A Retrospective Study on the Seroprevalence of West Nile Virus Among Donkeys and Mules in Bulgaria

Introduction

West Nile virus (WNV) infection is a zoonosis that became particularly important during the past two decades at a global scale. The causative agent belongs to the mosquito-transmitted arboviruses in the Flaviviridae family. The virus is also a member of Japanese encephalitis virus antigenic complex along with other closely related flaviviruses (Rathore and St John, 2020). WNV exhibits a remarkable life cycle that involves virus reservoirs (vertebrates, mainly birds), mosquito vectors (rare ticks), and final or incidental hosts.

The latter are infected during mosquito feeding, and might develop clinical disease ranging from flu-like symptoms to severe encephalitis (Habarugira et al., 2020). Humans and equids are defined as dead-end hosts, because they do not contribute to further virus spread in the population (Khare and Kuhn, 2022).

In 2010, eight seropositive cases (five donkeys and three horses) were reported for the first time in the northeastern region of Bulgaria (Di Sabatino et al., 2014). Five years later, the first case of West Nile neuroinvasive disease was laboratory confirmed in a human patient from Sofia, Bulgaria (ECDC, 2015). Since then, several local studies have provided evidence for circulation of the virus among reservoirs, vectors, and humans as incidental hosts.

A study conducted by Christova et al. (2020) confirmed Culex pipiens as the predominant mosquito species involved in the life cycle of WNV, and the affiliation of circulating strains to lineage 2. Another study by Trifonova et al. (2022) demonstrated high levels of antibodies against WNV in both wild resident and migratory birds, underscoring their significance as reservoir hosts.

WNV encephalitis of horses was first reported from Egypt in 1963 (Schmidt and Elmansoury, 1963). The same study also assessed different serological tests for WNV, and contributed to understanding of the role of equines as virus hosts. Information about the seroprevalence of the WNV among donkeys and mules in Europe is still scarce (de Heus et al., 2021). Investigations on these animal species may be a helpful epidemiological tool, especially in regions where vaccination in horses is administered (García-Bocanegra et al., 2012).

The lack of published data about the seroprevalence of the WNV among donkeys and mules on the basis of a large number of samples in Bulgaria has motivated this retrospective study.

Materials and Methods

The survey included a total of 200 serum samples collected in July–August 2015. The samples were collected for another global survey on the circulation of hepatitis viruses among equine population. The samples were obtained from 14 settlements in 6 districts of Bulgaria—5 in south Bulgaria and 1 in north Bulgaria. The districts were located in the following three regions: southeast (n = 3), southcentral (n = 2), and the northwest (n = 1). The tested animals were clinically healthy. They were 147 donkeys (60 males and 87 females) and 53 mules, aged from 2 to 30 years (mean 15.7 ± 6.7 years).

Description of the tested samples by districts, regions, species, and sex is presented in Supplementary Table S1. The blood samples were collected aseptically by jugular vein puncture in Vacutest collection tubes with clot activator (Vacutest Kima srl, Italy). After separation of the serum and centrifugation at 2000 g for 10 min, the samples were aliquoted into cryo tubes and stored at −80°C until the analysis in June 2023.

Sera were tested for antibodies against WNV glycoprotein E with a competitive enzyme-linked immunosorbent assay (ELISA) kit—Ingezim West Nile Compac (Ingenasa, Spain) according to the manufacturer's protocol. Before use, sera were thermally processed at 56°C for 45 min. The plates were spectrophotometrically read at 450 nm on ELISA reader (Biosan, Latvia) immediately after addition of the stop solution. For validation of results, the average optical density of positive and negative controls was used. Samples with inhibition percentage (IP) ≥40% were interpreted as positive, those with IP ≤30% as negative, whereas samples with IP between 30% and 40% as doubtful.

Positive sera were additionally tested by the virus neutralization test (VNT) in the OIE Reference Laboratory in Teramo (Italy) for detection of immunoglobulin G (IgG) neutralizing antibodies against the WNV and the closely related Usutu virus. In brief, serial twofold dilutions (starting from 1:5) were prepared in microtiter plates, and 10² tissue culture infective doses of antigen (strains Eg-101 and 939/01, respectively) were added to each dilution. Mixtures were incubated at 37°C for 1 h, and 10⁵ Vero cells were added to each well. The plates were incubated at 37°C for 5 days and checked for cytopathic effect. The antibody titer was defined as reciprocal of the highest dilution of the serum that showed 100% neutralization.

The positive sera were also tested in a competitive IgG- and immunoglobulin M (IgM)-ELISA for WNV (ID Screen^® West Nile Competition Multi-species and ID Screen West Nile IgM Capture, IDvet, Grabels, France). These tests were made in the reference laboratory following the company's instructions.

