The amino acid sequence coding for the ZIKV prM/M and E proteins was retrieved from the NCBI (National Center for Biotechnology Information, accession number: KU321639.1) and four vaccine constructs comprising the full-length prM/M and E proteins were designed as follows: ZK_ΔTMP (comprising a deletion of the TM region); LAMP/ZK_ΔTMP (comprising a deletion of the TM region and the addition of the LAMP sequence); ZK_ΔSTP (comprising a deletion of the TM and stem regions) and LAMP/ZK_ΔSTP (comprising a deletion of the TM and stem regions and the addition of the LAMP sequence). The constructs were reverse translated to DNA and optimized for expression in eukaryotic cells. The expression cassettes were commercially synthesized and inserted into the pcDNA3.1 mammalian expression vector (Genscript, USA).

Monolayers of BHK-21 cells (ATCC, USA) were transfected with the constructs using Lipofectamine 2000 (Invitrogen, USA) according to the manufacturer’s instructions. After 48h, cells were fixed using 80% acetone for 1h at room temperature (RT) and blocked with assay buffer containing 5% skimmed milk (Bio-Rad, USA) in PBS-T buffer (1X PBS with 0.05% Tween-20) for 1h at RT. For ZIKV E detection, the 4G2 monoclonal antibody (34) (1μg/mL) was added in assay buffer for 2h at 37°C. Plates were washed five times with PBS-T, followed by peroxidase-conjugated anti-mouse IgG antibody (Sigma Aldrich, USA) for 1h at RT. After a final wash step, the reaction was revealed with tetramethylbenzidine (TMB, KPL SureBlue Reserve, USA) substrate for 45 min at RT, stopped with 1N HCl and read at 450nm (OD 450nm) using a microplate spectrophotometer (Benchmark Plus, Bio-rad, USA). A plasmid coding for YFV proteins (pL/YFV) was used as a positive control (18). All samples were tested in duplicate, and the results were considered valid when intra-assay variability was below 20%.

The supernatant of transfected BHK21 cell culture was transferred to high binding, half area 96-well polystyrene plates (Costar, USA), and soluble ZIKV E proteins were allowed to bind to the plate surface for 16h at 4°C. Detection of immobilized ZIKV E proteins was performed following the ELISA protocol previously described.

Adult C57BL/6 mice (female and male), aged 8–10 weeks old, were purchased from the Faculdade de Medicina da Universidade de São Paulo (FMUSP, Brazil) animal facility and bred in specific pathogen-free conditions. Adult C57BL/6 mice with interferon α and β receptor depletion (IFNαβR^-/-) (B6.129S2-Ifnar1tm1Agt/Mmjax- https://www.jax.org/strain/010830), known to be the susceptible model of ZIKV infection, were used for adult infection assays. The IFNαβR^-/- mice were purchased from Instituto de Ciências Biomédicas da Universidade de São Paulo (ICB-USP, Brazil). Protocols were approved by the local Ethics Committee of Animal Care and Use (CEUA-ICB: 6912040918, CEUA-IMT: 363 and CEUA-FMUSP: 1763/2022).

Adult mice received two doses of immunization by intradermal (id) route at a 20-days interval with 50 μg of the DNA plasmids in a total volume of 50 μL of PBS delivered at the base of the tail. PBS solution was used as negative control. For adjuvant-associated immunization, mice received two doses of 50 μg of ZK_ΔSTP vaccine in the presence of commercial adjuvant formulations, such as Alum (i.e., aluminum hydroxide, 10 mg/mouse, Thermo Scientific, USA), MPLA (monophosphoryl lipid A, 10 μg/mouse, Sigma Aldrich, USA) or a combination of Alum (10 mg/mouse) + MPLA (10 μg/mouse). The solutions comprising the DNA vaccine along with each commercial adjuvants were freshly prepared immediately before immunization according to the manufacturers’ recommendation. Briefly, the plasmid DNA (50 μg) was diluted in each adjuvant suspension and gently homogenized by pipetting. Immunogenicity evaluation was performed at different time points after boost. Sera, inguinal lymph nodes (iLN) and spleens were also obtained at different time points.

