Discussion
Our survey of T. gondii in NPs across Costa Rica documented widespread infection and species-specific risks of infection. The seroprevalence obtained for captive NPs is similar to other studies in South America, such as 30.8% and 76% using MAT in captive Cebus apella, (Leite et al., 2008; Pires et al., 2012), and 79% inCebus sp. and 57% in Ateles sp. from zoos in São Paulo, Brazil (Bouer et al., 2010). The high prevalence of NPs in captivity may be due to management practices, including improperly washed fruits or vegetables, and raw or undercooked meat (Pires et al., 2012; Valentini et al., 2004), proximity of wild or domestic cats, and invasion of enclosures by infected birds and rodents that might be ingested by NPs (Pires et al., 2012; Valentini et al., 2004). These sources of infection would be similar for male and female NPs in captivity, consistent with our findings.
The 11.6% seroprevalence in wild Costa Rican NPs was lower than the 26.6% (n = 60) reported by Garcia et al (2005) using MAT in Brazil, and the prevalence we found for A. palliata (6.6%) is lower than documented for Alouatta caraya by Garcia et al (2005) in Paraná (17%, 3/17) and Molina et al (2014) in São Paulo, Brazil (75%, 15/20). However, it is very similar to findings by de Thoisy et al (2001), who found only 4% (2/50) seropositive Alouatta senilicus in French Guiana, and similar to this study had a higher proportion of positive females. Our data did not indicate a significant difference between sexes for any of the species, although this should be studied further given the sex bias observed in French Guiana and among domestic cats (Afonso et al., 2007).
In the case of C. imitator , the prevalence is equal to that reported by Garcia et al. (2005) in Cebus spp. with 30.2% (13/43). In fact, several studies have identified that carnivorous diet is a risk factor for T. gondii infection (Cabezón et al., 2011; Hejlicek et al., 1997). On the other hand, the prevalence found inA. geoffroyi (40%, 2/5) appears to be high, which is surprising considering its arboreal and herbivorous (frugivorous) behavior, characteristics that minimize its exposure to the parasite. No prevalence reports were found for wild Ateles sp . to compare this finding to. However, sample sizes for both A. geoffroyi andS. oerstedii were very small, so these findings should be interpreted with caution. The difference in seropositivity observed between howler (A. palliata ) and capuchin monkeys (C. imitator ) coincides with Garcia et al. (2005) and corresponds to behavioral characteristics between species. Howlers are predominantly folivorous, supplementing their diet with fruits, flowers and seeds, and obtaining most of the water they need from their food (reviewed by Wainwright, 2007). However, they can drink from water accumulated in branches, trunks or bromeliads (Glander, 1978; Gilbert & Stouffer, 1989), or search for it on the ground (Almeida-Silva et al., 2005). The source of infection for howlers would be water bodies infected with oocysts. Meanwhile, capuchins are the most omnivorous NPs, feeding on various sources like fruits, insects and small vertebrates such as birds, rodents, squirrels, coatis, bats, frogs and lizards (reviewed by Wainwright 2007 and Catão-Dias et al., 2013). They frequently go down to the undergrowth and ground while foraging and traveling (Feagle, 1999). In addition, Cebus spp. drink water directly from puddles (Fragaszy et al., 2004). These characteristics give capuchins greater exposure to oocysts in the soil / water or in invertebrates (transport hosts) and to tissue cysts present in infected vertebrates.
The low antibody titers obtained for the wild NPs (16 and 32) coincide with those reported by Garcia et al. (2005), who found mostly 16 and 32 for a single individual. While A. palliata had low antibody titers, two A. geoffroyi and two C. imitator had very high antibody titers (262,144 - 1,048,576). Molina et al (2014) reported values ​​of 25 for Callithrix penicillata (a very susceptible species) and higher titers (up to 1600) for A. caraya , arguing that differences in prevalence and titers could respond to differences in host susceptibility, contact rates or post-exposure time. Indeed,Alouatta ’s susceptibility in comparison to Cebus could be reflected in serological differences, since the probability of post-infection survival and, therefore, developing an immune response is naturally lower for howlers. Furthermore, if these individuals die, they would be excluded from the population and the sample, reducing the number of individuals captured with high titers. In contrast, titers of captive NPs were high in our study. Leite et al. (2008) reported antibody titers of 8000 by MAT for Cebus sp . in captivity, while Ekanayake et al. (2004) reported antibody titers >256 (up to 4096) in 3 of 21 positive free-ranging but urban macaques in Sri Lanka. Management factors along with the increased life expectancy forCebus sp. and Ateles sp. in captivity could explain the chance of infection and high titers observed in this group.
It is worth highlighting the large number of positive samples from the Gulf of Nicoya (Figure 1.A.). Tempisque River is one of the most important basins in the country, draining 10.6% of the territory and flowing into the Gulf of Nicoya (Gutiérrez et al 1985). Areas close to bodies of water could represent sources of infection for NP, because water can be contaminated at any point, transport oocysts long distances, and favor their survival (Lindsay & Dubey, 2009). Samples from protected areas with high levels of human contact also showed high prevalence and high titers in the two wild C. imitator mentioned before. In certain national parks and private reserves, feeding wild monkeys is a common practice, and capuchin and squirrel monkeys sometimes exhibit agonistic behaviors that include taking food directly from humans and coming down to the ground. Such opportunistic behavior (Campbell, 2013) may increase exposure to infectious agents (Wolfe et al., 1998; Ekanayake et al., 2004).
