Potential Animal Reservoir of Mycobacterium ulcerans: A Systematic Review

Mycobacterium ulcerans is the causative agent of Buruli ulcer, also known in Australia as Daintree ulcer or Bairnsdale ulcer. This destructive skin disease is characterized by extensive and painless necrosis of the skin and soft tissue with the formation of large ulcers, commonly on the leg or arm. To date, 33 countries with tropical, subtropical and temperate climates in Africa, the Americas, Asia and the Western Pacific have reported cases of Buruli ulcer. The disease is rarely fatal, although it may lead to permanent disability and/or disfigurement if not treated appropriately or in time. It is the third most common mycobacterial infection in the world after tuberculosis and leprosy. The precise mode of transmission of M. ulcerans is yet to be elucidated. Nevertheless, it is possible that the mode of transmission varies with different geographical areas and epidemiological settings. The knowledge about the possible routes of transmission and potential animal reservoirs of M. ulcerans is poorly understood and still remains patchy. Infectious diseases arise from the interaction of agent, host and environment. The majority of emerging or remerging infectious disease in human populations is spread by animals: either wildlife, livestock or pets. Animals may act as hosts or reservoirs and subsequently spread the organism to the environment or directly to the human population. The reservoirs may or may not be the direct source of infection for the hosts; however, they play a major role in maintenance of the organism in the environment, and in the mode of transmission. This remains valid for M. ulcerans. Possums have been suggested as one of the reservoir of M. ulcerans in south-eastern Australia, where possums ingest M. ulcerans from the environment, amplify them and shed the organism through their faeces. We conducted a systematic review with selected key words on PubMed and INFORMIT databases to aggregate available published data on animal reservoirs of M. ulcerans around the world. After certain inclusion and exclusion criteria were implemented, a total of 17 studies was included in the review. A variety of animals around the world e.g., rodents, shrews, possums (ringtail and brushtail), horses, dogs, alpacas, koalas and Indian flap-shelled turtles have been recorded as being infected with M. ulcerans. The majority of studies included in this review identified animal reservoirs as predisposing to the emergence and reemergence of M. ulcerans infection. Taken together, from the selected studies in this systematic review, it is clear that exotic wildlife and native mammals play a significant role as reservoirs for M. ulcerans.


Introduction
Sir Albert Cook, a British missionary doctor appointed at the Mengo Hospital in Kampala, Uganda, first noted the skin ulcer caused by Mycobacterium ulcerans in 1896. Later, in the late 1930s, two general practitioners, Drs. J. R. Searl and D. G. Alsop, working in rural Victoria, Australia, noticed a group of cases of mysterious skin ulcers around the town of Bairnsdale [1]. The cases were not published in the literature at the time and the causative organism was not identified or characterized. Professor Peter MacCallum and his colleagues first provided the detailed description of the disease in 1948, using presentation data of six patients in the Bairnsdale district, near Melbourne. They were the first to isolate M. ulcerans as the causative organism of the mysterious skin ulcer [2]. The first large cluster of M. ulcerans infection was identified in the Buruli County of Uganda (now called Nakasongola District) in the 1960s and the disease was termed 'Buruli ulcer' (BU) thereafter [3].
There have been several known outbreaks of Buruli ulcer around the world and each outbreak has its own unique characteristics in terms of epidemiology and the animals reported to be involved in transmission [4,5]. The World Health Organization (WHO) has classified BU as a neglected tropical disease [6]. Presently, BU has been reported (but not always microbiologically confirmed) in more than 30 countries spread over Africa, the Americas, Asia, and Oceania [7]. Australia is the only developed country with significant local transmission of BU, with foci of infection in tropical Far North Queensland [8,9], the Capricorn Coast region of central Queensland [10], the Northern Territory [11] and temperate coastal Victoria [10]. Non-human cases of M. ulcerans are prevalent in Australia only, where several cases of BU have been described in both native wildlife and domestic mammal species such as koalas (Phascolarctos cinereus) [12,13], common ringtail possums (Pseudocheirus peregrinus) [14,15], a mountain brushtail possum (Trichosurus cunninghami) [5,14,15], two horses [16], an alpaca [17], four dogs [18] and a cat [19]. Recent research in Victoria, Australia, has suggested the transmission of infection by mosquitoes, and possums with chronic BU as an important environmental reservoir of M. ulcerans in Victoria [14].

