Effects of Adjuvants on the Immunogenicity and Efficacy of a Zika Virus Envelope Domain III Subunit Vaccine

Zika virus (ZIKV), a mosquito-borne flavivirus, has attracted global attention due to its close association with congenital Zika syndrome and neurological diseases, and transmission through additional routes, such as sexual contact. Currently there are no vaccines approved for ZIKV, and thus, there is an urgent need to develop an effective and safe ZIKV vaccine. Domain III (DIII) of the ZIKV envelope (E) protein is an important vaccine target, and a vaccine developed using a mutant DIII of E (EDIII) protein protects adult and pregnant mice, and unborn offspring, against ZIKV infection. Here, we have used immunocompetent BALB/c mice treated with anti-interferon-α/β receptor 1 (Ifnar1) antibodies to investigate whether three adjuvants (aluminum (Alum), monophosphoryl lipid A (MPL), and MF59), either alone or in combination, could improve the efficacy of this EDIII subunit vaccine. Our data show that, although vaccine formulated with a single adjuvant induced a specific antibody and cellular immune response, and reduced viral load in mice challenged with ZIKV, the combination of Alum and MPL adjuvants led to a more robust and balanced immune response, stronger neutralizing activity against three recent ZIKV human strains, and greater protection against a high-dose ZIKV challenge. Particularly, the combination of Alum with MPL significantly reduced viral titers and viral RNA copy numbers in sera and tissues, including the male reproductive organs. Overall, this study has identified the combination of Alum and MPL as the most effective adjuvant for ZIKV EDIII subunit vaccines, and it has important implications for subunit vaccines against other enveloped viruses, including non-ZIKV flaviviruses.


Introduction
Zika virus (ZIKV) is a mosquito-borne flavivirus first isolated from a rhesus macaque in Uganda in 1947 [1]. Most ZIKV infections are asymptomatic or show mild clinical symptoms, such as fever, headache, or rash. However, some adults infected with ZIKV present with Guillain-Barré syndrome (GBS), a disorder of the nervous system that results in muscle weakness and/or paralysis [2,3]. In addition, infection during pregnancy may cause congenital Zika syndrome (CZS), which is associated Laboratory Animals, National Research Council Committee [28]. The protocols have been approved by the Institutional Animal Care and Use Committee (IACUC) of the New York Blood Center (permit numbers: 344 and 345).

Expression and Purification of a Recombinant ZIKV EDIII Protein
This protein was prepared as previously described [27]. Briefly, ZIKV DIII of E protein containing M375N and E377T mutations at residues 375 and 377 of E protein (hereinafter EDIII) was constructed using a mutagenesis kit (Thermo Fisher Scientific, Waltham, MA, USA) and a ZIKV wild-type plasmid expressing residues 298-409 (DIII) of E protein and a C-terminal Fc tag of human (hFc) IgG1 [26,27]. The recombinant EDIII protein was transiently expressed in the culture supernatant of 293T cells, and purified by protein A affinity chromatography (GE Healthcare, Chicago, IL, USA).

ZIKV Plaque-Forming Assay and Plaque Reduction Neutralization Test (PRNT)
Three recent ZIKV human strains, including R103451 (2015/Honduras), PAN2016 (2016/Panama), and PRVABC59 (2015/Puerto Rico), were used in the study. Briefly, viruses were grown in Vero E6 cells and detected for viral titers using a plaque-forming assay [27,32]. Mouse sera (about 50 µL) and tissues (about 20 mg for eye, and 40 mg for heart, spleen, muscle, and brain) collected 3 days post-challenge were also detected for ZIKV titers as described above, and the detection limits were about 20 plaque-forming unit (PFU)/mL for sera, 50 PFU/g for eye, or 25 PFU/g for heart, spleen, muscle, and brain tissues. Neutralizing antibodies in immunized mouse sera were detected by the PRNT as previously described [26,27]. Briefly, 100 PFU of ZIKV was incubated with 2-fold serial dilutions of mouse sera at 37 • C for 1.5 h, which were added to Vero E6 cells and incubated at 37 • C for 1 h. The cells were then overlaid with DMEM containing 1% carboxymethyl cellulose and 2% FBS, cultured at 37 • C for 4-5 days and further stained with 0.5% crystal violet. The neutralizing titer based on the serum neutralizations at a 50% plaque reduction (PRNT 50 ) was calculated using the CalcuSyn computer program [27,33,34].

