase at high dose (125 μM) (Fig. 6D). Thus, we suggest that augmentation of donor CD4 memory requires an unanticipated high dose of checkpoint peptide on APC to drive vaccine-generated effectors to total and TRM lung memory.

To determine if increasing Ag density also promoted CD4 helper activity, we measured the level of persistent anti-PR8 IgG in serum collected at 35 dpt from the mice given checkpoint Ag/*APC at 5 dpt, made with different doses of peptide. The titers of anti-PR8 IgG increased, although modestly, from low dose (1.25 μM) to the high doses (12.5 and 125 μM) and were ∼2-fold higher than background. Of interest, there was no change from medium dose (12.5 μM) to high dose (125 μM), suggesting some impacts of checkpoint Ab may require higher Ag density on APC than others.

Viral infection, including that with PR8 influenza, usually induces a vigorous CD4 T cell response. Following IAV infection, continuing viral replication generates both high levels of virus and viral Ag presentation and continuous activation of APC and other immune cells and perhaps T and B cells by pathogen recognition. Thus, naive CD4 T cells have the opportunity to recognize Ag initially. Then again, later at the checkpoint, CD4 effectors that develop and peak at ∼6 d can potentially recognize Ag again during the defined checkpoint we have identified, during which signals from Ag presentation are required to support memory formation. The recognition of Ag on activated APC with the help of multiple costimulatory interactions (e.g., CD28–CD80/CD86 and CD27/CD70) and autocrine IL-2, all at the effector checkpoint, are critical for CD4 effectors to become memory cells (22). In this study, we show that unlike influenza infection, inactivated vaccine (WIV), although it provides the strong signals to naive CD4 T cells, does not persist to provide strong Ag presentation through the effector stage. This is true even when we add additional Ag/*APC initially, which provides IL-6 signals during cognate interaction, to enhance the initial CD4 effector generation (27). We suggest that this is a likely reason that nonreplicating vaccines, such as most of those for influenza, are often ineffective in inducing memory CD4 T cells.

Given the dynamics of Ag presentation established in this study, we postulated that CD4 effectors generated by vaccines, like those generated by live infection, require Ag presentation at the checkpoint to become memory cells. We tested this by adding checkpoint Ag/*APC to provide the missing Ag presentation. Indeed, we found that although WIV vaccine (with or without Ag/*APC) presented Ag rapidly and effectively to naive Ag-specific CD4 T cells, presentation was gone after the first few days and lead to the generation of effectors but few memory cells. When we provided Ag/*APC at 5 dpt, there was enhanced CD4 memory generation, especially in the lung and preferentially among CD62LlowCD69+ lung TRM. This strategy also resulted in a better CD4 helper activity, including increased AbSC in the BM and enhanced serum Ab. Thus, these results support the concepts that nonreplicating vaccines fail to drive the robust development of memory from CD4 effectors in part because they do not provide Ag at the effector checkpoint and that CD4 effectors generated by vaccines as well as infection must go through a checkpoint in which they again recognize high levels of cognate Ag to form the bulk of memory cells, especially those in the lung.

Lung and other TRM have been characterized by the expression of CD69, CD103 (variable), CXCR6, and lack of CCR7, by their inaccessibility to i.v. Ab injection, and finally, by their lack of circulation, determined by parabiosis experiments (3538). Recently, reports have indicated that TRM can on occasion migrate and relocalize (41), and there is evidence that repeated Ag exposure can drive some circulating TEM to become TRM (42). CD4 TRM generated by influenza infection are CD69+ but often CD103 (43), but they share many transcriptional programs with CD8 TRM and are inaccessible to i.v. Ab, and these all correlate with each other (44). Lung CD4 TRM generated by influenza play a key role in protection and also mediate inflammation (43). In this study, we have used CD69 and CXCR6 expression and inaccessibility to anti-CD4. Ab identify lung TRM.

