Induction of CD8 T Cell Heterologous Protection by a Single Dose of Single-Cycle Infectious Influenza Virus

ABSTRACT

The effector functions of specific CD8 T cells are crucial in mediating influenza heterologous protection. However, new approaches for influenza vaccines that can trigger effective CD8 T cell responses have not been extensively explored. We report here the generation of single-cycle infectious influenza virus that lacks a functional hemagglutinin (HA) gene on an X31 genetic background and demonstrate its potential for triggering protective CD8 T cell immunity against heterologous influenza virus challenge. In vitro, X31-sciIV can infect MDCK cells, but infectious virions are not produced unless HA is transcomplemented. In vivo, intranasal immunization with X31-sciIV does not cause any clinical symptoms in mice but generates influenza-specific CD8 T cells in lymphoid (mediastinal lymph nodes and spleen) and nonlymphoid tissues, including lung and bronchoalveolar lavage fluid, as measured by H2-Db NP366 and PA224 tetramer staining. In addition, a significant proportion of X31-sciIV-induced antigen-specific respiratory CD8 T cells expressed VLA-1, a marker that is associated with heterologous influenza protection. Further, these influenza-specific CD8 T cells produce antiviral cytokines when stimulated with NP366 and PA224 peptides, indicating that CD8 T cells triggered by X31-sciIV are functional. When challenged with a lethal dose of heterologous PR8 virus, X31-sciIV-primed mice were fully protected from death. However, when CD8 T cells were depleted after priming or before priming, mice could not effectively control virus replication or survive the lethal challenge, indicating that X31-sciIV-induced memory CD8 T cells mediate the heterologous protection. Thus, our results demonstrate the potential for sciIV as a CD8 T cell-inducing vaccine.

IMPORTANCE One of the challenges for influenza prevention is the existence of multiple influenza virus subtypes and variants and the fact that new strains can emerge yearly. Numerous studies have indicated that the effector functions of specific CD8 T cells are crucial in mediating influenza heterologous protection. However, influenza vaccines that can trigger effective CD8 T cell responses for heterologous protection have not been developed. We report here the generation of an X31 (H3N2) virus-derived single-cycle infectious influenza virus, X31-sciIV. A one-dose immunization with X31-sciIV is capable of inducing functional influenza virus-specific CD8 T cells that can be recruited into respiratory tissues and provide protection against lethal heterologous challenge. Without these cells, protection against lethal challenge was essentially lost. Our data indicate that an influenza vaccine that primarily relies on CD8 T cells for protection could be developed.

INTRODUCTION

Influenza virus infection continues to pose a health and agricultural burden on humans and animals, chiefly for swine and poultry. Although infection in humans can be partially prevented by vaccination through currently licensed live attenuated and inactivated influenza vaccines (LAIV and IIV, respectively), where protection correlates with the induction of anti-hemagglutinin (anti-HA) neutralizing antibodies, the efficacy of these vaccines is not substantial. A recent systematic review of 17 randomized controlled trials and 14 observational studies found that the average efficacy of IIV and LAIV vaccines is only ca. 60 and 80%, respectively (1). Another concern has been raised for the elderly and for people with underlying medical conditions, for whom the efficacies of both types of influenza vaccines could be even lower (1, 2). In addition, there is minimal protection against infection with influenza virus that differs antigenically from the vaccine strains, including two currently circulating subtypes of influenza A virus (H1N1 and H3N2) and at least one of the two lineages of influenza B virus (Yamagata-like and Victoria-like) (2). As a result, influenza vaccines need to be reformulated annually to match the predominant circulating strains, which presents a tremendous economic burden and poses a challenge for influenza vaccine production and delivery logistics (3). Thus, development and evaluation of novel influenza vaccine approaches that can induce broad protective immunity is a focus of research.

CD8 T cells are the traditionally defined cytotoxic adaptive immune lymphocytes that can recognize specific peptides presented on host major histocompatibility complex class I (MHC-I) molecules via their surface T cell receptors. After being recruited to the infected tissues, cytotoxic T cells recognize and kill virally infected target cells through the release of effector molecules, such as perforin, Fas ligand, and granzymes (4, 5). During influenza virus infection, a majority of virus-specific CD8 T cells target influenza virus internal proteins, which are relatively conserved among different subtypes. In a C57BL/6 mouse model, two immunodominant CD8 T cell peptides have been identified within the viral nucleoprotein (NP) and the polymerase acidic (PA) proteins (NP366 and PA224, respectively) of X31 (H3N2) or PR8 (H1N1) viruses, since they share the same internal proteins (6). Studies by us and others have demonstrated that priming infection with one subtype of influenza virus could generate CD8 T cell immunity that is protective against infection of different subtypes (7–9). However, priming infection with wild-type (WT) viruses often leads to morbidity (weight loss) and pathological lesions. In addition, vaccination with live influenza virus without attenuation is not a practical approach for humans.

Genetic modification (mutation, deletion, and substitution) of viral components using molecular biology techniques represents a promising strategy for viral attenuation. Among these, single-cycle replicating viruses can be generated by manipulation of the viral genome (10). By removing the ability of a virus to spread after initial infection, bona fide immune stimulation could be elicited without causing the pathogenesis that is associated with nascent virus production. This method has been used for developing several single-cycle replicating viruses, such as herpes simplex virus, simian immunodeficiency virus, West Nile virus, rabies virus, lymphocytic choriomeningitis virus, and murine cytomegalovirus (11–16). One advantage of the single-cycle replicating virus is that it can trigger effective CD8 T cell responses (14, 16, 17). To fully minimize the potential damage to the host, while maintaining the ability to stimulate CD8 T cell immunity, we generated a single-cycle infectious influenza X31 virus (X31-sciIV) using plasmid-based reverse-genetics techniques that we have described previously (18–21). This virus has been modified so that the genomic HA segment, the receptor binding and fusion protein, has been replaced with green fluorescent protein (GFP). Thus, X31-sciIV can infect cells and undergo a single-cycle replication, but without HA transcomplementation it cannot form infectious progeny. Here, we observed that the one-cycle replication of X31-sciIV does not cause disease in mice, and yet priming with X31-sciIV induced influenza NP366 and PA224 peptide specific CD8 T cells in lymphoid and nonlymphoid tissues. In addition, these CD8 T cells express VLA-1 that is crucial for heterologous influenza protection, and they can produce cytokines in response to NP366 and PA224 peptide stimulation. Mice primed with X31-sciIV can be protected from lethal heterologous PR8 challenge. Depletion of CD8 T cells either after or before priming abolished protection against challenge. Our data indicate that sciIV immunization represents a novel vaccine approach for safely inducing functional heterologous CD8 T cell immunity.

MATERIALS AND METHODS

Animals.

Female C57BL/6 mice were ordered from The Jackson Laboratory and used at between 8 and 10 weeks old. Animals were maintained under specific-pathogen-free conditions at the University of Rochester vivarium facilities. All mouse experiments were approved by the University Committee of Animal Resources and complied with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Research Council. Infections were performed as previously described (19). Briefly, mice were anesthetized intraperitoneally (i.p.) with 2,2,2-tribromoethanol (Avertin; 240 mg/kg of body weight) and then inoculated intranasally (i.n.) with 30-μl portions of virus preparations as indicated. Morbidity was monitored by determining the body weight loss after infection, as well as based on clinical signs of infection (hunching, ruffling of fur, malaise, or respiratory distress). The percent body weight was determined relative to the starting weight.

Cells and virus.

Human embryonic kidney 293T cells (ATCC CRL-11268) and Madin-Darby canine kidney (MDCK) cells (ATCC CCL-34) were maintained in Dulbecco modified Eagle medium (DMEM; Mediatech, Inc.) supplemented with 10% fetal bovine serum (FBS; Atlanta Biologicals), and 1% PSG (penicillin, 100 U/ml; streptomycin, 100 μg/ml; l-glutamine, 2 mM [Mediatech, Inc.]). Cells were grown at 37°C in a 5% CO₂ atmosphere. MDCK cells constitutively expressing influenza virus H1N1 A/WSN/33 (WSN) have been previously described (18). MDCK HA-expressing cells were maintained in DMEM–10% FBS–1% PSG supplemented with hygromycin B at 200 μg/ml. After viral infections, the cells were maintained at 37°C in a 5% CO₂ atmosphere in DMEM containing 0.3% bovine serum albumin (BSA), 1% PSG, and TPCK (tolylsulfonyl phenylalanyl chloromethyl ketone)-treated trypsin (1 μg/ml; Sigma) (21).

Influenza H1N1 A/Puerto Rico/8/1934 (PR8) and influenza A/HKx31 (X31) viruses were prepared in eggs and used for i.n. infection in mice as described previously (7). WT X31 is a recombinant virus that contains the HA and NA segments of H3N2 A/Hong Kong/1/1968 and the remaining six segments from PR8 (22). Titers for WT and sciIV viruses were determined in MDCK or MDCK-HA cells by using a standard plaque assay or an HA assay.

Plasmids.

To generate viral RNA and protein expression plasmids for NA and HA of X31, MDCK cells were infected with X31 virus at a multiplicity of infection (MOI) of 0.1 for 24 h. RNA was isolated from infected cells by using TRIzol (Invitrogen), reverse transcribed (SuperScript II; Invitrogen), and amplified with specific primers and FastStart Taq DNA polymerase (Roche). PCR products were cloned into the pGEM-T vector (Promega) before subcloning into the pCAGGS protein expression plasmid using ClaI and NheI restriction enzymes (New England BioLabs [NEB]). HA and NA viral RNA containing the 3′ and 5′ noncoding regions (NCRs) were cloned into the pPolI viral RNA expression plasmid by using SapI restriction enzyme (20) (NEB). In plasmid pPolI HA(45)GFP(80), the internal open reading frame of HA was replaced with the entire coding region of green fluorescence protein (GFP), while conserving the 3′ and 5′ packaging signals (45 and 80 nucleotides, respectively) and NCRs of HA (23). All plasmids were generated by using standard cloning techniques and purified using a Wizard SV kit (Promega). The primers for the generation of the described plasmid constructs are available upon request. All plasmid constructs were verified by DNA sequencing (ACGT, Inc.).

