A Tetravalent Sub-unit Dengue Vaccine Formulated with Ionizable Cationic Lipid Nanoparticle induces Significant Immune Responses in Rodents and Non-Human Primates

Dengue virus has emerged as an important arboviral infection worldwide. As a complex pathogen, with four distinct serotypes, the development of a successful Dengue virus vaccine has proven to be challenging. Here, we describe a novel Dengue vaccine candidate that contains truncated, recombinant, Dengue virus envelope protein from all four Dengue virus serotypes (DEN-80E) formulated with ionizable cationic lipid nanoparticles (LNPs). Immunization studies in mice, Guinea pigs, and in Rhesus macaques, revealed that LNPs induced high titers of Dengue virus neutralizing antibodies, with or without co-administration or encapsulation of a Toll-Like Receptor 9 agonist. Importantly, LNPs were also able to boost DEN-80E specific CD4+ and CD8+ T cell responses. Cytokine and chemokine profiling revealed that LNPs induced strong chemokine responses without significant induction of inflammatory cytokines. In addition to being highly efficacious, the vaccine formulation proved to be well-tolerated, demonstrating no elevation in any of the safety parameters evaluated. Notably, reduction in cationic lipid content of the nanoparticle dramatically reduced the LNP’s ability to boost DEN-80E specific immune responses, highlighting the crucial role for the charge of the LNP. Overall, our novel studies, across multiple species, reveal a promising tetravalent Dengue virus sub-unit vaccine candidate.


Results
Lipid Nanoparticles enhance Dengue-2 envelope protein specific antibody responses in mice that possess virus neutralizing activity. We evaluated the ability of Merck LNPs to boost immune responses to a recombinant, truncated envelope protein from Dengue virus (DENV) serotype 2 (DEN2-80E) In previously reported preclinical studies evaluating DEN-80E antigens, formulations containing ISCOMATRIX TM adjuvant were found to result in the strongest DENV neutralizing antibody titers in both mice and non-human primates, and therefore it was selected as the comparator in the current studies 13,14 DENV's envelope specific antibodies that possess neutralizing activity are currently the main focus of Dengue vaccine clinical trials 26 ISCOMATRIX TM adjuvant contains ISCOPREP TM saponin, a purified fraction of Quillaia saponin, cholesterol and phospholipid 27 . Published studies suggest that the ISCOMATRIX TM adjuvant results in the induction of both Th1 and Th2 type responses 27 . The mechanism of action of ISCOMATRIX TM adjuvant is reported to involve a MyD88 dependent pathway for induction of CD8+ T cell responses 28 , as well, as inflammasome dependent and independent mechanisms for its in-vivo functionality 29 .
For murine studies, mice (n = 10 /group) received, 3 doses (one prime and two boosts) of DEN2-80E (1 μ g) formulated with PBS or the specified adjuvant, at 2-week intervals, (Fig. 1a). DEN2-80E antigen was not encapsulated within the LNP, but merely co-mixed prior to intramuscular injection. The LNPs were evaluated both alone and in combination with TLR-ODN. 2 weeks after the final dose, DEN2-80E-specific total serum IgG antibody titers were determined by end-point dilution ELISA. As expected, a statistically significant (p < 0.001) increase in DEN2-80E specific total IgG responses was observed in mice that received ISCOMATRIX TM adjuvant compared to the PBS control group (Fig. 1b). In contrast, a significant increase in titers was not seen in mice that received TLR9-ODN (2 μ g and 10 μ g). When TLR9-ODN was co-administered with or encapsulated within LNP, a dose depended increase in total IgG titers was observed with the two highest dose groups of TLR9-ODN (2 μ g and 10 μ g) formulated with LNP (25 μ g and 125 μ g) demonstrating statistical significance (p < 0.01) over the PBS group. Notably, mice that received 25 μ g or 125 μ g of LNP alone, showed statistically significant (p < 0.01) increases in DEN2-80E specific IgG response, suggesting these responses could be driven by LNPs themselves. Further isotyping of IgG antibodies was carried-out by assaying DEN2-80E specific IgG1 (Th2 biased) and IgG2a (Th1 biased). The isotope profile (IgG2a:IgG1 ratio) of antigen specific IgGs has been established as a marker to assess Th1 vs Th2 type immune response bias. Inspite of comparable total IgG titers, co-administration or encapsulation of TLR9-ODN (2 μ g and 10 μ g) with LNP (25 μ g and 125 μ g) led to a more Th1 type antibody response. In contrast, the LNP only immunized mice showed a Th2 type antibody profile (Fig. 1c). In addition to evaluating the total quality and quantity of DEN2-80E specific IgG responses in immunized mice, we evaluated the functional antibody response using a LiCor based DENV neutralization assay (as described in materials & methods). In this assay, sera from immunized mice were pre-incubated at various dilutions with DENV serotype 2 (DEN2) before adding the mixture to vero cells and calculated the neutralization titers (NT50). As expected, the group that received ISCOMATRIX TM adjuvant yielded statistically significant levels of DEN 2 virus neutralization titers (Fig. 1d). Notably, LNP by itself at 25 μ g and 125 μ g, or TLR-ODN (2 μ g and 10 μ g) co-administered or encapsulated with LNP (25 μ g and 125 μ g), also resulted in statistically significant DEN2 virus neutralization titers (nAb). Overall, these data strongly suggest that LNP adjuvant by itself was able to induce strong DEN2-80E specific antibody responses that were functionally effective at neutralizing Dengue Virus 2 infection. The immunization regime and dosing schedule for BALB/c mice vaccinated with 1 μ g of DEN2-80E antigen, formulated with various adjuvants are depicted (a,b). DEN2-80E antigen was not encapsulated within LNP, but, merely co-mixed and administered intramuscularly as specified. Total IgG responses to DEN2-80E determined by ELISA in DEN2-80E vaccinated mice with LNP by itself, LNP co-administered with, or encapsulating TLR9-ODN (c). Ratio of DEN2-80E specific IgG2a to DEN2-80E specific IgG1 titers determined from pooled mice sera from each group (d). Dengue virus serotype 2 specific neutralizing activity in individual immunized mouse sera was calculated using the LiCor based MicroNeut assay (e). All assays were performed in sera collected at 2 weeks post-dose 3, in duplicates, from 10 mice/group. *p < 0.01 and **p < 0.001 denote statistical significance compared to PBS group and #p < 0.05 denotes statistical significance compared to TLR9-ODN (only) group.