We found a striking enhancement of lung TRM generation when we added Ag/*APC at the time coinciding with the peak of the effector response that we previously correlated with the memory checkpoint (22, 23). The low TRM proportion without checkpoint Ag/*APC confirms our hypothesis that the transient Ag presentation by inactivated vaccine is unable on its own to support TRM differentiation. CD4 effector recognition of Ag at checkpoint (5 dpt) preferentially promoted lung TRM. Although lung TRM are commonly thought to be generated best only when vaccines are i.n. administrated, our studies using i.v. introduction of Ag/*APC indicate that the i.v. route was also effective in generating substantial populations of lung CD4 memory and TRM. We suggest CD4 effectors generated by inactivated vaccines may not be as fully differentiated as those generated by infection, regardless of the immunization route, and that introducing the Ag/*APC further drives their differentiation. TRM start to accumulate from the second week of response (45). The success of Ag/*APC delivered i.v. at the checkpoint in enhancing lung TRM is compatible with the concept that Ag delivery to the tissues may be important for generating TRM, providing more effective immunity (46) because i.v. Ag/*APC present Ag in the lung (22) and shown in Fig. 2 and the i.v. route has proven to be successful in other studies (47). The enhanced lung TRM generation raises the likelihood that one key facet of checkpoint Ag recognition is to induce a multifaceted program associated with effector migration, including CXCR3 expression (48), tissue residency depending on CXCR6 (49), and long-term survival (e.g., depending on IL-2– and IL-15–dependent manner) (43, 50). We plan to determine in future studies the more global induction of gene expression on lung TRM, induced by the addition of Ag/*APC at the checkpoint.

We found that the density of Ag on Ag/*APC at checkpoint was a limiting factor for memory CD4 T cell generation. As we increased the dose of HNT peptide from 12.5 to 125 μM, a marked increase in the generation of donor memory in both the spleen and lung and of lung CD69+ TRM occurred. This suggests that high levels of checkpoint Ag density may be needed to drive the vaccine-induced effectors to memory. The marked increase in donor lung memory and significant increase in spleen as well as the increase in the concentration of peptide used to load the APC likely reveals a need for high density TCR signaling not for longer duration of the Ag/*APC presentation because the short lifespan of APC is not impacted. This suggests high levels of Ag, connoting that continuing infection is needed. We also suggest that besides transient Ag presentation at the standard 12.5 μM dose used in this study, more extended stimulation could be required during the checkpoint interval if indeed the effectors generated by the vaccine are less differentiated than those elicited by infection.

In addition to the enhanced CD4 memory, there was a modest but significantly enhanced B cell response when we added Ag/*APC to the CD4 effectors at 5 dpt. There was an increase in total IgG Ab to influenza and a comparable increase in AbSC in the BM. We suggest the increased AbSC and Ab are a secondary consequence of enhanced CD4 effector helper activity. We found the checkpoint Ag/*APC strongly upregulated the expression of ICOS, a critical costimulatory molecule for helper function, on CD4 effectors. There was also a modest upregulation of CXCR5 and Bcl-6 coexpression. We have not yet determined if this enhanced B cell response was dependent on greater development of TFH- or non-TFH–mediated help. Published studies of Ab responses in human to inactivated vaccine are correlated with the induction of ICOS+ Th cells (51, 52), and non-TFH CD4 T cells can help B cells produce anti-influenza Ab generation, as we saw in earlier studies (5). Specific anti-PR8 IgG production was only modestly enhanced by checkpoint Ag/*APC, resulting in a 2.8-fold increase in serum Ab. Meanwhile, although memory CD4 generation was obviously Ag density dependent, serum IgG levels were not further increased when we increased Ag dose from medium to high. We suggest the modest effects of Ag/*APC given at 5 dpt on Ab production may indicate we have not yet fully optimized the signals we are providing, either Ag presentation itself and/or other as yet unidentified signals at checkpoint. We know the checkpoint Ag/*APC only provides Ag recognized by CD4 for <48 h (22) compared with a viral infection that lasts much longer (in this study, more than 11 dpt) and induces a much larger Ab response. Additional signals that we postulate might be useful at the checkpoint include longer Ag presentation to CD4 effectors, simultaneous stimulation of B cells, and higher levels of PAMP (53, 54). In agreement with the possibility that some signal(s) is still lacking, we note that in the WIV vaccine model, the checkpoint Ag/*APC (even at the highest Ag density, 125 μM) resulted in a smaller increase (∼6-fold) in CD4 T cell memory compared with the increase in the live infection model in which the memory population was dramatically increased equal or >20-fold in all organs (22). We assume that after vaccination, the total amount of Ag presentation to CD4 T cells is quite limited because we show the high level of Ag was only available for <3 d, and we suspect the same would be true if PAMP was due to recognition of viral RNA, which would present only when WIV was introduced but would then decay quickly. In contrast, following live infection with PR8, viral replication continues until effector CD8 and CD4 T cells enter the lung and clear the virus and infected cells (5). Thus, postinfection, high levels of Ag presentation and other infection-induced signals must persist for an extended time at least over 10 d. We therefore suggest that the effector population after WIV vaccine has received much less stimulation (Ag presentation, costimulatory, and inflammatory signals). We plan further studies to assess the functional impact on the protection of adding each of the possible limiting signals and once optimized to determine further correlates of protection in a more vaccine-like setting. Although the Ag/*APC addition has not fully recapitulated live infection, it is clear that adding Ag/*APC at the effector checkpoint induces multiple indications of further differentiation of the cells into more specialized effectors and substantial enhancement of lung memory, especially of the TRM subsets known to be important for protection against influenza.

