A Cell Culture (Vero)–Derived H5N1 Whole-Virus Vaccine Induces Cross-Reactive Memory Responses

Highly pathogenic strains of A/Influenza H5N1 virus continue to cause outbreaks in fowl and also result in human disease with a high fatality rate [1]. Preexposure vaccination with an appropriate vaccine is considered to protect against pandemic influenza viruses. However, the H5N1 virus has diversified genetically and has evolved into multiple clades, 5 of which have caused human infection to date [2, 3]. Therefore, production of pandemic vaccine can only be initiated once the exact candidate strain is determined. It would, however, be important to have H5N1 vaccines available that could be used in advance of a pandemic. These vaccines would be based on existing H5N1 strains, which may only be distantly related to the pandemic strain, and could be used as a priming vaccination in a prepandemic situation. A single booster vaccination, with the vaccine strain related to the pandemic strain, delivered in a pandemic situation would then ideally induce a rapid memory response, effectively neutralizing the pandemic strain

The current study was designed to determine whether a whole-virus H5N1 clade 1 strain vaccine could induce memory responses and responses to heterologous strains following administration of a priming regimen. We initiated a dose-finding and formulation study with 6 different formulations of a cell-culture (Vero)–derived whole-virus vaccine based on the clade 1 A/Vietnam/1203/2004 H5N1 strain. The results of our study demonstrated that a 7.5-μg nonadjuvanted formulation was the optimal dosage with respect to development of specific neutralizing antibody at 21 days after receipt of 2 vaccinations [4]. The study was then extended to investigate antibody persistence up to 17 months after priming and, specifically, to investigate the boosting effect with a single dose of a candidate vaccine based on the clade 2.1 A/Indonesia/05/2005 H5N1 virus strain. Such a prime-boost strategy is considered to be of particular value in inducing rapid immunity during a pandemic

MethodsThe vaccines used in these studies were cell culture (Vero)–derived whole-virus vaccines produced from the wild-type strains A/Vietnam/1203/2004 and A/Indonesia/05/2005. An open-label follow-up study was conducted; subjects who had been randomized to receive different doses and formulations (adjuvanted or nonadjuvanted) of Vero cell–derived clade 1 H5N1 influenza vaccine (A/Vietnam/1203/2004) in a 2-dose regimen received a single booster vaccination (12–17 months after priming) with a 7.5-μg nonadjuvanted formulation of a heterologous clade 2.1 vaccine (A/Indonesia/05/2005). The clinical data from the primary clade 1 vaccination study have been published elsewhere [4]

All subjects (n=141) at 1 study site were invited to participate in the follow-up study, and a total of 77 subjects were enrolled. The study protocol was approved by the Ethics Committee of the Medical University Vienna (Austria), and written consent was provided by the subjects

The primary end point was the number of subjects with antibody response to the booster vaccine strain (A/Indonesia/05/2005) at 21 days after boosting, defined as a microneutralization titer ⩾1:20. Secondary end points included number of subjects with seroprotective antibody response, defined as a microneutralization titer ⩾1:20 or a single radial hemolysis area ⩾25 mm [2] and a fold increase of antibody response (compared with baseline response before booster) at 7 and 21 days after boosting. In other studies with this vaccine, hemagglutination inhibition titers were measured and titers were significantly lower than those for microneutralization and single radial hemolysis assays [4]. Therefore, hemagglutination inhibition measurements were not reported in this study

Antibody assessment for the booster vaccination was performed before and 7 and 21 days after vaccination by neutralization assays for the A/Vietnam/1203/2004, A/Indonesia/05/2005, A/turkey/Turkey (clade 2.2), and A/Anhui (clade 2.3) strains. Serum samples were also tested by single radial hemolysis with the Vietnam strain. Neutralizing antibody determinations were performed in the laboratories of Baxter BioScience, with use of a fully validated assay (described elsewhere [4]), and were subject to stringent validity criteria, including a well-defined range of the virus titer and of the titer of the control serum (ie, an H5N1-positive guinea pig serum) [4]. Single radial hemolysis assays were performed as described elsewhere [5]

Safety analysis (systemic and local reactions and frequency and severity of adverse events) was performed for all 77 subjects. Adverse events were graded by the investigator for severity and for relatedness to the study product

