Trypanosoma cruzi, the etiologic agent of Chagas disease, has been detected in some 180 species belonging to 25 families of mammals in the Americas, with marsupials, edentates, and rodents being the most frequent sylvatic hosts (World Health Organization, 2002). Domestic transmission cycles mainly include humans, dogs and cats, and several species of triatomine bugs adapted to human dwellings (Pinto Dias, 2000). Transmission cycles of T. cruzi display large spatial and structural heterogeneity (Diotaiuti et al., 1995).

The Gran Chaco, a natural landscape unit of about 1,000,000 km² crossing over northern Argentina, Bolivia, Paraguay and southwestern Brazil, is one of the most endemic regions for Chagas disease. Natural infection by T. cruzi has been reported for armadillos Chaetophractus vellerosus, C. villosus, Cabassous unicinctus, Dasypus novemcinctus, Euphractus sexcinctus, and Tolypeutes matacus; opossums Didelphis albiventris and Lutreolina crassicaudata; short-tailed opossums Monodelphis domestica; foxes Lycalopex culpaeus and L. gymnocercus griseus; coatis Nasua nasua, and mice Calomys musculinus and C. laucha (Carcavallo and Martínez, 1968; Yeo et al., 2005). In a well-defined area in the semiarid Argentine Chaco, the only wild mammals found infected with T. cruzi were D. albiventris opossums, a few Conepatus chinga skunks and one Galictis cuja ferret among 230 mammals from 20 species (Pietrokovsky et al., 1991; Wisnivesky-Colli et al., 1992). In the Bolivian Chaco, sylvatic Triatoma infestans, Triatoma sordida and Triatoma guasayana have been found infected with T. cruzi and may be the putative sylvatic vectors of T. cruzi (Noireau et al., 2000). In the Argentine Chaco, however, the sylvatic vector of T. cruzi has not been firmly established. Since the 1990's the Chaco has been undergoing accelerated deforestation and change of land use patterns which may have affected the relationship between the domestic and sylvatic transmission cycle of T. cruzi to an unknown extent.

T. cruzi has been classified into two major phylogenetic lineages, T. cruzi I (TCI) and T. cruzi II (TCII), and several sublineages within TCII designated as IIa, IIb, IIc, IId and IIe (Anon., 1999). These lineages appear to be distributed differentially between triatomine species and hosts throughout the Americas. TCI was originally described from sylvatic hosts and predominates in domestic transmission cycles to the north of the Amazon basin, whereas TCII predominates in domestic cycles but has sometimes been found in sylvatic mammals as well (Brisse et al., 2000; Yeo et al., 2005). In the Argentine and Paraguayan Chaco, TCI infects D. albiventris opossums and much less frequently humans and T. infestans, whereas TCII typically infects T. infestans, domestic dogs and cats, humans and skunks (Luca d'Oro et al., 1993; Diosque et al., 2003; Yeo et al., 2005; Marcet et al., 2006). Based on isoenzyme and molecular markers, two studies concluded that domestic and sylvatic transmission cycles of T. cruzi overlapped partially throughout Argentina and in two rural areas within the Chaco region (Wisnivesky-Colli et al., 1992; Luca d'Oro et al., 1993; Diosque et al., 2003). Opossums were suggested as a possible bridge between sylvatic and domestic transmission cycles (Schweigmann et al., 1999; Diosque et al., 2004). Temporal variations in the sylvatic cycle of transmission of T. cruzi in the presence of significant environmental changes have not been investigated.

As part of a longitudinal study on the ecoepidemiology of Chagas disease in a well-defined rural area in northwestern Argentina under sustained vector surveillance and selective insecticide sprays, the present study sought to assess the host range and prevalence of T. cruzi in a wide variety of wild mammals, and to identify the parasite sublineages circulating in them. Furthermore, to assess the existence of long-term variations in the dynamics and intensity of sylvatic transmission of T. cruzi, we compared our results with data collected in the same area between 1984 and 1991, before both massive deforestation around one of the villages and sustained vector surveillance in all villages was initiated in 1992.

Fig. 1 Map showing capture sites of wild mammals in the Amamá area, 2002–2004.

Amamá was the only rural village in the Moreno Department that had ever been treated with residual insecticides (in 1985), but the absence of a vector surveillance system resulted in an exponential increase of domestic reinfestation and in renewed transmission (Gürtler et al., 1991, 2005). A second insecticide treatment conducted in 1992 included other highly infested neighboring rural villages, and sustained vector surveil-lance during 1993–2004 resulted in very low numbers of T. infestans and other triatomines infected with T. cruzi (Cecere et al., 1999, 2002; Marcet et al., 2006).

