Serological Evidence of Multiple Zoonotic Viral Infections among Wild Rodents in Barbados

Background: Rodents are reservoirs for several zoonotic pathogens that can cause human infectious diseases, including orthohantaviruses, mammarenaviruses and orthopoxviruses. Evidence exists for these viruses circulating among rodents and causing human infections in the Americas, but much less evidence exists for their presence in wild rodents in the Caribbean. Methods: Here, we conducted serological and molecular investigations of wild rodents in Barbados to determine the prevalence of orthohantavirus, mammarenavirus and orthopoxvirus infections, and the possible role of these rodent species as reservoirs of zoonotic pathogens. Using immunofluorescent assays (IFA), rodent sera were screened for the presence of antibodies to orthohantavirus, mammarenavirus (Lymphocytic choriomeningitis virus—LCMV) and orthopoxvirus (Cowpox virus—CPXV) infections. RT-PCR was then conducted on orthohantavirus and mammarenavirus-seropositive rodent sera and tissues, to detect the presence of viral RNA. Results: We identified antibodies against orthohantavirus, mammarenavirus, and orthopoxvirus among wild mice and rats (3.8%, 2.5% and 7.5% seropositivity rates respectively) in Barbados. No orthohantavirus or mammarenavirus viral RNA was detected from seropositive rodent sera or tissues using RT–PCR. Conclusions: Key findings of this study are the first serological evidence of orthohantavirus infections in Mus musculus and the first serological evidence of mammarenavirus and orthopoxvirus infections in Rattus norvegicus and M. musculus in the English-speaking Caribbean. Rodents may present a potential zoonotic and biosecurity risk for transmission of three human pathogens, namely orthohantaviruses, mammarenaviruses and orthopoxviruses in Barbados.


Introduction
Rodent species are the most abundant terrestrial animals (representing 43% of mammalian species) found globally and exist as a part of biological ecosystems, however they do pose significant threats to public health [1,2]. The threat of zoonotic infections from wild rodents is well-known, with the risk of infections involving various species of bacteria, fungi, parasites and viruses [1]. Most rodent-borne infections are transmitted through direct or indirect contact with infected rodents and their excreta [1,2]. For this reason, factors such as rainfall, topography, vegetation and occupation, which influence human exposure

Wild Rodents Trapping Survey
To understand the possible rodent reservoirs of orthohantaviruses, mammarenaviruses and orthopoxviruses in Barbados, a rodent trapping and sampling survey was conducted in 2019. A total of 160 rodents were trapped over 10 trapping nights from 15th January to 26th January 2019, at a total of 15 trapping sites around Barbados including chicken farms, recycling centres, horse stables, an agriproducts retail store, residential neighbourhoods, the national geriatric hospital, and sugarcane fields in parishes where previously recorded human orthohantavirus cases occurred (Table 1 & Figure 1) [30].  58 70 Key: N-non-scrotal male; SC-scrotal; PR-pregnant; NP-non-parous; *-with 4 female Mus musculus and 1 female Rattus norvegicus rodents, their reproductive status was either indiscernible or inadvertently not recorded. + -With one (1) Mus musculus rodent, the observed gender identification was not recorded.

Wild Rodents Trapping Survey
To understand the possible rodent reservoirs of orthohantaviruses, mammarenaviruses and orthopoxviruses in Barbados, a rodent trapping and sampling survey was conducted in 2019. A total of 160 rodents were trapped over 10 trapping nights from 15th January to 26th January 2019, at a total of 15 trapping sites around Barbados including chicken farms, recycling centres, horse stables, an agriproducts retail store, residential neighbourhoods, the national geriatric hospital, and sugarcane fields in parishes where previously recorded human orthohantavirus cases occurred (Table 1 & Figure 1) [30].   76.5% (52/68) were non-parous and with 5.9% (4/68) of the female M. musculus rodents the reproductive maturity was either indiscernible or not recorded (Table 1).