Challenge of Mice with ZIKV
The immunized mice were challenged with ZIKV as previously described [27,35]. Briefly, nine days post-last immunization of ZIKV EDIII protein with or without respective adjuvant(s), or PBS control, mice were pretreated with a blocking anti-interferon-α/β receptor 1 (Ifnar1) antibody (2 mg/mouse, Leinco Technologies, Fenton, MO, USA); and 24 h later, they were intraperitoneally (i.p.) challenged with ZIKV (strain R103451, 6 × 10 5 PFU; 200 µL/mouse). Mouse splenocytes were isolated at 3 days post-challenge and detected for cellular immune responses using flow cytometry analysis. Mouse sera and tissues were collected as described above, and detected for ZIKV titers and RNA copies using the plaque-forming assay and quantitative reverse transcriptase PCR (qRT-PCR), respectively.

qRT-PCR
This assay was performed to detect ZIKV RNA copies in sera and tissues of challenged mice as previously described [27,32]. Briefly, QIAamp MinElute Virus Spin Kit and RNeasy Mini Kit (Qiagen, Germantown, MD, USA) were used to extract RNAs from sera and tissues, respectively, which were quantified by qRT-PCR using Power SYBR Green PCR Master Mix, Ambion RNase Inhibitor, and MultiScribe Reverse Transcriptase (Thermo Fisher Scientific) in ViiA 7 Master Cycler PCR System (Thermo Fisher Scientific). The forward primer (5 -TTGGTCATGATACTGCTGATTGC-3 ) and reverse primer (5 -CCTTCCACAAAGTCCCTATTGC-3 ) were applied for amplification. A qRT-PCR standard curve was made via serial dilutions of a recombinant plasmid expressing the ZIKV membrane and E proteins, and a linear standard curve in the range of 10 1 -10 8 RNA copies (correlation coefficient: R 2 value > 0.98; detection limit: 10 1 RNA copies) was selected for calculation of ZIKV RNA copies in the test samples. Mouse sera (about 10 µL) and tissues, including eye, heart, spleen (about 20 mg), muscle, and testis (about 40 mg), were collected 3 days post-challenge, and detected for ZIKV RNA, so the detection limit was about 10 3 RNA copies/mL (for sera), 5 × 10 2 RNA copies/g (for eye, heart, and spleen), and 2.5 × 10 2 RNA copies/g (for muscle and testis) [32].

Statistical Analysis
All values were expressed as a mean with a standard error of the mean (s.e.m). Statistical significance of all data among different groups was calculated using one-way ANOVA: Tukey's multiple comparison test based on the GraphPad Prism Statistical Software. *, **, and *** indicate p < 0.05, p < 0.01, and p < 0.001, respectively.

3.1.
Comparison of the Adjuvanticity of Alum, MPL, MF59, and in Combination, for Immunization with ZIKV EDIII To compare the potency of adjuvants, BALB/c mice were immunized with a ZIKV non-neutralizing epitope-masked, EDIII subunit vaccine in the presence or absence of Alum, MPL, and Alum adjuvant in combination with the MPL adjuvant. In addition, MF59 was included as a control adjuvant. At seven days following the second dose, sera were collected and the antibody titers specific to ZIKV EDIII (containing a C-terminal hFc tag) or E (without fusion with hFc) were determined ( Figure 1). When used without an adjuvant, EDIII was immunogenic and induced EDIII-specific IgG antibodies (total IgG), while inclusion of Alum or MF59 adjuvant significantly improved the antibody response. Although MPL did not appear to enhance the antibody response to EDIII, when used in combination with Alum, the antibody response was significantly higher than for EDIII with Alum or MF59 alone ( Figure 2A). Since EDIII protein was fused with the hFc tag, IgG antibodies targeting hFc were also generated, showing a similar trend as those specific to EDIII ( Figure 2B). Nevertheless, when coating the ELISA plate with a ZIKV full-length E protein without hFc, significantly high-titer IgG antibodies were induced, particularly in the Alum and MPL-adjuvanted EDIII ( Figure 2C), suggesting that fusion of hFc to the EDIII subunit vaccine did not affect the generation of ZIKV-specific IgG antibodies. multiple comparison test based on the GraphPad Prism Statistical Software. *, **, and *** indicate p < 0.05, p < 0.01, and p < 0.001, respectively.

