Development of a potent Zika virus vaccine using self-amplifying messenger RNA.

Zika virus (ZIKV) is the cause of a pandemic associated with microcephaly in newborns and Guillain-Barre syndrome in adults. Currently, there are no available treatments or vaccines for ZIKV, and the development of a safe and effective vaccine is a high priority for many global health organizations. We describe the development of ZIKV vaccine candidates using the self-amplifying messenger RNA (SAM) platform technology delivered by cationic nanoemulsion (CNE) that allows bedside mixing and is particularly useful for rapid responses to pandemic outbreaks. Two immunizations of either of the two lead SAM (CNE) vaccine candidates elicited potent neutralizing antibody responses to ZIKV in mice and nonhuman primates. Both SAM (CNE) vaccines protected these animals from ZIKV challenge, with one candidate providing complete protection against ZIKV infection in nonhuman primates. The data provide a preclinical proof of concept that a SAM (CNE) vaccine candidate can rapidly elicit protective immunity against ZIKV.

INTRODUCTION

Zika virus (ZIKV) was originally identified in 1947 in the blood of a rhesus monkey from the Zika forest of Uganda, but the first human infection was not reported until 1952 (). Clinical cases of ZIKV infection were rarely reported until 2007, when an outbreak in the Federated States of Micronesia is believed to have infected an estimated 73% of the population (). This was followed by confirmed outbreaks in Micronesia and French Polynesia in 2013 to 2014, with an estimated 11% attack rate () and infections associated with neurological abnormalities, particularly the Guillain-Barre syndrome. By 2014, cases of ZIKV infection were reported in Brazil, and ZIKV infection of pregnant women was implicated in a 20-fold increase in the occurrence of neonatal microcephaly (). In February 2016, as the ZIKV epidemic continued to spread throughout the Americas and the Caribbean, the World Health Organization (WHO) declared ZIKV a Public Health Emergency of International Concern (WHO, 2016). Although the number of infections has decreased since the peak of infection in 2016, ZIKV is likely to continue to circulate endemically and cause sporadic outbreaks. Therefore, there is an ongoing need for a ZIKV vaccine that functions to protect women from infection conferred by the bite of an Aedes species mosquito or through sexual transmission, particularly during pregnancy when congenital transmission to the fetus can have devastating effects.

ZIKV is an enveloped RNA virus in the Flaviviridae family. Its ~11–kilo base pair (kbp) positive-sense RNA genome is translated into a single polypeptide, chain that is subsequently cleaved into structural proteins, including capsid (C), premembrane/membrane (prM), and envelope (E), and nonstructural proteins, including NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5. ZIKV buds into the lumen of the endoplasmic reticulum as an immature virion that displays 60 trimers of prM-E heterodimers arranged with icosahedral symmetry on its surface (). The “pr” portion of prM is cleaved by the host furin protease during egress from the cell, producing the mature infectious virion organized as 90 antiparallel dimers of M-E heterodimers that expose distinct surfaces of E, as compared with the immature form (, ). As in the case with other flaviviruses, ZIKV E protein has been identified as the primary target of potently neutralizing antibodies (–). Expression of flavivirus prM and E proteins is sufficient for the formation of noninfectious subviral particles (SVPs) (–). Gene-based delivery of ZIKV prM-E as vaccine antigens has been shown to protect against ZIKV challenge in animal models and is immunogenic in humans (–).

Self-amplifying messenger RNA (SAM) technology is an RNA-based vaccine platform that uses a 9-kbp self-amplifying RNA derived from an alphavirus genome (–). SAM-based vaccines can be rapidly produced by in vitro transcription and offer several advantages particularly suitable for rapid response in the event of an outbreak. The RNA is generated in a cell-free in vitro system, making it free of potential cellular contaminants and suitable for rapid scale-up. Self-amplification of SAM results in large amounts of subgenomic RNA in situ, leading to high expression levels of the vaccine antigen and the presence of double-stranded RNA (dsRNA), which activates interferon pathways, producing a self-adjuvanting effect. The delivery of SAM vaccines by a cationic nanoemulsion (CNE) has induced a potent immune response to multiple different antigens in both mice and nonhuman primates (NHPs) (–). CNE can be prepared and stockpiled separately from the RNA for later use (, ), and the SAM vaccine can be administered via needle and syringe, both of which can simplify the implementation and are particularly advantageous for the rapid response to pandemic outbreaks and emerging infectious diseases. We have used this platform to generate multiple vaccine candidates expressing the prM and E antigens of ZIKV. Here, we present the down selection of ZIKV vaccine candidates and the development of a potent ZIKV SAM (CNE) vaccine that induces protective immunity against ZIKV challenge in mice and NHPs. This study demonstrates how SAM technology can be used to rapidly develop and evaluate multiple vaccine candidates in response to emerging pathogens.

