|Home | About | Journals | Submit | Contact Us | Français|
In recent years, influenza viruses with pandemic potential have been a major concern worldwide. One unresolved issue is how infection or vaccination with seasonal influenza virus strains influences the ability to mount a protective immune response to novel pandemic strains. In this study, we developed a mouse model of primary and secondary influenza infection by using a widely circulating seasonal H1N1 virus and the pandemic strain of H1N1 that emerged in Mexico in 2009, and we evaluated several key issues. First, using overlapping peptide libraries encompassing the entire translated sequences of 5 major influenza virus proteins, we assessed the specificity of CD4 T cell reactivity toward epitopes conserved among H1N1 viruses or unique to the seasonal or pandemic strain by enzyme-linked immunospot (ELISpot) assays. Our data show that CD4 T cells reactive to both virus-specific and genetically conserved epitopes are elicited, allowing separate tracking of these responses. Populations of cross-reactive CD4 T cells generated from seasonal influenza infection were found to expand earlier after secondary infection with the pandemic H1N1 virus than CD4 T cell populations specific for new epitopes. Coincident with this rapid CD4 T cell response was a potentiated neutralizing-antibody response to the pandemic strain and protection from the pathological effects of infection with the pandemic virus. This protection was not dependent on CD8 T cells. Together, our results indicate that exposure to seasonal vaccines and infection elicits CD4 T cells that promote the ability of the mammalian host to mount a protective immune response to pandemic strains of influenza virus.
In the past year, as in previous years when a pandemic strain of influenza virus has emerged (19, 26, 31, 43, 45, 56), the outbreak of the influenza H1N1 virus of swine origin (14) was a major concern worldwide (reviewed in references 42, 44, and 67). For emerging pandemic influenza viruses, two critical questions need to be addressed. The first is how previous exposure to seasonal strains of virus and vaccines influences the ability to respond to the novel pandemic strain. The second issue is what components of the immune response are most critical for these effects. Recent experimental and epidemiological studies suggest that earlier exposures to distantly related seasonal viruses may have at least a partially protective effect. For example, clinical and epidemiological studies of the pandemic H1N1 virus infections worldwide suggested that rates of infection with the pandemic H1N1 2009 influenza virus differed significantly in different age groups, with children and young adults disproportionately susceptible to infection (4, 24). Depending on the study and region analyzed, individuals under the age of 25 years represented 45% to 60% of infected subjects, though the pathogenic effects of H1N1 virus infection were most pronounced in individuals more than 60 years old (4, 36). These findings, as well as recent immunological studies from our laboratory and other laboratories (11, 17, 20, 22, 25, 33, 39, 48, 51, 52, 55, 61, 62), suggest that previous encounters with vaccines or viruses provide immunological advantages and immunological memory in the population despite the “serological distance” between the hemagglutinin (HA) and neuraminidase (NA) proteins of seasonal and pandemic strains.
Although recent experimental work with ferrets and mice indicates that preexposure to a seasonal H1N1 virus can provide protective immunity to a later challenge with the 2009 H1N1 virus (27, 62), few studies have directly examined the scope or specificity of CD4 T cells that are cross-reactive for seasonal and pandemic H1N1 viruses. Understanding the specificity of CD4 T cells is essential for several reasons. First, cross-protective immunity requires that some fraction of the CD4 T cells elicited by seasonal viruses be specific for peptide epitopes that are shared by seasonal and pandemic strains. Such cross-reactive CD4 T cells, most commonly derived from highly conserved internal viral proteins, are thought to carry out several protective functions during a secondary infection, including rapid production of cytokines that can potentiate CD8 and B cell responses, direct cytolytic activity (reviewed in references 12, 37, and 38), mobilization of effectors (64), and rapid initiation of the innate antiviral response in the lung (59). Second, the ability of CD4 T cells to facilitate the production of high-affinity neutralizing antibodies may be linked to their protein specificity. Recent studies by Crotty and coworkers suggest that for large enveloped viruses, the antigen specificities of CD4 T cells and B cells must be physically contained within the same viral protein for optimal delivery of help (53). For neutralizing antibodies to influenza virus HA, this would mean that some CD4 T cells should be specific for the peptide epitopes that are genetically conserved in seasonal and pandemic virus HA proteins.
The study described here focuses on the specificity of influenza virus-specific CD4 T cells generated after infection with a seasonal strain of human H1N1 virus that was circulating widely in the United States for a decade (A/New Caledonia/20/99 [referred to below as A/New Caledonia]) and on the specificity of CD4 T cells generated after infection with the pandemic strain. We used a mouse model of primary and secondary infection with two inbred mouse strains that differ in their major histocompatibility complex (MHC) haplotypes, background genes, and susceptibilities to the pathological effects of pandemic H1N1 virus infection. We have characterized the prevalence and distribution of epitopes derived from conserved and virus-specific segments that are included in the response to live infection with these two types of influenza viruses. Using constructed peptide pools and cytokine enzyme-linked immunospot (ELISpot) assays to quantify and characterize influenza virus-specific CD4 T cells directly ex vivo, we provide evidence that even for distantly related viruses, the response to seasonal influenza virus partitions between two compartments: CD4 T cells dedicated to highly cross-reactive peptide epitopes and CD4 T cells with specificities that are unique to seasonal and pandemic influenza viruses. Secondary challenge with the pandemic virus is associated with rapid expansion of memory CD4 T cells specific for shared determinants, decreased viral titers, and accelerated neutralizing anti-influenza antibody responses to the serologically distant, novel pandemic strain.
