Po-Ling Chen

Background: Conformational stability has been shown to modulate the immunogenicity of structural class 1 viral fusion glycoproteins, yet the relationship between influenza A hemagglutinin (HA) stability and antigenicity remains poorly understood. Methods: Here, we screened a panel of 21 A/Hawaii/70/2019 (H1N1) HA mutants for expression, cleavage, and acid and conformational stability in cells. Twelve mutant HA proteins expressed from transfected plasmid constructs that showed relatively robust expression compared to WT also exhibited either altered stability or glycosylation deletion. mRNA-LNPs were generated containing these 12 HAs, along with the wild-type (WT) HA, to investigate the mutational effects on mRNA-LNP vaccine immunogenicity and protection in mice. Hemagglutination inhibition, microneutralization, and total serum IgG ELISAs were performed using a panel of H1N1 antigens to evaluate humoral immune responses at 28 days post vaccination. The vaccinated mice were then subjected to lethal challenge with a forward-drifted H1N1 virus. Results: Most mutant HA candidates elicited WT-like serological responses and provided protection against challenge, although noticeable decreases in serological responses and, to a lesser extent, protection were observed, especially among G72K- and R109E-vaccinated mice. Of note, substitutions at position E107 enhanced total IgG responses against the HA stalk. Modest, but detectable, increases in antibody-dependent cellular cytotoxicity (ADCC) were also observed, particularly with E107R. Although substitution-specific differences in weight loss were found in E107-vaccinated mice challenged with a mouse-adapted Cal/09 virus, these differences were minor, and protection against the forward-drifted challenge virus and a heterologous PR8 challenge virus was no different from that in mice vaccinated with WT HA. Conclusions: These findings demonstrate the ability of specific substitutions to alter components of humoral immunity by shifting HA domain-specific antibody responses, which could prove useful in the design and development of HA vaccine antigens possessing optimized immunogenicity.

While current influenza vaccines often lack broad protection against antigenically drifted strains, some modified hemagglutinin (HA) protein antigens have shown promise in eliciting broadly neutralizing antibodies against conserved epitopes. During infection, the mildly acidic environment of the late endosome triggers irreversible HA conformational changes resulting in a post-fusion structure with altered antigenicity. While enhancing the stability of other structural class I viral fusion protein antigens has been instrumental in improving the effectiveness of COVID-19 and RSV vaccines, the role of HA stability in influenza vaccine immunogenicity is relatively unclear. Here, we used the nucleoside-modified mRNA-LNP platform to test engineered HA antigens with specific acid-stabilizing mutations (E47K, K58I, R106K, and K153E) in the HA stalk. All mutations increased HA acid stability, but E47K and R106K did not increase immunogenicity. K153E and K58I, but not E47K and R106K, enhanced the cell-surface expression of the HA protein in vitro. In mice, K153E- and K58I-containing mRNA-LNP vaccines elicited increased neutralizing antibody titers against homologous virus. K153E conferred greater protection than wild-type vaccine against lethal heterologous A/PR/8/34 challenge at low doses (0.5–1.0 µg), despite the absence of neutralizing antibodies against the challenge strain. K153E also elicited greater expansion of antigen-specific antibody-secreting cells (ASCs) in the bone marrow, as well as cross-reactive T follicular helper (Tfh) cells in the spleen. For the vaccines studied, increased HA expression was a stronger correlate of mRNA-LNP enhancement than increased HA stability.

Because antigenic drift primarily generates amino-acid changes in the membrane-distal hemagglutinin (HA) head, broadly neutralizing antibodies (bNAbs) are being developed to target conserved epitopes in the membrane-proximal stem. Mutations to HA2 residue A44, a buried residue beneath the central stem epitope, in 2009 H1N1 viruses have been shown to cause resistance to stem-binding bNAbs. Here, we introduced A44V and A44T mutations into A/Tennessee/1-560/2009 (TN09) and A/Puerto Rico/15/2018 (PR18) and investigated their effects in cell culture, mice, and ferrets. In both virus strains, the mutations decreased HA and virus stability and decreased bNAb binding and neutralization in vitro. The mutations reduced pathogenicity and lung replication in DBA/2J mice. Ferrets were inoculated with PR18 wild-type (WT) or A44V virus, and the A44V mutation reduced day-1 and peak nasal virus titers. Airborne transmission in the A44V group occurred only after genotypic reversion (HA2-V44A) or acquisition of a distal re-stabilizing mutation (HA2-I77M). Compared to WT, an engineered PR18 virus containing both HA2 mutations (A44V and I77M) had similar growth and pathogenicity in mice in addition to decreased binding and neutralization by bNAbs. Overall, this work provides insight into the role of HA stability during HA stem-epitope remodeling that results in virus resistance to stem-binding bNAbs.

