Cross-reactive antibody response to the pandemic A (H1N1) 2009 influenza virus induced by vaccination with a seasonal trivalent influenza vaccine: a longitudinal study of three influenza seasons in Japan
Headquarters, Chemo-Sero-Therapeutic Research Institute (Kaketsuken), Kumamoto 860-8568
The cross-reactivity of antibody to the swine-origin pandemic influenza A (H1N1) 2009 virus induced by vaccination with a seasonal trivalent influenza vaccine was studied. Paired sera from a cohort of adult volunteers vaccinated with a trivalent seasonal influenza vaccine every year from 2006 to 2008 were collected each year and tested by hemagglutination inhibition (HI) for antibody against the pandemic influenza A (H1N1) 2009 virus. There was little increase in the geometric mean titer overall; a slight increase was detected in the sera obtained in the 2007–2008 season but not in the other two seasons. The proportion of individuals with HI antibody titers ≥ 1:40 did not change significantly from year to year. These results indicate that cross-reactivity of the antibodies induced by a trivalent seasonal vaccine to the pandemic influenza A (H1N1) 2009 virus is marginal.
seroconversion rate (percentage of participants with seroconversion, namely showing at least a four-fold increase in titer and titers of at least 1:40 after vaccination)
seroprotection rate (percentage of participants with titers ≥ 40)
On 11 June 2009, the World Health Organization declared an influenza pandemic on the basis of an outbreak of infection with swine-origin influenza A (H1N1) 2009 viruses. Genetic analysis of this virus revealed that North American swine viruses with genes from swine, humans, and avian influenza viruses had mixed with Eurasian swine influenza viruses. Other details of this virus's characteristics remain unknown, but the above unique characteristic could account for the ability of these viruses to spread in humans (1). Importantly, the HA gene of A(H1N1)pdm09 viruses differs substantially from that of the Russian-type seasonal influenza H1N1 viruses that was circulating at that time (2). In fact, apart from the elderly, fewer than 10% of the population had a HI antibody titer ≥ 1:40 to the pandemic viruses, suggesting that the antigenicity of the HA of A(H1N1)pdm09 viruses differed from that of seasonal H1N1 strains (3–8).
Since it would take several months to manufacture a vaccine against the influenza A(H1N1)pdm09 virus, this study's aim was to determine whether the immune response to the already available seasonal influenza vaccine could produce immunity to the A(H1N1)pdm09 virus equivalent to that of a specific A(H1N1)pdm09 vaccine and thereby prevent infection with this virus.
A study by the CDC of the serum neutralization antibody response to A(H1N1)pdm09 viruses after vaccination with seasonal trivalent influenza vaccines indicated that little or no cross-reactive antibodies were induced in any age group; in adults, they found only a twofold post-vaccination rise in the ratio of the GMT (6). Thus, vaccination with a seasonal trivalent influenza vaccine from 2009 was presumed to have no impact on preventing outbreaks of A(H1N1)pdm09 virus.
Since the preliminary study report from CDC, some further studies regarding cross-reactivity of antibodies have been published. With one exception (9), these studies reported that vaccination with a seasonal trivalent influenza vaccine induces little or no cross-reactivity of antibody to A(H1N1)pdm09 virus (7, 10, 11). Each of the studies was conducted during only one influenza season in a single cohort, or during more than one influenza season but in different cohorts; thus no studies have assessed individuals for cross-reactivity of antibody to A(H1N1)pdm09 virus induced by vaccination with multiple vaccines. Therefore, we conducted a longitudinal study to further assess development of cross-reactive antibody to A(H1N1)pdm09 virus using paired sera from the same cohort vaccinated with seasonal trivalent influenza vaccines in three consecutive influenza seasons (2006–2007, 2007–2008, and 2008–2009).
MATERIALS AND METHODS
The serum samples were collected each year and stored at below –20°C until assay. The donors were personnel of Kaketsuken (Kumamoto, Japan). Written informed consent was obtained from each participant, and the protocol of this study was in accordance with ethical guidelines for epidemiological studies and approved by Kaketsuken's Ethical Review Board.
In all three influenza seasons, the same physician verified whether the participants had already been vaccinated with the current season's vaccine prior to administering it. The participants were vaccinated in the upper arm with 0.5 mL s.c. of the seasonal trivalent influenza vaccine each year. In the 2009–2010 influenza season, some of the participants (eight volunteers) also received the monovalent H1N1 (2009) vaccine. Five participants received two doses of A(H1N1)pdm09 vaccine after the seasonal vaccine, and three participants received one dose before and one dose concomitantly with the seasonal vaccine.
Blood was collected twice, one sample before vaccination and the second 4 weeks later. In 2006–2007 and 2007–2008, blood was also collected 5 months after inoculation. In 2009–2010, blood was collected 3 weeks after each H1N1 (2009) vaccination.
The seasonal trivalent influenza vaccine was an inactivated non-adjuvanted split-virion vaccine prepared using embryonated chicken eggs as a virus growth substrate and contained 15 μg of HA of each antigen per dose. The A/H1N1 strains in the seasonal trivalent influenza vaccine differed each year, containing A/New Caledonia/20/99 in 2006–2007, A/Solomon Islands/3/2006 in 2007–2008 and A/Brisbane/59/2007 in 2008–2009. The A/H3N2 strains in the seasonal trivalent influenza vaccine were A/Hiroshima/52/2005 in 2006–2007, A/Hiroshima/52/2005 in 2007–2008 and A/Uruguay/716/2007 in 2008–2009. The B strains in the seasonal trivalent influenza vaccine were B/Malaysia/2506/2004 in 2006–2007, A/Solomon Islands/3/2006 in 2007–2008 and B/Florida/4/2006 in 2008–2009. In the 2009–2010 season, A(H1N1)pdm09 vaccine contained A/California/7/2009 strain, which was also prepared as a monovalent, inactivated non-adjuvanted split-virion vaccine.
The 31 participants consisted of 18 men and 13 women. As of November 2006, they were aged 22–58 years (9 in their 20s, 8 in their 30s, 11 in their 40s, and 3 in their 50s) with a mean age of 27.5 ± 1.9 years. There was no particular bias of sex or age in the participants. Of the 31 participants, 8 (6 men and 2 women) were vaccinated with A(H1N1)pdm09 vaccine in 2009–2010. As of November 2010, these participants were 27–53 years of age (one in his/her 20s, two in their 30s, one in his/her 40s and four in their 50s) with a mean age of 43.6 ± 6.4 years.
Antibodies to A(H1N1)pdm09 virus (A/California/7/2009 strain) were measured by an HI test (12) in Kaketsuken. Sera were treated with receptor-destroying enzyme (Denka-Seiken, Tokyo, Japan) at 37°C for 18 hrs and heat inactivated at 56°C for 30 mins. The sera were then incubated with a 5% suspension of chicken erythrocytes for 1 hr at 4°C, centrifuged and serially two-fold diluted. The HI titer was determined after a 1 hr incubation with 4 HA of whole inactivated viruses and a 0.5% suspension of chicken erythrocytes. The HI antibody titer to seasonal H1N1, A/New Caledonia/20/99 (2006–2007), A/Solomon Islands/3/2006 (2007–2008) and A/Brisbane/59/2007 (2009–2010) was measured at the laboratory of SRL, Tokyo, Japan.
The proportions of individuals with antibody titers of ≥ 1:40 and the GMT of HI antibodies were calculated. To calculate the GMT, a value of 5 was assigned to an HI antibody titer of < 1:10. McNemar's test was used to compare the percentage of individuals with pre- to post-vaccination changes in HI antibody titer between seasons. Similarly, the ratio of the GMT of HI antibodies and confidence intervals for the GMT were calculated to compare the GMT of HI antibodies. In addition, during the entire period of blood collection pair-wise comparisons were performed using generalized estimating equations for continuous response variables adjusted for multiplicity using the Tukey-Kramer method in order to assess changes in HI antibody titers to H1N1 (2009) viruses (A/California/7/2009 strain). The P level for statistical significance was 0.05, and a confidence coefficient of 95% was used for interval estimation. All statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC, USA).
Table 1 shows the GMTs of the HI antibodies to seasonal H1N1 viruses and A(H1N1)pdm09 virus. In all three influenza seasons we noted a significant increase in the GMT to seasonal H1N1 strains after vaccination; the increase in GMT varied from 1.46 to 1.96. In contrast, the amount of cross-reactivity to A(H1N1)pdm09 virus was low. The GMT increase was 0.91 in 2006–2007, 1.25 in 2007–2008 and 0.98 in 2008–2009. The change was significant only during the 2007–2008 season (1.25-fold; 95%CI = 1.06–1.48; Table 1).
Table 1. Geometric mean titer of the cross-reactive HI antibody to A(H1N1)pdm09 virus in participants vaccinated with seasonal influenza vaccines†
Influenza season and strain of influenza virus used in measurements
No. of participants
Geometric mean antibody titer‡
Ratio of GMT before and after vaccination§
†, All 31 participants were vaccinated with a seasonal trivalent influenza vaccine (2006–2007, A/New Caledonia/20/99; 2007–2008, A/Solomon Islands/3/2006; 2008–2009, A/Brisbane/59/2007).