Orthohantavirus IFA & RT-PCR Testing of Wild Rodents
Dried rodent blood was used to obtain rodent sera in all but two (2) rodents, where hearts were used to obtain the sera. Screening of sera from wild rodents trapped in seven different parishes in Barbados was conducted using IFA to identify seropositive rodents with orthohantavirus-specific IgG antibodies ( Table 2, Figures 1 and 2). Of the 160 rodents tested, 3.8% (6/160) were orthohantavirus IFA-positive. For mice, Mus musculus, 4.0% (6/150) were orthohantavirus IFA-positive, while for both rat species, R. norvegicus and Rattus rattus, none of them (0% (0/8) and 0% (0/2) respectively) were orthohantavirus IFA-positive (Table 2). The lack of orthohantavirus seropositivity among rats in this study is notable in comparison to a previous rodent survey in Barbados [31]. The orthohantavirus seropositive rodents identified in this current study were trapped at sites in St. John (66.7%, 4/6), St. Michael (16.7%, 1/6) and Christ Church (16.7%, 1/6), including residential neighbourhoods, a local geriatric hospital, sugarcane fields and chicken farms ( Figure 1). The two rodent sera (obtained from hearts) were both seronegative for all viral pathogens under investigation.
Tissues and sera from seropositive wild rodents were then tested using orthohantavirusspecific RT-PCR. Molecular investigations failed to detect orthohantavirus vRNA among wild rodents in Barbados, as none of the seropositive rodent sera nor tissues yielded positive orthohantavirus-specific RT-PCR results ( Table 2).
Mammarenavirus serosurveys of wild rodents in Barbados and in the Caribbean are non-existent. Thus, screening of wild rodent sera was conducted using mammarenavirus-specific IFA to identify seropositive rodents with orthohantavirus-specific IgG antibodies (Table 2). Of the 160 wild rodents trapped during the study, 2.5% (4/160) were seropositive for mammarenavirus infection using LCMV-specific IFA testing (Table 2). Among M. musculus rodents, 2.7% (4/150) were seropositive whilst 0% (0/8) of R. norvegicus and 0% (0/2) R. rattus rodents were seropositive for mammarenavirus infection ( Table 2). All the mammarenavirus-seropositive rodents were trapped in sugarcane fields and at a recycling centre in St. John (Table 2 & Figure 1). Tissues and sera from seropositive wild rodents were then tested usin orthohantavirus-specific RT-PCR. Molecular investigations failed to detec orthohantavirus vRNA among wild rodents in Barbados, as none of the seropositiv rodent sera nor tissues yielded positive orthohantavirus-specific RT-PCR results ( Table 2   Table 2. Serological and molecular survey study of wild rodents in Barbados.

Rodent Species
No. of Trapped Rodents Mammarenavirus serosurveys of wild rodents in Barbados and in the Caribbean ar non-existent. Thus, screening of wild rodent sera was conducted using mammarenaviru After identification of seropositive wild rodents using IFA, their sera and tissues were screened for mammarenavirus-specific vRNA using RT-PCR ( Table 2). None of the seropositive rodent sera nor tissues were positive following mammarenavirus-specific RT-PCR; thus, these molecular investigations failed to detect the presence of mammarenavirus vRNA among wild rodent species in Barbados (Table 2).

Discussion
Orthohantavirus, mammarenavirus and orthopoxvirus studies on rodents in Caribbean countries have been sparse or non-existent, and we present the first serological evidence of multiple orthohantavirus, mammarenavirus and orthopoxvirus infections in wild rodent species in both Barbados and the English-speaking Caribbean. Seroprevalence rates of 3.8% for orthohantavirus, 2.5% for mammarenavirus and 7.5% for orthopoxvirus infections were observed among wild rodents in Barbados during this study. This has potential biosecurity implications for farms, hospitals, residential homes, businesses involved in recycling and persons working in and living around sugarcane fields.
The last report of orthohantavirus infection among wild rodents in Barbados was in 2002, thus more updated data were required to understand the orthohantavirus infection status of wild rodents [31]. This study included wild M. musculus rodents to test the hypothesis of M. musculus as possible rodent vectors of orthohantavirus infections in Barbados and represents the first serological evidence of orthohantavirus infection of M. musculus in the Caribbean and the first evidence of mammarenavirus infection and orthopoxvirus infection among wild rodents in the Caribbean.

