Quantity and quality
The Zika virus epidemic prompted rapid development of several vaccines. Maciejewski et al. tested two similar DNA vaccines in nonhuman primates and found that although both induced neutralizing antibodies, only one was protective against a viral challenge. Further testing with samples from animals or from clinical trials with these vaccines revealed an important aspect regarding neutralizing antibody quality. The protective vaccine generated relatively more antibodies that could bind the mature form of Zika virions, whereas antibodies induced by the other vaccine were more sensitive to virion maturation state. Passive transfer experiments with sera normalized for mature virion neutralization showed equivalent protection between the two vaccines. These results highlight how antibody quality plays a major role in vaccine efficacy. Furthermore, thinking beyond conventional assays for antibody responses may lead to identification of correlates of protection.
Abstract
The emergence of Zika virus (ZIKV) in the Americas stimulated the development of multiple ZIKV vaccine candidates. We previously developed two related DNA vaccine candidates encoding ZIKV structural proteins that were immunogenic in animal models and humans. We sought to identify neutralizing antibody (NAb) properties induced by each vaccine that correlated with protection in nonhuman primates (NHPs). Despite eliciting equivalent NAb titers in NHPs, these vaccines were not equally protective. The transfer of equivalent titers of vaccine-elicited NAb into AG129 mice also revealed nonequivalent protection, indicating qualitative differences among antibodies (Abs) elicited by these vaccines. Both vaccines elicited Abs with similar binding titers against envelope protein monomers and those incorporated into virus-like particles, as well as a comparable capacity to orchestrate phagocytosis. Functional analysis of vaccine-elicited NAbs from NHPs and humans revealed a capacity to neutralize the structurally mature form of the ZIKV virion that varied in magnitude among vaccine candidates. Conversely, sensitivity to the virion maturation state was not a characteristic of NAbs induced by natural or experimental infection. Passive transfer experiments in mice revealed that neutralization of mature ZIKV virions more accurately predicts protection from ZIKV infection. These findings demonstrate that NAb correlates of protection may differ among vaccine antigens when assayed using standard neutralization platforms and suggest that measurements of Ab quality, including the capacity to neutralize mature virions, will be critical for defining correlates of ZIKV vaccine-induced immunity.
INTRODUCTION
Zika virus (ZIKV) is a mosquito-borne flavivirus that emerged in South and Central America in 2015 and is responsible for more than 240,000 confirmed human infections and considerable morbidity (Pan American Health Organization Zika Cumulative Cases; October 2019). ZIKV was first isolated in 1947 and, despite widespread transmission in Africa and Asia, was not linked to substantial human disease until outbreaks on the Pacific island of Yap in 2007 and French Polynesia in 2013 [reviewed in (1)]. Symptomatic ZIKV infections are uncommon but are typically characterized by mild fever, rash, headache, joint and muscle pain, and conjunctivitis (2, 3). Recent outbreaks identified new features of ZIKV pathogenesis, including associations with Guillain-Barré syndrome (4, 5), an ability to be transmitted sexually [reviewed in (6)], and a spectrum of congenital neurodevelopmental diseases, including microcephaly (7). A study across the United States’ territories found that congenital disease occurred in roughly 5% of women infected while pregnant (8). Because of its rapid spread and serious clinical manifestations, ZIKV was declared a public health emergency of international concern by the World Health Organization in February 2016.
Flaviviruses are small spherical virus particles composed of three structural viral proteins [capsid (C), envelope (E), and membrane (M)], a lipid membrane, and an ~11-kb single-stranded RNA genome of positive-sense polarity encoding a total of 10 structural and nonstructural proteins. Newly formed, noninfectious, immature virions incorporate the premembrane (prM) protein, which interacts with E to form heterotrimeric spikes (9). During viral egress, prM is cleaved by the cellular serine protease furin to create the small, viral membrane–anchored M protein present on mature viruses (10). The structure of the mature infectious form of the virion has been solved for several flaviviruses at high resolution, including ZIKV (11, 12). The relatively smooth surface of the mature virion is a highly ordered herringbone array of 90 antiparallel E protein homodimers oriented parallel to the viral membrane. The E proteins are anchored into the viral membrane by a helical stem and two antiparallel transmembrane domains and closely associate with the underlying M protein. E proteins play a critical role in viral entry and membrane fusion steps of the replication cycle and are the principal target of neutralizing antibodies (NAbs) after infection and vaccination [reviewed in (13)]. Studies of ZIKV and related flaviviruses have identified NAbs that bind quaternary surfaces of the mature virion composed of multiple E proteins (14–16). Antibodies (Abs) with these characteristics bind poorly to monomeric forms of the E protein and have been hypothesized to play a considerable role in protection from infection (17–19). However, virion maturation may be inefficient, resulting in the release of partially mature infectious virions on which E proteins exist as dimers or as trimers complexed with noncleaved prM. The extent of prM cleavage directly controls sensitivity to neutralization by many classes of Abs [reviewed in (20)] and raises the potential for viral conformational evasion of NAbs.
The rapid spread and unique pathogenic features of ZIKV prompted the development of multiple vaccine platforms, including live-attenuated viruses, inactivated ZIKV, heterologous viral vector expression systems, and synthetic nucleic acids [reviewed in (21)]. Expression of prM-E is sufficient for the generation and secretion of antigenic, noninfectious, subviral particles (SVPs) (22, 23), which have been previously shown to be immunogenic and protective in preclinical studies (24–31). DNA vaccines that encode flavivirus prM-E have proven to be immunogenic in clinical studies (32–34). Using this technology, we developed two DNA vaccine candidates (VRC5288 and VRC5283) that express ZIKV prM-E proteins (35). Both constructs encode prM-E from a French Polynesian ZIKV strain (ZIKV H/PF/2013) downstream of the prM signal sequence from Japanese encephalitis virus (JEV) (fig. S1), a modification suggested to increase protein expression in vitro (24). VRC5288 was further modified by substituting 38 amino acids at the C terminus of the E stem and transmembrane regions with the corresponding sequence of JEV. This modification was shown to promote more efficient release of ZIKV SVPs from transfected cells, in agreement with prior studies with dengue virus (DENV) (35, 36). VRC5288 and VRC5283 were shown to be immunogenic in mice and nonhuman primates (NHPs) (35). Clinical evaluation of both constructs confirmed their immunogenicity in humans (37) and a multicenter placebo-controlled randomized phase 2/2b clinical evaluation of VRC5283 is under way.
Humoral immunity is a critical feature of immunity to flaviviruses. The passive transfer of immune sera has been shown in multiple contexts to be sufficient for protection from flavivirus challenge (38–40). Abs contribute to protection via direct neutralization of infection. NAb activity has been established as a correlate of protection for several flavivirus vaccines (41–44). Studies with multiple ZIKV vaccine candidates demonstrated that NAb levels correlate with protection after viral challenge (35, 45–47). However, immune effector functions coordinated by invariant regions of the Ab heavy chain may also contribute to protection after virus infection or vaccination [reviewed in (48)]. In accordance, prior studies have shown that non- or poorly NAbs, including those Abs that do not directly recognize the virus particle, could provide protection in animal models of flavivirus infection (49–51). Recent studies of the ZIKV prM-E DNA vaccine GLS-5700 revealed protection after challenge, although this vaccine elicited E protein–binding Ab with low or undetectable neutralizing activity (29, 52). Further, a vaccine using nonstructural protein 1 (NS1) antigens incapable of eliciting NAbs was shown to be protective in animal models (53). In the current study, we sought to identify properties of vaccine-elicited Abs associated with protection from ZIKV infection using the candidate vaccines VRC5288 and VRC5283.
