Hyperattenuated Recombinant Influenza A Virus Nonstructural-Protein-Encoding Vectors Induce Human Immunodeficiency Virus Type 1 Nef-Specific Systemic and Mucosal Immune Responses in Mice

Abstract

We have generated recombinant influenza A viruses belonging to the H1N1 and H3N2 virus subtypes containing an insertion of the 137 C-terminal amino acid residues of the human immunodeficiency virus type 1 (HIV-1) Nef protein into the influenza A virus nonstructural-protein (NS1) reading frame. These viral vectors were found to be genetically stable and capable of growing efficiently in embryonated chicken eggs and tissue culture cells but did not replicate in the murine respiratory tract. Despite the hyperattenuated phenotype of influenza/NS-Nef viruses, a Nef and influenza virus (nucleoprotein)-specific CD8⁺-T-cell response was detected in spleens and the lymph nodes draining the respiratory tract after a single intranasal immunization of mice. Compared to the primary response, a marked enhancement of the CD8⁺-T-cell response was detected in the systemic and mucosal compartments, including mouse urogenital tracts, if mice were primed with the H1N1 subtype vector and subsequently boosted with the H3N2 subtype vector. In addition, Nef-specific serum IgG was detected in mice which were immunized twice with the recombinant H1N1 and then boosted with the recombinant H3N2 subtype virus. These findings may contribute to the development of alternative immunization strategies utilizing hyperattenuated live recombinant influenza virus vectors to prevent or control infectious diseases, e.g., HIV-1 infection.

Influenza viruses are segmented negative-strand RNA viruses belonging to the family Orthomyxoviridae. Influenza A virus consists of nine structural proteins and codes additionally for one nonstructural protein (NS1) with regulatory functions (7, 60). Due to the segmented nature of the viral genome, the mechanism of genetic reassortment can take place during mixed infection. This mechanism and the high mutation rate of the RNA genome of influenza viruses generate shift and drift antigenic variations in emerging viruses, allowing them to circumvent the preexisting immunity of the human population (29).

The development of reverse-genetics methods to manipulate the influenza virus genome has allowed the generation of influenza virus vectors that express foreign antigens (20, 46, 55). Several features suggest that influenza viruses could be attractive candidates for the development of effective vaccine vectors against various diseases: (i) influenza viruses induce strong cellular and humoral immune responses (2, 4, 5, 40); (ii) influenza virus does not contain a DNA phase in its replication cycle, and therefore chromosomal integration of viral genes can be excluded; (iii) since antibodies to different influenza virus subtypes show little cross-reactivity, preexisting immunity to the virus vector can be circumvented by booster immunizations with vectors belonging to different antigenic subtypes; and (iv) attenuated live influenza vaccines have already been developed (25, 34).

We previously demonstrated that mice immunized intranasally (i.n.) with recombinant influenza virus expressing a highly conserved human immunodeficiency virus type 1 (HIV-1)-specific neutralizing epitope inserted into the antigenic site B of the hemagglutinin (HA) molecule developed significant humoral immune responses at both the systemic and mucosal levels (13, 41). Other immunogenic influenza virus vectors expressing human and mouse malaria antigens in the stalk region of the neuraminidase (NA) have been reported elsewhere (38, 39, 50, 51).

Regardless of these promising data, the surface glycoproteins HA and NA of influenza viruses cannot be considered optimal targets for vector development. Repeated immunizations with the recombinant virus containing the same modified HA or NA is less effective for boosting the immune response due to preexisting immunity caused by the first immunization. Another limitation is the small size (about 10 amino acids) of the foreign amino acid sequence which can be introduced into the HA molecule (20, 46, 55). Although NA might tolerate insertions of longer sequences into its stalk region, this site is poorly presented to the immune system and is therefore less suitable for presentation of B-cell epitopes (20).

Since novel methods for plasmid-derived rescue of transfectant influenza viruses have been developed, theoretically all eight fragments of the influenza virus genome can be manipulated (15, 42). However, several features of the influenza virus NS gene indicate that it might be the preferred alternative for the introduction of desired foreign antigens. In contrast to other influenza virus proteins, the NS1 protein is variable in size among field and laboratory isolates of influenza A and B viruses (43). Due to the nonstructural nature of the NS1 protein, manipulations of the protein do not have a direct effect on the virion composition. At the same time, the NS1 protein is abundant in influenza virus-infected cells and is capable of inducing cytotoxic T lymphocyte (CTL) and specific antibody responses (3, 35). In addition, once the chimeric NS1 gene is rescued, it can be easily transferred to another influenza virus strain by methods of genetic reassortment.

We recently succeeded in establishing a reverse-genetics system using Vero cells, allowing us to obtain influenza viruses containing long deletions at the carboxyl end or even lacking the entire coding sequence of the NS1 protein (11, 19). Manipulation of the NS1 gene revealed its function to be that of an interferon (IFN) antagonist, and thus it appeared to be a powerful tool for the generation of attenuated influenza viruses (11, 56).

