The skin has been suggested as an attractive site for immunization due to the presence of potent antigen-presenting cells such as Langerhans and dermal dendritic cells (Glenn and Kenney, 2006; Hammond et al., 2001). To improve protective efficacy while reducing the antigen mass by targeting influenza antigens to the skin, intradermal (ID) immunization has been evaluated in clinical trials (Auewarakul et al., 2007; Belshe et al., 2004; Kenney et al., 2004; Khanlou et al., 2006; Van Damme et al., 2009). However, the conventional ID injection procedure requires highly trained medical personnel and is not well tolerated by vaccinees due to pain and discomfort at the site of injection (Auewarakul et al., 2007; Belshe et al., 2004; Kenney et al., 2004). Recent studies have demonstrated a promising alternative method that delivers inactivated whole-virion vaccines to the skin using microneedles, penetrating the outer layer of the skin (Kim et al., 2010; Kim et al., 2009; Quan et al., 2009; Zhu et al., 2009). This simple design could permit self-administration of vaccine by patients, possibly enabling vaccination campaigns to rapidly reach large populations (Prausnitz et al., 2009).

Conventional inactivated vaccines are produced from virus propagation in eggs. A new vaccine platform, virus-like particles (VLPs) produced in cell culture, has been shown to confer protection against highly pathogenic avian-origin influenza viruses in animal models, and can be manufactured without handling pathogenic live viruses (Bright et al., 2008; Haynes et al., 2009; Kang et al., 2009). In the present study, we investigated the immunogenicity and protective efficacy after a single vaccination using microneedles coated with dried H5 VLPs, in comparison with conventional intramuscular injection.

H5 VLPs derived from influenza A/Vietnam/1203/04 (A/VN/04) virus were produced in insect cells using recombinant baculovirus expression as previously described (Kang et al., 2009). Stainless steel microneedles were fabricated as arrays of 5 needles (Kim et al., 2010). The 700 μm length of microneedles used in this study is suitable for effective delivery of vaccine into mouse skin with a thickness of 500-600 μm (Azzi et al., 2005), because the whole microneedle is not fully inserted into the skin due to skin deformation during insertion. For vaccination in the skin, microneedles were coated on their surfaces with H5 VLPs in coating solution (1% carboxymethylcellulose (CMC) sodium salt as viscosity enhancer, 0.5% (w/v) Lutrol F-68 NF as surfactant, and 15% trehalose as stabilizer) and then air dried (Kim et al., 2010). A change in thickness of the microneedle was observed by bright field microscopy after coating with H5 VLPs and dissolution of coated H5 VLPs into PBS buffer (Fig. 1a). The amount of H5 VLPs coated onto each 5-microneedle array was 2.0±0.15 μg total proteins (approximately 0.2 μg HA) as determined after elution into PBS using a protein assay kit (Quan et al., 2009). Groups of mice (BALB/c, 6-8 weeks old, n=11 per group) were immunized using either i) microneedles without antigen (mock), ii) microneedles coated with 2 μg of H5 VLPs (MN), or iii) 2 μg of H5 VLPs in PBS buffer solution dissolved from coated microneedles given by intramuscular injection (IM). At weeks 3 and 7 after a single dose vaccination, antibody responses in sera were determined by quantitative ELISA using recombinant H5 HA protein as a coating antigen (Fig. 1b). Interestingly, at week 7 after a single immunization, 5-fold higher levels of H5 HA specific antibodies were observed in the microneedle vaccination group compared to the intramuscular control, which is significantly higher than those at week 3 (Fig. 1b).

The pattern of antibody isotypes may provide informative insight into the T helper type 1 or 2 immune responses (Hocart et al., 1989). Therefore, we determined antibody isotypes after a single microneedle or intramuscular vaccination. IgG1 isotype antibody was induced at significantly higher levels than IgG2a antibody after microneedle delivery (p=0.031, Fig. 1c). This IgG1 isotype-dominant pattern following H5 VLP microneedle immunization is similar to that observed with microneedle vaccination using whole inactivated A/Aichi/68 virus (H3N2) (Koutsonanos et al., 2009), but different from results obtained from 0.4 μg low vaccine dose of the A/PR/8/34 (H1N1) virus or VLPs (Quan et al., 2010; Quan et al., 2009), which predominantly induced IgG2a isotype antibody after microneedle vaccination of BALB/c mice. Interestingly, the IgG1 isotype antibody was previously shown to have higher neutralizing and similar HI titers, respectively, compared to the IgG2a isotype in immune sera of BALB/c mice (Hocart et al., 1989). Although it is not clear what factor(s) influences IgG1/IgG2a ratios, there are several potential parameters affecting the antibody isotype pattern, which includes mouse strains and types of vaccines (Hocart et al., 1989), stability and integrity of vaccine antigens (Quan et al., 2010; Quan et al., 2009), routes and doses of vaccines (Bright et al., 2008), and intrinsic immunogenicity properties of vaccines such as H5 VLPs of A/Vietnam/1203/04 observed in this study.