RESULTS
Vaccination with VRC5288 and VRC5283 confers unequal protection against ZIKV
DNA vaccines VRC5288 and VRC5283 were administered intramuscularly to rhesus macaques in two doses, 4 weeks apart using a needle-free delivery device. At weeks 0 and 4, groups of four animals received immunizations of 1, 0.3, or 0.1 mg of vaccine or a 1 mg dose of the empty parental control plasmid VRC8400 (Fig. 1A). The NAb response was measured at 0, 6, and 8 weeks after initial immunization using a reporter virus particle (RVP) ZIKV neutralization assay (Fig. 1, B to D, and fig. S2A). Both vaccines were similarly immunogenic for all three vaccine doses when compared separately at weeks 0, 6, and 8 using the dilution of serum required to inhibit infection by 50% (EC50) end point. Identical conclusions were reached when neutralization activity at week 8 was measured using a different cellular substrate or assay format (fig. S2, B to D).
(A) Zika virus (ZIKV) DNA vaccine study schedule for nonhuman primates (NHPs) is depicted. Animals (four per group) were immunized intramuscularly with VRC5283 or VRC5288 at a 1, 0.3, or 0.1 mg dose or mock-immunized with VRC8400 (control) at a 1 mg dose at week 0 (black arrow). The NHPs were boosted at week 4 (black arrow) with the same dose given at week 0 and challenged subcutaneously with ZIKV PRVABC59 at week 8 (white arrow). Sera samples collected at weeks 0, 6, 8, 10, and 12 were used in neutralization assays with ZIKV H/PF/2013 RVPs. Representative neutralization dose-response curves at weeks 0, 6, and 8 are shown for individual animals immunized with either (B) VRC5288 or (C) VRC5283. Each dot and associated error bar indicate the mean and range of technical duplicates, respectively. (D) Dilution of serum required to inhibit infection by 50% (EC50) was estimated from dose-response curves using nonlinear regression analysis. Each dot represents the mean EC50 neutralizing antibody (NAb) titer of two to five independent measurements for an individual animal, with each experiment performed in technical duplicates. The time of challenge (week 8) is denoted by a white arrow. The dotted line represents the limit of detection (LOD) for the assay. Any independent measurement below the LOD was assigned a value of half the LOD. Bar graphs and error bars indicate the geometric mean and geometric SD for each group. The mean EC50 NAb titers of VRC5288 and VRC5283 were compared within the same dose group and time point via a Kruskal-Wallis test, followed by a Dunn’s multiple comparisons test. Statistical analyses were nonsignificant (P ≥ 0.05). (E to K) NHPs were challenged with 103 FFU of ZIKV PRVABC59 at week 8 after vaccination. Viral loads (genome copies/ml) were measured up to 14 days after challenge by qRT-PCR. The dotted line represents the LOD (50 genome copies/ml). Any independent measurement below the LOD was assigned a value of half the LOD.
To evaluate VRC5288- and VRC5283-mediated protection, we subcutaneously challenged the animals at week 8 with 103 focus-forming units (FFU) of a Puerto Rican strain of ZIKV (PRVABC59). Viral load was measured by quantitative reverse transcription polymerase chain reaction (qRT-PCR) daily from 0 to 5 days and on days 7, 10, and 14 after challenge (Fig. 1, E to K). Although all 12 animals vaccinated with VRC5283 were protected from viremia, one or two animals in each VRC5288 dose group became viremic on at least 1 day after challenge. By comparison, all four animals in the VRC8400 control group had detectable viremia by day 3 (Fig. 1E). These results demonstrate that marked vaccine-mediated protection is possible at vaccine doses considerably lower than previously reported (35). To investigate whether vaccination conferred sterilizing protection, we measured NAb titers at 2 and 4 weeks after challenge (weeks 10 and 12, respectively) to detect an increase in titer as compared to week 8 (fig. S3). A ≥4-fold increase in EC50 NAb titers after challenge was defined as an anamnestic response, as described previously (54), and was detected in 9 of 12 animals that received VRC5288 (10- to 57-fold increase in NAb activity at 2 weeks after challenge), including all animals with detectable viremia. Three of 12 VRC5283-vaccinated animals (one from the 0.3 mg dose group and two from the 0.1 mg dose group) had a 38-, 10-, or 4-fold rise in NAb activity 2 weeks after challenge.
The complete protection against viremia conferred by VRC5283 at all tested doses prevented an estimate of a NAb correlate of protection for this vaccine candidate. To achieve breakthrough infections in vaccinated animals that were required for this analysis, we performed a second study consisting of two groups (six animals per group) in which two doses of 0.03 mg or 0.02 mg of VRC5283 were administered intramuscularly at 4-week intervals using a needle-free delivery device (Fig. 2). Vaccination was followed by challenge with 103 FFU of ZIKV PRVABC59 at week 8. The average EC50 NAb titer at the time of challenge was 835.9 and 250.1 for the 0.03 and 0.02 mg groups, respectively (Fig. 2A). Viral load was measured by qRT-PCR up to 14 days after challenge (Fig. 2, B to D). Although none of the animals receiving the 0.03 mg regimen developed detectable viremia (Fig. 2C), four of six NHPs receiving two doses of 0.02 mg of VRC5283 had detectable viremia on at least 1 day after challenge (Fig. 2D). Neutralization studies of postchallenge sera (Fig. 2, E to G) suggested that five animals in the 0.02 mg dose group, three of which were viremic, experienced an anamnestic humoral response (27- to 67-fold increase in NAb activity at 2 weeks after challenge) (Fig. 2G).
(A) NHPs (six per group) were immunized with VRC8400 at a dose of 0.03 mg or VRC5283 at a dose of 0.03 or 0.02 mg, following the same schedule depicted in Fig. 1A. At the indicated time points, EC50 NAb titers were calculated from the dose-response curves for all animals using nonlinear regression analysis. Each dot represents the average titer of two to three independent measurements for each animal, with each experiment performed in duplicate technical replicates. The time of challenge (week 8) is denoted by a white arrow. The dotted line represents the LOD for the assay. Any independent measurement below the LOD was assigned a value of half the LOD. Bar graphs and error bars indicate the geometric mean and geometric SD for each group. NAb titers measured after challenge (week 8) are represented with hatched gray bars. Mean EC50 NAb titers of the VRC5283 0.03 and 0.02 mg dose groups were compared pairwise within each time point via a Kruskal-Wallis test, followed by a Dunn’s multiple comparisons test. Statistical analyses were nonsignificant (P ≥ 0.05). (B to D) Animals were challenged with 103 FFU of ZIKV PRVABC59 at week 8. Viral loads (genome copies/ml) were measured up to 14 days after challenge by qRT-PCR. The dotted line represents the LOD (50 genome copies/ml). Any independent measurement below the LOD was assigned a value of half the LOD. (E to G) Anamnestic humoral response was measured for each animal by calculating the fold change in mean EC50 NAb titers between the time of challenge (week 8; denoted here as week 0) and at 2 or 4 weeks after challenge (weeks 10 and 12, respectively). The dotted line represents a ≥4-fold change, signifying an anamnestic response after infection. (H) Probability of becoming viremic after challenge was modeled as a function of EC50 NAb titers of the vaccinated NHPs at week 8 (time of challenge) using a probit regression model. EC50 NAb titers from animals that did not have detectable viremia were assigned a value of 0 (nonviremic) and those that did were assigned a value of 1 (viremic). NHP EC50 NAb titers for animals vaccinated with VRC5288 are denoted in red, and those vaccinated with VRC5283 are denoted in blue.
These studies revealed an unexpected contrast in the capacity of two highly related vaccines to protect against infection despite similar NAb titers before challenge. A neutralization titer that clearly distinguished protected from nonprotected animals was not observed, in agreement with prior studies with other vaccine candidates (45, 46). In animals receiving the VRC5288 vaccine, we estimated a reciprocal EC50 NAb titer of 1661 as sufficient to reduce the probability of infection by 70% (Fig. 2H). By comparison, the NAb titer required to achieve comparable protection was estimated to be roughly sixfold lower in animals receiving VRC5283 (reciprocal EC50 NAb titer of 272). This abundance of NAb is similar to that observed in a recent phase 1 clinical study of VRC5283, in which the vaccine was administered in three doses of 4 mg at 4-week intervals (37). The notably different amount of NAb required to achieve protection after immunization with similar constructs suggests that qualitative aspects of neutralizing activity need to be examined in greater depth to define a mechanistic correlate of protection that would be relevant across multiple vaccine platforms.