Reporter viruses are valuable tools for studying infections at the cellular level and in living animals. They also enable rapid, high-throughput antiviral drug screening and serological studies. We previously developed a bioluminescence-based reporter virus, rTN09-PA-Nluc, derived from influenza A/Tennessee/1-560/2009 (TN09, pH1N1) in which a NanoLuc (Nluc) reporter protein was fused to the PA protein. Reduced growth of rTN09-PA-Nluc in MDCK cells and mice was restored by mutations arising from mouse adaptation. Here, to test the hypothesis that the growth defect resulted from the PA-Nluc protein fusion, we generated the luciferase reporter virus rTN09-PA-Nluc/SG, which undergoes StopGo translation to yield separate PA and NLuc proteins along with a proportion of the PA-Nluc fusion. The rTN09-PA-Nluc/SG virus had greater protein expression and increased replication in MDCK cells compared to rTN09-PA-Nluc. The reporter virus encoding StopGo translation was superior to the virus without it in bioluminescence-based virus neutralization assays in vitro, providing results in 24 h as opposed to 3 days using unmodified influenza virus and standard neutralization assay protocols. However, the reporter virus encoding StopGo translation remained attenuated in mice. Mouse-adaptive mutations were needed for full virulence and efficient non-invasive imaging in mice. Overall, these findings demonstrate the benefit of incorporating StopGo translation into influenza reporter viruses for in vitro assays, yet mouse-adapted mutations appeared superior in mice.

The adaptation of egg-derived H7N9 candidate vaccine virus (CVV) in the mammalian cell line is an approach to developing a high-growth virus strain for the mass production of vaccine manufacturing. The adaptive mutations that occur in hemagglutinin (HA) are critical to the activity and potency of the vaccine virus. Previously, we identified a new mutation of A169S in the HA protein of an MDCK-adapted H7N9 vaccine virus (A/Anhui/2013, RG268); however, whether and how this mutation affects vaccine potency remain to be investigated. In this study, we serially passaged RG268 in MDCK cells and found that the HA titer and the TCID₅₀ of the passaged virus RG268-M5 were 4-fold (HA units/50 μL) and 3.5-fold (log₁₀ TCID₅₀/mL) higher than those of the original CVV. By inspecting tandem MS spectra, we identified a new glycosylation site at N167 near the receptor binding site of the HA protein of RG268-M5. Flow cytometry results revealed that RG268-M5 could efficiently infect MDCK cells and initiate viral protein replication as well as that of RG268. Though the new glycosylation site is in the antigenic epitope of viral HA protein, the HI assay result indicated that the antigenicity of RG268-M5 was similar to RG268. Additionally, immunizing mice with RG268-M5 mixed aluminum hydroxide could induce potent antibody responses against the homologous and heterologous H7N9 viruses in vitro whereas the titers were comparable with those from the RG268 group. These results provide in-depth structural information regarding the effects of site-specific glycosylation on virus properties, which have implications for novel avian influenza vaccine development.

Human infections with avian-origin H7N9 influenza A viruses were first reported in China, and an approximately 38% human mortality rate was described across six waves from February 2013 to September 2018. Vaccination is one of the most cost-effective ways to reduce morbidity and mortality during influenza epidemics and pandemics. Egg-based platforms for the production of influenza vaccines are labor-intensive and unable to meet the surging demand during pandemics. Therefore, cell culture-based technology is becoming the alternative strategy for producing influenza vaccines. The current influenza H7N9 vaccine virus (NIBRG-268), a reassortant virus from A/Anhui/1/2013 (H7N9) and egg-adapted A/PR/8/34 (H1N1) viruses, could grow efficiently in embryonated eggs but not mammalian cells. Moreover, a freezing-dry formulation of influenza H7N9 vaccines with long-term stability will be desirable for pandemic preparedness, as the occurrence of influenza H7N9 pandemics is not predictable. In this study, we adapted a serum-free anchorage-independent suspension Madin-Darby Canine Kidney (MDCK) cell line for producing influenza H7N9 vaccines and compared the biochemical characteristics and immunogenicity of three influenza H7N9 vaccine antigens produced using the suspension MDCK cell-based platform without freeze-drying (S-WO-H7N9), the suspension MDCK cell-based platform with freeze-drying (S-W-H7N9) or the egg-based platform with freeze-drying (E-W-H7N9). We demonstrated these three vaccine antigens have comparable biochemical characteristics. In addition, these three vaccine antigens induced robust and comparable neutralizing antibody (NT; geometric mean between 1016 and 4064) and hemagglutinin-inhibition antibody (HI; geometric mean between 640 and 1613) titers in mice. In conclusion, the serum-free suspension MDCK cell-derived freeze-dried influenza H7N9 vaccine is highly immunogenic in mice, and clinical development is warranted.

