Many arthropod-borne viruses (arboviruses) can cause viral encephalitis in vertebrates. Within the genus Flavivirus (family Flaviviridae), Saint Louis encephalitis virus (SLEV) re-emerged as a human pathogen in Argentina in 2002 and caused the first epidemic in 2005, and West Nile virus (WNV) was first isolated in Argentina from horses with encephalitis in 2006, although human deaths were not reported [1, 2]. The New World alphaviruses (family Togaviridae) western equine encephalitis virus (WEEV), eastern equine encephalitis virus (EEEV), and the Venezuelan equine encephalitis virus (VEEV) complex can cause febrile syndromes and sometimes fatal encephalitis in equines and humans. In Argentina, the circulation of VEEV has been confirmed in the enzootic part of the transmission network, but human and equine exposure has also been reported [3]. No activity has been reported for WEEV or EEEV since the mid-1980s. However, enzootic circulation of Madariaga virus (MADV, South American EEEV) occurs [4]. Viruses belonging to the genus Orthobunyavirus (family Peribunyaviridae) have expanded their geographical range from North America to South America. In Argentina, new strains have been isolated recently from cases of encephalitis and abortion in horses [5]. Originally classified as Cache Valley virus, they are now considered different variants of Fort Sherman virus (FSV) [6].

Equines are a valuable target for detecting the circulation of arboviruses because mosquitoes exploit them for blood meals, they share rural and urban transmission scenarios with humans, they develop robust and durable immune responses to infections, and they are susceptible to disease [7]. Draft horses in particular, are widely used by informal waste pickers in urban settlements in South America, they are not included in vaccination programs, and they are often malnourished and used for hard labor. Therefore, to identify zoonotic arboviruses circulating in the city of Santa Fe, we conducted a serosurvey on draft horses that move across the city during the day and spend the night in a well-vegetated suburban area.

We examined the arbovirus exposure pattern based on 222 blood samples collected in a three-year cross-sectional opportunistic survey (2013 and 2014, 73 samples; 2015, 76 samples). In addition, current transmission was surveyed with 47 paired samples (2 to 6 repeated samples) and seroconversion events were recorded. We analyzed the presence of neutralizing antibodies by plaque reduction neutralization tests (PRNT80) in Vero cell monolayers, using SLEV 78V-6507 [8], WNV E/7229/06 [2], FSV SFCrEq231 [5], Río Negro virus AG80-663 (RNV, VEEV subtype VI), WEEV 646, and EEEV Cba 55 strains [9]. Samples that neutralized at least the 80% of the inoculated plaque forming units (PFU) were considered positive, and they were titrated using twofold serial dilutions. A generalized linear model with a binomial distribution family was fitted to explore the contribution of sampling year (levels: I, 2013-2014; II, 2014-2015; III, 2015-2016) and “age class” (A, 0-5 years old [y.o.]; B, 6-10 y.o.; C, 11-15 y.o.; D, >15 y.o.) in the observed seroprevalence. A posteriori contrasts were performed to check the significance (α = 0.05) of the different levels of age and year. In addition, the association between titers of neutralizing antibodies and aging was explored for each virus by ANOVA by determining the covariation between the inverse of the antibody titer (log10 transformed) and the age class. Statistical analysis was performed using the computing environment R [10].

The prevalence and the titer range of neutralizing antibodies for viruses assayed were as follows: 76.8% FSV (152/198, 10– >1280), 60% SLEV (112/188, 10–320), 27% WNV (56/203, 10–640), 7.4% RNV (16/216, 40– >1280), 0% WEEV (0/195), and 0% EEEV (0/196). The distribution of seroprevalence rates for individual viruses, as well as the annual rates and data from paired specimens are summarized in Table 1.

Table 1 Annual infection rates and overall seroprevalence of arboviruses in an equine population from Santa Fe, Argentina
Full size table

Twelve seroconversions were detected during the study period: five for FSV, five for WNV, and two for SLEV (Table 2). Moreover, six seroconversions occurred in animals younger than 5 years old (4/5 for FSV, 1/5 for WNV, and 1/2 for SLEV). The recent exposure of horses indicates endemic circulation of SLEV, WNV and FSV.

Table 2 Arbovirus seroconversion events detected in a draft horse population from Santa Fe, Argentina
Full size table

Our study revealed a remarkably high prevalence of FSV, and it provides the first data about the exposure level of horses in Argentina. Importantly, FSV has been isolated from encephalitis and abortion cases in horses. Consequently, the biological and epidemiological characteristics of FSV should be investigated further to ascertain the potential health impact in horse and human populations. The seroprevalence of flaviviruses was notably higher than previously observed in farm horses from Santa Fe province (SLEV, 12.2%; WNV, 16.2%) [11]. This could be due to environmental factors that might modulate the composition of mosquito vector communities and host exposure. For instance, here, the horses tend to spend the night out in the open or in small crude stables without screens near a wooded nature reserve, which could increase the exposure to mosquitoes. Almost 23% of the samples exhibited heterotypic serological responses to both flavivirus. Following the criterion presented by Tauro et al. [11], we considered that sequential infections by different viruses (SLEV and WNV) can result in heterotypic responses. The samples that were positive for RNV were also analyzed for other South American members of the VEEV serocomplex, as cross-reactions are common between subtypes IF, IAB, and IV. Positive samples were not detected. Since all the samples screened for WEEV and EEEV were negative, we still lack data about the activity of these alphaviruses. However, the monitoring efforts should be sustained in order to evaluate whether “silent” sylvatic cycles are occurring and virus emergence is a potential epidemiological scenario. In this sense, the report of a fatal case of encephalitis caused by WEEV in Uruguay in 2009 and the recent detection of MADV in Culex mosquitoes in northeastern Argentina highlight the need to survey equine populations [4, 12]. In our study no significant differences in prevalence where observed when different time periods were compared (I, 2013-2014; II, 2014-2015; III, 2015-2016). Although there was also no significant effect of aging on antibodies titers (data not shown), when individuals were grouped into age classes, significant associations between prevalence and age were observed for WNV, SLEV, and FSV (Fig. 1), at least partially supporting the endemic circulation proposed above.

Fig. 1
figure 1

Seroprevalence by age in a population of draft horses from Santa Fe, Argentina. A, 0-5 years old (y.o.); B, 6-10 y.o.; C, 11-15 y.o.; D, >15 y.o.; RNV, Rio Negro virus; WNV, West Nile virus; SLEV, Saint Louis encephalitis virus; FSV, Fort Sherman virus. A significant pairwise odds ratio (p < 0.005) is indicated *. Error bars represent 95% confidence intervals.

Full size image

The use of draft animals in the urban environment is a frequent practice, not only in Santa Fe, but also in the rest of Argentina and many other developing countries. The recent activity and the endemicity of the analyzed viruses should encourage health authorities to promote the standardization and maintenance of larger-scale arboviral surveillance programs in Argentina and South America. In particular, draft horses are proposed as sentinels in locations where they are in common use that involves daily contact between equines and humans.