Immunogenicity of a Recombinant Influenza Virus Bearing Both the CD4+ and CD8+ T Cell Epitopes of Ovalbumin

Recombinant influenza viruses that bear the single immunodominant CD8+ T cell epitope OVA257−264 or the CD4+ T cell epitope OVA323−339 of the model antigen ovalbumin (OVA) have been useful tools in immunology. Here, we generated a recombinant influenza virus, WSN-OVAI/II, that bears both OVA-specific CD8+ and CD4+ epitopes on its hemagglutinin molecule. Live and heat-inactivated WSN-OVAI/II viruses were efficiently presented by dendritic cells in vitro to OT-I TCR transgenic CD8+ T cells and OT-II TCR transgenic CD4+ T cells. In vivo, WSN-OVAI/II virus was attenuated in virulence, highly immunogenic, and protected mice from B16-OVA tumor challenge in a prophylactic model of vaccination. Thus, WSN-OVAI/II virus represents an additional tool, along with OVA TCR transgenic mice, for further studies on T cell responses and may be of value in vaccine design.


Introduction
Recombinant influenza viruses represent a promising delivery system of foreign antigens that can be fused to viral proteins, such as hemagglutinin (HA), neuraminidase (NA), or nonstructural protein 1 (NS1), or can be expressed as an individual whole protein flanked with NA segmentspecific packaging signals [1][2][3][4][5]. Influenza viruses bearing model antigens have been used as tools to address several questions and long-standing issues in immune biology. In particular, recombinant influenza viruses bearing the single immunodominant CD8+ T cell epitope OVA 257−264 or the CD4+ T cell epitope OVA 323−339 of chicken ovalbumin (OVA) have been used in several studies to characterize T cell responses [6][7][8][9][10][11][12].
HA engineered to carry small polypeptides fused to its Nterminus is efficiently expressed and stably incorporated into retroviral particles without affecting its fusion activity [13].
Recently, this strategy was adopted to generate recombinant influenza viruses bearing relatively large domains of the protective antigens of Bacillus anthracis toxin (domain 1 , the region responsible for binding to the lethal factor and the edema factor (LEF) and domain 4, the receptorbinding domain (RBD)) fused to the N-terminus of the HA of A/Aichi/2/68 (H3N2) in the genetic background of A/WSN/33 virus (H1N1) [14,15]. The chimeric fusion proteins, named HA/LEF and HA/RBD, were functionally expressed during virus infection in mice and elicited antibody responses specific to both the viral HA and the protein domains of the Bacillus anthracis toxin. Here, by using a similar approach, we generated a recombinant influenza virus, WSN-OVA I/II , bearing both OVA-specific CD8+ and CD4+ epitopes in its HA, and characterized the OVA-specific T cell immune response in vitro and in vivo as well as its ability to protect mice from challenge with the OVA-expressing tumor cells B16-OVA in a prophylactic model of vaccination.  [16], and containing the GGGGD peptide arm, as described by Li et al. [14] ( Figure 1). The recombinant influenza WSN-OVA I/II virus bearing both of the OVA-specific CD8+ and CD4+ epitopes was generated by use of plasmid-driven reverse genetics, as described by Neumann et al. [16]. The presence of the inserted sequence was confirmed by sequencing.

Antigen Presentation In Vitro.
For the antigen presentation assays in vitro, CD8+ and CD4+ T cells were isolated from spleens and lymph nodes of OT-I and OT-II OVA TCR transgenic mice, respectively, by use of magnetic sorting (Miltenyi Biotec), according to the manufacturer's protocol. CD11c-positive dendritic cells (DCs) were isolated from the spleens of C57BL/6 mice by using MACS beads (Miltenyi Biotec) and the purity of the enriched CD11cpositive cell preparation ranged from 90% to 95%. DCs were resuspended to 10 7 cells/mL in serum-free RPMI 1640, and incubated with either live or heat-inactivated (HI) WSN-OVA I/II virus (1000 HAU/mL) for 1 h at 37 • C, 5% CO 2 . Heat inactivation was performed by incubating WSN-OVA I/II virus at 56 • C for 30 min in a water bath, and virus fusion activity and inactivation were determined as previously described [17]. After incubation, antigen-loaded DCs were washed and cultured with 10  In addition, the frequency of SIINFEKL-specific IFN-γproducing T cells was measured by using the ELISPOT assay on spleen-derived lymphocytes.