In the present study, we addressed the question of whether transfectant influenza A viruses can tolerate insertions of rather long sequences within the NS1 protein and whether such recombinant viruses would be attenuated and immunogenic in mice. The 137 C-terminal amino acids comprising multirestricted immunodominant regions of the HIV-1 Nef protein served as a model antigen (9, 21). Recombinant influenza viruses of the H1N1 and H3N2 subtypes (influenza/NS-Nef viruses), which express identical versions of a truncated NS1 protein (125 amino acids) fused to the 17-amino-acid self-cleaving 2A site of Picornavirus (48) and the 137 C-terminal amino acids of the Nef protein, were rescued in Vero cells by means of reverse genetics and the standard genetic reassortment methods. Both recombinant virus strains were genetically stable and showed normal growth characteristics in the usual cell substrates but did not replicate in mouse respiratory tracts. Surprisingly, despite the hyperattenuated phenotype, these influenza and NS-Nef virus clones induced significant Nef-specific B and CD8⁺-T-cell responses, detected in systemic and mucosal compartments, including the urogenital tracts, in mice after i.n. immunization.

MATERIALS AND METHODS

Cells.

Vero cells (ATCC CCL-81) were used for transfection experiments, selection and plaque purification of the rescued transfectant viruses, and virus titrations. The Vero cells were cultivated in serum-free medium (27). In addition, Madin-Darby canine kidney (MDCK) cells and 11-day-old embryonated chicken eggs were used for virus titrations. The MDCK cells were cultivated in Dulbecco's minimum essential medium (DMEM)-Ham's F12 (Biochrom KG) containing 2% fetal calf serum.

Construction of plasmids.

The transfectant virus was prepared using the existing plasmid clone of the influenza virus NS gene, pPUC19-T3/NS PR8, which contains a T3 promoter and the restriction site BpuA1 for plasmid linearization (11). The plasmid construct designated pPUC19-T3/NS/2A-Nef, containing the HIV-1/NL4-3 Nef-derived sequence (nucleotides 210 to 618 of the Nef gene), was obtained by a standard reverse transcriptase (RT)-PCR method using specific primers (information available upon request). Briefly, the purified PCR product has been blunt-end ligated into the plasmid construct pPUC19-T3/NS-124 between nucleotide positions 400 and 401 of the PR8 NS gene sequence. In the final step, a pair of complementary 5′ and 3′ 51-mer oligonucleotides (CODON Genetic Systems, Weiden, Austria) coding for the protease recognition sequence 2A (NFDLLKLAGDVESNLG/P) derived from foot-and-mouth disease virus (48) were hybridized and blunt-end ligated into the NS1 open reading frame upstream of the Nef-derived sequence.

Generation of recombinant influenza and NS-Nef viruses.

Generation of the A/PR/8/34 (PR8 wild-type [wt]) virus expressing the recombinant NS1 protein (PR8/NS-Nef) was performed according to the standard DEAE-dextran transfection protocol (33) with several modifications as described in detail earlier (11). Briefly, the synthetic ribonucleoproteins (RNPs) were generated by T3 RNA polymerase transcription from the linearized pPUC19-T3/NS/2A-Nef plasmid in the presence of purified influenza virus 25A-1 polymerase preparations (11).

In order to create the chimeric Aichi/NS-Nef virus, the RNA representing the recombinant NS segment of the PR8/NS-Nef virus was introduced into the genome of influenza A/Aichi/1/68 (H3N2) virus by a standard genetic reassortment performed on Vero cells utilizing rabbit polyclonal anti-PR8 virus hyperimmune serum for selection. Genotyping of the reassortants was performed by RT-PCR amplification and comparative restriction analysis of cDNA copies derived from each genome segment (the restriction map is available upon request).

Immunofluorescence.

Vero cells were infected with recombinant PR8/NS-Nef, Aichi/NS-Nef, or PR8 wt viruses at a multiplicity of infection (MOI) of 0.05. Twenty-four hours postinfection, the cells were trypsinized, washed, and fixed with acetone and then blocked with heat-inactivated goat serum followed by incubation of a 1:100 dilution of mouse monoclonal antibody (MAb) recognizing amino acid residues 179 to 195 of HIV-1 IIIB Nef (ARP387; this reagent was obtained from M. Harris and J. Neil through the NIBSC Centralised Facility for AIDS Reagents) or a 1:100 dilution of HIV-1_IR-CSF Nef mouse MAb recognizing epitopes on the C-terminal part of Nef (catalog no. 1539; this reagent was obtained from K. Krohn and V. Ovod through the AIDS Research and Reference Reagent Program, AIDS Division, National Institute of Allergy and Infectious Diseases, National Institutes of Health) for 40 min at 37°C (45). After being extensively washed with phosphate-buffered saline (PBS), the slides were incubated for 30 min with a 1:100 dilution of fluorescein isothiocyanate conjugated to goat anti-mouse immunoglobulin G (IgG) and then rewashed, mounted with 50% glycerol in PBS, and analyzed with a laser confocal microscope.

Analysis of viral protein synthesis.

Confluent monolayers of Vero cells in six-well plates were infected with recombinant influenza/NS-Nef viruses and PR8/NS-124 viruses at an MOI of 5. After 30 min, the inoculum was removed and RPMI medium was added. After 6 h of incubation at 37°C, the RPMI medium was replaced with 0.5 ml of cysteine-methionine-free minimal essential medium supplemented with [³⁵S]methionine-[³⁵S]cysteine (30 μCi/ml; Amersham) and incubated for 30 min. The cells were then washed two times with PBS and lysed directly in the dishes by adding 200 μl of electrophoresis sample buffer (62.5 mM Tris-HCl [pH 6.8], 2% sodium dodecyl sulfate, 5% 2-mercaptoethanol, 10% glycerol). The proteins were then analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis on 13% gels containing 5 M urea. After electrophoresis, the protein gels were fixed in a solution of 20% methanol and 5% acetic acid, rinsed with water, and dried for autoradiography.

