Clinical phase II and III studies of an AS03‐adjuvanted H5N1 influenza vaccine produced in an EB66® cell culture platform

Abstract Background We have developed an AS03‐adjuvanted H5N1 influenza vaccine produced in an EB66® cell culture platform (KD‐295). Objectives In accordance with Japanese guidelines for development of pandemic prototype vaccines, the phase II study was conducted in a double‐blind, randomized, parallel‐group comparison study and the phase III study was conducted in an open‐label, non‐randomized, uncontrolled study. Methods Healthy adult volunteers aged 20 ‐ 64 years enrolled in the phase II and III studies (N = 248 and N = 369) received KD‐295 intramuscularly twice with a 21‐day interval. After administration, immune response and adverse events were evaluated. In the phase II study, four different vaccine formulations were compared: MA (3.75 μg hemagglutinin [HA] antigen + AS03 adjuvant system), MB (3.75 μg HA + 1/2AS03), HA (7.5 μg HA + AS03), and HB (7.5 μg HA + 1/2AS03). In the phase III study, the MA formulation was further evaluated. Results In the phase II study, all four vaccine formulations were well‐tolerated and no SAE related to vaccination were observed. The MA formulation was slightly more immunogenic and less reactogenic among the vaccine formulations. Therefore, the MA formulation was selected for the phase III study, and it was well‐tolerated and no serious adverse drug reactions were observed. The vaccine fulfilled the three immunogenicity criteria described in the Japanese guidelines. Conclusions These data indicate that the MA formulation of KD‐295 was well‐tolerated and highly immunogenic and it can be considered a useful pandemic and pre‐pandemic influenza vaccine.


| INTRODUC TI ON
The most recent influenza pandemic of 2009-2010 remains fresh in our minds. Contrary to global expectations, the causative agent of the pandemic was an H1N1 virus. In the 2009 pandemic, various vaccines were used, including non-adjuvanted and adjuvanted subvirion and whole virion vaccines. 1 Among adults, the results of the vaccine use confirmed that even non-adjuvanted vaccines were highly immunogenic. This is because there was cross-reactivity in T helper epitopes between the H1N1 pandemic 2009 virus and previous seasonal H1N1 viruses. 2 In Japan, the local vaccine manufacturers produced monovalent non-adjuvanted split vaccine. At the same time, the Japanese government imported adjuvanted vaccines as a precaution in case of vaccine shortages, but many of these imported vaccines were left unused. However, pandemic threats, such as H5N1, have not disappeared and nobody knows what virus subtype will cause the next pandemic. At this moment, among the viruses with pandemic potential, viruses of avian origin, including the H7N9 subtypes, are a concern because of sporadic human infection. 3 As same with H5N1 subtypes, and unlike the H1N1 pandemic 2009 virus, immunogenicity of those viruses is very low in humans, which may be related to predicted poor T-cell immunogenicity. 4 Another important condition of a pandemic vaccine is timely manufacturing. In the case of the 2009 pandemic, the causative virus was first isolated in April 2009 and a candidate vaccine virus was generated in May. Usually, seasonal influenza vaccines are produced from spring to summer in Japan; therefore, the transition of production from seasonal vaccine to pandemic vaccine was relatively smooth in 2009. If a pandemic occurs in a period outside of seasonal vaccine production in the egg vaccine platform, more time will be needed to start the vaccine manufacturing because of egg supply. Furthermore, because of the damage of chickens by highly pathogenic avian influenza, there is a risk that egg supply will be stopped. To address these issues, we have been developing an AS03adjuvanted vaccine using H5N1 influenza virus antigen derived from a duck cell line (EB66 ® ). In the previous phase I study, we confirmed that the vaccine was well-tolerated and elicited a broadly cross-reactive antibody response. 5 In this paper, we report further evaluation of AS03-adjuvanted H5N1 influenza vaccine formulations produced in an EB66 ® cell culture platform, KD-295, in phase II and III studies to assess its immunogenicity and safety. In addition, phase II study data were registered and released in JapicCTI-121788, and phase III study data were registered and released in JapicCTI-121936.