The captured animals were transported to the field laboratory and examined clinically before and after anesthesia with tiletamine chlorhydrate and zolacepam chlorhydrate (Zelazol®, Fort Dodge Sanidad Animal, La Plata, Argentina) or ketamine clorhydrate (Vetaset ®, Fort Dodge) combined with xylacine (Ronpun ®, Bayer) in doses appropriate for each species according to Carpenter et al. (2001). Rodents (mice and cavies) and mouse opossums were not anaesthetized. Opossums were assigned to a given age class based on tooth formula, eruption and tooth wear according to Schweigmann et al. (1999). Each animal was weighed, sexed and measured from snout to base of the tail. Samples of blood (except from cavies and marmosets), hair, feces and ectoparasites were preserved for future studies. Fresh blood smears from a sample of mice were examined microscopically under 400× magnification. Whenever possible, the animals were tatooed in the ears after examination and then released at the capture site without incurring in any harm, with exception of mice which were euthanized.

Feces from the infected xenodiagnosis bugs were cultured in diphasic medium at 28 °C as described by Lauricella et al. (2005). Parasite cultures were cryop-reserved and DNA extracted as described by Marcet et al. (2006). For the first two infected opossums, 11 one-month-old female Balb-c mice were inoculated intraperitoneally with feces collected from the infected xenodiagnosis bugs and then hemocultured.

Mammal infections with T. cruzi were confirmed by kDNA-PCR amplification of the 330 base pair (bp) fragment from the minicircle DNA of the kinetoplastid genome using the primers and cycling conditions previously published (Schijman et al., 2003). PCR tests were carried out under conditions that prevented DNA carryover contamination. Each PCR run included less than 13 samples, 100 fg of T. cruzi DNA as positive control, and sterile water instead of DNA as a negative control. Aliquots of 12 μl of PCR products were visualized under UV light after electrophoresis in 2.5% agarose gels containing ethidium bromide.

Culture samples of each infected mammal were typified to the sublineage level by PCR strategies targeted to spliced-leader DNA, 18s rDNA, 24s alfa rDNA and A10 genomic markers with the incorporation of Taq platinum (Invitrogen, USA) as described by Marcet et al. (2006). Parasites were also genotyped to lineage level directly from feces of infected xenodiagnosis bugs fed on an opossum.

Multiple logistic regression analysis implemented in Stata 9.0 (StatCorp, College Station, TX) was used to assess the relationship between infection prevalence and age, capture season, and year of survey. The binary response variable was infection on a given year. The predictors were survey year (categorical variable, for years 1988–1991), and age class (categorical variables with five levels, after pooling age classes 1 and 2, and age classes 6 and 7 to increase sample size). Reference levels were age classes 1 and 2 pooled, survey year 1988, and summer. Interaction terms were added to the model and dropped from the final equation if found not significant at the 5% level.

A total of 521 mammals from at least 13 identified species was captured with an overall effort of 10,718 trap-days and additional manual captures that yielded approximately half of the total catch (Table 1). Rodents (mice, moles and cavies) comprised most of the catch (n = 282) followed by skunks (n = 92), three species of armadillos (n = 90), and Didelphis opossums (n = 42). Most opossums (73%) and skunks (69%) were captured around the more forested Trinidad, Mercedes and Villa Matilde villages, whereas only 22% of opossums and 18% of skunks were captured around Amamá (Fig. 2). The mean catch of mammals per unit effort using National traps in Amamá was half of that in the rest of the villages, whereas the catch with Sherman traps was similar between village clusters.

Table 1 Prevalence of Trypanosoma cruzi infection as determined by xenodiagnosis in wild mammals captured in the Amamá area between 2002 and 2004

Fig. 2 Capture of sylvatic mammals per unit effort according to type of trap (Sherman, National) in communities with different landscape. March 2003, July 2003, November 2003 and July 2004. Other communities: La Curva, San Luis, Lote S, Pampa Pozo. Numbers on (more ...)