RESULTS

Generation of ZIKV SAM candidate constructs

Nine ZIKV SAM constructs were designed on the basis of wild-type (WT) sequences of the ZIKV structural proteins C, prM, and E to potentially enhance expression, antigen processing, antigen secretion, and SVP production (Table 1 and Fig. 1). The WT-prM-E construct expressed the ZIKV prM-E sequence from the Brazilian Natal strain (KU527068.1). The CO-prM-E construct was a codon-optimized (CO) version of WT-prM-E. To potentially enhance antigen secretion, CO-prM-E was further modified by mutating the prM signal sequence to the PQAQA enhanced signal sequence (ESS) to generate CO-prM-E.ESS.1 () or by replacing with the immunoglobulin G (IgG) signal sequence to generate CO-prM-E.ESS.2. In addition, we designed three codon-optimized constructs by including the ZIKV C protein in the background of CO-prM-E to potentially enhance the formation and antigenic quality of SVPs. CO-C-prM-E.1 carried the C and prM-E sequences with the native C/prM cleavage site, where the cleavage is rather inefficient in the absence of the virus-encoded protease. CO-C-prM-E.2 was identical to CO-C-prM-E.1, except that the prM signal sequence was mutated to the PQAQA ESS. To enable efficient cleavage between C and prM, CO-C-prM-E.3 was generated by insertion of a 2A self-cleaving sequence in the native cleavage site between C and prM in CO-C-prM-E.1. Last, we also included two constructs, VRC5283 and VRC5288, that have previously been tested as DNA vaccines based on the French Polynesian H/PF/2013 strain (, ). The prM-E amino acid sequence is identical between the Natal strain and the H/PF/2013 strain. VRC5283 SAM and DNA plasmid expressed the codon-optimized prM-E sequence but contained the prM signal sequence from the Japanese encephalitis virus (JEV), which was designed to increase antigen secretion and promote SVP production (). VRC5288 was identical to VRC5283 except that the C-terminal 98 amino acids of the ZIKV E protein were replaced with those from the JEV. This C-terminal domain comprises the stem and transmembrane region of the E protein and theoretically should not be surface exposed in the mature virus particles. The JEV stem and transmembrane substitution has been demonstrated to improve SVP secretion for ZIKV and other flaviviruses (, ). All constructs were generated by synthesizing the necessary inserts and cloning them into the SAM vector backbone (Fig. 1).

Table 1

Description of the SAM vaccine candidates.

Name	Description
WT-prM-E	Wild-type (WT) prM-E sequencefrom ZIKV Natal strain (Brazil)KU527068.1.
CO-prM-E	Identical to WT-prM-E, except thatthe prM-E sequence was codonoptimized (CO).
CO-prM-E.ESS.1	Identical to CO-prM-E, except thatthe prM signal peptide “cregion” contained theenhanced signal sequence(ESS) PQAQA mutation.
CO-prM-E.ESS.2	Identical to CO-prM-E, except thatthe prM signal sequence wasreplaced with theimmunoglobulin G signalsequence.
CO-C-prM-E.1	Identical to CO-prM-E, except thatit additionally contained thecodon-optimized capsid (C)protein with the native C/prMcleavage site.
CO-C-prM-E.2	Identical to CO-C-prM-E.1, exceptthat the prM signal peptide “cregion” contained theenhanced signal sequence(ESS) PQAQA mutation.
CO-C-prM-E.3	Identical to CO-C-prM-E.1, exceptthat a 2A sequence wasinserted in the native cleavagesite between C and prM.
VRC5283	CO prM-E sequence from the ZIKVFrench Polynesian H/PF/2013strain, with the prM signalpeptide replaced with that ofJEV.
VRC5288	Identical to VRC5283, except thatboth the prM signal peptideand the last 98 amino acids of Eprotein were replaced withthose of JEV.

Fig. 1

Schematic of ZIKV SAM constructs.

Schematic of amino acid sequence differences of SAM constructs designed to express ZIKV prM-E or C-prM-E.

In vitro characterization of the ZIKV SAM candidate constructs

We first evaluated the in vitro characteristics of the ZIKV SAM constructs. SAM RNAs were transcribed in vitro from template DNA constructs using T7 polymerase, and RNA integrity was validated by agarose gel electrophoresis (fig. S1). To assess the ability of RNAs to launch and amplify, we measured the presence of dsRNA amplifying intermediates in cells. Baby hamster kidney (BHK) cells were transfected with 0.1 μg of RNA, and after 16 to 20 hours, cells were stained using an anti-dsRNA antibody. RNA potency, calculated as the percentage of dsRNA-positive cells, was determined by flow cytometry. Each ZIKV SAM RNA showed a potency greater than 40%, and cells positive for ZIKV SAM dsRNA showed comparable mean fluorescence intensity (MFI), suggesting an efficient launch of self-amplification (Fig. 2A). To assess protein production, BHK cells were transfected with 4 μg of RNA, and expression levels of prM-E were determined 24 hours after transfection by immunoblot using the pan-flavivirus E–specific monoclonal antibody 4G2. WT-prM-E, CO-prM-E, VRC5283, or VRC5288 RNA expressed E antigen at levels higher than CO-C-prM-E.3 or CO-prM-E.ESS.2 RNA (Fig. 2B). Moreover, E antigen expression was barely detectable in cells transfected with the CO-C-prM-E.1, CO-C-prM-E.2, or CO-prM-E.ESS.1 RNA. Consequently, the four constructs with the highest levels of antigen expression (WT-prM-E, CO-prM-E, VRC5283, and VRC5288) were selected for evaluation in animal studies.