‡, A value of 5 was substituted for a hemagglutination inhibition antibody titer of < 1:10 in the GMT calculation. Blood was collected before and 4 weeks after vaccination.
§, A significant rise in titer of antibodies to the pandemic A(H1N1)pdm09 after vaccination was noted only during the 2007–2008 flu season.
The proportion of individuals with an HI antibody titer of ≥ 1:40 to A(H1N1)pdm09 virus remained unchanged (Table 2) whereas the proportion to the seasonal H1N1 strains increased significantly in all three influenza seasons (2006–2007 [P= 0.025], 2007–2008 [P= 0.005] and 2008–2009 [P= 0.025]).
Table 2. The percentage of participants with titer of cross-reactive HI antibody ≥ 1:40 to A(H1N1)pdm09 virus before and after vaccination with seasonal influenza vaccines†
Influenza season and strain of influenza virus used in measurements
No. of participants
HI antibody titer ≥ 1:40
Before vaccination (%)
After vaccination (%)
†, All 31 participants were vaccinated with a seasonal trivalent influenza vaccine (2006–2007, A/New Caledonia/20/99; 2007–2008, A/Solomon Islands/3/2006; 2008–2009, A/Brisbane/59/2007). Blood was collected before and 4 weeks after vaccination.
‡, McNemar's test was used for before-and-after comparison of the percentage of individuals with different hemagglutination inhibition antibody titers.
Analysis of changes in the HI antibody titer to A(H1N1)pdm09 virus over the 3 years of the study revealed significant differences between the titer before vaccination during the 2007–2008 season and the titer before vaccination during the 2008–2009 season (P= 0.0106) and between the titer before vaccination during the 2007–2008 season and the titer after vaccination during the 2008–2009 season (P= 0.0260; Fig. 1).
Of this study's 31 participants, eight were vaccinated with A(H1N1)pdm09 vaccine (A/California/7/2009) in 2009–2010. No individual had an HI antibody titer of ≥ 1:40 before vaccination. Seroconversion occurred in five (62.5%) and six (75.0%) individuals after the first and second dose of the vaccination, respectively. In addition, the GMT (95% CI) for these eight participants was 7.11 (5.01–11.87) before vaccination, 47.57 (20.08–112.67) after the first dose and 67.27 (32.01–141.39) after the second dose. This confirmed that the individuals in this cohort are strong responders to A(H1N1)pdm09 vaccine.
This study investigated whether cross-reactive antibodies to A(H1N1)pdm09 virus were induced by vaccination with a seasonal trivalent influenza vaccine over a 3-year period from 2006 to 2009. The GMT against A(H1N1)pdm09 virus was slightly increased only in the 2007–2008 season. However, in the 2007–2008 season, the proportion of individuals with an HI antibody titer of ≥ 1:40 was similar to that in the 2006–2007 and 2008–2009 influenza seasons. Therefore, these results indicate that vaccination with a seasonal trivalent influenza vaccine does not elicit antibodies cross-reactive to the A(H1N1)pdm09 virus at the clinically significant level defined by the criteria of the European Medicines Agency (13).
These results are consistent with previous findings. The CDC reported that little or no cross-reactive neutralizing antibodies were present in paired sera collected before the 2009 pandemic from adults vaccinated with a seasonal influenza vaccine (6). Lee et al. also showed that a trivalent seasonal influenza vaccine used in the southern hemisphere induced very weak cross-reactive antibodies to A(H1N1)pdm09 (10). These studies suggest that seasonal vaccines are ineffective against A(H1N1)pdm09 virus because they stimulate little or no cross-reactive antibody.
On the other hand, Xie et al. reported an increase in cross-reactive antibodies to A(H1N1)pdm09 virus in 120 adults (aged 20 years and over) and 59 elderly individuals after vaccination with a seasonal trivalent non-adjuvanted influenza vaccine (9). However, they noted an increase in the HI antibody titer to A(H1N1)pdm09 virus (concomitant with an increase in seroprotection and seroconversion rates) when only two of the five strains of A(H1N1)pdm09 viruses, A/California/7/2009 and A/South Carolina/18/2009, were used in the HI test. They conducted their study during the 2009–2010 pandemic season, whereas the other studies, including the present one, were conducted using sera collected before the 2009 pandemic.