Passive transfer of vaccine-immune sera identifies a role for Ab quality in protection
To explore the role of Abs in the protection conferred by vaccination with VRC5288 and VRC5283, vaccination and passive transfer studies in a highly sensitive, lethal mouse model of ZIKV infection were performed (55). To evaluate direct protection, we intramuscularly immunized AG129 mice with two doses of 50 μg of VRC5288 or VRC5283 vaccine administered 3 weeks apart (10 animals per group); we used VRC8400 as a negative control and administered on the same schedule. At week 5 after initial vaccination, sera were collected to measure NAb titers. Both vaccines elicited robust NAb responses, with mean reciprocal EC50 titers of 3676 and 7982 for VRC5288 and VRC5283 vaccines, respectively (Fig. 3A). Animals were thereafter challenged subcutaneously at week 6 with 100 FFU of ZIKV PRVABC59. In contrast to animals receiving the control VRC8400 vector, viral RNA was detected in only two animals in the VRC5288-immunized group and none of the VRC5283-immunized group (Fig. 3B). All vaccinated animals survived ZIKV challenge (Fig. 3C).
(A to C) AG129 mice (10 per group) were immunized with 50 μg of VRC8400, VRC5288, or VRC5283 at weeks 0 and 3, followed by challenge with 100 FFU of ZIKV PRVABC59 at week 6. (A) EC50 NAb titers were measured at week 5. Each dot represents the average EC50 NAb titer for each mouse serum sample measured independently two to four times. Results for three of the mice immunized with VRC5288 are not depicted, as nonlinear regression analysis from several independent experiments could not predict the EC50 NAb titer with confidence. (B) Viral loads were measured by qRT-PCR at 5 days after challenge. (C) Survival of the mice was assessed up to 30 days after challenge. (D to F) Antibody passive transfer studies were performed in AG129 mice (8 to 10 per group), followed by challenge with 100 FFU of ZIKV PRVABC59. IgG was purified from pooled sera collected from VRC5288- or VRC5283-immunized C57BL/6 mice (25 per group) or convalescent ZIKV–infected BALB/c mice (n = 60) for passive transfer studies. The antibodies were processed to remove endotoxin before passive transfer studies. Control IgG is commercially available, endotoxin-free purified mouse IgG. Concentrations were normalized to a reciprocal EC50 NAb titer of 4000 for vaccine-immune IgG and 1000 for convalescent IgG and then administered intraperitoneally 1 day before challenge. (D) Sera were collected 1 day after passive transfer of antibodies to measure in vitro neutralization. Each dot represents the average EC50 NAb titer of one to three independent measurements for each animal. (E) Viral loads were measured by qRT-PCR at day 5 after challenge. (F) Survival of the mice was assessed up to 36 days after challenge. For graphs depicting NAb titers (A and D), horizontal lines and error bars indicate the EC50 geometric mean and geometric SD for each group, respectively. For graphs depicting viral load (B and E), horizontal lines and error bars indicate the mean and SD for each group, respectively. The dotted line represents the LOD for the assay. Any independent measurement below the LOD was assigned a value of half the LOD. Statistics were calculated via a Kruskal-Wallis test, followed by a Dunn’s multiple comparisons test (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).
To define the role of Ab-mediated protection in this model, we purified immunoglobulin G (IgG) from C57BL/6 mice vaccinated with VRC5288 or VRC5283 and convalescent ZIKV infection–immune BALB/c mice as a positive control. The concentration of vaccine-elicited NAb was normalized for activity (EC50 NAb titer of 4000) and then passively transferred into AG129 mice in groups of 10. Purified IgG from convalescent controls was administered to achieve an EC50 NAb titer of 1000 (eight animals per group). Serum neutralizing titers were confirmed in individual animals 1 day after transfer (Fig. 3D), followed by challenge with 100 FFU of ZIKV PRVABC59. Analysis of viral RNA on day 5 revealed significantly reduced viremia in animals receiving VRC5283 IgG (P = 0.0046), whereas VRC5288 IgG conferred little to no protection as compared to control animals that received commercially available, non-ZIKV–specific, purified mouse IgG (P > 0.99) (Fig. 3E). Passively transferred Abs obtained from VRC5288-immunized animals were less protective than VRC5283-elicited Abs, despite nearly identical measures of NAb activity. Convalescent ZIKV–immune sera were considerably more protective than Abs elicited by vaccination, as demonstrated by greater survival rates after challenge (Fig. 3F), although we cannot rule out a contribution of NS1-reactive Abs in this context. These studies identify a critical role for NAb quality in defining the protective capacity of vaccine-immune sera.
VRC5288 and VRC5283 Abs are characterized by a similar capacity for E protein binding and effector function
The neutralization potential of Abs is markedly influenced by the display of their cognate epitopes on infectious virions. Flavivirus Abs frequently fail to neutralize infection because they bind E protein epitopes displayed infrequently on intact virions. Alternatively, some Abs recognize quaternary epitopes present only on intact virus particles. To investigate the discordance between the NAb abundance and the degree of protection in vaccinated NHPs, we measured serum Ab binding to a monomeric soluble form of E and ZIKV SVPs using an enzyme-linked immunosorbent assay (ELISA) (Fig. 4, A and B). Both VRC5283- and VRC5288-elicited Abs bound soluble E and SVPs at similar titers (average 0.95- and 0.49-fold difference in binding for VRC5283 and VRC5288, respectively, for all vaccine dose groups) (Fig. 4C). Ab-binding titers measured using SVPs or soluble E were also similar among vaccine groups (P > 0.99). Similar conclusions were reached when soluble E binding was measured using a Luminex assay format (fig. S4). These studies suggest that differences in the ability to elicit Abs with a quaternary mode of recognition are not sufficient to explain the differences in protection conferred by these vaccines.
Total antibody titers from the NHP vaccine sera at week 8 (time of challenge) were measured by ELISA using either concentrated ZIKV subviral particles (SVPs) generated from the VRC5283 construct or polyhistidine-recombinant ZIKV (strain SPH2015) soluble envelope protein (sE). Representative graphs are shown for the ELISA data obtained with (A) VRC5283 1 mg and (B) VRC5288 1 mg sera. (C) End point dilution for the total Ab binding titer measured using ZIKV SVP and sE at de-escalating doses for VRC5288 and VRC5283 NHP sera at week 8 is shown. Each dot represents the mean end point dilution of two independent measurements for an individual animal, with each experiment performed in technical duplicates. The dotted line represents the LOD for the assay. Any independent measurement below the LOD was assigned a value of half the LOD. Bar graphs and error bars indicate the geometric mean and geometric SD for each group. The mean end point dilution measured using the SVP or sE was compared within the same dose group and time point via a Kruskal-Wallis test, followed by a Dunn’s multiple comparisons. Statistical analyses were nonsignificant (P ≥ 0.05). (D) Phagocytic score for antibody-dependent cellular phagocytosis (ADCP) was measured for the NHP sera (four per group) from the 1 mg dose groups for weeks 0, 6, 8, 10, and 12. ZIKV E protein–coated Alexa Fluor 488 beads were incubated with sera for 2 hours, after which the Ab-bead complexes were added to monocytic THP-1 cells and incubated overnight. Alexa Fluor 488+ cells, indicative of ADCP, were detected by flow cytometry. One animal in the VCR5283 group was not measured at weeks 10 and 12. Each point represents the group geometric mean of the animals’ phagocytic score at the respective time point, and error bars indicate the geometric SD. The dotted vertical line represents the time of challenge. Statistics comparing VRC5288 and VRC5283 groups were calculated via two-way analysis of variance (ANOVA), followed by a Sidak’s multiple comparisons test (**P < 0.01). OD450, optical density at 450 nm.