Only four of 501 specimens examined by xenodiagnosis (including four opossums recaptured) were positive for T. cruzi, with an overall prevalence of 0.8% (Table 1). Three (7.9%) D. albiventris opossums and one (1.1%) C. chinga female skunk, all adults or preadults, were found to be infected by T. cruzi. The three infected opossums (two females and one male, aged approximately 7–18 months old) were each caught in the secondary forest around a different village, close to (100–450 m) permanent water bodies and between 230 and 600 m from the closest house (Fig. 1). One of the infected opossums was recaptured eight months later at about 1 km from the site of first capture, and still was xenodiagnosis-positive. The infected skunk was trapped at 300 m from the closest house and very close to a canal. None of the three xenodiagnosis-negative opossums that were recaptured between 3 and 5 months apart converted parasitologically. In opossums, the annual prevalence in 2002–2003 (6.7%, 1/15) and 2004 (8.7%, 2/23) was not significantly different (Fisher's test, P > 0.5). The prevalence of infection increased significantly with age class from 0% (juveniles) to 17% (older adults), and from summer (0%) to spring (14%)(Fig. 3), although these prevalences are based on few infected animals. Based on the age-prevalence curve for the 2002–2004 period, the force of infection λ was 1.2 per 100 opossum-months after weaning (95% confidence interval, CI, 0.53–1.94%). The xenodiagnosis-positive opossums infected 54% of 51 xenodiagnosis bugs, whereas the skunk infected 88% of eight bugs. Among 44 mammals examined by fresh blood films, a T. cruzi-like flagellate was detected in a field mouse but subsequent xenodiagnosis and anatomopathological analysis were negative.

Fig. 3 Comparison of prevalence of Trypanosoma cruzi infection as determined by xenodiagnosis in opossums by age class (A) and seasons (B) in 1988–1991 (taken from Schweigmann et al., 1999) and 2002–2004 in the Amamá area.

We compared our results with data collected similarly in the study area between 1984 and 1991 (Wisnivesky-Colli et al., 1992; Pietrokovsky et al., 1991; Schweigmann et al., 1999). The overall prevalence of T. cruzi in opossums fell significantly from 31.9% (1984–1987, n = 72) and 35.9% (1988–1991, n = 409) to 7.9% in 2002–2004 (χ²-test, d.f. = 1, P = 0.02 and P = 0.006, respectively). The prevalence of T. cruzi infection in opossums increased significantly with age class from 3.7% in age class I to 68.7% in classes VI–VII in 1988–1991, and from 0% in classes II–IV to 16.7% in classes VI–VII in 2002–2004 (Fig. 3A). For the 1988–1991 period, the odds of infection relative to age classes I–II increased from 6.1 (95% confidence interval, CI = 2.5–14.5) in class III, 9.5 (CI = 4.2–21.4) in class IV, 14.0 (CI = 6.0–32.6) in class V, to 32.9 (CI = 10.4–104.7) in classes VI–VII. Significant differences between years were only observed for 1989 (OR = 2.3, CI = 1.2–4.2) (χ² = 83.2, d.f. = 7, n = 409, P < 0.0001). No significant interaction between age class and year of survey was detected. The average force of infection per 100 opossum-months after weaning declined more than sixfold from 8.2 in 1988–1991 (CI, 4.66–11.70%). The annual force of infection was very similar in 1988 (λ = 5.5%, CI, 3.6–7.4), 1989 (λ = 10.2%, CI, 6.8–13.5), 1990 (λ = 9.7%, CI, 5.0–14.5), and 1991 (λ = 6.6%, CI, 0.3–13.5). Infection prevalence increased significantly from summer to spring in both study periods (Fig. 3B). All opossums captured in spring were adults or sub-adults. The infection prevalence in skunks also decreased five-fold from 5.6% (1984–1987, n = 36) and 4.1% (1985–1989, n = 49) to 1.1% in 2002–2004, although not significantly so (Fisher's test, P = 0.29 and P = 0.20, respectively).

Infection with T. cruzi was confirmed by kDNA-PCR in all four xenodiagnosis-positive individuals. The three opossums were infected with TCI whereas the skunk was infected with T. cruzi IIc, yielding the same band patterns reported by Yeo et al. (2005) (Fig. 4). Opossum parasites genotyped from feces of infected xenodiagnosis bugs (directly or from cultured feces) or after passage through mice gave the same results (lineage I) in two opossums.

Fig. 4 Amplification of the intergenic region of the mini-exon genes (A), the D7 DNA ribosomal 24S-alfa domain (B) and A10 genomic marker (C); lanes (1) negative control; 2–5 Didelphis albiventris T. cruzi I (2) and (3) are the same individual on different (more ...)