Scientific RepoRts | 6:34215 | DOI: 10.1038/srep34215 DEN2 env specific CD4+ and CD8+ T cell responses are enhanced by Lipid Nanoparticles. Cell mediated immune responses were first evaluated by employing an IFN-γ ELISPOT assay. 2 weeks after the final vaccine dose (Fig. 1a), spleens from 5 randomly selected mice from each vaccination group were processed, and pooled splenocytes were incubated with overlapping peptides that cover the amino acid sequence of the DEN2-80E antigen. As compared to the un-adjuvanted DEN2-80E + PBS group, the number of IFN-γ + spot forming units (SFUs), where significantly higher in the DEN2-80E + ISCOMATRIX TM adjuvant group. As seen with the nAb titers, vaccination with TLR9-ODN + DEN2-80E vaccination did not result in a considerable increase in IFN-γ + SFUs. However, co-administration of DEN2-80E with LNP (in the presence or absence of TLR9-ODN) dramatically enhanced the IFN-γ + SFU formation, which demonstrated the LNP's ability to enhance cell mediated immune responses to DEN2-80E antigen (Fig. 2a). Since ELISPOT assays do not distinguish between antigen specific CD4+ T cell or CD8+ T cell responses, we evaluated cell-mediated immune responses by flow cytometry intracellular staining (ICS) in pooled spleens from each group of vaccinated animals. The majority of the groups had undetectable levels of DEN2-80E specific T cell responses: mice vaccinated with DEN2-80E + PBS, DEN2-80E + TLR9-ODN, and the mice vaccinated with DEN2-80E + lower doses of LNP, LNP + TLR9-ODN and TLR9-ODN encapsulated in LNP. Vaccination with DEN2-80E + ISCOMATRIX TM adjuvant resulted in minimal induction of antigen specific CD4+ and CD8+ T cells. However, vaccination with DEN-80E formulated with TLR9-ODN within LNP, LNP + TLR9-ODN, and LNP by itself (at the highest doses), yielded a substantial increase in DEN2-80E specific CD4+ T cells that express IFN-γ and TNF-α (Fig. 2b) and CD8+ T cell responses that express IFN-γ and TNF-α (Fig. 2c). The most significant frequency of DEN2-80E specific CD4 and CD8+ T cells were observed in the group that received the highest dose of LNP + TLR9-ODN + DEN2-80E. While LNP + DEN2-80E vaccination by itself was able to induce T cell responses, co-administration of TLR9-ODN significantly enhanced the frequency of both CD4+ and CD8+ T cell responses specific to DEN2-80E.

Evaluation of cytokine and chemokine profiles in vaccinated mice.
Having established the potency of DEN2-80E specific immune responses when combined with various adjuvants, we evaluated cytokine and chemokine profiles in vaccinated mice. Using a multiplex assay platform, we evaluated 16 cytokines and chemokines in pooled immunized mouse sera from each group, at two time points, 4 hrs and 24 hrs post intramuscular administration of adjuvant by themselves and in combination with DEN2-80E antigen. Shown in Fig. 3 is a heat map of fold change for each cytokine and chemokine as compared to PBS treated mice at each time point assayed. Mice injected with TLR9-ODN, TLR9-ODN + LNP, TLR9-ODN in LNP, and ISCOMATRIX TM adjuvant by itself or in combination with DEN2-80E antigen, resulted in elevation of pro-inflammatory cytokines including IL-6 and TNF-α , and the Th1-type cytokine, IL-12. The inflammasome specific cytokine L-1β was increased in mice injected with DEN2-80E alone or in combination with other adjuvants, at both 4 hrs and 24 hrs post injection. This supports the notion that the Dengue envelope protein might be contributing to the inflammasome activation, previously reported upon in-vitro infection and in Dengue virus infected patients 30,31 . LNP alone, or in combination with DEN2-80E, did not induce significant levels of IL-1β and TNF-α , but a minimal increase in IL-6 was seen at 4 hrs and 24 hrs post injection. Other chemokines such as (i) Eotaxin that attracts eosinophils and neutrophils, and (ii) Interleukin-8 (IL-8), a chemokine that significantly attracts neutrophils and other innate immune cells, were significantly upregulated in sera of all mice injected with all adjuvants + /− antigen. However, macrophage specific chemokines such Macrophage Inflammatory Protein-1 (MIP-1a), Macrophage Derived Chemokine (MDC), and monocyte specific, Monocyte Chemotactic Protein-1 (MCP-1) were significantly increased in all LNP containing immunization groups.