Fig. 2

In vitro characterization of ZIKV SAM constructs.

(A) Potency of the ZIKV SAM constructs. The RNA was in vitro transcribed, and 0.1 μg of the RNA from each vaccine construct was electroporated into BHK cells. Cells were collected 24 hours later and stained with an anti-dsRNA antibody. The upper and lower panels show the percentage and MFI of dsRNA-positive cells determined using flow cytometry, respectively. Error bars represent the SD of technical triplicates. Shown are the representatives of at least two experiments. (B) ZIKV SAM RNA expression of prM-E protein. BHK cells were electroporated with 4.0 μg of the indicated SAM RNA. Twenty-four hours later, cells were collected and lysates were analyzed by immunoblotting using antibody 4G2 specific to flavivirus E protein or to β-tubulin. Shown are the representatives of at least two experiments.

ZIKV SAM (CNE) vaccines induce a potent protective immune response in mice

To test the immunogenicity of the selected ZIKV SAM vaccine candidates, 10 BALB/c mice per group were immunized intramuscularly with 1.5 or 15 μg of WT-prM-E, CO-prM-E, or VRC5283 RNA. VRC5288 RNA was only evaluated at 1.5 μg. All of the RNAs were formulated with CNE [i.e., SAM (CNE) vaccine]. The ratio of CNE to the RNA, termed the nitrogen:phosphate (N:P) ratio, is described as the molar ratio of protonatable nitrogen atoms in the cationic lipid of CNE to phosphates on the RNA. For these constructs, the optimal ratio was determined to be 6.3:1 (fig. S2). All subsequent references to dosage are based on the amount of RNA with the CNE formulation that remains at a 6.3:1 ratio with the RNA. For comparison, one group of mice was also immunized with 50 μg of VRC5283 DNA vaccine intramuscularly using electroporation, previously shown to be immunogenic (). Mice were immunized at days 0 and 21 and challenged intraperitoneally at day 49 with 100 focus-forming units (FFUs) of ZIKV strain PRVABC59 (Fig. 3A). Neutralizing antibody activity was examined at days 0, 14, and 35 using a ZIKV reporter virus particle (RVP) neutralization assay, as previously described (). When administered at 1.5 μg, WT-prM-E, CO-prM-E, and VRC5283 SAM (CNE) vaccines elicited neutralizing antibody responses in most animals at day 35, whereas VRC5288 SAM (CNE) vaccine elicited detectable neutralizing activity in only 3 of 10 animals (Fig. 3B). At this dose, VRC5283 SAM (CNE) vaccine elicited the highest neutralizing antibody titers among the four SAM vaccines at day 35 (P < 0.05; Fig. 3B). This trend was preserved when ZIKV SAM (CNE) vaccines were administered at a dose of 15 μg. Of note, VRC5283 DNA and VRC5283 SAM (CNE) vaccines elicited significantly higher neutralizing activity than the WT-prM-E SAM (CNE) vaccine at day 35 (P < 0.05; Fig. 3C). Mice that received 15-μg doses of WT-prM-E, CO-prM-E, and VRC5283 SAM (CNE) vaccines had significantly higher neutralizing activity than mice that received 1.5-μg doses (P < 0.0001; Fig. 3, B and C).

Fig. 3

ZIKV SAM (CNE) vaccines are immunogenic and protective in mice.

(A) BALB/c mice were immunized with ZIKV SAM (CNE) vaccines or VRC5283 DNA vaccine at the indicated doses at days 0 and 21 and subsequently challenged with 100 FFU of ZIKV at day 49. (B and C) Neutralizing antibody titers were determined by RVP neutralization assay at days 0, 14, and 35. Horizontal lines and error bars represent the mean log10 reciprocal EC50 (half-maximal neutralization of infection) dilution ± SD of 10 mice per group, respectively. The dotted line represents the limit of confidence (LOC), a reciprocal titer of 60. Any replicates below LOC were assigned a value of 30 (0.5 LOC). Significant difference by two-way analysis of variance (ANOVA) with Tukey’s multiple comparison posttest between vaccines at the same doses as indicated: *P ≤ 0.05, **P ≤ 0.01, and ****P ≤ 0.0001. (D) Viral loads were determined by qRT-PCR at day 3 after challenge. Horizontal line and error bars represent mean log10 FFU equivalent/ml ± SD for 10 mice per group, respectively. Dotted line represents the LOC. Any replicates below the LOC were assigned a value of 0.5 LOC. #P < 0.0001 between all other groups by one-way ANOVA with Tukey’s multiple comparison posttest.

