Investigating the Potential Role of North American Animals as Hosts for Zika Virus

Introduction

Zika virus (ZIKV) is an emerging mosquito-borne pathogen (genus Flavivirus, family Flaviviridae) related to viruses such as West Nile, yellow fever, and dengue virus. The positive-sense RNA virus was first isolated from a rhesus monkey in Uganda in 1947 (Dick et al. 1952) and since then sporadic human cases have been documented in Africa and Asia. The first major outbreak of ZIKV documented outside the continent of Africa occurred on Yap Island in 2007 (Duffy et al. 2009) and subsequent outbreaks were reported in French Polynesia and the Pacific Islands in 2013 (Musso et al. 2014).

The recent outbreak in Brazil began in 2015 and has rapidly spread to more than 40 countries throughout the Americas with reports of active transmission (Lucia 2016). The virus follows an epidemiological pattern in urban areas that is similar to that of chikungunya and dengue viruses (Musso et al. 2015), and there are concerns that ZIKV will move into North America and establish a sylvatic cycle, expanding the geographic range of the current epidemic. Therefore, it is important to understand the ecology and transmission routes of ZIKV to develop diagnostic and therapeutic countermeasures against the virus.

Currently little is known about the enzootic cycle for ZIKV and whether animal infections are important in transmission of the virus to humans. Nonhuman primates play an important role in maintaining a sylvatic cycle for arboviruses such as dengue, yellow fever, and chikungunya viruses (Inoue et al. 2003, Althouse et al. 2015). However, the epidemiology and potential animal reservoirs for ZIKV are currently not well understood.

The first reports of ZIKV included the isolation of the virus from a rhesus monkey in Africa in April 1947 (Dick et al. 1952, McCrae and Kirya 1982). Since then, in Africa, antibodies against ZIKV have been detected in goats, rodents, zebras, hippos, impalas, sheep, wildebeest, lions, water buffalo, and elephants (Fagbami 2009, Haddow et al. 2012). Other serological evaluations outside Africa identified antibodies in ducks, cows, bats, rats, goats, sheep, orangutans, horses, and buffalo (Darwish et al. 1983, Wolfe et al. 2001, Vorou 2016). More recently, a study in Brazil identified a possible reservoir in neotropical monkeys (Favoretto et al. 2016).

Competent mosquito vectors (Aedes aegypti and Aedes albopictus) (Marcondes and Ximenes Mde 2016) for ZIKV do exist in North America, leading to the concern that the virus will establish itself in the continental United States and potentially develop a sylvatic transmission cycle. Therefore, it is critical to understand the epidemiology and ecology of ZIKV by identifying potential animal reservoirs in North America.

Here, we describe a series of experimental infections designed to investigate whether animals with certain taxonomic groups in North America have the potential to serve as ZIKV amplifying hosts or reservoirs. The objectives of this study were to identify potential animal reservoirs within North America with a focus on wildlife and livestock, and to examine ZIKV pathogenesis in neonatal and adult animals. The results from this study gave insight into the likelihood that ZIKV will have an amplifying host/reservoir and establish an endemic cycle or follow more of a cyclical re-emergence.

Materials and Methods

Animals selected for the study represented a diversity of species found in North America with ages ranging from neonates to adults (Table 1), and were obtained from commercial sources (chickens, ducks, frogs, mink, snakes, and hamsters), wild caught (armadillos, rabbits, pigeons, groundhogs, deer mice, raccoons, and sparrows), and from local suppliers (calves and goats). All animals were housed in the animal biosafety level 3 facility at Colorado State University for the duration of the study, and all procedures were approved by the Colorado State University Institutional Animal Care and Use Committee (14-5140A).

Two strains of ZIKV were used for the inoculation: FSS13025, isolated from a human in Cambodia in 2010, and PRVA BC59, isolated from a human from Puerto Rico in 2015. These strains of virus were selected as recent isolates of the Asian lineage of ZIKV. Animals were challenged with one of two ZIKV strains by subcutaneous and intradermal injection of ∼10⁵ plaque-forming units (pfu) of virus. This relatively high dose was chosen to ensure against false negative assessment of host susceptibility.