Log-transformed microneutralization titer values were analyzed within a mixed model accounting for the effects of vaccine group (6 separate groups determined by the 6 doses and formulations investigated in the primary vaccination study), time, age, sex, and vaccine group-time interaction. The logarithm of the baseline titer value was included in the model as a continuous covariate. Baseline was defined as the prevaccination titer on the day of the first vaccination for the Vietnam and Indonesia strains. Compound symmetry covariance structure was assumed, and the Kenwood-Roger correction for the degrees of freedom was applied. The microneutralization titers on days 7 and 21 after boosting were compared with the prevaccination titer (on the day of the booster) and, for the Vietnam and Indonesia strains, with the day 42 titer (after the first vaccination) within the mixed model

ResultsThe prime-boost vaccinations were well tolerated. All local and systemic reactions after the booster vaccination were generally mild and transient; none were severe. The rate of local reactions up to 7 days after boosting was 17%. All injection-site reactions were mild, and no swelling, induration, or redness was reported. The rate of systemic reactions was 44% (mostly headache and malaise); these were generally mild and transient, and none were severe

Prior to boosting, the neutralizing geometric mean titers (GMTs) were determined against 4 H5N1 strains (ie, Vietnam, Indonesia, turkey/Turkey, and Anhui strains) (Table 1). At this time, antibody levels against all strains were still above the baseline level measured prior to the primary vaccination schedule, with the maximum GMTs measured in the 15-μg nonadjuvanted group against clades 1, 2.1, 2.2, and 2.3. The GMTs measured for the 7.5-μg nonadjuvanted formulation were also higher than those obtained for the adjuvanted formulations, with 1 exception (A/turkey/Turkey). Following boosting with the 7.5-μg nonadjuvanted Indonesia strain vaccine, there was a rapid and substantial increase in GMTs against all 4 viruses. Seven days after boosting, significant increases (P<.001) in neutralizing antibody levels were seen for all formulations used in priming. The optimal dose and formulation for the priming vaccination (ie, 7.5-μg nonadjuvanted vaccine) also resulted in very high GMTs against the Indonesia, Vietnam, turkey/Turkey, and Anhui strains 7 days after boosting with Indonesia (GMT, 102.9, 70.5, 90.7, and 51.5, respectively). A slightly higher titer was obtained following priming with the 30-μg adjuvanted formulation for the Vietnam strain (GMT, 75.5). This enhanced antibody response following priming with the 30-μg adjuvanted formulation was more pronounced 21 days after boosting, with the highest GMTs obtained at this time for the Indonesia, Vietnam, turkey/Turkey, and Anhui strains, respectively. However, a significant dose effect at 7 days (P=.037) and 21 days (P=.016) was only seen for responses to the Vietnam strain after boosting. No significant dose effect was observed for responses to any other strains. There was also no significant difference between responses to adjuvanted or nonadjuvanted preparations for any strain at either 7 or 21 days

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 1

Reverse cumulative distribution of immune response measured by microneutralization titers (booster vaccination regimen)

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 1

Microneutralization Geometric Mean Titers (GMTs) for A/Vietnam/1203/2004 (clade 1), A/Indonesia/05/2005 (clade 2.1), A/turkey/Turkey (clade 2.2), and A/Anhui (clade 2.3) at Days 0, 7, and 21 (Booster Vaccination Regimen)

Figure 1 demonstrates reverse cumulative distribution curves of neutralizing antibody titers at 7 and 21 days after boosting. These curves demonstrate that within 7 days of receiving the booster with the Indonesia strain, 81%–100% of subjects have substantial titers of neutralizing antibody (ie, ⩾1:20) against the booster strain. After 21 days, almost 100% of the total number of subjects tested (73 [97%] of 75) demonstrated this level of neutralizing antibody to the booster strain. Similar results were obtained for the clade 1, clade 2.2, and clade 2.3 strains, with 72 (96%) of 75, 72 (96%) of 75, and 67 (83%) of 75 subjects demonstrating this level of neutralizing antibody to the Vietnam, turkey/Turkey, and Anhui strain, respectively

