Tuberculosis (TB) is the leading cause of death from a bacterial infectious disease in humans. Caused by Mycobacterium tuberculosis (Mtb), approximately one third of the global population is thought to be infected, with high burdens in lower income countries, including much of Africa and Asia (1). The socioeconomic, and public health costs can be staggering; for example, in South Africa, where 244,053 people were reported with TB in 2016, the national TB program budget (US$244 million) was a significant proportion of the national budget (2). However, these figures neglect the potential impact of human TB on other incidental host species, especially livestock, and wildlife.

Many high human TB burden countries are dependent on animal-related industries such as agriculture and tourism to support their economy. With encroachment of human settlements into land previously used for agriculture or natural habitats, there are increased opportunities for disease transmission at the interface between animals and people. Despite the emerging field of “One Health” and growing knowledge of the zoonotic risks of animal diseases (3), few studies have assessed the impact of human diseases on animals (4). Tuberculosis, caused by Mtb, has been reported in cattle in rural areas of Africa, including the Eastern Cape Province of South Africa, as well as in captive wildlife and pets (5–9). However, infections with Mtb have been discovered only recently in free-ranging wildlife in Asia (10–12). Although animals are typically considered dead-end hosts for Mtb, there is evidence that infected elephants are capable of spreading infection to other elephants and different species, including humans (13–16). Therefore, discovery of animal cases of Mtb infection and disease, especially in free-range, multi-host settings, could have significant implications for species management, public health and veterinary disease control, and conservation endeavors. This case of Mtb disease in a free-ranging African elephant highlights the importance of applying the One Health paradigm to address anthroponoses where important human pathogens, such as Mtb, can be introduced into wildlife populations (4).

At necropsy, an estimated 80% of the left lung and 40–50% of the right lung consisted of multifocal to coalescing encapsulated cavities (10–15 cm in diameter) (Figure 1). The lungs contained a mixture of cavitating lesions and miliary focal granulomas (Figure 2). Impression smears showed clusters of acid-fast positive bacilli. The primary histological finding was multifocal pyogranulomatous pneumonia. Granulomas comprised central foci of variably mineralized necrotic debris and clusters of acid-fast positive bacilli encapsulated in variably thick layers of macrophages and epithelioid cells (many of which contained haematoidin pigment), mixed with small numbers of multinucleate giant cells, lymphocytes, and plasma cells (Figure 3). Similar lesions were found in the bronchial lymph nodes.

High intensity of the test line in the Chembio DPP VetTB assay suggested the presence of IgG antibodies to MTBC-specific fusion antigen ESAT-6/CFP10 (Figure 4). In addition, the presence of high intensity lines for the MTBC-specific antigens including CFP10 protein, ESAT-6/CFP10 fusion protein, and DID65 fusion (MPT70/PstS1/CFP10) protein (17), using the multi-antigen print immunoassay (MAPIA), supported a presumptive diagnosis of TB (Figure 5) (18, 22).

Mycobacterium tuberculosis was isolated from lung, pooled head (retropharyngeal, mandibular), thoracic (mediastinal, tracheobronchial), and mesenteric lymph nodes, which were all the samples collected for mycobacterial culture due to the presence of gross lesions. Spoligotyping to confirm infection (rather than laboratory contaminant) and characterization of the separate culture isolates was performed. The isolates were characterized as belonging to the SIT33/LAM3/F11 family (23, 24) (octal code 776177607760771). Compared to other Mtb isolates previously reported from elephants, the whole genome sequence was unique (BioProject ID is PRJNA430907; BioSample accession is SAMN08380889; http://www.ncbi.nlm.nih.gov/biosample/8380889). However, it clustered in the LAM3/F11 family commonly found in human TB patients in South Africa (24) (Figure 6; Table S1).

This is the first confirmed case of Mtb disease in a free-ranging African elephant and suggests that this anthroponosis may be a greater threat to wildlife populations in Africa than previously recognized. Previously, a presumptive diagnosis of TB was made in a free-ranging African elephant with past human contact in Kenya, but culture of affected tissues was not performed and infection and identity of the suspected MTBC organism were not confirmed (25). Cases of Mtb disease have been reported in free-ranging Asian elephants in Sri Lanka and India, supporting the emergence of this important anthroponotic disease (11, 12, 26).