Results

Fourteen out of the competitive ELISA-tested 200 serum samples showed a positive result (7% [95% confidence interval {CI} = 3.83–11.75]). The IPs of sera varied between 51.5% and 95.3%. Except for one sample, all others were confirmed by competitive IgG-ELISA in the reference laboratory. None of tested sera showed IgM antibodies. VNT confirmed two of ELISA-positive samples with titers 1:5. Neutralizing antibodies against the Usutu virus were not detected in tested samples (titer <1:10). The seroprevalence of WNV based on positive results of both used methods was assessed as 1% (2/200 [CI = 0.12–3.61]), with further details presented in Table 1.

The highest seroprevalence of 5% (1/20) was observed in the Sliven district, located in the southeastern region of Bulgaria, followed by Pleven district (northwestern region) with 4.8% (1/21). Positive samples by VNT were not detected in the other investigated districts. Supplementary Fig. S1 illustrates their geographical location.

Discussion

This retrospective study fills the information gap about the seroprevalence of WNV in incidental hosts (donkeys and mules) in our country. Out of 200 samples, 14 (7%) tested positive using a competitive ELISA; however, only two (1%) samples were confirmed by VNT conducted in the reference laboratory. Discrepancies in the results between the tests may be due to the lower specificity of the competitive ELISA, leading to cross-reactions with other closely related flaviviruses (Maeki et al., 2019). Nevertheless, we can exclude circulation of Usutu virus among donkeys and mules in our study as neutralizing titers below 1:10 were detected.

Owing to the lack of actual data on the WNV seroprevalence among equids in Bulgaria, the obtained results may be compared with those from human surveys during the same period. A nation-wide serological survey was carried out in 2015 on 1451 samples from all 28 districts of Bulgaria (Christova et al., 2017). The authors identified 22 positive samples (1.5%) with IgG-ELISA, whereas 6 out of 1451 were confirmed to possess neutralizing antibodies (0.4%). The seroprevalence rates found in Vidin, Ruse, and Silistra districts (7.5%, 6%, and 6%, respectively) are comparable with the 4.8% seroprevalence observed in the Pleven district in our study.

These districts border the Danube River in the north, and Romania, where a large number of human outbreaks and deaths have been recorded in the past two decades (Bakonyi and Haussig, 2020). A study conducted in Romania affirms that the lower area of the Danube Delta serves as a conducive environment for the circulation of the WNV. This is substantiated by a seroprevalence of 33.53% observed in tested horses (Savuta et al., 2007). The seroprevalence rate among donkeys in Turkey, Bulgaria's southern neighbor, varies significantly within the Anatolia region ranging from 1.28% (3/234) in eastern Anatolia (Mehmet et al., 2017) to as high as 20% (14/70) in northeastern Anatolia (Yildirim et al., 2018).

This variability can likely be attributed to the diverse climatic conditions prevailing in the respective study areas. The established seroprevalence in donkeys living in the “Zasavica” Nature Reserve in the Vojvodina province, Serbia (located at the northwestern boundary), was 86.79% (46/73) as of July 2022 (Lupulovic et al., 2023). Notably, all ELISA-positive samples were subsequently validated by VNT, with titers exceeding 1:256.

On the basis of this finding, the authors concluded that WNV was circulating within the donkey population in the study territory. In 2022, nine outbreaks among equids were reported from Greece (southern boundary) with no specification of their species (ECDC, 2023). To the best of our knowledge, no published data are currently available concerning the seroprevalence in donkeys from north Macedonia, another neighboring country of Bulgaria.

None of mule samples was positive in this study. The samples originated mainly from the Smolyan district (n = 52) and one sample was from Haskovo district. On the contrary, García-Bocanegra et al. (2012) found higher individual seroprevalence in mules (9.6%) than in donkeys (4.9%) in southern Spain, but the difference was not statistically significant. The observed difference in both studies may be attributed to the distinct geographical locations of the tested mules and associated contributing factors. Other serological surveys in equids in Europe are mainly focused on WNV circulation in horses or made no species differentiation (Ganzenberg et al., 2022), making direct comparisons and discussions challenging.

The low titers of 1:5 in the serum neutralization test and the lack of IgM antibodies suggest a prior infection. It should also be noted that higher antibody titers are detectable after vaccination that is not applied among equids in our country. Therefore, these species can serve as sentinel animals for WNV monitoring.

The discrepancy with competitive ELISA was found out only for one sample interpreted as positive in our laboratory and negative in the reference laboratory. The IP of this sample was 86.13%, almost twice higher than the threshold of 40%, so according to manufacturer's instructions, it did not require confirmation as doubtful sample. Until arriving at the reference laboratory, sera underwent two thaw–freeze cycles. Even with a highly sensitive assay as ELISA and low antibody titers, a partial degradation that may influence the ultimate result is possible (Rusenova et al., 2022).

This study has some limitations, for example, lack of samples from all regions of the country, an uneven distribution of number of samples by districts, and a small number of samples confirmed by both ELISA and VNT. These limitations did not permit the performance of statistical analyses related to risk factors. Despite the limitations, it demonstrates that donkeys are exposed to WNV infection and seroconvert, contributing to our understanding of virus circulation among donkeys in settlements in north and south Bulgaria. Furthermore, this research might be a strong foundation for future investigations on WNV seroprevalence and on potential risk factors among equids in the country.