Serum strain-specific or cross-reactive neuraminidase inhibiting antibodies against pandemic А/California/07/2009(H1N1) influenza in healthy volunteers

Background Pre-existing antibodies to influenza virus neuraminidase may provide protection against infection influenza viruses containing novel hemagglutinin (HA). The aim of our study was to evaluate serum neuraminidase-inhibiting (NI) antibodies against А/California/07/2009(H1N1) [H1N1/2009pdm] and А/New Caledonia/20/1999(H1N1) [H1N1/1999] influenza viruses in relation with the age of participants and hemagglutination-inhibition (HI) antibody levels. Anti-H1N1/2009pdm neuraminidase and anti-H1N1/1999 neuraminidase antibody levels were measured in total 219 serum samples from Russian healthy peoples of various ages examined before and a year after pandemic strain appearance. We adjusted peroxidase-linked lectin micro-procedure to measure NI antibody titers using the reassortant A/H7N1 influenza viruses based on A/equine/Prague/1/56(H7N7). Also, HI antibody titers were estimated against H1N1/2009pdm, H1N1/1999 and a panel of seasonal A/H1N1 influenza viruses. Results In sera samples collected during the fall of 2010, mean titers of specific HI and NI antibodies to H1N1/2009pdm were 2–2.1 times lower than antibody levels against H1N1/1999. Of the 163 individuals examined, 58 (35.6%) had NI anti-H1N1/2009pdm antibody titers > 1:20, compared to 93 (57.1%) who had NI anti-H1N1/1999 antibody titers > 1:20. There were low correlations between HI and NI antibody levels against either H1N1/1999 or H1N1/2009pdm in the same serum samples. The 24 adults born between 1957 and 1977 expressed very low levels of NI antibodies to A/H1N1 influenza viruses. Persons with low HI anti-H1N1/2009pdm titers but positive to seasonal A/H1N1 demonstrated significantly higher NI anti-A/H1N1 antibody titers than unexposed subjects. In 2005 cross-reactive NI anti-H1N1/2009pdm antibody titers > 1:20 were detected among 7.1% of young people. Conclusions Our study confirmed that contact with seasonal influenza viruses may have contributed to generating the cross-reacting anti-H1N1/2009pdm NI antibodies which were detected in the sera of 18-20 years old people examined before the pandemic virus active circulation. The lowest levels of antibodies to the neuraminidase of N1 subtype were in the group of participants born during the circulation of influenza A/H2N2 or A/H3N2 viruses. The low correlation between HI and NI antibody titers suggests that NI antibody detection can be used as an additional test to evaluate the immune response after influenza infections or immunizations.


Background
In April 2009, the World Health Organization registered the first 21st century pandemic caused by a type А/ H1N1 influenza virus (the genus Influenzavirus, the family Orthomyxoviridae) not previously isolated from animals or humans. Pandemic influenza, in contrast to seasonal influenza, affected young people more frequently than elderly [1]. In the USA, 79% of laboratory-confirmed cases of pandemic H1N1/2009pdm infections were from persons younger than 30 years, and 2% from age group older than 60 years [2]. A number of publications have analyzed pre-existing neutralizing antibodies and T-cell immunity against H1N1/2009pdm [2,3], although little is known about pre-existing cross-reactive anti-neuraminidase (NA), or neuraminidase-inhibiting (NI) antibodies to pandemic A/H1N1 in humans. Several animal studies, including a plasmid DNA vaccine model, suggest that NI antibodies could provide partial protection from lung infection and even from lethal challenge with highly pathogenic А/H5N1 influenza viruses [4][5][6][7].
In humans, NI antibodies play a role in decreasing the severity of natural infection caused by influenza A shift or drift variants [8]. Previously it was shown that immunization with seasonal influenza strains induced cross-reactive serum antibody to the NA of antigenically distinct H1N1/2009pdm, mostly in elderly individuals [9,10].
The aim of our current study was to examine the presence of homologous and cross-reactive NI antibodies against H1N1/2009pdm in serum samples collected in the fall of 2010 from healthy Russian people for more detailed estimation of the overall and age-specific influenza immunity.

