Maternal antibodies provide partial protection from postnatal Zika viremia in nonhuman primates

Zika virus (ZIKV) will remain a public health threat until effective vaccines and therapeutics are made available in the hardest hit areas of the world. Recent data in a nonhuman primate model showed that infants postnatally infected with ZIKV were acutely susceptible to high viremia and neurological damage, suggesting the window of vulnerability extends beyond gestation. We addressed the susceptibility of two infant rhesus macaques born healthy to dams infected with Zika virus during pregnancy. Passively acquired neutralizing antibody titers dropped below detection limits between 2 and 3 months of age, while binding, possibly non-neutralizing antibodies remained detectable until viral infection at 5 months of age. Post-infection acute serum viremia was substantially reduced relative to adults infected with the same dose of the same stock of a Brazilian isolate of ZIKV (n=11 pregnant females) and another stock of the same isolate (n=4 males and 4 non-pregnant females). Virus was never detected in cerebrospinal fluid nor in neural tissues at necropsy two weeks after infection, suggesting reduced viral burden relative to adults and published data from infants. However, viral RNA was detected in lymph nodes, confirming some tissue dissemination. Though protection was not absolute, our data suggest infants born healthy to infected mothers may harbor a modest but important level of protection from postnatally acquired ZIKV for several months after birth, an encouraging result given the potentially severe infection outcomes of this population.


1
Zika virus (ZIKV) emerged in Brazil in 2015 and maternal infection during pregnancy was 2 astutely correlated with an increase in newborns with microcephaly [1, 2], a profound 3 developmental defect that results in infants with reduced brain size and cognitive 4 capacity. ZIKV was initially discovered in 1947 in the Zika forest of Uganda during 5 surveillance for yellow fever virus [3]. Soon thereafter, it became clear that human 6 infections with ZIKV in that region were not uncommon [4,5] but disease associated with addressed. A recent macaque study showed that fetal infection after subcutaneous delivered via caesarian section at approximately gestational day 155 (full term) and 1 housed in a primate nursery until 5 months of age and then infected with ZIKV 2 subcutaneously with 10^4 PFU of a Brazilian isolate (Rio-U1/2016 GenBank KU926309), 3 which was passage twice in Vero cells post-virus isolation. The animals were euthanized 4 fourteen days later.

6
Viral load measurements. Viral RNA was amplified and quantified as described 7 previously [23]. Briefly, RNA was manually extracted from fluid samples (CSF or blood 8 serum) using the High Pure Viral RNA Kit (Roche). RNA was then subjected to reverse 9 transcription and quantitative PCR using primers and a fluorescently conjugated probe 10 on an Applied Biosystems 7900 instrument.

12
Plaque Reduction Neutralization Test (PRNT) 80 measurements. Neutralizing antibody 13 quantification by plaque reduction neutralization test (PRNT) endpoint 80% PRNT titers 14 were determined in infant macaque plasma, where each sample was tested in duplicate.

15
Plasma samples were heated to 56°C for 30 minutes to inactivate complement, serially (represented as the mean final virus-serum dilution from both replicates) that inhibited at 1 least 80% of plaques is reported as the antibody titer.

3
Detection of ZIKV-specific IgG in rhesus plasma. High-binding 96-well ELISA plates 4 (Greiner; Monroe, NC) were coated with 40 ng/well of 4G2 monoclonal antibody, 5 produced in a mouse hybridoma cell line (D1-4G2-4-15,ATCC; Manassas, VA), diluted 6 to 0.8 ng/uL in 0.1M carbonate buffer (pH 9.6) and incubated overnight at 4°C. Plates 7 were blocked in 1X Tris-buffered saline containing 0.05% Tween-20 and 5% normal goat 8 serum for 1 hour at 37°C, followed by an incubation with diluted ZIKV (strain 9 PRVABC59, BEI; Manassas, VA) for 1 hour at 37°C. Optimal virus dilution was plus 2 x standard deviations (SD). For samples considered positive, their OD values for 23 the serial dilution were entered into Prism v8 (GraphPad Software; San Diego, CA) to transformed data was then analyzed using a sigmoidal dose-response nonlinear 1 regression model. Any sample considered negative was assigned an ED50 of 12.5, the 2 lowest dilution tested, because ED50 cannot be accurately calculated below the lowest 3 dilution tested. Zika-specific IgG binding was reported in Log10 ED50.