Orthohantavirus Infection among Wild Rodents
Given the known rodent hosts present in Barbados, Rattus and Mus spp., SEOV infections would be expected given its worldwide distribution [32]. Orthohantavirus prevalence among wild rodents varies by geographic location from 2.5% in Brazil, 5.6% to 7.7% in Paraguay, 4.6% to 5.6% and as high as 23.5% in Panama, and even higher rates of 29% and 28% in Grenada and Barbados respectively [31,[33][34][35][36][37]. The overall orthohantavirus seropositivity rate of 3.8% observed among all rodents in this study, and among R. norvegicus (0%) and M. musculus (4%), were like those observed in the USA, Kuwait, China and Europe [38][39][40][41]. The lack of orthohantavirus seropositivity among wild rats in this study may be due to the low numbers of rats trapped and the low percentage of males, as adult male rats are known to have higher orthohantavirus seroprevalence rates [13]. Orthohantavirus seropositive M. musculus rodents were identified by IFA testing, highlighting the possible risk these may play in orthohantavirus transmission in Barbados. The lack of genetic evidence of orthohantavirus infection among wild rodents, and the lack of demonstration of onward transmission, limits the conclusion of their potential as orthohantavirus reservoirs in Barbados. Limited orthohantavirus research has been conducted in the Caribbean. The first serological evidence of orthohantavirus infections in the Caribbean involved the detection of anti-orthohantavirus antibodies in suspected leptospirosis patients and rodents in Barbados [31]. In this study, 12% of 60 patients presenting with febrile illness possessed orthohantavirus-specific immunoglobulin M (IgM) [31]. The identity of the circulating orthohantavirus strain(s) and their source in Barbados has remained unknown. Evidence of Maporal, Caño Delgadito and Araraquara orthohantavirus infections in rodents, and the presence of multiple rodent hosts in Venezuela, Brazil and Columbia [42][43][44][45] along with a recent HPS outbreak in French Guiana in 2016, enhance the risk of new and more lethal orthohantavirus strains entering the Caribbean region via international and inter-regional trade and travel [46].
Orthohantavirus transmission is influenced by environmental and climatic factors including rainfall, topography and vegetation [3,4,47]. High rainfall can be associated with increased orthohantavirus transmission, as higher infection rates were observed during the wet season compared to the dry season since rainfall can permit moist soil which facilitates rodent burrowing, breeding, survival and the proliferation of vegetation and food for rodents [4,47]. The rainy season in Barbados has been associated with higher human orthohantavirus prevalence rates, with peaks occurring in the months of August and September [30]. A lower orthohantavirus seroprevalence of 3.8% was observed in this study compared to the previous rodent study, with a 29% seroprevalence rate. This may be due to a few differences between the former and currently reported rodent serosurveys, namely (1) sample size, 68 rodents vs. 160 rodents, (2) the proportion of rodent species, predominantly Rattus spp. and no M. musculus rodents vs. predominantly M. musculus rodents, and (3) timing of sampling, wet season (July-August 2000) vs. dry season (January 2019) [31]. These differences may have resulted in the lower orthohantavirus seroprevalence rate observed, and the lack of detection of orthohantavirus vRNA among wild rodents.
Conversely, excessive rainfall and/or extreme weather events including flooding can result in the reduction of rodent populations, reduced risk of orthohantavirus transmission, and reduced orthohantavirus seroprevalence [48]. However, no major flooding events were recorded during 2018 in Barbados. Other climatic factors influencing orthohantavirus transmission include atmospheric moisture variability and temperature [48], so tropical climatic conditions such as high temperature and humidity in Barbados could influence the survival of orthohantaviruses in the environment and their transmission. Further research on the influence of abiotic factors in orthohantavirus transmission is therefore necessary to understand the orthohantavirus ecology in Barbados and the Caribbean. Future studies to obtain genetic evidence of the orthohantavirus strain(s) present are necessary to determine the true reservoir status of rodents, and perhaps should be conducted in synch with peak periods of human orthohantavirus infections to permit higher chances of obtaining genetic sequences from rodents. The inclusion of bats in future rodent serosurveys should also be strongly considered given the detection of orthohantaviruses in various regions around the world [49][50][51][52][53].
The specific mammarenavirus infecting M. musculus in this study is unknown, as antibody cross-reactivity can be observed among mammarenaviruses and only FRNT and genetic evidence can confirm the identity of mammarenaviruses [70,71]. One possible mammarenavirus that may be present among wild M. musculus in Barbados is LCMV, which is a teratogen causing birth defects when acquired by pregnant mothers. Though LCMV has a world-wide distribution, viruses of the LCM-LAS complex are found primarily in Africa, whereas the Tacaribe complex viruses are found in North and South America. The Tacaribe complex includes the following viruses: Tacaribe, Allpahuayo, Amapari, Bear Canyon Virus, Chapare, Cupixi, Flexal, Guanarito, Junin, Latino, Machupo, Oliveros, Paraná, Pichinde, Pirital, Sabiá, Tamiami, and Whitewater Arroyo.
Mammarenavirus studies in the Caribbean, however, have been limited even though the Tacaribe virus has been isolated since the 1950s in Trinidad and Tobago, and some virological studies conducted with Jamaican fruit bats (Artibeus jamaicensis) and mosquitoes have discovered the presence of this virus in Trinidad and Tobago [57,72]. The role of A. jamaicensis as a possible reservoir host was explored and found both to be susceptible to infection by Tacaribe virus and to exhibit disease pathology, suggesting this species may not be the reservoir for Tacaribe virus [73]. This virus was also isolated from lone star (Amblyomma americanum) ticks in Florida, USA, and as this tick is known to feed on humans, the ability of the lone star tick to transmit this virus to people requires further investigation [74]. At present, the reservoir host of the Tacaribe virus remains unknown. Future studies should explore the source of mammarenavirus infections in wild rodents.