Results and discussion
We estimated HI and NI antibodies against both H1N1/ 2009pdm and seasonal H1N1/1999 influenza viruses in 163 sera samples obtained in the fall of 2010 ( Figure 1 The age distribution of A/H1N1-specific antibodies was analyzed in several age groups of participants. Persons in group 1 were born prior to 1957; group 2: 1957-1976; group 3: 1977-1999; and group 4, after 2000 (Table 1). Only group 1 expressed NI antibodies against H1N1/2009pdm in significantly higher mean titers than HI antibodies ( Table 1). The highest mean titers of NI antibodies against both H1N1/2009pdm and H1N1/1999 were detected in group 3. Figure 2 presents the seroprotection rates (proportions of subjects with antibody titers ≥ 1:40) [11] among groups of participants. Again participants in group 3 (1977-1999 years of birth) had the levels of herd immunity to both HA and NA of seasonal H1N1/1999 even higher than children born after 2000 (group 4), when А/New Caledonia/20/99(H1N1) became dominant (Fisher's test: p = 0.0001; see Figure 2). This phenomenon could be attributed to higher social   Figure 2).
To determine a possible origin of NI antibodies against H1N1/2009pdm, we evaluated antibody titers ≥1:40 supposed to be protective in three groups separated on the basis of pre-existing HI antibody levels against H1N1/ 2009pdm and seasonal A/H1N1 viruses (n = 159, 4 sera were excluded from analysis because there was no HI data with all tested A/H1N1 antigens). Group I were persons with low HI titers (≤ 1:10) against H1N1/ 2009pdm and no antibodies to other A/H1N1 viruses (≤ 1:20); group II were persons with low HI titers (≤ 1:10) against H1N1/2009pdm and positive (HI titers ≥ 1:40) for any of the other tested A/H1N1 influenza viruses (see Methods); and group III included all persons with HI titers against H1N1/2009pdm ≥ 1:20 (Table 2). Group I subjects, who were negative to all A/H1N1 viruses used in the HI test, had a higher median age (45 years) than group II (17 years) and group III (15 years), and a wider variation of age (Kolmogorov-Smirnov's test: p < 0.005; see Table 2).
To reveal true cross-reactive NI antibodies against H1N1/2009pdm, we also investigated levels of antibodies against H1N1/2009pdm in serum samples collected from 18-20-year-old people in the fall of 2005, long before H1N1/2009pdm arose ( Figure 3). None of these subjects had detectable anti-H1N1/2009pdm HI antibodies. Only 7.1% (4 of 56) had NI antibody titers > 1:20 against H1N1/2009pdm compared to 41.1% (23 Mean titers of HI antibodies against H1N1/2009pdm are lower than NI antibodies (p = 0.001). 2 Mean titers of HI antibodies against H1N1/1999 are lower than NI antibodies (p = 0.006). 3 Mean titers of antibodies against NA of pandemic influenza virus H1N1/2009pdm in are lower in group 2 than in group 3 (p = 0.002). 4 Mean titers of HI antibodies against H1N1/2009pdm are higher in group 3 than in group 1 (p = 0.0002). 5 Mean titers of HI antibodies against H1N1/1999 are higher in group 3 than in groups 1, 2, and 4 (p < 0.0001). 6 Mean titers of NI antibodies H1N1/1999 are higher in group 3 than in groups 1, 2, and 4 (p < 0.001).   [12] previously reported that in 2010 the threshold level of herd immunity against H1N1/2009pdm was not achieved in Hamburg, Germany.