Mycobacterium bovis is endemic in KNP, where multiple host species are infected (27, 28). In the case described in this report, M. bovis was initially suspected as the most likely causative agent, although M. bovis had not previously been reported in elephants in the park or elsewhere in southern Africa. Investigation of the source of the Mtb in our case was confounded by the likelihood that the elephant may have been infected for many years prior to death, based on the extensive pulmonary pathology. Chronic Mtb infection has been documented for at least 9 years in captive elephants (18, 29). Despite these constraints, a Mtb surveillance program for elephants has been implemented in KNP, which includes retrospective serological testing using DPP VetTB as well as prospective bronchoalveolar lavage for mycobacterial culture.

Since the Mtb isolate from the African elephant clustered with the F11 strain commonly found in human TB patients in South Africa (24), we hypothesize that indirect contact may have occurred through human-derived contaminated food or infectious biological discharge. The carcass was found close to a popular rest stop and staff camp which are frequently visited by local elephants. However, since elephants can move over vast distances and show very little respect for man-made boundaries, there could be other potential sources of infection.

This case is similar to those found in Asia with no identifiable source of Mtb found in free-ranging Asian elephants from parks where tourists have no direct contact with animals (11, 12, 26). However, it is possible that these elephants may have had access to human waste in the park or human settlements along the boundaries. Alternatively, although less likely, Mtb may have been transmitted from another infected animal (29–31).

Transmission of Mtb between humans and from humans to animals has been linked to prolonged close contact, primarily through aerosols (32, 33). A report from Spain described three separate cases of Mtb, between 2007 and 2009, in cattle herds that were linked to active tuberculosis in farm staff (34). Similarly, Mtb has been isolated from cattle and domestic pigs in a number of European countries, as well as an African elephant, agouti and tapir in a Polish zoo (35). In North American zoos, the prevalence of TB in captive Asian elephants is higher than in African elephants, which may reflect a closer association of humans with this species, especially in range countries (36). Studies of TB in captive elephants in India have shown the highest seroprevalence in temple elephants that have high human contact (37). However, recent reports describing Mtb transmission between captive wildlife species over long distances challenge this paradigm. Diagnosis of human TB in a chimpanzee, black rhinoceros, Rocky mountain goats, lar gibbon, giraffe, and tapir has been associated with Mtb-infected Asian elephants in the same zoos but with no documented contact (29, 38, 39).

Indirect contact through environmental contamination is recognized as a risk factor for inter-species transmission of M. bovis, another member of the MTBC. Cattle have been infected by sharing grazing and water sources contaminated by secretions from M. bovis-infected badgers, fallow and red deer, white-tailed deer, and wild boar (39–42). Transmission of M. bovis has occurred in cattle that were exposed to uneaten feed from infected deer after 140 days (43).

Since elephants frequently use their trunk to investigate their environment, it is possible that the elephant in the present case became infected through aerosolization of bacteria on contaminated food or domestic waste from a Mtb-infected human. Urine from infected humans may also contaminate the environment and be a source of Mtb. Studies have shown that pathogenic mycobacteria in fresh human urine could survive up to 2 weeks at 30°C and up to 6 weeks at 15°C (44). Inadequate treatment of waste water may also be a potential source, as Mtb was found in stream water in Slovakia (45). Some of the rivers in KNP also serve as water sources for local communities upstream of KNP, and therefore could be a potential source of Mtb through contamination by human waste.

Although the exact location, number and duration of potential exposures are unknown for this elephant, it is possible that it was a single and brief event. Although frequent and prolonged exposure to a person with TB disease increases the risk of TB infection in humans, infection after casual contact can occur as demonstrated by a human case after a visit to a work site three times, for <15 min per visit (46, 47). Similarly, three workers with no other TB risk factors were infected with Mtb after occupational exposure to contaminated medical waste at a treatment facility (48). The possibility of infection following a single time exposure or limited indirect contact, especially in free-ranging animals, has significant implications for our perspectives on zoonotic and anthroponotic TB transmission.

The knowledge gaps evident in this case regarding a human pathogen in a wildlife species highlights the One Health focus required to address the issue of TB at the human-animal interface. Wildlife TB is recognized as a serious barrier to animal conservation efforts in South-East Asia (49). This threat to human, captive and free-ranging wildlife health is a growing concern globally (50), with the significance of this problem shown by the Government of Nepal's endorsement of The Nepal Elephant Tuberculosis Control and Management Action Plan (2011–2015) (51). Emerging diseases with potential to affect multiple species are of particular concern since wildlife may become reservoirs that impact domestic animal and human health, and biodiversity (52).