Our study shows a dramatic decrease in the prevalence and incidence of T. cruzi infection in sylvatic hosts over nearly two decades. To our knowledge, this study may be the first to assess the existence of long-term variations in the dynamics and intensity of sylvatic transmission of T. cruzi in a well-defined area, and to provide approximate estimates of force of infection of opossums based on age-prevalence curves. Opossums were still the main sylvatic reservoir hosts of T. cruzi, followed by skunks. No other mammal was found infected with T. cruzi, in spite of greater spatial coverage, sampling intensity, and inclusion of other common hosts (rodents) that have not been studied as extensively in the Gran Chaco. An extensive xenodiagnosis survey of four species of armadillos also gave negative results for T. cruzi in three provinces in northeastern Argentina (Martínez et al., 1983), unlike hemoculture surveys in the Paraguayan Chaco (Yeo et al., 2005). Interestingly, high infection prevalences in Didelphis opossums (36%) were also recorded in the neighboring Chaco province between 1999 and 2002 (Diosque et al., 2004), and were as high as 52% elsewhere in the Americas (Grisard et al., 2000; Telford and Tonn, 1982; Travi et al., 1994).

The prevalence of T. cruzi in opossums and skunks assessed by xenodiagnosis was stable between 1984 and 1991 and decreased four- or five-fold in 2002–2004. Because parasitologic and serologic methods were found to be equally sensitive for detecting T. cruzi infections in Didelphis (Fernandes et al., 1991; Grisard et al., 2000) and all of the infections we detected were confirmed by PCR, our xenodiagnosis-based estimates may be very close to the actual prevalence. The infection prevalence increased significantly with age and from summer to spring as in the past (Schweigmann et al., 1999), suggesting that the same underlying processes occurred though at different intensities. Vertical transmission (including infections acquired congenitally and during lactation) and horizontal transmission (through urine or anal gland secretions) appear to be very unlikely routes of infection in Didelphis sp. (Jansen et al., 1994; Grisard et al., 2000; Telford and Tonn, 1982; Travi et al., 1994). The observed age-prevalence curves in opossums, increasing with time of exposure (age) after weaning, are consistent with vector-mediated transmission.

The striking decline in the prevalence and incidence of T. cruzi in opossums and skunks occurred in parallel in both Amamá and Trinidad-Mercedes villages. In principle, these changes may be related to: (i) massive or selective changes in landscape over two decades; (ii) sustained vector surveillance of domestic and peridomestic habitats and selective insecticide sprays, either alone or combined with landscape changes; (iii) random fluctuations in the prevalence of infection in wild mammals. The first hypothesis is based on the major environmental changes that took place especially around Amamá since 1990, which included massive deforestation for cattle-ranching and an increase in the number of households from 1993 to 2002 (31% increase in Amamá, 25% in Trinidad and 22% in Mercedes). Around Amamá, wire fencing and selective clearing of shrubs and some tree species was followed by controlled fires and seeding pastures for cattle; some sections underwent indiscriminate deforestation with heavy machinery. Several houses were relocated along the paved and main dirt roads, and the village ended up enclosed by a wired fence and surrounded by homogeneous grassland, unlike other study villages (Fig. 1). This affected the goat-raising practices of local peasants and their subsistence economy, and probably also reduced refuge availability for opossums and other wildlife in the vicinity of houses. In theory, deforestation exerts a large negative impact on wildlife abundance, but the impact on ecologically adaptable synanthropic mammals (such as opossums) was expected to be much lower, if any at all. Deforestation may also augment the contact rate of opossums and skunks with peridomestic and domestic sites in search for food, though this increase was not perceived and reported by local villagers. Degradation of sylvatic habitats, including selective extraction of larger, older trees (more likely to have the type of tree holes used by opossums) may have also affected the abundance of triatomine bugs and contact rates between opossums and bugs through reduction of suitable habitats for such encounters. However, the local sylvatic vectors of T. cruzi remain to be identified conclusively, and the impacts on triatomine populations of deforestation and changes in wildlife abundance (often assumed and theoretically reasonable) are not supported by direct evidence (Walsh et al., 1993). Because domestic dogs were identified as a significant source of mortality for D. albiventris opossums both locally (Schweigmann, N.J., unpublished results, 1995) and in the pampas grasslands (Perez Carusi, L. unpublished results, 2006), reduced availability of suitable habitats or refuges and increased opossum mortality may explain, at least in part, the much smaller catch of opossums and skunks in Amamá.