Influenza virus A/New Caledonia/20/99 was prepared in the allantoic cavities of embryonated chicken eggs, as described previously (46). Influenza virus A/California/04/2009 E3 (referred to below as A/California; generously provided by David Topham at the University of Rochester) is known to contain alterations in the HA gene resulting in three amino acid differences from the original strain isolated from humans (27).
Female BALB/c (Ad, Ed) and A/J (Ak, Ek) mice were purchased from Jackson Laboratories (Bar Harbor, ME). The mice were maintained according to institutional guidelines in the specific-pathogen-free facility at the University of Rochester and were used at the age of 7 to 16 weeks. Mice in all groups were age matched.
Mice were anesthetized by intraperitoneal injection with 20 mg/ml of tribromoethanol at 200 to 250 μl per mouse. For primary infection, mice were infected intranasally either with 5 × 104 50% egg infective doses (EID50) of A/New Caledonia/20/99 in 30 μl phosphate-buffered saline (PBS) or with 1.5 × 104 EID50 of A/California/04/2009 in 30 μl PBS; then they were euthanized on day 10 postinfection. For rechallenge studies, mice were infected intranasally with 5 × 104 EID50 of A/New Caledonia/20/99, rested for 8 to 9 weeks, and then challenged with 1.5 × 104 EID50 of A/California/04/2009 administered intranasally in 30 μl PBS. Mice were euthanized on day 5 and day 10 postinfection, and the spleen, lungs, and mediastinal lymph nodes were excised and used as a source of CD4 T cells for in vitro assays. Serum was collected from individual mice by cardiac puncture.
The peptides used for ELISpot assays were 17-mer peptides, overlapping by 11 amino acids, that encompassed the entire sequences of the hemagglutinin (HA) and neuraminidase (NA) proteins from the A/New Caledonia/20/99 influenza virus (H1N1), the nucleocapsid (NP) and matrix (M1) proteins from the A/New York/348/2003 influenza virus (H1N1), and a nonstructural (NS1) protein from the A/New York/444/2001 (H1N1) virus, as described previously (41, 47; K. Richards, F. Chaves, and A. J. Sant, unpublished data). Peptide arrays were obtained from the NIH Biodefense and Emerging Infections Research Repository (NIAID) as follows: NR-2602 for the HA protein and NR-2606 for the NA protein of influenza virus A/New Caledonia/20/1999, NR-2611 for the NP protein and NR-2613 for the M1 protein of influenza virus A/New York/348/2003, and NR-2612 for the influenza virus A/New York/444/2001 NS1 protein. Amino acid sequences for the NP, M1, and NS1 proteins are highly conserved between these viruses and the A/New Caledonia virus. Peptides were reconstituted at 10 mM in PBS with or without added dimethyl sulfoxide to increase the solubility of hydrophobic peptides and 1 mM dithiothreitol (DTT) for cysteine-containing peptides. Single peptides were used at a final concentration of 10 μM and peptide pools at a final concentration of 2 μM for each peptide in the pool. Peptides unique to A/New Caledonia/20/99 or A/California/04/2009 were synthesized in our own facility using an Apex 396 system (AAPPTec, Inc., Louisville, KY). Peptides were precipitated 3 times in ether to remove organic solvents and impurities before use.
Mice were euthanized at the indicated times postinfection, and single-cell suspensions from the spleen and draining mediastinal lymph nodes were collected and processed in culture medium (Dulbecco's modified Eagle medium [DMEM] plus 10% fetal bovine serum [FBS]). Splenocytes were depleted of red blood cells by treatment with ACK lysis buffer (0.15 M NH4Cl, 1 mM KHCO3, 0.1 mM Na2 EDTA in H2O [pH 7.2 to 7.4]) for 5 min at room temperature; then they were washed and enriched for CD4 T cells by antibody- and complement-mediated lysis. Monoclonal antibody (MAb)-producing cell lines obtained from the American Type Culture Collection included 3.155 (anti-CD8), RA3/3A1/6.1 (anti-B220), and M5/114 (anti-I-Ad, anti-I-Ed,k) for BALB/c mice and 14-4-4S (anti-I-Ek) for A/J mice. After incubation for 30 min at 4°C in the pool of monoclonal antibodies at 2 × 107 cells/ml, cells were washed, resuspended in a 1:20 dilution of guinea pig complement (Low-Tox-M; Cedarlane Laboratories, Burlington, Ontario, Canada) at a concentration of 2 × 107 cells/ml, and incubated at 37°C for 30 min. Viable cells were purified by density gradient centrifugation with Lympholyte-M (Cedarlane Laboratories). For antigen-presenting cells (APC), splenocytes from naïve mice were depleted of T cells by incubation at 2 × 107 cells/ml with a monoclonal antibody supernatant from the J1j.10 cell line (anti-Thy-1.2) and complement treatment. ELISpot assays were performed as described previously (41, 46). Briefly, 96-well MultiScreen HTS filter plates (Millipore, Billerica, MA) were coated with 2 μg/ml of a purified rat antibody against mouse gamma interferon (IFN-γ) (clone AN18) or interleukin 2 (IL-2) (clone JES6-1A12) (BD Biosciences, San Jose, CA) in PBS at room temperature for 2 h or overnight at 4°C. Plates were washed and blocked for 1 h at room temperature with culture medium. A total of 300,000 CD4 T cells were cocultured with 500,000 syngeneic T cell-depleted splenocytes as APC and recall peptides in a total volume of 200 μl for at least 16 h at 37°C under 5% CO2. Plates were washed and developed as described previously (46) using Vector Blue substrate kit III (Vector Laboratories, Burlingame, CA). Substrate was prepared in 100 mM Tris, pH 8.2. After drying, the plates were processed for spot counting using an ImmunoSpot reader, series 2A, with ImmunoSpot software, version 3.2 (Cellular Technology Ltd., Cleveland, OH). Data were calculated and presented as spots per million CD4 T cells, with background values subtracted.
