A TLR9 agonist synergistically enhances protective immunity induced by an Alum-adjuvanted H7N9 inactivated whole-virion vaccine

ABSTRACT

Antigen sparing is an important strategy for pandemic vaccine development because of the limitation of worldwide vaccine production during disease outbreaks. However, several clinical studies have demonstrated that the current aluminum (Alum)-adjuvanted influenza vaccines fail to sufficiently enhance immune responses to meet licensing criteria. Here, we used pandemic H7N9 as a model virus to demonstrate that a 10-fold lower amount of vaccine antigen combined with Alum and TLR9 agonist can provide stronger protective effects than using Alum as the sole adjuvant. We found that the Alum/CpG 1018 combination adjuvant could induce more robust virus-specific humoral immune responses, including higher total IgG production, hemagglutination-inhibiting antibody activity, and neutralizing antibody titres, than the Alum-adjuvanted formulation. Moreover, this combination adjuvant shifted the immune response toward a Th1-biased immune response. Importantly, the Alum/CpG 1018-formulated vaccine could confer better protective immunity against H7N9 challenge than that adjuvanted with Alum alone. Notably, the addition of CpG 1018 to the Alum-adjuvanted H7N9 whole-virion vaccine exhibited an antigen-sparing effect without compromising vaccine efficacy. These findings have significant implications for improving Alum-adjuvanted influenza vaccines using the approved adjuvant CpG 1018 for pandemic preparedness.

KEYWORDS:

• Aluminum hydroxide
• CpG 1018
• TLR9 agonist
• H7N9
• whole-virion vaccine

Introduction

Global equity in vaccination coverage is essential for limiting pandemic outbreaks and preventing the transmission of new strains, according to the experience of the coronavirus disease 2019 (COVID-19) pandemic [1,2]. To achieve this goal, more investment in vaccine manufacturing capacity, novel vaccine product development, and new adjuvant formulation creation is necessary for future pandemics. Additionally, Tregoning et al. suggested that efforts to improve existing vaccine platforms are also possible approaches to maintain the availability of multiple vaccine platforms, thereby increasing manufacturing capacity [3]. To date, several existing vaccine types, including inactivated whole-virion, split-virion, subunit, and recombinant hemagglutinin vaccines, with safe profiles have been used against both seasonal and pandemic influenza [4,5]. Therefore, further improving and refining these existing vaccine platforms will strengthen the response to future influenza pandemics.

The epidemic of avian influenza continues to be a global public health problem. From March 2013 to January 2023, 1,568 cases of human infection with H7N9 avian influenza were reported in China, with 616 deaths (case fatality rate: 39.3%). Notably, the fifth wave of the epidemic, which occurred from October 2016 to September 2017, accounted for approximately half of the total infections. After the implementation of the poultry vaccine strategy in China starting in September 2017, there have been only four cases of human infection with H7N9 virus as of May 2023 [6]. However, the H7N9 virus isolated in 2019 formed a distinct clade with several mutations in the hemagglutinin gene that may be associated with antigenic drift [7]. Additionally, several mutations in the HA and NA genes of H7N9 viruses were identified, leading to enhanced human receptor binding and reduced susceptibility to antiviral drugs [7–9]. To improve vaccine efficacy, adjuvants are necessary for inactivated whole-virion and split-virion H7N9 vaccines, as shown in various clinical studies conducted for pandemic preparedness [10–12].

Aluminum salts are common adjuvants used in various licensed vaccines [13]. Although aluminum salts are also used in the development of avian influenza vaccines, their efficacy and dose-sparing effect remain controversial. Some clinical trials have shown that the addition of aluminum to whole- or split-virion influenza vaccines does not sufficiently enhance antibody responses to meet licensing criteria [14–17] and even reduces immunogenicity [18]. Similarly, animal experiment results showed that the Th2-polarized immunity induced by aluminum-formulated vaccines has no beneficial role in viral clearance [19]. Moreover, aluminum has no or a lower adjuvant effect on influenza vaccines at low HA doses than the MF59 adjuvant [11,20,21]. Therefore, further advances in adjuvant technology are required to improve the induction of immune responses by aluminum-adjuvanted influenza vaccines. Combination with immune potentiators, such as CpG oligonucleotides or monophosphoryl lipids, is a promising approach to improve the immunogenicity of aluminum-formulated vaccines, to regulate the Th1/Th2 profile and to reduce the required antigen dose in preclinical and clinical studies [13,22–24].

CpG oligonucleotides are potent immunostimulants that act through Toll-like receptor 9 (TLR9) signalling and are used in various developed vaccines [25,26]. Of these, CpG 1018 is currently licensed for use in the hepatitis B virus (HBV) vaccine HEPLISAV-B [27]. Despite its limited immunostimulatory effects on influenza vaccines in clinical and animal studies [28,29], CpG has shown great potential as a combination adjuvant. The combination of experimental CpG enhances the immunogenicity of both recombinant hemagglutinin and split-virion vaccines that were formulated with various adjuvants, such as aluminum hydroxide (Alum), calcium phosphate, or MF59 [22,24]. Nevertheless, the efficacy of these combinations may vary depending on the specific virus strain or adjuvant used. Furthermore, previous studies have suggested that the influence of CpG on the binding of antigens to Alum depends on buffer conditions and the nature of the antigen [30], and not all Alum/CpG combination adjuvants may exert a synergistic effect on all types of vaccines [31]. Therefore, the impact of clinically approved CpG 1018 on the immune response induced by Alum-adjuvanted inactivated whole-virion (WV) vaccines is unexpected, even though similar studies have been conducted on split-virion and recombinant HA vaccines.