There have been a number of previous studies suggesting that continuous Ag presentation can enhance T cell response (5557). What is clear from our studies is that for optimum CD4 responses following PR8 infection, it is not continuous Ag that is required but rather a requirement for Ag, transiently, at a defined step in CD4 effector development (22, 23). At this checkpoint, effectors that fail to recognize Ag undergo passive cell death and those that recognize Ag can be induced to undergo multiple effector fates, one of which is transition to memory (22). The generation of memory requires autocrine IL-2 production, induction of upregulated CD25 and CD127, and downregulation of the proapoptotic protein Bim, consistent with the need for preventing default apoptosis and effector contraction by Ag-dependent survival pathways to promote memory. Most relevant in this study is that vaccine-induced effectors, exposed to Ag, costimulatory molecular ligation, and an undetermined level of inflammation for <3 d, also need the checkpoint Ag recognition signals, and they can be supplied at least in part by short-term exposure to Ag/*APC, which implies that they may be able to be readily supplied by strategies that give inactivated vaccine repeatedly at this time. Whereas infection maintains high levels of Ag, for most vaccine formulations, the levels of Ag decrease with time, consistent with our hypothesis that a lack of strong Ag presentation at the checkpoint is a major reason for the ineffectiveness of modern vaccines using highly purified proteins.

Multiple studies support the concept that the generation of more durable and universal effective vaccines for influenza must include vigorous development of CD4 memory in secondary lymphoid organs and tissue that can provide heterosubtypic CD4 T cell recall (3, 36, 58). We suggest that strong T cell memory is critical for protecting against influenza and other ssRNA viruses, including coronaviruses such as the new SARS Cov-2, because the viruses evolve rapidly and are likely to generate new strains that evade pre-existing neutralizing Ab. But in influenza outbreaks, epitopes for T cell recognition were almost always conserved (31). We suggest that vaccine strategies that provide strong Ag recognition at the CD4 effector stage checkpoint will induce more and better CD4 T cell memory generation and that inducing CD4 memory should become an integral part of the most effective immunization strategies. Although we achieved an increased generation of lung CD4 memory and a small increase in persistent Ab by the addition in this study of Ag/*APC strategy, the degree of enhancement is still less than optimal compared with the memory generation in live virus infection. We suggest that this implies we still need to better define what additional signals to further optimize strong TFH differentiation are necessary to induce the generation of high affinity long-lived AbSC and B cell memory and more extensive CD4 memory. Moreover, it remains unclear what exact approaches work best to provide Ag and ensure its presentation by activated APC in a vaccine setting in humans. It is well appreciated that the repeated addition of protein Ag can improve Ab responses (55). Adjuvants that greatly prolong Ag half-life and also activate or target APC may be needed both initially and, we suggest, at the checkpoint. Higher doses of vaccine may be needed. The route of Ag administration also deserves consideration to ensure Ag presentation occurs in sites of desired tissue residency, so i.n. aerosol delivery or installation now widely used for FluMist and over the counter allergy treatments is an attractive possibility to include for respiratory viral infection. Certainly, there may be resistance to the concept of the need for two closely spaced doses of vaccine, but if the need for yearly immunization could be reduced, we suggest it would be worth the cost and inconvenience.