Comparison of the Adjuvanticity of Alum, MPL, MF59, and in Combination, for Immunization with ZIKV EDIII
To compare the potency of adjuvants, BALB/c mice were immunized with a ZIKV nonneutralizing epitope-masked, EDIII subunit vaccine in the presence or absence of Alum, MPL, and Alum adjuvant in combination with the MPL adjuvant. In addition, MF59 was included as a control adjuvant. At seven days following the second dose, sera were collected and the antibody titers specific to ZIKV EDIII (containing a C-terminal hFc tag) or E (without fusion with hFc) were determined ( Figure 1). When used without an adjuvant, EDIII was immunogenic and induced EDIIIspecific IgG antibodies (total IgG), while inclusion of Alum or MF59 adjuvant significantly improved the antibody response. Although MPL did not appear to enhance the antibody response to EDIII, when used in combination with Alum, the antibody response was significantly higher than for EDIII with Alum or MF59 alone ( Figure 2A). Since EDIII protein was fused with the hFc tag, IgG antibodies targeting hFc were also generated, showing a similar trend as those specific to EDIII ( Figure 2B). Nevertheless, when coating the ELISA plate with a ZIKV full-length E protein without hFc, significantly high-titer IgG antibodies were induced, particularly in the Alum and MPL-adjuvanted EDIII ( Figure 2C), suggesting that fusion of hFc to the EDIII subunit vaccine did not affect the generation of ZIKV-specific IgG antibodies. Immunization and challenge schedule. Mice were immunized intramuscularly (i.m.) with PBS (background control) or with ZIKV EDIII without an adjuvant, or with aluminum salts (Alum), monophosphoryl lipid A (MPL), Alum + MPL, or MF59 adjuvant and boosted at day 21 after the initial dose. Sera were collected at seven days after the boost (day 28) and used to determine ZIKV E or EDIII-specific antibody, and neutralizing antibody levels. Mice were injected with an anti-interferonα/β receptor 1 (Ifnar1) antibody (in order to become susceptible to subsequent ZIKV infection) [36,37] at nine days after the boost (day 30) and subsequently challenged with ZIKV (strain R103451, 6 × 10 5 plaque-forming unit (PFU)) the following day (day 31). Sera and tissue samples were collected three days post-challenge (day 34) and used to determine ZIKV titers and viral load (RNA copies). Splenocytes were also tested for the ZIKV EDIII-specific cellular immune response. Immunization and challenge schedule. Mice were immunized intramuscularly (i.m.) with PBS (background control) or with ZIKV EDIII without an adjuvant, or with aluminum salts (Alum), monophosphoryl lipid A (MPL), Alum + MPL, or MF59 adjuvant and boosted at day 21 after the initial dose. Sera were collected at seven days after the boost (day 28) and used to determine ZIKV E or EDIII-specific antibody, and neutralizing antibody levels. Mice were injected with an anti-interferon-α/β receptor 1 (Ifnar1) antibody (in order to become susceptible to subsequent ZIKV infection) [36,37] at nine days after the boost (day 30) and subsequently challenged with ZIKV (strain R103451, 6 × 10 5 plaque-forming unit (PFU)) the following day (day 31). Sera and tissue samples were collected three days post-challenge (day 34) and used to determine ZIKV titers and viral load (RNA copies). Splenocytes were also tested for the ZIKV EDIII-specific cellular immune response. combination of Alum with MPL adjuvant induced significantly higher IgG1 and IgG2a antibody titers than either Alum or MPL alone, or when MF59 was the adjuvant ( Figure 2D,E). When used in combination, MPL and Alum adjuvants induced a more balanced antibody response, indicated by a low IgG1 (Th2)/IgG2a (Th1) ratio. In contrast, Alum generated a Th2-biased IgG1 antibody response, while EDIII alone, or when formulated with MF59, showed a slight Th2 bias ( Figure 2F). The above data demonstrate that the combination of Alum with MPL adjuvant enhanced the levels of EDIIIspecific IgG antibodies (IgG1 and IgG2a) and produced a more balanced IgG response.

The Effect of Alum in Combination with MPL Adjuvant on Anti-ZIKV Neutralizing Antibody Titers
Next, we compared the levels of anti-ZIKV neutralizing antibodies induced by EDIII vaccine when formulated with the different adjuvants. To this end, sera collected at seven days after immunization with the second dose of EDIII were tested for neutralization against three human ZIKV We also determined the levels of the IgG1 and IgG2a antibodies induced by EDIII at seven days after the second dose. EDIII alone induced a good IgG1 response while the IgG2a response was noticeably weaker ( Figure 2D,E). When used alone, Alum induced higher levels of IgG1 (but not IgG2a) antibodies than EDIII without an adjuvant, while MPL induced significantly higher levels of IgG2a (but not IgG1) antibodies. When used in combination, Alum with MPL significantly improved the ability of EDIII to induce the IgG1 and IgG2a antibody responses ( Figure 2D,E). Moreover, the combination of Alum with MPL adjuvant induced significantly higher IgG1 and IgG2a antibody titers than either Alum or MPL alone, or when MF59 was the adjuvant (Figure 2D,E). When used in combination, MPL and Alum adjuvants induced a more balanced antibody response, indicated by a low IgG1 (Th2)/IgG2a (Th1) ratio. In contrast, Alum generated a Th2-biased IgG1 antibody response, while EDIII alone, or when formulated with MF59, showed a slight Th2 bias ( Figure 2F). The above data demonstrate that the combination of Alum with MPL adjuvant enhanced the levels of EDIII-specific IgG antibodies (IgG1 and IgG2a) and produced a more balanced IgG response.