The protective efficacy of the vaccines was assessed by quantifying viral load by quantitative real-time polymerase chain reaction (qRT-PCR) 3 days after ZIKV challenge (Fig. 3A). All vaccine regimens resulted in significantly reduced viral load compared with unvaccinated mice (P < 0.0001; Fig. 3D). WT-prM-E, CO-prM-E, and VRC5283 SAM (CNE) vaccines were protective against ZIKV challenge at both the 1.5- and 15-μg doses (Fig. 3D). At the 1.5-μg dose, all three vaccine groups had only one animal with detectable viral load. In contrast, viral load was detectable in all but two animals vaccinated with 1.5 μg of VRC5288 SAM (CNE) vaccine, which elicited little to no neutralizing antibody activity. At the 15-μg dose, mice vaccinated with VRC5283 SAM (CNE) vaccine had no detectable viral load, whereas one mouse in each of the WT-prM-E and CO-prM-E SAM (CNE) vaccine groups had low but detectable viral load. Similar to its SAM (CNE) counterpart, the VRC5283 DNA vaccine also demonstrated complete protection, consistent with previous studies (). Overall, WT-prM-E, CO-prM-E, and VRC5283 SAM (CNE) vaccines elicited the most protective responses and were advanced into the next stage of evaluation.

ZIKV SAM (CNE) vaccines induce a potent protective response in NHPs

To determine the lead vaccine candidate, the immunogenicity and efficacy of VRC5283 SAM (CNE) and CO-prM-E SAM (CNE) vaccines were tested in rhesus macaques. On days 0 and 28, four groups of eight animals were vaccinated intramuscularly by needle and syringe with phosphate-buffered saline (PBS) (placebo), 75 μg of VRC5283 SAM (CNE) vaccine, or 75 μg of CO-prM-E SAM (CNE), or by PharmaJet needle-free device with 4 mg of the VRC5283 DNA vaccine (Fig. 4A). A dose of 75 μg of SAM (CNE) was selected because it had previously been demonstrated to be immunogenic in NHPs (). Neutralizing antibody activity in sera was assessed on days 0, 28, and 56 to determine preexisting immunity and antibodies elicited by the first and second vaccinations, respectively. At day 56, animals were challenged subcutaneously with 1000 FFUs of ZIKV strain PRVABC59. Viremia in sera was assessed on days 3, 4, 5 and 7 after challenge by qRT-PCR. Neutralizing antibody activity was also assessed on days 70 and 84 to measure the anamnestic response to ZIKV challenge.

Fig. 4

ZIKV SAM (CNE) vaccines are immunogenic and protective in NHPs.

(A) Rhesus macaques (n = 8 per group) were immunized with the indicated vaccines at days 0 and 28 and subsequently challenged with 1000 FFU of ZIKV at day 56. (B) Neutralizing antibody activity was determined by ZIKV RVP assay at days 0, 28, and 56. Horizontal line and error bars represent the mean log10 reciprocal EC50 dilutions ± SD of eight animals per group, respectively. Statistical difference is shown between vaccines at each time point by two-way ANOVA with Tukey’s multiple comparison posttest. Dotted line represents the LOC. Any replicates below the limit of detection were assigned a value of 0.5 LOC. (C) Viral loads were determined by qRT-PCR at days 3, 4, 5, and 7 after challenge. Each animal is depicted as a single line. Dotted line represents the LOC. Any replicates below the LOC were assigned a value of 0.5 LOC. (D) Postchallenge anamnestic antibody response. Fold change in EC50-neutralizing antibody activity is shown relative to day 56. Each animal is depicted as a single line. Dotted line indicates a fold change of 4.

A single dose of each vaccine candidate elicited neutralizing antibodies that were significantly boosted by a second dose (P < 0.0001; Fig. 4B). Following the first vaccination (day 28), the CO-prM-E group showed significantly lower neutralizing activity than those in the DNA VRC5283 group (P < 0.001), but there was no significant difference following the second vaccination (day 56). There was no significant difference in the neutralizing activity in the VRC5283 SAM (CNE) and DNA vaccine groups after the first or second vaccination.

After ZIKV challenge, the PBS-immunized control animals were uniformly infected with peak viremia occurring at day 4 or 5 after challenge (Fig. 4C). All control animals subsequently developed neutralizing antibodies in response to ZIKV infection (Fig. 4D). Compared with the placebo, the CO-prM-E vaccine demonstrated considerable protection with only one animal showing a low level of viremia at day 7 after challenge (Fig. 4C). Correspondingly, this animal had a 23-fold rise in neutralizing antibody activity after challenge (Fig. 4D). Two additional animals in that group had a greater than fourfold change in neutralizing antibody activity despite having undetectable viremia. The VRC5283 DNA vaccine also demonstrated protection with only one animal (0BC) having detectable viremia in only one of three qRT-PCR replicates on a single day (Fig. 4C). This animal had an eightfold rise in neutralizing antibody, supporting the conclusion that this animal had low-level infection. Remarkably, the VRC5283 SAM (CNE) vaccine showed complete protection from ZIKV infection with no animals having detectable viremia based on qRT-PCR (Fig. 4C) or a greater than a fourfold increase in neutralizing antibody titers after challenge (Fig. 4D). These data show that the VRC5283 SAM (CNE) vaccine consistently elicited sterilizing immunity to ZIKV infection in mouse and NHP challenge models.