Reduction in ionizable cationic lipid content impacts LNP mediated increase in D2-80E specific responses in immunized mice. As described in our previous publications 32, 33 , the lipid nanoparticle is composed of four main components: Merck proprietary ionizable cationic lipids, 1,2-distearoyl-glycero-3-p hosphocholine (DSPC) phospholipid, cholesterol, and a PEGylated lipid, PEG (2000)-DMG. To specifically evaluate if reduction in the charge of the lipid nanoparticle would impact its ability to boost antigen specific immune responses, we manufactured LNPs, termed Reduced Cationic Lipid LNPs, (RCLs). RCLs have the same components as the original LNPs, but the content of ionizable cationic lipid is significantly reduced. We immunized BALB/c mice (n = 5) (similar vaccination regime as described in Fig. 1a) using DEN2-80E antigen formulated with three concentrations of LNP or RCL (25 μ g, 75 μ g and 125 μ g) as adjuvants. As shown in Fig. 4a, a dose dependent increase in DEN2-80E specific IgG responses was observed in the LNP + DEN2-80E vaccinated groups, as compared to the DEN2-80E + PBS vaccinated group. While the 25 μ g dose of LNP was insufficient, the 75 μ g and 125 μ g doses of LNP were sufficient to induce statistically significant increases in DEN2-80E specific total IgG responses. However, vaccination with DEN2-80E and RCL (25 μ g, 75 μ g and 125 μ g) did not yield a statistically significant increase in DEN2-80E specific IgG titers (Fig. 4a). We next evaluated T cell responses by ICS in pooled spleens from all 5 mice per group as described in section 1.2. Evaluation of DEN2-80E specific CD4+ T cell response (Fig. 4b) and CD8+ T cell responses (Fig. 4c) revealed that LNP + DEN2-80E resulted in a dose dependent increase in detectable DEN2-80E specific CD4+ T cells and CD8+ T cells that expressed IFN-γ and TNF-α . In comparison, groups that received DEN2-80E + RCL had a lower frequency of DEN2-80E specific CD4+ T cell and CD8+ T cell responses. Overall, these data suggest that a reduction in the cationicity of the LNPs substantially reduced the LNP's ability to boost B-cell and T-cell responses to DEN2-80E in vaccinated mice.

LNP boosts neutralizing antibody responses against all four Dengue virus serotypes in Guinea pigs.
As previously mentioned, a successful Dengue vaccine should elicit neutralizing antibodies (nAbs) against all four serotypes of the virus. Therefore, building on our murine model using monovalent DEN2-80E + LNP, we evaluated if LNPs would be able to boost nAbs against all four DENVs in Guinea Pigs immunized with a tetravalent Figure 2. T cell responses to DEN2-80E in mice. 2 weeks after final vaccine dose, spleens from immunized mice were collected, processed, and total DEN2-80E specific T cell response was evaluated. IFN-γ producing spots per million (SFU) splenocytes were determined by IFN-γ ELISPOT assay. Bars represent average value from pooled spleens from 5 randomly selected mice out of 10 immunized mice/group, after background subtraction from non-stimulated cells (a). DEN2-80E specific CD4+ T cells that express IFN-γ , and TNF-α (b), and DEN2-80E specific CD8+ T cells that express IFN-γ , and TNF-α (c) were evaluated by intracellular FACS assay (FACS ICS). Bars represent T-cell events per million splenocytes from 5 randomly selected mice out of 10 immunized mice/group after background subtraction from unstimulated cells.
DENV-80E (DEN1-4-80E) vaccine. We utilized a pre-optimized dose of 3 μ g, 3 μ g, 3 μ g, and 6 μ g of DEN1-80E, DEN2-80E, DEN3-80E, and DEN4-80E respectively, formulated with varying doses of LNP (ranging from 25 μ g to 1500 μ g). Previous studies have determined that the DEN4-80E component is the least immunogenic and is therefore dosed at twice the concentration of the other three DEN-80E proteins in order to yield a more balanced nAb responses to all four serotypes. As controls, we utilized 15 units of ISCOMATRIX TM adjuvant and 45 μ g of Merck aluminum adjuvant (MAA). As depicted in Fig. 5a, Guinea pigs (n = 4) were vaccinated three times, two weeks apart, by the intramuscular route. 2 weeks after the last dose, sera from guinea pigs were collected and Dengue virus neutralization titles were determined. As shown in Fig. 5b, formulation with MAA did not significantly increase the nAb titers against any of the four DENV serotypes. Expectedly, ISCOMATRIX TM adjuvant led to statistically significant increases in nAb titers against all four DENV serotypes. Formulation of LNPs with DEN1-4-80E resulted in a dose dependent increase in nAbs titers against all four serotypes, wherein the 750 μ g and 1500 μ g doses were statistically significant. This confirmed the potential for LNP's to boost nAbs against all four serotypes of DENV when vaccinated with the tetravalent DEN1-4-80E antigens.

Evaluation of LNP containing tetravalent Dengue vaccine in non-human primates.