Western blot analysis.

Western blot analysis was performed by electrophoretic transfer of the proteins from the 16% polyacrylamide gel to a polyvinyl-difluoride membrane (Millipore) for 2 h. After being blotted, the membrane was incubated for 1 h with the specific rabbit anti-NS hyperimmune serum (kindly donated by A. Garcia-Sastre, Mount Sinai School of Medicine, New York, N.Y.) or with HIV-1_IR-CSF Nef mouse MAb diluted 1:2,000 in PBS containing 0.1% Tween 20 and 1% skim milk. After being washed, the membrane was incubated for 1 h with alkaline phosphatase-conjugated goat anti-rabbit or goat anti-mouse IgG diluted 1:20,000. Following additional washing steps, blots were developed by adding a standard staining solution containing nitroblue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate.

Mice and immunizations.

Six- to 8-week-old female BALB/c mice (Forschungsinstitut für Versuchstierzucht, Himberg, Austria) were housed under conventional conditions and were provided with standard diet and water ad libitum. The mice were divided randomly into groups and were immunized i.n. in the absence or presence of ether anesthesia with live PR8/NS-Nef and Aichi/NS-Nef viruses according immunization schemes described in detail in the respective figure legends. Control groups of mice were immunized by the same immunization procedures with either the PR8/NS-124 or the PR8 wt virus. The intervals between immunizations were always 3 weeks.

Viral replication in murine respiratory tracts.

To determine viral replication in mouse respiratory tracts, BALB/c mice (nine mice/group) were immunized under ether anesthesia (see the legend to Fig. 4) and three mice per group were sacrificed at days 2, 4, and 6 after i.n. inoculation of the viral stocks. The lungs were aseptically removed, and group-specific pools of the organs were made. Tissue-derived extracts were prepared by grinding the tissue samples in a homogenizer to a 10% (wt/vol) suspension with glass sand in PBS containing penicillin, streptomycin, and gentamicin. (13). The suspensions were centrifuged at 3,000 × g for 5 min, and the supernatants were assayed for infectious virus particles in plaque assays, utilizing Vero cells.

FIG. 4.

Virus load in respiratory tract tissue of immunized mice. BALB/c mice were immunized i.n. under ether anesthesia with 10⁶ PFU of influenza PR8/NS-Nef, Aichi/NS-Nef, or PR8/NS-124 virus or 10³ PFU of the PR8 wt virus/mouse. Three mice of each group were sacrificed at days 2, 4, and 6 postinfection. The pooled lungs were homogenized, and the tissue extracts were assayed for infectious virus particles in plaque assays using Vero cells. The results are presented as PFU per milliliter of 10% (wt/vol) tissue extracts.

Serum collection.

Blood was collected from the murine retroorbital venous plexus 2 weeks after the second and third immunizations and was allowed to clot for 4 h at room temperature. Then the samples were spun in a microcentrifuge, and the sera were removed and stored at −20^°C.

ELISA.

An enzyme-linked immunosorbent assay (ELISA) protocol was performed as described earlier (13) utilizing a glutathione-S-transferase–Nef fusion protein (GST-Nef; 1 μg/ml in carbonate buffer; pH 9.6) as a coating antigen (22) (this reagent was obtained from G. Reid through the NIBSC Centralised Facility for AIDS Reagents). Serial dilutions of sera in PBS-Tween containing 1% skim milk were added to the coated plates, and the mixtures were incubated for 2 h at room temperature. Bound antibodies were detected with goat anti-mouse IgG γ-chain-specific antibody conjugated with horseradish peroxidase (Sigma). The plates were stained with o-phenylenediamine dihydrochloride as a substrate.

Isolation of lymphocyte populations.

Spleens and lymph nodes draining the respiratory tracts (mediastinal and tracheobronchial lymph nodes) and urogenital tracts of immunized BALB/c mice (three mice per group) were collected at day 10 following the last immunization and were used for isolating single-cell lymphocyte populations. The spleens and draining lymph nodes were mechanically dissociated into single-cell suspensions by means of cell strainers (Falcon). The erythrocytes present in the cell suspensions were lysed with Tris-buffered ammonium chloride. The urogenital tracts (vagina, cervix, uterine horns, and urethrae) were aseptically removed, minced, washed, and dissociated enzymatically in DMEM containing a mixture of collagenase (0.8 mg/ml; Sigma) and dispase (0.8 mg/ml; Boehringer Mannheim) according to a protocol described previously (13).

Cell separation.

In some experiments, single-cell suspensions were depleted of CD8⁺ cells by utilizing saturating concentrations of biotinylated rat anti-mouse CD8a MAb (5H10-1; BD PharMingen). The cells were then incubated with magnetic microbeads conjugated with streptavidin using the magnetic cell separation technique (Miltenyi Biotec, Bergisch Gladbach, Germany) (37). The efficiency of the cell separation was controlled by flow cytometry. Cell suspensions were always depleted up to 95% of the labeled cells (data not shown).

ELISPOT assay.