Taken together, the results in this study show clearly that the WIV vaccine (enhanced with Ag/*APC) briefly but effectively presents Ag and induces effectors at the initiation of naive CD4 T cell response, suggesting that it is particularly the lack of Ag presentation, perhaps with other signals, at the checkpoint that is largely responsible for the poor efficacy of nonreplicating inactivated and component vaccines in generating strong CD4 memory and long-term immunity. Results in this study provide proof of principle that recognition of Ag by effector CD4 T cells at their fate-determining checkpoint, even those generated by nonreplicating vaccines can promote their effective transition to memory and especially the generation of total lung memory and lung TRM. This provides support for using prolonged provision of abundant Ag recognition as part of a vaccine strategy that should be able to induce better CD4 immunity and hence broader, more durable protection, especially to heterosubtypic strains.

This work was supported by grants from the National Institutes of Health, including National Institute on Aging R21AG058758 (to S.L.S.) and P01AG021600 (Project 1 to S.L.S.) and National Institute of Allergy and Infectious Diseases R21AI128606, R21AI146532, R01AI118820 (to S.L.S.), and U19AI109858 (Project 2 to S.L.S.).

The online version of this article contains supplemental material.

Abbreviations used in this article:

AbSC

Ab-secreting cell

*APC

TLR agonist-activated APC

BM

bone marrow

BMDC

BM-derived dendritic cell

*BMDC

TLR agonist-activated BMDC

dLN

draining lymph node

dpt

day posttreatment

HA

hemagglutinin

IAV

influenza A virus

i.n.

intranasal(ly)

LN

lymph node

mLN

mediastinal LN

PAMP

pathogen-associated molecular pattern

TCM

central memory cell

TEM

effector memory cell

TFH

T follicular helper

Tg

transgenic

TRM

tissue-resident memory cell

WIV

whole inactivated influenza virus

WIV + Ag/*APC

*APC along with WIV during priming.