Generation and characterization of MDCK X31-HA cell lines.

The pCAGGS X31 HA and pCB7 hygromycin B resistance plasmids were used to cotransfect MDCK cells using Lipofectamine 2000 transfection reagent (Invitrogen) at a ratio of 3:1, as previously described (19). Cell clones were selected after serial dilution and testing for complementation of sciIV infection and immunofluorescence assay (IFA). For complementation infections, cells were seeded 1 day prior to infection (3 × 10⁵ cells; 12-well plate format), and WSN-sciIV was used for infection at an MOI of 0.001. GFP expression was observed by fluorescence microscopy (Leica DM-IRB) at various times postinfection. Images were captured (Cooke Sensicam QE), pseudocolored, and merged using Adobe Photoshop CS4 (v11.0) software. Tissue culture supernatants from complementation experiments were collected at various times postinfection, clarified, and titrated on MDCK-HA cells to determine the HA titer. For IFA, cells were fixed with 4% formaldehyde (Polysciences, Inc.), washed with 1× phosphate buffer saline (PBS), and blocked with 2% BSA in 1× PBS. Primary incubation with mouse monoclonal antibody against WSN HA (2G9, 1 μg/ml) (18) or anti-X31 sera (1:1,000) was done at 37°C for 1 h. After three washes, fluorescein isothiocyanate (FITC)-conjugated rabbit anti-mouse antibody (1:140; Dako) mixed with DAPI (4′,6′-diamidino-2-phenylindole; 1:1,000; Research Organics) was added, followed by incubation for 30 min in the dark at 37°C. Cells were washed, mounted in 1× PBS, and visualized and imaged by fluorescence microscopy.

X31-sciIV rescue.

To generate X31-sciIV, ambisense (pDZ) reverse-genetics plasmids containing PR8 PB2, PB1, PA, NP, M, and NS (20) were used together with pPolI X31 NA, pPolI HA(45)GFP(80) (23), and pCAGGS X31 HA to cotransfect a mixture of 293T and MDCK-HA cells (1:1 ratio). At 48 h posttransfection, tissue culture supernatants from transfected cells were clarified of cell debris and used to infect MDCK-HA cells. Infection was monitored by determining the GFP expression, and X31-sciIV in supernatants at 3 days postinfection was plaque purified prior to stock amplification in MDCK X31-HA cells. Virus stocks were divided into aliquots and maintained at −80°C.

Hemagglutination assay.

A standard hemagglutination assay was carried out to estimate the production of X31-sciIV virus in parental and X31-HA expressing MDCK cells at various times postinfection. Briefly, 50-μl portions of infected tissue culture supernatants were 2-fold serially diluted with 1× PBS in a 96-well V-bottom plate, followed by incubation with an equal volume of 1% turkey red blood cells (RBC) for 45 min. The plates were then incubated for 30 min on ice and observed for hemagglutination. The HA titer was determined to be the reciprocal of the last dilution at which RBC were agglutinated.

Priming and challenging.

Mice were inoculated i.n. with 10⁵ PFU of X31-sciIV or WT X31 virus. At day 10 after priming, bronchoalveolar lavage (BAL) fluid was obtained by washing the respiratory tract using a 1-ml syringe loaded with 5% RPMI 1640 cell culture medium (Sigma). Lung, spleen, and mediastinal lymph node (MLN) tissues were harvested for single-cell suspension preparation and antibody staining. For challenge, mice were primed i.n. with 10⁵ PFU of X31-sciIV, rested for 2 weeks, and then infected with PR8 WT (3,000 PFU/mouse for a lethal dose and 3 PFU/mouse for a nonlethal dose).

Flow cytometry.

Unconjugated anti-CD16/32 was obtained from eBioscience. Live/Dead fixable violet fluorescent reactive dye was purchased from Molecular Probes/Invitrogen. FITC-, phycoerythrin (PE)-Cy7-, or Alexa Fluor 700-conjugated anti-CD3 antibodies were obtained from Biolegend. Alexa 488-conjugated anti-CD49a antibody was prepared as described previously (24). Allophycocyanin (APC)-Cy7-conjugated anti-CD8a (53-6.7) was purchased from BD Pharmingen. PE-conjugated H2-Db NP366 and APC-conjugated H2-Db PA224 tetramers were acquired from the National Institutes of Health (NIH) tetramer core facility at Emory University. Lung tissues were smashed, and single-cell suspensions were obtained by Histopaque 1083 purification (Sigma). For staining, freshly isolated cells were washed with staining buffer (1× PBS with 1% FBS) and blocked with unlabeled anti-CD16/32 for 20 min, followed by staining with Live/Dead violet dye and the respective antibodies for 30 min at 4°C. The cells were then washed twice and resuspended in staining buffer before the samples were run in the LSRII machine (BD Biosciences, San Jose, CA). All flow cytometry data were analyzed using FlowJo software (Tree Star, San Carlos, CA).

