Figure 1

Figure 1. Geographic distribution of Mycobacterium africanum cases in white-nosed coatis, Costa Rica, 2024–2026. Red triangles indicate the 7 distinct geographic locations where coatis infected with M. africanumwere...

During 2024–2026, the National System of Conservation Areas (Costa Rica) wildlife officers identified 8 free-ranging male white-nosed coatis in Guanacaste Province with nodular or ulcerative cutaneous lesions and variable clinical deterioration and referred them to wildlife rehabilitation facilities (Figure 1; Appendix Table 1). Clinical signs ranged from localized cutaneous masses and lymphadenitis to extensive ulcerative skin lesions, severe pneumonia, and poor body condition. Several animals received empirical amoxicillin (15 mg/kg) and enrofloxacin (10 mg/kg); however, all deteriorated and died within days of admission. We performed complete necropsies on 6 of the 8 coatis. We examined only archived tissues for the remaining 2 coatis, which died in 2024 with similar lesions.

Figure 2

Figure 2. Representative pathologic and imaging findings in white-nosed coatis infected with Mycobacterium africanumlineage 6, Costa Rica. A) Dorsoventral radiograph from 1 coati revealing 2 well-defined, rounded, soft-tissue masses, 1...

Among the 6 coatis that underwent complete necropsy, all had nodular or ulcerative cutaneous lesions and granulomatous pulmonary disease. We identified additional lesions on serosal surfaces and other internal organs, including the diaphragm, liver, spleen, intestine, and lymph nodes. