Introduction

Highly pathogenic H5N1 avian influenza virus resulted in a windstorm of disease outbreaks in birds in China and other Asian countries in 2004. It broke through the species barrier and infected humans, resulting in high mortality [1, 2]. The high variability of AIV and its incidental outbreak on hosts make it difficult to explain AIV’s ecological characteristics [3, 4] and provide a significant challenge for AI’s prevention and control. The use of vaccines is still the most effective means to control avian influenza virus (AIV). AIV could escape host immune system through continuous antigenic mutation, which makes the development of AI vaccine become an endless task for researchers worldwide and new directions for future influenza vaccines are being explored. Injecting human and animal subjects with naked DNA as a vaccine, which contains special foreign DNA, has shown great efficiencies and advantages [5–7].

The type A influenza virus genome has eight RNA segments encoding 11 proteins which provide alternative antigens for influenza virus vaccines. Several antigens have been used as vaccine candidate proteins. The surface glycoprotein hemagglutinin (HA) contains the major antigenic determinants and antibodies to HA can neutralize viral infectivity. Another surface glycoprotein neuraminidase (NA) is a tetrameric enzyme which cleaves terminal sialic acid residues from oligosaccharide chain [8] and then releases progeny viruses from infected cells. Antibodies to NA result in reduction of pulmonary virus titers which reduce viral infectivity. The proteins such as nucleoprotein (NP), matrix protein (M1), and the ion channel protein (M2) are highly conserved between and within different subtypes [9–11] and, therefore, are attractive target antigens for vaccines that provide cross-strain protection [12–16].

Attempts to control influenza virus have been made by immunization with DNA vaccines [17–20], but these papers are concerned with one or two genes. The present study reports a comparison of the relative immunogenicity and protection against a lethal influenza virus challenge after immunization. However, these antigens other than HA and NA could only provide weak protection. So, we used multiple immunizations to improve the protective immunity of plasmid DNAs containing NP, M1, and M2. In this study, we explored an appropriate immunization alternative or combination to provide some clues as to how to prevent highly pathogenic avian influenza (HPAI) virus.

Materials and methods

Viruses, mice, immunization, and infection

Influenza viruses used in this study included a mouse-adapted A/PR/8/34(PR8) (H1N1) and a H5N1 influenza virus A/Chicken/Henan/12/2004 isolated in 2004 which was passaged and adapted in mouse as described previously [21]. These viruses were frozen at −70°C until use. The H5N1 viruses were used in a biosafety level 3 containment facility in the Wuhan Institute of Virology.

Specific-pathogen-free female BALB/c mice, aged 6–8 weeks old, were purchased from the Center for Disease Control and Prevention in Hubei Province, China. Mice were bred in the Animal Resource Center at the Wuhan Institute of Virology, Chinese Academy of Sciences. All mice were maintained in specific-pathogen-free conditions prior to infection.

In vivo electroporation was carried out according to the method described by Aihara and Miyazaki [22]. Mice were immunized with plasmid DNA dissolved in 30 μl of Tris–EDTA buffer at a dosage of 50 μg by injection into the quadriceps muscles. After injection in the right quadriceps muscle, a pair of electrode needles with 5 mm apart was inserted into the muscle to cover the DNA injection sites and electric pulses were delivered using an electric pulse generator (Electro Square Porator T830 M; BTX, San Diego, CA). Three pulses of 100 V each, followed by three pulses of the opposite polarity, were delivered to each injection site at a rate of one pulse per second. Each pulse lasted for 50 ms. A booster was given 2 weeks after the first immunization. The non-immunized mice were set up as controls. Two weeks after the last immunization, the mice were anesthetized and challenged with 20 μl of the viral suspension containing 5LD50 influenza virus A/Chicken/Henan/12/2004(H5N1) or A/PR/8/34 (H1N1) by intranasal drip.

Plasmid DNAs and peptide

Plasmids pCAGGSP7/HA, pCAGGSP7/NA, pCAGGSP7/NP, pCAGGSP7/M1, and pCAGGSP7/M2 were constructed by cloning the PCR products of HA, NA, NP, M1, and M2 gene from the A/Chicken/Henan/12/2004(H5N1) influenza virus strain into the plasmid expression vector pCAGGSP7, respectively, as described previously [23]. The plasmids were propagated in Escherichia coli XL1-blue bacteria and purified using QIAGEN purification kits (QIAGEN-tip 500). The peptide RAVKLYKKLKRE for M1 protein [24] and the peptide TYQRTRALV for NP protein [25], which were used for IFN-γ ELISPOT assay, were synthesized by Shanghai Sangon Biological Engineering Technology & Services Co., Ltd.

Specimens

Three days after the challenge, five mice from each group were randomly taken out for sample collection. The mice were anaesthetized with chloroform and then bled from the heart with a syringe. The sera were collected from the blood and used for IgG antibody assay. After bleeding, the mice were incised ventrally along the median line from the xiphoid process to the point of the chin. The trachea and lungs were taken out and washed 3 times by injecting with a total of 2 ml of PBS containing 0.1% BSA. The bronchoalveolar wash was used for virus titration after removing cellular debris by centrifugation [26].

Antibody (Ab) assays

Titers of IgG Abs against the respective viral proteins were measured by using nitrocellulose membranes, to which HA, M1, M2, and NP molecules in the H5N1 virus-infected Madin Darby canine kidney (MDCK) cells were transferred after separating each lysate on SDS-PAGE. The MDCK cells grown in 25-cm2 bottles were inoculated with diluted virus suspension at a MOI of 0.01. After incubation at 37°C for 1 h, virus suspension was removed and 5 ml of DMEM with 2% serum was added. The infected cells were lysed in the SDS loading buffer 36 h after infection and the lysates separated by SDS-PAGE, before being transferred to a nitrocellulose membrane. The membrane was blocked with the non-fat milk, dried, and cut into longitudinal slips of 2 mm width. Serial twofold dilutions of sera from each group of immunized mice were prepared and each diluted serum sample was used to incubate each membrane slip. The highest serum dilution, giving a positive staining of slip at the site corresponding to the molecular weight of each viral protein, represented the immunoblotting Ab titer of each molecule. The serum neutralization activity was measured by hemagglutination inhibition assays [27]. Receptor destroying enzyme (RDE)-treated sera were serially diluted (twofold) in V-shaped 96-well plates. Four HA units of viral antigen were added to the test and incubated at room temperature for 15 min, followed by the addition of 0.5% RBCs and incubation at room temperature for 30 min. The reciprocal of the highest serum dilution that completely inhibits hemagglutination was considered the HI titer. The inhibition assay of NA activity by Ab (NI assay) was performed according to the method described by WHO [27]. The NI assay was performed using fetuin (Sigma) as substrate. The A/Chicken/Henan/12/2004(H5N1) virus grown in the 10-day-old chicken embryos and stored as allantoic fluid at −80°C was employed as enzyme source. The NI titer of the mouse antiserum was defined as the dilution of the serum inhibiting 50% of NA activity.

Virus titrations

The bronchoalveolar washing, diluted 10-fold serially starting from a dilution of 1:10, inoculated to MDCK cells at 37°C for 2 days, so as to examine cytopathic effect. The virus titer of each specimen, expressed as the 50% tissue culture infection dose (TCID50), was calculated by the Reed-Muench method. The virus titer in each experimental group was represented by the mean ± SD of the virus titer per ml of specimens from five mice in each group.

IFN-γ ELISPOT assay

Spleen cells were isolated from mice for ELISPOT assays at 2 weeks after the last vaccination. According to the instruction manual (U-CyTech, Netherlands), 96-well PVDF plates (Millipore, Bedford, MA) were coated with 100 μl of 10 μg/ml rat anti-mouse IFN-γ Ab in PBS and incubated at 4°C overnight. The plates were washed 3 times with sterile PBS and then blocked with 200 μl of blocking solution R and incubated at 37°C for 1 h. Next, 1 × 105 lymphocytes isolated from the spleen cells were added to the wells in triplicate, stimulated with 2 μg/ml of influenza virus peptide, and incubated at 37°C for 18 h. The lymphocytes were then removed, and 100 μl of biotinylated anti-mouse IFN-γ Ab was added to each well and incubated at 37°C for 1 h. Subsequently, 100 μl of properly diluted Streptavidin-HRP conjugate solution was added and incubated at room temperature for 2 h after washing 5 times with PBS. Finally, the plates were treated with 100 μl of AEC substrate solution and incubated at room temperature for 20 min in the dark. The reaction was stopped by washing with dematerialized water. The plates were air-dried at room temperature and read using an ELISPOT reader (Bioreader 4000; Bio-sys, Germany). Medium backgrounds were consistently <10 SFC per 106 splenocytes.

