Estimating the geographic distribution of human Tanapox and potential reservoirs using ecological niche modeling

Tanapox virus, a member of the genus Yatapoxvirus, is a zoonotic virus that causes mild febrile illness and nodular skin lesions in humans and non-human primates [1, 2]. The virus is sufficiently distinct from the related orthopoxviruses –including variola, and monkeypox— that traditional smallpox vaccination offers no cross-reactive immunity to Tanapox [3]. The disease was first observed among humans living in the flood plain of Kenya’s Tana River in 1957 [4]. Additional instances of human disease have been reported in animal handlers at primate facilities in the United States and among international travelers who visited tropical Africa [5–7]. As well, Tanapox was reported extensively from the Democaratic Republic of Congo (DRC) during the period of heightened surveillance post-smallpox eradication.

Clinical disease in humans begins with a febrile prodrome (fever 38-39°C) that can include headache, backache, prostration, and lymphadenopathy. This period lasts for 1–2 days and is followed by lesion development. Lesions evolve to become firm, deep-seated, elevated nodules reaching 20 mm in diameter, which frequently break or become ulcerated, often leading to secondary infections. A patient may have only a single nodule, presumably at the site of virus entry into the skin, or Tanapox lesions can occur in crops of 2–10 [8]. Lesions typically resolve after 6 to 8 weeks with scarring often present [9]. The disease is rarely reported and may be confused with scabies, insect bites, or tropical ulcers [2]. An additional factor for low reporting may could be that patients with limited resources may also avoid seeking medical care as the disease has a relatively benign course.

Over a 40-year period from 1957 to 2003 human Tanapox disease has been sporadically identified across the breadth of Africa; spanning more than 6,000 kilometers from Sierra Leone to Tanzania [4–8]. The area covered includes a range of climatic aspects and several distinct ecological regions [10]. Tanapox is a zoonosis, with non-human primates highlighted as prospective reservoirs [11] (Additional file 1). Studies conducted in Kenya in 1971, after a large outbreak of Tanapox, revealed an association between the prevalence and activity patterns of biting arthropods and Tanapox occurrence leading to the suggestion that arthropods may play a role in outbreaks of human disease [12]. It is conceivable that biting insects could play a role in transmitting the virus mechanically, by moving the virus passively via mouthparts from one person to another.

The sporadic nature of human Tanapox renders traditional prospective epidemiology studies for determination of risk impracticable. Human Tanapox is not a reportable disease in endemic regions, so traditional surveillance systems will not reliably enumerate cases. Medical staff in these areas may be insufficiently trained to recognize the symptoms of Tanapox, and modern diagnostic testing is generally not available. Novel methods for determining disease risk are needed. Here we focus on identifying presumptive zoonotic risks by searching for ecologic commonalities between where Tanapox occurs and the likely reservoir species that live there.

Due to the high diversity of mammalian taxa which interact with humans in these regions it can be very difficult to determine particular taxa involved in disease transmission. For the purpose of this study we have assumed that primates are the likely reservoir based on laboratory findings that indicated only primates show evidence of productive infection after being experimentally inoculated [4]. Other taxa examined, which did not demonstrate Tanapox virus-induced illness, include calves, lambs, pigs, goats, rabbits, guinea-pigs, and mice [4]. Many species of primates exist in the locations where human disease occurs. Human-primate interactions are also common in these locations, where primates may be sold, kept as pets, or serve as agricultural pests [13]. The ecological niche modeling (ENM) technique known as the Genetic Algorithm for Rule-Set Production (GARP) relates ecological information from the landscape of disease to create a set of rules that define factors associated with disease presence [14]. The final result, an (ENM) can be rendered on a map depicting the potential geographic distribution for the pathogen. This study uses the Desktop version of GARP to model the predicted presence of human Tanapox from case locations available in the literature and records from Centers for Disease Control and Prevention (CDC) and World Health Organization (WHO) surveillance activities. In addition, we evaluate a list of potential primate reservoirs to determine which have ranges that coincide with human Tanapox.

The ENM of human Tanapox predicts a potential distribution over much of tropical Africa. Areas predicted in 8 or more models included the forested areas of the Congo basin and West Africa, coastal areas of East Africa, and areas surrounding The African Great Lakes (Lake Albert, Lake Tanganyika and Lake Malawi). While predicted areas in DRC and coastal Kenya would be expected based on input data, the remaining areas indicate potential undescribed presence. In general the model predicts that Tanapox could be found in areas of the continent with high temperatures and rainfall.

Three cases of Tanapox that were reported in travelers returning from Africa were not included in construction of the model because the precise geographic location of the presumed exposure is not known [5–7]. The general locations visited by the travelers included, southeast Republic of Congo, Bagamoyo district of Tanzania, and Sierra Leone. All three of these areas were broadly predicted areas of Tanapox presence (Figure 2).