1
Wiley
,
J. A.
,
R. J.
Hogan
,
D. L.
Woodland
,
A. G.
Harmsen
.
2001
.
Antigen-specific CD8(+) T cells persist in the upper respiratory tract following influenza virus infection.
J. Immunol.
167
:
3293
3299
.
2
Allie
,
S. R.
,
J. E.
Bradley
,
U.
Mudunuru
,
M. D.
Schultz
,
B. A.
Graf
,
F. E.
Lund
,
T. D.
Randall
.
2019
.
The establishment of resident memory B cells in the lung requires local antigen encounter.
Nat. Immunol.
20
:
97
108
.
3
Teijaro
,
J. R.
,
D.
Verhoeven
,
C. A.
Page
,
D.
Turner
,
D. L.
Farber
.
2010
.
Memory CD4 T cells direct protective responses to influenza virus in the lungs through helper-independent mechanisms.
J. Virol.
84
:
9217
9226
.
4
Wolf
,
A. I.
,
K.
Mozdzanowska
,
W. J.
Quinn
III
,
M.
Metzgar
,
K. L.
Williams
,
A. J.
Caton
,
E.
Meffre
,
R. J.
Bram
,
L. D.
Erickson
,
D.
Allman
, et al
.
2011
.
Protective antiviral antibody responses in a mouse model of influenza virus infection require TACI.
J. Clin. Invest.
121
:
3954
3964
.
5
McKinstry
,
K. K.
,
T. M.
Strutt
,
Y.
Kuang
,
D. M.
Brown
,
S.
Sell
,
R. W.
Dutton
,
S. L.
Swain
.
2012
.
Memory CD4+ T cells protect against influenza through multiple synergizing mechanisms.
J. Clin. Invest.
122
:
2847
2856
.
6
Chiu
,
C.
,
P. J.
Openshaw
.
2015
.
Antiviral B cell and T cell immunity in the lungs.
Nat. Immunol.
16
:
18
26
.
7
Eickhoff
,
C. S.
,
F. E.
Terry
,
L.
Peng
,
K. A.
Meza
,
I. G.
Sakala
,
D.
Van Aartsen
,
L.
Moise
,
W. D.
Martin
,
J.
Schriewer
,
R. M.
Buller
, et al
.
2019
.
Highly conserved influenza T cell epitopes induce broadly protective immunity.
Vaccine
37
:
5371
5381
.
8
Denison
,
M. R.
,
R. L.
Graham
,
E. F.
Donaldson
,
L. D.
Eckerle
,
R. S.
Baric
.
2011
.
Coronaviruses: an RNA proofreading machine regulates replication fidelity and diversity.
RNA Biol.
8
:
270
279
.
9
Nomaguchi
,
M.
,
N.
Doi
,
T.
Koma
,
A.
Adachi
.
2018
.
HIV-1 mutates to adapt in fluxing environments.
Microbes Infect.
20
:
610
614
.
10
Vider-Shalit
,
T.
,
M.
Almani
,
R.
Sarid
,
Y.
Louzoun
.
2009
.
The HIV hide and seek game: an immunogenomic analysis of the HIV epitope repertoire.
AIDS
23
:
1311
1318
.
11
Grifoni
,
A.
,
J.
Sidney
,
Y.
Zhang
,
R. H.
Scheuermann
,
B.
Peters
,
A.
Sette
.
2020
.
A sequence homology and bioinformatic approach can predict candidate targets for immune responses to SARS-CoV-2.
Cell Host Microbe
27
:
671
680.e2
.
12
Sallusto
,
F.
,
J.
Geginat
,
A.
Lanzavecchia
.
2004
.
Central memory and effector memory T cell subsets: function, generation, and maintenance.
Annu. Rev. Immunol.
22
:
745
763
.
13
Sallusto
,
F.
,
D.
Lenig
,
R.
Förster
,
M.
Lipp
,
A.
Lanzavecchia
.
1999
.
Two subsets of memory T lymphocytes with distinct homing potentials and effector functions.
Nature
401
:
708
712
.
14
Reinhardt
,
R. L.
,
A.
Khoruts
,
R.
Merica
,
T.
Zell
,
M. K.
Jenkins
.
2001
.
Visualizing the generation of memory CD4 T cells in the whole body.
Nature
410
:
101
105
.
15
Román
,
E.
,
E.
Miller
,
A.
Harmsen
,
J.
Wiley
,
U. H.
Von Andrian
,
G.
Huston
,
S. L.
Swain
.
2002
.
CD4 effector T cell subsets in the response to influenza: heterogeneity, migration, and function.
J. Exp. Med.
196
:
957
968
.
16
Iijima
,
N.
,
A.
Iwasaki
.
2014
.
T cell memory. A local macrophage chemokine network sustains protective tissue-resident memory CD4 T cells.
Science
346
:
93
98
.
17
Turner
,
D. L.
,
K. L.
Bickham
,
J. J.
Thome
,
C. Y.
Kim
,
F.
D’Ovidio
,
E. J.
Wherry
,
D. L.
Farber
.
2014
.