Here, we evaluated the impact of CpG 1018 on an Alum-formulated H7N9 inactivated WV vaccine in mice at different doses. The immunogenicity, T-cell immune response induction, and vaccine efficacy of the vaccine were examined. We show that Alum combined with CpG 1018 synergistically promotes antibody production and protects against H7N9-induced body weight loss in mice immunized with the H7N9 WV vaccine. The addition of CpG 1018 to the Alum-adjuvanted H7N9 WV vaccine has an antigen-sparing effect without compromising vaccine immunogenicity and promotes a Th1-polarized profile. These findings may enable the practical use of CpG 1018 in Alum-adjuvanted influenza vaccines.

Materials and methods

H7N9 viruses and vaccine bulk

H7N9 A/Anhui/1/2013 (NIBRG-268), A/Guangdong/17SF003/2016 reassortant virus (CBER-RG7D), and A/Gansu/23277/2019 (IDCDC-RG64A) reassortant viruses, all generated using reverse genetics, were obtained from the National Institute of Biological Standard and Control (NIBSC), the Centre for Biologics Evaluation and Research (CBER), and the Centres for Disease Control and Prevention (CDC), respectively. All H7N9 reassortant viruses were propagated in Madin-Darby canine kidney (MDCK) cells with OptiPRO SFM medium supplemented with 4 mM glutamine and 2 μg/mL TPCK-treated trypsin (Sigma). All experiments with H7N9 reassortant viruses were conducted in a biosafety level 2 (BSL-2) laboratory. The bulk of the inactivated H7N9 whole-virion vaccine made from CBER-RG7D virus was produced using an MDCK cell-based manufacturing system by a PIC/S GMP Bioproduction Plant at the National Health Research Institutes (NHRI), Taiwan. HA antigen content in the vaccine bulk was measured by single radial immunodiffusion assay using reference antiserum and antigen from NIBSC, codes 18/112 and 18/196.

Overall experimental design

To evaluate the potential of Alum or Alum/CpG 1018 to reduce the required dosage of the H7N9 WV vaccine, we administered varying amounts of the vaccine to BALB/c mice. Specifically, the mice received the H7N9 WV vaccine containing 0.015, 0.15, or 1.5 μg of HA protein, in combination with 30 μg of Alum, 10 μg of CpG 1018, or no additional adjuvant. As recommended in previous studies [13,32], we selected a mouse-appropriate dosage that was one-tenth of the human dose for both the antigen and adjuvant. Hence, we utilized one-tenth of Alum-adjuvanted H7N9 WV vaccine [11], which contained 1.5 μg of HA and 30 μg of Alum, as a reference dose in this study.

BALB/c mice were intramuscularly immunized twice with the H7N9 WV vaccine in combination with Alum or CpG 1018. ELISA was used to determine the antigen-specific antibody response and immunoglobulin G subclasses. Hemagglutination inhibition (HI) and microneutralization assays were used to analyze vaccine-induced protective antibody responses. Finally, to assess the vaccine-induced immune response in protection against influenza, immunized mice were exposed to modest or high lethal doses of CBER-RG7D virus four weeks after the last vaccination.

Immunization of mice

Female BALB/c mice were obtained from the National Laboratory Animal Centre (Taipei, Taiwan). All animals were housed at the Animal Centre of the NHRI, which has been accredited by AAALAC International. Six- to eight-week-old female BALB/c mice were immunized intramuscularly with 50 μL of the H7N9 WV vaccine twice at 2 weeks apart. For vaccine formulation, the H7N9 WV antigen was mixed with or without Alhydrogel (2% w/v suspension of aluminum hydroxide; Brenntag AG) and CpG 1018 in PBS buffer in an equal final volume. The CpG 1018 oligonucleotide (5’-TGACTGTGAACGTTCGAGATGA-3’) was synthesized by GeneDerix and resuspended in ddH₂O. Unadjuvanted and adjuvanted vaccines were rotated under constant mixing on a rotating mixer for 2 h at room temperature before administration. All animal experiments were conducted according to an IACUC-approved protocol (protocol number: NHRI-IACUC-110044).

Immunoassay

The specific antibody response against H7N9 was determined by ELISA. In brief, 50 μL of 0.5 μg/mL HA protein (H7N9 WV vaccine bulk) in 0.1 M carbonate buffer (pH 9.5) was coated onto 96-well microplates by overnight incubation at 4°C. The coated plates were washed twice with 0.05% Tween 20 in PBS and then blocked with 3% BSA in PBS at room temperature for 1 h. Diluted sera from immunized animals were added to the wells and incubated for 2 h at room temperature. HRP-conjugated goat anti-mouse IgG (1:10000; Thermo Scientific), HRP-conjugated rabbit anti-mouse IgG1 (1:5000; Invitrogen), and HRP-conjugated rabbit anti-mouse IgG2a (1: 5000; Invitrogen) were used as the secondary antibodies. The assay was developed by using the TMB substrate set (BioLegend). The absorbance was measured using a SpectraMax M2 microplate reader (Molecular Device) at 450 nm.