The embryonated egg-based platform currently produces the majority of seasonal influenza vaccines by employing a well-developed master donor virus (MDV, A/PR/8/34 (PR8)) to generate high-growth reassortants (HGRs) for A/H1N1 and A/H3N2 subtypes. Although the egg-based platform can supply enough seasonal influenza vaccines, it cannot meet surging demands during influenza pandemics. Therefore, multi-purpose platforms are desirable for pandemic preparedness. The Vero cell-based production platform is widely used for human vaccines and could be a potential multi-purpose platform for pandemic influenza vaccines. However, many wild-type and egg-derived influenza viruses cannot grow efficiently in Vero cells. Therefore, it is critical to develop Vero cell-derived high-growth MDVs for pandemic preparedness. In this study, we evaluated two in-house MDVs (Vero-15 and VB5) and two external MDVs (PR8 and PR8-HY) to generate Vero cell-derived HGRs for five avian influenza viruses (AIVs) with pandemic potentials (H5N1 clade 2.3.4, H5N1 clade 2.3.2.1, American-lineage H5N2, H7N9 first wave and H7N9 fifth wave). Overall, no single MDV could generate HGRs for all five AIVs, but this goal could be achieved by employing two in-house MDVs (vB5 and Vero-15). In immunization studies, mice received two doses of Vero cell-derived inactivated H5N1 and H7N9 whole virus antigens adjuvanted with alum and developed robust antibody responses.

Background: Influenza vaccine manufacturers traditionally use egg-derived candidate vaccine viruses (CVVs) to produce high-yield influenza viruses for seasonal or pandemic vaccines; however, these egg-derived CVVs need an adaptation process for the virus to grow in mammalian cells. The low yields of cell-based manufacturing systems using egg-derived CVVs remain an unsolved issue. This study aimed to develop high-growth cell-derived CVVs for MDCK cell-based vaccine manufacturing platforms. Methods: Four H7N9 CVVs were generated in characterized Vero and adherent MDCK (aMDCK) cells. Furthermore, reassortant viruses were amplified in adherent MDCK (aMDCK) cells with certification, and their growth characteristics were detected in aMDCK cells and new suspension MDCK (sMDCK) cells. Finally, the plaque-forming ability, biosafety, and immunogenicity of H7N9 reassortant viruses were evaluated. Results: The HA titers of these CVVs produced in proprietary suspension MDCK (sMDCK) cells and chicken embryos were 2- to 8-fold higher than those in aMDCK cells. All H7N9 CVVs showed attenuated characteristics by trypsin-dependent plaque assay and chicken embryo lethality test. The alum-adjuvanted NHRI-RG5 (derived from the fifth wave H7N9 virus A/Guangdong/SP440/2017) vaccine had the highest immunogenicity and cross-reactivity among the four H7N9 CVVs. Finally, we found that AddaVax adjuvant improved the cross-reactivity of low pathogenic H7N9 virus against highly pathogenic H7N9 viruses. Conclusions: Our study indicates that cell-derived H7N9 CVVs possessed high growth rate in new sMDCK cells and low pathogenicity in chicken embryo, and that CVVs generated by this platform are also suitable for both cell- and egg-based prepandemic vaccine production.

Outbreaks of infection by novel avian influenza virus strains in humans cause public health issues worldwide, and the development of vaccines against such novel strains is the most effective method for the prevention of these virus outbreaks. All types of vaccines must be tested for potency before use; thus, quantitative potency assays are needed for influenza vaccines. The single radial immunodiffusion (SRID) assay is considered the gold standard for quantification of influenza virus antigens, and the SRID reference reagents are essential for the determination of vaccine potency. However, it remains debatable whether reference reagents derived from egg-based vaccine platforms can be used to precisely quantify non-egg-derived vaccines; thus, influenza vaccine production using cell-based platforms has attracted increasing attention. To evaluate the utility of reference reagents derived from a cell-based influenza vaccine platform, we prepared cell-based reference reagents from MDCK cell-grown viruses and compared them with egg-derived reference reagents. A primary liquid standard (PLS) was purified from cell-derived candidate influenza vaccine viruses, and hemagglutinin (HA) antigen content was determined by a densitometric method. The produced PLS could be stored at 4°C for more than 10 months. We also established a simple HA protein purification method for goat antiserum preparation, and the performance of the resulting antiserum was compared to that of standard reagents obtained using different production platforms. The results of this study indicate that these reference reagents can be used for both cell-based and egg-based production platforms and that the differences between these two types of platforms are negligible.