Materials and Methods
The PRISMA guidelines developed by the Centre for Review Dissemination (CRD) were used as the methodology for the systematic review [20]. A review protocol was registered with PROSPERO international prospective register of systematic reviews, which can be viewed online [21]. The systematic literature review was conducted using online databases MEDLINE and INFORMIT to aggregate all the published literature. Initially, MEDLINE was used to retrieve all the scientific information concerning the research topic. INFORMIT was searched with same search strategies adopted for MEDLINE. The following key words were chosen after a series of trial searches in order to ensure an adequate number of relevant articles were reviewed: (Buruli OR 'Mycobacterium ulcerans') AND (Host OR Vector OR Reservoir OR Animal), accessed on 6 May 2018. The title and abstract of each of the articles were initially scanned to ensure that the included articles met the aim and scope of the systematic review. Articles that were deemed irrelevant to the aim of this systematic review or out of the research scope were excluded. For those articles that were not clear by the title and abstract, the full text was retrieved and further analyzed in order to determine if they met the inclusion and exclusion criteria below. The studies that reported only experimental or laboratory exposure of M. ulcerans in animals were excluded. The search strategy exclusively focused on potential animal reservoirs, not the vectors. The detection of the causative agent had to be confirmed by culture of bacteria and/or PCR. To be considered positive a sample needed to be confirmed either by culture of bacteria or positive for IS 2404 and reconfirmed by KR and IS 2606. Undoubtedly, PCR targeting IS 2404 is highly specific for detecting M. ulcerans in clinical specimen [22]. However, for detecting M. ulcerans from environmental samples, confirmatory PCR targeting two additional insertion sequences, IS 2606 and the ketoreductase B domain (KR), is essential to differentiate M. ulcerans from other environmental mycobacteria that may carry IS 2404 and other non-mycolactone-producing mycobacteria [22]. Thus, IS 2404-PCR used in conjunction with IS 2606 and KR-PCR confirms that the detected organism is M.
ulcerans. There were no language restrictions. Risk of bias was assessed by one reviewer on the basis of independent factors such as sample size, location and nature of infection.

Results of the Literature Search and Method of Inclusion
The total number of discovered articles in MEDLINE database was 351. Three hundred and fourteen articles were excluded after reading the title and abstracts as they were not relevant to the research question. Full texts of thirty-seven studies were retrieved in portable document format (PDF) for further analysis. Of these remaining 37 studies, 19 were excluded as they clearly did not meet inclusion criteria (i.e., they were review articles, focused on vectors rather than on animal reservoirs, or pertained to laboratory or experimental exposure). One additional duplicate article was excluded as well. The remaining 17 studies from the PubMed database were included for systematic review. There were no additional articles in INFORMIT that did not appear in the initial MEDLINE search results. The flow chart for study selection process is shown in Figure 1.

Results of the Literature Search and Method of Inclusion
The total number of discovered articles in MEDLINE database was 351. Three hundred and fourteen articles were excluded after reading the title and abstracts as they were not relevant to the research question. Full texts of thirty-seven studies were retrieved in portable document format (PDF) for further analysis. Of these remaining 37 studies, 19 were excluded as they clearly did not meet inclusion criteria (i.e., they were review articles, focused on vectors rather than on animal reservoirs, or pertained to laboratory or experimental exposure). One additional duplicate article was excluded as well. The remaining 17 studies from the PubMed database were included for systematic review. There were no additional articles in INFORMIT that did not appear in the initial MEDLINE search results. The flow chart for study selection process is shown in Figure 1.

Basic Characteristics of Selected Studies
Out of the 17 included studies, ten were conducted in Australia, two in Ghana and one was conducted in each of Ivory Coast, North America, United States, Benin and Japan. The basic characteristics of selected studies for review are shown in Table 1

Basic Characteristics of Selected Studies
Out of the 17 included studies, ten were conducted in Australia, two in Ghana and one was conducted in each of Ivory Coast, North America, United States, Benin and Japan. The basic characteristics of selected studies for review are shown in Table 1 below.

Discussion on Possible Reservoirs and Vectors of Mycobacterium ulcerans by Country
This systematic review assessed the potential animal reservoir of M. ulcerans around the world recorded to date. This is essential for understanding the epidemiology and mode of transmission of the disease, which subsequently aids in prevention, control and elimination strategies.

