Serological Positivity against Selected Flavivirus and Alphavirus in Free-ranging Bats and Birds from Costa Rica Evidence Exposure to Arboviruses Seldom Reported Locally in Humans

Arboviruses have two ecological transmission cycles, sylvatic and urban. For some, the sylvatic cycle has not been thoroughly described in America. To study the role of wildlife in a putative sylvatic cycle, we sampled free-ranging bats and birds in two arbovirus endemic locations and analyzed them using molecular, serological, and histological methods. No current infection was detected, and no significant arbovirus-associated histological changes were observed. Neutralizing antibodies were detected against selected arboviruses. In bats, positivity in 34.95% for DENV-1, 16.26% for DENV-2, 5.69% for DENV-3, 4.87% for DENV-4, 2.43% for WNV, 4.87% for SLEV, 0,81% for YFV, 7.31% for EEEV, and 0.81% for VEEV was found. Antibodies against ZIKV were not detected. In birds, PRNT results were positive against WNV in 0.80%, SLEV in 5.64%, EEEV in 8.4%, and VEEV in 5.63%. An additional retrospective PRNT analysis was performed using bat samples from three additional DENV endemic sites resulting in a 3.27% prevalence for WNV and 1.63% for SLEV. Interestingly one sample resulted unequivocally WNV positive confirmed by serum titration. These results suggest that free-ranging bats and birds are exposed to not currently reported hyperendemic-human infecting Flavivirus and Alphavirus, however, their role as reservoirs or hosts is still undetermined.

Arboviruses are maintained in nature through two different transmission cycles, sylvatic and urban [7]. In a sylvatic transmission cycle, a vertebrate reservoir or host capable of maintaining high levels of viremia to become a source of infection for the arthropod vectors is required [7]. In Costa Rica, clinical infections caused by DENV, CHIKV, and ZIKV in human beings have been reported circulating every year in a sustained manner [8][9][10]. Also, there is serological evidence of SLEV, VEEV, EEEV, and YFV in some wild and domestic animals [11,12,13], and molecular evidence of other arboviruses as well [14]. Thus, Costa Rica is a hyperendemic country for the presence of arbovirus and there is a continuous increasing interest in the search for possible reservoirs in wildlife.
Bats and birds have a wide distribution and diversity in Costa Rica. Bats represent 45% of the total species of mammals in the country and encompass 115 different species belonging to 9 different families [15]. Costa Rica also possesses an extraordinary diversity of birds, over 922 different species have been reported in the territory, including also migratory birds [16].
For some arboviruses, birds are involved in the sylvatic cycle [17]. Furthermore, bats have been also implicated in maintaining arboviruses in the wild, though inconclusively [18].
To study the role of wildlife in a putative sylvatic cycle in our country, we sampled free-ranging bats and birds in two arbovirus endemic locations and analyzed them using molecular, serological, and histological methods. No current infection was detected, and no significant arbovirus-associated histological changes were observed, though neutralizing antibodies were detected against several hyperendemic arboviruses but also several non-currently circulating in humans arboviruses in our country such as WNV. Our results suggest that free-ranging bats and birds are exposed to not hyperendemic-human infecting Flavivirus and Alphavirus, however, their role as reservoirs or hosts is still undetermined.

Materials and Methods
Between August 2017 and November 2018, and extensive sampling was performed in two different locations in Costa Rica, Santa Cruz, Guanacaste (10°16′00″N 85°39′00″W) and Talamanca, Limón (09°44'28"N, 82°50'46"W) (Figure 1). At each site, eight households (sampling units) were selected and sampled. Birds and bats were collected using mist nets between 15:00-17:00 and 18:00-20:00, respectively. At least five birds and ten bats from each sampling unit were taxonomically identified [19,20] measured, weighed, and euthanized by an intramuscular anesthesia overdose [21] (Supplementary Tables 1-4). All the protocols were approved by the Institutional Committee for the Care and Use of Animals from the University of Costa Rica (CICUA-042-17), Committee of Biodiversity of the University of Costa Rica, and the collection permits from the National System of Conservation Areas (SINAC ACT-PIM-070-17, R-SINAC-PNI-ACLAC-054-2018). After the euthanasia, a blood sample was taken from the heart. The sample was centrifuged, and the serum was stored at 4 ºC until it arrived at the laboratory. Also, a complete necropsy procedure was performed in each animal; all tissues were collected and placed in 10% buffered formalin. Additionally, individual samples of lung, spleen, heart, kidney, liver, brain, eye, knee, intestine, stomach (ventriculus/proventriculus), urinary bladder, and gonads were taken and placed in RNAlater (Life Technologies, Thermo Fisher Scientific Inc.) and held at -80°C. A pool of organs also was taken and placed in RNA later. RNA extraction was performed using TRIzol reagent (Invitrogen, Carlsbad CA, USA), following the manufacturer's instructions. Complementary DNA (DNAc) was obtained using RevertAidTM RT Kit K#1691 (Thermo Scientific, MA, USA) according to the manufacturer's instructions.