A protocol for the immediate ex vivo CD8⁺ gamma IFN (IFN-γ) enzyme-linked immunospot (ELISPOT) assay (49) was adapted utilizing the following synthetic peptides: Nef182-196 (Nef peptide; EWRFDSRLAFHHVAREL), comprising the conserved H-2K^d-restricted CTL epitope of the Nef protein (57), and the NP147-155 peptide (NP peptide; TYQRTRALV), an H-2K^d-restricted immunodominant CTL epitope of the influenza A virus nucleoprotein (39). Briefly, threefold serial dilutions of cell populations derived from murine spleens, draining lymph nodes, and urogenital tracts were transferred to wells coated with anti-IFN-γ MAb (R4-6A2; BD PharMingen). The cells were incubated for 22 h at 37°C and 5% CO₂ in DMEM containing 10% fetal calf serum, interleukin-2 (30 U/ml), penicillin, streptomycin, and 50 μM 2-mercaptoethanol in the presence of synthetic peptides. A biotinylated anti IFN-γ MAb (XMG1.2; BD PharMingen) was utilized as a conjugate antibody, and then the plates were incubated with streptavidin peroxidase (0.25 U/ml; Boehringer Mannheim Biochemica). Spots representing IFN-γ-secreting CD8⁺ cells were developed utilizing the substrate 3-amino-9-ethylcarbazole (Sigma) containing hydrogen peroxide in 0.1 M sodium acetate, pH 5.0. The spots were counted with the help of a dissecting microscope, and the results were expressed as the mean number of IFN-γ-secreting cells ± standard error of the mean (SEM) of triplicate cultures. Cells incubated in the absence of synthetic peptides developed <10 spots/10⁶ cells. Since depletion of CD8⁺ cells usually resulted in >92% reduction of spot formation, cell separation was omitted in most assays (data not shown).

RESULTS

Influenza virus tolerates a long insertion into the NS1 gene.

Several transfectant influenza viruses containing truncated forms of the NS1 protein have been generated previously in our laboratory (11). One of these transfectant viruses, designated PR/NS-124, contains the N-terminal NS1-specific 125 amino acids, since a stop codon has been introduced at nucleotide position 400 of the NS gene (11). This virus appeared to be slightly attenuated, immunogenic in mice, and capable of growing in embryonated chicken eggs and Vero and MDCK cells. Relying on the features of the PR8/NS-124 virus, we decided to insert the 137 C-terminal amino acids of the HIV-1 Nef protein comprising multirestricted immunodominant regions into the same position (i.e., 400) without deleting any part of the NS genome segment (Fig. 1). To provide posttranslational separation of the target protein (Nef fragment) from the N-terminal portion of the influenza virus NS1 protein, 51 nucleotides encoding the 2A cleavage site (NFDLLKLAGDVESNLG/P) of the foot-and-mouth disease virus (48) were introduced upstream of the Nef sequence (Fig. 1).

FIG. 1.

Schematic map of the NS genes and gene transcripts for wt A/PR/8/34 virus and recombinant PR8/NS-Nef virus. The NS gene of the PR8 wt virus (wild type NS gene) is shown as an open bar with the length in nucleotides indicated above. The position of the insertion of foreign nucleotide sequences is indicated by an arrow. The insertion of 51 nucleotides (nt) encoding the 2A cleavage site (2A; dark shaded bar) adjacent to 411 nt encoding the 137 C-terminal amino acids (aa) of the HIV-1 NL 43 Nef protein (Nef; light shaded bar) and the position of the stop codon TAA are indicated (recombinant NS gene). The viral NS1 open reading frame (ORF) is shown with the length in amino acids indicated above (NS1 mRNA). Viral NEP mRNA—equal for the wt and recombinant viruses—is shown, with the in-frame mRNA sequence shared between the NS1 and NEP ORFs indicated by an open bar. The solid bar represents the unique ORF of the NEP mRNA transcript (NEP mRNA).

Following RNP transfection of Vero cells, the plasmid-derived recombinant NS RNA was successfully rescued into viral particles. The resulting transfectant virus based on A/PR/8/34 (H1N1) was designated PR8/NS-Nef. Another recombinant vector was obtained in Vero cells by genetic reassortment of PR8/NS-Nef (H1N1) virus and the mouse-adapted influenza A/Aichi/1/68 (H3N2) virus strain. The Aichi/NS-Nef (H3N2) virus inherited gene segments encoding surface glycoproteins from the Aichi strain and all other genes, including the recombinant NS-Nef gene, from the recombinant PR8/NS-Nef virus. Genotyping of reassortants was performed by RT-PCR amplification and comparative restriction analysis of cDNA copies derived from each genome segment. The genetic stability of influenza/NS-Nef virus vectors was confirmed by sequence analysis of NS genes following five serial passages in Vero cells. No revertants were found (data not shown).

Nef antigen is expressed in infected cells.

The presence of the Nef antigen in infected Vero cells was determined by using two MAbs, HIV-1_IR-CSF Nef and HIV-1_IIIB Nef. Twenty hours postinfection, the acetone-fixed infected Vero cells were analyzed by immunofluorescence. A distinct immunofluorescence signal was observed in cells infected with either the PR8/NS-Nef or Aichi/NS-Nef virus (Fig. 2A and C), whereas no immunufluorescence signal was detected with the control viruses PR8/NS-124 and wt PR8 (Fig. 2B). The Nef fragment was predominantly found in the form of granules distributed in the cytoplasm of infected Vero cells (Fig. 2D). No difference in intracellular localization of the Nef antigen was detected in the cytoplasm of Vero cells infected with either the PR8/NS-Nef or the Aichi/NS-Nef virus irrespective of the Nef-specific MAbs used.

FIG. 2.