Statistics

The results of the test groups were evaluated by Student’s t-test; if P-value is less than 0.05, the difference was considered significant. The survival rates of the mice in the test and control groups were compared by using Fisher’s exact test.

Results

Comparison of the ability of various viral DNA vaccines to provide protection against lethal H5N1 influenza virus infection

Two hundred and sixty mice were randomly divided into groups of 13. Mice were immunized with 50 μg of HA, NA, NP, M1 DNA, or M2 DNA by electroporation, or remained unimmunized to serve as the negative control. To compare the ability of various DNA plasmids to protect against homologous virus infection, the different DNA plasmids were immunized for different times. The mice were immunized 1–5 times, 2 weeks apart (Table 1). Two weeks after the final vaccination, the mice were challenged with lethal avian influenza H5N1 virus. The mice were observed for 21 days after the challenge; body weights and survival were evaluated as the ability of each plasmid DNA to protect mice against a lethal H5N1 influenza virus challenge.

Table 1 Protection and antibody responses in mice immunized with plasmid DNAs encoding various viral proteins from the H5N1 influenza virusa
Full size table

As shown in Table 1, the survival rate of the mice immunized once with 50 μg of HA-DNA against virus challenge was 100%. The survival rates of the mice immunized once and twice with 50 μg of NA-DNA against virus challenge were 80 and 100%, respectively. The survival rates of the mice immunized 3–5 times with 50 μg of NP-DNA, against virus challenge were 60, 60, and 85%, respectively. A higher survival rate was observed in the mice immunized 5 times compared with 4 or 3 times, but the difference was not significant. The mice immunized with the M1 plasmid DNA showed 0, 15, and 25% protection rate, respectively after 3, 4, and 5 times of immunization. The survival rates slightly increased with immunization times. The mice immunized with the M2 plasmid DNA showed 0, 5, and 5% survival rate, respectively after 3, 4, and 5 times of immunization. The unimmunized mice died within 10 days after challenge.

The residual lung virus titers and the bodyweight loss rates of the mice immunized with HA DNA or NA DNA groups were significantly lower than those of mice in the control group (P < 0.05). On the other hand, these mice immunized with NP, M1, or M2 DNA suffered more weight loss than the HA, NA DNA groups. The mice immunized 3–5 times with 50 μg of NP-DNA had bodyweight loss rates significantly lower than those of mice in the control group (P < 0.05). Meanwhile, the mice immunized with M1, M2, or NP DNA did not show a significant decrease of lung viral titers as compared with the control group.

The results suggest that both HA and NA DNA are more protective against influenza virus infection, and the HA DNA is followed by the NA DNA. The protection effect of NP DNA is better than that of the M1 DNA. But the M2 DNA could not effectively protect mice (Table 1).

Antibody responses in the mice immunized with various DNA vaccines

As mentioned above, BALB/c mice were immunized with 50 μg of HA, NA, NP, M1, M2 DNA by electroporation. The sera were examined for specific IgG Abs to each protein by immunoblotting assay, HI assay, or NI assay. In the mice immunized with HA DNA, anti-HA IgG Ab responses were detected in the sera by immunoblotting assays. The specific antibodies were further confirmed by HI assay. In the mice immunized with NA DNA, a high anti-NA IgG Ab response, in which the secondary response was much higher than the primary one, was detected in the serum by NI assay. The mice immunized with Ml, M2, NP all induced detectable Ab responses by immunoblotting assay (Table 1), and the Ab titers increased with the immunization times. Thus, the mice immunized with 50 μg of various DNA all induced specific Ab response.