Australia
Out of 17 studies included in this review, 10 were conducted in Australia. In Australia, the disease is more prevalent in the southeastern state of Victoria and in Far North Queensland. After the detection of M. ulcerans infection in four koalas in 1980 at Raymond Island, Australia [13], the entire island was searched for koalas in the following year. Thirty-six male and 51 female koalas were captured and examined. Of these, 18 out of 87 animals had skin wounds and 11 were found positive for M. ulcerans. Diagnosis was made on pathological and bacteriological examination; the PCR-based method used for the identification of M. ulcerans from clinical and environmental samples was only implemented in 1996 [30]. Non-human cases of M. ulcerans in Australia have been reported in marsupial species such as koalas [13], ringtail and brushtail possums [14,15,31], horses [16], alpacas [17], dogs [18] and cats [19]. A study conducted by Fyfe and colleagues between 2007-2009, at Point Lonsdale, a small coastal town south east of Melbourne, Australia, which is also endemic for BU, found that 43% of ringtail possum and 29% of brushtail possum faecal samples were positive for M. ulcerans DNA [14]. Only 1% of faecal samples from non-endemic area possums were positive for M. ulcerans DNA in this study, suggesting terrestrial mammals such as possums are potential reservoirs of M. ulcerans in southeast Australia. Several studies have identified possums (both ringtail and brushtail) as potential reservoirs since then [5,15]. In Australia, other than the southeastern state of Victoria, BU is also prevalent in Far North Queensland [8]. Inspired by the evidence of possums as potential reservoirs of M. ulcerans in Victoria, a study conducted by Roltgen and colleagues (2013) in northern Queensland, Australia, detected M. ulcerans DNA from two bandicoot faecal samples, suggesting the possibility that bandicoots are a potential reservoir of M. ulcerans in Far North Queensland [9].

Africa
Out of the 17 studies included in this review, four were conducted in West African countries: two in Ghana [23,25], one in the Ivory Coast [24] and one in Benin [27]. Durnez and colleagues (2006) caught 326 rodents and 222 shrews from endemic and non-endemic villages of Benin and tested for M. ulcerans, but no specific DNA was detected from any of their samples [27]. Despite their results, they suggested the necessity of more intensive research focusing on small mammals in Africa. Willson reported positive PCR with IS 2404 only from tadpoles and fishes from Ghana [25]. Similarly, two faecal specimens from Thryonomys swinderianus (agouti) were reported positive for M. ulcerans in a study conducted by Bi Diangoné Tian and colleagues (2014) from the Ivory Coast [24]. They suggested agouti, which are closely related to Australian possums, could be a potential reservoir of M. ulcerans in Africa. However, RT-PCR targeting IS 2606 was not conducted to confirm M. ulcerans. A faecal survey of domestic animals in rural Ghana for M. ulcerans conducted by Tobias and associates suggested no evidence of association between domestic animals and M. ulcerans in endemic and non-endemic villages in Ghana [23]. Unlike Australia, not a single study in Africa has reported the presence of M. ulcerans-positive DNA or cases in non-human species, suggesting that transmission dynamics may be different in Africa and Australia or, alternatively, a host animal is yet to be identified in Africa.

Other Countries
No study has reported M. ulcerans DNA or cases in non-human species in any country other than Australia. A study conducted by Heckert in 1997 at Chesapeake Bay, USA detected a new Mycobacterium species from wild striped bass [29]. This new isolate was closely related to M. marinum, M. ulcerans, and M. tuberculosis. Similarly, Sakaguchi and associates reported an atypical mycobacterial infection in an Indian flap-shelled turtle (Lissemys punctata punctata), imported from India to Japan in an aquarium [26]. A PCR assay targeting the rpoβ gene revealed the isolate had 89-100% homology to M. ulcerans and M. marinum. Again, this study could not differentiate M. ulcerans from mycolactone-producing M. marinum (MPMM). Appleyard and Clark (2002) reported a new Mycobacterial species, namely 'Mycobacterium visibilis' from three cats initially suspected of having M. ulcerans infection [28].

Conclusions
Human cases of BU have been reported in more than 30 countries from Africa, America, Asia and Oceania. Since the implementation of PCR-based methods for the detection and identification of M. ulcerans from clinical and environmental samples, there has been a significant increase in overall knowledge of BU. There is no record of direct human-to-human transmission of M. ulcerans, unlike tuberculosis and leprosy. Australia is the only country where non-human cases of BU have been identified, with small mammals, especially possums and, to some extent, bandicoots, being implicated as potential reservoirs of M. ulcerans. Despite there having been several outbreaks in African countries, no non-human cases have been recorded so far and there is no evidence of any animal acting as a potential reservoir for this organism. None of the studies included in this review discussed strain variation of M. ulcerans in different geographical regions leading to an increase or decrease in susceptibility among animal or human population. Compared to other mycobacteria, such as M. tuberculosis, there is very little genetic diversity among isolates of M. ulcerans. Some variation among the strains of M. ulcerans from Africa, the Americas, Asia and the Western Pacific has been recorded; however, the linkage between these various strains and virulence in human or animal population has not been recognized so far. Remarkable differences in the type of mycolactone produced by M. ulcerans in different geographical location has been recorded. African strains produce more mycolactone variant A and B, whereas strains from Australia produce more mycolactone variant C. However, this variation has nothing to do with host susceptibility to M. ulcerans; rather, it determines cytopathogenecity and thus clinical presentation of disease.
This systematic review suggests the need for extensive laboratory and field research focusing on domestic animals and wildlife to elucidate their roles in BU-endemic countries.