DISCUSSION

Despite the waning epidemiology, development of a vaccine for ZIKV infection remains a public health priority to protect against congenital ZIKV infection in future outbreaks. Here, we describe the preclinical selection and development of a potent ZIKV vaccine from nine constructs using SAM technology expressing various forms of the ZIKV prM-E antigen. One ZIKV SAM vaccine candidate provided complete protection from ZIKV challenge in both mice and NHPs. This is the first time that a SAM vaccine has been demonstrated to be protective in the NHP model. We used a CNE formulation as a delivery vehicle. This is a technology that has been well characterized and is known to generate a rapid immune response when administered. CNE is stable at 2° to 4°C for years and can easily be mixed with RNA at bedside by trained medical personnel. These properties make SAM CNE particularly suitable for rapid response to emerging outbreaks (–). Other RNA-based ZIKV vaccine candidates have been previously described (, ), but vaccines based on a self-amplifying RNA platform have the potential to express higher levels of antigen compared with conventional mRNA-based platforms, and the self-adjuvanting effect of amplifying RNA intermediates can further enhance vaccine immunogenicity ().

The SAM platform allows rapid screening of diverse vaccine design concepts. We tested nine SAM constructs designed to express ZIKV prM-E antigens that varied in prM signal sequence, codon optimization, inclusion of the C protein, cleavage sites, and the sequence of the E transmembrane and stem domain. All designs produced similar levels of dsRNA intermediates, indicating an equal ability to produce transcripts, although protein expression varied. The prM signal sequence played a key role in protein expression, but the mutation made to enhance expression by including the enhanced signal peptide sequence (PQAQA) in CO-prM-E.ESS.1 failed to drive antigen expression. The cleavage region of the prM signal peptide is hydrophobic, and the PQAQA mutation changes the hydrophobicity, which has been previously reported to enhance E expression (). Consistent with our findings, it has been shown to increase cleavage at the C-prM junction, resulting in increased E expression, when C protein was present (). However, our result also indicated that when C protein was absent, the mutation could not improve the expression of prM and E proteins. Substitution of the native ZIKV signal sequence used in CO-prM-E and WT-prM-E with the IgG (CO-prM-E.ESS.2) or JEV signal sequence (VRC5283) had little impact on protein expression. Inclusion of the C protein with the native cleavage sites (CO-C-prM-E.1 and CO-C-prM-E.2) decreased protein expression. This was expected, as cleavage at the native site is inefficient without the viral protease. Expression was rescued by the mutation of the cleavage site to include a P2A site (CO-CprM-E.3), but overall expression was not increased compared with constructs lacking C. In future studies, this vaccine approach could be used to determine whether the inclusion of C enhances the protective efficacy of a SAM vaccine by affecting the functional quality of antibody responses.

The in vitro antigen expression data allowed us to down select four vaccine designs from nine vaccine candidates, and three SAM (CNE) vaccines, WT-prM-E, CO-prM-E, and VRC5283, elicited robust neutralizing antibody responses and were protective in ZIKV challenge animal models. Despite encoding similar sequences, expressing similar levels of intracellular antigen in vitro, and eliciting similar levels of neutralizing antibody activity, only the VRC5283 SAM (CNE) vaccine consistently provided complete protection in NHPs (75-μg dose) and in mice (15-μg dose). Although WT-prM-E and CO-prM-E were based on the Natal strain and VRC5283 is based on the H/PF/2013 strain, the prM-E amino acid sequence is identical, and CO-prM-E and VRC5283 were codon optimized using the same algorithm. Thus, the only difference between CO-prM-E and VRC5283 is that the native prM signal sequence used in CO-prM-E was replaced with the corresponding JEV prM signal sequence in VRC5283. It is intriguing to speculate that the JEV prM signal sequence may enhance prM-E production in vivo, play a role in SVP formation, or affect the topology of SVPs and qualitative aspects of the antibody response (). These findings demonstrate the importance of evaluating multiple vaccine candidates in vivo even when only subtle differences can be detected in vitro.