Having confirmed that LNPs generated strong neutralizing antibodies against Dengue envelope antigen(s) in mice and Guinea pigs, and enhanced DENV env specific T-cell responses in mice, we investigated the safety and efficacy of this vaccine formulation in non-human primates. Specifically, we utilized a well-established Rhesus macaque model 14 . For this study animals received 10 μ g, 10 μ g, 10 μ g, and 20 μ g of the DEN1-80E, DEN2-80E, DEN3-80E and DEN4-80E antigens respectively formulated alone or with adjuvant. As outlined in Fig. 6a, the animals were vaccinated at 0, 1 and 6 months. The vaccination formulation administered to each group of animals is listed in Fig. 6b-each group contained 3-4, weight and sex matched animals. All vaccine formulations were injected intramuscularly, into the right deltoid. Using a Draize scoring scale from 0-4, (criteria are described in Supplementary  Figure 1), animals were evaluated for the appearance of erythema and edema at the injection site for up to 7 days, post dose 1 (Fig. S1a), dose 2 ( Fig. S1b), and dose 3 (Fig. S1c). No vaccination-induced local site reactogenicity was observed in any of the vaccine groups, including animals that received the highest dose of LNP and RCL, suggesting that the Dengue vaccine formulation containing LNP was well tolerated (Supplementary Figure 1).
Using a 23-plex Luminex panel, we evaluated cytokine and chemokine profiles in sera from immunized animals, prior to dosing (0 hr) and at 4, 8, 24 and 48 hrs post dose 1. The majority of the cytokines and chemokines showed no significant changes at 48 hrs post injection (data not shown). Shown in Fig. 6c are fold changes of key cytokines and chemokines in each animal at 4, 8, and 24 hrs, calculated over the 0 hr time point of each animal. We chose to focus on the cytokines and chemokines that showed more consistent trends between animals within a group and the ones of biological relevance. The inflammasome specific cytokine IL-1β , was not significantly upregulated in any treatment group including in the LNP (high dose) treated animals, unlike what was observed in mice. The pro-inflammatory cytokine, IL-6 was mildly upregulated in most treatment groups  The immunization regime and dosing schedule for BALB/c mice vaccinated with 1 μ g of DEN2-80E antigen, formulated with PBS or LNPs (25 ug/75 ug/125 ug) or RCL (25 ug/75 ug/125 ug) are depicted (a). 2 weeks post dose 3, sera was assayed for DEN2-80E specific total IgG responses on an ELISA platform (b). 2 weeks post dose 3, spleens from all 5 mice/group were isolated, processed, and pooled to assess T cell responses. DEN2-80E specific CD4+ T cells that express IFN-γ , and TNF-α (c) and DEN2-80E specific CD8+ T cells that express IFN-γ , and TNF-α (d) were evaluated by intracellular FACS assay. Bars represent T-cell events per million splenocytes from 5 mice/group assayed in duplicate after non-stimulated subtraction. ***p < 0.001; *p < 0.05 as compared to PBS. The immunization regime, dosing schedule, and vaccine formulation details for Guinea pigs vaccinated with tetravalent Dengue vaccine (Tetra DEN-80E) is depicted (a,b). Dengue virus serotype 1 (D1), serotype 2 (D2), serotype 3 (D3), serotype 4 (D4) specific neutralizing activity from individual immunized Guinea pig sera was calculated using the LiCor based assay and depicted as LiCor microtneut NT50 titers (c). Each dot/symbol represents an individual animal. The geometric mean titers (GMT) of serotype specific neutralization are depicted on top of the graph above each group. DEN1 titers: *p < 0.001 as compared to DEN80E + PBS group; DEN2 titers: **p < 0.001 as compared to DEN80E + PBS group; DEN3 titers: ***p < 0.001 as compared to DEN80E + PBS group; DEN4 titers: ****p < 0.001 as compared to DEN80E + PBS group. including in the LNP (High) group, with or without DEN1-4-80E. However, the levels were far less significant as compared to the sustained upregulation observed in animals in the ISCOMATRIX TM adjuvant + DEN1-4-80E group (Group 3) and MF59/AddaVax TM + DEN1-4-80E group (Group 9). Sera from animals in the LNP high dose (group 1), LNP(High) + DEN1-4-80E (Group 6), RCL + DEN1-4-80E (Mid and High) (Groups 7 and 8), MF59/ AddaVax + DEN1-4-80E (Group 9) displayed an increase of the pro-inflammatory cytokine, TNF-α at 4 hrs. The Th1 type cytokine IL12 was also mildly upregulated in LNP containing vaccine formulations. The chemokine IL-8 was highly upregulated in all vaccination groups except the DEN1-4-80E only (Group 2). The monocyte chemotactic protein (MCP-1) was significantly upregulated in LNP (High) (Group 1) and LNP(High) + DEN1-4-80E (Group 6) at 4, 8, and 24 hrs, whereas DEN1-4-80E mixed with ISCOMATRIX TM adjuvant and MF59/AddaVax TM also led to increase in MCP-1.
In addition to cytokines and chemokines, a variety of safety biomarkers were evaluated in sera samples collected from immunized animals at day 1, 3 and 7-post dose 1 (PD1). Evaluation of Alanine Transaminase (ALT) (Fig. S2a), Aspartate Transaminase (AST) (Fig. S2b), Total bilirubin (Tbil) (Fig. S2c), Direct bilirubin (Dbil) (Fig. S2d), and C-Reactive Protein (CRP) (Fig. S2e), in sera samples at day 1, post-dose 1 revealed no significant elevation in immunized animals. These biomarkers were unaltered at day 3 and day 7 (PD1) samples as well (data not shown). Assessment of blood lipid content (lipemia), hemolytic activity and icteric activity showed no elevation post vaccination (data not shown). Furthermore, we evaluated C3a and CBb levels as a measure of complement activation in immunized animals (Supplementary Figure 3). Measurement of C3a (S3a) and CBb (S3b) in plasma samples from immunized animals at 0, 4, 24 and 72 hrs, post dose 3, demonstrated that LNP or RCL injection does not lead to a significant level of complement activation. Overall, through the exhaustive list of biomarkers evaluated, we conclude that the LNP containing Dengue vaccine formulations did not increase the level of biomarkers associated with liver injury, systemic inflammation, or complement activation.

