Cryptosporidium is one of the most important parasitic diarrheal agents in humans in the world, is among the top four causes of moderate-to-severe diarrheal disease in young children in developing nations, and is problematic as an opportunistic co-infection with HIV due to increased morbidity and mortality [1,2]. Cryptosporidium is well adapted to zoonotic, waterborne, and foodborne transmission, with a life cycle occurring in suitable hosts and transmission by the fecal-oral route [3]. Zoonoses represent the majority of diseases emerging globally with potential to expand to new host systems, [4] yet despite these health threats, few studies have examined the ecology and epidemiology of this pathogen in rural tropical forest systems characterized by high rates of overlap among humans, domesticated animals, and wildlife [5,6].

In Tanzania, agriculture represents over a quarter of the national income and 80 percent of its labor force [7], but natural resources are declining, affected by desertification and soil degradation from recent droughts. This process has resulted in a high rate of loss of forest and woodland habitat [8]. The resulting fragmented landscape increases human-wildlife contact in these areas, elevating the risk for disease transmission. The Greater Gombe Ecosystem (GGE), Tanzania, in particular, is vulnerable to habitat disturbance, and this has both ecological and financial implications since it is home to diverse wildlife, including endangered chimpanzees (Pan troglodytes schweinfurthii), that are important contributors to the national economy through tourism [7].

Gombe National Park, established in 1968, is a small 35-km² forest reserve located 16-km north of Kigoma in Western Tanzania (4°40′S 29°38′E). The park is 1500-m above sea-level with hills sloping westward from a rift escarpment to Lake Tanganyika [9]. It is home to a number of non-human primate species, including baboons (Pabio anubis), and a well-known wild chimpanzee population studied continuously for over 50 years [10,11]. There are three chimpanzee communities (Kasekela, Mitumba and Kalande); two of which, (Kasekela, and Mitumba) are habituated [12]. The habitat ranges of these two communities overlap slightly permitting opportunity for member contact. Their habitats have differing degrees of human encroachment [13]. Kasekela, the larger community (∼ 65 individuals), is situated at the center of the park in less disturbed forest, whereas Mitumba, the smaller Northern community (∼ 25 individuals), is in close proximity to Mwamgongo (4°40′S, 29°34’60′ E), a village home to ∼5000 inhabitants and their livestock. Another village borders the park to the South, but not along the Eastern ridge, due to high elevation and historic soil depletion. Human presence in the park is limited to researchers, tourists, park management staff, local field assistants and members of their families. The park border is not fenced and therefore villagers and their untethered animals (goats, sheep and dogs) are able to enter the park [14]). Mitumba chimpanzees are frequently reported raiding agricultural fields to the east in the Northern village, Mwamgongo, especially during the dry season (I. Lipende, personal communication). There is little evidence that the chimpanzee population has emigrated outside its established habitat for over 20 years, and immigration events are rare.

Death from infectious diseases is the leading cause of mortality for Gombe chimpanzees [15,16]. The chimpanzees have experienced SIVcpz-associated mortality and morbidity, with SIVcpz prevalence ranging between 9–18% and a 10–16-fold higher age-corrected death hazard for infected individuals [17]. Cryptosporidium is of special concern in this chimpanzee population, as SIVcpz illness may be complicated by Cryptosporidium co-infection, and mirror clinical features observed in human HIV/Cryptosporidium co-infections [2], that report Cryptosporidium infection rates from 8–30% [18,19]. To improve our understanding of this relationship, and highlight potential management options, we investigated risk factors for Cryptosporidium infection and assessed cross-species transmission potential among people, non-human primates, and domestic animals in the GGE, Kigoma District, Tanzania.

Six hundred and eighty-four fecal specimens were screened for Cryptosporidium including 254 human, 99 domestic animal (n = 76 goat, n = 14 sheep, n = 9 dog) and 331 wildlife (n = 251 chimpanzee, n = 80 baboon) specimens. Cryptosporidium spp, were detected by PCR from 40 (5.8%) fecal samples but was not detected in any water samples (n = 42). The infection rate of Cryptosporidium was highest among 21/131 (16.0%) nonhuman primates tested, compared to 7/73 (9.6%) livestock and 8/185 (4.3%) humans. No significant differences in frequency were observed between chimpanzees and baboons (Table 1, Fisher’s exact test p = 0.457) or between the two chimpanzee communities (Table 1, Fisher’s exact test p = 0.7655). Of the 8 cases of Cryptosporidium detected in humans, 7 (87.5%) resided in Mwamgongo village and one (12.5%) in Mitumba camp. No human cases were detected in the Kasekela camp. Sheep had the highest occurrence of Cryptosporidium (22%) compared to goat (9%) and dogs (0%) but small sample size prevented evaluation of significance.