The VRC5283 DNA vaccine has previously been shown to protect NHPs from ZIKV infection and is currently in phase 2 clinical trials (, ). Here, we demonstrate that 75 μg of SAM (CNE) vaccine elicited similar neutralizing activity and protection as 4 mg of DNA vaccine expressing the same protein sequence. The VRC5283 DNA vaccine elicited a higher neutralizing activity after a single administration compared with the CO-prM-E SAM (CNE) vaccine. However, the dose of DNA vaccine used was more than 50-fold higher than the dose of the SAM vaccine (4 mg versus 75 μg, respectively) and delivered using different methods and, therefore, is not an equal comparison. Clinical trials of the ZIKV DNA vaccine expressing the prM-E proteins have used doses between 2 and 8 mg (, ), and the ability of the SAM (CNE) vaccine to produce protection similar to that of the DNA vaccine in preclinical models suggests an opportunity for dose reduction. In general, more than 200 times the amount of SAM RNA can be produced from a single DNA template. This would be a substantial advantage for manufacturing, particularly for vaccines for emerging and reemerging pathogens where a large number of vaccine doses are rapidly needed for immediate use or stockpiling. SAM RNA may be effective at lower doses for a number of reasons. Whereas DNA needs to enter the nucleus, RNA only needs to enter the cytoplasm to be transcribed (–). The SAM replicon is designed for enhanced amplification of RNA transcripts from which antigen can be expressed and produces dsRNA, which provides a self-adjuvanting effect through activation of pattern recognition receptors (e.g., TLR3, RIG-I, and MDA-5) likely contributing to the efficacy of the SAM vaccine. Last, SAM vaccines can be administered via needle and syringe, further simplifying its implementation in a rapid response to emerging outbreaks.

The SAM technology is uniquely poised to be an effective vaccine manufacturing and delivery platform for emerging infectious diseases such as ZIKV where a rapid response is required. SAM vaccines have now been shown to be effective against many vaccine targets (–), yet the properties of the drug substance and manufacturing approach are the same regardless of antigen, which facilitates process development and the regulatory pathway (). The completely synthetic nature of the platform permits the rapid scale up and production of vaccine upon demand, with the absence of cellular contamination. The ease of gene synthesis allows rapid evaluation of multiple vaccine design concepts, which is particularly important for emerging pathogens that have not been previously studied in depth. The synthetic delivery component of this technology also has the advantage of circumventing antibody responses to vector structural proteins commonly seen with viral vectors, while still eliciting antigen-specific immune responses with similar magnitude and composition. Consistent with previous reports, immune responses to vector nonstructural proteins, if any, do not affect antibody responses or the boosting effect on antibody responses by subsequence immunizations in this study (–). The self-amplifying nature produces high and prolonged antigen expression as well as adjuvanting effect, giving rise to potent antigen-specific immune responses (). The ability to produce the vaccine antigen endogenously within a cell can also elicit robust T cell responses in addition to B cell responses, as demonstrated in immunogenicity animal models for vaccines to cytomegalovirus, influenza, and human immunodeficiency virus (HIV) (–). CNE provides an effective nonviral delivery system capable of inducing immune responses comparable to adjuvanted subunit vaccines and viral vectors. CNE has demonstrated a long storage life (i.e., over 3 years) at 4°C, and its bedside mixing formulation enables the stockpiling of the raw materials of both CNE and RNA, a key feature of the rapid response (). Although human data on CNE are still pending, it has demonstrated efficacy in multiple preclinical models (–). The data shown here demonstrate that a ZIKV SAM vaccine delivered by CNE is remarkably effective in animals, including NHPs. Clinical studies will determine whether this preclinical result can translate to efficacy in humans and whether the SAM vaccine approach can protect women of child-bearing age, and in turn fetuses, during a ZIKV outbreak.

MATERIALS AND METHODS

DNA vector constructs

The antigen DNA sequences of ZIKV vaccine constructs in this study were synthesized by GenScript (Piscataway, NJ) and cloned into a SAM vector backbone (). Codon optimization was carried out by GenScript for expression in human cells. The antigen sequences of WT-prM-E, CO-prM-E, CO-prM-E.ESS.1, CO-prM-E.ESS.2, CO-C-prM-E.1, CO-C-prM-E.2, and CO-C-prM-E.3 were based on the Natal strain of ZIKV (GenBank accession no. KU527068.1). VRC5283 and VRC5288 sequences were based on the H/PF/2013 French Polynesian ZIKV isolate (GenBank accession no. AHZ13508.1), and VRC5288 E transmembrane domain is based on the JEV strain SA(V) (GenBank accession no. BAA14218.1) (). The C-prM-E amino acid sequences of the Natal and H/PF/2013 strains are identical. In the DNA vaccine version of VRC5283, the antigen coding sequences were cloned into the mammalian expression vector VRC8400 (–) and under the control of the CMV immediate early promoter ().

RNA synthesis

In vitro RNA transcription was performed as previously described (). Briefly, DNA plasmids encoding the RNA vaccines were linearized by restriction digestion with Bsp QI at the precise 3′ end of SAM sequences and purified by phenol-chloroform extraction. Linearized DNA templates were in vitro transcribed using T7 RNA polymerase (MEGAscript Kit T7; Ambion AM1333) and purified by LiCl precipitation. RNAs were capped using a vaccinia capping kit (ScriptCap m7G; Epicentre Biotechnologies SCCE 0610), purified again by LiCl precipitation, and dissolved in nuclease-free water (AM9937, Ambion). The quality and integrity of the RNA were analyzed by 1% agarose gel electrophoresis. SAM RNA was stored in the liquid form at −80°C. Under these conditions, SAM RNA is stable for at least 1 year.