We next evaluated the ability of vaccine-elicited Ab to coordinate phagocytosis of E protein–decorated particles by the monocytic cell line THP-1 (Fig. 4D and fig. S5). Sera collected at the time of challenge from animals receiving the 0.3 and 0.1 mg doses of VRC5288 and VRC5283, but not from unvaccinated controls, supported Ab-dependent cellular phagocytosis (ADCP) at similar magnitude. By comparison to VRC5283, ADCP activity was an average of 0.13-fold greater in samples collected from VRC5288-vaccinated animals at time of challenge for all doses, suggesting that differences in ADCP activity are not sufficient to account for unequal protection by the two vaccines.
Neutralization by vaccine-immune, but not convalescent NAbs, is sensitive to the maturation state of the virion
To characterize differences in the quality of the NAb response to VRC5288 and VRC5283, we next explored the ability of ZIKV-reactive Abs to neutralize virions with varying amounts of uncleaved prM. Prior studies with West Nile virus (WNV) and DENV demonstrated that the prM content of the virion modulates neutralization sensitivity through changes in the availability of otherwise poorly accessible epitopes (56–59). Virion maturation state–sensitive patterns of neutralization can be detected by comparing the sensitivity of the heterogeneous population of virus particles released from transfected cells under standard conditions (Std RVPs) with a more homogeneously mature population of virions produced in the presence of overexpressed human furin (+Furin RVPs) (56). Studies with a ZIKV-specific monoclonal Ab (mAb) that binds the poorly accessible E domain II fusion loop (60) demonstrated differences in the neutralization of ZIKV Std and ZIKV +Furin RVP preparations (fig. S6, A and B); increasing virion maturation markedly reduces the proportion of virions sensitive to neutralization (P = 0.0006). Using ZIKV Std and ZIKV +Furin RVPs, we measured the maturation sensitivity of six human convalescent sera collected between 3 and 7 weeks after symptom onset from natural ZIKV infection and from sera obtained at later time points for two of the individuals (table S1). Neutralization activity was similar for both ZIKV Std and ZIKV +Furin RVPs (average fold change of −0.38) (Fig. 5A). Analysis at a later time post-ZIKV symptom onset did not change this characteristic of the NAb response (fig. S6C). Last, convalescent sera collected from control VRC8400-vaccinated NHPs 2 weeks after challenge with 103 FFU of ZIKV PRVABC59 also equivalently neutralized ZIKV with varying amounts of uncleaved prM (Fig. 5B).
The EC50 NAb titer of human convalescent ZIKV-immune sera and NHP sera collected 2 weeks after ZIKV infection was measured using standard (Std) or mature (+Furin) ZIKV RVP preparations. (A) EC50 NAb titers were measured for ZIKV convalescent human sera collected from volunteers NIH.1 to NIH.6 (described in table S1). (B) NHPs immunized with the control vaccine VRC8400 from the study in Fig. 1 were challenged with 103 FFU of ZIKV PRVABC59 at week 8. The EC50 NAb titers of sera from 2 weeks after challenge (week 10) were measured using either Std or +Furin ZIKV RVP preparations. Bar graphs and error bars indicate the geometric mean and geometric SD, respectively. Each dot represents an independent measurement of EC50 NAb titer for the individual volunteer or NHP. Average fold change in EC50 NAb titers between Std and +Furin ZIKV preparations is denoted above the respective bar graphs. Statistics calculated via Kruskal-Wallis test, followed by a Dunn’s multiple comparisons test, were nonsignificant (P ≥ 0.05). (C) Sera from the NHP vaccine study described in Fig. 1 and a previously published NHP study (35) were used to measure NAb activity pairwise against Std or +Furin ZIKV RVPs at time of challenge (week 8) and 2 weeks after challenge (week 10). Each dot represents the average EC50 NAb titer from one to two independent measurements for an individual animal. The horizontal line and error bars denote the group geometric mean and geometric SD, respectively. The dotted line represents the LOD for the assay. Statistics were calculated via Kruskal-Wallis test, followed by a Dunn’s multiple comparisons test. (D) Sera from human volunteers immunized with VRC5288 via a needle and syringe or VRC5283 via a needle and syringe or needle-free device were assayed in neutralization assays with Std or +Furin ZIKV RVPs at 12 weeks after immunization. Each dot represents the average EC50 NAb titer from two independent measurements for an individual volunteer. The horizontal dotted line represents the LOD for the assay. Any independent measurement below the LOD was assigned a value of half the LOD. Statistics were calculated via Kruskal-Wallis test, followed by a Dunn’s multiple comparisons test (*P < 0.05, **P < 0.01, and ****P < 0.0001).
To determine whether the NAb responses elicited by VRC5288 and VRC5283 vaccination were sensitive to the virion maturation state, we studied the sera from the vaccinated NHPs described in the first experiment (Fig. 1) and those included in a published study of these vaccine candidates (35). Sera collected at 8 weeks after initial vaccination were evaluated in paired neutralization studies with ZIKV Std and ZIKV +Furin RVPs (Fig. 5C and figs. S6, D to F, and S7). In contrast to sera obtained from infected humans or experimentally challenged VRC8400-vaccinated NHPs detailed above, sera from vaccinated NHPs had a significantly reduced capacity to neutralize mature forms of ZIKV relative to those that retained uncleaved prM (P = 0.0058 and P = 0.0145 for VRC5288- and VRC5283-vaccinated animals, respectively) (Fig. 5C). Maturation state differences in NAb sensitivity were greater with VRC5288 as compared to VRC5283 (mean, 9.8-fold and 3.7-fold change, respectively). Similar comparative studies of sera collected 2 weeks after challenge revealed equivalent NAb activity irrespective of the virion maturation state or subsequent protection from challenge, defined by either viremia or an apparent anamnestic response (Fig. 5C and figs. S6, G to I, and S7).
We next evaluated whether human vaccination resulted in a NAb response that was similarly sensitive to the presence of prM on the virion using sera from volunteers of phase 1 clinical studies of VRC5288 and VRC5283 vaccine candidates (37). Samples from volunteers who received three 4 mg doses of vaccine at 4-week intervals (fig. S8A), delivered by either needle and syringe or a needle-free device, were selected. Sera collected 4 weeks after the last vaccination were analyzed by neutralization assay with ZIKV Std and ZIKV +Furin RVPs (Fig. 5D and fig. S8). As observed in vaccinated NHPs, the Ab response of volunteers receiving either DNA vaccine candidate was markedly sensitive to the virion maturation state and had less potent neutralizing activity when assayed against mature virions. NAbs elicited by vaccination with the VRC5288 vaccine were incapable of neutralizing mature forms of the virion in most (78.9%) recipients, despite detectable neutralization using standard preparations of RVPs.
NAb activity against the mature virus particle predicts vaccine-mediated protection
The influence of the virion maturation state on the neutralization activity of vaccine-elicited Abs raised the possibility that NAbs specific to mature forms of the virion may better predict protection from ZIKV infection. To evaluate this, we characterized serum Abs purified from C57BL/6 mice vaccinated with VRC5288 or VRC5283 DNA for their neutralization capacity against ZIKV Std and ZIKV +Furin RVPs. As observed with vaccinated human and NHP sera, purified ZIKV vaccine–induced mouse polyclonal IgG was influenced by the maturation state of the virion (19.7- and 63.4-fold reduced sensitivity against +Furin ZIKV RVPs for VRC5283 and VRC5288, respectively). Ab concentrations were next normalized on the basis of their capacity to neutralize mature forms of the virion (EC50 titer of 300) and passively transferred into AG129 mice; this required transfer of more total IgG from the VRC5288 group than from VRC5283 animals. Equivalent NAb titers against mature ZIKV RVPs were confirmed among animals receiving VRC5288 and VRC5283 IgG 1 day after transfer (Fig. 6A), after which mice were challenged with 100 FFU of ZIKV PRVABC59. ZIKV viremia was measured at day 5 after challenge. Viral burden was reduced similarly among animals receiving VRC5283- and VRC5288-elicited NAbs (Fig. 6B). In contrast to studies with NAbs normalized by activity against standard preparations of ZIKV RVPs (see Fig. 3), transfer of IgG with an equivalent capacity to neutralize mature virions resulted in similar protection for both VRC5283- and VRC5288-elicited NAbs (Fig. 6C). These data suggest that the capacity to neutralize mature forms of the virion better predicts protection from infection.