Immunofluorescence analysis. Vero cells were infected with recombinant PR8/NS-Nef (A), PR8 wt (B), and Aichi/NS-Nef (C) viruses with an MOI of 0.05. Twenty-four hours after infection, cells were collected, fixed on cover slides with acetone, and incubated with anti HIV-1_IIIB Nef mouse MAb followed by incubation with fluorescein isothiocyanate conjugated to goat anti-mouse IgG, as described in Materials and Methods. An analogous immunofluorescence signal was observed when infected cells were incubated with anti-HIV-1_IR-CSF Nef mouse MAb (data not shown). (D) The intracellular localization of the Nef antigen is visible at a higher magnification.

We next evaluated whether the expressed Nef antigen present in the cytoplasm existed in a cleaved form distinct from the NS1 N-terminal sequence according to the putative function of the 2A cleavage site. For this purpose, electrophoresis of ³⁵S-labeled viral proteins as well as Western blot analysis was performed. Figure 3A shows that the cleaved form (approximately 16.5 kDa) of the Nef-derived polypeptide was present in infected Vero cells 6 h postinfection. In contrast, only uncleaved NS-Nef antigen (32 kDa) was detected in the cell lysate obtained 24 h postinfection from Vero cells, as determined by Western blot analysis utilizing either the Nef-specific MAb (Fig. 3B) or the NS1-specific polyclonal rabbit antiserum (Fig. 3C). In these experiments, neither cleaved Nef fragment nor truncated NS1 polypeptide was found (Fig. 3B and C).

FIG. 3.

Sizes of the wt and recombinant NS1 proteins. (A) Confluent monolayers of Vero cells were infected with recombinant PR8/NS-Nef and PR8 wt viruses at an MOI of 5. Six hours postinfection, the cells were labeled with [³⁵S]methionine and [³⁵S]cysteine (Amersham). Cell extracts were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis on 13% polyacrylamide gels containing 5 M urea. Lanes 1 and 2, mock-infected Vero cell extracts; lane 3, PR8 wt-infected Vero cell extract; lane 4, PR8/NS-Nef-infected Vero cell extract. (B and C)Viral proteins derived from the infected Vero cells 24 h postinfection were separated on 16% polyacrylamide gels, transferred to a nitrocellulose membrane, and detected by Western blot analysis with anti HIV-1_IR-CSF Nef mouse MAb (B) and rabbit anti-NS hyperimmune serum (C). Lanes 1 and 4, PR8/NS-Nef; lanes 2 and 3, PR8 wt.

Recombinant influenza and NS-Nef viruses replicate efficiently in Vero and MDCK cells and embryonated chicken eggs.

The reproduction capacity of the recombinant influenza PR8/NS1-Nef and Aichi/NS-Nef viruses has been evaluated in Vero and MDCK cell lines and 11-day-old embryonated chicken eggs. Virus yields in supernatants of infected Vero cells (MOI, 0.01) and the allantoic fluid from infected embryonated chicken eggs (infection dose, 10² PFU/egg) were determined 48 h postinfection in a standard plaquing assay on Vero cells. Both influenza and Nef recombinant virus strains in Vero cells reached titers of up to 2 × 10⁷ PFU/ml, which were comparable to those of the PR8 wt strain (10⁷ PFU/ml) and the transfectant PR8/NS-124 virus as reported earlier (11). The yields of recombinant strains and the PR8/NS-124 virus were approximately 2 log units lower in MDCK cells than those achieved on Vero cells. In contrast, PR8 wt virus displayed 1.2-log-unit-higher titers on MDCK cells than in Vero cells. The growth rate of recombinant influenza/NS-Nef viruses in embryonated eggs did not significantly differ from the Vero-adapted PR8 wt strain (8 × 10⁷ PFU/ml), reaching the levels of 2 × 10⁷ PFU/ml for the PR8/NS-Nef virus and 7 × 10⁷ PFU/ml for the Aichi/NS-Nef virus. Thus, in spite of the fact that most of the Nef product had not been dissociated from the N-terminal part of NS1, recombinant influenza/NS-Nef viruses were capable of replicating efficiently in all three production substrates tested.

Recombinant influenza/NS-Nef viruses are completely attenuated in mice.

Next, we examined the potential of the recombinant PR8/NS-Nef and Aichi/NS-Nef vectors to replicate in the respiratory tracts of BALB/c mice. Female BALB/c mice (nine mice/group) were infected i.n. with 10⁶ PFU of the PR8/NS-Nef, Aichi/NS-Nef, or PR8/NS-124 virus or with 10³ PFU of the PR8 wt virus per mouse under ether anesthesia. The maximum virus titer of mice immunized with the wt PR8 was detected 4 days postimmunization and was 1.2 × 10⁶ PFU/ml of a 10% (wt/vol) lung tissue extract (Fig. 4). In the group of mice immunized with the PR8/NS-124 virus, the maximum virus load in the lungs (1.5 × 10⁵ PFU/ml) was detected at day 4 postinfection. In contrast, replication of both recombinant strains, PR8/NS-Nef and Aichi/NS-Nef, was completely abrogated in mouse lungs (Fig. 4) or nasal tissues (data not shown). In all animal experiments, no pathological events, such as body weight loss or pneumonia, were associated with Nef-expressing vectors.

Single immunization with an influenza virus vector induces a CD8⁺-T-cell response that is not boosted after repeated immunization with the same virus.