For the identification of viral ARN, a generic RT-PCR using the protocol designed for amplification of partial NS5 segment of Flavivirus [22] and the partial amplification of nsP4 for Alphavirus [23] was performed. The amplicons were analyzed by electrophoresis and interpreted using the equipment QIAxcel Advanced (QIAGEN) and the Software QI-Axcel ScreenGel 1.6.0.
A plaque reduction neutralization test (PRNT) assay was performed with a unique serum dilution (1:10) of each sample to detect neutralizing antibodies. Neutralization assays were performed in 96-well, flat-bottomed tissue culture plates because of the limited sera volume. Briefly, each sample was heat-inactivated for 30 minutes at 56°C.). The serum was . The virus-serum mix was incubated 1h at 37°C. Then, 50 µL of this mixture was added to Vero Cells monolayer previously seeded and incubated for 1h at 37°C in a 5% CO2 atmosphere. After adsorption, the mixture was removed from Vero cells, and 100 µL of MEM with 2% of FBS and 1.5% of carboxymethylcellulose was added. (Sigma-Aldrich, USA). Plates were incubated for 72h at 37°C in a 5% CO2 atmosphere. After the incubation period, MEM containing CMC was removed, and plates were fixed during an hour with formalin (3.7%) and stained with crystal violet (1%). A 90% reduction of foci number relative to the average of the viral control (no sera) was considered positive.
A total of 140 birds were collected (89 males and 51 females), representing 37 genera and 43 species (Supplementary Tables 3-4). Serum neutralization analysis (PRNT 1:10) was performed against WNV and SLEV in 124 serum samples and against EEEV and VEEV in 71 serum samples due to sample volume constraints. In birds, PRNT resulted positive against WNV in 0.80% (1/124), SLEV in 5.64% (7/124), EEEV in 8.4% (6/71), and VEEV in 5.63% (4/71). Cross-reactions were also observed in 2 individuals presenting antibodies against EEEV and VEEV (Table 4). All RT-PCR analyses performed in both blood and pool tissues collected from bats and birds were negative for the detection of viral ARN from flaviviruses and alphaviruses. Histopathological findings were nonspecific, inflammatory, and degenerative changes frequently associated with the presence of diverse parasites (protozoa, metazoans). Thus, histological analysis of tissues did not show any significant findings related to arboviral infections

Discussion and Conclusions
The sylvatic cycle from some arboviruses has not been thoroughly described. Birds have been implicated in participating in the virus cycle of some arboviruses [30][31][32][33]. Nevertheless, even though representing one of the most abundant mammals and diverse species in many tropical and neotropical areas, bats are still not directly implicated as part of the virus cycle of the most important arboviruses infecting humans such as DENV, ZIKV, and WNV, but have been described as accidental hosts [18,24]. In many areas, the identification of wildlife susceptible to infection and thus been able to participate as hosts or reservoirs is not fully elucidated, and surely, just identifying exposure does not translate into being part of the virus cycle. Nevertheless, because of the importance of this identification and to further study the role of wildlife in a putative sylvatic cycle in our country, we sampled free-ranging bats and birds in two arbovirus endemic locations and analyzed them using molecular, serological, and histological methods. No current infection was detected, and no significant arbovirus-associated histological changes were observed. Though neutralizing antibodies were detected against several arboviruses. These results demonstrate that free-ranging bats and birds in our country are exposed to not hyperendemic-human infecting Flaviviruses and Alphaviruses.