Hemagglutination inhibition (HI) assay

HI titres in mouse sera were determined according to a harmonized protocol [33]. Briefly, mouse sera were pretreated with receptor-destroying enzyme (RDE II, Denka Seiken) and preabsorbed with turkey red blood cells (TRBCs). After removing the TRBCs by centrifugation, the sera were twofold serially diluted from the initial 1:10 dilution and mixed with 4 hemagglutinating units of formaldehyde-inactivated H7N9 virus in a volume of 50 μL at room temperature for 1 h. Next, 50 μL of 0.5% TRBC suspension was added and incubated at room temperature for 40 min. Finally, assay plates were tilted, and the TRBC flow pattern was read. The HI titre was proportional to the highest serum dilution that completely inhibited hemagglutination. Sera that failed to inhibit hemagglutination at the initial dilution of 1:10 were assigned an HI value of 5.

Microneutralization assay

MDCK cells were seeded (3 × 10⁴ cells/well) in 96-well plates for 24 h to form a monolayer. Preimmunization sera and antisera against H7N9 were pretreated at 56°C for 30 min to destroy heat-labile nonspecific viral inhibitory substances. The sera were diluted to an initial dilution of 1/10 with DMEM, added to a well containing 200 TCID₅₀ of H7N9 in a volume of 0.2 mL, and then incubated at 35°C for 2 h. Subsequently, the virus-serum mixture was inoculated onto MDCK cell monolayers, and the cells were incubated at 35°C. Quadruplicates were prepared for each serum dilution. The cytopathic effect in each well was recorded after 4–5 days of incubation. The 50% neutralization (NT₅₀) titre was calculated using the Reed-Muench formula. Neutralization titres below the starting dilution of 1:10 were assigned a value of 5 for calculation purposes.

Cytokine production assay

Seven days after the final vaccination, the mice were sacrificed, and splenocytes were collected and plated at a density of 5 × 10⁶ cells per well in 24-well plates. The cells were stimulated with 5 μg/mL recombinant H7 ectodomain (A/Guangdong/17SF003/2016) produced in the ExpiCHO expression system (ThermoFisher). After stimulation for 3 days at 37°C, the supernatant was harvested and assayed for cytokine production. The levels of secreted mouse IFN-γ, IL-5, IL-13 and IL-2 were evaluated by ELISA using the matching antibody set (Invitrogen) in accordance with the manufacturer’s instructions.

Animal challenge

Four weeks after the last vaccination, the mice were challenged intranasally with a 2- or 10-fold 50% minimum lethal dose (MLD₅₀) of CBER-RG7D in a 20 μL volume under isoflurane anesthesia. Subsequently, the mice were monitored daily for weight loss and survival and were euthanized and scored as dead if more than 20% of their body weight was lost. To determine the viral load in the lung, lung tissues were homogenized in 2 mL of PBS containing 200 U/mL penicillin and 200 μg/mL streptomycin using a gentleMACS® Dissociator (Miltenyi Biotec). After centrifugation at 600 × g for 10 min at 4 °C, the clarified supernatant was harvested for viral RNA quantification.

Quantification of viral RNA load

Clarified supernatant of homogenized lung tissue from H7N9-infected mice was harvested for viral load detection. RNA extraction was carried out on tissue supernatant with QIAamp Viral RNA Mini Kit (Qiagen). RNA extracts were reverse-transcrbed using FIREScript RT cDNA synthesis kit (Solis Biodyne) and Uni12 primer [34]. Viral RNA was quantified by real-time PCR in a QuantStudio 6 Flex Real-Time PCR System (ABI) using the Power SYBR® Green PCR Master Mix (ABI). The viral copy number was estimated by the standard curve method with primers specific for H7N9 HA gene (Forward: TGAAAATGGATGGGAAGGCC, Reverse: TGCCGATTGAGTGCTTTTGT) [35].

Statistical analysis

Statistical data were generated using GraphPad Prism software. The statistical significance of differential findings between experimental groups was determined by unpaired Student’s t test, one-way ANOVA, or two-way ANOVA with Tukey’s or Sidak’s posttest. Significant differences in Kaplan‒Meier survival curves were analyzed with a log-rank test. Differences were considered statistically significant if the p value was ≤ 0.05.

Results

Alum/CpG 1018 robustly enhanced the immunogenicity of the H7N9 WV vaccine

Administration of the H7N9 WV vaccine alone induced H7N9-specific total IgG antibody production (Figure 1(a)) in a dose-dependent manner. The Alum-formulated vaccine significantly increased the antibody titre compared with the H7N9 WV alone vaccine, regardless of antigen dose. In contrast, CpG 1018 alone had no obvious immunostimulatory effect on the H7N9 WV vaccine at a dose of 0.15 μg HA (Figure S1), consistent with previous studies [22,24,28,29]. Importantly, Alum/CpG 1018-formulated vaccines synergistically increased the antibody titres compared to those with Alum- or CpG 1018-adjuvanted vaccines. We also successfully confirmed this synergistic effect of the Alum/CpG 1018 adjuvant system on the quadrivalent split-virion influenza vaccine (Figure S2). After the first immunization, the IgG antibody titre gradually increased to a peak at week 8, slightly decreased to a level similar to that at week 4, and then persisted until week 20 (Figure 1(b)). Similar to a previous study [24], the synergistic effect of CpG 1018 addition was not found when it was combined with MF59-like squalene oil-in-water emulsion (SWE) adjuvant.