To characterize the insert (Nef peptide)- and vector (NP peptide)-specific CD8⁺-T-cell response, female BALB/c mice were immunized once or twice i.n. without narcosis with 10⁶ PFU per animal of the PR8/NS-Nef, Aichi/NS-Nef, PR8/NS-124, or PR8 wt virus according to immunization protocols outlined in Fig. 5 to 7.

FIG. 5.

Quantification of Nef peptide-specific and NP peptide-specific IFN-γ-secreting cells in murine spleen cells. Three BALB/c mice per group were immunized once or twice i.n. in the absence of anesthesia with 10⁶ PFU of influenza PR8/NS-Nef, Aichi/NS-Nef, PR8/NS-124, or PR8 wt virus/mouse as indicated. The booster immunization (imm.) was performed 21 days after priming. The single-cell suspensions obtained 10 days after immunization from the spleens of the mice were assessed for Nef peptide-specific (A) or NP peptide-specific (B) IFN-γ-secreting CD8⁺ T cells in an ELISPOT assay. Shown are the mean numbers of antigen-specific IFN-γ-secreting cells + SEM of triplicate cultures.

FIG. 7.

Quantification of Nef peptide-specific and vector NP peptide-specific IFN-γ-secreting cells in urogenital single-cell populations of immunized mice. Three BALB/c mice per group were immunized twice i.n. in the absence of anesthesia with 10⁶ PFU of influenza PR8/NS-Nef, Aichi/NS-Nef, or PR8/NS-124 virus/mouse as indicated. The booster immunization (imm.) was performed 21 days after priming. The single-cell suspensions derived 10 days after the second immunization from digested urogenital tracts (vagina, cervix, uterine horns, and urethra) of immunized mice were assessed for Nef peptide-specific (A) or NP peptide-specific (B) IFN-γ-secreting CD8⁺ T cells in an ELISPOT assay. Shown are the mean numbers of antigen-specific IFN-γ-secreting cells + SEM of triplicate cultures.

Mice immunized once with either the PR8/NS-Nef or Aichi/NS-Nef virus induced significant numbers of Nef peptide-specific CD8⁺ T cells in single-cell suspensions derived from spleens (139 ± 4 spots in PR8/NS-Nef-immunized mice; 137 ± 39 spots in Aichi/NS-Nef-immunized mice) (Fig. 5A) and lymph nodes (173 ± 23 spots in PR8/NS-Nef-immunized mice; 160 ± 25 spots in Aichi/NS-Nef-immunized mice) (Fig. 6A). No relevant Nef peptide-specific CD8⁺-T-cell response was determined in either compartment of mice immunized with the PR8/NS-124 and PR8 wt viruses (the number of spots was always lower than 13) (Fig. 5A and 6A).

FIG. 6.

Quantification of insert Nef peptide-specific and vector NP peptide-specific IFN-γ-secreting cells in lymph nodes draining the respiratory tracts of immunized mice. Three BALB/c mice per group were immunized once or twice i.n. in the absence of anesthesia with 10⁶ PFU of influenza PR8/NS-Nef, Aichi/NS-Nef, PR8/NS-124, or PR8 wt virus/mouse as indicated. The booster immunization (imm.) was performed 21 days after priming. The single-cell suspensions obtained 10 days after immunization from lymph nodes draining the respiratory tracts (mediastinal and retrobronchial lymph nodes) of the mice were assessed for Nef peptide-specific (A) or NP peptide-specific (B) IFN-γ-secreting CD8⁺ T cells in an ELISPOT assay. Shown are the mean numbers of antigen-specific IFN-γ-secreting cells + SEM of triplicate cultures.

When vector (NP peptide)-specific CD8⁺-T-cell responses were compared, similar numbers of specific CD8⁺ spleen cells were found in all groups of mice tested (Fig. 5B). In contrast to the systemic compartment (spleens), significant differences were obtained in the mucosa-associated respiratory lymph nodes. The replication-competent PR8/NS-124 virus induced a markedly higher frequency of NP peptide-specific CD8⁺ T cells, whereas recombinant influenza and NS-Nef viruses and the pathogenic PR8 wt virus induced lower magnitudes of NP peptide-specific CD8⁺ T cells (Fig. 6B).

Attempts to boost the antigen-specific CD8⁺-T-cell response by a second immunization using the same viral strain were not successful. As shown in Fig. 5 and 6, the frequency of antigen-specific CD8⁺ T cells clearly decreased after repeated exposure of mice to the same virus in spleens as well as draining lymph nodes. These results were consistent, irrespective of the virus used for immunization as well as the specificity measured (Nef peptide versus NP peptide).

Strong secondary CD8⁺-T-cell response is induced after priming with a recombinant H1N1 subtype vector and boosting with the H3N2 subtype vector.

Since both recombinant vectors, PR8/NS-Nef and Aichi/NS-Nef, contained the same recombinant NS gene and therefore expressed the same truncated Nef antigen, we investigated whether successive immunizations with these vectors would induce a Nef peptide-specific secondary CD8⁺-T-cell response. In these experiments, one group of mice was primed i.n. with the PR8/NS-Nef virus and boosted 21 days later with the Aichi/NS-Nef virus. Another group of mice was immunized with the same viruses but in the reverse order. The data shown in Fig. 5 and 6 indicate that the sequence in which the respective recombinant vectors were used for priming and boosting appeared to be crucial, since we consistently observed that priming with Aichi/NS-Nef (H3N2) followed by boosting with PR8/NS-Nef (H1N1) induced significantly lower numbers (approximately the range of the primary CD8⁺-T-cell response) of the Nef peptide- and NP peptide-specific CD8⁺ T cells in spleens and draining lymph nodes than did the reverse order of immunization. A strong secondary antigen-specific CD8⁺-T-cell response was detected in both of the compartments tested after the mice were primed with the recombinant PR8/NS-Nef (H1N1) vector followed by a boost with the H3N2 subtype Aichi/NS-Nef vector. In this case, Nef- and NP peptide-specific secondary responses were approximately 1.5 to 3 times higher than after a single immunization (Fig. 5 and 6).