Lung niches for the generation and maintenance of tissue-resident memory T cells.
Mucosal Immunol.
7
:
501
510
.
18
Hamada
,
H.
,
M. L.
Garcia-Hernandez
,
J. B.
Reome
,
S. K.
Misra
,
T. M.
Strutt
,
K. K.
McKinstry
,
A. M.
Cooper
,
S. L.
Swain
,
R. W.
Dutton
.
2009
.
Tc17, a unique subset of CD8 T cells that can protect against lethal influenza challenge.
J. Immunol.
182
:
3469
3481
.
19
Mbawuike
,
I. N.
,
P. R.
Wyde
.
1993
.
Induction of CD8+ cytotoxic T cells by immunization with killed influenza virus and effect of cholera toxin B subunit.
Vaccine
11
:
1205
1213
.
20
Bhide
,
Y.
,
W.
Dong
,
I.
Gribonika
,
D.
Voshart
,
T.
Meijerhof
,
J.
de Vries-Idema
,
S.
Norley
,
K.
Guilfoyle
,
S.
Skeldon
,
O. G.
Engelhardt
, et al
.
2019
.
Cross-protective potential and protection-relevant immune mechanisms of whole inactivated influenza virus vaccines are determined by adjuvants and route of immunization.
Front. Immunol.
10
:
646
.
21
Krammer
,
F.
2019
.
The human antibody response to influenza A virus infection and vaccination.
Nat. Rev. Immunol.
19
:
383
397
.
22
Bautista
,
B. L.
,
P.
Devarajan
,
K. K.
McKinstry
,
T. M.
Strutt
,
A. M.
Vong
,
M. C.
Jones
,
Y.
Kuang
,
D.
Mott
,
S. L.
Swain
.
2016
.
Short-lived antigen recognition but not viral infection at a defined checkpoint programs effector CD4 T cells to become protective memory.
J. Immunol.
197
:
3936
3949
.
23
McKinstry
,
K. K.
,
T. M.
Strutt
,
B.
Bautista
,
W.
Zhang
,
Y.
Kuang
,
A. M.
Cooper
,
S. L.
Swain
.
2014
.
Effector CD4 T-cell transition to memory requires late cognate interactions that induce autocrine IL-2.
Nat. Commun.
5
:
5377
.
24
Obst
,
R.
,
H. M.
van Santen
,
R.
Melamed
,
A. O.
Kamphorst
,
C.
Benoist
,
D.
Mathis
.
2007
.
Sustained antigen presentation can promote an immunogenic T cell response, like dendritic cell activation.
Proc. Natl. Acad. Sci. USA
104
:
15460
15465
.
25
Pulendran
,
B.
,
J. Z.
Oh
,
H. I.
Nakaya
,
R.
Ravindran
,
D. A.
Kazmin
.
2013
.
Immunity to viruses: learning from successful human vaccines.
Immunol. Rev.
255
:
243
255
.
26
Schotsaert
,
M.
,
A.
García-Sastre
.
2017
.
Inactivated influenza virus vaccines: the future of TIV and QIV.
Curr. Opin. Virol.
23
:
102
106
.
27
Plotkin
,
S.
2014
.
History of vaccination.
Proc. Natl. Acad. Sci. USA
111
:
12283
12287
.
28
Jackson
,
L. A.
,
S. J.
Holmes
,
P. M.
Mendelman
,
L.
Huggins
,
I.
Cho
,
J.
Rhorer
.
1999
.
Safety of a trivalent live attenuated intranasal influenza vaccine, FluMist, administered in addition to parenteral trivalent inactivated influenza vaccine to seniors with chronic medical conditions.
Vaccine
17
:
1905
1909
.
29
Koroleva
,
M.
,
F.
Batarse
,
S.
Moritzky
,
C.
Henry
,
F.
Chaves
,
P.
Wilson
,
F.
Krammer
,
K.
Richards
,
A. J.
Sant
.
2020
.
Heterologous viral protein interactions within licensed seasonal influenza virus vaccines.
NPJ Vaccines
5
:
3
.
30
Brahmakshatriya
,
V.
,
Y.
Kuang
,
P.
Devarajan
,
J.
Xia
,
W.
Zhang
,
A. M.
Vong
,
S. L.
Swain
.
2017
.
IL-6 production by TLR-activated APC broadly enhances aged cognate CD4 helper and B cell antibody responses in vivo.
J. Immunol.
198
:
2819
2833
.
31
Jones
,
S. C.
,
V.
Brahmakshatriya
,
G.
Huston
,
J.
Dibble
,
S. L.
Swain
.
2010
.
TLR-activated dendritic cells enhance the response of aged naive CD4 T cells via an IL-6-dependent mechanism.
J. Immunol.
185
:
6783
6794
.
32
Jelley-Gibbs
,
D. M.
,
D. M.
Brown
,
J. P.
Dibble
,
L.
Haynes
,
S. M.
Eaton
,
S. L.
Swain
.
2005
.