Results
The recombinant influenza WSN-OVA I-II virus bearing both the OVA-specific CD8+ and CD4+ epitopes was generated by introducing the coding sequences into the HA gene of A/WSN/33 virus, after the coding region of the HA signal peptide (SP), and the GGGGD peptide arm, as described by Li et al. [14] ( Figure 1). The rescued virus replicated in tissue culture with a titer ranging from 5 × 10 6 to 10 7 plaqueforming units (PFU)/mL. We first determined whether the OVA-specific T cell epitopes of WSN-OVA I−II virus were efficiently presented to antigen-specific T cells. To this end, the capacity of splenic DCs to process and present both live and HI WSN-OVA I-II viruses was assessed by using an in vitro proliferation assay of naïve OVA-specific CD8 + T and CD4+ T cells derived from OT-I and OT-II TCR-transgenic mice, respectively. Live virus-loaded DCs efficiently processed and presented both OVA-specific epitopes on MHC class I and II molecules, whereas there was no response with uninfected DCs (Figure 2(a)). CFSE-labelled transgenic T cells were also used to monitor antigen-specific proliferation due to the progressive halving of the dye fluorescence following cell division. WSN-OVAI/II virus-loaded DCs were responsible for vigorous proliferation of OT-I and OT-II cells within the CD8+ and CD4+ gated populations, respectively, compared to unloaded-or WSN virus-loaded DCs (Figure 2(b)). Furthermore, our data indicate that the OVA-specific MHC class I-restricted epitope located on the HA of whole HI WSN-OVA I/II virus was efficiently cross-presented by DCs to CD8+ T cells (Figure 2(a)). The binding and fusogenic functions of HI WSN-OVA I/II virus were intact as assessed by standard hemagglutinating and hemolytic assays, respectively, thus confirming the importance of a functional HA to access the cytoplasm of DCs [17]. The higher T cell proliferation of OT-I compared to OT-II cells induced by DCs loaded with HI WSN-OVA I/II virus, bearing equimolar amounts of both epitopes, correlates with previous data showing that the cell-associated antigen is more efficiently presented to CD8+ than to CD4+ T cells [19].
To determine the immunogenicity of the WSN-OVA I/II virus in C57BL/6J mice, we first determined its pathogenicity relative to the wild-type WSN virus. None of the inoculated mice died, even those inoculated with doses of WSN-OVA I/II virus as high as 10 6 PFU, whereas the wild-type WSN virus killed all of the mice at this dose (data not shown). Pulmonary viral loads measured 3 and 6 days after infection in mice inoculated with 3 × 10 5 PFU of WSN-OVA I/II virus (average titres of 4.1 and 2.3 log 10 TCID 50 /lung, resp.) were about four logs lower than those observed for mice inoculated with 5 × 10 2 PFU of the wild-type WSN virus ( Figure 3). Moreover, we could not detect virus in the nasal turbinates of the WSN-OVA I/II -infected mice, whereas detectable viral titres were measured in the nasal turbinates of the WSN-infected animals on day 3 and 6 after infection (data not shown). Although the recombinant virus did not replicate as efficiently as the WSN virus, we used the above viral doses to immunize mice and measured comparably high levels of NP-specific CD8+ T cells for both viruses in the MLN and in the spleen at day 9 after infection ( Figure 4(a)). The OVA-specific CD8+ T cell response was also vigorous, as determined by ex vivo IFN-γ release by lymphoid cells after the addition of specific peptides (Figure 4(a)) indicating that the attenuated recombinant virus was efficiently immunogenic for CD8+ T cells. To assess whether the engineered OVA-specific CD4+ T cell epitope was also presented in the context of infection, we measured the OVA 323−339 -specific CD4+ T cells secreting IFN-γ in the enriched CD4+ T cell population from the MLN of these mice. WSN-OVA I/II -infected mice produced an average of 120 spot-forming cells (SFC)/million cells, whereas WSNinfected mice gave no OVA-specific response (Figure 4(b)). Although the insertion of foreign epitopes into the HA of WSN-OVA I/II virus was responsible for the attenuated phenotype, the double OVA recombinant virus was still replicating and inducing a remarkable OVA-specific T cell response in mice. Recombinant viral vectors expressing the immunodominant OVA-specific CD8+ T cell epitope induce protective immunity against challenge with lethal doses of malignant melanoma cells expressing OVA [20,21]. Moreover, recombinant influenza viruses expressing a tumor-associated antigen have been previously shown to elicit efficient antitumor responses and to induce tumor inhibition in experimental murine cancer models [22,23]. We, therefore, examined the ability of WSN-OVA I/II virus to protect mice against tumors by using the tumor cell line B16-OVA (OVA-transfected B16 cells), which cause pulmonary metastases when given intravenously. Immunization with WSN-OVA I/II virus protected mice from lung metastases, as measured 12 days after tumor challenge, compared to nonimmunized mice ( Figure 5(a)). The antitumor immunity was OVA-specific, as all of the mice immunized with the parental WSN virus developed tumors. We also found that WSN-OVA I/II virus-immunized mice produced a stronger OVA 257−264 (SIINFEKL-)specific CD8+ T cell response than did either the nonimmunized mice or the wild-type WSN virus-immunized mice, as measured by ex vivo ELISPOT with pooled spleens on day 12 posttumor challenge ( Figure 5(b)). The higher OVA 257−264specific CD8+ T cell response detected in WSN-OVA I/II virus-immunized mice that were subsequently challenged with B16-OVA cells correlated with the presence of OVAspecific memory T cells in these mice before tumor challenge.

