A Zika Virus-Like Particle Vaccine Mitigates Early Pregnancy Loss In Rhesus Macaques

Abstract

Zika virus (ZIKV) is an arthropod-borne Orthoflavivirus that caused a major outbreak in the Americas in 2015-16. In Brazil, up to 46% of ZIKV positive pregnancies resulted in congenital Zika syndrome (CZS). CZS is characterized by a wide range of neurologic birth defects and miscarriage in up to 7.6% of affected pregnancies. With no current licensed ZIKV vaccines, we sought to evaluate a Zika virus-like particle (VLP) vaccine candidate in a rhesus macaque (RM) pregnancy model. VLPs were produced in mammalian cells expressing the pre-membrane-envelope region of the Asian-lineage ZIKV strain PRVABC59, which belongs to the Asian ZIKV lineage that is associated with outbreaks of congenital disease. To evaluate vaccine protection against adverse pregnancy complications, two cohorts of female RM were vaccinated with ZIKV-VLP with adjuvant Alhydrogel (alum) or adjuvant alone prior to mating. At gestational day (GD) 30 (early first trimester), pregnant animals were challenged with ZIKV-DAK 41524, an African-lineage strain shown to induce 1st-trimester fetal demise in 78% (n=11/14 animals) of RM, making it an ideal and stringent model for evaluating ZIKV vaccines. Within the vaccinated cohort, 2 of 3 animals reached the study endpoint of GD 90 with no observed adverse pregnancy outcomes. The third animal experienced pregnancy loss at GD 49 (18 d post infection), although no infectious virus was detected in placental or fetal tissues. In the unvaccinated cohort, two animals had severe adverse events. One animal experienced preterm labor, and another developed early-onset hydrops fetalis with widespread ZIKV-RNA detected via RNAscope and extensive placental damage. These results confirm a significant risk for early pregnancy loss in RM infected with ZIKV-DAK 41524. This model can be further used to understand the complexities of placental immunological features underlying stillbirth and miscarriage following infection. Our findings indicate that this ZIKV-VLP vaccine candidate protected pregnant macaques against fetal demise associated with highly pathogenic ZIKV challenge.

Figure 1.. ZIK-VLP vaccine adjuvanted with alum induces ZIKV- neutralizing antibodies and limits ZIKV replication and spread in nonpregnant macaques.

A, Study schematic for cohorts 1–3. Non-pregnant rhesus macaques (average ages: 9.6, 10.9, and 11.4 years for cohorts 1–3, respectively) were immunized subcutaneously (s.q.) in the right arm with either PBS (control), ZIKV-VLP alone (10⁵ FEE), or 10⁵ FEE ZIKV-VLP adjuvanted with 500 μg alhydrogel (alum). Animals received a prime and boost 21 days apart, followed by challenge with ZIKV strain PRVABC59 (ZIKV-PR). Challenge was performed with a total of 10⁵ FFU administered via five s.q. injections per arm (10⁴ FFU per site) to mimic mosquito bites. Blood was collected at designated timepoints to assess plasma viremia and immune responses. Schematic was created in BioRender. Jaeger, H. (2025) https://BioRender.com/wnxj8so. B, ZIKV-binding IgG titers were measured in plasma via ELISA. The lower limit of quantification (LLOQ) was 1:50 (dashed line). C, ZIKV-neutralizing antibody responses were assessed using focus reduction neutralization tests (FRNT₅₀) with heat-inactivated plasma. Titers were calculated using nonlinear regression. The LLOQ was a 1:50 plasma dilution. D, Plasma viremia was quantified via RT-qPCR. Data represent the mean of three technical replicates. The limit of detection (LOD) was 100 ZIKV RNA copies/mL, as determined by standard curve. E, Axillary lymph node (LN) biopsies were collected under ultrasound guidance at day 7 post-infection and viral RNA was quantified per 1 μg of tissue. F, G Viral loads at necropsy (28 dpi) were detected via one-step RT-qPCR in triplicate. The LOD was 100 copies of vRNA/μg.

Figure 2.. ZIKV-VLP vaccination in pregnant macaques: study design and fetal outcomes.