Unexpected prolonged presentation of influenza antigens promotes CD4 T cell memory generation.
J. Exp. Med.
202
:
697
706
.
33
Anderson
,
K. G.
,
K.
Mayer-Barber
,
H.
Sung
,
L.
Beura
,
B. R.
James
,
J. J.
Taylor
,
L.
Qunaj
,
T. S.
Griffith
,
V.
Vezys
,
D. L.
Barber
,
D.
Masopust
.
2014
.
Intravascular staining for discrimination of vascular and tissue leukocytes.
Nat. Protoc.
9
:
209
222
.
34
Powell
,
T. J.
,
T.
Strutt
,
J.
Reome
,
J. A.
Hollenbaugh
,
A. D.
Roberts
,
D. L.
Woodland
,
S. L.
Swain
,
R. W.
Dutton
.
2007
.
Priming with cold-adapted influenza A does not prevent infection but elicits long-lived protection against supralethal challenge with heterosubtypic virus.
J. Immunol.
178
:
1030
1038
.
35
Mackay
,
L. K.
,
A. T.
Stock
,
J. Z.
Ma
,
C. M.
Jones
,
S. J.
Kent
,
S. N.
Mueller
,
W. R.
Heath
,
F. R.
Carbone
,
T.
Gebhardt
.
2012
.
Long-lived epithelial immunity by tissue-resident memory T (TRM) cells in the absence of persisting local antigen presentation.
Proc. Natl. Acad. Sci. USA
109
:
7037
7042
.
36
Zens
,
K. D.
,
J. K.
Chen
,
D. L.
Farber
.
2016
.
Vaccine-generated lung tissue-resident memory T cells provide heterosubtypic protection to influenza infection.
JCI Insight
1
: e85832.
37
Schenkel
,
J. M.
,
D.
Masopust
.
2014
.
Tissue-resident memory T cells.
Immunity
41
:
886
897
.
38
Kumar
,
B. V.
,
W.
Ma
,
M.
Miron
,
T.
Granot
,
R. S.
Guyer
,
D. J.
Carpenter
,
T.
Senda
,
X.
Sun
,
S. H.
Ho
,
H.
Lerner
, et al
.
2017
.
Human tissue-resident memory T cells are defined by core transcriptional and functional signatures in lymphoid and mucosal sites.
Cell Rep.
20
:
2921
2934
.
39
Schmiel
,
S. E.
,
J. A.
Yang
,
M. K.
Jenkins
,
D. L.
Mueller
.
2017
.
Cutting edge: adenosine A2a receptor signals inhibit germinal center T follicular helper cell differentiation during the primary response to vaccination.
J. Immunol.
198
:
623
628
.
40
McKinstry
,
K. K.
,
S.
Golech
,
W. H.
Lee
,
G.
Huston
,
N. P.
Weng
,
S. L.
Swain
.
2007
.
Rapid default transition of CD4 T cell effectors to functional memory cells.
J. Exp. Med.
204
:
2199
2211
.
41
Klicznik
,
M. M.
,
P. A.
Morawski
,
B.
Höllbacher
,
S. R.
Varkhande
,
S. J.
Motley
,
L.
Kuri-Cervantes
,
E.
Goodwin
,
M. D.
Rosenblum
,
S. A.
Long
,
G.
Brachtl
, et al
.
2019
.
Human CD4 + CD103 + cutaneous resident memory T cells are found in the circulation of healthy individuals.
Sci. Immunol.
4
:
eaav8995
.
42
Van Braeckel-Budimir
,
N.
,
S. M.
Varga
,
V. P.
Badovinac
,
J. T.
Harty
.
2018
.
Repeated antigen exposure extends the durability of influenza-specific lung-resident memory CD8+ T cells and heterosubtypic immunity.
Cell Rep.
24
:
3374
3382.e3
.
43
Strutt
,
T. M.
,
K.
Dhume
,
C. M.
Finn
,
J. H.
Hwang
,
C.
Castonguay
,
S. L.
Swain
,
K. K.
McKinstry
.
2018
.
IL-15 supports the generation of protective lung-resident memory CD4 T cells.
Mucosal Immunol.
11
:
668
680
.
44
Szabo
,
P. A.
,
M.
Miron
,
D. L.
Farber
.
2019
.
Location, location, location: Tissue resident memory T cells in mice and humans.
Sci. Immunol.
4
:
eaas9673
.
45
Mackay
,
L. K.
,
A.
Rahimpour
,
J. Z.
Ma
,
N.
Collins
,
A. T.
Stock
,
M. L.
Hafon
,
J.
Vega-Ramos
,
P.
Lauzurica
,
S. N.
Mueller
,
T.
Stefanovic
, et al
.
2013
.
The developmental pathway for CD103(+)CD8+ tissue-resident memory T cells of skin.
Nat. Immunol.
14
:
1294
1301
.
46
Kim
,
T. S.
,
S. A.
Gorski
,
S.
Hahn
,
K. M.
Murphy
,
T. J.
Braciale
.
2014
.