LNP containing tetravalent Dengue vaccine formulation enhances Dengue Virus Neutralizing
Antibodies in Rhesus macaques. Having established safety profiles, the magnitude and durability of neutralizing antibodies (nAb) against each of the four Dengue serotypes was evaluated in serum samples from vaccinated animals. The LiCor-based micro-neutralization assay was used to determine nAb titres against all four serotypes Fig. 7a-d), every 4 weeks, longitudinally, out to 48 weeks from the initial vaccination date. As previously observed formulation of DEN1-4-80E with the ISCOMATRIX TM adjuvant (group 3) resulted in the highest geometric mean nAb titres against all four serotypes 14 . As observed in Guinea pigs, the nAb titres in NHPs that received LNP high + DEN1-4-80E (group 6) were comparable to those seen in the ISCOMATRIX TM adjuvant group (red line), and significantly higher than titres observed in the DEN1-4-80E + PBS (Group 2) or DEN1-4-80E + MF59/AddaVax TM (group 9) vaccinated animals. The responses were directly proportional to the dose of LNP used i.e. nAb titres were highest in animals that received the high dose (red line), followed by mid dose (orange line), and then low dose (cyan line) of LNP + DEN1-4-80E. Interestingly, as seen in mice, a reduction in the cationicity of the LNP (RCL), inhibited the LNP's ability to enhance nAb titres. The magnitude of nAbs titers to DEN1-4-80E vaccination waned over time, as seen at week 24 prior to the third dose. However, measurement of nAb titres 4 weeks after the third vaccine dose (week 28), showed a significant increase in nAb titres against all four serotypes both in group 3 and group 6. Interestingly, DEN1-4-80E + MF59/AddaVax TM (group 9/green line) also showed a modest increase in nAbs against all four Dengue serotypes, inspite of undetectable responses before the third dose. The nAb titres obtained from group 3 and group 6, declined slightly following the third dose, but were still detected out to week 48 (~6 months) after the last dose. Overall, these results demonstrate that LNP (High) + DEN1-4-80E vaccination led to significant nAb titres against all four Dengue virus serotypes, comparable to DEN1-4-80E+ ISCOMATRIX TM adjuvant (group 3), and significantly higher than all other vaccine formulations. DEN2 env specific CD4+ and CD8+ T cell responses are enhanced by Lipid Nanoparticles. In addition to evaluating nAb in the vaccinated Rhesus macaques (Fig. 6b), we evaluated T cell responses. 4 weeks post-dose 3 (Week 28), Peripheral Blood Mononuclear Cells (PBMCs) were collected from all vaccinated animals. One million PBMCs from each animal were individually incubated with peptides pools representing DEN-80E protein from each of the four serotypes. Using a FACS ICS assay, CD3 + CD4+ T cells and CD3 + CD8+ T cells that expressed cytokines such as IFN-γ , IL-2, and TNF-α specific to DEN1-80E, DEN2-80E, DEN3-80E, and DEN4-80E were measured. Surprisingly, no significant increase in CD4+ or CD8+ T cell responses specific to DEN1-80E, DEN2-80E, DEN3-80E antigens were observed in any immunized animals. However, both CD4+ and CD8+ T cells were detected against DEN4-80E antigen in some immunized animals. Shown in Fig. 8 (Fig. 8a) and CD8+ T cell responses (Fig. 8b) that were positive for IFN-γ , IL-2, and TNF-α . All the other vaccinated animals, including, groups containing ISCOMATRIX TM adjuvant or MF59/AddaVax TM with DEN1-4-80E antigen yielded minimal levels of detectable CD4+ T cell or CD8+ T cells responses. The multifunctional quality of the T cell responses from this assay was also evaluated i.e. DEN4 specific individual T cells that express one, two or all three cytokines. As depicted as piecharts in Fig. 8c, only LNP (high) + DEN1-4-80E vaccinated animals showed DEN4-80E specific multi-functional CD4+ and CD8+ T cell responses that expressed all three cytokines-IFN-γ , IL-2, and TNF-α . Interestingly, as seen in mice, RCL (high) + DEN1-4-80E did not result in a significant increase in DEN4 specific immune responses, further confirming that the ionizable cationic lipid content of LNPs is a crucial factor needed to boost antigen specific T cell Scientific RepoRts | 6:34215 | DOI: 10.1038/srep34215 responses (LNP vs RCL). We have previously reported that ISCOMATRIX TM adjuvant + DEN1-4-80E vaccination in Non-Human Primates induces appreciable amounts of antigen specific T cell responses [28]. Therefore, to further confirm the T cell responses assayed via FACS ICS (Fig. 7), we analyzed PBMCs collected from animal at week 28 and week 40, from group 2, group 3, and group 6 in an IFN-γ ELISPOT assay after incubation with DEN1/2/3/4-80E specific peptide pools. The ELISPOT assay is considered more sensitive than the FACS ICS assay, therefore, this served as an ideal confirmatory assay. The number of IFNγ + spot forming units (SFUs) per million cells was mock subtracted (DMSO treatment) and graphed in Supplementary Figure 4a (week 28) and Supplementary Figure 4b (week 40). At both time points, IFNγ + SFU specific for DEN1, DEN2, DEN3, and DEN4 were undetectable in the PBS group, minimally detectable in ISCOMATRIX TM adjuvant + DEN1-4-80E group, but readily detected in LNP (high) + DEN1-4-80E immunized animals. Confirming the FACS ICS data, the DEN4-80E specific responses were much higher in the LNP (high) + DEN1-4-80E vaccinated group as compared to DEN1/2/3-80E antigens. Overall, these data suggested that LNPs have the potential to enhance cell-mediated immune responses to a sub-unit recombinant antigen in non-human primates.