A-C, Female rhesus macaques (average age: 15 years) were immunized s.q. in the right arm with either 500 μg alum alone (n=5) or 10⁵ FFE of ZIKV-VLP adjuvanted with 500 μg alum (n=5) for a total of three doses. Animals then were placed into time-mated breeding (TMB) and were subsequently challenged with ZIKV strain Dakar 41524 (ZIKV-DAK). Animals were challenged with 10⁵ FFU, administered across ten total s.q. sites with 5 sites per arm (10⁴ FFU per site) to mimic multiple mosquito bites. Blood samples were collected at designated timepoints to assess serum viremia and immune cell activation. Animals were also monitored daily for signs of viremia and weight loss and underwent non-sedated ultrasound assessments every other day to monitor fetal viability. Females were divided into three groups, resulting in different average days post-vaccinated to challenge: A, Group 1 (230 dpv) B, Group 2 (198 dpv) and C, Group 3 (158 dpv). *In alum-only control group (C), one fetus was euthanized following preterm labor with bradycardia (C1), and another experienced fetal demise due to features consistent with hydrops fetalis (14–16 dpi; C2). In the VLP + Alum group (V), one case of hydrops fetalis met humane endpoint criteria at 18 dpi (V2). Animals that did not become pregnant were released from the study and are denoted by R. Image created in BioRender. Jaeger, H. (2025) https://BioRender.com/dqz39uj

Figure 3.. Dynamics of ZIKV-Specific Antibo, Maternal Viremia, and Fetal Survival.

A, ZIKV-binding IgG antibody titers were quantified in maternal macaque plasma at the indicated timepoints by ELISA. The dashed line represents the average baseline (day 0) titer, determined to reflect background/nonspecific binding. The lower limit of quantification (LLOQ) was a 1:20 dilution. B, Maternal plasma viremia was measured by RT-qPCR, with each data point representing the average of three technical replicates. The limit of detection (LOD) was 500 copies ZIKV RNA/mL plasma; samples below the LOD are plotted at 0 copies/mL. C, Kaplan-Meier survival curve of fetal outcomes stratified by treatment group.

Figure 4.. Longitudinal peripheral blood innate immune cell phenotype and activation.

Rhesus macaque PBMC isolated at the indicated timepoints were stained with antibodies directed at specific cellular markers and analyzed for cell phenotype using flow cytometry. Changes in the longitudinal frequency of both total and activated (CD169+) classical monocytes (A), non-classical monocytes (B), and intermediate monocytes (C) were quantified. Dendritic cells (CD16−/CD14−) were separated into myeloid dendritic cells (D), and plasmacytoid dendritic cells (not shown) were quantified as total and activated (CD169+). Lines represent individual animals referenced in the legend with Alum (teal) and VLP-Alum (pink).

Figure 5.. Adaptive humoral responses show differences in early timepoints post-ZIKV inoculation.

A, ZIKV-binding IgG antibody titers were quantified in maternal macaque plasma at the indicated timepoints by ELISA. The dashed line represents the average baseline (day 0) titer for all animals (Figure 3A), determined to reflect background/nonspecific binding. The limit of detection (LOD) was a 1:20 dilution. B, the longitudinal development of ZIKV-neutralizing antibodies was quantified in ZIKV neutralization assays using the ZIKV-PR strain as the infectious virus and heat-inactivated macaque plasma at indicated timepoints post-inoculation. The 50% focus reduction neutralization titers (FRNT₅₀) were determined using non-linear regression. The LOD was a 1:50 plasma dilution indicated by the dashed lines (black). Vertical dashed lines indicate the timing of demise of each fetus.

Figure 6.. Representative ultrasound and necropsy findings in ZIKV-infected pregnancies.

A-D, Ultrasound images from cases C2F and V2F (gestational days 46–48) show features of hydrops fetalis, including scalp edema (A), ascites (B), pleural effusion and cardiomegaly (C), and generalized subcutaneous edema spanning the entire fetus (D). E, Necropsy confirmed these findings, with the red asterisk denoting hepatosplenomegaly and the red arrow highlighting subcutaneous edema and spinal cord tenting. F, A large placental bleed with chorionic and amniotic membrane separation was observed in case V1 at 35 days post-infection. G, The placental edge folded over consistent with a circumvallate placenta was also noted in this animal. All ultrasound images were obtained by board-certified maternal-fetal medicine (MFM) physician J.O.L.

Figure 7.. Pregnancy dynamics measured for endpoint cases longitudinally show normal growth dynamics when compared to institutional controls.

A, Maternal body weight was measured at indicated timepoints for each animal and graphed as a percentage of starting body weight. Longitudinal ultrasound measurements of fetal growth across gestation including (B) Biparietal Diameter (BPD), (C) Head Circumference (HC), (D) Abdominal Circumference (AC), and (E) Femur length (FL) in the six ZIKV-challenged fetuses with Alum control animals (n=3) in teal and VLP +Alum (n=3) in pink were plotted against historical controls, from ONPRC published data used to calculate the logarithmic regression for 50^th percentile (black solid line) and for 10^th and 90^th (dashed line) (52). F, Amniotic Fluid Index data was obtained by standard measurement of four quadrants and plotted against historical control with the logarithmic regression representing for 50^th percentile (black solid line) and for 5^th and 95^th (dashed line). G, UtA blood flow (cQtA/kg) corrected for maternal body weight using UtA diameter, maternal UtA cross-sectional area and volume of blood flow through the maternal UtA as previously described (43). H, Doppler measurements were also used to calculate the umbilical artery Pulsatility Index (PI). Graphs were again plotted against ONPRC historical data. Graphs were created using GraphPad Prism v10.2.