The Asian strain of locally isolated ZIKV (BR-PE243/2015) was used for the infection assays. After in vitro virus expansion, the concentration of the viral stock was determined through classical plaque assay and expressed as particle forming units (PFU)/mL. Adult IFNαβR^-/- mice previously immunized with ZK_ΔSTP or PBS control were infected two weeks after boost intravenously (iv) with 10⁵ particle forming units (PFU)/mL of ZIKV (100 μL). Survival and weight gain were monitored for up to 3 weeks. Sera were collected at 6- or 7-days post infection (dpi), while brains were obtained after euthanasia at 10 or 21 dpi and processed for viral load measurement. Mice were initially maintained until 21 dpi to observe clinical signs of infection, weight gain and mortality rates and possible natural recovery. The clinical signs of infection included observed neurological symptoms (hind limb paralysis, tremors and ataxia) and behavioral changes (changes in activity levels, exploration and social interactions). It is important to note that a non-lethal dose of virus was adopted to perform the infection challenge experiments. Therefore, no severe clinical signs of disease were observed to justify euthanasia. Mice that succumbed to infection presented weight loss, weakness and decreased social interaction. When immunization was performed in the presence of the Alum adjuvant formulation, the time point was reduced in half to confirm the vaccine’s protective and prophylactic effect.

Cellular response was assessed in spleens through ELISPOT using the IFN-γ ELISPOT set (BD Biosciences, USA) as previously described (26) 0.5x10⁶ cells/well were stimulated with 2 µg/mL of recombinant ZIKV E protein (Native Antigen, UK). The anti-CD3 and anti-CD28 antibodies (1 µg/mL each) (BD Bioscience, USA) were used as positive control. Unstimulated cells were used as baseline. Spots were quantified using AID iSpot FluoroSpot Reader System and AID EliSpot Version 7.0 software (AID-Autoimmun Diagnostika GMBH, Germany). Assays were performed in duplicate, baseline values were subtracted from all samples, and the results were expressed as the average number of spots forming cells (SFCs) per 10⁶ cells.

To measure equilibrium binding affinity, anti-ZIKV E IgG antibodies were titrated through ELISA. High binding 96-well polystyrene plates (Costar, USA) were coated with 2 µg/mL of recombinant ZIKV E protein (Native Antigen, UK) for 16h at 4°C, and blocked with assay buffer (5% skimmed milk in PBS-T buffer) for 30 min at RT. Samples were serially diluted (1:50 – 1:25.600) and added to the plates. Plates were washed five times with PBS-T followed by peroxidase-conjugated anti-mouse IgG antibody (Sigma Aldrich, USA) for 1h at RT. After a final wash step, the reaction was revealed with TMB (KPL SureBlue ReserveTM) substrate for 30min at RT, stopped with 1N HCl and read at 450nm (OD 450nm) using a microplate spectrophotometer (Benchmark Plus, Bio-rad, USA). All samples were tested in duplicate, and the results were considered valid when intra-assay variability was below 20%. Antibody titers are expressed as EC₅₀ and were calculated using a three-parameter nonlinear model. The data was plotted using GraphPad Prism version 7.0.

ZIKV-specific neutralizing antibodies were assessed by PRNT following a modified protocol described in detail elsewhere (35). Mice sera were heat-inactivated (56°C for 30 min), serially diluted (1:50 – 1:25.600) in culture media (DMEM, GIBCO, USA) and incubated with 50 PFU of locally isolated ZIKV for 1h at 37°C under gentle agitation. The antibodies-virus complexes were transferred to a monolayer of Vero cells, and DMEM containing 1% (w/v) carboxymethyl cellulose was overlaid. The cells were incubated at 37°C, 5% CO₂ for 7 days. Plaque number was calculated after fixation of the cells with 10% formaldehyde and staining with 0.2% crystal violet. The cut-off for PRNT positivity was defined based on a 50% reduction in plaque counts (PRNT50) compared to the positive control (ZIKV infected cells). ZIKV-specific neutralizing antibody titers were estimated using a four-parameter non-linear regression. Samples were considered positive when neutralizing antibody levels were ≥1:100.

The frequency of GC B cells from iLNs was assessed one week after boost as previously described (26). 2x10⁶ cells were stained with anti-CD3 (PerCP-Cy5.5 – Clone: 17A2), anti-B220 (APC – Clone: RA3-6B2), anti-CD95 (PE-Cy7 – Clone: Jo2) and anti-GL-7 (FITC – Clone: GL-7) antibodies (all from BD Biosciences, USA). Viability was determined using live/dead (Texas Red, Invitrogen, USA). A total of 500.000 events were collected and analyzed by flow cytometry (LSR Fortessa, BD Biosciences, USA) using FACS-Diva (BD Bioscience, USA) and FlowJo 10.0.6 (BD Bioscience, USA) software. Fluorescence minus one (FMO) controls were included to confirm proper compensation and to define positive signals.