ELISPOT assay.

The enzyme-linked immunospot (ELISPOT) assay was described previously (25). Briefly, 5 μg of anti-mouse gamma interferon (IFN-γ) antibody (R4-6A2; BD Pharmingen)/ml was coated in 96-well ELISPOT plates (Millipore). The plates were washed with RPMI 1640 cell culture medium containing 10% FBS before freshly prepared spleen cells (0.5 × 10⁶ cells/well) were added. The cells isolated from mice were stimulated with H2-Db-restricted 9-mer NP366 and 10-mer PA224 peptides at a final concentration of 5 μg/ml for 16 h. Biotinylated anti-IFN-γ antibody (XMG1.2, 1:2,000; BioLegend) was added to plates for 1 h of incubation. The plates were washed, and 1:1,000 diluted alkaline phosphatase-conjugated streptavidin (Jackson Laboratory) was added, followed by incubation for 30 min, before the spots were developed using a Vector Blue alkaline phosphatase substrate kit III (Vector Laboratory). ELISPOT plates were analyzed by using a CTL ImmunoSpot plate reader and counting software (Cellular Technology Limited).

CD8 T cell depletion.

For CD8 T cell depletion, two strategies (depletion after and before priming) were applied. For depletion after priming, groups of mice were primed i.n. with 10⁵ PFU of X31-sciIV and rested for 2 weeks. Two days before challenge with a lethal dose of PR8 virus (3,000 PFU per mouse), mice were injected i.p. with 200 μg of anti-CD8 antibody (clone 2.43; BioXCell, West Lebanon, NH), or 200 μg of rat IgG2b isotype control (eBioscience). Antibody injection was performed again on the day of challenge and at days 2, 4, and 6 after challenge. Animals were then monitored daily for weight loss and survival. At day 3, BAL samples were collected and used for viral titration by the 50% cell culture infectious dose (TCID₅₀). For depletion before priming, groups of mice were injected i.p. with 200 μg of anti-CD8 or isotype control as described above. Antibodies were injected again on the day of priming and at days 2, 4, and 6 after priming with 10⁵ PFU of X31-sciIV. At day 28, mice were challenged with 3,000 PFU of PR8 virus as described above and monitored for weight loss and survival. At days 3 and 6 after challenge, lung samples were collected for viral titration. For both depletions, flow cytometric analysis was performed for verification of depletion, staining with anti-CD8a (clone 53-6.7; BD Pharmingen) antibody, which does not interfere with the depletion antibody.

TCID₅₀ titration.

BAL samples were centrifuged, and supernatants were used directly for viral titration. Lung samples were homogenized to obtain supernatants. Briefly, samples were 10-fold serially diluted in DMEM containing 1% BSA and TPCK-treated trypsin (1 μg/ml) in 96-well plates. Next, 50 μl portions from each resulting well were used to infect confluent MDCK cells in triplicate. After 1 h of incubation at 37°C, the cells were washed and cultured with serum-free DMEM. To visualize the cytopathic effect (CPE), cells were fixed at 72 h postinfection and stained with 1% crystal violet in 20% ethanol for 30 min at room temperature. The TCID₅₀ titer was determined based on the method of Reed and Muench.

Statistical analysis.

Statistical significance was determined by using a two-tailed, unpaired Student t test to compare two appropriate groups. For multiple comparisons, one-way analysis of variance (ANOVA) was used for statistical analysis. A P value of <0.05 was considered statistically significant.

RESULTS

X31-sciIV generation using influenza reverse-genetics techniques.

We have shown previously that sciIV can be pseudotyped to carry influenza HAs from various subtypes (2, 20). To recover pseudotyped viruses displaying the major surface antigen of X31 for future analysis on heterologous protection, we sought to first generate MDCK cells that constitutively express HA from X31. Corroborating our previous findings, only MDCK cells that express HA are able to complement an HA-deficient sciIV, whereas parental MDCK cells do not, as observed by the spread of GFP fluorescence after WSN-sciIV, although it appeared that WSN-sciIV replicated better in WSN-HA MDCK cells than in X31-HA MDCK cells (Fig. 1A). Importantly, X31-HA MDCK cells were antigenically accurate, since sera from mice recovering from X31 viral infection were able to specifically bind to X31-HA cells, but not WSN-HA or parental MDCK cells, as detected by IFA (Fig. 1B). In contrast, 2G9 monoclonal anti-WSN HA antibody only recognized WSN-HA MDCK cells (Fig. 1B). For X31-sciIV rescue, we selected and used one X31 HA-expressing MDCK cell line that can support a high level of WSN-sciIV virus production, as assessed by hemagglutination assay (data not shown). Using methods similar to those used to generate WSN-sciIV (18) or pandemic influenza virus H1N1 A/California/04/2009 (pH1N1)-derived pH1N1/E3-sciIV (19), X31-sciIV was engineered. After plaque purifying and amplifying stocks of X31-sciIV, we verified that X31-sciIV can only propagate in HA-expressing MDCK cells, but not in parental MDCK cells, as indicated by the GFP expression (Fig. 2A). Thus, the rescued X31-sciIV virus has HA protein on the surface but does not contain functional HA gene in its genome, preventing subsequent rounds of replication in normal target cells lacking HA (MDCK). However, in HA-expressing MDCK cells, X31-sciIV can propagate as efficiently as the WT virus (Fig. 2B).