Of the 30 high-burden countries listed by the World Health Organization, 22 are elephant range countries, as well as supporting other iconic and endangered wildlife populations, including African wild dogs, mountain gorillas, Malaysian tigers, and black and Sumatran rhinoceros (1, 53). Due to the logistical difficulties in locating carcasses of wild animals for disease surveillance, most of the available information on risk of inter-species TB transmission originates from studies with captive wildlife (16). In Nepal, more than 10 working elephants have died of TB with the isolate identified as belonging to a Mtb lineage found among human patients in that country (30). Studies of TB in elephants in India, Thailand, Sri Lanka, Malaysia, Nepal and Laos have also confirmed transmission between humans and elephants (11, 30, 33, 54–56). Mtb has also been found in other wildlife in these countries, including free-living hanuman langurs in India and free-ranging and captive macaques in four Asian countries (10, 57).

Data on Mtb cases in African wildlife is limited compared to that in Asian species. Mycobacterium tuberculosis was diagnosed in eight species of wildlife at the National Zoological Gardens in South Africa (7). Since that report, there have been two deaths of African elephants at this zoo attributed to Mtb (unpubl. data). In addition, an increase in cases of Mtb documented in captive wildlife facilities in South Africa (including antelope, primates and warthogs) between 2002 and 2011, suggests greater spill-over in parallel with the growing human epidemic (58).

Disease surveillance is essential to determine the presence and extent of Mtb infection in wildlife, however this is limited by the logistical challenges associated with acquiring fresh samples from carcasses. This can be overcome by ante-mortem testing, although there is a paucity of diagnostic tests available for TB detection in wildlife. Serological assays and trunk wash sampling for mycobacterial cultures have been used to identify infected elephants (13, 15, 48). As demonstrated in this case, serological tests using MTBC-specific antigens for antibody detection could provide a presumptive diagnosis of TB in elephants, with reportedly higher sensitivity as compared to trunk wash culture (18, 22, 29, 55). Surveys of TB in elephants using serological tests have reported high prevalence in at-risk populations. For example, in the Malaysian peninsula, elephants had a seroprevalence of 20.4% and their handlers, tested with an interferon-gamma release assay, had a prevalence of 24.8% (33). Similarly, 13% of Nepal's and 15% of India's captive elephant populations were seropositive in recent surveys (35, 51, 54). High prevalence in these populations highlights the significant risk that humans present to captive and, potentially, free-ranging wildlife at interfaces.

Advances in molecular epidemiological and other diagnostic techniques have facilitated understanding of inter-species transmission (29, 58). Whole genome sequencing provides a tool for determining potential sources, number and timing of introduction of pathogens to a population. Although such tools have been commonly used in disease outbreaks in humans (59), they have been under-utilized in conservation efforts for evaluation of risks at interfaces. In addition to Mtb, other human pathogens such as methicillin-resistant Staphylococcus aureus, influenza A virus, and protozoal organisms such as Giardia and Cryptosporidium, present under-recognized biological threats to wildlife and domestic animal species (4). Since many outbreaks occur in developing countries where there is a lack of diagnostic capacity, strengthening collaboration between human and veterinary diagnostic centers would provide expertise and maximize use of limited resources (3). As the human-animal interface increases in complexity and sheer size globally, a greater focus on building multi-disciplinary teams with expertise in veterinary medicine, conservation, animal management, ecology, epidemiology, public health, and infectious diseases are needed to address the threat that human pathogens present to animals, especially threatened and endangered species. Without a One Health approach to managing diseases such as TB, it is unlikely that public health goals will be achieved, since there is a possibility of spill-over and spill-back between species, and this may result in devastating consequences for ecosystems such as loss of biodiversity and impact on environmental health (52, 60).

MM, PB, GH, and LvS conducted the post-mortem examination. EM performed histopathological examination. MM, AS-G, and KL performed serological tests. ER, RW, and AD performed mycobacterial cultures and speciation. SR-A and AD performed whole genome sequencing and analyzed whole genome sequence data. MM, PB, ER, AD, PvH, SP, RW, WW, and KL had key input into project design. All authors assisted with writing and editing the manuscript.