5
Behavioral observations. We employed a battery of age-appropriate behavioral tests that 6 are designed for use in infant nonhuman primates. These tests were performed to 7 identify any effects prenatal exposure to ZIKV. Both infants received neurobehavioral 8 tests modelled upon testing tools used for human infants [34,35] and adapted for use in 9 nonhuman primates [36]. Tests were administered every two weeks, from 14 days of 10 age until euthanasia at 20 (F10) or 21 weeks (F09). Each infant's scores were compared 11 descriptively against the mean and standard deviation across seven control animals 12 reared in the same fashion and tested by the same behavioral technician. Data for 13 control animals were available at three time points.

15
During the first month of life, a Neonatal Behavioral Assessment (NBA) tool was employed.

25
F09 was the only animal in our studies to clear blood viral RNA prior to day 5. In infant F10, the viral load remained below 1,000 copies per milliliter but remained detectable 1 until day 7 ( Figure 3A). These acute viral loads contrast with those of 11 pregnant 2 females infected with the same stock of the same strain of the virus at the same dose 3 and route ( Figure 3B) as well as four non-pregnant females ( Figure 3C) and four adult 4 males ( Figure 3D) infected with a separate stock of the same dose and strain of the 5 virus. Area under the curve (AUC) analyses showed that F09 had a total viremia lower 6 than all other animals in our previous studies with the exception of a single pregnant 7 female that had a slightly lower peak viremia and cleared virus from blood far earlier 8 than was typical for our pregnant animals. F10 had total viremia far lower than any 9 animal in any cohort tested at our facilities ( Figure 3E). At necropsy, we performed RT 10 PCR for ZIKV RNA on serum, CSF, multiple brain regions (frontal cortex, parietal lobe, 11 occipital lobe, temporal lobe, brain stem, optic nerve, cerebellum, choroid plexus, and 12 subcortical white matter), and axillary lymph nodes and virus was detected only in the 13 axillary lymph in both animals ( Figure 3F). These data contrast sharply from a recent 14 study that found infants born to healthy dams and infected postnatally showed viral loads 15 that peaked between 10^6 and 10^7 viral copies per milliliter, which is approximately one 16 log higher than that demonstrated by the adults and where viral RNA was detected in 17 several neurological sites two weeks post infection [32]. Neither infant in this study 18 showed signs of potential virus-induced pathology at necropsy. F10 harbored a choroid 19 plexus cyst that resulted in unilateral hydrocephalus in the brain, but such cysts are 20 common, are generally considered of little consequence and are not likely viral in origin.

22
Antibody responses. We next examined humoral responses in the infants to see if they 23 might explain the strikingly low viral loads. We used a plaque reduction neutralization 24 test (PRNT) to assess neutralizing antibodies in serum after birth and after infection in 25 both infants. Both showed detectable levels of neutralization at birth, which quickly waned below the limit of detection by 2 to 3 months. Neutralizing antibodies reemerged 1 after infection and continued to rise until euthanasia at 2 weeks post infection ( Figure   2 4A). To measure binding antibodies, we employed a whole virion ELISA assay using 3 plasma samples collected throughout the infants' lives both before and after infection.

4
Binding IgG titers decreased between birth and 3-4 months of age, consistent with the 5 expected kinetics of passively-transferred maternal IgG, but remained detectable until 6 viral inoculation at five months, and then rose after infection, similar to the neutralizing 7 antibody titers ( Figure 4B).

16
Here, we report the results of a small study describing results from two infant macaques 17 born to dams infected with ZIKV during the third trimester. One dam had relatively high

3
Both infants in our study harbored detectable levels of anti-ZIKV neutralizing antibodies 4 at birth that declined between one-and four-months post birth. We interpret these data 5 to suggest these antibodies were passively acquired from the dams as opposed to 6 mounted directly by the infants. ZIKV-binding IgG also declined after birth but remained 7 detectable between three and five months of age, when the animals were infected.

8
When we infected the infants with ZIKV, they exhibited low peak viremia that was rapidly

22
suggesting that even a brief period of serum viremia is sufficient for tissue dissemination, 23 which may result in consequences not tested in our study, including inflammation.
Taken together, our data suggest that infants born healthy to ZIKV infected mothers 1 maintain a level of protection from ZIKV that dampened acute viral loads and limited 2 tissue dissemination of the virus. We propose that passively acquired maternal