FIG 1.

Generation of MDCK cells expressing X31 H3 HA. To establish X31 HA-expressing MDCK cell lines, the X31 HA mRNA gene was cloned into the pCAGGS vector and cotransfected with the pCB7 vector containing the hygromycin B gene. After transfection, hygromycin B-resistant clones were subcultured and screened by infection using WSN-sciIV virus complementation (A) and IFA (B). (A) WSN-sciIV virus complementation. Parental, WSN-expressing, and X31 HA-expressing MDCK cells were infected with WSN-sciIV at an MOI of 0.01. At 48 h postinfection, cell monolayers were imaged with a fluorescence microscope to observe GFP expression. Images are shown at ×10 magnification. Scale bar, 40 μm. (B) IFA. Parental, WSN-expressing, and X31 HA-expressing MDCK cells were fixed and stained with 2G9 (WSN) or X31 sera, followed by treatment with rabbit anti-mouse IgG FITC (green) and DAPI for nuclear staining (blue). The cell monolayers were imaged with a fluorescence microscope. Images are shown at ×40 magnification, pseudocolored and overlaid using Photoshop software. Scale bar, 10 μm.

FIG 2.

Viral rescue of X31-sciIV by reverse genetics. A mixture of 293T cells and X31 HA-expressing MDCK cells was cotransfected with pPolI X31 NA, pPolI HA(45)GFP(80), pCAGGS X31 HA, and ambisense pDZ rescue plasmids encoding PR8 PB2, PB1, PA, NP, M, and NS. (A) At 48 h posttransfection, tissue culture supernatants were harvested for infecting X31 HA-MDCK cells and, at 48 h postinfection, GFP expression was observed using fluorescence microscopy. Images are shown at ×40 magnification. Scale bar, 10 μm. (B) Rescued X31-sciIV virus was then plaque purified, and the viral growth on parental and X31 HA-expressing MDCK cells (MOI = 0.001) was compared to that of WT X31 virus using a hemagglutination assay.

Lack of X31-sciIV pathogenicity in C57BL/6 mice.

Mouse body weight loss is one of the clinical signs that can be observed and measured after influenza virus infection. It is also well correlated with the severity of mouse lung lesions and is commonly used to evaluate influenza virus pathogenicity in mice, together with mortality (26). To examine the pathogenicity of X31-sciIV, groups of 10 C57BL/6 mice were inoculated i.n. with 10⁵ PFU of X31-sciIV. For 2 weeks postinfection, animals inoculated with X31-sciIV remained active and did not show any signs of sickness (data not shown). There was also no weight loss observed after inoculation with X31-sciIV, whereas typical weight loss and full recovery was seen in animals infected with the same dose of WT X31 virus (Fig. 3). Instead, slight weight increase was seen in X31-sciIV virus-infected mice in a pattern similar to that of the PBS inoculation control. As expected, global lung lesions were not seen in any of the mice infected with X31-sciIV at any of the time points examined (not shown). Similar results were obtained when mice were inoculated with 10⁶ PFU of X31-sciIV (data not shown). Together, the data indicate that X31-sciIV virus does not cause disease and pathogenesis.

FIG 3.

X31-sciIV infection did not cause weight loss in mice. Female C57BL/6 mice were i.n. inoculated with 10⁵ PFU of X31-sciIV virus, 10⁵ PFU of WT X31 virus, or PBS control. The body weight of each mouse was recorded daily for 2 weeks, and values are expressed as the percentage of the original body weight measured before inoculation. The data are averages plus the standard deviations (SD) from 10 mice per group and are representative of two independent experiments.

CD8 T cell immune responses after X31-sciIV infection in vivo.

Single-cycle replicating viruses from disparate viral families, including Arenaviridae and Rhabdoviridae (negative-sense, single-stranded RNA) and Herpesviridae (double-stranded DNA), have been shown to be able to induce virus specific CD8 T cell immune responses (14, 16, 17). To investigate whether X31-sciIV can similarly elicit CD8 T cell responses, we infected mice with 10⁵ PFU of X31-sciIV and, at day 10 after inoculation, various tissue cells were prepared and stained with anti-CD3, anti-CD8, antibodies and H2-Db-restricted NP366 and PA224 peptide specific tetramers. Compared to the robust increase in lung and BAL fluid CD8 T cells after X31 infection, the total numbers of lung and BAL CD8 T cells were only slightly increased at day 10 after X31-sciIV priming, each with respect to mock-primed mice (Fig. 4A). However, the influenza NP366 and PA224 peptide specific CD8 T cells could be clearly detected in the lung and airway tract after X31-sciIV priming (Fig. 4B and C), and even the PA224 tetramer showed a much lower level of staining for lung and BAL cell samples from X31-sciIV-primed mice (Fig. 4B). In addition to airway tissues, influenza virus-specific CD8 T cells were present in MLN and spleen after X31-sciIV priming (Fig. 4D). This finding indicates that X31-sciIV can induce acute virus-specific CD8 T cell responses and that these cells can also be recruited into respiratory tissues, where infection occurred.