The A/New Caledonia/20/99 and A/California/04/2009 E3 viruses were used to infect the T cell-depleted splenocytes used as APC. APC were washed in serum-free medium and were incubated with the virus at different multiplicities of infection (MOI) under serum-free conditions for 1 h at 37°C. After incubation, cells were washed and resuspended in a culture medium containing 10% serum. A total of 500,000 infected APC were cocultured with 300,000 CD4 T cells in a total volume of 200 μl in a 96-well MultiScreen HTS filter plate for at least 16 h at 37°C under 5% CO2.
Overnight monolayer cultures of the Madin-Darby canine kidney (MDCK) cell line were infected with the A/New Caledonia/20/99 virus, and total RNA was isolated by TRIzol (Invitrogen, Carlsbad, CA). The cDNA was isolated from RNA by reverse transcription-PCR (RT-PCR) using reverse transcriptase with random primers (AffinityScript QPCR cDNA synthesis kit; Stratagene, La Jolla, CA). The NP gene was generated by PCR performed with Pfu Turbo polymerase (Stratagene) using forward primer 5′-AAAAAA CATATG GCG TCC CAA GGC ACC AAAC-3′ and reverse primer 5′-TTTTTT CTCGAG TTA ATT GTC GTA CTC CTC TGC ATT GTC TCCG-3′ (Integrated DNA Technologies [IDT], Coralville, IA), which also incorporated XhoI and NdeI restriction sites, and was cloned into the pET28a expression vector (Novagen) with T4 DNA ligase (Fermentas). After sequence confirmation, the recombinant plasmid was transformed into competent cells [One Shot BL21(DE3)pLysS; Invitrogen] for expression of NP protein. The NP protein was purified using Ni2+ column chromatography with a histidine binding purification kit (Novagen). Fractions containing NP protein were pooled and dialyzed to a low-salt buffer (50 mM Tris-Cl [pH 7.5], 1 mM EDTA, 200 mM NaCl, 50% glycerol, 1 mM DTT), and the NP protein was stored in small aliquots at −80°C until use.
Sera were collected from individual mice, and the presence of NP-specific antibodies was determined by an enzyme-linked immunosorbent assay (ELISA). Ninety-six-well polystyrene flat-bottom plates (Costar) were coated overnight at 4°C with 200 ng/100 μl of purified NP protein per well. Wells were rinsed with wash buffer (0.05% Tween 20 [Sigma-Aldrich] in PBS) and were then incubated with blocking buffer (3% bovine serum albumin [BSA] in PBS) for 1 h at room temperature. The blocking buffer was removed, and diluted serum samples (in 0.5% BSA-PBS) were added to the plates and incubated for 2 to 3 h at room temperature. The wells were washed with PBS and were incubated for 1 h at room temperature with 100 μl/well alkaline phosphatase-conjugated goat anti-mouse secondary antibody (SouthernBiotech, Birmingham, AL) diluted in 1% BSA-PBS at a 1/1,000 dilution. One p-nitrophenyl phosphate substrate tablet (5 mg/tablet; Sigma) was dissolved in 15 ml diethanolamine substrate (DSB) buffer, pH 9.8. Subsequently, wells were washed with wash buffer, and 100 μl of substrate per well was added and was developed at room temperature (25 to 40 min). Absorbance at 405 nm was read using SoftMax Pro software and a VMax plate reader.