Strong secondary Nef-specific CD8⁺-T-cell response is detectable in a distant mucosal site.

We next addressed the question of whether the antigen-specific cellular immune response detected in the respiratory tract, which is a site of immunization, could also be detected according to the concept of a common mucosal immune system in a distant mucosal site. For this purpose, single-cell suspensions derived from the urogenital tract were obtained from immunized mice. Two immunizations were necessary before significant numbers of Nef peptide-specific CD8⁺ T cells could be detected. As expected, the strongest Nef peptide-specific CD8⁺-T-cell response was detected when the mice were primed i.n. with PR8/NS-Nef (H1N1) virus and subsequently boosted with the Aichi/NS-Nef (H3N2) virus (342 ± 18 IFN-γ secreting cells/10⁶ cells; Fig. 7A). This immunization protocol was also found to induce the strongest NP peptide-specific CD8⁺-T-cell response (Fig. 7B).

Nef-specific IgG is detected in sera of immunized mice.

As described for the detection of T-cell responses, mice were utilized to assess the Nef-specific serum antibody response. The reactivity of mouse serum IgG with the GST-Nef fusion peptide was tested by ELISA. The mice received a third i.n. immunization, since no significant Nef-specific IgG was detected 2 weeks after the second immunization. Nef-specific antibodies were detected only in groups of mice which had been successively immunized with H1N1 and H3N2 vectors (Fig. 8). The highest level of Nef-specific IgG compared with the control group, which had been immunized three times i.n. with PR8 wt virus, was detected in mice immunized twice i.n. with 10⁶ PFU of the PR8/NS-Nef (H1N1) virus and boosted with 10⁶ PFU of the Aichi/NS-Nef (H3N2) virus (Fig. 8).

FIG. 8.

Nef-specific IgG in sera of mice. Mice were primed i.n. with either 10⁶ PFU of the PR8/NS-Nef (H1N1) or Aichi/NS-Nef (H3N2) virus/ml and were boosted 3 weeks later with the same vector. The third immunization was performed following three more weeks utilizing the vector of a different subtype. The control group was immunized three times with the PR8 wt virus. The reactivities of serum samples (obtained 2 weeks after the third immunization) with the GST-Nef fusion peptide were determined by ELISA. OD 492, optical density at 492 nm.

DISCUSSION

It is now well accepted that live viral vaccines administered by mucosal routes efficiently stimulate humoral and cell-mediated immune responses in both mucosal and systemic compartments. The main problems associated with the application of infectious viruses as vectors include residual virulence, potential side effects, poorly controllable persistence, and preexisting immunity to viral vectors. In this respect, implementation of live attenuated influenza vaccines as vectors seems to be a promising strategy for mucosal immunization against various pathogens (17, 47).

The major challenge in generating an effective live vector is to obtain an optimal balance among attenuation, safety, adequate antigen expression, and immunogenicity. Recombinant influenza virus vectors expressing foreign antigens in the context of HA or NA described so far were mostly virulent in mice, since they replicated efficiently in murine respiratory tracts (20, 46, 55). We suggest in the present study an alternative approach enabling the generation of attenuated influenza virus vectors containing insertions of foreign amino acid sequences in the NS1 protein. This strategy is based on previous findings documenting decreasing virulence of influenza viruses with increasing deletions at the carboxyl end of the NS1 protein due to impairment of its type I IFN′s antagonist function (11, 19).

The recombinant influenza virus A/PR/8/34 containing a modified NS gene segment (PR8/NS-Nef) was rescued by utilizing the NS reverse-genetics system in Vero cells (11). This PR8/NS-Nef virus codes for the N-terminal NS1-specific 125 amino acids followed C terminally by 17 amino acids of the 2A cleavage site derived from foot-and-mouth disease virus fused to the last 137 C-terminal amino acids of HIV-1 (NL43) Nef. Elimination of the first 68 amino acids of the Nef protein was performed with the purpose of excluding the myristoylation site and other domains associated with pathogenic properties of the multifunctional HIV-1 Nef protein (1, 23).

One of the advantages of the NS rescue system is the possibility of transferring the engineered NS gene to other influenza virus subtypes by the simple method of genetic reassortment. In this manner, another vector (Aichi/NS-Nef) belonging to the H3N2 subtype but containing the same recombinant NS gene was obtained. Recombinant PR8/NS-Nef (H1N1) and Aichi/NS-Nef (H3N2) viruses were confirmed to be genetically stable. Both vectors displayed normal growth characteristics in IFN-deficient Vero cells and were only slightly attenuated in MDCK cells or embryonated chicken eggs. These findings indicate that despite the fact that the majority of the Nef antigen was not cleaved from the NS1 protein at the 2A cleavage site (Fig. 3), the resulting fusion peptide was capable of accomplishing its IFN antagonist function (18, 19).