Homologous and heterologous protection against lethal H5N1 and H1N1 influenza virus challenge in the mice immunized with the plasmid DNA mixtures containing NP- and M1-expressing DNAs

To explore if the mixture of M1 and NP DNA could provide enhanced protection, we immunized the mice with the mixture of 50 μg M1 and 50 μg NP DNA 3, 4, and 5 times by electroporation, 20 mice per group, using unimmunized mice as the negative control. Two weeks after the final vaccination, the mice were challenged with a lethal dose (5 × LD50) of avian influenza virus A/Chicken/Henan/12/2004 (H5N1) by intranasal drip. The mice were observed for 21 days after the challenge; body weights and survival were evaluated as the ability of the plasmid DNA mixture to protect mice against a lethal challenge with H5N1 influenza virus.

The survival rates of the mice immunized 3, 4, and 5 times were 70, 85, and 95%, respectively; the unimmunized mice all died after the challenge. The survival rates of the immunized groups increased with immunization times, but there were no significant differences among the groups, which was the same as the lung virus titers. The weight losses of the immunized groups were still lower than the control group (P < 0.05) (Table 2).

Table 2 Homologous and heterologous protection against lethal H5N1 or H1N1 (PR8) virus challenge in mice immunized with plasmid DNA mixture of M1 and NP
Full size table

In the mice immunized with plasmid DNA mixtures (NP and M1), the sera were examined for specific IgG Abs to each protein by immunoblotting assay. High anti-NP or anti-M1 IgG Abs responses were induced in the sera and the Ab titers increased with the immunization times. The level of the anti-NP or anti-M1 Ab response in the mice immunized with the plasmid mixtures was not significantly different from that in the mice immunized with a single NP-expressing DNA or M1-expressing DNA (Table 2).

To test the ability of M1 and NP DNA mixtures to protect mice against heterologous H1N1 influenza virus, the mice immunized 5 times were challenged with a lethal dose (5 × LD50) of influenza A/PR/8/34 (H1N1) by intranasal drip. The results showed that the immunized mice survived 80%; all the mice in the control group died. The highest weight loss rate of the immunized group was 15.5%, which was lower than that of the control group (38.1%, Table 2). No differences in the lung titers were observed between the immunization group and the control group. It showed the mixture of M1 and NP DNA could not only protect the mice effectively against the challenge of the homologous virus but also against the heterologous virus.

In another experiment, 10 mice immunized with the M1 and NP DNA mixtures at a dosage of 50 μg for each component DNA by electroporation 5 times, were challenged with a lethal dose (5 × LD50) of avian influenza virus A/Chicken/Henan/12/2004(H5N1) 180 days after the final vaccination. The immunized mice showed a 50% survival rate, as compared with the 0% survival rate of the unimmunized mice. In the meantime, the sera were examined for specific IgG Abs to each protein by immunoblotting assay. Anti-NP or anti-M1 IgG Abs responses were still detected in the sera and the Ab titers decreased compared to the level detected 2 weeks after the mice were immunized 5 times. The lung virus titers of the immunized mice showed no significant differences with the control group. But the weight loss of the immunized group were lower than the control group (P < 0.05) (Table 3). This suggests that inoculation of the mixture of M1 and NP plasmid DNA could provide long-term protection against H5N1 virus challenge.

Table 3 Long-lasting protection against lethal H5N1 virus challenge in mice immunized with plasmid DNA mixture of M1 and NPa
Full size table

Cell-mediated immunity

Cellular immune responses to DNA vaccine were assessed by measuring IFN-γ secretion in mouse splenocytes. BALB/c mice, aged 6–8 weeks old, were immunized with NP, M1, or a mixture of NP and M1 DNA at a dosage of 50 μg for each component DNA by electroporation 3 times (on days 0, 14, and 28), 4 times (on days 0, 14, 28, and 42), and 5 times (on days 0, 14, 28, 42, and 56), as described above. Splenocytes were harvested 14 days after the final immunization. The number of IFN-γ-secreting splenocytes was calculated as the average number of spots in the triplicate stimulant wells.