FIG 4.

CD8 T cell response induced by X31-sciIV. Female C57BL/6 mice were i.n. inoculated with 10⁵ PFU of X31-sciIV virus, 10⁵ PFU of WT X31 virus, or PBS control. At 10 days after inoculation, the spleens, mediastinal lymph nodes (MLN), lungs, and bronchoalveolar lavage (BAL) tissues were harvested for lymphocyte preparation and FACS staining. (A) Total CD8 T cells in the lungs and BAL fluid. The total number of lung/BAL CD8 T cells was obtained by multiplying the total number of lung or BAL lymphocytes counted by the percentage of CD8 T cells in the total gated lymphocyte population. The data represent averages ± the SD. The differences in the total CD8 T cells in the lungs or BAL fluid among three groups of animals were compared and analyzed by using one-way ANOVA. (B) Tetramer analysis of influenza virus-specific CD8 T cells in the lungs and BAL fluid. Lung or BAL single-cell suspensions were stained with Live/Dead dye, anti-CD3, anti-CD8 antibodies, and H2Db/NP366 and H2Db/PA224 tetramers. The data are representative of three independent experiments. (C) Total NP366- and PA224-specific CD8 T cells in lungs and BAL fluid. The total number of lung or BAL CD8 T cells was obtained by multiplying the total number of lung or BAL CD8 T cells by the percentage of tetramer-positive cells. The data represent averages ± the SD. The differences in the influenza virus NP366- or PA224-specific CD8 T cells in the lungs or BAL fluid among three groups of mice were compared and analyzed by using one-way ANOVA. (D) Tetramer analysis of influenza virus-specific CD8 T cells in lymphoid tissues. At day 10 after inoculation, the spleens and MLNs were collected for single-cell suspension preparation and staining with Live/Dead dye, anti-CD3, anti-CD8 antibodies, and H2Db/NP366 and H2Db/PA224 tetramers. The data are representative of three independent experiments.

Characterization of X31-sciIV induced influenza virus-specific CD8 T cells.

VLA-1 is an integrin that can mediate the adhesion of activated CD8 T cells to epithelial collagen (27). It has been previously shown to be associated with the survival of memory CD8 T cell in the airway and the efficacy of protection against heterologous influenza virus rechallenge (28). To evaluate whether i.n. immunization with X31-sciIV can induce CD8 T cell effectors similar to those induced by WT infection, VLA-1 expression was determined on influenza virus-specific respiratory CD8 T cells. At day 10 after priming, we analyzed lung and BAL cell samples for VLA-1 expression. We found that after X31-sciIV priming, ca. 50% of NP366 tetramer-specific lung CD8 cells expressed VLA-1, which was similar to that seen with WT X31 infection (Fig. 5A). Surprisingly, the proportion of NP366 tetramer specific cells expressing VLA-1 was higher in BAL fluid compared to that with WT X31 infection (Fig. 5A). VLA-1 expression was also detectable on lung or BAL PA224 tetramer-specific CD8 T cells after X31-sciIV priming, but at lower proportions compared to that seen with WT X31 infection (Fig. 5B). To further examine whether the CD8 T elicited cells can produce cytokines, we stimulated spleen cells from primed animals with NP366- and PA224-specific MHC-I peptides and used an IFN-γ ELISPOT assay to measure antigen-specific CD8 T cell responses. Our data showed that moderate frequencies of CD8 T cells generated by X31-sciIV immunization demonstrate IFN-γ cytokine responses to NP366 peptides compared to that from WT-X31 infection (Fig. 5C). Activation of CD8 cells stimulated with PA224 peptide could also be detected, but to a lesser extent than that observed with NP366 (not shown). This result agreed with the low number of PA224-specific CD8 T cells that were induced by priming with X31-sciIV (Fig. 4B and C). Nevertheless, these data suggest that functional influenza virus-specific CD8 T cell responses can be elicited after X31-sciIV priming.

FIG 5.

VAL-1 expression and IFN-γ response by CD8 T cells after X31-sciIV priming. Female C57BL/6 mice were intranasally inoculated with 10⁵ PFU of X31-sciIV or WT X31 virus control. At 10 days postinoculation, various tissues, including spleen, lung, and bronchoalveolar lavage (BAL), were harvested for single-cell suspension preparation. (A) VLA-1 expression on NP366-specific lung and BAL CD8 T cells. Lung single-cell suspensions were stained with Live/Dead dye, anti-CD3, anti-CD8, anti-CD49a antibodies, and H2Db/NP366 tetramers. The plots are representative of three independent experiments. (B) VLA-1 expression on PA224-specific lung and BAL CD8 T cells. BAL single-cell suspensions were stained with Live/Dead dye, anti-CD3, anti-CD8, anti-CD49a antibodies, and H2Db/PA224 tetramers. The plots are representative of three independent experiments. (C) Peptide-specific IFN-γ production. Spleen cell suspensions were stimulated with NP366 or PA224 peptides. The IFN-γ response was measured by an ELISPOT assay. The data are averages ± the SD from three independent experiments. The differences in IFN-γ response by spleen NP366- or PA224-specific cells from three groups of mice were compared and analyzed by using one-way ANOVA.