Mice were euthanized, and lung tissue was collected and stored at −80°C prior to assay. Lung tissue was first homogenized in a cold Kontes Tenbroeck tissue grinder (Kimble Chase, Vineland, NJ) in 1 ml of sterile cold Hanks balanced salt solution (HBSS) buffer (GIBCO) until the tissue was completely crushed and then incubated on ice for 30 min. The supernatant was collected and was serially diluted in PBS plus 0.1% gentamicin (GIBCO). Overnight monolayer cultures of MDCK cells in 6-well plates were infected with 0.5 ml/well of diluted samples of virus for 1 h at 37°C under 5% CO2 to allow viral uptake and were then washed with HBSS. Equal volumes of 1.2% SeaKem agarose (Cambrex, Rockland, ME) and 2× minimum essential medium (GIBCO) supplemented with 0.4% 1× phenol red and 0.5 μg/ml tosylsulfonyl phenylalanyl chloromethyl ketone (TPCK) trypsin (Worthington, Lakewood, NJ) were mixed and added to infected-cell monolayers at 2 ml/well. The plates were incubated at 37°C under 5% CO2 for 48 h and were stained with 1% crystal violet in 10% normal buffered formalin (NBF) for 2 h to overnight. After removal of the stain and agar, plates were washed and dried, and influenza virus plaques were counted.
Sera were first heat inactivated at 56°C for 30 min and then serially diluted 2-fold in serum-free DMEM. Two hundred tissue culture infective doses (TCID50) of virus particles were added to the serum dilutions, and plates were incubated at 37°C under 5% CO2 for 1 h. The serum-virus mixture was transferred to 96-well plates containing monolayers of MDCK cells and was incubated at 37°C under 5% CO2 for 1 h. Unbound virus was removed by washing with serum-free medium. Plates were incubated with 200 μl serum-free medium containing 0.5 μg/ml TPCK trypsin at 37°C under 5% CO2 for 4 days; then they were stained with 1% crystal violet in 10% NBF for 1 h. The microneutralization titers were measured as the reciprocal of the highest serum dilution at which all of the culture wells were negative for cytopathic effect.
An anti-CD8 antibody (clone 2.43; BioXCell, West Lebanon, NH) was used for CD8 depletion in vivo. Two hundred micrograms of an anti-CD8 or isotype control IgG2b (BioXCell) antibody was injected intraperitoneally 2 days prior to infection, on the day of infection, and 2 days after infection. After infection, mice were observed daily for weight loss and survival.
Statistical significance was evaluated using an unpaired Student t test. A P value of <0.05 was considered statistically significant.
To address how previous exposure to a widely circulating seasonal human H1N1 virus influences the impact of confrontation with the pandemic strain, we first evaluated the overall pathogenicity of the pandemic virus in mice. Two different inbred strains of mice (BALB/c and A/J) were tested, because inbred strains of mice have been shown to differ in their susceptibilities to infection with H1N1 viruses (1, 58) and because our laboratory has shown that the specificity of the CD4 T cell responses can differ dramatically with the MHC alleles expressed (41). A/J and BALB/c mice are of different genetic backgrounds and MHC haplotypes (H-2k and H-2d, respectively). Both strains express I-A and I-E molecules, analogous to the HLA-DQ and HLA-DR class II molecules that are expressed in humans; thus, each strain offers 2 different peptide-presenting MHC class II molecules to the responding CD4 T cells. Mice were infected with the A/California virus using the natural intranasal route of infection, and their weight and health were monitored over time. Figure 1 shows that A/J mice displayed considerably more virus-induced clinical pathology in response to infection with the pandemic A/California H1N1 virus than did BALB/c mice, whether weight loss over time (Fig. 1A) or the percentage of mortality after infection (Fig. 1B) was analyzed. This result is consistent with recently published work that has evaluated differences in the pathogenicity of H1N1 and other influenza viruses in inbred mice (1, 6, 58).
Because A/J mice showed readily detectable susceptibility to the A/California virus, they were used for the protection experiments. One cohort of mice was infected with the seasonal A/New Caledonia influenza virus, and after virus clearance and contraction of the primary response (8 weeks postinfection), they were challenged with the pandemic H1N1 A/California virus. A cohort of mice not subjected to the original infection with the A/New Caledonia virus but infected with the A/California virus was analyzed in parallel in order to directly compare primary A/California virus infection with the secondary challenge. As can be seen in Fig. 2, previous infection with, and clearance of, the seasonal H1N1 A/New Caledonia strain protected mice against the pathological effects of A/California virus infection, with only a modest (<10%) weight loss (Fig. 2A). Naïve A/J mice infected with the A/California virus without any preexposure to the seasonal strain displayed significant weight loss (30% by day 7 postinfection). The mice previously exposed to the seasonal strain also cleared the pandemic virus more rapidly from the lung (Fig. 2B). Most importantly, there was no pandemic virus-induced lethality in the mice that had established immunological memory from the seasonal H1N1 strain (Fig. 2C). Therefore, a single infection with, and clearance of, a seasonal H1N1 virus can provide immunological protection from a lethal dose of the distantly related pandemic strain. This conclusion is in agreement with recently published work using different H1N1 virus strains in ferrets (39) and C57BL/6 mice (27, 39) and with recent studies showing the ability of the commercial FluMist vaccine to protect C3H and C57BL/6 mice from future infection with the pandemic H1N1 virus (62).