Discussion
Here, we engineered a recombinant influenza WSN-OVA I/II virus to bear the OVA-specific CD4+ and CD8+ T cell epitopes on its viral HA. This virus showed an attenuated phenotype and was highly immunogenic in mice. The effective proliferation of OVA-specific CD8+ T cells in vitro with HI WSN-OVA I/II virus-loaded DCs is in agreement with previous data showing that inactivated influenza viruses that maintain their fusogenic properties are efficiently crosspresented by DCs [17]. The presence of both OVA T cell epitopes on the same viral HA protein provides an additional tool, along with OVA-specific TCR transgenic mice, to investigate antigen processing and presentation by nonreplicating influenza virus-loaded DCs, in the context of MHC class I and class II. The double OVA recombinant virus thus represents an attractive delivery system for use in studies that explore new vaccination strategies for inducing virus-specific memory CD8+ T cells. Moreover, the finding that foreign epitopes located on the HA of inactivated recombinant influenza virions are efficiently cross-presented to specific CD8+ T cells offers interesting opportunities to further explore the use of recombinant influenza virusbased nonreplicating immunogens. In particular, this feature could be used to develop virosomes, as antigen delivery systems, since antigen encapsulation inside the virosomal particles is usually required for cytosolic delivery upon membrane fusion and for proteasome processing to induce CTL responses [24]. Our results also show that immunization with WSN-OVA I/II virus protects mice from B16-OVA tumor challenge. Here, we did not directly address the question of whether the double OVA recombinant virus has advantages over the existing recombinant viruses that bear the single OVA epitopes. Although the induction of OVA 257−264 (SIINFEKL-)specific CD8+ T cells in the context of viral infection may be sufficient to provide protective immunity, it is plausible that the simultaneous induction of tumorspecific CD4+ T cells may participate in interactions that shape the protective response against tumors [21]. This is particularly true if tumor-specific immune responses are induced with nonreplicating immunogens and thus in the presence of a poor inflammatory response. Recent studies suggest that abnormalities of the immune response in tumor-bearing hosts correlate with the induction of immune tolerance toward tumor-associated antigens and with the functional suppression of antitumor T cell effectors due to myeloid-derived suppressor cells [25,26]. Thus, combining WSN-OVA I/II virus infection and adoptive transfer of OVAspecific TCR transgenic T cells in mice bearing OVAexpressing tumors may help dissect the immunoregulatory mechanisms responsible for tumor-induced tolerance and immune suppression and may provide useful information for rational design of cancer immunotherapy.  Figure 5: Protective and antigen-specific immunity to melanoma B16 cells expressing OVA. C57BL/6J mice were immunized once with either the parental WSN virus or the recombinant WSN-OVA I/II virus. Three weeks after immunization, mice were challenged i.v. with 3 × 10 5 B16-OVA cells or with PBS. The number of pulmonary metastases (a) and the frequency of SIINFEKL-specific IFN-γ-producing T cells, as measured by using the ELISPOT assay on spleen-derived lymphocytes (b) was determined 12 days after the tumor cell inoculation. All of the experiments included 8 mice per group and were repeated three times with similar results. * P < 0.001 and * * P < 0.0001, comparing WSN-OVA I/II -versus WSN-infected and naïve mice.
Susceptibility for Avian Influenza, IMECS). The study was also funded by ERATO (Japan Science and Technology Agency), by a grant-in-aid for Specially Promoted Research from the Ministries of Education, Culture, Sports, Science, and Technology, by grants-in-aid from Health, Labor, and Welfare of Japan, and by National Institute of Allergy and Infectious Disease Public Health Service research grants.