Landscape changes were qualitatively smaller and more gradual in Trinidad and Mercedes, and therefore may not explain the marked apparent decrease in the local abundance of opossums (a decrease that occurred despite using the same capture methods, a sizable capture effort, and some of the hunters as in previous decades). Local householders reported seeing opossums and opossum tracks much less frequently than in the past, whereas the apparent abundance of skunks and foxes reportedly increased strikingly following a waning fur trade. Whether these apparent local changes in omnivorous mammals are interrelated is unclear. Opossum abundance may also have decreased through increased predation or pathogens. Increase of mesopredators (such as foxes) is predicted to decrease the equilibrium infection prevalence and abundance of infected hosts in a simple susceptible-infectious model, irrespective of whether predators tend to attack infected or healthy prey (Ostfeld and Holt, 2004). For a stable vector population, the lower abundance of opossums would reduce its rate of contact with triatomine bugs. A mathematical model of T. cruzi transmission (Cohen and Gürtler, 2001) suggests that a dramatic reduction in the abundance of a primary reservoir host with high and long-lasting infectiousness to the vector (such as dogs or its putative sylvatic equivalent, opossums) would be sufficient to drive down the transmission risk and reduce the equilibrium prevalence of infection in bugs and hosts when other parameters are held constant.

The second hypothesis (effects of sustained vector surveillance and selective insecticide sprays) implies a strong link between domestic and sylvatic transmission cycles, with the domestic cycle “spilling over” T. cruzi and/or bugs to the sylvatic cycle. Before 1992, only Amamá experienced a transient low-risk period between 1986 and 1988 following the first residual spraying with insecticides made in the Moreno Department, while Trinidad and all other rural villages were highly infested and with ongoing active transmission until late 1992 (Gürtler et al., 1991). Interestingly, transmission of T. cruzi to opossums was remarkably stable during an extended period, as shown by (i) the nearly constant mean prevalence of infection in opossums in 1984–1987 and 1988–1991 (31.9–35.9%), and (ii) the very similar age-prevalence curves recorded annually between 1988 and 1991. The 1992 community-wide insecticide campaign included other neighboring villages, and sustained vector control actions during 1993–2004 reduced to marginal levels the abundance of T. infestans infected with T. cruzi in domestic and peridomestic habitats, and the transmission of T. cruzi to domestic dogs dropped from 65% at baseline to 8.9% and 4.7% at 7.5 and 10 years after sustained vector surveillance, respectively (Cecere et al., 1999, 2002; Cardinal et al., in press). Moreover, during 2002–2004 the mean prevalence of T. cruzi in opossums was again stable though at very low levels; the infected wild mammals were captured around the villages that had the lowest infestation, and only marginal prevalence of T. cruzi (<1%) was detected in T. infestans and other triatomine bugs (Cardinal et al., in press). Therefore, the likelihood of potentially infective contacts between T. cruzi-infected domestic or peridomestic bugs from the study villages and sylvatic mammals was probably too remote to explain the occurrence of infected mammals. The “spill-over” of domestic or peridomestic bugs into the forest and subsequent creation of sylvatic foci of transmission is an unexplored possibility in our system, but is supported by at least one example where such “spill-over” was detected in Didelphis opossums and T. infestans (Diotaiuti et al., 1995). The stable patterns observed in opossum infections between 1984 and 1991 and between 2002 and 2004 does not support that the major long-term decline in the prevalence of T. cruzi may be explained by random fluctuations (third hypothesis).

Opossums and skunks are nocturnal, omnivorous, insect-eating mammals with a limited average home range (<50 ha), short mean life expectancy (12–18 months and perhaps 2–4 years, respectively), and high fertility leading to fast population turnover. D. albiventris opossums are sometimes considered semi-nomads, with each individual using multiple refuges in tree holes or burrows. Both opossums and skunks may acquire T. cruzi infection by eating infected T. infestans, either when they approach houses in search for food, or when adult bugs eventually disperse by flight into sylvatic habitats, or by contamination with bug feces. The probability of experimental opossum infection with T. cruzi by ingestion or contamination was about 6–8% (Rabinovich et al., 2001). Regardless of the exact mechanism and habitat where vector-mediated transmission occurs, if domestic or peridomestic T. infestans infected with T. cruzi were implicated in opossum infections, the observed major decrease in the extent and abundance of infected bugs would imply an insignificant risk of infection for local opossums. The crucial piece of evidence that may cast light on the origin of wild mammal infections is the diversity of T. cruzi lineages in domestic and sylvatic hosts and in bugs.