The Effect of Alum in Combination with MPL Adjuvant on Anti-ZIKV Neutralizing Antibody Titers
Next, we compared the levels of anti-ZIKV neutralizing antibodies induced by EDIII vaccine when formulated with the different adjuvants. To this end, sera collected at seven days after immunization with the second dose of EDIII were tested for neutralization against three human ZIKV strains (R103451, Vaccines 2019, 7, 161 7 of 16 PAN2016, and PAVABC59). For all three strains, Alum and MPL as the adjuvants induced a significantly stronger neutralizing antibody response than EDIII alone, while MPL or MF59 only enhanced the neutralizing antibody response for the R103451 strain ( Figure 3A-C). Importantly, for all three ZIKV strains, the neutralizing antibody titers induced by EDIII were significantly higher when Alum was used in combination with MPL than for Alum, MPL, or MF59 as the sole adjuvant ( Figure 3A-C). In contrast, sera injected with PBS only induced a background-level neutralizing antibody ( Figure 3A-C). These data show that although single adjuvants increased the anti-ZIKV neutralizing antibody response, the combination of Alum with MPL adjuvant consistently induced the highest level of neutralizing antibodies against all of the ZIKV strains investigated.
Vaccines 2019, 7, x 7 of 15 strains (R103451, PAN2016, and PAVABC59). For all three strains, Alum and MPL as the adjuvants induced a significantly stronger neutralizing antibody response than EDIII alone, while MPL or MF59 only enhanced the neutralizing antibody response for the R103451 strain ( Figure 3A-C). Importantly, for all three ZIKV strains, the neutralizing antibody titers induced by EDIII were significantly higher when Alum was used in combination with MPL than for Alum, MPL, or MF59 as the sole adjuvant ( Figure 3A-C). In contrast, sera injected with PBS only induced a background-level neutralizing antibody ( Figure 3A-C). These data show that although single adjuvants increased the anti-ZIKV neutralizing antibody response, the combination of Alum with MPL adjuvant consistently induced the highest level of neutralizing antibodies against all of the ZIKV strains investigated.

The Effect of Alum in Combination with MPL Adjuvant on the ZIKV EDIII-Specific Cellular Immune Response
Splenocytes isolated from immunized mice three days after ZIKV challenge (strain R103451) were used to investigate if the adjuvants enhanced the ZIKV EDIII-specific cellular immune response. Specifically, we determined the levels of secretion of IFN-γ (Th1), IL-4 (Th2) and IL-17 (Th17) by CD45-positive (CD45 + )-CD4 + T cells, and IFN-γ and IL-4 by CD45 + -CD8 + T cells. For both T cell subsets, only low levels of these cytokines were detected when mice were immunized with EDIII alone ( Figure 4A-E), suggesting that EDIII without an adjuvant did not induce a strong EDIII-specific T cell response. When used individually, Alum or MF59 increased the secretion of IL-4 by CD45 + -CD4 + T cells, while increased expression of IL-17 was only seen for MPL ( Figure 4B,C). For CD45 + -CD8 + T cells, inclusion of Alum significantly increased IFN-γ expression, while IL-4 expression was increased by Alum and MF59 ( Figure 4D,E). When used in combination, Alum with MPL adjuvant significantly enhanced expression of IFN-γ and IL-4 cytokines tested in both CD45 + -CD4 + and CD45 + -CD8 + T cells over that observed when the adjuvants (Alum, MPL, and/or MF59) were used individually ( Figure 4A,B,D,E). These data indicate that although individual adjuvants enhanced the EDIII-specific cellular immune response, the combination of Alum with MPL had a greater effect, particularly in CD45 + -CD8 + T cells. Figure 3. Detection of neutralizing antibody titers in immunized mouse sera. Sera collected at seven days following the second dose of EDIII were tested by a plaque reduction neutralization test (PRNT) assay for neutralizing antibodies against ZIKV human strains R103451 (A), PAN2016 (B), and PRVABC59 (C). PBS shows the background level of neutralizing antibodies. Neutralizing activity for the indicated adjuvants is presented as the neutralizing antibody titer based on the serum neutralizations at a 50% plaque reduction (PRNT 50 ). Data are expressed as the mean ± s.e.m (n = 5). Significant differences (*: p < 0.05; **: p < 0.01; ***: p < 0.001) among different groups are shown.