As indicated in Fig. 1, compared with the unimmunized control groups, a significant number of NP- and M1-specific IFN-γ-secreting splenocytes were detected in all the immunized groups. The number of IFN-γ-secreting splenocytes in M1 DNA group increased with the immunization times; at least a twofold increase was achieved in the mice immunized 5 times compared with those immunized 3 times. Similar results were seen in the NP DNA group. The mice immunized with NP DNA 5 times produced a higher level of IFN-γ-secreting splenocytes than those immunized 3 or 4 times. With the same immunization times, the NP DNA group produced more spots than the M1 DNA group.

Fig. 1
figure 1

Detection of IFN-γ secreted from splenocytes by ELISPOT assays. The splenocytes were stimulated with 2 μg/ml of peptide. Data shown are the mean ± SD of three independent experiments from different animals. Significance was defined as a P-value of <0.05. M1: splenocytes of mice immunized with M1 DNA stimulated with M1 peptide. M1 + NP(M1): splenocytes of mice immunized with the mixture of M1 and NP DNA stimulated with M1 peptide. NP: splenocytes of mice immunized with NP DNA stimulated with NP peptide. M1 + NP(NP): splenocytes of mice immunized with the mixture of M1 and NP DNA stimulated with NP peptide. Control: mice with no immunization

Full size image

The splenocytes of the mixture of M1 and NP group were stimulated with the M1 peptide or NP peptide. The results showed that the number of M1 or NP peptide-specific IFN-γ-secreting splenocytes increased with immunization times. Also, the number of IFN-γ-secreting splenocytes in the mixture of M1 and NP group stimulated with the NP peptide was higher than that of the group stimulated with the M1 peptide (Fig. 1).

Further, compared to the same immunization times, regardless of 3, 4, or 5 times, when stimulated with the NP peptide, the mixture of M1 and NP groups elicited more spots than the NP groups. The same situation was found in the M1 and NP mixture group when stimulated with the M1 peptide. But there was no significant number of IFN-γ-secreting cells produced when the NP groups were stimulated with the M1 peptide; in the same way, the M1 groups showed no spots when stimulated with the NP peptide. Only a few non-specific spots were detected in the control groups (spots ≤10/106 cells). The spots of plate background containing splenocytes in the absence of mitogens or nominal antigens were the same as those in the control groups. And the number of positive non-specific IFN-γ spots (concanavalin stimulated) was up to 2000/106 cells (data not shown).

In another experiment, the mice were immunized with the M1 and NP DNA mixtures at a dosage of 50 μg for each component DNA by electroporation 5 times. Splenocytes were harvested 180 days after the final immunization. The splenocytes of the mixture of M1 and NP group were stimulated with the M1 peptide or NP peptide. The results showed that IFN-γ-secreting splenocytes of M1 or NP peptide specific still exist although the data showed the number of specific IFN-γ-secreting splenocytes decreased (Fig. 2).

Fig. 2
figure 2

Long-lasting specific IFN-γ secreted from splenocytes in mice immunized with plasmid DNA mixture of M1 and NP 5 times. Splenocytes were harvested 180 days after the final immunization. M1: stimulated with M1 peptide. NP: stimulated with NP peptide. Control: mice with no immunization

Full size image

Discussion

The outbreaks of highly pathogenic avian influenza subtype H5N1 have occurred annually since 2003 and caused human infection, which accelerates the efforts to devise protective strategies against the widespread viruses. Plasmid DNA vaccines have been considered as alternatives to traditional inactivated influenza virus vaccine that has been a main measure to prevent avian influenza for a long time. An effective influenza DNA vaccine could be developed by determining the most protective viral protein-expressing DNAs and then by preparing mixtures of the effective protein-expressing DNAs. The results of the present study show that vaccination of mice with the HA or NA DNA successfully induced the antibody responses that protected the immunized mice against subsequent influenza virus challenges (Table 1). The M1- and NP-expressing DNA provided partial protection for the mice, and M2-expressing DNA failed to effectively protect the mice.

A protective effect of M2-specific antibody has been demonstrated after using the monoclonal antibody [28]. In this paper, mice were immunized with H5N1 M2 DNA 3, 4, and 5 times. Although we detected the anti-M2 antibody, the vaccine failed to protect the mice from influenza virus infection. The failure of M2-specific Abs to protect mice from the lethal viral challenge may be due to the relatively low level of Ab titer induced by M2 DNA vaccine in this study. On the other hand, M1 DNA vaccine could partially protect mice from the influenza virus challenge, inducing cellular immunity responses and a high level of antibody responses. The group immunized 5 times survived more than the groups immunized 3 and 4 times, which suggested the potential to enhance protection effect of M1 when combined with a certain adjuvant. The mice immunized with 50 μg of NP-DNA 5 times received 85% survival rate against a lethal dose of influenza virus challenge. The DNA mixture of M1 and NP could provide almost complete protection (95%).