Figure 8.. ZIKV tissue distribution in dams.

A, Graphical depictions of viral loads from S. Table 2 were detected via one-step RT-qPCR. The limit of detection (LOD) was 500 copies of ZIKV vRNA. Graphs represent tissues that had at least one animal with positive viral detection and are separated into five tissue types. A-E, Lymphoid, Genitourinary, Musculoskeletal, Gastrointestinal and Pulmonary, and Nervous System tissues were separated for each tissue. Graphs were created using GraphPad Prism v10.2.

Figure 9. Infectious virus and placental vRNA loads.

A, infectious virus was isolated from tissue homogenate, plasma, or fetal fluids. 400μl was added to C6/36 cells for 3 days and then titered on vero cells in a focus forming assay where + = virus detected (<10⁵ FFU/mL), ++ = high virus titer detected (10⁵-10⁸ FFU/mL), - = not detected, nt = not tested. Table created in BioRender. Jaeger, H. (2025) https://BioRender.com/4sig92p. B, Total RNA was isolated from mapped placental cotyledons and quantified using a one-step RT-qPCR for ZIKV RNA, summarized in Table 2. Percent positivity was determined by the number of cotyledons with vRNA detected over the LOD divided by total amount of cotyledons tested for each animal. Graph created using GraphPad prism v10.2.

Figure 10.. ZIKV infection induces placental abnormalities including infarction, abruption, and vascular lesions in early demise cases.

Gross and histopathological examination of placentas from animals C1 (A, D, G), C2 (B, E, H), and V1 (C, F, I), collected at necropsy on GD34, GD47, and GD92, respectively. C1 (A) shows a large placental abruption outlined in white, consistent with an event occurring within 24 hours prior to delivery. In C2 and V1 (B, C), white asterisks mark infarctions, and a white arrowhead in V1 (C) indicates a circumvallate placenta. The corresponding hematoxylin and eosin-stained sections are shown in D–I, where the abruption identified grossly in C1 is confirmed histologically (D), C2 shows leukocytoclastic vasculitis and spiral artery remodeling (E, black asterisks), and V1 demonstrates a microinfarction (F). Additional histopathological features are shown in G–I, where C1 contains red blood cells (RBCs) without nuclear clearing indicated by the black arrowhead (G), C2 displays RBCs with nuclear clearing (H), black arrowhead), typically observed at 2–3 per high-power field, and V1 exhibits moderate villous knotting (I). Scale bars are shown in black; all histology images were acquired at 40× magnification.

Figure 11.. ZIKV vRNA was detected via ISH staining only in case C2.

(A–C) Representative sagittal sections of FFPE-blocked C2 fetus stained with a ZIKV-specific ISH probe (Warp Red) and counterstained with hematoxylin and lithium blue. More intact brain regions were visible on the left side (A, C), while major organs and musculoskeletal structures were better preserved on the right. Each column corresponds to the low-magnification overview images shown in A–C. ZIKV+ cells were detected in multiple major organs, including the lungs (D), heart (E), kidney (G), gastrointestinal tract (H), liver (J), and musculoskeletal regions such as the foot (K). Viral RNA was also detected throughout the central nervous system, including the frontal lobe (F), meninges (I), and skin and soft tissue surrounding the skull. Scale bars are shown in black; all histology images were acquired at 40× magnification. (M) Quantification of total cell counts (blue nuclei) and ZIKV+ cells (Warp Red staining) was performed using Halo AI (Indica Labs).5 sections were stain from right side (B) where the 5 data points represent 5 separate slides. The kidney, lungs and brain were quantified from the left side and only have one data point representing the one slide that was stained. Graph was created using GraphPad Prism v10.2.

Figure 12.. Hydrops fetalis cases show signs of hypoxia.

(A–H) Representative sagittal sections from formalin-fixed paraffin-embedded (FFPE) fetuses C2F (A, C, E, G) and V2F (B, D, F, H) stained with Masson’s Trichrome. (A, B) Low-magnification images of the heart; corresponding high-magnification views shown in (E, F). Cyan arrow in (E) indicates a proangiogenic response within the C2 fetal myocardium. (C, D) Low-magnification images of the liver; corresponding high-magnification views shown in (G, H). Cyan arrows in (G, H) identify portal vessels lacking fibrosis—absence of blue staining confirms minimal collagen deposition. (I, J) Fetal placental sections stained for CD31 with DAB chromogen (brown). Black arrows indicate fetal capillaries within the villous stroma. (K, L) Fetal placental sections stained for CD68 with DAB chromogen (brown). Black arrows highlight fetal macrophages (Hofbauer cells) within the villous stroma. All images acquired at 40× magnification; scale bars shown in black.