Discussion
A successful Dengue vaccine that protects against all four serotypes of Dengue viruse is urgently needed. Preclinical and clinical evaluation of sub-unit and live attenuated Dengue vaccines has met with varying degrees of safety and immunogenic protection 34,35 . Most recently, Sanofi Pasteur's recombinant, live, attenuated Chimeric Yellow Fever-Tetravalent Dengue Vaccine (CYD-TDV), that expresses preM and env proteins of Dengue virus, became the first Dengue vaccine to be approved for human use for the prevention of Dengue virus infection by two countries; Mexico 36 and Philippines 37 . Clinical trials in Asia and in Latin America reported that Dengue serotype 1, 3 and 4 specific seroconversion was successfully induced in individuals who received this vaccine, but insufficient seroconversion was observed for Dengue serotype 2 [38][39][40][41] .
Live-attenuated, viral vectored Dengue vaccines can result in an imbalance of equipotent serotype specific immune responses and thereby impacts vaccine-induced protection against all four serotypes of Dengue virus 42 . This issue can be circumvented with a subunit vaccine where one can strategically manipulate the Dengue antigen content/concentration included in the formulation. For instance, extensive research from our group has demonstrated that doubling the amount of recombinant envelope protein for Dengue virus serotype 4 in the sub-unit vaccine formulation compared to the env protein concentration of serotypes 1, 2 and 3, is required for generating a more balanced neutralizing antibody (nAb) response against all four serotypes in murine and non-human primate vaccination studies 14 . In addition, sub-unit Dengue vaccines offer substantial advantage over live attenuated vaccines since they are safer (to vaccinate immune suppressed and pediatric population), and provide long-term vaccine stability (low risk of genetic mutations).
Generation of nAbs against all four Dengue viruses has been a key objective of current Dengue vaccine candidates 19 . However, emerging evidence suggests that T cell responses are also crucial for protection against Dengue virus infection 43 . Based on our recent discovery that specific ionizable cationic lipid nanoparticles (LNPs) can boost B-cell and T-cell responses to sub-unit antigens, we evaluated a novel tetravalent subunit Dengue virus vaccine formulation containing ionizable cationic LNPs. Here we report that ionizable cationic LNPs significantly enhanced the generation of nAbs to dengue envelope proteins in mice, Guinea pigs, and Rhesus macaques, to levels comparable to the well-characterized ISCOMATRIX ™ adjuvant.
The lack of reliable animal models for Dengue virus infection and disease outcomes has been a significant limiting factor in the field 44 . Evaluation of safety and immunogenicity of Dengue vaccine candidates have largely relied on the use of Rhesus macaques. However, the Rhesus macaque model does not recapitulate the disease outcome of Dengue virus infection as observed in humans. We have extensively reported that nAbs titres induced , Peripheral Blood Mononuclear Cells (PBMCs) were collected from immunized animals and incubated with DEN1-80E, DEN2-80E, DEN3-80E, DEN4-80E specific peptide pools, and subjected to FACS ICS analysis. DEN4-80E specific CD4+ T cells that express IFN-γ , IL-2, and TNF-α (a) and DEN4-80E specific CD8+ T cells that express IFN-γ , IL-2, and TNF-α (b) were calculated as T-cell events per million PBMCs. Each dot/symbol represents an individual animal. The multi-functionality of the T-cell responses was calculated using the SPICE software and represented here as pie charts. Out of the three cytokines assessed (IFN-γ , IL-2, and TNF-α ) in antigen specific T-cells, expression of one cytokine by T-cells from a vaccinated group is assigned blue, two cytokines is assigned orange and three cytokines is assigned red (c). *p < 0.05 as compared to DEN80E + PBS group. upon DEN1-4 80E + ISCOMATRIX ™ adjuvant vaccination is sufficient to reduce viremia against Dengue virus challenge 14,45 . LNP (High) + DEN1-4-80E vaccination induced antibody titers that were comparable to DEN1-

4-80E + ISCOMATRIX ™ adjuvant vaccination in quality (neutralizing all four serotypes of Dengue viruses) and
durability. Therefore, we chose not to challenge the LNP + DEN1-4-80E immunized animals, as we fully expect that the animals would have no detectable viremia. In addition, we would not be able to investigate the impact of LNP + DEN1-4-80E vaccination on Dengue virus induced disease severity.
