Increasing Hemagglutination Inhibition Antibodies against Two Lineages of Type B In�uenza Virus in 2017-2018 Winter Season in Beijing, China

In�uenza virus circulates every year with lower activity than that of in�uenza A virus in China. During 2017 to 2018 winter season, a sharp surge of in�uenza activity dominated by type B/Yamagata lineage virus caused unprecedented medical burden in Beijing. The research aimed to understand the underlying mechanism for this circulation and be prepared for epidemics in the future. Sera samples collected from the patients in 2016-2017 and 2017-2018 �u seasons were tested for pro�ling hemagglutinin inhibition (HI) antibodies against both prevailing Victoria and Yamagata lineages of type B in�uenza viruses. It showed that the seroprevalence against both lineages of the virus in 2017-2018 winter was higher than that in 2016-2017, while no difference of the seroprevalence was observed between the two viruses. Meanwhile, signi�cant elevated geometric mean titer (GMT) against both lineages of in�uenza B viruses was found in the specimens collected during 2017-2018 �u season than that from 2016-2017, suggesting the viruses might undergo antigenic changes. These results also suggested that lower GMT against both type B variants in 2016-2017 might serve as an immunological niche for the dominating of B/Yamagata virus in China during 2017-2018 winter season. Our �ndings have implication that there was a signi�cantly elevation of HI antibodies to in�uenza viruses B in 2017-2018, compared in 2016-2017. On the other hand, the low level of HI antibodies to both B/Y and B/V in 2016-2017 could contribute to the severe B/Y epidemic in 2017-2018 to some extent.


Introduction
Since October 2017, information from the World Health Organization (WHO) has highlighted the increasing level of in uenza activity compared with previous u seasons ( WHO. 2018).It dramatically reached the peak and caused a huge medical burden in Beijing and such situation had been lasting till April, 2018 (Kochanek et al. 2018).Sharply increasing number of patients with in uenza-like-illness (ILI) suggested that the circulating virus might have evaded the herd immunity or mutated to achieve a higher transmissibility (Bedford et al. 2015).This predication soon was con rmed by the evidence from China CDC that Yamagata lineage of type B in uenza virus (Flu B) was the predominant strain resulting the epidemic, although few cases were reported to be type A in uenza virus (Flu A) infection including pdmH1N1/2009 and H3N2 (CDC.2018).
Flu B virus rstly identi ed in 1940 was associated with considerable hospital admissions and deaths worldwide annually (FRANCIS. 1940;Chen and Holmes. 2008).During early 1980s, it evolved into two lineages, designated as B/Yamagata (B/Y) and B/Victoria (B/V), showing distinct antigenicity and transmission pattern (Chen R et al. 2008;Kanegae et al. 1990;Vijaykrishna et al. 2015).Since then, Flu B viruses have co-circulated with Flu A virus during epidemic season and the dominant lineage has changed over years in different geographical locations (Ye et al. 2018;Yang et al. 2018).Several studies have pointed out the differences of the epidemiology between B/Y and B/V lineages, such as younger average ages of persons preferred being attacked by and higher transmissibility of B/V viruses, compared with B/Y viruses (Vijaykrishna et al. 2015;Caini et al. 2019;Yoshihara et al. 2019).
For a long time, much effort has been done on Flu A virus, whereas Flu B virus was relatively less investigated, resulting in poor lineage matches with recommended in uenza vaccine strains and the control options require further improvement (Htwe et al. 2019;Noh et al. 2018).HI tests measure the presence of speci c antibodies of IgG in sera that inhibit virus-mediated agglutination of erythrocytes (Rowe et al. 1999).This is a sensitive assay that is affected, in some level, by nonspeci c hemagglutinin inhibitors in the sera.However, it is a particularly reliable method for in uenza surveillance, providing con rmative result for suspect individuals with successive sera samples, also pro le the herd immunity against upcoming in uenza variant, in which case tiers ≥ 1:40 are considered protective (Skevaki et al. 2012).In the present study, serological assay against two lineages of Flu B viruses circulating in China was determined to investigate the cross-sectional HI antibodies in de ned samples.Our ndings provided insights on low level herd immunity to in uenza B virus prior to the B/Y lineages epidemic would in some level contribute to this epidemic.