| Study designs and subjects
The phase II study was conducted in adults between the ages of 20 and 64 years in a randomized, double-blinded (all involved were blinded), comparative fashion from 2 April to 6 November 2012 to further assess immunogenicity and safety of the vaccine, and to determine the appropriate dosage to be evaluated in the phase III study. After selection of one formulation, the phase III study was performed from August 23, 2012 to March 10, 2013 in an unblinded, uncontrolled study enrolling adults between the age of 20 and 64 years.
In both studies, the selection criteria were healthy adults aged 20-64 who agreed with written informed consent. Exclusion criteria included no history of H5N1 infection or vaccination. These studies were conducted in Tokyo, Osaka, and Kagoshima in Japan.
Prior to clinical studies, related documents such as the clinical trial protocol and informed consent form were reviewed by the IRB within each hospital. The studies were conducted in accordance with the Helsinki Declaration, GCP, and other relevant regulations.
Written informed consent was obtained from participants prior to enrollment.
Regarding the allocation method to each group, first the sponsor distributed the investigational drugs allocated randomly to the study sites. The principal investigator or subinvestigator entered the information about subjects as of obtaining written informed consent and about the study sites into the Electronic Data Capture system (Medidata RaveTM). The investigational drug allocation system (Medidata BalanceTM) featured in the Electronic Data Capture system allocated the each subject to MA group, MB group, HA group, and HB group on 1:1:1:1 ratio by using a minimization method. Age, gender, stratification, study, and study sites were used as the adjustment factors.
In the phase III study, the MA formulation (3.75 μg HA + AS03) was evaluated based on the results of the phase I/II study. In both studies, the investigational vaccine KD-295 was administered intramuscularly at a dose volume of 0.5 mL given twice at an interval of 21 ± 7 days.

| Immunological evaluation
The immunogenicity evaluation protocols of the phase II and III studies were the same. Blood samples were taken before the first

| Safety evaluation
The safety evaluation protocols of the phase II and III studies were the same.

| Study population
Demographic data in the FAS of the phase II and III studies are shown in Table 1. The mean ± SD for age in the FAS was 39.1 ± 11.1 years and 37.8 ± 11.2 years overall in the phase II and III studie,s respectively. No population imbalance was observed among the groups.

| Immunogenicity
HI antibody against the vaccine strain was measured using horse and chicken erythrocytes in the phase II study. In the measurement using horse erythrocytes, the HI antibody response to the vaccine strain after the second vaccination (Day 42) in the FAS fulfilled all three criteria of immunogenicity described in the guidelines in all groups.

| Safety
Adverse events and adverse drug reactions occurring during the study period (Day 0-Day 201) are summarized in Table 4. Two SAEs (thyroid cancer and acute abdomen) and one pIMD (pasuda disease) occurred in the phase 3 trial, but in all cases a causal relationship with vaccination was denied. Through both studies, no serious adverse drug reactions (death, SAE, IAE, and pIMD related to vaccination) were reported. Among unsolicited adverse drug reactions, injectionsite pruritus showed the highest incidence in both studies. The majority of cases of injection site pruritus were grade 1. grade 3 unsolicited adverse drug reactions occurred in one case (dehydration) in the phase II study the MA group and in two cases (positional vertigo and malaise) in the phase III study. The incidence rates were low.    adverse drug reactions, that with the highest incidence was fatigue, which was expressed in 40%-60% of patients in each group. The next highest were headache and myalgia (30%-40%). In addition, the incidence rate of grade 3-solicited systemic adverse events and solicited systemic adverse drug reactions was low.