Figure 1. Effects of adjuvants on the H7N9 WV vaccine-induced antibody responses. BALB/c mice (n = 5 per group) were intramuscularly immunized twice with the H7N9 WV vaccine in combination with aluminum hydroxide (Alum) or CpG 1018. Serum samples were collected for humoral immune response evaluation at the indicated timepoint after the first immunization. (a, b) The levels of total IgG antibodies against H7N9 WV were assessed by ELISA. (c) H7N9-specific hemagglutination-inhibition (HI) antibodies were quantified by hemagglutination inhibition assay. The gray dashed line indicates a 10-fold initial dilution of serum samples. The black dashed line represents a ≥ 4-fold rise in HI titre, also called 4-fold seroconversion. (d) Vaccine-induced neutralizing activity against H7N9 was evaluated by microneutralization assay. The dashed line indicates a 20-fold initial dilution of serum samples. The log₁₀-transformed IgG titre and log₂-transformed HI and NT titres of sera collected at week 6 after the first immunization were analyzed by two-way ANOVA with Tukey’s posttest. * indicates comparisons of the adjuvant effect among the same antigen dose groups. # indicates comparisons between the various antigen doses that were adjuvanted with Alum/CpG 1018. @ indicates comparisons between the 1.5 μg HA with Alum and 0.15 μg HA with Alum/CpG 1018 groups. *^/#P < 0.05, **^/##/@@P < 0.01, ***P < 0.001, ****^/####P < 0.0001.

Moreover, vaccine-induced protective antibody responses were analyzed by hemagglutination inhibition (HI) and microneutralization assays. After administration of the H7N9 WV alone vaccine, all animals in the 0.015 μg HA group had HI titres below or equal to the detection limit, and 40% of animals from the 0.15 and 1.5 μg HA groups achieved 4-fold seroconversion at week 6 (Figure 1(c)). In contrast, the Alum-adjuvanted groups showed higher HI titres with 4-fold seroconversion and neutralization (NT₅₀) titres, relative to those of the groups without adjuvant treatment (Figure 1(c,d)). Within the Alum/CpG 1018 groups, the addition of CpG 1018 further increased the HI and NT₅₀ titres in the 0.15 and 1.5 μg HA groups but not in the 0.015 μg HA group. Notably, the 0.15 μg HA group with the Alum/CpG 1018 combination adjuvant had a significantly higher specific antibody level, HI titre and NT₅₀ titre than the reference group at week 6 after vaccination. These results supported a dose-sparing effect of the Alum and CpG 1018 combination adjuvant on the H7N9 WV vaccine, with a 10-fold reduction in antigen usage. Importantly, the combination of Alum and CpG 1018 more potently increased specific antibody levels, HI titres and NT₅₀ titres.

Alum/CpG 1018 shifted the antibody response toward IgG2a dominance

IgG2a plays an important role in viral clearance during infection [36,37]. To evaluate the effect of CpG 1018 on IgG isotype switching, the levels of the IgG1 and IgG2a isotypes were measured by ELISA. After two doses of vaccine, the Alum-adjuvanted vaccine markedly increased antigen-specific IgG1 titres in the 0.015 and 0.15 μg HA groups, and this raised IgG1 titre was offset in the groups with CpG 1018 addition (Figure 2(a)). However, this finding was not observed in the 1.5 μg HA group. On the other hand, the antigen-specific IgG2a titres significantly increased in the Alum-adjuvanted H7N9 WV groups (Figure 2(b)), regardless of the antigen dose. The titre was further increased after the combination of Alum with CpG 1018, which was similar to the observation for the total IgG titre (Figure 1(a)). The results suggested that CpG 1018 coadjuvantation redirected the Alum-induced humoral immunity toward the IgG2a-dominated antibody response.

Figure 2. Effects of adjuvants on the H7N9 WV vaccine-induced IgG isotype antibody titres. BALB/c mice (n = 5 per group) were intramuscularly immunized twice with the H7N9 WV vaccine in combination with aluminum hydroxide (Alum) or CpG 1018. Serum samples were collected for humoral immune response evaluation at week 6 after the first immunization. (a, b) Antigen-specific immunoglobulin G subclasses, IgG1 (a) and IgG2a (b), in mouse serum were quantified by ELISA. The log₁₀-transformed IgG1 and IgG2a titres were analyzed by two-way ANOVA with Tukey’s posttest. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Alum/CpG 1018 induced a Th1-polarized immune response

IgG subclass switching is associated with Th1- and Th2-polarized immune responses. To examine the impact of CpG 1018 on Alum-induced T-cell responses, we assessed T-cell responses after immunization with the H7N9 WV vaccine with or without Alum or CpG 1018. Splenocytes from the immunized mice were stimulated with recombinant H7 protein. The secreted levels of Th1-type cytokines (IL-2 and IFN-γ) and Th2-type cytokines (IL-5 and IL-13) were measured by ELISA. The CpG-adjuvanted group produced the lowest amounts of IL-2, IFN-γ, and IL-13 among the groups, and the level of IL-5 was even undetectable in this group (Figure 3(a–d)). In contrast, vaccination with WV/Alum induced higher levels of IL-2, IFN-γ, IL-5 and IL-13. The addition of CpG 1018 to Alum profoundly suppressed the levels of IL-5 and IL-13 (Figure 3(c, d)), while the levels of IL-2 and IFN-γ were maintained or slightly decreased (Figure 3(a, b)). The results indicated that Alum and CpG1018 offset each other’s activity, especially in regulating the Th2-mediated immune response. Vaccine-induced trends in these cytokine levels were observed in mice receiving 0.15 or 1.5 μg of HA antigen. Furthermore, analysis of the Th1/Th2 ratio showed that the H7N9 WV vaccines adjuvanted with CpG 1018 or Alum/CpG 1018 produced Th1-biased responses (Figure 3(e,f)). IFN-γ/IL-5 and IFN-γ/IL-13 ratios were increased by approximately 21.2- and 6.3-fold, respectively, in the 0.15 μg HA/Alum/CpG 1018 group compared with those in the 0.15 μg HA/Alum group. Therefore, these results indicated that the addition of CpG 1018 shifts immune responses toward a Th1 bias.