Heterologous protection by X31-sciIV priming with a single dose.

Influenza CD8 T cells are known to be crucial for heterologous influenza virus protection. Our observation that influenza virus-primed CD8 T cells could be generated in the airways and lungs after X31-sciIV priming prompted us to test whether X31-sciIV priming could provide protection against heterologous infection. We primed mice with 10⁵ PFU of X31-sciIV and challenged them with a lethal dose of PR8 virus 14 days later. X31-sciIV virus-primed mice initially showed a weight loss after lethal PR8 challenge (Fig. 6A), but they did not show the severe clinical symptoms, including hunched posture and lethargy, that were experienced by PBS-primed control mice (data not shown). Furthermore, mice primed with X31-sciIV virus regained body weight after day 6 of challenge, and all virus-primed mice survived (Fig. 6A and B), whereas all PBS-primed control mice succumbed to infection by day 10 (Fig. 6B). Thus, with X31-sciIV priming, mice were protected from lethal PR8 heterologous challenge. Such protection conferred by X31-sciIV priming was also evident when a nonlethal dose of PR8 virus was used for challenge, which led to an absence of morbidity compared to PBS-primed mice (Fig. 6C).

FIG 6.

X31-sciIV provides protection from heterologous influenza virus challenge. Female C57BL/6 mice were primed with either 10⁵ PFU of X31-sciIV virus, 10⁵ PFU of WT X31 virus, or PBS control and rested for 14 days before challenge with a lethal dose of PR8 virus. Animal body weight loss (A) and survival (B) in each group (n = 10) were monitored daily for 2 weeks after challenge. (C) Weight loss after a nonlethal PR8 challenge. Female C57BL/6 mice were primed with either 10⁵ PFU of X31-sciIV, WT X31 virus or PBS control and rested for 14 days and then rechallenged with a nonlethal dose of PR8 virus. Weight loss of the animals in each group (n = 10) was then monitored daily for 2 weeks. The data presented are averages ± the SD and are representative of two independent experiments.

CD8 T cells induced by X31-sciIV are crucial for heterologous challenge protection.

To examine the potential involvement of CD8 T cells in heterologous protection after X31-sciIV priming, we depleted CD8 T cells in X31-sciIV-primed animals by using anti-CD8 antibody before and after lethal PR8 challenge. Antibody blockade led to an undetectable amount of CD8 T cells in the lungs at day 3 postchallenge compared to the isotype control (data not shown). We found that animals with CD8 depletion had significantly higher BAL virus titers compared to animals depleted with isotype control antibody (Fig. 7A). In fact, with CD8 T cell depletion, X31-sciIV-primed animals had BAL virus titers only slightly lower than those seen in animals primed with a PBS control, suggesting that CD8 T cells were major contributors in controlling virus lung replication after heterologous challenge. In addition, mice with CD8 depletion showed continuous weight loss and could not recover after lethal PR8 challenge (Fig. 7B). The mortality of the mice in CD8 depletion group reached 100% at day 9 after challenge, although compared to the PBS-primed control animals these mice demonstrated slightly prolonged survival (Fig. 7C), suggesting additional antiviral functions of antibodies or immune cells. In contrast, gradual recovery was observed at day 6 after lethal PR8 challenge in all X31-sciIV-primed mice with isotype depletion, and no deaths occurred (Fig. 7B and C). Thus, the data indicate that CD8 T cells induced by X31-sciIV priming in this model are necessary for protection against heterologous influenza virus challenge.

FIG 7.

CD8 T cells mediated heterologous protection induced by X31-sciIV. Female C57BL/6 mice were primed with 10⁵ PFU of X31-sciIV or PBS and rested for 14 days before challenge with a lethal dose of PR8 virus. X31-sciIV primed animals were either depleted with CD8 antibody or isotype control 2 days before the challenge, on the day of the challenge, and at days 2, 4, and 6 after the challenge. (A) Virus detection in BAL samples. BAL washes were performed at day 3 after lethal PR8 challenge in PBS- or X31-sciIV-primed mice with CD8 or isotype control depletion. BAL supernatants (n = 5 per group) were titrated on MDCK cells grown in 96-well plates. The differences in virus titer in the BAL samples from three groups of mice were compared and analyzed by using one-way ANOVA. (B and C) Morbidity and mortality. Weight loss (B) and survival (C) of the animals in each group (n = 10) were monitored daily after the challenge for 2 weeks. The data presented are averages ± the SD.