Because the antibodies to NP and M1 did not have neutralizing activity and generally could not mediate protection [29], the anti-NP and -M1 antibody levels were not related with survival. An enhanced T-cell response was important for increased protection effect. In fact, the number of IFN-γ-secreting splenocytes (Fig. 1) correlated with survival. Several studies also indicated that cellular immune responses were very important in clearing influenza A virus infections in mice and humans [30].

The lung virus titers on day 3 did not show a difference among the M1-DNA, NP-DNA, the mixture of M1 and NP DNA group, and the control group (Tables 1 and 2). The previous research showed, in mouse models, effector CTLs were first detectable in the lung on day 7 and the number reached the peak by day 9 or 10 [31]. The accumulation of CTLs correlated with clearance of the virus, which occurred by day 8 or 9, since the T-cell recall response was delayed in its action. So the lung virus titers, 3 days post-infection, in the immunized groups were close to the control group, which explained the survival rates increased with the immunization times but lung virus titers did not.

It is interesting that the DNA mixture induced a greater number of IFN-γ-secreting splenocytes than M1 or NP alone. Lee and Sung inoculated mice with vector DNA pCI-neo and pCIN-NP; the co-administration of pCI-neo significantly increased the NP-specific CTL activity to approximately 61%. It showed that co-administration of pCI-neo augmented T-helper-cell response [32]. In another ELISPOT assay (data not shown), BALB/c mice were immunized with NP, M1, a mixture of NP and pCAGGSP7 vector, a mixture of M1 and pCAGGSP7 vector DNA at a dosage of 50 μg for each component DNA by electroporation 3 times. The number of IFN-γ-secreting cells of the mice immunized with a mixture of NP and pCAGGSP7 vector (293 spots/106) was more than the NP DNA group (178 spots/106). Likewise, the M1 and pCAGGSP7 vector DNA group produced 137 spots/106, but the M1 group produced 76 spots/106. This was in agreement with Xie’s study [33]. These data indicated that IFN-γ-secreting splenocytes were more efficiently induced and expanded in mice following immunization with the mixture of NP and M1 DNA vaccine versus NP or M1 alone. But for the serum antibody level, the differences between the simple M1\NP DNA plasmid and the plasmid mixture were not obvious.

The cross-protective cell-mediated immunity has been found in chicken H5N1 and H9N2 [34]. The notion of a “universal” vaccine for highly pathogenic strains is attractive. The NP sequence of A/Chicken/Henan/12/2004 (H5N1) differs by approximately 26 amino acids from A/PR/8/34 (H1N1), and the M1 amino acid sequence of A/Chicken/Henan/12/2004 (H5N1) is identical to A/PR/8/34 (H1N1). The present study shows that the mixture of M1 and NP DNA not only succeeded in providing protection against the homologous virus, but also against the heterologous virus. A broadly cross-protective vaccination may be usable in animals as well as humans.

Gene vaccination appears to be very effective in inducing long-lasting memory responses [7, 35, 36]. The memory cells generated are responsible for maintaining a long-lasting CTL response. Our data also showed that the serum IgG and IFN-γ cytokines response lasted for 6 months in the mice immunized with the mixture of M1 and NP 5 times.

Further vaccine development will call for a multi-component vaccine, which is more likely to be effective in eliciting the versatile immune responses and less likely to cause immune failure to the mutant viruses. This study directly compared different genes of H5N1 AIV in eliciting protective immunity in the form of DNA vaccine and provided a basis for selecting multiple-component AIV vaccines. On the other hand, we used multiple immunizations to improve the protective immunity of plasmid DNAs. Thus, a multi-component vaccine with an appropriate immunization time should be developed to provide an alternative approach against potential H5N1 influenza virus pandemic. These studies are currently in progress.