Antibody passive transfer studies were performed in AG129 mice (10 per group). IgG was purified from pooled sera collected from VRC5288- or VRC5283-immunized C57BL/6 mice and processed to remove endotoxin before passive transfer studies. Control IgG is commercially available, endotoxin-free purified mouse IgG. Using +Furin ZIKV RVPs, VRC5288- and VRC5283-immune IgG was normalized to a reciprocal EC50 NAb titer of 300 and administered intraperitoneally 1 day before challenge. (A) Sera were collected 1 day after passive transfer of Ab to measure in vitro neutralization against +Furin RVPs. Each dot represents the average EC50 NAb titer of two independent measurements for each animal. Animals were challenged with 100 FFU of ZIKV PRVABC59 at 1 day after Ab transfer. (B) Viral loads were measured by qRT-PCR at day 5 after challenge. (C) Survival of the mice was assessed up to 31 days after challenge. In (A), horizontal lines and error bars indicate the geometric mean and geometric SD for each group, respectively. In (B), horizontal lines and error bars indicate the mean and SD for each group, respectively. The dotted lines represent the LOD for the respective assays. Any independent measurement below the LOD was assigned a value of half the LOD. Statistics were calculated via a Kruskal-Wallis test, followed by a Dunn’s multiple comparisons test (**P < 0.01 and ****P < 0.0001).
DISCUSSION
The spread of ZIKV to the Western Hemisphere and its unanticipated linkage to severe congenital disease stimulated the development of multiple vaccine candidates. Despite the rapid pace of development, evaluation of ZIKV vaccines is occurring against the backdrop of a waning epidemic (1). Therefore, defining correlates of immune protection will play a critical role in licensing strategies for ZIKV vaccine candidates. The two DNA vaccine candidates evaluated in our current study have been shown to be immunogenic in mice, NHPs, and humans (35, 37). Using data pooled from NHP immunization studies with both vaccines, we previously estimated that a vaccine-elicited reciprocal EC50 NAb titer of 1000, when measured using a standard ZIKV RVP assay, would provide a 70% probability of protection against ZIKV infection (35). However, this interpretation was limited by the assumption that NAbs elicited by both vaccines contribute equally to protection. In this study, we demonstrated that despite the amino acid sequence similarity of the two DNA vaccines, the level of NAb activity required to achieve protection from viremia in an NHP model by VRC5283 was substantially lower than the level required by vaccination with VRC5288. Equivalent protection after VRC5288 vaccination requires roughly sixfold higher neutralizing activity than VRC5283 when assayed using standard preparations of ZIKV RVPs that retain uncleaved prM.
The protection conferred by the passive transfer of Abs before ZIKV challenge identifies the importance of humoral immunity for protection (38). Multiple characteristics of the Ab response may contribute to protection from infection including the mechanism of neutralization, the epitopes targeted by binding and neutralizing Abs, and Fc-mediated effector functions. Our studies did not identify differences in the ability of Abs elicited by either DNA vaccine to coordinate Fc receptor–dependent phagocytosis. Although further studies are warranted, our results suggest that differences in this Ab effector capacity do not explain the variable vaccine potency. Similarly, comparable titers of Abs capable of binding a soluble form of the E protein or virus-like particles containing E protein were present in NHP sera collected after vaccination with either vaccine candidate. Relationships between the capacity of vaccine-elicited serum Abs to bind (ELISA) or neutralize (EC50 determined by standard RVP neutralization assay) ZIKV were similar between VRC5283 and VRC5288 recipients.
Our neutralization studies suggest that Abs elicited by these two DNA vaccines differ from one another and those in convalescent sera. Our studies do not identify the Ab specificities elicited by vaccination nor infection. Future epitope mapping and serum depletion or competition studies are required to address this limitation. However, the analysis of neutralization studies with structurally diverse forms of the virion was illuminating. Flaviviruses may be structurally heterogeneous due to inefficient cleavage of the prM protein and has been shown for multiple flaviviruses to change the antigenicity of infectious virions. The impact of virion maturation on the antigenic structure of infectious virions and neutralization potency can be studied by manipulating the prM content of ZIKV RVPs by coexpression of furin during virion production. Use of these techniques revealed that the accessibility of epitopes recognized by many Abs, including the immunodominant E protein domain II fusion loop, is reduced on mature forms of the virion, resulting in decreased sensitivity to neutralization (50, 56, 61). We have demonstrated here that Abs elicited by natural or experimental ZIKV infection are largely insensitive to the presence of prM on the virus particle, whereas prM-E vaccine–elicited Abs, to varying degrees, have a markedly reduced capacity to bind mature virions in numbers sufficient to mediate neutralization. In neutralization studies, immune sera from VRC5288-vaccinated animals were more sensitive to the prM content of virions than were Abs elicited by VRC5283. These Abs were also less potent in passive transfer studies in mice when normalized by the ability to neutralize standard preparations of ZIKV RVPs. By comparison, immune sera from both vaccines were similarly protective when standardized for neutralization activity against mature forms of the virion. Abs that bind epitopes displayed on the E protein dimers of virus particles may contribute more to protection than those specific for epitopes presented efficiently only when prM cleavage is incomplete. Although the proportion of infectious ZIKV produced in vivo that retains noncleaved prM is unknown, a recent study of DENV isolated directly from six human individuals demonstrated that mature forms of the virion predominate (62). A capacity to efficiently neutralize mature virions may be an important component of vaccine-elicited protection from ZIKV. Although standard neutralization assay formats revealed immunogenicity of VRC5288 in humans (37), our studies suggest that these circulating Abs have little capacity to neutralize mature ZIKV. Thus, the use of mature forms of ZIKV in neutralization assays will provide a more relevant correlate of vaccine-mediated protection.
Our study also demonstrated that Abs elicited by vaccination with these two DNA vaccine candidates differed from those in convalescent sera. The structure of the ZIKV SVP antigens produced by these DNA vaccines is not yet available at high resolution. Early studies demonstrated that SVPs are heterogeneous in structure and size (63). Electron microscopy data of SVPs from tick-borne encephalitis virus revealed one species of this population to be smaller than infectious virions yet composed of antiparallel dimers like those found on infectious virions (64). Similar findings were reported recently for DENV (65). Experiments with multiple flaviviruses suggest that the antiparallel dimer is the primary antigenic unit of the virion targeted by NAbs [reviewed in (66)]. Although SVPs have been shown in numerous contexts to be immunogenic, our studies suggest that the structure of SVPs produced by VRC5283 and VRC5288 may differ from each other and from infectious virions, which, in turn, may play a role in defining the functional properties or specificity of vaccine-elicited Abs. We did not determine the structural basis of differences in the immune response to VRC5283- and VRC5288-encoded SVP antigens. Although substitutions in the C terminus of VRC5288 from ZIKV to JEV residues were shown to increase SVP release and, when introduced in the context of DENV SVPs, have minimal impact on antigenicity (36), these changes have the potential to affect the structure and conformational dynamics of SVPs in vivo in yet unappreciated ways. Similarly, immunization of mice with two modified mRNA ZIKV prM-E vaccines that differed in the signal sequence of prM also revealed differences in their protective capacity (67). These studies highlight a need for structure-guided antigen design for flavivirus prM-E vaccines and raise the possibility that SVP antigens exploited in numerous vaccine development programs will elicit a different quality of Ab response than inactivated or live-attenuated vaccine approaches that include the mature infectious form of the virus. As our study explored the impact of virion maturation on just two of the many ZIKV vaccine candidates developed to date, additional studies are warranted.