The Effect of Alum in Combination with MPL Adjuvant on the ZIKV EDIII-Specific Cellular Immune Response
Splenocytes isolated from immunized mice three days after ZIKV challenge (strain R103451) were used to investigate if the adjuvants enhanced the ZIKV EDIII-specific cellular immune response. Specifically, we determined the levels of secretion of IFN-γ (Th1), IL-4 (Th2) and IL-17 (Th17) by CD45-positive (CD45 + )-CD4 + T cells, and IFN-γ and IL-4 by CD45 + -CD8 + T cells. For both T cell subsets, only low levels of these cytokines were detected when mice were immunized with EDIII alone ( Figure 4A-E), suggesting that EDIII without an adjuvant did not induce a strong EDIII-specific T cell response. When used individually, Alum or MF59 increased the secretion of IL-4 by CD45 + -CD4 + T cells, while increased expression of IL-17 was only seen for MPL ( Figure 4B,C). For CD45 + -CD8 + T cells, inclusion of Alum significantly increased IFN-γ expression, while IL-4 expression was increased by Alum and MF59 ( Figure 4D,E). When used in combination, Alum with MPL adjuvant significantly enhanced expression of IFN-γ and IL-4 cytokines tested in both CD45 + -CD4 + and CD45 + -CD8 + T cells over that observed when the adjuvants (Alum, MPL, and/or MF59) were used individually ( Figure 4A,B,D,E). These data indicate that although individual adjuvants enhanced the EDIII-specific cellular immune response, the combination of Alum with MPL had a greater effect, particularly in CD45 + -CD8 + T cells.

The Effect of Alum in Combination with MPL Adjuvant on Protective Efficacy of ZIKV EDIII Vaccine, Including Reproductive Organs
To investigate if the adjuvants increased protection against ZIKV infection, immunized mice were challenged with a high-dose ZIKV (strain R103451, 6 × 10 5 PFU/mouse) one day after administering an anti-Ifnar1 antibody (Figure 1). Viral load was determined by a plaque-forming assay and a ZIKV RNA copy number determined by qRT-PCR assay using sera and tissues collected three days post-challenge, an optimal time point to detect ZIKV infection in anti-Ifnar1 antibodytreated immunocompetent mice, including BALB/c (our unpublished data) [37]. Similar levels of ZIKV were seen in sera and most tissues (eye, heart, spleen, and muscle) for mice immunized with EDIII without an adjuvant or with PBS ( Figure 5A-E), whereas only in brain tissue, a significant reduction in viral titers was observed for mice immunized with EDIII alone ( Figure 5F). These results suggest that immunization with EDIII is unable to elicit adequate protection against ZIKV challenge. Mice immunized with EDIII formulated with a single adjuvant (Alum, MPL, and/or MF59) had significantly reduced viral titers compared to those immunized with EDIII alone (sera, heart, spleen, muscle, and brain) ( Figure 5A,C-F) and those immunized with PBS (sera, eye, heart, spleen, muscle, and brain) ( Figure 5A-F), suggesting that inclusion of an adjuvant increased protection. Importantly, in sera and all of the tissues tested, the combination of Alum with MPL adjuvant resulted in significantly lower, or undetectable, viral titers than other immunization conditions with a single adjuvant (Alum, MPL, and/or MF59) ( Figure 5A-F). These results demonstrate that, although single adjuvants can increase protection, the greatest level of protection against a high-dose ZIKV challenge is achieved when EDIII vaccine is formulated with both Alum and MPL adjuvants.