All animals were monitored daily for overt clinical signs of disease. For small animals (deer mice, hamsters, garter snakes, and frogs), blood was collected from half of each animal group every day for 7 days, then on 14, 21, and 28 days postinoculation (dpi); for all other animals, blood samples were collected daily from every animal on days 0–5, then 7, 14, 21, and 28 dpi. For these animals, body temperatures were also recorded at the time of each blood sample collection. Level of infectious virus in sera and tissues (kidney, liver, spleen) was determined by plaque assay on Vero cells (Westaway 1966), and RT-PCR was performed on selected samples to detect viral RNA in sera and tissues using primers to the Zika 3′ UTR (forward 5′-GGACTAGTGGT TAGAGGAGACCC-3′ and reverse 5′-CGTGGTGGAAAC TCATGGAGT-3′) (GenBank KX601168.1). Neutralizing antibodies were assayed by plaque reduction neutralization test (PRNT) using a 90% cutoff (Lindsey et al. 1976).

Results

Overt clinical signs, including fever, were not detected in any of the animals during the study. Results of virus and antibody titer detection for all species are outlined in Table 1. Infectious virus was detected by plaque assay only in sera from armadillos and frogs with viral titer ranging from 2 × 10² to 2.2 × 10³ pfu/mL. Viremia was detected in some frogs between days 1 and 3, but not thereafter. Virus was not detected by plaque assay in kidney, liver, or spleen samples collected at necropsy from any animal. ZIKV RNA was detected in frog and armadillo sera at 2 dpi by RT-PCR, confirming the plaque assay results, but not in any other animal at 2 dpi. All sera collected before inoculations were uniformly negative for neutralizing antibodies to ZIKV by PRNT, but postinoculation, antibodies were detected in individual goats, pigs, ducks, frogs, and cottontail rabbits by 7 dpi (Tables 1 and 2). One pig had detectable antibodies as early as 5 dpi. Antibody responses were more robust at 7 dpi with titers dropping by 28 dpi.

During the challenge study, we observed unexpected births from the wild caught cottontail rabbits (three total) and deer mice (one total). Based on the time from virus inoculation to birth, the pregnant animals were experimentally challenged in the second or third trimester. The mouse births included five offspring in the litter. For the rabbits, two live births included five to six offspring per litter and a stillbirth was observed with the third rabbit. Neither infectious virus nor viral RNA was detected in sera from the stillborn offspring, and maternal antibodies were not detected for any of the offspring by PRNT. Overt neurological disorders were not observed in liveborn offspring and gross abnormalities were not identified at necropsy in any of these offspring from infected mothers.

Discussion and Conclusions

ZIKV continues to spread throughout the Americas and will likely become endemic. It has been observed that ZIKV follows transmission patterns similar to chikungunya and dengue viruses, leading to concerns that ZIKV will continue to spread globally (Reisen 2013, Musso et al. 2015) Predicting emergence patterns and developing countermeasures against urban transmission are critical to understanding the transmission cycles of these arboviruses between humans and animals. The purpose of this study was to identify potential animal reservoirs and/or amplifying hosts in North America that could sustain the transmission cycle of ZIKV. Our results support previous findings of antibody detection in goats and ducks, and also identified that pigs, rabbits, and frogs seroconverted after infection. Furthermore, this study identified low levels of infectious virus in frogs and armadillos.

There was no apparent influence of age on ZIKV susceptibility between juvenile and adult animals. In addition, there were no signs of neonatal abnormalities in offspring born after inadvertent experimental infection of pregnant deer mice and cottontail rabbits. However, these observations are limited to viral exposure later in gestation compared with exposure during early gestation, where the risk of congenital infections may be higher.

A limitation of this study was the small number of animals per species; thus, a more extensive evaluation of those animals that developed antibody titers and a detectable viremia is warranted. Specifically, the potential role of amphibians as a reservoir needs further evaluation. Although the viremia titers observed in frogs were low and likely below the threshold required for efficient transmission to mosquitoes, it could be that other species of amphibians would have a higher magnitude of viremia. The potential importance of amphibians in Zika ecology would also require bridge vectors that feed on both amphibians and primates.

The results from these experiments indicate that none of the animals tested are likely to be competent hosts for ZIKV, but rabbits and pigs may potentially serve as sentinel species in North America where virus is transmitted by A. albopictus, which will feed on these species. Use of sentinel species as a surveillance tool, as was previously demonstrated with birds and West Nile virus (Komar 2001), is of great value in monitoring rapid global spread of zoonotic pathogens (Halliday et al. 2007). The results of this study contribute to our understanding of viral dynamics as well as developing appropriate surveillance programs and countermeasures for ZIKV in North America.