Sera samples and ethics approval
Two batches of serum samples from Beijing Tsinghua Changgung Hospital (BTCH) and Beijing Children Hospital (BCH), respectively, were used in the present study.Batch one was 190 sera samples collected during Nov, 2016to Apr, 2017(2016-2017) from outpatients in BTCH with the age ranging from 0.5 to 94 years old.Batch two was 283 sera samples collected from outpatients presenting acute respiratory infections (ARI) from BCH during the Dec, 2017 to Mar, 2018Mar, (2017Mar, -2018)), with the age ranging from 0.5 to 17 years old.Samples in Batch one was subdivided into 3 groups according to ages and diagnosis, group 1 from 69 adults aged from 18 to 94 years old and excluded respiratory diseases, group 2 from 75 patients aged from 0.5 to 17 years old present other than respiratory symptoms, group 3 from 46 patients aged from 0.5 to 17 years old and presenting ARI.All sera samples were remaining from routine clinical tests and were stored at -40 °C till use.The studies did not involve any health-related patient information and were approved by the Ethics Committee of both Beijing Children's Hospital and Beijing Tsinghua Changgung Hospital (Approval No. 2018-k-130 / 17120-0-01).
Representative in uenza viruses for HI assay Type B in uenza viruses isolated from China during the two winter seasons were selected for HI assay on the basis of data collected by the World Health Organization (WHO) as well as our own study (Zhu et al. 2019).For Victoria lineage, B/Jiangxi_yushui/11102/2014 which was phylogenetically and antigenically close to the recommended vaccine strainB/Brisbane/60/2008, belonging to V1A sublineage, was selected as the antigen for HI.For Yamagata lineages, the representative virus B/Beijing/BTCH_71/2018, belonging to Y3 sublineage isolated in early 2018 in Beijing was used (Zhu et al. 2019).The phylogenetic relationship of the selected viruses was presented (Fig. 1).Viruses were ampli ed in 9-11-day-old speci c pathogen free (SPF) chicken embryonated eggs at 35 °C for 2 days.The allantoic uid were collected and cleared by low-speed centrifugation.The titer of the virus was determined by HA assay and viral antigen aliquots was stored at -80 °C till use.All experiments associated with live viruses were performed in the biosafety level 2 plus laboratory.
Figure 1.Phylogenetic tree of the HA gene of type B in uenza viruses used in the study.
Phylogenic tree was constructed using the partial HA genes (370-610NT covering the most important positions distinguishing the Victoria and Yamagata lineages) from the clinic isolates and full-length HA sequences from the GenBank as well as GISAID.Strains labeled with black triangle represented as B/Victoria and B/Yamagata lineage viruses were used as antigens in the present study.

Hemagglutinationinhibition (HI) assay
Serum samples were examined for the presence of antibodies against the hemagglutinin of the selected type B in uenza viruses by HI assay as described previously (Bodewes et al. 2011;Wang et al. 2012).Brie y, serum samples were heat inactivated at 56 °C for 1 hour and then were absorbed by 20% (v/w) turkey red blood cell (TRBC) to reduce the non-speci c binding.2-fold serial dilutions with 1:10 staring dilution of pretreated serum samples were subsequently incubated with 8 hemagglutination units of in uenza virus or phosphate-buffered saline (PBS) for 30 min at 37 °C, and subsequently, 1% TRBC was added.HI pattern were read after incubation for 30 min at 22 °C.The highest dilution of serum that still gave complete inhibition of the hemagglutination was recorded as the titer.Serum samples were considered negative when they failed completely to inhibit agglutination of TRBC by any of the selected viruses.Serum samples collected from mice before and after immunization with each of the type B in uenza viruses were used as negative and positive controls, respectively.Considering previous exposures to Flu B viruses or vaccinations in individual's life would have interference on HI antibody pro le, we took ≥ 1:80 as cut-off limit for "positive reaction" of recently infection.

Statistics
Data analyses were performed using SPSS software (version 17.0; SPSS Inc., Chicago, IL, USA).Nonparametric test was used for comparing the age difference of children and GMT of the sera samples.Correlation of HI titers against two lineages of Flu B virus was employed Spearman tests.The Pearson Chi-square test or Fisher's exact tests were performed to evaluate HI antibodies seroprevalence to Yamagata and Victoria lineages of Flu B virus, with cut-off value as HI titers ≥ 1:80.Two-sided at the 5% level of signi cance was considered statistically remarkable in all tests.