Figure 3. T-cell responses induced by the adjuvanted H7N9 WV vaccine. BALB/c mice (n = 4 per group) were intramuscularly immunized twice with the H7N9 WV vaccine in combination with aluminum hydroxide (Alum) or CpG 1018. Splenocytes were collected at day 7 after the second immunization, and the levels of secreted IFN-γ, IL-2, IL-5 and IL-13 (a–d) were evaluated after restimulation with recombinant H7 protein. (e, f) The ratios of IFN-γ to IL-5 and IFN-γ to IL-13 were calculated. The log₁₀-transformed cytokine level and IFN-γ/IL-5 and IFN-γ/IL-13 ratios were analyzed by two-way ANOVA with Tukey’s posttest. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Alum/CpG 1018 conferred robust protection against H7N9 challenge

To assess the role of the vaccine-induced immune response in protection against influenza, mice were first exposed to a lethal dose (2 MLD₅₀) of H7N9 virus. The body weight of the mice was monitored daily after challenge as an indicator of disease progression. Upon 2 MLD₅₀ virus challenge, the survival rate of the PBS (control) group was 25% (Figure 4(a)). In contrast, all mice in the Alum- and Alum/CpG 1018-adjuvanted H7N9 vaccine groups survived, regardless of antigen dose. Moreover, the vaccinated mice also showed less body weight loss relative to that of the control group (Figure 4(b)), suggesting that both Alum- and Alum/CpG 1018-adjuvanted H7N9 vaccines were sufficient to confer protection against challenge with a modest dose (2 MLD₅₀) of H7N9 virus. Furthermore, two-way ANOVA indicated that the protection from H7N9-induced weight loss was an adjuvant-dependent effect (two-way ANOVA (within 1.5 μg HA groups): time effect, F (14, 196) = 36.85, P < 0.0001; adjuvant effect, F (1, 14) = 9.858, P = 0.0072; interaction effect, F (14, 196) = 5.614, P < 0.0001; two-way ANOVA (within 0.15 μg HA groups): time effect, F (14, 196) = 31.18, P < 0.0001; adjuvant effect, F (1, 14) = 5.355, P = 0.0364; interaction effect, F (14, 196) = 4.498, P < 0.0001). Sidak’s posttest showed that, even with varying antigen dosages, there was no significant difference in body weight change between groups adjuvanted with the same adjuvant at each time. Notably, on day 3 to day 4 postchallenge, the Alum/CpG 1018 combination adjuvant significantly protected against body weight loss compared with the Alum adjuvant in the groups administered either 0.15 or 1.5 μg HA antigen.

Figure 4. Protective efficacy of the Alum-adjuvanted H7N9 vaccine with or without CpG against H7N9 challenge. BALB/c mice (n = 8 for each group) were intramuscularly immunized twice with PBS or adjuvanted H7N9 WV vaccines. Four weeks after the final immunization, mice were intranasally challenged with (a, b) 2-fold MLD₅₀ or (c, d) 10-fold MLD₅₀ of CBER-RG7D H7N9 virus. (a, c) Survival rate was monitored daily after H7N9 challenge. Significant differences between the PBS group and the other groups were calculated using the log-rank test, **p < 0.01, ***p < 0.001. (b, d) Body weight change (%) of the mice was monitored daily after H7N9 challenge as an indicator of disease progression. Only surviving mice are shown in body weight results that are presented as the mean ± SD. Differences between vaccinated groups in panel b were calculated with two-way ANOVA with Sidak’s posttest. However, significant differences (on day 3 to day 9 postinfection) in panel d were calculated using Student’s t test due to missing data resulting from death. * indicates comparisons between the 1.5 μg HA with Alum and 0.15 μg HA with Alum/CpG groups. ^# indicates comparisons between the 0.15 and 0.015 μg HA groups with Alum/CpG. ^@ indicates comparisons between the 1.5 μg HA with Alum and 0.015 μg HA with Alum/CpG groups. *^/#/@P < 0.05, **^/##P < 0.01, ***P < 0.001, ****P < 0.0001. (e) Viral RNA loads (n = 6 for each group) in the lungs of H7N9-infected mice at day 3 postchallenge were quantified by RT-qPCR. Data are presented as the mean ± SD. Differences between groups were calculated using one-way ANOVA with Tukey’s posttest. ****P < 0.0001.

Next, we utilized a higher challenge dose (10 MLD₅₀) to more rigorously evaluate the impact of vaccine-induced immunity on survival and body weight change. Moreover, to evaluate the potential of the Alum/CpG 1018 combination adjuvant, the HA antigen dose was further reduced to 0.015 μg. Following virus challenge with the higher dose, all control mice died, and an 87.5% survival rate was observed after administration of 1.5 μg HA/Alum or 0.015 μg HA/Alum/CpG 1018 vaccines (Figure 4(c)). In contrast, all mice in the 0.15 μg HA/Alum/CpG 1018 group survived during the observation period. According to log-rank comparison, the three vaccinated groups did not significantly differ from each other. As shown in Figure 4(d), 0.15 μg HA/Alum/CpG exhibited the best prophylactic efficacy on weight loss among the vaccinated groups. Interestingly, mice in the 1.5 μg HA/Alum group lost significantly more body weight on days 3 and 4 postchallenge than mice in the 0.015 μg HA/Alum/CpG 1018 group, although both groups exhibited comparable HI titres and neutralizing antibody levels (Figure 1(c,d)). These findings highlighted the importance of CpG 1018 addition to the Alum-adjuvanted vaccine to induce better protective efficacy against H7N9 challenge. Of note, mice receiving 0.15 μg HA/Alum/CpG 1018 had a significant improvement in disease progression compared with those in the 1.5 μg HA/Alum group upon challenge with 2 (Figure 4(b)) and 10 MLD₅₀ (Figure 4(d)), indicating that the combination of Alum with CpG 1018 could not only achieve a dose-sparing effect for the H7N9 WV vaccine but also confer better protective immunity against challenge.