RNA potency assay

To determine the efficiency of SAM to launch the self-amplification cycle, approximately 1.0 × 106 BHK cells were electroporated (one pulse, 120 V, 25 ms) with 0.1 μg of SAM RNA, and 4.0 μg of mouse thymus carry RNA (Clontech Laboratories). Cells were seeded in six-well plates with 1% fetal bovine serum (FBS) growth media and incubated at 37°C and 5% CO2 for 16 to 18 hours. Cells were collected, fixed, and permeabilized with Cytofix/Cytoperm (BD Biosciences) and then stained with allophycocyanin (APC)–conjugated anti-dsRNA antibody (J2 monoclonal IgG2a antibody; Bioclass). APC conjugation on anti-dsRNA IgG2a was carried out using the Zenon Allophycocyanin Labeling Kit (Invitrogen). Percentage of dsRNA-positive cells was measured by flow cytometry on an LSR II flow cytometer (BD Biosciences).

Immunoblot analysis

Approximately 1.0 × 106 BHK cells were electroporated (one pulse, 120 V, 25 ms) with 4.0 μg of SAM RNA and seeded in six-well plates with 1% FBS growth media. At 24 hours after electroporation, cells were lysed under nonreducing conditions and whole-cell lysates subjected to SDS–polyacrylamide gel electrophoresis and blotted to a polyvinylidene difluoride membrane. ZIKV E protein and β-tubulin were detected with the pan-flavivirus E–specific monoclonal antibody 4G2 (GeneTex) at a dilution of 1:120 and with anti-tubulin rabbit antibody (Abcam) at a dilution of 1:500, respectively, followed by IRDye 800CW–conjugated donkey anti-mouse IgG secondary antibody (LI-COR) at a 1:30,000 dilution. Protein bands were visualized on the Odyssey fluorescent imager (LI-COR).

CNE formulation preparation

CNE was prepared as described previously (, ). Briefly, squalene, 1,2-dioleoyloxy-3-(trimethylammonium)propane (DOTAP), and sorbitan trioleate were combined and heated to 37°C. The resulting oil phase was then combined with an aqueous phase consisting of polysorbate 80 in 10 mM citrate buffer at pH 6.5. The aqueous phase and oil phase were mixed for 15 min and then homogenized using a rotor-stator premixer to produce a “coarse emulsion” for 45 min. The homogenizer jacket was precooled using chilled water (2° to 8°C), and operations were performed under a nitrogen overlay to prevent squalene oxidation. The coarse emulsion was then passed through an M110EH-30 Microfluidizer (Microfluidics, Newton, MA) at a homogenization pressure of 100,000 Kpa seven times. The resulting fine emulsion was equilibrated at room temperature overnight before low-bioburden filtration and sterile filtration. The final theoretical concentrations of squalene, DOTAP, sorbitan trioleate, and polysorbate 80 were 39, 4.0, 4.7, and 4.7 mg/ml, respectively. The formulation was stored at 2° to 8°C, a condition where CNE stability has been demonstrated to be at least 3 years.

RNA/CNE complexation

To generate SAM (CNE) vaccine materials for immunization in animal studies, the SAM RNA was complexed with CNE at different N:P ratios. The N:P ratio was determined by calculating the number of moles of protonatable nitrogens per milliliter of CNE and calculating the number of moles of phosphate present in the SAM molecule, assuming a constant of 3 nmol of phosphate per microgram of the RNA. To obtain a desired N:P ratio, the RNA was first prepared at an appropriate concentration, added to an equal volume of CNE, mixed, and then allowed to complex on ice for 30 to 120 min. The RNA was complexed to CNE at the N:P ratio of 6.3:1 unless indicated otherwise. To obtain the N:P ratio of 6.3:1, the RNA was prepared at 300 μg/ml, mixed with an equal volume of CNE at bedside, and formulations were diluted to the desired dosing concentrations before administration.

Animal experiments

BALB/c mice were obtained from the Jackson laboratory and housed in the animal facility of the Vaccine Research Center, National Institute of Allergy and Infectious Diseases (NIAID), National Institutes of Health (NIH), Bethesda, MD. Rhesus macaques (Macaca mulatta) were housed and all experiments were performed at Bioqual Inc. (Rockville, MD). All animal experiments were reviewed and approved by the Institutional Animal Care and Use Committee of the Vaccine Research Center (VRC), NIAID, NIH. All animals were housed and cared for in accordance with the local, state, federal, and institutional policies in an American Association for Accreditation of Laboratory Animal Care–accredited facility at the NIH or Bioqual Inc.

Vaccination and challenge of mice

All dosing information for SAM RNA in these studies refers to the amount of RNA injected. Eight- to 12-week-old BALB/c mice (10 per group) were immunized by intramuscular injection of 1.5 or 15 μg of SAM (CNE) vaccine at days 0 and 21. For DNA vaccination, mice were injected with 50 μg of VRC5283 DNA followed by electroporation at days 0 and 21 (BTX AgilePulse, Holliston, MA). Sera were collected at day 0 before the first immunization, day 14, and day 35, binding antibody titers were determined by enzyme-linked immunosorbent assay (ELISA), and neutralizing antibody titers were determined by RVP neutralization assay as described below. At day 49, mice were challenged with 100 FFUs of Puerto Rican ZIKV strain PRVABC59 by intraperitoneal administration. At day 3 after challenge, sera were collected, and viral loads were determined by qRT-PCR analysis.