MATERIALS AND METHODS
Study design
The objective of this study was to identify a serological correlate of protection for the ZIKV DNA vaccines VRC5288 and VRC5283 in NHPs. Rhesus macaques were distributed to immunization groups based on age, gender, and weight before injections of vaccines. Animals were challenged with ZIKV and bled for NAb and viral load quantitation. To compare the role of Abs in the protection conferred by vaccination with VRC5288 and VRC5283, vaccination and passive transfer studies were performed in the AG129 lethal mouse model of ZIKV infection. Male and female AG129 mice were randomly assigned to groups, and animals were either directly vaccinated or administered IgG purified from vaccinated or infected mice. Mice were challenged with ZIKV and bled for NAb and viral load quantitation. Sample sizes were chosen empirically to ensure adequate statistical power. Investigators were not blinded with respect to the vaccination status of animals before laboratory studies. ZIKV RVP assays were performed to measure neutralizing activity using virion preparations that varied with respect to the extent of prM cleavage. Each individual sample was measured multiple times as detailed in the legend of each figure. Valid estimates of neutralization potency were defined by assays in which nonlinear regression estimates of the EC50 had an R2 ≥ 0.85 and the 95% confidence range of the estimated EC50 was less than fourfold in either direction. All valid measurements were included in our analysis; no outliers were excluded. Primary data are reported in data file S1.
Ethics statement
This study was carried out in accordance with the recommendations and guidelines of the National Institutes of Health (NIH) Guide to the Care and Use of Laboratory Animals. The protocol was approved by the Animal Care and Use Committee of the Vaccine Research Center (VRC), National Institute of Allergy and Infectious Diseases (NIAID) at the NIH. Mice were housed in a facility fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC). Animal procedures were conducted in strict accordance with all relevant federal and NIH guidelines and regulations.
DNA vector constructs
ZIKV DNA vaccine plasmids VRC5288 and VRC5283 were previously described (35). Both vaccines are based on the ZIKV H/PF/2013 virus isolate (GenBank accession AHZ13508.1). Plasmids encode the ZIKV structural proteins prM and E under the control of the cytomegalovirus immediate early promoter for mammalian cell expression. Both plasmids also contain the prM signal sequence from JEV at the 5′ end of the prM-E sequence. VRC5288 differs from VRC5283 in that the C-terminal 98 amino acids of the E protein, which encode the stem and transmembrane regions, were exchanged with the corresponding JEV sequence, resulting in 38 amino acid differences. The inserts were cloned into the mammalian expression vector VRC8400 (32, 68, 69). The empty plasmid VRC8400 was used as a negative control.
Animal experiments
AG129 mice were obtained from Marshall BioResources. BALB/c and C57BL/6 mice were obtained from the Jackson laboratory. Mice were bred in-house in the animal facility at the VRC, NIAID, NIH, Bethesda, MD, USA. Rhesus macaques (Macaca mulatta) were used in the NHP studies. Macaques were housed, and all experiments were performed at BIOQUAL Inc. All animal experiments were reviewed and approved by the Animal Care and Use Committee of the VRC, NIAID, NIH. All animals were housed and cared for in accordance with local, state, federal, and institutional policies in an AAALAC-accredited facility at the NIH or BIOQUAL Inc.
NHP immunization and challenge
In the first study, 28 rhesus macaques were distributed into groups of four based on age, gender, and weight. VRC5288 and VRC5283 were evaluated at de-escalating doses (1, 0.3, and 0.1 mg), and a seventh group received the control vaccine VRC8400 at 1 mg dose. All animals received split-dose injections of the respective vaccine in each quadricep (for example, two 0.5 mg injections for the 1 mg dose group) at weeks 0 and 4 using a needle-free delivery system (PharmaJet). Animals were challenged at week 8 with 103 FFU of ZIKV PRVABC59 administered subcutaneously. Animals were bled daily up to day 5 and on days 7, 10, and 14 after challenge for viral load quantification. Animals were bled, and serum NAb titers were measured by RVP neutralization assay at weeks 0, 6, 8, 10, and 12.
A second study was performed with VRC5283 in NHPs to allow breakthrough at lower vaccine doses to measure a serological correlate of protection for the vaccine. Eighteen NHPs were evenly distributed into groups of six. The vaccine was evaluated at lower doses (0.03 and 0.02 mg) in two of the groups. The third group received the control vaccine VRC8400 at the 0.03 mg dose. All animals received split-dose injections of their respective vaccine in each quadricep at weeks 0 and 4 using a needle-free delivery system. The animals were then challenged with 103 FFU of ZIKV PRVABC59 administered subcutaneously. Animals were bled daily from days 3 to 5 and on days 7 and 14 after challenge for viral load quantification. Animals were bled, and serum NAb titers were measured by RVP neutralization assay at weeks 0, 6, 8, 10, and 12.
Sera from a prior study evaluating ZIKV DNA vaccine immunogenicity in NHPs at higher doses (35) were included in our analysis of maturation state sensitivity along with sera from the two-dose de-escalation studies detailed above. Briefly, in this published study, 30 rhesus macaques were evenly distributed into groups of six based on age, gender, and weight. VRC5288 and VRC5283 were evaluated at a 4 or 1 mg dose, and VRC8400 was evaluated at a 4 mg dose. All NHPs received two injections at a 4-week interval, with the exception of the animals in the VRC5288 1 mg-dose group, which received only a single injection. Vaccine-immune sera from NHPs that received two doses of VRC5283 or VRC5288 were reanalyzed here for their ability to neutralize standard and +Furin preparations of RVPs in paired experiments.
Vaccination and challenge of mice
Three groups of 10-week-old AG129 mice (10 per group) were immunized with 50 μg of VRC8400, VRC5288, or VRC5283 by intramuscular injection, followed by electroporation (AgilePulse In Vivo Electroporation System, BTX) at weeks 0 and 3. Sera were collected at week 5 to measure the NAb titer. Mice were challenged at week 6 with 100 FFU of ZIKV PRVABC59 subcutaneously in the footpad. Viremia was measured 5 days after challenge by qRT-PCR. Survival was monitored daily up to denoted days after challenge.
Passive transfer of purified immunoglobulin and challenge of mice
Vaccine-immune sera for use in passive transfer studies was pooled from 25 VRC5288- or VRC5283-immunized C57BL/6 mice (sera collected at weeks 2, 5, and 6 after immunization). Convalescent sera were pooled from 60 ZIKV-infected BALB/c mice (100 FFU of ZIKV PRVABC59 administered intraperitoneally, sera collected at weeks 6 to 8 after infection). Mouse serum was processed to remove endotoxin using Pierce High-Capacity Endotoxin Removal Resin (Thermo Fisher Scientific). Commercially available, endotoxin-free mouse IgG (Sigma-Aldrich) was used as a negative control and administered as a milligram-dose equivalent to the maximum group IgG dose in a given experiment. For the passive transfer experiment shown in Fig. 3, four groups of 10-week-old AG129 mice (8 to 10 per group) were administered purified IgG (Protein G Sepharose 4 Fast Flow, GE Healthcare) from VRC5288-immune sera, VRC5283-immune sera, ZIKV convalescent sera, or commercially available negative control IgG. Using neutralization data obtained with standard ZIKV RVP preparations, VRC5288- and VRC5283-immune IgG was normalized to an EC50 NAb titer of 4000, whereas convalescent IgG was normalized to an EC50 NAb titer of 1000. For the passive transfer experiment shown in Fig. 6, three groups of 10-week-old AG129 mice (10 per group) were administered purified IgG from VRC5283-immune sera, VRC5288-immune sera, or commercially available negative control IgG. Using neutralization data obtained with +Furin RVP preparations, VRC5288- and VRC5283-immune IgG was normalized to an EC50 NAb titer of 300. In both experiments, mice were challenged with 100 FFU of ZIKV PRVABC59 subcutaneously in the footpad 1 day after transfer of IgG. Sera were collected at the time of challenge to measure the EC50 NAb titer against standard (Fig. 3) or + Furin (Fig. 6) ZIKV RVPs. Viremia was measured 5 days after challenge by qRT-PCR. Survival was monitored daily up to denoted days after challenge.