| D ISCUSS I ON
The phase II and phase III clinical studies revealed that MA formulation of KD-295 (3.75 μg HA + AS03) was well-tolerated and highly immunogenic and that no serious adverse drug reactions were observed. Therefore, the MA formulation can be considered as a useful pandemic and pre-pandemic influenza vaccine.
Although most human cases of avian influenza to date have been associated with direct contact with infected birds, as causative agents for the next pandemic, viruses including H5, H7, and H9 subtypes are still of concern as viruses with pandemic potential. 7 Among them, the H5 subtype was the first target for vaccine development, and since then many types of vaccine have been developed. The first developed H5N1 vaccine was a non-adjuvanted split vaccine, and it was reported that two doses of 90 μg HA of the vaccine-induced neutralization antibody titers reaching 1:40 or greater in 54 percent of study subjects. 8 Non-adjuvanted and adjuvanted whole virion vaccines were then developed, and their immunogenicity in humans with 7.5-15 μg HA antigen dose was much better than that of the split vaccine. 9 We also conducted a clinical study with an egg-derived, alum-adjuvanted whole virion H5N1 vaccine. 6 However, although the vaccine was immunogenic, it could not meet one of the three criteria of CHMP guidelines.
Therefore, we decided to develop a more immunogenic vaccine with a platform other than chicken eggs to have flexibility in the vaccine production.
As a result of immunological evaluation in the phase II study, all vaccine groups fulfilled the three required criteria described in the Japanese guidelines based on the HI antibody titers measured using horse erythrocytes after administration of two doses of the vaccine.
When vaccine strain derived from avian influenza virus such as H5N1 is used as an antigen, HI antibody titers' measurement using horse erythrocytes has more sensitive in detection of the antigen than using chicken erythrocytes. 10 This is why HI antibody titers measured using not chicken, but horse erythrocytes were used for primary evaluation.
Although not statistically significant, GMFR was higher in the groups given the vaccine with standard AS03 (HA, MA) dose.
Although all the vaccine formulations were well-tolerated, the MA formulations were less incidence of solicited systemic adverse events (pyrexia, headache, and chills) than the HA formulations.
Therefore, the MA formulation was selected to be evaluated in the phase III study.
In the phase III study, the MA formulation containing a standard dose of AS03 and 3.75 μg HA antigen was further confirmed to be  Feldstein et al 13 compared human immunogenity data of several H5N1 vaccines and concluded that adjuvanted H5N1 vaccines induced high theoretical efficacy and that AS03-adjuvanted vaccine was more immunogenic than MF59-adjuvanted vaccine. It has also been confirmed that the AS03 could increase immunogenicity of H7N1, H7N9, and H9N2 antigens. 14-17 Based on these data, this study has limitations because it does not directly compare KD-295 with other vaccines, but KD-295 appears to be more effective than other licensed H5N1 vaccines.
Marichal et al proposed that alum-adjuvant induces neutrophil migration and cell death, and subsequently DNA released from the host cell activates innate immunity as DAMPs. 18 AS03-adjuvant reportedly activates not only innate immunity, but also adaptive immunity comprehensively, induces production of various cytokines, and contributes to enhancing the antigen-specific antibody production of B cells. 19,20 It also reported that α-tocopherol plays an important role in these immune responses. In fact, omission of α-tocopherol from AS03 modified the innate immune response and lead to lower antibody responses. 19 Therefore, at this time, vaccines with emulsion-type adjuvant, including AS03 with α-tocopherol, would be the promising choice for both pre-pandemic and pandemic avian influenza vaccines.
When compared with the safety profile of our alum-adjuvanted H5N1 whole virion vaccine, 6 the greatest difference between the TA B L E 6 (Continued) two vaccines is local (injection site) pain. With the alum-adjuvanted vaccine, 10%-45% of participants reported local pain after administration of the vaccine, whereas 70%-80% of subjects reported it with the AS03-adjuvanted vaccine. Local pain is a common adverse drug reaction with emulsion-type adjuvanted of vaccines. 9 However, it is self-limiting and leads to no further complication.
The incidence rates of other local and systemic adverse events observed in the current studies were generally higher than those of alum-adjuvanted vaccines; however, most events were grades 1 and 2, and the highest rate of grade 3 was pyrexia yet at only 1.6% which led to no further complication. In general, as was the case in the phase I study, the vaccine was well-tolerated in both the phase II and III studies. can be considered as a useful pandemic and pre-pandemic influenza vaccine. In addition, KD-295 is currently approved in Japan.