The Effect of Alum in Combination with MPL Adjuvant on Protective Efficacy of ZIKV EDIII Vaccine, Including Reproductive Organs
To investigate if the adjuvants increased protection against ZIKV infection, immunized mice were challenged with a high-dose ZIKV (strain R103451, 6 × 10 5 PFU/mouse) one day after administering an anti-Ifnar1 antibody (Figure 1). Viral load was determined by a plaque-forming assay and a ZIKV RNA copy number determined by qRT-PCR assay using sera and tissues collected three days post-challenge, an optimal time point to detect ZIKV infection in anti-Ifnar1 antibody-treated immunocompetent mice, including BALB/c (our unpublished data) [37]. Similar levels of ZIKV were seen in sera and most tissues (eye, heart, spleen, and muscle) for mice immunized with EDIII without an adjuvant or with PBS ( Figure 5A-E), whereas only in brain tissue, a significant reduction in viral titers was observed for mice immunized with EDIII alone ( Figure 5F). These results suggest that immunization with EDIII is unable to elicit adequate protection against ZIKV challenge. Mice immunized with EDIII formulated with a single adjuvant (Alum, MPL, and/or MF59) had significantly reduced viral titers compared to those immunized with EDIII alone (sera, heart, spleen, muscle, and brain) ( Figure 5A,C-F) and those immunized with PBS (sera, eye, heart, spleen, muscle, and brain) ( Figure 5A-F), suggesting that inclusion of an adjuvant increased protection. Importantly, in sera and all of the tissues tested, the combination of Alum with MPL adjuvant resulted in significantly lower, or undetectable, viral titers than other immunization conditions with a single adjuvant (Alum, MPL, and/or MF59) ( Figure 5A-F). These results demonstrate that, although single adjuvants can increase protection, the greatest level of protection against a high-dose ZIKV challenge is achieved when EDIII vaccine is formulated with both Alum and MPL adjuvants. , and brain (F). Sera and tissues were collected at three days post-challenge with ZIKV (strain R103451), and viral load determined by a plaque-forming assay. Mice immunized with PBS served as the control. The detection limits for the samples were ~20 PFU/mL (sera), ~50 PFU/g (eye), and ~25 PFU/g (heart, spleen, muscle, and brain). Data are expressed as the mean ± s.e.m (n = 5). Significant differences (*: p < 0.05; **: p < 0.01; ***: p < 0.001) among different groups are shown.
Consistent with viral load, mice immunized with EDIII alone had only slightly lower levels of ZIKV RNA in sera and tissues (eye and spleen) compared to those immunized with PBS ( Figure  6A,B,D), confirming that an adjuvant is required to induce protection. Similarly, inclusion of a single adjuvant (Alum, MPL, and/or MF59) resulted in a significantly lower level of ZIKV RNA in sera and tissues (eye, heart, spleen, or muscle) compared with immunization with EDIII alone (Figure 6A-E). Furthermore, the levels of ZIKV RNA in sera and all of the tissues, when Alum and MPL adjuvants were used in combination for immunization, were significantly lower than when a single adjuvant (Alum, MPL, and/or MF59) was used ( Figure 6A-E). These results are consistent with the levels of viral load ( Figure 5) and confirm that protection against ZIKV is greatest when Alum and MPL adjuvants are used in combination.
ZIKV can replicate in the reproductive organs for long periods; therefore, we also determined the amount of ZIKV RNA in testes of immunized and challenged mice. Similar to the results in other tissues, inclusion of a single adjuvant reduced the level of viral RNA more than immunization with EDIII alone or with PBS ( Figure 6F). Moreover, viral RNA copies were the lowest in testis of mice receiving EDIII formulated with Alum and MPL in combination ( Figure 6F). Importantly, our results show that formulation of EDIII with a combination of Alum and MPL adjuvants can reduce viral RNA copies in male reproductive organs. Figure 5. Detection of ZIKV titers in challenged sera (A), eye (B), heart (C), spleen (D), muscle (E), and brain (F). Sera and tissues were collected at three days post-challenge with ZIKV (strain R103451), and viral load determined by a plaque-forming assay. Mice immunized with PBS served as the control. The detection limits for the samples were~20 PFU/mL (sera),~50 PFU/g (eye), and~25 PFU/g (heart, spleen, muscle, and brain). Data are expressed as the mean ± s.e.m (n = 5). Significant differences (*: p < 0.05; **: p < 0.01; ***: p < 0.001) among different groups are shown.
Consistent with viral load, mice immunized with EDIII alone had only slightly lower levels of ZIKV RNA in sera and tissues (eye and spleen) compared to those immunized with PBS ( Figure 6A,B,D), confirming that an adjuvant is required to induce protection. Similarly, inclusion of a single adjuvant (Alum, MPL, and/or MF59) resulted in a significantly lower level of ZIKV RNA in sera and tissues (eye, heart, spleen, or muscle) compared with immunization with EDIII alone (Figure 6A-E). Furthermore, the levels of ZIKV RNA in sera and all of the tissues, when Alum and MPL adjuvants were used in combination for immunization, were significantly lower than when a single adjuvant (Alum, MPL, and/or MF59) was used ( Figure 6A-E). These results are consistent with the levels of viral load ( Figure 5) and confirm that protection against ZIKV is greatest when Alum and MPL adjuvants are used in combination. Figure 6. Detection of ZIKV RNA in challenged sera (A), eye (B), heart (C), spleen (D), muscle (E), and testis (F). Sera and tissues were collected at three days post-challenge with ZIKV (strain R103451) and viral RNA copies determined by qRT-PCR. Mice immunized with PBS served as the control. The detection limits were ~10 3 RNA copies/mL (sera), ~5 × 10 2 RNA copies/g (eye, heart, and spleen), and ~2.5 × 10 2 RNA copies/g (muscle and testis). Data are expressed as the mean ± s.e.m (n = 5). Significant differences (*: p < 0.05; **: p < 0.01; ***: p < 0.001) among different groups are shown.