PRNT is the gold standard for the serological diagnosis of previous arboviral infection [17]. Using PRNT with chimeric viruses we demonstrated the presence of antibodies against several Flaviviruses (DENV1-4, SLEV, YFV, WNV) and Alphaviruses (EEEV, VEEV) in bats and birds in our country. However, as expected cross-reactions with different Flaviviruses and Alphaviruses were prevalent in our microneutralization assays. Cross-reactions make it difficult to establish unequivocally the exposure to the etiologic virus. This phenomenon is explained by the similarities between the immunodominant epitopes within the family Flaviviridae and Togaviridae [17]. The envelope structural glycoprotein is the main target for the neutralizing antibodies in Flaviviridae and E2 structural protein for Togaviridae [34]. To determine against which specific arbovirus the animal was exposed, titration of the serum is necessary [17]. Unfortunately, the scant amount of serum available in most of our samples did not allow this. However, studies suggest that the serological evidence in a PRNT allows determining that bats were exposed against a given arbovirus in a period between 10 days to months before the analyses [17]. In the retrospective serological study with samples that were preliminary positive for DENV, serum titration allowed the identification and confirmation of a single bat from an urban area (Valle Central) unequivocally positive to WNV (Molossus sinaloe) with a titer >1:640 [35].
On the other hand, viral RNA was not evidenced in the blood or tissues from any of the sampled animals. Because arboviral infections are generally transient, the chance of detecting viral RNA from sera or tissues is extremely low [36]. In this study, we did not detect any active viremia in bats and birds. This is not the first evidence of arboviruses in bats in Costa Rica. A previous study demonstrated serological prevalence in 21.2% (51/241) and positivity by RT-PCR in 8.8% (28/318) bats for DENV RNA [24]. The retrospective PRNT assays were performed with the samples obtained from this previous project. However, these results contrast with our findings, in which we cannot evidence viral RNA in the animal blood and/or tissues. In the previous report, it was speculated ingestion of mosquitoes as the route of infection, since most of the collected samples were from insectivorous animals. Bats as a dead-end host for DENV were then suggested [24]. The majority of the bats collected in this study are frugivorous, and we did not detect any viral RNA in any organ including intestines. Furthermore, during histological evaluation on the intestines no evidence of ingested arthropods was observed. Thus, infection through the ingestion of mosquitoes might be unlikely. Mosquito feeding therefore seems to be the more plausible route of infection.
In other locations in the American continent, bats have been subjected to extensive evaluations for arbovirus infection with different outcomes. In Brazil, 103 samples from free-ranged bats were analyzed via nested-PCR and hemagglutination inhibition test. All the samples were negative for arbovirus; thus, the authors concluded that bats do not constitute a reservoir for these viruses in the studied areas [37]. In Trinidad, a serological analysis demonstrated a prevalence of 2.9% (11/384) against VEEV using an ELISA in bats [36]. In Colombia, VEEV viral RNA and antigen detection were demonstrated via RT-PCR and immunohistochemistry (IHC)in the brain tissues from two frugivorous bats, suggesting that these species might be hosts for this zoonotic agent [38]. Finally, a new strain of encephalitogenic VEEV (Tonate Virus) was isolated from bats in French Guiana [39]. Neutralizing antibodies against SLEV have been reported in bats from Ohio, USA? [40]. WNV has never been reported in bats in Central America [41], there are only two reports of the presence of neutralizing antibodies against WNV in the USA [42,43]. YFV and ZIKV infection has not been reported so far in bats in America [44]. We found only one bat with neutralizing antibodies against YFV, but with a clear cross-reaction against DENV-1, DENV-2 and WNV. Only one of the three bats positive against WNV was positive exclusively and unequivocally to this agent. In the retrospective assay, one Molossus sinaloe was confirmed through serum titration (1:640) as seropositive against WNV. Thus, this is the first evidence of WNV exposure in bats in Central America. This is the first time that SLEV, WNV, EEEV, and VEEV exposure have been demonstrated in wild birds in Costa Rica. Overall reported seroprevalence against those agents is low in America. A study conducted in Yucatán México revealed a seroprevalence of 4.3% (11/257) against WNV through ELISA [30]. In Trinidad, one study reported a seroprevalence of 1.4% (2/140) against SLEV through hemagglutinin inhibition test [45]. WNV and SLEV can cross-react [31] and serum titration was not performed. Only one species, Empidonax virescens is migratory and have neutralizing antibodies against SLEV and cross reaction between VEEV and EEEV: suggesting a non-local exposure. None of the species identified in this study has been reported as a reservoir for SLEV and WNV [17]. Therefore, we conclude that some of our sampled birds were exposed to these viruses, but since we did not detect any active viremia, we were not able to identify the putative reservoirs for these viruses in our country. Thus, further sampling efforts in exemplar numbers and areas must be conducted.
Taken together, we can conclude that free-ranging bats and birds are exposed to not currently reported hyperendemic-human infecting Flaviviruses such as WNV and some Alphaviruses, however, their role as reservoirs or hosts is still undetermined. They may have a role as a dead-end host, but no evidence supports that the sampled species during this study might function as reservoirs.