Single radial hemolysis data are reported for the Vietnam strain in direct comparison to the microneutralization data. These analyses confirmed the effective result of the booster vaccination, with 83.3%–100% of subjects achieving seroprotective levels of antibody to the Vietnam strain, as determined by the single radial hemolysis assay at 21 days after boosting. Fourteen (82.4%) of 17, 13 (86.7%) of 15, 10 (83.3%) of 12, 13 (100%) of 13, 6 (100%) of 6, and 12 (100%) of 12 subjects had seroprotective single radial hemolysis titers after a priming schedule with 3.75-μg adjuvanted, 7.5-μg adjuvanted, 7.5-μg nonadjuvanted, 15-μg adjuvanted, 15-μg nonadjuvanted, and 30-μg adjuvanted formulations, respectively

DiscussionThis study was designed to assess antibody persistence after a 2-dose priming regimen of different doses of an adjuvanted or nonadjuvanted H5N1 clade 1 strain vaccine and, specifically, to investigate the effect of boosting with a single dose of a heterologous clade 2.1 strain pandemic virus vaccine. Modeling of pandemic spread has shown that induction of an immune response to a pandemic strain in a significant proportion of the population within 2 weeks after the beginning of an outbreak is critical to interrupt virus transmission [6]. Conventional strategies requiring 2 vaccinations received 21 days apart at pandemic onset are unlikely to result in rapid protection of the population. Accessibility to vaccine consisting of a new pandemic strain would be further complicated by long manufacture and delivery periods

It has been reported that some H5N1 vaccine formulations may overcome strain mismatch by inducing cross-clade neutralizing antibody responses [4, 7, 8]. However, prepandemic vaccination is unlikely to induce full protection if a pandemic occurs years later, because of decreasing circulating antibody titers. An alternative strategy has been investigated in a trial with an adjuvanted H5N3 vaccine, which demonstrated that after a 2-dose priming, boosting with a third dose of the homologous H5N3 vaccine induced crossreactive antibodies to a range of H5N1 strains [9–11]. These reports suggested that prepandemic priming of the population followed by a single booster vaccination might be a useful strategy

Our study describes the first investigation of an approach using a clade 1 H5N1 vaccine for priming followed by boosting with a clade 2 H5N1 vaccine. This strategy involves use of a whole-virus vaccine formulation, because these are generally more immunogenic than split or subunit vaccines in immunologically naive individuals [12]. Earlier egg-derived whole-virus vaccines were associated with enhanced reactogenicity, particularly fever. However, this cell culture–derived whole-virus vaccine was well tolerated, with only 1 subject developing an elevated temperature (38°C) after vaccination. A phase I/II study reported elsewhere [4] and a number of unpublished studies (Baxter, unpublished data) involving >1000 subjects demonstrate that the vaccine is well tolerated in adult and elderly populations

The immunogenicity of the whole-virus vaccine was best demonstrated with respect to neutralizing antibody responses. In the other reported phase I/II study [4], vaccination resulted in substantial responses, not only to the clade 1 vaccine strain, but also to the widely divergent clade 2.1 strain (A/Indonesia/05/2005). However, 12–17 months after primary vaccination, only low titers for the Vietnam, Indonesia, turkey/Turkey, and Anhui strains could be detected. This concurs with reports that 16 months after 2 vaccinations with an MF-59 oil-in-water adjuvanted H5N3 vaccine, no detectable H5 antibody could be detected either by hemagglutination inhibition or microneutralization, although residual titer could be measured by single radial hemolysis assay [10]

Following boosting with the Indonesia strain, there was a rapid and substantial increase in neutralizing antibody titer to all 4 strains. The antibody titer 21 days after the single booster vaccination was significantly higher than the titers achieved 21 days after the 2 priming vaccinations for the Vietnam and Indonesia strains. There was no significant difference between priming antigen doses and formulation. These findings are in agreement with those reported for H5N3 and H5N1 vaccine prime-boost studies [10, 11, 13–15]

Our findings support the concept of prepandemic priming with an existing H5N1 strain vaccine, which may not be an exact match for an emerging H5N1 pandemic strain. Boosting with a single dose of the optimally or closely matched pandemic strain would result in a rapid (7 day) high titer response to the pandemic strain, which could be superior to the response obtained following 2 vaccinations with a later pandemic strain vaccine without initial priming vaccinations. This strategy would also result in a substantial increase in the availability of vaccine, in addition to the more rapid induction of protective immunity early in a pandemic and significant global public health benefits

• Received February 23, 2009.
• Accepted May 5, 2009.

• © 2009 by the Infectious Diseases Society of America