Human sera
ZIKV convalescent human sera were obtained with informed consent at the NIH VRC (NIAID Institutional Review Board protocol number VRC 200 (ClinicalTrials.gov NCT00067054) after the symptomatic phase of illness and confirmed to be ZIKV positive by PCR (table S1) (70). None of the individuals reported previous infection with DENV, WNV, or yellow fever virus (YFV). NIH.1 and NIH.3 reported being previously vaccinated for YFV.
Vaccinated human sera samples were collected as previously described from the phase 1 clinical trials for VRC5288 and VRC5283 (NCT02840487 and NCT02996461, respectively) (37). Briefly, all sera for the current study were obtained from participants immunized at 0, 4, or 8 weeks intramuscularly with 4 mg of vaccine. A group of 20 participants ages 18 to 35 years old were immunized with VRC5288 via a single needle syringe dose. Two groups of participants ages 18 to 50 years old were immunized with VRC5283 via a split-dose needle and syringe (n = 15) or a needle-free injection (n = 14). Samples assayed in this study were collected 12 weeks after the initial vaccination. Two volunteers immunized via needle syringe, one with VRC5288 and one with VRC5283, were excluded from this study because they failed to develop a NAb response.
Cell lines and viruses
Mammalian human embryonic kidney (HEK) 293T and Vero cells were grown in Dulbecco’s modified Eagle medium with GlutaMAX (Thermo Fisher Scientific) supplemented with 7% fetal bovine serum (FBS) and penicillin-streptomycin (PS; 100 U/ml). Raji B lymphocytes expressing DCSIGNR (Raji-DCSIGNR) were cultured in RPMI 1640 medium (Thermo Fisher Scientific) supplemented with 7% FBS and PS (100 U/ml) (71). All cells were maintained at 37°C and 7% CO2. ZIKV strain H/PF/2013, collected during the 2013 French Polynesia outbreak (72), was used for the focus reduction neutralization test (FRNT) neutralization assays. Virus stocks were propagated by infecting Vero cells and collecting supernatant 2 to 4 days after infection. Virus was clarified, passed through a 0.2-μm membrane filter (Millipore), and stored at −80°C for later use. The Puerto Rican ZIKV strain PRVABC59 was used in challenge studies (34).
Quantitative reverse transcription polymerase chain reaction
Viral loads in NHPs were measured by qRT-PCR. ZIKV RNA was quantified for the NHP studies by BIOQUAL Inc. ZIKV RNA was purified from NHP serum by QIAGEN columns according to the manufacturer’s protocol. For qRT-PCR, 10% total extracted RNA was added to SensiFAST Probe Lo-ROX One-Step Kit (Bioline) along with the following primers and probe targeting the conserved NS5 region: forward, ACATGGTGCGCATCATA; reverse, AGCACTCCAGGTGTAGACCC; FAMProbe, CATGGACTACCTATCCACCCAAGTTCGCTACTT. Cycling conditions were performed at 48°C for 30 min and held at 95°C for 10 min and then 40 cycles of 95°C for 30 s and 60°C for 1 min using a 7500 Sequence Detection System (Applied Biosystems). Standard curves were generated using viral RNA with a known concentration and performing serial dilutions. The limit of detection (LOD) was determined to be 50 genome copies/ml. Values are the average of one to three technical replicates.
Viral loads in mice were measured by qRT-PCR. ZIKV RNA was extracted from 50 μl of mouse serum using Agencourt RNAdvance Blood Kit (Beckman Coulter Life Sciences) and an automated liquid handler and eluted into 33 μl of water. qRT-PCR was performed using the following primers and probe in a 20-μl reaction: forward, GGAAAAAAGAGGCTATGGAAATAATAAAG; reverse, CTCCTTCCTAGCATTGATTATTCTCA; FAMProbe, AGTTCAAGAAAGATCTGGCTG; and a 10-μl viral RNA input volume, TaqMan Fast Virus 1-Step Master Mix. Cycling conditions were performed at 50°C for 5 min and 95°C for 20 s and then 45 cycles of 95°C for 5 s and 60°C for 30 s, using a Bio-Rad CFX96 Real-Time System. Standard curves were generated using serial dilutions of the inoculation stock that had been titered via focus assay. The LOD was determined empirically to be 0.5 FFU equivalents/ml. Values are the average of technical duplicates or single measurements.
RVP production
RVPs were generated via genetic complementation of a plasmid expressing the ZIKV structural genes (CprM-E) with a green fluorescent protein (GFP)–expressing WNV subgenomic replicon (70, 73, 74). HEK293T cells were cotransfected with the CprM-E and replicon plasmids at a 3:1 ratio by mass. Cotransfection of the plasmids with 0.5 μg of plasmid expressing the human furin protease was used to generate mature ZIKV +Furin RVP preparations (56). Transfected cells were incubated at 30°C, and supernatants were harvested on days 3 to 6 after transfection. Supernatants were passed through a 0.2-μm filter (Millipore) and stored at −80°C. The infectious titer of ZIKV RVP preparations was determined by infecting Raji-DCSIGNR cells with twofold serial dilutions of RVPs. Cells were fixed 36 to 48 hours after infection using 2% paraformaldehyde, and GFP-positive cells were measured by flow cytometry.
RVP neutralization assays
ZIKV RVP neutralization assays were performed using Raji-DCSIGNR or Vero target cells as previously described (35, 70, 74, 75). Briefly, serial dilutions of mAbs or sera were incubated with RVPs for 1 hour at 37°C to allow for steady-state binding. Immune complexes were used to infect Vero or Raji-DCSIGNR cells and performed in technical duplicates. Infection of cells was carried out at 37°C. Cells were fixed 36 to 48 hours after infection using 2% paraformaldehyde, and GFP-positive cells were measured by flow cytometry (FACSCelesta, BD). Nonlinear regression analysis was performed to estimate the dilution of sera required for half-maximal infectivity (EC50). The LOD was set as the reciprocal initial dilution of sera (1:60 for NHPs, 1:30 for humans, and 1:60 or 1:30 for mice, depending on the experiment). Titers measured below the initial dilution were assigned a titer of one-half the LOD (30 for NHPs, 15 for humans, and 30 or 15 for mice).
FRNT neutralization assay
FRNT assays were performed using NHP sera as previously described (35). Serial dilutions of sera were incubated with 100 FFU per well of ZIKV H/PF/2013 for 1 hour at 37°C. Immune complexes were then added to Vero cells and incubated for 4 hours at 37°C. Opti-MEM (Thermo Fisher Scientific) containing 1% dissolved methylcellulose (Sigma-Aldrich) was then added to the plates in equal volume, leading to a final methylcellulose concentration of 0.5%. Infection occurred at 37°C for 40 to 48 hours. Foci were visualized by sequential incubations with 500 ng/ml of the mouse-derived ZIKV E domain III protein–specific mAb ZV-67 (60), horseradish peroxidase (HRP)–conjugated goat anti-mouse IgG (Sigma-Aldrich), and TrueBlue peroxidase substrate (SeraCare). Foci were quantitated using the ImmunoSpot S6 Macroanalyzer (Cellular Technologies). The resulting data were analyzed by nonlinear regression analysis as detailed above.