Viral load quantification was performed in sera and brain tissues of ZIKV-infected mice through qRT-PCR/Taqman following a modified protocol described in detail elsewhere (36). Briefly, viral RNA was extracted from serum and brain tissues using the QIAamp Viral RNA kit (Qiagen, Germany) and the RNeasy Mini Kit, (Qiagen, Germany), respectively, according to the manufacturer’s instructions. ZIKV RNA was amplified using the TaqMan Fast Virus 1–8 Step Master Mix reagent (Applied Biosystems, USA) in a 7500 Fast Dx Real-Time PCR Instrument (Applied Biosystems, USA). The following primers and probes were used to quantify viral RNA: 5′-CCGCTGCCCACACAAG-3′ (forward primer); 5′-CCACTAACGTTCTTTTGCAGACAT-3′ (reverse primer), and FAM-AGCCTACCTTGACAAGCAGTCAGACACTCAA-MGB (probe). Infective viral particle amount in the samples was estimated after interpolation to a ZIKV standard curve run in each plate. The standard curve used in the PCR assays consisted of the same viral inoculum used to infect the animals (in the challenge experiments) and ranged from 10⁶ to 10² PFU/mL. The virus concentration corresponding to the last point of the dilution curve (10² PFU/mL) was considered as the assay limit of detection (dashed line). Positive and negative controls corresponded to purified viruses (from cell culture) and water, respectively.

Data were analyzed using the GraphPad Prism software (version 7.0). Comparison between groups was performed using non-parametric statistical tests. P values were considered statistically significant when p≤ 0.05. * Mann-Whitney; # Kruskal-Wallis; § Multiple-t-test.

Previous studies have demonstrated that prM/M and E proteins co-expression induces nAb against ZIKV (12, 14, 37) and others flaviviruses (38–40). Removal of antigen membrane anchors of E protein, such as the TM portion, often improves vaccine immunogenicity by favoring extracellular protein secretion (18, 19). Similarly, depletion of the E protein stem portion was also described to improve protective nAb production (13, 41, 42). We have previously demonstrated that humoral and cellular responses against viral antigens can be remarkably improved by the modulation of molecular adjuvants added to the expression cassette due to improved antigen presentation (23–26, 43). In this work, we designed four DNA vaccines coding for the ZIKV prM/M and E proteins along with the tPA-SP leader sequence (Figure 1A). The relevance of deletions of the TM (ΔTMP) and stem (ΔSTP) regions of envelope, as well as the presence of the LAMP molecular adjuvant, for vaccine immunogenicity were evaluated.

Aiming to define the best construct candidate for the in vivo studies, the vaccine antigens expression efficiency of all four constructs was assessed in vitro. While the ZIKV E protein was successfully detected intracellularly in transfected cells for all candidates (Supplementary Figure 1), the ZK_ΔSTP construct remarkably induced the highest antigen expression (Figure 1B). Additionally, the cell-secreted soluble E protein was also detected in the supernatants of cells transfected with ZK_ΔSTP (Figure 1C).

Aiming to identify the most promising vaccine candidate, we assessed the immunogenicity of each construct in adult C57BL/6 mice. Mice were immunized via intradermal (id) route with two doses of each DNA vaccine at a 20-days interval, and humoral and cellular responses were interrogated one week after boost. Our research group has been largely using id administration in maternal and neonatal immunization protocols, especially due to easiness of access and vaccine bioavailability in neonatal mice (23–26).

Vaccination with all four DNA constructs induced ZIKV-specific T cell response, but only vaccines without stem increased the frequency of IFN-γ-secreting cells in comparison to the PBS control group (Figure 2A). Mice immunization with ZK_ΔSTP vaccines showed higher levels of anti-ZIKV E IgG antibodies (Figure 2B), which suggests that deletion of the full membrane-anchoring regions of the E protein is fundamental to humoral response induction. The ZK_ΔSTP vaccine induced the strongest production of anti-ZIKV antibodies among the tested vaccines in comparison to the control group and LAMP/ZK_ΔSTP (Figure 2B). High levels of anti-ZIKV IgG2c antibody subclass were also detected upon ZK_ΔSTP immunization (Supplementary Figure 2). Noteworthy, vaccine formulations associated with LAMP did not show enhanced immunogenicity, when compared to their counterparts.