Orthopoxvirus IFA Seropositive Rodents
We also report the first serological evidence of orthopoxvirus infection among wild rodents in the Caribbean, with an orthopoxvirus seroprevalence rate of 7.5% observed among wild rodents. This is on the lower end of orthopoxvirus seroprevalence rates observed compared to studies of wild rodents in Europe and Africa, including Finland (33%), Germany (32%), Uganda (42%), Hungary (18%), Turkey (0.3%) and Serbia (3.2%) [67,[75][76][77]. Evidence of orthopoxvirus infection outside of Europe is limited, and thus this adds to the limited knowledge of orthopoxvirus infections worldwide and is of importance, as orthopoxviruses continue to be emerging zoonotic agents [28,78,79]. In Europe, bank voles (Clethrionomys glareolus) and wood mice (Apodemus sylvaticus) constitute the main reservoirs [27], whereas CPXV was sporadically detected in rats (Rattus norvegicus) [80,81]. Domestic cats can play a role in the transmission of CPXV to humans [82,83]. Direct transmission of CPXV from rodents to humans has also been documented, including rats and mongoose [81,[84][85][86][87]. CPXV infection of immunocompetent persons usually results in localized lesions mainly on fingers, hands, or the face [88]. However, in immunocompromised patients, severe generalized CPXV infections have been documented [89,90].
The lack of smallpox vaccinations in the late 1970s, and an increasing trend of keeping wildlife as pets, have been offered as the reasons for increasing human cases of smallpox in Europe [28,81,86]. Public health education regarding the zoonotic pathogens potentially transmissible by rodents (mice and rats) in Barbados can be updated to include these zoonotic viral pathogens, and biosecurity awareness training given to at-risk persons including cow farmers, dairy workers, ancillary staff, sanitation workers, veterinarians, and the general public.

Strengths
Several key findings were observed in this study, including (1) the identification of a new rodent species in the Caribbean, M. musculus, with serological evidence of orthohantavirus infection and (2) serological evidence of multiple viral human pathogens, namely orthohantavirus, mammarenavirus and orthopoxvirus infections among wild rats and mice, R. norvegicus and M. musculus respectively, in Barbados.

Study Limitations
This study had some limitations and would benefit from some improvements, including (1) larger sample sizes for rodents (particularly rats), (2) a change in the timing of rodent sampling which ideally should coincide with a variation in rainfall and climatic conditions (dry and wet seasons) and preferably over multiple years and (3) the absence of sampling of other possible virus zoonosis reservoirs such as bats [49,53,91].

Recommendations
To limit the potential exposure to viral pathogens from wild rodents, enhanced public awareness should be conducted to inform persons of potential transmission risks of mammarenavirus, orthohantavirus and orthopoxvirus infections and to improve biosecurity at residences, on farms, in hospitals and in other places of business.