HI titers against two lineages of Flu B viruses in sera from 2016-2017
The HI titers to the two lineages of Flu B viruses in sera from 2016-2017 were compared.Result indicated that the frequency with HI titers over the cut-off value in the 3 subgroups did not show difference, either to B/Y, or to B/V lineage virus (Table 1).HI titers in group 3, the sera from children present ARI did not increase as compared to the group 2, sera from children without ARI (P = 0.47 for B/Y and P = 0.78 for B/V).Hence, we put group 2 and group 3 together (n = 121) to represent HI value in children (CHI) from 2016-2017.The HI titers in CHI group were also compared to subgroup 1 of adults and no difference was found (P = 0.65 for B/Y and P = 0.49 for B/V).The results indicated a lower epidemic of Flu B in 2016-2017 winter, with the HI titers in children and adult showing no difference previous the B/Y epidemic year.2).On the other hand, a moderate correlation was found between HI titers to B/V and B/Y in 2016-2017 (r = 0.58) and 2017-2018 (r = 0.49) (Fig. 2).Antigenic drift of B viruses from 2016-2017 the 2017-2018 Geometric mean titer (GMT) was usually used for evaluating the in uenza vaccine e cacy and potency in de ned population.In the study, the GMT of HI titer against both lineages of Flu B viruses in sera collected in 2017-2018 (74.16 against B/Y and 67.57against B/V, respectively) was much higher than that collected in 2016-2017 (31.87 against B/Y and 26.44 against B/V, respectively).Signi cant increase in the GMT to both lineages could be seen in 2017-2018, as compared to 2016-2017.However, we did not nd difference as the GMT to B/Y and B/V either in 2016-2017 (P = 0.14), or in 2017-2018 (P = 0.11).
Given the same antigens used in the HI assay for sera samples from two different years, we predict that Flu B viruses might undergo antigenic drift and it is the reason for the B/Y lineage virus evades herd immunity.Collectively, these data suggested weak HI response in 2016-2017 to both Flu B viruses may contribute to the severe epidemic of Flu B in winter of 2017 (Table 3), and the antigenic drift produce a transition and higher HI titers against most recent/homologues infection.

Discussion
The HI assay serves as a golden method for serologic diagnosis of in uenza, if suitable viral antigen was employed.In the context that an abrupt spread of in uenza B/Y lineage virus, we performed this HI antibody screening study using cryopreserved sera samples, aiming to evaluate the HI antibody transition after the outbreak of B/Y lineage.Setting 1:80 as cut-off value, we compared the positivity of HI titer and GMT to both B/Y and B/V lineages in 2017-2018 to previous year of 2016-2017.Result indicated that the positivity of HI titer in 2017-2018 was signi cantly higher than that in 2016-2017, and the GMT was also signi cantly increased in 2017-2018 than the previous year.However, we did not nd any difference as the HI titers to B/Y and B/V in the two consecutive years.In 2016-2017, positivity and GMT of HI titer to both B/Y and B/V were at a low level.Signi cantly, in 2017-2018, positivity and GMT of HI titers to B/Y and B/V increased simultaneously, though many reports indicated that B/Y was the leading reason for severe spreading of in uenza during the time (Dayan. 2018).The HI antibody to B/Y lineage with percentage of 66.1% exceeds reported positivity of antigen tests to Flu B at around 10-20%, and reported positivity of RT-PCR assay to Flu B at 38% in ILI cases during 2017-2018 winter, suggesting recessive infection would be a primary reason for acquiring the HI antibodies under Flu B epidemic.
Sudden surge of Flu B/Y virus during the 2017-2018 winter may re ect changes of the viral tness (Virk et al. 2019), as well as herd immunity against this leading strain.The prevailing B/Y variant become to dominate might re ect the vaccine effectiveness (VE) or natural situation cannot provide adequate protective immunity against this strain.In this study, low GMT of HI to both B/Y and B/V in previous epidemic year (2016)(2017) add value to this assumption and reinforce the emergence of the B/Y epidemic.HI titer could not only serve as an indicator for the protective immunity after the immunization, but also hint infection/transmission status in a well-de ned population (Bodewes et al. 2011).Flu B had been detected at very low incidence in ILI patients during 2018-2019 seasonal u epidemic (unpublished data by our group), which could be attributed to previous year's development of HI antibodies to B/Y and B/V lineages.
In China, general public has gradually developed a strong "vaccine hesitancy".People including high risk individuals are reluctant or refuse to be vaccinated despite the availability of in uenza vaccine due to a variety of reasons.In uenza vaccination coverage rate in China was at 1.5% of the total population between 2004 and 2014 (Yang et al. 2016) First of all, we used preserved clinical samples from children with ARI to perform the crosssectional study, the sera were not so typical and previous infection or vaccination might affect production of HI titers.Secondly, the sera samples preserved in 2016-2017 were not su cient enough, we had to make a composition using samples from children with or without ARI together, after excluding difference between the two groups, and to compare to those from 2017-2018.Lastly, since we did not perform HI assay to Flu A, we are not sure the in uence of AbbreviationsHI Hemagglutinin inhibition GMT Geometric mean titer WHO World Health Organization ILI In uenza-like-illness Flu B Type B in uenza virus Flu A Type A in uenza virus B/Y B/Yamagata B/V B/Victoria BTCH Beijing Tsinghua Changgung Hospital BCH Beijing Children Hospital ARI Acute respiratory infections

Figure 2
Figure 2 Comparison of HI titer distributions against two lineages of Flu B viruses by Spearman tests.We can see categories of HI titer to both lineages increased signi cantly in 2017-2018 than in 2016-2017.Moderate correlation can be found between B/Y and B/V lineage as the HI titers in 2016-2017 (a) and 2017-2018 (b).