Nevertheless, both influenza virus vectors were replication deficient in mice. This hyperattenuation phenotype of both recombinant viruses indicates that introduction of additional amino acids downstream of position 125 of the NS1 protein could have affected some function of the NS1 protein, since PR8/NS-124 virus encoding the same size NS1 protein grew efficiently in mouse respiratory tracts (Fig. 4). This might be attributed to the low efficiency of the 2A cleavage site, although the direct effect of the Nef polypeptide interacting with some intracellular components cannot be excluded (23).

Despite the results showing that both vectors were completely attenuated in mice, our data indicate that influenza/NS-Nef virus vectors were capable of inducing a primary CD8⁺-T-cell response directed to the Nef polypeptide in spleens and in the lymph nodes draining the respiratory tract. Moreover, in accordance with the previously published results for the influenza virus lacking the NS1 open reading frame (56), the vector (NP peptide)-specific CD8⁺-T-cell responses in spleens and respiratory lymph nodes of mice induced by both vectors were in the range of those induced by the virulent PR8 wt virus.

We also addressed the question of whether primary insert (Nef peptide)-specific CD8⁺-T-cell response could be boosted by a second immunization. In fact, mice primed with the PR8/NS-Nef (H1N1) vector and subsequently boosted with the Aichi/NS-Nef (H3N2) vector induced approximately 1.5- to 3-times-higher magnitudes of Nef or NP peptide-specific CD8⁺ T cells than the primary response (Fig. 5, 6). In addition, a significant secondary Nef or NP peptide-specific CD8⁺-T-cell response was detected in a distant mucosal site—the urogenital tracts of immunized mice (Fig. 7). There is no plausible explanation of why this order of immunization (H1N1-H3N2) and not the reverse (H3N2-H1N1) was effective in boosting secondary CD8⁺-T-cell responses. Thus, in addition to data demonstrating that the immunogenicity of influenza virus vectors could be strongly enhanced by boosting immunizations with antigenically distinct vaccinia virus vectors expressing the same foreign antigen (38, 39, 51), our results indicate that it is possible to achieve a similar effect by utilizing influenza virus vectors belonging to different antigenic subtypes.

It might be possible that the generation of secondary responses in the urogenital tract after i.n. immunization, especially in the absence of anesthesia, is acting through the nasal-associated lymphoid tissue (13, 59), a site of primary deposition of the recombinant vector, since no virus particles were detected in the lungs of mice immunized with recombinant influenza/NS-Nef vectors. The migratory behavior of immune cells as postulated by the concept of the common mucosal immune system has been most frequently documented for mucosal B cells but only rarely for activated mucosal T cells (36). However, the selective induction and homing of antigen-specific CD8⁺ T lymphocytes expressing α4β7 adhesion molecules derived from the blood circulation upon mucosal immunization with attenuated simian immunodeficiency virus was reported recently (8). I.n. immunization of DNA-lipid complexes resulted in encoded antigen-specific CTLs also being localized in the distal genital lymph nodes (28). In addition, mice immunized i.n. with an adenovirus vector expressing glycoprotein B of herpes simplex virus (HSV) and challenged intravaginally up to 18 months later with a heterologous HSV type 2 (HSV-2) developed strong anti-HSV-2 CTL memory responses in the distal urogenital tract (16).

It is of considerable interest that in mice, replication-deficient PR8/NS-Nef and Aichi/NS-Nef viruses were capable of inducing an antibody response to the viral HA (hemagglutination inhibition titers ranged between 16 and 32 [data not shown]) as well as to the intracellularly localized Nef antigen, although Nef-specific IgG antibody titers were rather weak and three successive immunizations with both influenza virus vectors were necessary. The immunogenic potential of hyperattenuated influenza and NS-Nef virus vectors might be explained by the fact that viruses containing truncated forms of the NS1 protein induce high levels of type I IFNs in vivo (18, 19). In fact, we have found that Nef-expressing vectors, as well as PR8/NS-124 virus, induced markedly higher levels of type I IFNs in serum following intraperitoneal and even i.n. immunization of mice than the corresponding wt parent viruses (data not shown). The immunomodulating properties of type I IFNs have been demonstrated elsewhere (10, 30, 54).

The Nef protein is an important accessory product of HIV, possessing important biochemical functions, and it is essential for viral pathogenicity in vivo (26). Delivery vectors expressing selected Nef or Nef-derived sequences were suggested to be promising for HIV-1 vaccine design due to the high immunogenic potential of the Nef protein (6, 9, 21, 52). These vectors include DNA-based vaccination, live viruses, viruslike particles, bacteria, lipopeptide, and/or polytope vaccine constructs (12, 14, 24, 31, 32, 53, 57, 58). Hyperattenuated influenza and NS-Nef viruses are capable of inducing Nef-specific B- and T-cell immunity and thus might be considered as an alternative mucosal vaccination approach against human HIV-1 infection. Moreover, it might be possible to create a recombinant NS gene comprising selected dominant and/or subdominant protective (44) conserved B-cell and CTL epitopes. The desired recombinant NS gene could be transferred by genetic reassortment to several live influenza virus vaccinal strains of different subtypes for subsequent boosting immunizations. These influenza virus vectors might be used in combination with any other vector expressing analogous antigens to ensure maximal booster effect. Generation of attenuated influenza virus NS vectors offers the possibility of obtaining novel recombinant vaccines with a nearly optimal balance of safety and immunogenicity directed against a broad range of pathogens.