Next, we determined hemagglutination inhibition (HI) antibody responses using 1% horse red blood erythrocyte (Fig. 1d). HI titers over 60 were detected in the microneedle vaccination group, which is 4-fold higher than the titers in the mock control and 1.5 fold higher than the IM group (p=0.03). As high doses of H5 vaccine were reported to be required in humans for moderate immunogenicity, this observation suggests that a MN H5 vaccine could bring significant improvement compared to other non-adjuvanted vaccines tested thus far in clinical trials. The low antibody response reported in a study by Patel et al. is likely related to the small doses of antigen injected intradermally (Patel et al., 2010). In other clinical trials indicating low immunogenicity of H5 vaccines, prime-boost immunizations with 90 μg HA doses of baculovirus-expressed or 30 μg of inactivated subunit vaccines produced in eggs were needed to induce antibody responses that were expected to be protective in 54 to 58% of individuals vaccinated (Nicholson et al., 2001; Treanor et al., 2006; Treanor et al., 2001).

For challenge studies to determine protective efficacy against lethal infection at 20 weeks after a single vaccination, mice (n=5 out of 11 vaccinated mice) were inoculated with 20×LD₅₀ of wild type A/Vietnam/1203/04 (H5N1) by intranasal instillation. All mice in the mock control group showed severe losses in weight and reached the euthanasia endpoint between 6 and 7 days after challenge (Fig. 2a). Mice in the IM group also experienced severe weight losses and showed a significant delay in recovering body weight after lethal challenge infection (Fig. 2b). In contrast, the H5 VLP microneedle group (MN) did not lose body weight and showed 100% protection compared to partial protection (60%) by IM immunization. Therefore, microneedle vaccination in the skin with H5 VLPs induced enhanced protective efficacy compared to intramuscular injection.

To better characterize the protection afforded by microneedle vaccination with H5 VLPs, lungs were harvested at week 20 post vaccination from additional groups of mice (n=6 out of 11 vaccinated mice) euthanized at day 4 post lethal challenge with reassortant A/Vietnam/1203/04 (H5N1) virus (20×LD₅₀), homogenized and titrated by plaque assay (Fig. 3a) as described previously (Quan et al., 2007). The reassortant A/Vietnam/1203/04 virus was generated to contain HA and NA genes from A/Vietnam/1203/04 (H5N1) and internal genes from A/PR/8/34 as described previously (Bright et al., 2008; Hoffmann et al., 2002). The H5 VLP microneedle vaccination group showed lung viral titers that were 20-fold lower than the mock control and 6-fold lower than the intramuscular immunization group. Thus, these results suggest that the improved protection by microneedle vaccination with H5 VLPs was partially due to more effective early control of viral replication in the lungs.

One of the main goals of vaccination is to induce long-lasting immune memory that mediates protection upon exposure to infectious agents. We tested whether a single microneedle vaccination in the skin with H5 VLPs could induce increased levels of memory B cells that are rapidly differentiating into antibody secreting plasma cells upon viral infection (Fig. 3b). Spleen cells were collected from the same groups of mice at day 4 after challenge as described for assay of lung viral titers. Single cell suspensions from spleens were subjected to in vitro culture with inactivated H5N1 A/Vietnam/1203/04 viral antigen coated on the plates. Importantly, with antigenic stimulation by coated viral antigen on the plates, spleen cells collected from the group of mice that were immunized in the skin using microneedles showed significantly higher levels of virus-specific antibodies secreted into culture supernatants compared to those by intramuscular immunization after 6 days but not after 2 days, of culture (Fig. 3b). Therefore, these results suggest that a single microneedle vaccination with H5 VLPs is more effective in generating memory B cells that can differentiate into antibody-secreting cells compared to conventional intramuscular immunization. Regarding the cytokine-secreting splenocyte responses, intramuscular immunization with H5 VLPs showed higher IL-4 and similar IFN-γ levels of secreting cells, respectively, compared to those obtained after microneedle vaccination (data not shown), indicating that IL-4 and IFN-γ secreting cellular responses have less correlation with improved protection by microneedle vaccination in the skin with H5 VLPs.