Discussion
Protein-based subunit vaccines are generally safer than vaccines based on live attenuated or inactivated viruses but usually have low immunogenicity and efficacy [20,21]. However, vaccine performance can be improved by structure-based designs via masking non-neutralizing or unfavorable epitopes, or via antigen conjugation [27,38,39]. In addition, the choice of adjuvant can greatly influence the immunogenicity and efficacy of vaccines, including subunit vaccines [29,40,41].
Our previous studies demonstrated that masking of a non-neutralizing epitope (surrounding residue 375) of a ZIKV DIII of E protein subunit vaccine (i.e., EDIII) improved immunogenicity and protection of adult and pregnant mice, and unborn offspring, against ZIKV infection [27]. Here, we extended these studies to investigate the influence of licensed adjuvants on the immunogenicity and protection of mice against high-dose ZIKV challenge following immunization with this EDIII subunit vaccine. Vaccines prepared with a single adjuvant (Alum, MPL, or MF59) significantly improved the antibody response (total IgG, IgG1, and/or IgG2a) over immunization without an adjuvant. However, the highest IgG titers were obtained when Alum and MPL were used in combination.
ZIKV can replicate in the reproductive organs for long periods; therefore, we also determined the amount of ZIKV RNA in testes of immunized and challenged mice. Similar to the results in other tissues, inclusion of a single adjuvant reduced the level of viral RNA more than immunization with EDIII alone or with PBS ( Figure 6F). Moreover, viral RNA copies were the lowest in testis of mice receiving EDIII formulated with Alum and MPL in combination ( Figure 6F). Importantly, our results show that formulation of EDIII with a combination of Alum and MPL adjuvants can reduce viral RNA copies in male reproductive organs.

