Date Published: January 31, 2017
Publisher: Public Library of Science
Author(s): Lark L. Coffey, Patricia A. Pesavento, Rebekah I. Keesler, Anil Singapuri, Jennifer Watanabe, Rie Watanabe, JoAnn Yee, Eliza Bliss-Moreau, Christina Cruzen, Kari L. Christe, J. Rachel Reader, Wilhelm von Morgenland, Anne M. Gibbons, A. Mark Allen, Jeff Linnen, Kui Gao, Eric Delwart, Graham Simmons, Mars Stone, Marion Lanteri, Sonia Bakkour, Michael Busch, John Morrison, Koen K. A. Van Rompay, Karol Sestak.
Animal models of Zika virus (ZIKV) are needed to better understand tropism and pathogenesis and to test candidate vaccines and therapies to curtail the pandemic. Humans and rhesus macaques possess similar fetal development and placental biology that is not shared between humans and rodents. We inoculated 2 non-pregnant rhesus macaques with a 2015 Brazilian ZIKV strain. Consistent with most human infections, the animals experienced no clinical disease but developed short-lived plasma viremias that cleared as neutralizing antibody developed. In 1 animal, viral RNA (vRNA) could be detected longer in whole blood than in plasma. Despite no major histopathologic changes, many adult tissues contained vRNA 14 days post-infection with highest levels in hemolymphatic tissues. These observations warrant further studies to investigate ZIKV persistence and its potential clinical implications for transmission via blood products or tissue and organ transplants.
Emerging mosquito-borne Zika virus (ZIKV, Flaviviridae, flavivirus) was first detected in Brazil in 2015 and has since spread to at least 60 countries in South and Central America, Oceania and Asia . In 2015 and 2016, more than 4,800 imported cases have been reported in travelers returning to the continental United States. An additional 35,000 cases were reported in US Territories. Local mosquito-borne transmission in Florida was first documented in July 2016 with 210 local cases reported as of December 28, 2016 . ZIKV is also transmitted sexually, via blood transfusion, and possibly organ transplants [3–6]. The World Health Organization declared the ongoing ZIKV outbreaks a public health emergency on February 1, 2016  based on explosive geographic spread, clinical and epidemiological associations with microcephaly [8–11] and other neurological defects (reviewed in ), and lack of a licensed vaccine or specific therapies. ZIKV was historically considered a mild febrile illness with very little research dedicated to understanding infection dynamics and mechanisms of disease; therefore, very little is known about human ZIKV infections. Similar to other viral diseases, the availability of relevant animal models would be extremely useful for understanding human ZIKV infection dynamics and pathogenesis and to test intervention approaches such as vaccines, drugs, or other strategies with a goal of interrupting all modes of ZIKV transmission and treating infection. Such pre-clinical screening studies can guide clinical trials with the ultimate goal of ending the ZIKV pandemic. New murine models developed since March 2016 require immunodeficient mice [13–20], which is a major limitation. Since fetal development and placental biology of humans is more similar to non-human primates (NHP) than to rodents, NHP may better model adult infections and ZIKV-induced fetal neuropathogenesis than mice. Except for the first isolation of ZIKV in a rhesus macaque in Uganda in 1947 and experimental infection of several macaques with mouse-brain passaged ZIKV in the 1950s [21, 22], there were no published reports of ZIKV infection of non-human primates, and no animals had been experimentally inoculated with contemporary outbreak strains until 2016. In recent months, several studies using Asian or South-American lineage strains injected subcutaneously in rhesus or cynomolgus macaques showed infection marked by viremia, and that prior infection or immunization provided protection from subsequent challenge [23–26]. In this study, we monitored and euthanized 2 non-pregnant macaques 14 days post-intravenous inoculation of a 2015 Brazilian strain of ZIKV to characterize clinical disease, the time course of virus levels in plasma, whole blood, urine, and saliva. We also measured the kinetics of neutralizing and binding antibodies and ZIKV RNA distribution and histopathologic changes in tissues. We demonstrate that viral RNA could be detected longer in whole blood than in plasma and that despite no major histopathologic changes, many macaque tissues contained viral RNA 14 days post-infection.
After intravenous inoculation of ZIKV, the 2 rhesus macaques developed a short viremia, with plasma vRNA levels peaking at 5.7 log10 RNA copies/ml and remaining detectable to 7 dpi in plasma and 10 dpi in whole blood. Infectious virus was detectable in plasma to 4 dpi. Clearance of ZIKV from plasma coincided with the development of robust neutralizing antibody responses beginning 5 dpi that peaked at 2460 on 14 dpi when the experiment was terminated. The kinetics of vRNA in plasma, saliva, and urine in these macaques following intravenous inoculation with Brazilian ZIKV parallel data from other recent rhesus macaque studies [23–26] that used ZIKV isolates from areas other than Brazil and that were inoculated subcutaneously. In one of those studies , which used the same inoculation dose, qRT-PCR assay, and plasma volumes, plasma vRNA was detected intermittently at low levels to 21 dpi in non-pregnant animals; by contrast, neither of the adult animals in our study had detectable plasma vRNA at >0.7 log10 copies/ml, the Aptima® cut-off value, after 10 dpi. These differences may be due to a combination of inoculation route, methodological, viral, and host factors. Our intravenous inoculation likely resulted in antigen presentation to many lymph nodes simultaneously, possibly promoting faster innate immune responses with clearance of circulating vRNA, whereas subcutaneous inoculation used in the other study may have resulted in slower dissemination through the draining lymph node, as indicated by the broader window of peak viremia, 2 to 6 dpi, observed in that study. Another possible explanation for the disparity in length of plasma vRNA detection across the 2 macaque studies is differences in the 2015 Brazil and 2013 French Polynesia ZIKV genomes, although Asian lineage outbreak strains share >99% genome-wide nucleotide identity [31, 32]. Lastly, considering the small animal numbers in these pilot studies, other host factors including macaque genetics or other intra-animal differences, could have contributed. Control non-vaccinated macaques in two vaccine studies that were challenged with 3 or 6 log10 2015 Puerto Rico ZIKV or a different Brazilian ZIKV strain than we used here (BeH815744, accession number KU365780) developed vRNA levels in plasma for 6 to 7 days that peaked 3 to 5 dpi at about 6 log10 copies/ml [25, 26], consistent with our observations. The pattern of transient viremia in all of the macaque studies and the absence of clinical disease including febrile responses parallels observations from humans where the majority of infections are sub-clinical [33–35].