Based on the superior performance of the ZK_ΔSTP regarding antigen expression in vitro and induction of virus-specific immune response in vivo, this construct was selected for further validation. Mice were immunized as described before and immunogenicity of the selected vaccine was evaluated at different time points (Figure 3A). Our data on the frequency of IFN-γ-secreting cells shows that ZK_ΔSTP induced T cell response both in the short (10 days after boost) and the long-term (3 months after boost) assessment (Figures 3B, C)

As for B cell response, our data show that ZK_ΔSTP induced high titers of anti-ZIKV E IgG antibodies in the first month after immunization. However, these levels sharply decreased from 90 days on after immunization (Figure 3D). While this finding could indicate the loss of the long-lasting protective response, high nAb titers were still observed throughout the evaluation time (Figure 3E). This data was corroborated by the induction of GC B cells in the immunized animals (Figure 3F, Supplementary Figure 3). GC are transiently formed microanatomical sites within the B cell zone in the follicles of secondary lymphoid organs, where B cells proliferate and where antibodies undergo iterative rounds of gene somatic hypermutation coupled to multiple selection and differentiation pathways of affinity maturation. Therefore, GC B cell are known as the source of the high-affinity, class-switched antibodies required for protective immunity (44, 45). The increase in the frequency of GC B cells in draining iLN one week after the vaccination boost suggest that ZK_ΔSTP activated this cell population, which might be involved in efficient nAb generation.

Adult C57BL/6 IFNαβR-/- mice, known as the susceptible model of ZIKV infection, were used for the viral challenge assays. Prior to the challenge experiments, ZK_ΔSTP vaccine immunogenicity was confirmed in this mice lineage. Production of virus-specific and neutralizing antibodies were identified in sera obtained at day zero before infection (Supplementary Figures 4A, B). To assess vaccine protective efficacy, mice were challenged with 10⁵ PFU of ZIKV two weeks after the ZK_ΔSTP vaccine boost and evaluated regarding weight gain, viremia and brain viral load for up to 3 weeks (Figure 4A).

The data showed that one mouse from non-immunized (PBS) group sustained lethality (20%), while survival was maintained in mice vaccinated with ZK_ΔSTP. PBS group mice also exhibited marked weight loss in the first week after infection and presented a slower recovery, when compared to animals immunized with the ZK_ΔSTP vaccine (Figure 4B). Noteworthy, the ZK_ΔSTP immunized mice exhibited a weight gain profile comparable to the non-infected group (mock) and showed no clinical signs of infection. Additionally, viremia was absent in the vaccinated mice 1 week after infection (Figure 4C). Protective effect of vaccination was also prominent in the brain tissues (Figure 4D). We observed a considerable decrease in viral load in the brains at 21dpi (samples from 42% of animals were under the detection limit), when compared to the PBS group. Interestingly, the presence of ZIKV in the brain tissues of PBS group was observed even 3 weeks post infection (Figure 4D). Despite the promising protective effect observed, the lack of a sterile immunity indicates that additional strategies might be needed to increase antibody titers over time and to ensure optimal immunity.

Aiming to enhance the magnitude and lasting of the vaccine humoral response, immunization of the ZK_ΔSTP DNA vaccine along with commercially available adjuvants was explored. Animals were immunized following the same protocol previously described and in association with Alum, MPLA or a combination of both (Alum+MPLA). Immunization with MPLA alone and its combination with Alum (Alum+MPLA) did not increase virus-specific antibody titers. In fact, those groups exhibited a similar IgG profile to what was observed for the non-adjuvant control group (animals that received the ZK_ΔSTP DNA vaccine only), in which antibody levels decreased 2 months after boost (Figure 5A). Conversely, immunization in the presence of Alum induced high titers of anti-ZIKV E IgG antibodies over time (Figure 5A), as well as increased anti-ZIKV nAb titers at short and long-term periods, when compared to the non-adjuvant group (Figure 5B). Interesting to mention that, even though Alum induced a neutralizing response similar to the control group during the first month after immunization, long-term nAb levels were significantly increased later on (60 and 90 days after boost). This finding reflects an important improvement of vaccine humoral response induction (Figure 5B). Based on these results, the immunization with ZK_ΔSTP in association with Alum was further evaluated for protection against ZIKV infection.

Adult IFNα/βR-/- mice were challenged for ZIKV infection following the same protocol previously adopted (Figure 6A), and sera and brains were collected at 6 and 10dpi, respectively. As previously observed, 2 animals sustained lethality (25%) in the non-vaccinated group (PBS), while no animal loss was observed in the immunized groups. Additionally, weight gain rate was maintained within normal limits for these groups (when compared to the mock control) (Figure 6B). It is important to mention that, while immunization with ZK_ΔSTP alone and its combination with Alum lead to elimination of viremia within one week after infection (Figure 6C), only the immunization in the presence of the adjuvant was capable of ensuring complete absence of virus in the brain at short period (Figure 6D). Together, these findings confirm the protective effect of this vaccine approach.