To evaluate the role of the Alum/CpG1018 combination adjuvant in viral clearance, the viral RNA in lung tissue were analyzed. Upon 10 MLD₅₀ virus challenge, up to 7.2 log (copies/mL) of viral RNA load were detected in the lung on day 3 postchallenge (Figure 4(e)). Consistent with the survival rate in Figure 4(c), the viral RNA load in the 0.15 μg HA/Alum/CpG 1018 group was similar to that in 1.5 μg HA/Alum group, and the viral RNA load showed at least a 2.6 log reduction compared with the control group.

Alum/CpG 1018 increased the cross-reactivity of H7N9 vaccine

To investigate the potential cross-protective immunity of the adjuvanted WV vaccine against various H7N9 strains, we evaluated the cross-reactive HI titre against WHO-selected H7N9 vaccine candidates, including A/Anhui/1/2013 (NIBRG-268) and A/Gansu/23277/2019 (IDCDC-RG64A) H7N9 reassortant viruses derived from the first epidemic wave or the infected case in March 2019, respectively. Consistent with the findings shown in Figure 1(c), the homologous HI titres generated by the Alum/CpG 1018-adjuvanted vaccine were higher than those elicited by the Alum-adjuvanted vaccine (left part in Figure 5). Overall, the geometric mean HI titres against the Anhui virus were approximately 2-3-fold lower than the titres against the homologous strain (left and middle parts in Figure 5), irrespective of the adjuvant type. Notably, only the Alum/CpG 1018 adjuvant system induced heterologous HI titres against the Anhui virus that met the seroprotective threshold HI titre of ≥40, whereas approximately 50% of the animals in the Alum-adjuvanted groups did not achieve this level (middle part in Figure 5). Moreover, the cross-reactivity against the Gansu strain was remarkably lower than that against the Anhui strain, with all animals in both the Alum and Alum/CpG 1018-adjuvanted groups exhibiting HI titres below the detection limit (right part in Figure 5).

Figure 5. Effects of adjuvants on the H7N9 WV vaccine-induced cross-reactive antibody responses. BALB/c mice (n = 8 for each group) were intramuscularly immunized twice with PBS or adjuvanted H7N9 WV vaccines. Serum samples were collected for cross-reactivity evaluation at week 4 after the first immunization. Homologous hemagglutination-inhibition (HI) titres against CBER-RG7D (Guangdong) and heterologous HI titres against NIBRG-268 (Anhui) and IDCDC-RG64A (Gansu) were quantified by hemagglutination inhibition assay. The gray dashed line indicates a 10-fold initial dilution of serum samples. The black dashed line represents a ≥ 4-fold increase in HI titre. Significant differences between the homologous strain and heterologous strains were calculated using one-way ANOVA with Tukey’s posttest. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Discussion

In this study, we attempted to investigate the contribution of CpG 1018 to the immunogenicity of the Alum-adjuvanted H7N9 WV vaccine. When mice were exposed to the vaccine with the Alum/CpG 1018 combination adjuvant, a more robust humoral immune response was found in the 1.5 and 0.15 μg HA groups, including virus-specific IgG, HI, and NT₅₀ titres, compared to that in the Alum-adjuvanted groups (Figure 1). The addition of CpG 1018 to the Alum-adjuvanted influenza vaccine shifted the immune response toward a Th1-biased IgG2a-dominated humoral response (Figures 2 and 3). Furthermore, this combination adjuvant required only one-tenth of HA antigen (0.15 μg HA) to confer optimal protection against body weight loss upon challenge with a higher dose (10 MLD₅₀) of H7N9 virus (Figure 4). However, the reference group (1.5 μg HA/Alum) exhibited suboptimal protection, even in the presence of an HI titre ≥ 40. This study showed that the addition of CpG 1018 to the Alum-adjuvanted influenza vaccine can synergistically promote antibody production and protect against viral challenge in an antigen-sparing manner.

The use of combination adjuvants is a common approach for developing new adjuvants with diverse mechanisms of action to stimulate the immune response, resulting in a more effective and durable immune response and antigen dose sparing [38,39]. Previous studies have shown that combination with other adjuvants is necessary for the immunostimulatory capacity of experimental CpG in recombinant HA and split-virion influenza vaccines [22,24]. However, this is the first study to demonstrate that the addition of the approved adjuvant CpG 1018 has antigen-sparing ability in Alum-adjuvanted WV vaccines in mice. We demonstrated that the combination of Alum and CpG 1018 efficiently increased HI and NT titres, induced Th1-type immune responses, and provided protection against H7N9 challenge. Relative to the reference HA dose (1.5 μg), these benefits of the Alum/CpG 1018 combination were found in the mice receiving 1.5 or 0.15 μg HA doses but were not observed in the group with 0.015 μg of HA, indicating that this combination enables at least tenfold dose sparing. That is, the Alum/CpG 1018 adjuvant formulation can increase dose availability tenfold while maintaining the same vaccine productivity for pandemic preparedness.