Among the advantages of using an ENM approach to identify geographic locales with the potential to support pathogen transmission, is the additional information that can be obtained from parsing out which ecologic features most influence the model’s predictive capacity. The identification of temperature (maximum and minimum) and precipitation as the environmental variables with the most influence may prove useful in determining the fauna responsible for disease transmission. The initial reports of Tanapox indicate flooding coinciding with an abundance of mosquito activity prior to the 1957 outbreak [17]. This has led to a theory that insects might be involved in transmission. It has additionally been suggested that Tanapox nodules are more commonly found on the extremities because those areas are often not covered by clothing and susceptible to biting insects [8]. Temperature and precipitation are often linked to insect development, and their influence here may support the theory of arthropod transmission [18].

Several of the primate ENMs used in this study have a high degree of overlap with the ENM created from human Tanapox. In general, areas of West, Central, and Coastal East Africa appear to have environmental aspects suitable for both the highlighted Cercopithecus and Cercocebus genera and human Tanapox. No single species, however, was able to be highlighted as a reservoir. Additionally, the primates with the highest degree of spatial overlap have ranges restricted to West and Central Africa, and therefore are unable to cause human disease in Kenya. This may indicate that the ENMs for both Tanapox and the hypothesized primates are simply ‘finding the forests’ or that a pool of primate species may host the virus. A scenario where overlapping primate ranges with similar environmental characteristics like rainfall and temperature may allow the pathogen to spread from one side of the continent to the other.

Primates have been implicated in Tanapox disease based on both infections at animal facilities and observations of primates in areas where human disease occurred. Cercopithecus and Cercocebus monkeys were noted as present in areas with frequent Tanapox infections in Kenya and have been found to have neutralizing antibodies to the Tanapox virus in laboratories [4, 19]. The genus Cercopithicus contains the most widespread primates in Africa, and are often found in areas that have undergone environmental change associated with human activity including clearing forest for roads, logging, and crop cultivation [20]. Cercopithicus nictitans are known to disperse throughout narrow riverine areas, and have a distribution that closely mirrors river basins. Cercocebus mangabeys are noted to be mainly terrestrial and often inhabit areas close to rivers; areas often mentioned in Tanapox outbreaks in Kenya. These primates inhabit the habitat surrounding human settlements and are often found raiding farms, hunted for meat, or kept as pets [20].

Omnivorous primates known to subsist on fruits and insects in undisturbed forest, will raid crops in forest fragments created by encroaching agriculture. Periods of heightened crop raiding closely mirror peak agriculture production following seasonal rains (primates may prefer cultivated maize to wild fruits) [21]. Humans in these areas have additional exposure to primates while entering forest fragments for daily activities (collecting firewood or water) [21]. The red-tailed monkey, Cercopithecus ascanius, show a proclivity for crop raiding and residents in communities bordering Kibale National Park in Uganda were more likely to harbor bacteria similar to those found in C. ascanius, than other forest primate species [22]. Primates that live in close proximity to humans and have behavior that would facilitate human-primate interactions may similarly facilitate Tanapox transmission and warrant future investigation.

This type of modeling has several limitations, and caution is necessary during interpretation. This analysis is based on the location of human cases, which can often be improperly reported and serve only as a proxy to the complex ecological interactions necessary to model the actual pathogen [23]. The coarse spatial resolution (1 km) of the environmental data used in modeling is also a limiting factor in the spatial resolution in the final map. The ENMs for primates relied on data available from online databases and models could only be constructed for species with sufficient data. While results are given at the species level, they should be interpreted in a broader context. GARP models predict areas of potentially suitable habitat or potential niche, not necessarily the current range or realized niche of a species [24]. This may be especially so with primates, where complex social and behavioral characteristics may govern habitat selection. This method of investigation is not meant to serve as a replacement for biologic or epidemiologic studies, but instead serves as a way to renew research interests in a long-neglected disease.

The creation of an ENM for Tanapox predicts a range that includes areas of historical presence, recent anecdotal reports from travellers, and an expanded area of favourable ecological habitat. Also presented here is a short list of primates that have ecological ranges concordant with human Tanapox and may serve as suitable disease reservoirs. In the fifty years since the disease was first isolated, only a handful of articles have been published on the subject. This study aimed to use ecological modeling techniques to provide additional evidence to an often-neglected disease burden in humans. The results of this research are a step to improving the ability of healthcare workers in resource-limited areas of Africa to diagnose and enhance surveillance for cases of Tanapox. It is hoped this research can be used as a guide to maximize future research by limiting the time and effort spent sampling in areas with low environmental suitability and highlighting a short list of pirmates that might serve as sylvatic hosts or reservoirs of this diseae.