In our study, the participants born between 1977 and 1999 were the most likely to be infected to pandemic H1N1/2009pdm compared to those born before 1957. This finding confirmed the extensive published data about the high frequency of H1N1/2009pdm infection in children, teenagers, and adults younger than 35 years during the 2009 pandemic, varying between 20-60% in different countries [13]. High anti-H1N1/2009pdm NA antibody levels, rather than anti-HA antibody levels, in participants older than 53 years can be attributed to contact with earlier circulating A/H1N1 viruses. Examin The low correlation between HI and NI antibody titers we found can be attributed to the autonomy of the serum immune response to both surface glycoproteins of the influenza virus, widely reviewed in scientific literature [14][15][16]. Other reasons for such divergence may be different times of persistence of NI and HI antibodies, according to individual anamnesis of the surveyed influenza infections [17,18], and the ability of anti-NA antibodies to interact with a wider spectrum of viruses than anti-HI antibodies. Thus, some persons with no antibodies against HA of influenza pandemic virus in the fall of 2010 still may have been exposed to H1N1/2009pdm. Nevertheless, high levels of antibodies against H1N1/ 2009pdm NA among participants not expressing anti-H1N1/2009pdm HA of pandemic virus, but positive to epidemic H1N1 viruses HA, compared with completely negative individuals examined during the same epidemic period, may, to some extent, confirm the relationship between exposure to epidemic A/H1N1 viruses and development of antibodies cross-reactive with H1N1/ 2009pdm NA. However, the most reliable data about truly cross-reactive antibodies to H1N1/2009pdm may be obtained using sera collected long before the appearance of pandemic virus. During the 2005 epidemic season, the 7.1% of the 56 examined in our study 18-20-year-old volunteers had anti-pandemic virus NI antibodies in titers > 1:20.
The most convincing data concerning the protective action of pre-existing anti-NA antibodies were obtained in the 1970s using a large cohort of volunteers with no or low levels of anti-HA antibodies against pandemic influenza. One such study, by Monto et al. [8], showed that, prior to the Hong Kong A/H3N2 influenza pandemic, only 12% of the population had high NI antibody titers to NA of N2 subtype, while 72% had none. The relatively low level of herd immunity to A/H2N2 NA in the population correlated with the wide distribution of the new pandemic subtype A/H3N2 virus in 1968. However, this herd immunity was nonetheless sufficient to moderate the severity of the pandemic: the frequency of confirmed A/H3N2 influenza infection, determined by elevated levels of HI antibodies, was reversely proportional to pre-existing levels of anti-N2 antibodies. The possible effect of HA was eliminated because the sera were obtained before the virus with the novel H3 HA had appeared. Volunteers between 20 and 45 years of age who had anti-N2 antibodies in titers ≥ 1:16 were 2.0-2.6-fold less likely to develop respiratory infection symptoms than persons with low NI antibody titers. The authors suggested that neuraminidase antibody can protect not so much against infection as against symptoms of influenza thus permitting the individual to 'up-date' his antibody status from time to time without suffering clinical influenza [8]. The levels of protective NI antibodies still unclear although protective HI antibody titers defined as ≥ 1:40 [11]. The low correlation between HI and NI antibody titers we found suggests that NI antibody detection can be used as an additional test to evaluate the immune response after influenza infections or immunizations.

Ethics statement
In our retrospective study we used only serum samples left as a part of routine tests. These serum samples were provided by the Diagnostics Laboratories (Saint Petersburg, Russian Federation).
The peroxidase-linked lectin micro-procedure previously reported by Lambré et al. [22] was adjusted to assay NI antibodies using diagnostic A/H7N1 reassortant viruses. Serum samples were heated at 56°C for 30 min, serially diluted in PBS-BSA with the pH = 6.9 (typically, seven 2-fold dilutions starting at 1:10). Sixty μL of serum dilutions were incubated with an equal volume of pre-diluted virus containing 128 HA units for 30 min at 37°C. After incubation, 100 μL of the mixture was added to the 96-well plates (Sarstedt AG & Co, Nümbrecht, Germany) coated with 150 μL of 50 μg/mL fetuin (Sigma-Aldrich, St. Louis, MO, USA). After 2 hours incubation at 37°C the plates were washed, and NA activity was measured by incubating with peroxidaselabeled peanut lectin (2.5 μg/mL; Sigma-Aldrich, St. Louis, MO, USA) for 1 h at room temperature, washing, and adding 100 μL of peroxidase substrate. The reaction was stopped after 5 min with 100 μL of 1 N sulfuric acid. OD values were measured at 450 nm using the universal microplate reader (ELx800, Bio-Tek Instruments Inc, Winooski, VT, USA). NI titers were expressed as the inverse of the dilution that gave 50% OD 450 of positive control (virus without serum).