We identified three species of Cryptosporidium (C. hominis, C. suis and C. xiaoi) in this population (Table 1) based on RFLP and sequence analyses of the SSU rRNA gene. C. hominis was detected in all human cases and all 7 cases from sheep and goats were C. xiaoi. Six of the 12 positive chimpanzees from Kasekela were genotyped as C. suis (a Cryptosporidium species predominantly associated with pigs but has been found in a few human cases). This species was not detected in the Mitumba chimpanzee community (Fisher’s two-tailed exact test; p-value = 0.0537). The porcine species was also not found in the specimens from baboons, humans or domestic animals in the village. The remaining Kasekela chimpanzees (n = 6) and the 4 Mitumba chimpanzees had C. hominis. All baboons (n = 5 individuals) also had C. hominis. GP60 subtyping of a subset (n = 16) of the C. hominis positive samples identified a common subtype IfA12G2 in humans and nonhuman primates. The subtype sequence was identical to two sequences in GenBank; a human C. hominis IfA12G2 sequence from South Africa (GenBank accession number JN867334) and a sequence recovered from an olive baboon in Kenya (GenBank accession number JF681172).

We used data from the survey to identify potential risk factors for Cryptosporidium infection (Table 2). Among 95 respondents (100%), villagers in Mwamgongo were at greater risk for infection when living with a person who was positive for Cryptosporidium (OR = 9.722; 95% CI 1.741–54.279; p = 0.011). Persons living with Cryptosporidium-positive livestock tended to have a greater odds of infection (OR = 4.750; 95% CI 0.944–23.908; p = 0.059). Other factors related to behaviors, including location, occupation (either agricultural or forestry), and not boiling water for consumption were not statistically significant. Although presence of clinical signs was not statistically significant, when reviewing survey data for the Cryptosporidium positive patients, 4/8 reported having diarrhea; 2 sought treatment at the village clinic (Flagyl and Paracetemol). Four of 8 households (50%) reported at least one additional member of the household experiencing gastrointestinal symptoms, including diarrhea and cramping. Interestingly, affected individuals did not report consuming water from an open source (100%), which appeared protective (OR = 0.156; 95% CI 0.085–2.875) but not statistically significant (p = 0.162). 4% of study respondents reported boiling their water before use. No infected individuals reported boiling their drinking water. An association between season and Cryptosporidium infection was not observed in humans or non-human primates (Tables 2 and 3). Chimpanzee demographic factors such as age and sex were not risk factors for Cryptosporidium infection and evidence of diarrhea was not a reliable predictor of Cryptosporidium illness. Kasekela chimpanzees tended to have a higher likelihood (OR = 7.062; 95% CI 0.398–125.251, p = 0.07) of infection with C. suis as compared to the Mitumba community (Table 3).

Of 90 humans residing in camps within Gombe National Park, only one Cryptosporidium infection was observed (1%). In contrast, Cryptosporidium infection in residents of Mwamgongo village reached 10% during the drier months. These frequencies are comparable to those reported in some studies from children and adults without HIV, ranging from 0–18% [25–27], though frequencies as high as 32% have been reported elsewhere [28]. Site and HIV prevalence would seem to be important factors in human occurrence rates. Although HIV testing was beyond the scope of this study, a recent country report [9] indicates a low HIV prevalence (< 1%) from the Kigoma region where the study occurred. Infection was not statistically associated with gastrointestinal illness or stool consistency for humans or wildlife, findings consistent with earlier studies [28–32]. Surprisingly, unsafe drinking water (i.e. untreated/unboiled water, open water source) was not found to increase risk of Cryptosporidium infection, which could be the result of degradation or improper tapping of the water line, exposing water to environmental contamination [33]. This warrants further study as consumption of contaminated ground water has been repeatedly associated with Cryptosporidium infection [34,35].

Cryptosporidium was detected in baboons (11%) and the two chimpanzee communities (15–21%) at moderate frequencies. Similar frequencies of Cryptosporidium have been detected previously in other African primates. Habituated mountain gorillas in Bwindi Impenetrable National Park (BINP), Uganda had Cryptosporidium in (11%) of specimens sampled; 73% of positive specimens were detected from human-habituated gorillas [36]. Genetic characterization of this population determined that the gorillas and the local human community both carry C. parvum [37]. Eleven percent of red colobus and black-and-white colobus in Kibale National Park (KNP), Uganda were infected with Cryptosporidium; genetic sequences from some humans and colobus from KNP were identical, while two strains from red colobus in the forest interior were infected with a divergent subclade, suggesting the possibility of separate zoonotic and sylvatic cycles [28]. These findings support the zoonotic spillover potential of Cryptosporidium among humans, wildlife and livestock.