Discussion
Protein-based subunit vaccines are generally safer than vaccines based on live attenuated or inactivated viruses but usually have low immunogenicity and efficacy [20,21]. However, vaccine performance can be improved by structure-based designs via masking non-neutralizing or unfavorable epitopes, or via antigen conjugation [27,38,39]. In addition, the choice of adjuvant can greatly influence the immunogenicity and efficacy of vaccines, including subunit vaccines [29,40,41].
Our previous studies demonstrated that masking of a non-neutralizing epitope (surrounding residue 375) of a ZIKV DIII of E protein subunit vaccine (i.e., EDIII) improved immunogenicity and protection of adult and pregnant mice, and unborn offspring, against ZIKV infection [27]. Here, we extended these studies to investigate the influence of licensed adjuvants on the immunogenicity and protection of mice against high-dose ZIKV challenge following immunization with this EDIII subunit vaccine. Vaccines prepared with a single adjuvant (Alum, MPL, or MF59) significantly improved the antibody response (total IgG, IgG1, and/or IgG2a) over immunization without an adjuvant. However, the highest IgG titers were obtained when Alum and MPL were used in combination.
As stated above, adjuvants influence the immune response to vaccines. Alum is a potent inducer of the Th2 (IgG1) immune response [23,42]. Similarly, MF59 generally induces a slightly-biased Th2 response as seen for vaccines for influenza virus and MERS-CoV [29,43]. In contrast, MPL induces expression of IFN-γ and IL-2 and, therefore, has a more Th1 bias and generates a more balanced (Th1/Th2) immune response [22,25]. Moreover, a combination of MPL and Alum adjuvants (i.e., AS04) may shift the immune responses from Th2 to Th1 and generate a more balanced immune response [44]. Consistent with these observations, we found that MF59 and Alum, when used individually as the adjuvant for EDIII, induced a Th2-biased immune response, while, when MPL was the sole adjuvant or used in combination with Alum, a more balanced (Th1/Th2) immune response was generated.
In the current study, we observed that Alum, MPL, and MF59 adjuvants enhanced production of anti-ZIKV neutralizing antibodies; however, the combination of Alum with MPL generated significantly higher neutralizing antibody titers than the individual use of adjuvants, and these results were consistent for three emerging human ZIKV strains. In addition, Alum and MPL, when used in combination, enhanced secretion of IFN-γ (Th1) and IL-4 (Th2) by CD4 + , particularly CD8 + , T cells. Viral vaccines (e.g., against HBV and HPV) formulated with a combination of Alum and MPL adjuvants (i.e., AS04) stimulate antigen-specific T and B cells, resulting in stronger antibody (including neutralizing antibodies) and cellular responses than vaccines formulated using Alum alone [22,44]. Using the ZIKV EDIII protein as a model antigen, our study has confirmed these observations by showing that a combination of Alum and MPL adjuvants enhanced the immunogenicity of EDIII over that observed with single adjuvants (Figures 2-6), whereas the Alum and MPL combination alone without EDIII was unable to elicit ZIKV-specific immune response and protection against ZIKV infection [26,27].
ZIKV does not normally infect immunocompetent mice, including BALB/c and C57BL/6 [36,53], but does infect immunocompromised mice deficient for interferon receptors (interferon-α/β receptor (Ifnar) or A129, or interferon-α/β/γ receptor, AG129) [36,53,54]. To carry out challenge studies, we used BALB/c mice treated with anti-Ifnar1 antibodies, which also render mice susceptible to ZIKV infection [36,37]. Formulation of EDIII with individual adjuvants improved protection against high-dose ZIKV challenge, and was consistent with our virus neutralization studies results, that is, the combination of Alum with MPL significantly improved protection further. This was confirmed by showing that a combination of Alum with MPL as the adjuvant significantly reduced viral titers and viral RNA copies in sera and tissues compared to other immunization regimes. ZIKV can be transmitted through sexual contact, the virus can persist in female or male (testes and seminal vesicles) reproductive organs for a prolonged period, and viral RNA can be detected in semen for up to one year after ZIKV infection [53,[55][56][57][58][59]. Thus, vaccines that reduce viral load in reproductive organs will likely be more effective in reducing virus transmission during sexual contact. Importantly, our results show that EDIII formulation using Alum and MPL in combination significantly reduced viral RNA in the testes of ZIKV-challenged mice, thereby confirming that this adjuvant combination is most likely to yield greater vaccine efficacy than single adjuvants alone. In contrast, EDIII alone without adjuvant(s) resulted in the highest viral titers and viral RNA copies in sera and tissues tested among all vaccination groups, including reproductive organs, which were only slightly lower than those in PBS-injected control mice. It has been shown that protection against ZIKV infection is positively correlated with serum neutralizing antibody titers [26,27]; thus, the high-level ZIKV in the EDIII alone-immunized mice identified above may be partially due to the relatively low serum neutralizing antibody titer induced in these mice.
It should be noted that the EDIII protein used in this study was fused with a C-terminal hFc tag. Fc-fusion proteins have been approved by the FDA for human therapy, and their safety and pharmacokinetic activity have been evaluated extensively [60][61][62][63]. Fusion of Fc tag to recombinant proteins may promote the proteins to form dimeric structures, significantly improving their solubility, stability, and avidity, as well as increasing their immunogenicity and/or efficacy [64][65][66]. Fc-fused vaccine design has been applied in subunit vaccines against HIV, influenza virus, Ebola virus, and other emerging or reemerging viruses [66][67][68][69][70]. In terms of ZIKV vaccines, a Fc-fused E protein dimer (ZE-Fc) has been shown to mimic native dimeric status of E protein and induce higher neutralizing antibodies than a monomeric E protein without Fc (monZE) [71]. We have also found that hFc-fused ZIKV EDIII proteins form dimeric structures capable of eliciting high-titer antibody responses, neutralizing antibodies, and protection against high-dose ZIKV infection (Figures 2, 3, 5 and 6) [27], consistent with previous findings.
In conclusion, using immunocompetent BALB/c mice, we compared the immunogenicity, neutralizing activity, and protective efficacy (after injection with anti-Ifnar1 antibodies) of a ZIKV EDIII subunit vaccine when formulated with different adjuvants (individually or in combination). Although EDIII formulated with Alum, MPL, or MF59 induced a specific antibody and cellular immune response, and reduced viral load, the greatest vaccine efficacy was achieved when Alum and MPL adjuvants were combined in a single vaccine. Overall, this study has identified Alum combined with MPL as the adjuvant of choice for ZIKV subunit vaccines, and has important implications for subunit vaccines against other enveloped viruses, including non-ZIKV flaviviruses.
Author Contributions: L.D. and C.S. initiated and designed the study. X.W., W.T., and X.Z. performed the experiments and analyzed the data. Y.Z., L.D., and C.S. wrote, edited, and approved the manuscript.
Funding: This research was funded by the National Institutes of Health (NIH) grant (R21AI137790) and intramural funds of the New York Blood Center (NYB616). ZIKV strains (R103451, PAN2016, and PAVABC59) were obtained from BEI Resources.

Conflicts of Interest:
The authors declare no conflict of interest.