It is known that many CD8 T cells recognize epitopes contained within conserved influenza virus proteins and that CD8 T cells contribute significantly to a protective immune response to influenza (reviewed in references 10, 12, 15, and 65). Although we have not mapped CD8 T cell specificity in the A/J strain, it is likely that after infection with the seasonal A/New Caledonia strain of influenza virus, these mice have memory CD8 T cells that can cross-react with the epitopes generated by the pandemic strain, killing infected cells in the lung and thus providing a protective effect. To determine whether such cross-reactive CD8 T cells are the cells primarily responsible for the protection observed in the challenge experiments, mice that had been infected with the seasonal virus 8 weeks previously were depleted of CD8 T cells by MAb treatment in vivo at 2 days before infection, on the day of infection, and 2 days after the challenge with the A/California virus. Control mice were treated with an isotype control antibody. No CD8 T cells were detectable in anti-CD8-treated mice even weeks after this antibody administration (data not shown). Challenged control or CD8-depleted mice were monitored for weight loss and virus-induced morbidity for 14 days after challenge with the pandemic H1N1 influenza virus. As can be seen from Fig. 3, immunological protection from virus-induced pathology and death from the A/California virus infection, although slightly diminished, was not lost upon depletion of CD8 T cells prior to and during the secondary virus challenge. All mice in the CD8-depleted group recovered from the challenge with a lethal dose of the A/California virus. This suggests that other components of the immune system, including the CD4 memory cells, are sufficient to protect from a pathogenic challenge with the pandemic strain. The remainder of this study therefore focused on the specificity of the CD4 T cell repertoire elicited by seasonal and pandemic H1N1 viruses and its potential role in facilitating antibody responses.
In order to investigate the role that CD4 T cells might play in secondary infections with the pandemic strain after priming with the seasonal strain, we used the scheme outlined in Fig. 4 A to analyze CD4 T cell memory from the primary response to the seasonal strain and the boosting that took place in the challenged animals at an early time point postchallenge. Cytokine ELISpot assays were performed with CD4 T cells at 8 weeks after infection with the A/New Caledonia virus, and at day 5 after challenge with the pandemic strain, using pools of overlapping peptides representing the entire sequences of 5 major viral proteins (HA, NA, NP, M1, and NS1). In parallel, long-term memory in the mice exposed to the seasonal strain of H1N1 virus was evaluated, as was the primary CD4 T cell response to the pandemic virus. Figure 4B shows the results of these analyses. Previous exposure of the immune system to the seasonal strain of H1N1 virus leads to a limited but detectable persistent memory CD4 T cell population (hatched bars) that is influenza virus peptide specific and broadly reactive to many viral proteins. The mice previously infected with the seasonal strain and then challenged with the pandemic strain (filled bars) showed rapid expansion of CD4 T cells in response to the influenza virus-derived peptides corresponding to each of the viral proteins tested. The response of CD4 T cells isolated from mice previously infected with the seasonal strain was much more robust than the response to primary infection with the A/California virus (open bars) at this early time point. Gamma interferon production by the CD4 T cells isolated from previously primed mice was somewhat enriched. At later time points (not shown), CD4 T cells from both groups of mice showed comparable levels of virus-specific reactivity. This result in the secondary challenge experiments is consistent with data showing that memory cells are more quickly recruited for the response with cytokine production, are more sensitive to antigen than naïve cells, and are less dependent on costimulation (5, 35, 49), likely through their enhanced T cell receptor (TcR) signaling capacity (reviewed in reference 18).
We next investigated the fine specificity of the CD4 T cells elicited by seasonal and pandemic H1N1 viruses. Although some individual epitopes that are shared by seasonal and pandemic strains have been defined by our laboratory and other laboratories, there has been no published work on how the CD4 repertoire elicited by seasonal virus infection partitions its specificities between cross-reactive and virus-specific peptides. In order to evaluate this, the sequences of the peptides in overlapping peptide sets representing the viral proteins expressed by these two viruses were compared and assigned to two groups, as shown in Tables S1 and S2 in the supplemental material. One subset of peptides (“conserved”) was composed of sequences that either were identical for the pandemic and seasonal strains or differed by only a single amino acid (see Table S1). These were considered potential shared and cross-reactive epitopes. Peptides with more than a single substitution between A/California and A/New Caledonia were considered potentially unique but were further evaluated as substitutions that we expected were likely to change CD4 T cell recognition. Peptides differing by more than three amino acids, and frequently by more than five, were considered “unique” (see Table S2 in the supplemental material). In studies for which results are not shown, these peptides were further analyzed empirically and iteratively for cross-reactivity, and any peptides that were found to be cross-reactive (i.e., stimulatory when CD4 T cells from mice infected with the alternative virus were tested) were eliminated from the peptide pools (unpublished data). Peptides of each category identified by this scheme were pooled into protein-specific pools. Tables S1 and S2 show the compositions of the pools of peptides; it can be seen that the total number of peptides in each pool differs depending on the size and degree of sequence conservation in each protein. For example, NP peptides were enriched in the “conserved” category, while HA-derived peptides were more highly represented in the unique pools. In order to draw more generalizable conclusions about the distribution of CD4 T cell specificities among the conserved versus virus-unique epitopes, both A/J and BALB/c mice were studied. Groups of A/J or BALB/c mice were infected with the A/New Caledonia/20/99 or A/California/04/09 virus, and 10 days postinfection, CD4 T cells were purified. Peptide-specific responses using the pools of peptides were evaluated using an ELISpot assay in which IFN-γ-producing cells specific for the different peptide pools were quantified.