Similar patterns have been observed with other directly or environmentally transmitted enteric pathogens. At KNP, Giardia was detected in red colobus in forest fragments (5.7%), but not detected in undisturbed forest (0%) [38]. Genetic analyses of recovered strains determined that red colobus were infected with Giardia duodenalis assemblages associated with humans and livestock, suggesting complex cross-species transmission [39]. Escherichia coli strains from gorillas in BINP and chimpanzees in KNP that overlapped with humans were more similar to strains collected from resident humans and livestock compared to strains collected from gorillas and chimpanzees living in undisturbed forest [40,41]. Additionally, in KNP, genetic similarity between human/livestock and primate bacteria increased three-fold with a moderate to high increase in anthropogenic disturbance of forest fragment. Bacteria harbored by humans and livestock were more similar to those of monkeys that entered human settlements to raid crops, than to bacteria of other primate species [42]. These findings reinforce the notion that habitat overlap and anthropogenic disturbance increase the risk of interspecies transmission between wildlife, humans and livestock and that transmission can occur both via direct physical contact with or ingestion of contaminated feces and by indirect exposure via a shared (potentially contaminated) watershed. Although water samples screened in this study were negative for the parasite, waterborne outbreaks of Cryptosporidium as a result of human and animal fecal contamination are common [43–45]. Watershed sampling for Cryptosporidium in this study was opportunistic using smaller volumes of water than advocated by standard screening protocols [46] and provided only limited inter-seasonal sampling [47–49]. Therefore, our negative results do not assure that waterborne transmission is not important in this system. Future studies using more comprehensive watershed sampling would help to resolve this aspect of Cryptosporidium transmission.

The results of our RFLP and sequence analyses of the SSU rRNA gene suggest multiple potential zoonotic pathways for Cryptosporidium transmission in this study system. The village data reinforces our understanding that species of Cryptosporidium vary in their zoonotic potential [50]. Our data suggest that there is less likelihood that the C. xiaoi affecting the livestock is capable of causing illness in humans, considering the close proximity of livestock to humans in this community [i.e., animals often residing in homes], where animal-human contact is quite high though it has been documented in other studies [51]. Cryptosporidium hominis was also not detected among the domesticated animals in this study, but C. hominis has been found in other studies to be a zoonotic species, affecting both humans and domesticated animals [52,53]. Homes with positive livestock had a tendency for increased risk of human infection suggesting contribution of environmental factors or behaviors that may place the household at increased risk.

We anticipated that the Kasekela chimpanzee community would have a lower occurrence of Cryptosporidium compared to Mitumba, which shares a natural border with Mwamgongo village. Although not statistically significant, there was a higher frequency of Cryptosporidium recovered in Kasekela. The occurrence of C. hominis was not statistically higher for Mitumba compared to Kasekela. However, the results demonstrated that 50% (6/12) of the Cryptosporidium recovered from the Kasekela community are C. suis, a species associated with pigs. Domestic pigs are not found in the park or village due to religious preference (predominantly Muslim communities). However, bush pigs (Potaochoerus larvatus), native to the Gombe forest habitat, are common in Kasekela but have not been observed frequently in Mitumba forest. Thus bush pigs may serve as the reservoir host for C. suis in this system. Chimpanzees may be infected through contact with this animal (i.e., chimpanzees occasionally consume bush pigs) or by indirect contact with infective feces on the forest floor or the contamination of shared water sources. This putative pathway of transmission is supported by the fact that C. suis was not recovered in the Mitumba chimpanzee community, the baboons, domesticated livestock or village inhabitants. C. suis has zoonotic potential having been previously identified in an HIV+ patient in Lima, Peru, [54] and from patients in Henan, China and England [55,56]. Similarly, Salyer et al [28] found that while Cryptosporidium may be transmitted frequently among domesticated animals, humans and wildlife in areas of overlap, there may be host-parasite specific dynamics that occur in the absence of these regular interactions creating separate transmission cycles.

The appearance of the C. hominis species and its subtype IfA12G2C among the humans, baboons and chimpanzee communities demonstrates the zoonotic transmission potential of this parasite species among these closely related host species and points to a dominance of anthropozonotic transmission in this system. This subtype has been previously reported among captive olive baboons in Kenya [57], and is prevalent in humans in Africa [58]. Additionally, common primate behaviors may increase the likelihood for animal to human transmission. For example, baboons raid camp food reserves and homes, potentially transmitting etiologic agents via infected feces to humans. The Mitumba chimpanzees are reported to raid agricultural fields just outside the park boundaries, which can transmit diseases from the potentially contaminated feces of livestock or exposed human sewage.

Our results are based on small sample sizes, that if increased could alter the frequency and predictions of Cryptosporidium subtypes, infection and illness. The finding of C. suis in the Kasekela chimpanzee community is novel. We presume these chimpanzees experienced cross-species transmission from bush pigs in Gombe forest, because domesticated pigs are absent from the area. Unfortunately, we do not have access to fecal specimens from the local bush pig population to compare to specimens recovered from Kasekela chimpanzees. Despite the high overlap observed between people and livestock in villages in this region, our results suggest that the transmission dynamics of Cryptosporidium for humans and livestock are distinct but the dominance of C. hominis in humans and non-human primates suggest the potential for cross-species transmission. Our findings highlight the complex nature of zoonotic parasite transmission and stress the need for further studies in similar systems.