To further evaluate whether naive or X31-sciIV-primed CD8 T cells are involved in heterologous protection in our model, mice were depleted of CD8 T cells before and during priming. FACS analysis of the lungs, spleen, and MLN at 10 days after priming confirmed that >95% of the CD8 T cells could be depleted (data not shown). Mice were rested for 28 days after priming, allowing the naive CD8 T cell compartment to repopulate, and were then challenged with PR8. At 6 days postchallenge, significantly more virus was recovered from the lungs of anti-CD8-treated mice compared to isotype control-treated mice (Fig. 8A). There was no significant difference in the virus titers at 3 days after challenge, and both groups of mice lost weight at similar rates, although only X31-sciIV-primed mice treated with the isotype control regained weight (Fig. 8B). Furthermore, CD8-depleted mice all succumbed to influenza virus infection by day 9 (Fig. 8C), an observation similar to our findings when depleting upon challenge. Altogether, these findings demonstrate that heterologous protection conferred by X31-sciIV priming is achieved in the presence of antigen-specific memory CD8 T cells. This approach also demonstrated that there is a minimal contribution of naive CD8 T cells to heterologous protection in our vaccine model.

FIG 8.

Antigen-specific CD8 T cells from X31-sciIV priming are recalled for protection upon challenge. Female C57BL/6 mice were either depleted with CD8 antibody or isotype control 2 days before priming with 10⁵ PFU of X31-sciIV, on the day of the priming, and at days 2, 4, and 6 after the priming. Mice were rested for 28 days before challenge with a lethal dose of PR8 virus. (A) Virus detection in lungs. Lungs were excised at 3 and 6 days after lethal PR8 challenge. Lung homogenates (n = 3 per group) were titrated on MDCK cells grown in 96-well plates. The differences in the lung virus titer at day 3 or 6 between the isotype control or anti-CD8-depleted mice were compared by using an unpaired Student t test. (B and C) Morbidity and mortality. Weight loss (B) and survival (C) of the animals in each group (n = 4) were monitored daily after the challenge for 2 weeks. The data presented are averages ± the SD.

DISCUSSION

Influenza virus is known for its remarkable rapid antigenic evolution and for the constant emergence of novel zoonotic strains, both of which allow this virus to evade vaccine-induced protection (29, 30). Currently approved influenza vaccines only show adequate protection when the dominant circulating virus matches the seed viruses used in the vaccines. To overcome this limitation, novel types of influenza vaccines that can induce broad heterologous protection against different viral isolates need to be developed.

Influenza virus attenuation using molecular approaches has been a focus for novel influenza vaccine development. One excellent example is a virus with nonstructural protein 1 (NS1) deletion or truncation, which primarily disrupts the interferon antagonism function of NS1 (31, 32). The NS1-altered influenza virus is significantly attenuated but remains immunogenic, and immunization with this virus can provide protection against homologous virus challenge (33, 34). One shortcoming of using this NS1-modified virus as a vaccine is that the virus is not fully attenuated, so safety is a concern. Similar issues exist for genetic attenuation targeting other influenza virus proteins, such as the NP and M2 proteins (35, 36). Compared to live attenuated influenza virus or the attenuation mutations mentioned above, the X31-sciIV generated in the present study using a technique we reported previously (18, 23) is self-limiting and safe.

Preexisting CD8 T cell immunity has been shown capable of inhibiting virus replication, alleviating the severity of disease and preventing death in case of lethal infection (4). The critical role of CD8 T cells in influenza virus protection has been extensively studied in animal models and human subjects using either laboratory-adapted viruses or clinical isolates (5, 7, 37–46). In the late 1970s, seminal studies using a mouse model have demonstrated that transferred influenza virus-specific cytotoxic T cells could provide protection against lethal challenge and that the killing function of these cells is not restricted to a particular virus strain (41, 42). It was further shown by several independent laboratories that the majority of influenza virus-specific CD8 T cells are generated toward influenza virus internal proteins (47–50). Although we are gaining further understanding of the mechanism of CD8 T cell-mediated protection against influenza virus infection, influenza vaccines that solely rely on CD8 T cell immunity are rarely reported and studied. As a result, such a vaccine is not commercially available.

Here, using X31-sciIV virus derived from the H3N2 X31 WT virus, we show that sciIV immunization could elicit CD8 T cell response and the induced influenza virus-specific CD8 T cells could express the VLA-1 protective marker and have biological activities such as cytokine production. In addition, sciIV-induced CD8 T cells were similarly distributed to infection sites and noninfection sites as seen by WT influenza virus infection, although, as expected, at a lower level. Finally, memory CD8 T cell immunity induced by X31-sciIV was shown to be critical for heterologous protection. In fact, without CD8 T cell immunity, the protection against rechallenge is nearly lost. Hence, our findings suggest that lung airway CD8 T cells offer significant protection. However, a direct comparison has not yet been tested to evaluate the protective differences between upper-airway CD8 T cells (induced by vaccines such as LAIV) and lower-airway CD8 T cells (as elicited in the present study). Our studies also suggest strategies that generate lung airway CD8 T cells should be considered as vaccine approaches. In particular, the sciIV approach used here demonstrates that influenza vaccines whose protection is predominantly mediated by CD8 T cells could be developed. Since these CD8 T cells generated by sciIV priming target influenza virus internal proteins, an sciIV influenza vaccine could possibly offer broad protection.