ZIKV prM-E–inducible cell line
A tetracycline-inducible cell line expressing ZIKV prM-E was generated using the tetracycline-regulated expression (T-REx) system of cell lines that stably expresses the tetracycline repressor from the pcDNA6/TR plasmid (Invitrogen). The open reading frame from plasmid VRC5283, encoding the JEV signal sequence upstream of ZIKV prM-E, was PCR-amplified and cloned into the pT-REx-DEST30 mammalian expression vector using Gateway technology (Invitrogen) to generate the plasmid pT-REx-ZIKV prM-E. Parental T-REx–Chinese hamster ovary (CHO) cells maintained in F12 medium supplemented with 7% FBS and blasticidin S HCl (0.01 mg/ml; Gibco) were transfected with the pT-REx-ZIKV prM-E plasmid using Lipofectamine 3000 (Invitrogen). The following day, medium was replaced with F12 medium supplemented with 7% FBS, blasticidin S HCl (0.01 mg/ml), and Geneticin (0.5 mg/ml; Gibco) to select for transfected cells. After confirming expression in the resulting polyclonal cell line, cells were sufficiently diluted in 96-well plates to achieve a single cell per well, and monoclonal cell lines were expanded in F12 medium supplemented with 7% FBS, blasticidin S HCl (0.01 mg/ml), and Geneticin (0.5 mg/ml). To generate ZIKV SVPs for use in ELISA experiments, T-REx–CHO ZIKV prM-E cells were induced with tetracycline (500 ng/ml) and maintained in F12 medium supplemented with 1% FBS, blasticidin S HCl (0.01 mg/ml), and Geneticin (0.5 mg/ml) at 37°C. Parental T-REx–CHO cells were induced with tetracycline (500 ng/ml) and maintained in F12 medium supplemented with 1% FBS and blasticidin S HCl (0.01 mg/ml) at 37°C as a negative control. Supernatants were collected on days 3 to 6, filtered through a 0.2-μm membrane, concentrated using an Amicon Ultra-15 100-kDa filter (Millipore), and stored at −80°C until use.
Particle-capture ELISA
Microplates (384 wells; Corning) were coated with mAb ZV-67 (3 μg/ml) in 0.1 M carbonate-bicarbonate buffer (pH 9.4) (Thermo Fisher Scientific) and incubated at 4°C overnight. Plates were blocked with 5% skim milk in 1× phosphate-buffered saline with 0.1% Tween 20 (PBST) at 37°C for 1 hour and then washed once with 1× PBST. Concentrated ZIKV SVP was diluted 1:20 in 5% skim milk in 1× PBST, added to mAb ZV-67–coated plates, and incubated at 37°C for 1 hour. Plates were then washed with 1× PBST six times. MAb ZV-67 was biotinylated using the DSB-X Biotin Protein Labeling Kit (Thermo Fisher Scientific) according to the manufacturer’s protocol. Serial twofold dilutions of biotinylated mAb ZV-67 (3 μg/ml starting concentration) or week 8 NHP vaccine sera (starting with an initial 1:50 dilution) in 5% skim milk in 1× PBST were added to plates and incubated at 37°C for 1 hour. Plates were washed with 1× PBST six times and incubated with high-sensitivity streptavidin-HRP (Thermo Fisher Scientific) or goat anti-rhesus IgG (H + L)–HRP (SouthernBiotech) at 37°C for 1 hour. The plates were washed with 1× PBST six times. The colorimetric assay was developed using 3,3′,5,5′-tetramethylbenzidine (TMB) (SeraCare) and stopped with the addition of 1 N HCl. Signal was measured at 450 nm using BioTek Synergy H1. Background signal obtained with concentrated supernatant from the parental T-REx–CHO cell line was subtracted from the respective sample measured with the ZIKV antigen. The end point dilution was determined as the final dilution greater than four SDs above the mean signal obtained in the absence of sera or Ab.
Soluble E protein ELISA
Recombinant soluble ZIKV E (sE) (strain SPH2015) with a C-terminal polyhistidine tag (Sino Biological) was used in an indirect ELISA. Microplates (384 wells; Corning) were coated with sE (2.4 μg/ml) or bovine serum albumin (2.4 μg/ml; Sigma-Aldrich) in 0.1 M carbonate-bicarbonate buffer (pH 9.4) (Thermo Fisher Scientific) and incubated at 4°C overnight. Plates were blocked with 5% skim milk in 1× PBST at 37°C for 1 hour and then washed once with 1× PBST. Serial twofold dilutions of biotinylated mAb ZV-67 (3 μg/ml starting concentration), rabbit anti-His tag mAb (0.2 μg/ml starting concentration) (GenScript), or week 8 NHP vaccine sera (starting with an initial 1:50 dilution) in 5% skim milk in 1× PBST were added to plates and incubated at 37°C for 1 hour. Plates were washed with 1× PBST six times and incubated with high-sensitivity streptavidin-HRP (Thermo Fisher Scientific) or goat anti-rhesus IgG (H + L)–HRP (SouthernBiotech) at 37°C for 1 hour. The plates were washed with 1× PBST six times. The colorimetric assay was developed using TMB (SeraCare) and stopped with the addition of 1 N HCl. Signal was measured at 450 nm using BioTek Synergy H1. The end point dilution was calculated similarly to the particle capture ELISA.
Ab-dependent cellular phagocytosis
A previously described ADCP assay was adapted to include recombinant ZIKV E protein (76). Recombinant ZIKV E protein (Meridian Life Science) was biotinylated and conjugated to streptavidin-coated Alexa Fluor 488 beads (Invitrogen). ZIKV E–coated beads were incubated with 1:10 dilutions of serum samples in cell culture medium for 2 hours at 37°C. THP-1 (American Type Culture Collection TIB-202) human monocytic cells were added to bead-Ab immune complexes (2.5 × 104 cells per well) and incubated at 37°C overnight. Cells were fixed with 4% paraformaldehyde and analyzed on an IntelliCyt iQue Screener PLUS flow cytometer. The phagocytic score was determined using the following calculation: (percentage of Alexa Fluor 488+ cells) × (Alexa Fluor 488 geometric mean fluorescent intensity of Alexa Fluor 488+ cells)/10,000.
Antigen-specific IgG quantitation
Recombinant ZIKV E protein (Meridian Life Science) was coupled to MagPlex beads (Luminex) according to previously published protocols (77). Serum samples were diluted and incubated with antigen-coupled beads for 2 hours. After bead washing, IgG was detected using phycoerythrin-labeled secondary Abs (0.65 μg/ml; SouthernBiotech). The median fluorescent intensity of 50 beads per region was analyzed on an IntelliCyt iQue Screener PLUS flow cytometer.
Statistical analysis
All data were graphed and analyzed using GraphPad Prism version 7. Relationships between EC50 NAb titers and protection were modeled using probit regression. Fold changes between end point sera dilutions measured using ZIKV sE and ZIKV SVP in Fig. 4C were calculated using the equation (SVP − sE)/sE. Fold changes between untransformed NAb titers measured using ZIKV Std and ZIKV +Furin RVPs in Fig. 5 and figs. S6 to S8 were calculated using the equation (Std − Furin)/Furin. Statistical significance between different groups was calculated on log10-transformed data using a Mann-Whitney test or a Kruskal-Wallis test; P values presented were adjusted for multiple comparisons using Dunn’s procedure. P < 0.05 was considered statistically significant.
SUPPLEMENTARY MATERIALS
stm.sciencemag.org/cgi/content/full/12/547/eaaw9066/DC1
Fig. S1. Comparison of the VRC5283 and VRC5288 DNA vaccine constructs.
Fig. S2. Assay comparison for measuring VRC5288- and VRC5283-elicited NAb titers.
Fig. S3. ZIKV infection elicits a variable anamnestic response in NHPs.
Fig. S4. ZIKV DNA vaccines VRC5288 and VRC5283 elicit similar binding titers against monomeric E at time of challenge.
Fig. S5. ADCP increases after ZIKV challenge.
Fig. S6. ZIKV DNA vaccines VRC5288 and VRC5283 elicit a maturation state–sensitive NAb response in NHPs.
Fig. S7. ZIKV virion maturation sensitivity disappears after challenge in VRC5288- and VRC5283-immunized NHPs.
Fig. S8. VRC5288 and VRC5283 elicit a maturation state–sensitive response in humans.
Table S1. ZIKV-convalescent human sera history.
Data file S1. Primary data.
This is an article distributed under the terms of the Science Journals Default License.
REFERENCES AND NOTES
- Copyright © 2020 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works