As can be seen in Fig. 5, both A/J and BALB/c mice mounted robust CD4 T cell responses to the genetically conserved peptides when either the seasonal A/New Caledonia virus or the pandemic A/California virus was used to infect the mice. The seasonal A/New Caledonia strain elicited somewhat more peptide-reactive CD4 T cells than the pandemic strain at the time point tested. The protein-specific hierarchy differed between the two strains, as would be expected for strains with differing MHC class II molecules that select different peptides for presentation to CD4 T cells. For example, CD4 T cells isolated from A/J mice responded preferentially to conserved peptides from HA and NP, with low numbers of CD4 T cells responding to M1-derived peptides, while CD4 T cells from BALB/c mice had the greatest reactivity to conserved peptides from NP, with few T cells elicited by NS1-derived peptides. Importantly, both strains of mice produced CD4 T cells specific for genetically conserved peptides from HA and NA, which may be the most potent for facilitating the production of high-affinity, isotype-switched antibodies (53).
Figure 6 shows the CD4 T cell responses of mice primed with the different viruses and tested for reactivity to “unique” peptide pools constructed with either A/New Caledonia-derived peptides or A/California-derived peptides. These experiments revealed that substantial numbers of CD4 T cells elicited by both strains are, in fact, virus specific. Both A/J and BALB/c mice produced a subset of CD4 T cells that were selectively reactive to the virus strains that elicited them. For each strain, HA and NA were a major focus of the virus-specific responses, a result consistent with the overall genetic diversity of HA and NA and their genetic sources in the pandemic strain, classical and Eurasian swine influenza viruses, respectively (21, 42, 57, 67). NP-reactive CD4 T cells specific for each virus were also abundant in the response. Importantly, these studies also point out that influenza virus proteins offer many potentially unique peptide epitopes that can be used to distinguish CD4 T cells elicited by seasonal and pandemic H1N1 viruses. We did note some residual cross-reactivity between HA and NP peptides from the pandemic A/California and seasonal A/New Caledonia strains that had been selected as “unique” when CD4 T cells from the draining lymph nodes of BALB/c mice were analyzed. It should be possible to identify those cross-reactive peptides and remove them from the pool by testing them individually or in smaller pools, so that a truly “virus-specific” peptide pool can be generated for future functional and phenotypic studies in both mice and humans.
After finding that the primary CD4 T cell response elicited by both H1N1 viruses partitions between reactivity to conserved and virus-specific epitopes, we sought to examine the specificity of the rapidly boosted CD4 T cells identified in the experiments for which results are shown in Fig. 5. To evaluate peptide specificity, mice were infected with the seasonal A/New Caledonia strain, rested for 8 weeks, and then challenged by infection with the pandemic A/California strain. Control groups of mice represented the 8-week memory population elicited by the seasonal A/New Caledonia strain and mice infected only with the A/California strain. The CD4 T cells were isolated either at the 8-week memory time point or on day 5 postinfection and were tested for reactivity with conserved versus unique peptide pools using IFN-γ and IL-2 ELISpot assays as a readout of peptide-reactive cells (Fig. 7). As expected, by far the most dramatically influenza virus reactive cells at this early time point were those CD4 T cells specific for the highly conserved epitopes. The total number of cytokine-producing cells in the primed animals exceeded that in the primary response by more than 20-fold. This result shows that CD4 T cells specific for genetically conserved regions of each influenza virus protein are rapidly recalled when the pandemic virus is encountered.
The findings that many CD4 T cells were specific for peptide epitopes shared by the seasonal and pandemic strains and that previous exposure to the seasonal H1N1 virus protects mice from lethal infection with the pandemic strain raised the issue of how CD4 memory cells might facilitate protective responses to the pandemic strain. CD4 T cells have been both speculated and demonstrated to play multiple roles in protective immunity to influenza, including the mobilization of effectors to the lung, the production of cytokines, and direct cytotoxic effects on infected host cells (reviewed in references 37 and 60). Perhaps the most widely accepted role for CD4 T cells in protective immunity to influenza is facilitation of the production of high-affinity neutralizing antibodies. We wanted to determine whether rapid expansion of memory CD4 T cells upon secondary infection might be associated with accelerated antibody production, which may contribute to protection. To evaluate this issue, sera were collected from mice infected only with the pandemic strain and from mice that had been preexposed to a seasonal H1N1 virus before challenge with the pandemic strain. Because cross-reactive serum antibody to NP remained at high levels in mice exposed to A/New Caledonia (data not shown), HA-specific antibody production was evaluated using a microneutralization assay. Control experiments (Fig. 8) showed no detectable cross-reactivity of neutralizing antibodies elicited by A/New Caledonia infection with A/California viruses. When sera from mice preexposed to A/New Caledonia and then challenged with the pandemic A/California strain were analyzed, enhancement of the kinetics of production of neutralizing antibody to A/California was observed, despite the apparent lack of B cell priming for A/California HA by infection with the A/New Caledonia virus. This result suggests that preexposure of the animals to A/New Caledonia potentiated the subsequent antibody response to A/California by priming cross-reactive CD4 T cells that could be recalled by the A/California infection. It is significant that this boost in antibody production occurs early, at a time when it might be available to neutralize pandemic virus progeny during viral replication in the lung, thus attenuating the virus life cycle.