Table 1
Summary and comparing of HI titers to B/Y and B/V lineages in 2016-2017The percentages were estimated as a fraction of total cases belonging to each category.P values were estimated using χ 2 tests or Fisher's exact tests (between B/Y and B/V of each group).† .CHI: summary of group 2 and group 3 to represent samples from children.Adults: samples from adult in group 1 ‡ .B/Beijing/BTCH_71/2018 was used as antigen representative of B/Yamagata lineage virus.§ HI titer against prevailing Flu B viruses was compared using sera samples from2016-2017 and 2017-2018.All the samples from children aged 0.5-17 years and no difference being found (P = 0.77).There were 28 sera samples in 2016-2017 CHI (n = 121) groups being found as HI positive against both lineages of Flu B viruses, with the frequency at 23.14%, and no difference being found between B/Y and B/V lineage.However, there were 66.08% showed HI positive against B/Y lineage and 59.72% showed HI positive against B/V lineage in the 2017-2018, with no signi cant difference was observed to B/Y and B/V lineages in this season, though the HI positivity to B/Y lineage did mildly increase than B/V (66.08% vs 59.72%, P = 0.054).It was signi cantly increased as the HI titers in 2017-2018 than in 2016-2017, either to B/Yamagata (P < 0.01) and B/Victoria (P < 0.01), indicating that HI antibody in children increased simultaneously, to both B/Y and B/V viruses with the spreading of B/Y lineage (Table

Table 2
Comparing prevalence of HI titers to B/V and B/Y lineages in2016-2017 and 2017-2018The percentages were estimated as a fraction of total cases belonging to each category.P values were estimated using χ 2 tests or Fisher's exact tests (age groups), and p values of 0.05 or less are shown in bold.† .B/Beijing/BTCH_71/2018 was used as the representative of B/Yamagata lineage virus.‡ .B/Jiangxi_Yushui/11102/2014 was used as the representative of B/Victoria lineage virus Figure Comparison of HI titer distributions against two lineages of Flu B viruses by Spearman tests.We can see categories of HI titer to both lineages increased signi 2017-2018 than in 2016-2017.Moderate correlation can be found between B/Y and B/V lineage as the HI titers in 2016-2017 (a) and 2017-2018 (b).

Table 3
GMT of HI titers against two lineages of Flu B viruses in sera from children in 2016-2017 and 2017-2018.based on the HI titer.P values were estimated using nonparametric tests and p values of 0.05 or less are shown in bold.† .B/Beijing/BTCH_71/2018 used as the representative of B/Yamagata lineage virus.‡ .B/Jiangxi_Yushui/11102/2014 used as the representative of B/Victoria lineage virus.
(Rota et al. 1992latAndrews et al. 2015)t of herd immunity against in uenza virus relies on the natural infection.immunizationcoverageandlack of B/Y antigen in current trivalent in uenza vaccine (TIV) in China (before 2018-2019 u season) may provide favorable conditions for the dominance and prevalence of B/Y virus.Long-term co-circulating viruses of Flu A H1/H3 subtype accompanied by decreased HI antibody to both Flu B virus may serve as an immunological niche for the B/Y lineage virus spreading in China during the 2017-2018 winter.Our ndings have implications that quadrivalent in uenza vaccines (QIV) should be timely updated for better prevention against seasonal in uenza epidemics.It has been well documented that antisera raised against Victoria lineage of Flu B virus shared low cross reaction with that against Yamagata lineage(Rota et al. 1992).However, distinct immunodominance of the two lineages of Flu B virus may affect or elicit more potent cross reactive or protective antibodies based on previous history of infection needs more in-depth investigations.Such scenario has been illustrated by pdm/H1N1/2009 virus which replaced previous seasonal H1N1 strain and established in human population apt to induce cross protection antibody response(Victoraet al. 2015;Andrews et al. 2015).Under such background, HI assay could be used as evaluation of infection and herd immunity but not suitable for diagnosis in clinical context as cross reaction between B/Y and B/V lineages from previous infection could not be excluded.In summary, our study indicates that there was a signi cantly elevation of HI antibodies to in uenza viruses B in children with ARI in 2017-2018, compared with children in 2016-2017.On the other hand, the low level of HI antibodies to both B/Y and B/V in 2016-2017 could contribute to the severe B/Y epidemic in 2017-2018 to some extent.Limitations exist in the study.