Distinct dendritic cell subsets dictate the fate decision between effector and memory CD8(+) T cell differentiation by a CD24-dependent mechanism.
Immunity
40
:
400
413
.
47
Thompson
,
E. A.
,
P. A.
Darrah
,
K. E.
Foulds
,
E.
Hoffer
,
A.
Caffrey-Carr
,
S.
Norenstedt
,
L.
Perbeck
,
R. A.
Seder
,
R. M.
Kedl
,
K.
Loré
.
2019
.
Monocytes acquire the ability to prime tissue-resident T cells via IL-10-mediated TGF-β release.
Cell Rep.
28
:
1127
1135.e4
.
48
Glennie
,
N. D.
,
V. A.
Yeramilli
,
D. P.
Beiting
,
S. W.
Volk
,
C. T.
Weaver
,
P.
Scott
.
2015
.
Skin-resident memory CD4+ T cells enhance protection against Leishmania major infection.
J. Exp. Med.
212
:
1405
1414
.
49
Wein
,
A. N.
,
S. R.
McMaster
,
S.
Takamura
,
P. R.
Dunbar
,
E. K.
Cartwright
,
S. L.
Hayward
,
D. T.
McManus
,
T.
Shimaoka
,
S.
Ueha
,
T.
Tsukui
, et al
.
2019
.
CXCR6 regulates localization of tissue-resident memory CD8 T cells to the airways.
J. Exp. Med.
216
:
2748
2762
.
50
Mackay
,
L. K.
,
E.
Wynne-Jones
,
D.
Freestone
,
D. G.
Pellicci
,
L. A.
Mielke
,
D. M.
Newman
,
A.
Braun
,
F.
Masson
,
A.
Kallies
,
G. T.
Belz
,
F. R.
Carbone
.
2015
.
T-box transcription factors combine with the cytokines TGF-β and IL-15 to control tissue-resident memory T cell fate.
Immunity
43
:
1101
1111
.
51
Bentebibel
,
S. E.
,
S.
Lopez
,
G.
Obermoser
,
N.
Schmitt
,
C.
Mueller
,
C.
Harrod
,
E.
Flano
,
A.
Mejias
,
R. A.
Albrecht
,
D.
Blankenship
, et al
.
2013
.
Induction of ICOS+CXCR3+CXCR5+ TH cells correlates with antibody responses to influenza vaccination.
Sci. Transl. Med.
5
: 176ra32.
52
Miyauchi
,
K.
,
A.
Sugimoto-Ishige
,
Y.
Harada
,
Y.
Adachi
,
Y.
Usami
,
T.
Kaji
,
K.
Inoue
,
H.
Hasegawa
,
T.
Watanabe
,
A.
Hijikata
, et al
.
2016
.
Protective neutralizing influenza antibody response in the absence of T follicular helper cells.
Nat. Immunol.
17
:
1447
1458
.
53
Turner
,
J. S.
,
Z. L.
Benet
,
I. L.
Grigorova
.
2017
.
Antigen acquisition enables newly arriving B cells to enter ongoing immunization-induced germinal centers.
J. Immunol.
199
:
1301
1307
.
54
Castiblanco
,
D. P.
,
R. W.
Maul
,
L. M.
Russell Knode
,
P. J.
Gearhart
.
2017
.
Co-stimulation of BCR and toll-like receptor 7 increases somatic hypermutation, memory B cell formation, and secondary antibody response to protein antigen.
Front. Immunol.
8
:
1833
.
55
Tam
,
H. H.
,
M. B.
Melo
,
M.
Kang
,
J. M.
Pelet
,
V. M.
Ruda
,
M. H.
Foley
,
J. K.
Hu
,
S.
Kumari
,
J.
Crampton
,
A. D.
Baldeon
, et al
.
2016
.
Sustained antigen availability during germinal center initiation enhances antibody responses to vaccination.
Proc. Natl. Acad. Sci. USA
113
:
E6639
E6648
.
56
Davies
,
B.
,
J. E.
Prier
,
C. M.
Jones
,
T.
Gebhardt
,
F. R.
Carbone
,
L. K.
Mackay
.
2017
.
Cutting edge: tissue-resident memory T cells generated by multiple immunizations or localized deposition provide enhanced immunity.
J. Immunol.
198
:
2233
2237
.
57
Wirth
,
T. C.
,
H. H.
Xue
,
D.
Rai
,
J. T.
Sabel
,
T.
Bair
,
J. T.
Harty
,
V. P.
Badovinac
.
2010
.
Repetitive antigen stimulation induces stepwise transcriptome diversification but preserves a core signature of memory CD8(+) T cell differentiation.
Immunity
33
:
128
140
.
58
Strutt
,
T. M.
,
K. K.
McKinstry
,
Y.
Kuang
,
C. M.
Finn
,
J. H.
Hwang
,
K.
Dhume
,
S.
Sell
,
S. L.
Swain
.
2016
.
Direct IL-6 signals maximize protective secondary CD4 T cell responses against influenza.
J. Immunol.
197
:
3260
3270
.

The authors have no financial conflicts of interest.

Supplementary data