Besides identifying specific areas where there could be elevated risk of infection, our data indicates that environmental variables such as forest cover and precipitation could be associated with exposure risk. Seropositivity in A. palliata was higher when there was a higher percentage of forest cover and less annual precipitation. Forest cover can protect oocysts from sunlight, allowing them to remain viable where protected for 1-1.5 years (Ruiz et al., 1973; Frenkel et al., 1975). Smith & Frenkel (1995) and Almería et al. (2004) found greater seroprevalence in hares and other mammals sampled in forested areas versus more arid grasslands, arguing that shadow and relative humidity provided by forest cover act on oocyst conservation by decreasing the evaporation rate and desiccation of oocysts.
Wet seasons tend to increase oocyst survival (Frenkel et al., 1975). In fact, T. gondii seroprevalence in cats (Afonso et al., 2013) and humans (Hubálek 2005) has been associated with rainy and warm episodes (North Atlantic Oscillation), and in wild ruminants (Gamarra et al., 2008) with humid areas. Contrary to expectations, in this study precipitation was inversely related with seropositivity in both NP species. Costa Rica is a tropical country with high relative humidity and stable temperature overall. It is possible that increased precipitation in Costa Rica results in greater runoff, transporting oocysts towards the coasts and away from the animals. In recent past years, low rainfall due to the El Niño phenomenon has generated severe droughts and forest fires in some areas of Costa Rica. Among many other animals, A. palliata were severely affected by water and food shortages, with high mortality due to dehydration and starvation, as well as injuries due to troops fighting for food. Because behavior change driven by droughts can increase exposure to parasites present in the scarce sources of water, the risk of disease increases especially for animals weakened by starvation and dehydration.
In the present study, the source of exposure of wild NP could be wild and not domestic cats, which might explain the low effect of human population density on seropositivity. Contact with humans has been associated with high seropositivity in macaques (Ekanayake et al., 2004), which can become infected by ingesting human food from the ground in areas frequented by domestic cats (Tenter et al., 2000). Domestic cat population estimates as well as sampling in areas where humans feed and interact with wildlife should be included in future studies.
Additionally, Toxoplasma genotypes produce different degrees of virulence in humans and mice (Xiao & Yolken, 2015), and given the high diversity recently found in Central and South America (Ajzenberg et al., 2004; Lehmann et al., 2004; Rajendran et al., 2012; Vitaliano et al., 2014; Rego et al., 2018; Vethencourt et al., 2019) the observed differences in seroprevalence between wild and captive NP might be due to different genotypes. Little is known about the genotypes circulating in wildlife, and associations between strain type, lesion patterns and clinical outcome have not been reported in wildlife frequently. Thus, future studies that focus on genotyping and virulence of T. gondii isolates in wildlife and domestic animals from wild and anthropized environments would be of great value.
In the present study, MAT was used as it is simple and fast, doesn´t require specific modifications, can be used for large sample sizes and diverse host species, and can be run on serum and plasma (Desmonts & Remington, 1980; Shaapan et al., 2008; Dubey, 2010). An antibody titer of 1/25 is often considered as evidence of exposure to T. gondiiin many mammals (Dubey et al., 1995) and 1/5 in birds (Dubey et al., 2016). However, there is no antibody titer that is considered specific for primates and our analysis required that we be able to maximize sensitivity of detection given the high susceptibility and low seroprevalence in some NP species, thus we reported all antibody titers of 1/16 or higher (Garcia et al., 2005).
Among the reagents used for MAT is mercaptoethanol which destroys immunoglobulin M (IgM) and therefore MAT only detects IgG (Seefeldt et al., 1989). Because IgG are maintained for life (Dubey, 2010), their detection indicates exposure to T. gondii at some point during the animal´s life; inability to discriminate recent infections precludes assessing longitudinal shifts in seropositivity. However, IgG presence can be used to detect the spatial influences of environmental factors. Further analyses through longitudinal sampling and detection of IgM are recommended.
In conclusion, our study documented widespread T. gondiiinfection in NP and species-specific risks of infection for the first time in Costa Rica. The high seroprevalence and titers found in captive capuchin and spider monkeys may be due to management practices, the proximity of cats or intermediate hosts, and the increased life expectancy in captivity for these species. The low seroprevalence and titers in wild NP varied between species. The difference in seropositivity observed between wild howler (A. palliata ) and capuchin monkeys (C. imitator ) in this study agrees with behavioral and dietary characteristics, in which capuchins are more exposed to oocysts while foraging on the ground, and by ingesting invertebrates (transport hosts) and vertebrates (tissue cysts). However,Alouatta ’s susceptibility compared to Cebus might also explain the observed serological differences, due to decreased survival. Our data indicated that specific areas could represent an elevated risk of infection (i.e. water runoff and human interaction), and environmental variables such as abundant forest cover and low precipitation could be associated with higher exposure risk in wild NP. Surveillance of T. gondii in NP is required to better understand the infection status, genotypes and drivers involved in wild and captive NP, including individuals in the process of reintroduction, so that biosecurity measures are improved, avoiding the release of infected individuals, and develop novel strategies to protect wild populations.