An advantage of the current study design is that it is longitudinal rather than cross-sectional. This prevents the potential confounding that is inherent in a cross-sectional design, by stopping immune response factors from affecting the results. In addition, it assessed amounts of cross-reactive antibodies to A(H1N1)pdm09 virus. Although we strongly believe that very little cross-reactive antibody to A(H1N1)pdm09 virus is induced by immunization with a seasonal vaccine, one limitation of this study is that the number of participants (n= 31) was small and they were all from a single institution.
Several researchers have investigated the direct effectiveness of a seasonal influenza vaccine in protecting against natural infection and morbidity with A(H1N1)pdm09 virus and reported conflicting results. Johns et al. conducted a study of 1205 pandemic H1N1 confirmed cases from April–October 2009 using the matched case-control method (14). After controlling for sex, age, and history of seasonal influenza vaccination during the 2008–2009 influenza season, their analysis revealed that seasonal trivalent inactivated influenza vaccine prevented infection by A(H1N1)pdm09 virus with an effectiveness of 44% (95% CI = 32–54). In addition, Orellano et al. assessed and confirmed the effectiveness of seasonal trivalent inactivated vaccine in preventing hospitalizations for A(H1N1)pdm09 influenza virus infection in 6866 cases in Argentina (15).
In contrast to these results, Skowronski et al. reported that seasonal trivalent influenza vaccines exacerbates A(H1N1)pdm09 virus infection (16). They suggested one possible explanation for their unexpected finding, namely that the results were an artifact of selection bias or confounding. However, Rosella et al. stated that it is unlikely that an unidentified confounder could have explained these findings (17). Moreover, Kelly et al. could find no evidence for harm or benefit from seasonal influenza vaccination in 2009–2010 (18). However, recently conducted meta-analyses have discovered that a seasonal trivalent vaccine was generally beneficial before the A(H1N1)pdm09 vaccine was available, though cross-protection was less than the direct effect of strain-specific vaccination against A(H1N1)pdm09 (19).
If vaccination with a seasonal vaccine is clinically effective, cross-reactive T cell responses could provide an explanation. Iorio et al. reported that cellular (activation of interferon gamma-producing CD8+ T cells) but not humoral immune responses (HI antibody titer) against H1N1 (2009) virus were induced after vaccination with the seasonal trivalent subunit adjuvanted influenza 2007–2008 vaccine in 17 healthy adults. They considered that this cellular immune cross-reactivity might, at least in part, explain the low pathogenicity of A(H1N1)pdm09 virus (20). De Groot et al. also suggested that T helper cells induced by seasonal influenza infection or vaccination might explain the enhanced antibody response to the influenza A (H1N1) 2009 vaccine (21). Moreover, Weinfurter et al. reported that, in non-human primates, cross-reactive T cells generated by priming with a seasonal H1N1 virus are involved in rapid clearance of 2009 pandemic H1N1 influenza virus after challenge (22).
Recently, our group (23) and other groups (24) have reported that prior vaccination with a seasonal trivalent influenza virus inhibits the antibody response to pandemic virus elicited by the pandemic influenza vaccine. Although the explanation for this inhibitory effect remains unclear, one possible explanation is “original antigenic sin,” which is the notion that an antibody response to a new variant is inhibited when individuals immunologically primed with other strains are re-infected with a related but different new variant. Such inhibitory mechanisms might limit the immune response against A(H1N1)pdm09 after natural infection and thus be involved in exacerbation of natural infection.
Although induction of cross-reactive humoral immune responses and protective effects of a prior vaccination with a seasonal vaccine against A(H1N1)pdm09 virus infection remain controversial, prior vaccination with a seasonal vaccine might still have some positive or negative impact on the outcome of any subsequent A(H1N1)pdm09 virus infection. Therefore, in a pandemic situation prior vaccination with a seasonal vaccine should be considered carefully.
The authors wish to thank Kaketsuken staff members Kayo Ibaragi, Shigemi Yamamoto, Keiko Tazoe, Mariko Miyata, Emiko Sato, Akiko Saeki, Takayuki Masaki, Seiichi Harada, and Nobuo Mon’nai for their great contribution to the preparation of vaccines used in this study and for the blood sample collection required by this study.
Shingo Uno, Kazuhiko Kimachi, Keiichiro Miyazaki, Ayano Oohama, Junko Kei, Tomohiro Nishimura, Koichi Odoh and Yoichiro Kino are employees of Kaketsuken. At the time of the clinical study of seasonal trivalent influenza vaccines, Fujio Matsuo was an employee of Kaketsuken. This study was planned and conducted by Kaketsuken and its results were analyzed by Kaketsuken. Data analysis for this study was performed by Statcom. Kaketsuken provided all the funding for this study.