Vaccination and challenge of NHPs

Rhesus macaques (eight animals per group) were randomized by sex, age, and body weight and immunized by intramuscular injection with 4 mg of VRC5283 DNA vaccine using a needle-free PharmaJet device (Golden, CO) or with 75 μg of SAM vaccine or PBS control by needle and syringe at days 0 and 28. Immunized animals were challenged by subcutaneous injection with 1000 FFUs of ZIKV PRVABC59 at day 56. Blood samples were collected weekly for analysis of total binding antibody titers by ELISA and neutralizing antibody titers by RVP neutralization assay. Blood samples were also collected on days 3, 4, 5 and 7 after challenge for determination of viral load by qRT-PCR analysis.

Particle-based anti-ZIKV total binding antibody ELISA

Partially purified ZIKV SVPs produced by transfection of plasmid VRC5283 in HEK293F cells (2 μg/ml) were added to 96-well Nunc MaxiSorp plates and incubated at 4°C overnight. Serial dilutions of sera from immunized animals were added to the plates and incubated at room temperature for 1 hour. After washing, horseradish peroxidase (HRP)–conjugated goat anti-mouse IgG, Fcγ specific (Jackson ImmunoResearch Laboratories), was added to the plate, incubated at room temperature for 1 hour, and washed with PBS with Tween 20. The assay was developed using 3,3′,5′,5-tetramethylbenzidine HRP substrate (KPL, MD, USA), stopped by the addition of 0.5 M H2SO4, and then measured at 450 nm (SpectraMax Plus 384, Molecular Devices, CA, USA). The limit of confidence (LOC) is a reciprocal end point dilution of 30. Any values below the LOC were set at 1/2 LOC.

Neutralizing antibody assay

ZIKV RVP neutralization assays were performed as previously described (). Briefly, RVPs were produced by complementation of a green fluorescent protein (GFP)–expressing West Nile virus lineage II subgenomic replicon with the ZIKV C-prM-E structural genes from strain H/PF/2013 (). RVPs were titered and then sufficiently diluted in neutralization assays to ensure antibody excess at informative points on the dose-response curves. The diluted RVPs were incubated with serial dilutions of mouse or macaque sera in technical duplicates at 37°C for 1 hour to allow for steady-state binding, followed by infection of Raji cells expressing the flavivirus attachment factor DC-SIGNR (Raji-DCSIGNR). The number of infected GFP-positive cells was determined by flow cytometry 24 to 48 hours later. The resulting data were analyzed by nonlinear regression analysis with a variable slope to determine the dilution of sera required for half-maximal neutralization of infection (EC50 titer) (Prism 6 software; GraphPad). The initial dilution of sera (1:60; based on the final volume of RVPs, cells, and sera) was set as the LOC. Neutralization titers predicted as <60 were reported as a titer of 30 (1/2 the LOC). All measurements represent the average of at least two independent experiments.

Quantitative real-time PCR

ZIKV viral RNA was extracted from 140 μl of NHP serum using the Qiagen QIAamp Viral RNA Kit, following the manufacturer’s protocol, and eluted in 33 μl of water. ZIKV viral RNA was extracted from 50 μl of mouse serum using Agencourt RNAdvance Blood Kit and an automated liquid handler, and also eluted in 33 μl of water. qRT-PCR was performed using three technical replicates of each sample. A standard curve for the qRT-PCR assay was generated using serial dilutions of the inoculation stock that had been titered via plaque assay. Standard curves were used to quantitate the viral load in FFU equivalents per milliliter. LOC was determined empirically to be 0.5 FFU equivalents/ml for mouse sera and 0.05 FFU equivalents/ml for NHP sera. For the NHP qPCR, we detected virus in 9 of 12 replicates across six plates at 0.05 FFU equivalents/ml. For a 20-μl qRT-PCR, 10 μl of RNA input volume, TaqMan Fast Virus 1-Step Master Mix, and the following primers and probe (IDT, IL, USA) were used: forward primer, GGAAAAAAGAGGCTATGGAAATAATAAAG; reverse primer, CTCCTTCCTAGCATTGATTATTCTCA; and FAM probe, AGTTCAAGAAAGATCTGGCTG. Cycling conditions were 50°C for 5 min, 95°C for 20 s, and then 45 cycles of 95°C for 5 s followed by 60°C for 30 s. The PCR was carried out using a Bio-Rad CFX96 Real-Time System.

Statistical analyses

All data were graphed, and statistics were performed on log10-transformed data. Differences among the groups were determined using one-way or two-way analysis of variance (ANOVA) with a Tukey’s multiple comparison test, as indicated.