Didelphis opossums have predominantly been found infected with TCI throughout the Americas (reviewed by Brisse et al., 2000; Yeo et al., 2005). In our study area, however, three of 18 opossums, two ferrets and a skunk were infected with zymodemes equivalent to T. cruzi IId up to the early 1990s' (Wisnivesky-Colli et al., 1992; Luca d'Oro et al., 1993), based on current parasite divisions (Barnabé et al., 2000). Current results show opossums only infected by TCI, whereas only 3 domestic or peridomestic T. infestans were found infected by T. cruzi (out of 389 bugs from the study villages examined for infection) and had lineage I (Marcet et al., 2006). For our area therefore, it was highly unlikely that the opossums acquired TCI from domestic or peridomestic habitats. In the absence of data on the putative vector(s) implicated in transmission to sylvatic hosts, the long-term reduction of T. cruzi infection in opossums may be tentatively explained by environmental degradation causing a large reduction of old trees and an apparent reduction of opossum abundance, possibly through marked changes in wildlife composition and abundance. Identifying the precise mechanisms linking these processes to the decline in opossum infections requires further investigation.

Our study provides the first record of a mammal (a skunk) infected with T. cruzi IIc in Argentina. This lineage has also been found infecting very few T. infestans and dogs in the study area (Cardinal et al., unpublished data). In Paraguay, T. cruzi IIc has also been isolated from two species of armadillos and a short-tailed opossum, but not from T. infestans (Yeo et al., 2005). The co-occurrence of two species of sympatric wild mammals infected with two different lineages of T. cruzi may imply either (i) the simultaneous existence of two independent sylvatic transmission cycles mediated by different vector species, or (ii) the host selecting for certain parasite lineages. The former hypothesis is supported by a compendium of evidence suggesting that terrestrial mammals that excavate burrows in open terrain (such as armadillos and skunks) are closely associated with TCII, whereas mammals of arboreal habits (such as opossums) are closely associated with lineage I (Yeo et al., 2005). Persistence of TCI in opossums and the occurrence of a different sublineage in skunks (IId up to the 1990s' and IIc more recently) also suggest that these two hosts are involved in separate transmission cycles over decades. The very low frequency of infected skunks suggests they may be incidental hosts of T. cruzi. Although parasite isolation methods may select for certain strains and lineages of T. cruzi, no lineage selection was recorded among 145 isolates from opossums in spite of using different methods (Fernandes et al., 1991). Experimental infections of Didelphis opossums with TCII were considered less persistent than infections with TCI (Fernandes et al., 1994). In addition, mixed infections with different lineages and sublineages contradict the hypothesis of strict sublineage selection by the host.

In conclusion, our study shows a clear-cut decline in T. cruzi infection in sylvatic hosts in the context of various degrees of deforestation combined with other changes in wildlife composition and apparent reduction of opossum abundance. Regardless of the underlying processes, the decline in sylvatic T. cruzi infections indicates a diminishing risk of introduction of sylvatic strains and lineages into the local domestic transmission cycle. Whether this is a generalized pattern in the Gran Chaco remains to be investigated.

We are grateful to Francisco Petrocco, Juan M. Gurevitz, Leonardo Lanati, Evelin Grijalva, M. Cruz Pino, Paula Marcet, Margarita Bisio and Rebecca Dieter for assistance in field or laboratory work. Reference strains of T. cruzi I and T. cruzi IIa, IIb, IIc, IId and IIe were kindly provided by Patricio Diosque and Miguel Angel Basombrío (Instituto de Patología Experimental, Universidad Nacional de Salta, Argentina) and by Michel Tibayrenc (UR62 “Genetics of Infectious Diseases”, IRD Centre, Montpellier, France). This project received major support by NIH Research Grant # R01 TW05836 funded by the Fogarty International Center and the National Institute of Environmental Health Sciences (NIEHS) to UK and REG. Additional funding from Agencia Nacional de Promoción Científica y Técnica de Argentina and University of Buenos Aires (to REG), WHO-TDR ID 20285, CONICET and Bunge & Born Foundation (to AGS). AGS, MJL and REG are members of CONICET Researcher's Career.