The evaluation of cell mediated immune responses in mice and Rhesus macaques revealed that the LNPs were far superior to ISCOMATRIX ™ adjuvant in generating poly-functional CD4+ T cell responses and CD8+ T cell responses in immunized animals. In mice, we observed that the frequency of DEN2-80E specific CD8+ T cell responses, but not CD4+ T cell responses, was slightly reduced in groups that received formulations containing LNP and TLR9-ODN (Fig. 2c). This unique observation warrants extensive investigation of the specific subsets of innate immune cells that uptake the antigen upon co-administration with LNP versus LNP + TLR9-ODN. We hypothesize that cell types that are recruited to the injection site, uptake antigen and migrate to the local draining lymph nodes may be different between LNP and TLR9-ODN containing groups. In particular, the cross-presenting innate immune cells such as the CD8+ dendritic cells. In Rhesus macaques, inspite of minimal DEN-80E specific CD4+ T cell responses induced upon DEN-80E + ISCOMATRIX ™ adjuvant and DEN-80E + LNP (high) vaccination, we observed significant Dengue virus specific nAb titers. However, it has been established in the literature that CD4+ T cells are not required for generation of Dengue vaccine specific nAb or CD8+ T cell responses 46 . In LNP (high) + DEN1-4-80E vaccinated non-human primates, out of the four antigens (DEN1-4-80E), we only observed CD8+ T cells responses to DEN4-80E. This may be because the MHC-I specificity for Dengue antigens has not been characterized in Rhesus macaques. Furthermore, the Dengue virus non-structural proteins NS1, NS3 and NS5 have been characterized to be the main epitopes of CD8+ T cell responses in humans 47 and Cynomolgus macaques 48 vaccinated with live-attenuated Dengue virus vaccines. Therefore, inclusion of NS1, NS3 and NS5 Dengue virus non-structural proteins, as additional antigens in the LNP + DEN1-4-80E vaccine formulation might be beneficial. The ability of our ionizable cationic LNPs to boost DEN-80E specific CD8+ T cell responses in mice and, more importantly, in non-human primates is of utmost significance. Generation of CD8+ T cell responses to recombinant protein antigens in non-human primates is exceptionally rare. To our knowledge, the only known published report demonstrating this effect required the direct conjugation of a Toll-Like Receptor 7/8 (TLR7/8) agonist to the protein antigen 49 . Here, we report that the LNP has the ability to induce multi-functional, CD8+ T cell responses (Fig. 8b) without the need for direct conjugation to the protein. This offers significant manufacturing advantages to the vaccine formulation. More importantly, there was no requirement for inclusion of any known innate immune agonists in the vaccine formulation to yield significant B-cell or T-cell responses against the antigen.
We postulate that the LNP has strong immune-modulatory properties, possibly enabling the activation innate immune pathways. Although not experimentally confirmed at this time, we believe that our LNP is able to induce such strong CD8+ T cell responses through its ability to either specifically deliver the co-administered antigen to cross-presenting DCs or through activating these cells to enhance antigen presentation. Furthermore, we are committed to uncovering the mechanism of action behind the LNP's ability to boost such robust B-cell, CD4+ T cell and more specifically, CD8+ T cell responses to sub-unit antigens. It has been reported that lipid nanoparticles can activate specific Toll-Like Receptor pathways 50,51 as well as NLRP3-inflammasome pathways 50,52 . Current and future studies are aimed at utilizing transgenic mice that are defective in specific immune pathways (specific TLR knock out mice, NLRP3 inflammasome receptor knock out mice, MyD88 receptor knock out mice), and in-vitro immune cell assays to investigate which innate immune pathway is required for the functionality of our lipid nanoparticles. Such specific understanding will enable the design and development of lipid nanoparticles as effective delivery vehicles and immune-modulators in vaccine formulations.
By modulating the charge density of the ionizable cationic lipid within the LNP, we demonstrate that the ionizable cationic lipid content is crucial for the LNP's ability to boost immune responses to the Dengue antigen. It is possible that the reduction in LNP charge density affects the biophysical properties of DEN-80E antigen-LNP interactions and therefore might impact the LNP's ability to boost immune responses in-vivo. Therefore, future studies will be aimed at evaluating the physio-chemical properties of LNP, RCL, and alternative ionizable cationic lipids when co-formulated with the DEN-80E antigens.
We did not observe prolonged, significant expression of inflammatory cytokines in mice and non-human primates injected with LNP-containing vaccine formulations. Importantly, we observed no significant increase in safety biomarkers evaluated in LNP + DEN-80E vaccinated non-human primates, including no visible reactogenicity at the local site of injection. Administration of select lipid nanoparticles and liposomes via intravenous (IV) route has been reported to lead to allergy and anaphylaxis due to complement activation 53,54 . Our evaluation of markers of complement activation (C3a and CBb) in the plasma of vaccinated animals at numerous time points demonstrated no significant elevation suggesting that the chosen LNP formulation, dose range of LNP, and the intramuscular route of injection (as compared to IV), supports our LNP + DEN80E formulation as a well-tolerated vaccine.