Several clinical studies on influenza vaccines have shown that aluminum adjuvants fail to sufficiently enhance immune responses to meet licensing criteria (seroconversion factor, seroconversion rate, and seroprotection rate) for assessing the clinical protection of seasonal influenza vaccines [14–17]. Although it remains uncertain whether the criteria are applicable to animal experiments and nonseasonal influenza studies, this study further used the vaccine adjuvanted with Alum alone as a reference to evaluate the potency of the Alum/CpG 1018 combination adjuvant in the H7N9 WV vaccine. Our results demonstrated that the addition of the approved adjuvant CpG 1018 to the Alum-based H7N9 WV vaccine increased vaccine efficacy to meet the seroprotective threshold HI titre of ≥ 40, making it simultaneously superior to the reference group. Thus, it is worthwhile for further research to prove the immunogenicity and dose-sparing effect of Alum/CpG 1018-adjuvanted influenza vaccines in a ferret model and in clinical trials.

Adjuvants play a crucial role in eliciting both humoral and cellular immune responses through vaccination. Optimizing distinct adjuvant formulations for various antigens is notably important. Based on Wack’s report [24], we observed Alum alone, when formulated with seasonal trivalent influenza vaccine, only induced minimum HI titres. However, the addition of CpG dramatically enhanced H1N1- and H3N2-specfic HI titres. Conversely, MF59 alone elevated HI titres against H1N1, H3N2, and B viruses, yet the inclusion of CpG in the MF59 adjuvant failed to further enhance HI titres against H1N1 and B viruses. These results indicated the significance of both adjuvant formulations and antigen choice in achieving elevated HI titres. Moreover, the CpG (CpG1268) used in Wack’s study differs from that in our report (CpG1018). This variation in CpG sequences could result in distinct effects within various formulations (MF59 vs. Alum). Our results highlight that the Alum/CpG adjuvant system has a superior adjuvant effect over MF59-like SWE adjuvant (Figure 1(b)). In addition to the possible suitability of Alum/CpG for H7N9 vaccine, this divergence could also raise from the use of different buffer systems (phosphate vs. histidine buffers) in vaccine formulation. The efficacy of aluminum-containing vaccines can be affected by factors such as adsorption intensity, phosphorus-aluminum molar ratio, pH, ionic strength, and antigen type [40]. Notably, phosphate has been shown to enhance humoral and cellular immune responses triggered by aluminum hydroxide-containing vaccines, accompanied with a decrease in adsorption intensity [41]. Furthermore, the adsorption intensity of Alum to CpG oligonucleotide diminished upon the addition of phosphate, and the use of CpG oligonucleotide reduced the adsorptive capacity of Alum for certain antigens [30]. Thus, besides the ratio of Alum, CpG oligonucleotide and antigen in vaccine formulation, buffer composition should be optimized to ensure adequate adsorption of antigen and CpG oligonucleotide, as well as to facilitate optimal immune response stimulation.

The use of combination adjuvants, despite concerns about cost and potential side effects, can synergistically activate a more potent immune response through multiple mechanisms [38,39]. Additionally, the dose-sparing effect of these adjuvants may ultimately reduce the cost of vaccination. Although the cost of a vaccine may vary depending on the amount of antigen used and the addition of adjuvants, the cost-effectiveness of adjuvanted vaccines primarily depends on their effectiveness. Cost-effectiveness analyses of seasonal influenza vaccines have shown that adjuvanted vaccines are cost-effective primarily due to their effectiveness in older individuals at the highest risk of influenza-related complications and mortality [42,43]. In a pandemic situation, the use of more effective vaccines, especially for H5N1 and H7N9, which have high fatality rates, can significantly impact outbreak control and healthcare utilization. Adjuvants can increase the availability of low-dose whole-virus influenza vaccines several times or even tens of times that of traditional doses [44], which is critical for epidemic control during influenza pandemics. Regarding the safety of the Alum/CpG 1018-adjuvanted H7N9 vaccine, although it requires further investigation, a phase II clinical trial of the Alum/CpG 1018-adjuvanted COVID-19 vaccine has already demonstrated the safety of this adjuvant system [45].

Recent studies (Kang et al. and Strohmeier et al.) have reported the potential of using CpG 1018 as an effective adjuvant to enhance the efficacy of recombinant HA and NA vaccines [46,47]. However, the immunostimulatory effect of CpG adjuvant alone on influenza vaccines remains controversial. Several animal studies, including this study, have demonstrated that 10 μg of CpG alone is insufficient to improve the immunogenicity of influenza vaccines [22,24]. Additionally, a clinical study revealed that 1 mg of CpG 7909 failed to improve humoral responses induced by a trivalent split influenza vaccine [29]. Kang et al. reported the immune response induced by 0.5 μg of recombinant HA and 40 μg of CpG 1018, but this did not meet the common licensing criteria (HI titre of ≥ 40), and only a 40% survival rate was observed after adjuvanting with CpG 1018 [46]. On the other hand, Strohmeier et al. showed that at least 3 μg of recombinant NA antigen was required when adjuvanted with CpG alone to induce robust protection against virus challenge in a mouse model [47]. However, Strohmeier et al. noted that the antigen dosage of 3 μg used in their study corresponded to 1/15th of the human dose, which is a frequently used dosage of recombinant antigen in mouse studies. In light of this, they suggested that raising the NA antigen dosage above 45 μg in future human trials may not be feasible. In contrast, our study has demonstrated that the Alum/CpG 1018 adjuvant system provides sufficient protection from virus challenge in a mouse model with only 0.15 μg of HA in the WV antigen. Notably, this dosage is equivalent to only 1/100th of the commonly used human antigen dosage. Despite the differences in vaccine platforms between inactivated whole-virion and recombinant protein vaccines, our study overall provides evidence that the Alum/CpG 1018 adjuvant system has a higher immunostimulatory effect and better antigen-sparing potential than CpG alone.

