Control of maternal Zika virus infection during pregnancy is associated with lower antibody titers in a macaque model

Abstract

Introduction: Zika virus (ZIKV) infection during pregnancy results in a spectrum of birth defects and neurodevelopmental deficits in prenatally exposed infants, with no clear understanding of why some pregnancies are more severely affected. Differential control of maternal ZIKV infection may explain the spectrum of adverse outcomes.

Methods: Here, we investigated whether the magnitude and breadth of the maternal ZIKV-specific antibody response is associated with better virologic control using a rhesus macaque model of prenatal ZIKV infection. We inoculated 18 dams with an Asian-lineage ZIKV isolate (PRVABC59) at 30-45 gestational days. Plasma vRNA and infectious virus kinetics were determined over the course of pregnancy, as well as vRNA burden in the maternal-fetal interface (MFI) at delivery. Binding and neutralizing antibody assays were performed to determine the magnitude of the ZIKV-specific IgM and IgG antibody responses throughout pregnancy, along with peptide microarray assays to define the breadth of linear ZIKV epitopes recognized.

Results: Dams with better virologic control (n= 9) cleared detectable infectious virus and vRNA from the plasma by 7 days post-infection (DPI) and had a lower vRNA burden in the MFI at delivery. In comparison, dams with worse virologic control (n= 9) still cleared detectable infectious virus from the plasma by 7 DPI but had vRNA that persisted longer, and had higher vRNA burden in the MFI at delivery. The magnitudes of the ZIKV-specific antibody responses were significantly lower in the dams with better virologic control, suggesting that higher antibody titers are not associated with better control of ZIKV infection. Additionally, the breadth of the ZIKV linear epitopes recognized did not differ between the dams with better and worse control of ZIKV infection.

Discussion: Thus, the magnitude and breadth of the maternal antibody responses do not seem to impact maternal virologic control. This may be because control of maternal infection is determined in the first 7 DPI, when detectable infectious virus is present and before robust antibody responses are generated. However, the presence of higher ZIKV-specific antibody titers in dams with worse virologic control suggests that these could be used as a biomarker of poor maternal control of infection and should be explored further.

Keywords: ZIKV; Zika virus; congenital Zika syndrome (CZS); macaque model; maternal ZIKV infection; maternal antibody response; pregnancy.

Figure 1

Individual maternal virologic outcomes. (A) Maternal plasma viral RNA (vRNA) loads throughout pregnancy determined by RT-qPCR. The lower limit of detection (LLoD) for the assay (150 copies/mL) is represented by a dotted line. All vRNA loads below the LLoD are set equal to the LLoD. (B) Area under the curve (AUC) values calculated from maternal plasma vRNA loads throughout pregnancy. (C) Timing of peak plasma vRNA loads is denoted by a histogram. (D) Peak plasma vRNA loads as determined by RT-qPCR. (E) Infectious virus titers determined via plaque assay from blood collected on the same day as peak vRNA load. Infectious virus titers are expressed as plaque forming units (PFU)/mL. “NA” = not applicable and indicates that the dam did not have any sample remaining from the day of peak vRNA load to test. “ND” = infectious virus was not detected on the day of peak vRNA load. (F) Duration of plasma vRNA burden is denoted as a histogram indicating the last day where plasma ZIKV vRNA levels were above the LLoD. (G) Duration of detectable infectious virus is denoted by a histogram indicating the last day where infectious virus was detected in the blood via plaque assay. “NA” = not applicable and indicates that the dam did not have any sample remaining to test. (H) The percentage of maternal-fetal interface (MFI) biopsies from the placenta, decidua, chorionic plate, and all three combined (total) that were vRNA-positive at delivery. The LLoD for tissue samples is 3 copies/mg of tissue. “NT” = not tested because the animal had a natural birth.

Figure 2

Maternal virologic outcomes based on virologic control status. (A) Area under the curve (AUC) values of the viral RNA (vRNA) loads graph. (B) Duration of plasma vRNA burden defined as the last day with a plasma vRNA load above the lower limit of detection (150 copies/mL). (C) Percentage of vRNA-positive chorionic plate, decidual, placental, and total biopsies from the time of delivery. (D) Peak plasma vRNA loads. (E) Duration of infectious virus in the plasma determined via plaque assay. Dam 044-104 (non-controller) did not have any remaining sample to assay. (F) vRNA to infectious virus ratios were determined for each dam from the day of peak vRNA load. The ratio was determined by dividing the vRNA copies/mL by the infectious virus PFU/mL. Dams with undetectable infectious virus titers were assigned a ratio of 0. Dams 044-103 (controller), 044-104 (non-controller), and 044-122 (non-controller) did not have any sample remaining from the day of peak plasma vRNA load to conduct a plaque assay. Statistically significant differences between virologic control groups were determined using a Mann-Whitney U test (*p<0.05, **p<0.01, ***p<0.001).

Figure 3

ZIKV-specific IgM antibody dynamics. (A) The sample to calibrator ratio for each dam was determined at several time points by dividing the optical density reading at 450 nm (OD450) of the experimental sample by the OD450 of the kits calibrator sample. A ratio that is ≥ 1.1 indicates a positive IgM sample (dotted line), a ratio between 1.1 and 0.8 indicates a sample that is borderline positive, and a ratio that is< 0.8 indicates a negative sample (solid line). Dams were divided based on their virologic control status. (B) The percentage of dams in each group that were considered to be positive for ZIKV-specific IgM at each time point was determined by dividing the number of positive animals within the group by the total number of animals within the group. Statistically significant differences between virologic control groups were determined using a Mann-Whitney U test (*p<0.05, **p<0.01, ***p<0.001).

Figure 4

ZIKV-specific IgG binding and neutralizing antibody dynamics. ZIKV-specific IgG binding antibody EC₉₀ and EC₅₀ titers were estimated from the raw IgG binding antibody curves ( Supplementary Figure 2 ) for each animal at multiple timepoints post-infection. Similarly, ZIKV-specific neutralizing antibody PRNT₉₀ and PRNT₅₀ titers were estimated from the raw neutralizing antibody curves ( Supplementary Figure 4 ). The (A) EC₉₀ and (B) EC₅₀ titers for the two virologic control groups were compared at each timepoint tested. The limit of detection is denoted by a dotted line on the EC₉₀ figure. The (C) PRNT₉₀ and (D) PRNT₅₀ titers for the two virologic groups were compared at each timepoint tested. The limit of detection is denoted by a dotted line on the PRNT₉₀ figure. Statistically significant differences between the virologic control groups were determined using a Mann-Whitney U test (*p<0.05, **p<0.01, ***p<0.001).

Figure 5

IgM and IgG linear epitope counts. The total number of linear epitopes, when defined as multiple adjacent reactive peptides, across the entire viral polyprotein was quantified at (A) 13-17 and 21-24 days post-infection (DPI) for IgM with dams separated based on virologic control group. The overall IgM linear epitope count within each region of the ZIKV polyprotein was also determined at (B) 13-17 DPI and (C) 21-24 DPI. For this, all dams were considered as a single population and colors correspond to individual regions of the ZIKV polyprotein. The total number of linear epitopes across the entire viral polyprotein was quantified at (D) 27-31 and 108-135 DPI for IgG with dams separated based on virologic control group. The total number of linear IgG epitopes within each region of the ZIKV polyprotein was determined at (E) 27-31 and (F) 108-135 DPI.