Rodent Trapping and Sample Collection
A prospective cross-sectional wild rodent survey was conducted during January 2019 to investigate the serological and molecular evidence of orthohantavirus, mammarenavirus and orthopoxvirus infection status. In total, 160 rodents were trapped using Sherman and Tomahawk traps at 15 different urban, farm and rural locations around Barbados (Figure 1). Traps were baited with peanut butter, banana, bacon and English potato (for a source of moisture for the trapped animal). Care was taken to ensure the traps were shaded to minimize heat exhaustion of trapped rodents. Traps were collected the following morning after being set the preceding afternoon/evening. Trapping was conducted for at least two nights per site with a minimum of 8 traps per location. Trapped rodents were euthanized by chloroform inhalation on a saturated cotton wad followed by cervical dislocation. For each rodent, its species, sex, and biometrics (weight, tail and body length) were recorded.
Dissections were conducted to collect tissue samples (liver, kidney, spleen, heart, lung and ear). Blood samples were obtained from the body cavities by allowing the blood to be absorbed in the absorption zones of pre-cut Nobuto blood filter papers. The filter papers with blood were allowed to dry completely before being wrapped in tissue paper and stored in a cool, dark place. After drying and storing the filter papers, a 1 cm 2 piece of the absorption zones of the filter papers with rodent blood was cut, diluted with 350 µL of Dulbecco medium + 0.2% bovine serum albumin (BSA) (pH 7.2) and shaken overnight with gentle rocking at 4 • C, to elute the blood sample. In the absence of a filter paper blood sample, the heart was used, and no dilution was necessary. The eluate was then tested using the respective virus-specific IFA.

Serological Testing of Rodent Sera
Each rodent blood sample was screened by indirect IFA for orthohantavirus infection using three separate orthohantavirus strain antigens (PUUV, SEOV and HTNV) with infected and non-infected cells mixed in a ratio of 1:3, to ensure specificity of the readouts as described previously (Figure 2) [67]. For each rodent sample, screening was done for orthohantavirus-(SEOV and PUUV), mammarenavirus-(LCMV) and orthopoxvirus-(CPXV)-specific IFA, by placing 20 uL of each sample on the respective IFA slide well. For negative controls, 20 uL of PBS was added (along with anti-mouse IgG). For the positive control well, 20 uL of the respective positive control sera was added (SEOV-or PUUV-positive human sera (1/20 dilution) for orthohantavirus, LCMV-positive mouse monoclonal antibody (1/2 dilution) and CPXV-positive human sera). The slides were then incubated in the moist chambers for 30 min at 37 • C. Each slide was rinsed with cold PBS (4 • C) solution to remove excess antibodies. This was repeated 3 times with cold PBS for 5 min each. The final rinse was performed in double distilled water (ddH 2 O) for 5 min and air dried with a fan. Then 20 µL of polyclonal rabbit anti-mouse fluorescein isothiocyanate (FITC) conjugate (diluted 1:100), or polyclonal sheep anti-rat FITC conjugate (diluted 1:2000) was placed in the rodent (mouse or rat) sample wells, anti-mouse or anti-rat negative control well and LCMV positive control well, or orthohantavirus positive control well. A 20 µL aliquot of anti-human IgG FITC conjugate (diluted 1:100) was added to the human samples, anti-human negative control, and positive control wells. The slides were incubated in the moist chamber for 30 min at 37 • C, washed 3 times with cold PBS washing solution, then rinsed once with ddH 2 O to remove excess conjugate and air dried using a fan. The slides were protected from light and stored at 4 • C until reading with a UV microscope. After observation of slides using the microscope, a note of seropositivity or negativity was recorded for each sample, as well as any unspecific or unclear reactions. Reciprocal titres of >16 by IFA were considered positive results.

RNA Extraction and Molecular Testing of Rodent Sera and Tissue
Lung and kidney samples from seropositive rodents were lysed and homogenized using Roche MagNA lyser (Basel, Switzerland) and Trizol ™ reagent (ThermoFisher Scientific, Waltham, MA, USA), and RNA was extracted following the manufacturer's instructions. RNA yield was determined using NanoDrop (ThermoFisher Scientific, Waltham, MA, USA) based on absorbance measurements at 260 nm and 280 nm. Pan-orthohantavirus and mammarenavirus RT-PCRs were conducted as previously described [92,93]. Acknowledgments: The authors would like to thank the entire staff at the Best-dos Santos Public Health Laboratory for provision of laboratory storage access, the Ministry of Health for approval to trap wild rodents at various sites around Barbados, the several business owners and residential homeowners who permitted rodent sampling on their premises, Veterinary Services, the Ministry of Agriculture (Barbados) for providing incineration services, the Barbados Water Authority for access to Bowmanston Pumping Station environs for trapping and sample processing, and the technical assistance and the staff at the Virology Department of the University of Helsinki, Finland for IFA screening of rodent sera for orthohantavirus, mammarenavirus and orthopoxvirus infections.