Following natural virus infection, mice undergo general IgG subclass switching to IgG2a [48,49], which is known for its potent Th1-biased antibody response in pathogen elimination [36,37]. Therefore, an effective adjuvant for vaccines against infectious diseases not only promotes vaccine efficacy but also polarizes the immune response toward the Th1 profile. Although Alum is a Th2-type adjuvant [19], the combination of Alum and CpG 1018 could elicit a Th1-biased immune response, as demonstrated by high IgG2a titres (Figure 2(b)) and high IFN-γ/IL-5 and IFN-γ/IL-13 ratios (Figure 3(e,f)). Notably, elderly individuals display an age-related shift from Th1-polarized to Th2-polarized immune responses following vaccination [50]. Given its Th1-promoting properties, the Alum/CpG 1018 combination adjuvant may serve as a suitable adjuvant system to meet the vaccination requirements of elderly individuals with Th2-prone immunity. However, our results showed that Alum/CpG 1018 combination adjuvant was more effective in preventing body weight loss compared to Alum adjuvant (Figure 4(b, d)), but not in survival rate and virus clearance (Figure 4(a, c, e)). These findings suggest that the immune response profile elicited by a vaccine may influence the prognosis, and further investigations are warranted to fully comprehend the implications of these observations.

In addition to a robust virus-specific antibody response, the prophylactic effect of Alum/CpG 1018-adjuvanted vaccines against H7N9 infection may be partly attributable to the induction of a Th1-biased immune response. When challenged with a higher dose (10 MLD₅₀), the reference HA dose (1.5 μg) with Alum and the 100-fold lower HA dose (0.015 μg) with Alum and CpG 1018 conferred equivalent survival rates; however, the group with CpG 1018 addition exhibited stronger protection against body weight loss (Figure 4(c, d)). Although the HI, NT₅₀, and total IgG titres in the two groups were comparable (Figure 1), the discernible difference between the groups was the IgG subclass evoked (Figure 2). The Alum/CpG 1018 combination adjuvant induced an IgG2a-dominated antibody response, whereas Alum alone triggered IgG1 dominance. Overall, these results suggested that the addition of CpG 1018 to the Alum-adjuvanted vaccine may elicit immune responses with better antiviral activity.

In this study, we also evaluated the cross-protective potential of Alum/CpG 1018-adjuvanted WV vaccines. Our results demonstrated that Alum/CpG 1018-adjuvanted WV vaccines may be more potent cross-protect against the Anhui strain than the Alum-adjuvanted WV vaccine. Consistent with a previous study [51], an AddaVax-adjuvanted H7 vaccine derived from A/Guangdong/17SF003/2016 provided cross-reactivity and protection against Anhui-like H7N9 challenge in mice. However, we found that neither the Alum/CpG 1018- nor the Alum-adjuvanted WV vaccines provided cross-protection against the recently emerged H7N9 A/Gansu/23277/2019 virus. The Gansu variant exhibited remarkable genetic divergence from the A/Guangdong/17SF003/2016 H7N9 virus, with 19 substitutions in the head (Table S1) and stem regions (Table S2) of hemagglutinin, along with acquired potential glycosylation sites at N123 and N149 in the head region. These findings align with the observed cross-reactivity of antiserum from infected ferrets and hyperimmunized sheep (NIBSC 18/112 (Figure S3)). It is evident that adjuvant development alone cannot achieve comprehensive enhancement of cross-protection against all viruses. Therefore, in the face of significantly mutated viruses, the update of new vaccine strains for the production of inactivated WV vaccines remains crucial.

Aluminum adjuvants have been used in humans for a long time, and the safety profile of CpG 1018 at a high dose (3 mg/dose) is recognized in the use of HBV vaccine. In this report, we demonstrated that the formulation of aluminum hydroxide with CpG 1018 can not only reduce the antigen used but also elicit better protective immunity against the H7N9 virus than that with Alum alone, and this information could be used to solve the dilemma of insufficient production capacity of pandemic vaccines. Moreover, the Th1-polarized and IgG2a-dominated immune responses induced by the Alum/CpG 1018-formulated vaccine may be more helpful to combat future pandemics. Therefore, this Alum/CpG 1018 adjuvant system provides a platform to strengthen the vaccine efficacy and production capacity of existing aluminum-formulated vaccines for various diseases, especially avian influenza and elderly individual-used seasonal influenza vaccines.

Acknowledgments

The authors are grateful to the Bioproduction Plant at the NHRI for supplying the H7N9 vaccine bulk. TTT and SJL designed the experiments, analyzed the data, and prepared the manuscript. KMC, RYC and JRC aided in manuscript review. TTT, KMC and IHC performed the experiments. TTT and RYC developed reagents. TTT, KMC, RYC, JRC, and SJL provided supervision and oversaw final manuscript preparation.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The authors declare that the data supporting the findings of this study are available within this paper or are available from the corresponding author upon reasonable request.

Additional information

Funding

This study was funded by National Health Research Institutes [grant Nos. IV-109-GP-02, IV-110-GP-02, and IV-111-GP-02 to L.S.J.) of Taiwan.