Influenza pandemics are very difficult to anticipate, both with regard to timing and with regard to species origin, and can be devastating in their impact. The unpredictability of the emergence of new pandemic strains of influenza virus in human populations is a particularly significant concern at this time, because new influenza vaccines are difficult to produce with current technology and generally require at least 6 months to prepare and test before distribution (reviewed in references 23, 34, and 66). Accordingly, there is great interest in understanding if and how the encounter with seasonal viruses and vaccines influences the ability of the immune system to respond to novel pandemic strains of influenza virus. In the study reported here, we have specifically evaluated the ability of a previously widely circulating seasonal strain of influenza A H1N1 virus to potentiate responses to the novel pandemic strain of H1N1 that arose in Mexico in the spring of 2009. In evaluating this issue, we examined two inbred mouse strains that express different MHC molecules and background genes for their CD4 T cell reactivities toward epitopes conserved in seasonal and pandemic influenza viruses, the impact of CD4 memory on B cell responses to the pandemic strain, and the ability of the encounter with a seasonal strain of influenza virus to protect against the pathological effects of the pandemic strain.
A number of important results were revealed by these studies. First, our studies revealed that different inbred mouse strains are differentially susceptible to the pathological consequences of infection with the H1N1 strain of the virus: A/J mice are more susceptible than BALB/c and C57BL/10 mice (not shown). Other studies have shown that the mouse genetic background can dramatically impact susceptibility to influenza A virus strains (1, 6, 7, 58). In one particularly systematic study of protective immunity to A/Puerto Rico/8/34 (A/PR8) (58), a mouse-adapted laboratory strain of influenza virus, A/J and DBA/2 mice displayed the most susceptibility to infection, while C57BL/6, BALB/c, CBA, and FVB/NJ mice were among the more resistant strains and ultimately recovered from infection. This study suggested that the genetics of susceptibility are complex and involve multiple stages of influenza virus replication, the innate response, and the adaptive immune response. In the BALB/c mice that we have studied for this report, the activity of CD8 T cells does not appear to be responsible for the resistance to infection, since removal of these cells by antibody-mediated depletion did not render the mice more susceptible to the pathological effects of virus infection, such as weight loss and virus-induced mortality (data not shown). This finding is in agreement with recent studies showing that exacerbated inflammatory responses often characterize mouse strains highly susceptible to influenza infection (6, 7, 58). Studies on human subjects have also suggested that susceptibility to influenza is shared by close relatives and is thus a heritable property (2, 29). In our studies, although we have not determined the mechanisms that underlie the rapid weight loss induced by A/California/04/09 infection, we have found that mice of the most susceptible strain, A/J, become highly resistant to virus-induced weight loss and mortality if they are preexposed to a nonpathogenic seasonal strain of H1N1 virus, suggesting that immunological memory overrides the natural susceptibility of this mouse strain to the pandemic virus. The results of CD8 depletion experiments indicate that this protection does not depend on CD8 T cells, although it is likely that in normal animals or humans, CD8 T cells contribute significantly to protection. It is likely that memory CD4 T cells can contribute to protection in the secondary challenge by multiple mechanisms. As discussed above, they may facilitate early production of neutralizing antibodies and are likely to provide a direct effector function in response to influenza (reviewed in references 37 and 60). Recent studies (59) suggest that memory CD4 T cells can induce rapid production of a number of innate inflammatory cytokines and chemokines within the lungs of influenza-infected animals, which may dramatically reduce early viral replication and facilitate the recruitment of other effector cells (64). There is also evidence for direct cytotoxic effects of CD4 T cells in response to influenza (8, 63). Memory CD4 T cells of the appropriate specificity and phenotype are likely to contribute on multiple levels to increased resistance to future challenge with pathogenic strains of influenza virus and to contribute to more-effective responses to vaccination.