Overall, this the first study reporting a unique, tetravalent, Dengue virus sub-unit vaccine formulation containing ionizable cationic lipid nanoparticles, that is highly efficacious and well-tolerated, with the promise of serving as a next-generation Dengue virus vaccine. The level of safety and efficacy reported here warrant the need for continued evaluation of LNP containing sub-unit vaccine formulations against Dengue virus and other emerging Flaviviruses such as Zika virus. Care and Use Committee. Indian Rhesus macaques of either sex, weighing more than 3 kg, were weight and sex matched for each group. All animals were Flavivirus naïve i.e. sera negative for Dengue virus 1,2,3,4 and West Nile virus. Vaccines were administered to Rhesus macaques as single injection per dose, intramuscularly on the right deltoid, in a total volume of 500 μ l. The composition of vaccines administered per group is listed in Fig. 6b. Lipid Nanoparticles, Adjuvants & Vaccine Antigens. Lipid nanoparticle (LNP), alone or encapsulating TLR9-ODN (LNP in TLR9-ODN), were prepared by rapid precipitation process as previously described 33 . The lipid components of LNP adjuvants comprised an asymmetric ionizable amino lipid (Lipid D, [47]), distearoylphosphatidylcholine (DSPC), cholesterol, and poly(ethylene glycol)2000-dimyristoylglycerol (PEG2000-DMG). LNP and LNP's Reduced Cationic Lipid version (RCL), were prepared at varying molar ratios of the constituent lipids. LNP was prepared at a molar ratio of 58:30:10:2 ionizable amino lipid, cholesterol, DSPC and PEG-lipid 16 55 . Saponin based ISCOMATRIX TM adjuvant was manufactured and supplied as sterilized bulk by CSL Limited, Australia under a license agreement. DEN-80E antigen for all four serotypes was expressed using stably transformed Drosophila S2 cells, and purified as previously descried 13 . For mice studies, 1 μ g of DEN2-80E antigen was utilized. Tetravalent vaccine formulations of DEN1, DEN2, DEN3, DEN4 serotype specific "80E" env protein (DEN1-4-80E) content were: 3 μ g, 3 μ g, 3 μ g, 6 μ g for Guinea pig studies and 10 μ g, 10 μ g, 10 μ g, 20 μ g for Rhesus macaques studies. Safety Parameters. Rhesus macaques were injected with specified vaccine formulations on the right deltoid, and monitored for signs of edema and erythema by a certified veterinarian present on site at the New Iberia Primate Research Facility (NIRC), Louisiana. Local site reactogenicity was scored on a scale of 0-4 as indicated in the Draize scoring criteria in Supplementary Figure 1. Circulating safety biomarkers were evaluated from animals in all groups by collecting whole blood into serum or plasma collection tubes at indicated time points. Sera samples were utilized for assessing ALT, AST, Tbil, Dbil and C-Reactive Protein levels at indicated time points, and plasma samples were utilized for evaluation of complement activation markers, C3a and CBb. Both serum specific and plasma specific biomarkers were evaluated by established protocols in-house, at Merck Research Laboratories in West-Point, Pennsylvania, as previously described 56-58 . Cytokine & Chemokine Analyses. Sera from immunized mice and Rhesus macaques were collected at  Fig. 6c. B-cell responses. For assessing DEN2-80 specific immune responses, sera were collected from immunized mice 2 weeks post dose 3 (as depicted in Fig. 1a). Endpoint ELISA's were utilized to assay DEN2-80E specific total IgG, IgG1 and IgG2a titres in mouse sera. Briefly, 96 well plates (NUNC #430341) were coated with overnight with 1 μ g/mL of DEN2-80E antigen, blocked with a buffer containing PBS + 1% BSA (Sigma #A7030), and stained with limiting dilutions of sera from immunized mice. After washing, Goat anti-mouse IgG (Invitrogen, Carlsbad, CA, #626520), Goat anti-mouse IgG1 (S. Biotech #1070-05), or Goat anti-mouse IgG2a Abs (S. Biotech #1080-05), conjugated with HRP, were used as secondary antibodies at 1:6000 dilution. The reaction was developed by addition of TMB substrate (Virolabs, VA), and plates were read using PerkinElmer Envision plate reader measured absorbance at 450 nm.
T cell responses. Peripheral Blood Mononuclear Cells (PBMCs) were collected from immunized Rhesus macaques at indicated time points, and kept frozen until use. Spleens from vaccinated mice, two weeks post dose 3, were collected and processed to isolate splenocytes. PBMCs and splenocytes were stimulated in RPMI based R10 media at 37′ c, as previously described elsewhere 25 .
Dengue virus neutralization titres were assayed using an infra-red dye (IRD) based high-throughput and sensitive neutralization assay that was developed and well-established by by our group (described in detail 18,61 ). Briefly, vero cells were seeded at 1.5 × 10^4 cells/well in a 96wp format. 1000 pfu/ml of DEN1-4 virus was pre-incubated with dilutions of heat-inactivated sera from immunized mice, Guinea pigs, and Rhesus macaques, for 1 hr at 37′ C, and added on to vero cells. 4 days later cells were fixed, and incubated with anti-Dengue specific 4G2 mouse antibody (SDIX) followed by biotinylated anti-mouse IgG antibody Catalogue no. BA-2000) (Vector labs Inc., Burlingame, CA), followed by a cocktail of DRAQ5 dye (Cell Signaling, MA), and IRDye 800CW Streptavidin (Li-Cor). Stained plates were washed, air-dried and were read using an infrared Odyssey SA imaging system (Li-Cor Biosciences, NE). Serum end-point neutralization titers or NT50 values calculated i.e. the reciprocal of the highest serum dilution that reduces the 800 nm/700 nm fluorescence integrated intensity ratio greater than 50% when compared to virus control included on each assay plate. If a sample failed to neutralize at this dilution, a titer of 1:5 was assigned to that sample and this dilution was used to calculate the geometric mean of that group.