In examining the peptide specificity of the elicited CD4 T cells, we found that a surprisingly large fraction of the CD4 T cell response was dedicated to genetically conserved peptides that either were identical in the seasonal and pandemic strains or differed by only a single amino acid. Significantly, reactivity to conserved peptides, which are likely to be recalled upon challenge with the pandemic strain, was not limited to the highly conserved NP, NS1, and M1 proteins but was also readily apparent for conserved segments of HA and NA. Recent studies on responses to vaccinia virus suggest that CD4 T cell help to antigen-specific B cells may be best conveyed by T cells of the same protein specificity as the antigen-specific B cells (53). This “linked” CD4 T cell help for B cells is probably due to the fact that for some viruses, the antigen internalized through the immunoglobulin receptor in peripheral lymphoid tissue may be an isolated viral protein rather than an infectious virus particle, thus leading to the display of peptides derived from that single protein on the cell surface for recruitment of cognate CD4 T cell help. If this is true for influenza virus, the memory HA- and NA-specific CD4 T cells we have detected here that are reactive to conserved segments of HA and NA may be particularly important in facilitating a protective antibody response to the HA and NA proteins. Our studies revealed that mice previously exposed to the seasonal strain displayed an accelerated CD4 T cell response that was readily detectable at 5 days after infection with the pandemic strain, much earlier than that observed in the primary response. Although the pools of peptides tested in the secondary challenge were unselected and were composed of all viral peptides, it is likely that most of the rapidly responding CD4 T cells were specific for conserved epitopes and thus represented the recall of memory cells. Our preliminary experiments (not shown) using individual peptides representing either new or conserved epitopes suggest that memory CD4 T cells are preferentially expanded in the recall response. Although our experiments described here have shown preferential reactivity to cross-reactive CD4 T cell epitopes during secondary responses to related influenza viruses, they have not explicitly evaluated the phenomenon of “original antigenic sin”. This phenomenon was described many decades ago to explain preferences in antibody responses to influenza (13) and has been studied intermittently by others since, to examine biases in antibody and T cell responses in challenge experiments toward the strain of influenza virus that was originally encountered (28, 30, 32, 40, 68). Although this is a potentially important complication of “preemptive” vaccination for protection from novel pandemic influenza virus strains, in order to address this issue appropriately, one would need to compare carefully both the affinity and the fine specificity of CD4 T cells elicited in the primary response and the secondary challenge, which has not been done here.
Additionally, and importantly, mice preexposed to a seasonal influenza virus displayed accelerated production of neutralizing-antibody responses to the new challenge pandemic strain, despite the apparent lack of serological cross-reactivity between the HA molecules expressed by the two viruses. The combined CD4 T cell data and serological analysis suggest that an expanded population of primed CD4 T cells can facilitate new B cell responses that have not been encountered before. According to this model, there may be limiting numbers of CD4 T cells capable of providing help during T-B cell interactions, as has recently been suggested by adoptive transfer experiments (50). In addition to providing help during challenge with live infection by the novel virus, helper cells in the memory population of CD4 T cells specific for HA and NA may be critical for “dose-sparing” efforts that are needed for rapid deployment of limited doses of vaccines to novel influenza virus strains (reviewed in references 3, 9, 16, and 54). The relative contribution of this early antibody production to protection during the secondary challenge is not known at this time, but the local production of neutralizing antibody may help limit viral loads during replication in the lung.
The studies reported here also revealed that the CD4 T cell responses elicited by each virus include specificities unique to the respective strain. This result indicates that responses to the pandemic virus can be distinguished from responses derived from long-term memory. This will allow studies of subclinical encounters with this pandemic strain and analyses of unique phenotypic characteristics of CD4 T cells elicited by this virus. Recent studies on CD8 T cell differentiation suggest that the gene expression patterns and functionality of T cells continue to evolve with each restimulation (69), suggesting that CD4 T cells specific for epitopes that are boosted repeatedly through life, such as the conserved epitopes, may be quite distinct from those specific for novel epitopes that have expanded only once, and it will be important to determine whether this is the case for responses to influenza. At this time, it is not clear whether the most “experienced” memory CD4 T cells are the most useful for facilitating antibody responses or for contributing to other, less well characterized effector functions of CD4 T cells during the immune response to influenza. Dissection of this issue will be facilitated by the availability of well-characterized peptide reagents that allow distinction between novel and long-term memory CD4 T cells. Additionally, if the CD4 T cells elicited by this particularly novel virus display unique functional characteristics, the use of these peptide reagents would allow phenotypic characterization of the pandemic-virus-specific CD4 T cells.
In conclusion, we have found that CD4 T cells displaying broad peptide and antigen specificity are elicited by seasonal viruses and that these CD4 T cells are specific both for epitopes that are shared with the pandemic strain and for epitopes unique to one of the strains. The genetically conserved epitopes are found within all of the viral proteins tested, and CD4 T cells specific for these epitopes show rapid recall responses during a secondary challenge. This recall response is associated with accelerated neutralizing-antibody production and protection from lethal infection with the pandemic H1N1 strain of influenza virus. In humans in whom multiple class II molecules are expressed, we would expect an even broader repertoire of CD4 T cells and a correspondingly broader repertoire of influenza virus-specific cells expanded by each encounter with seasonal vaccines or viruses. A substantial fraction of the immune responses to seasonal and pandemic influenza viruses will be cross-reactive, and thus, a seasonal influenza virus can elicit cross-reactive memory cells that can be mobilized for different roles in protection against infection with pandemic strains, or that can be used to facilitate rapid antibody responses to novel vaccine candidates.
This work was supported by grants HHSN266200700008C and AI51542 to A. J. Sant from the National Institutes of Health.
We thank John Treanor for comments on the manuscript and Scott Leddon and Katherine Richards for editorial assistance in the preparation of the manuscript. We also thank Francisco Chaves for the synthesis and preparation of peptides for these studies.
†Supplemental material for this article may be found at http://jvi.asm.org/.
Published ahead of print on 5 October 2011.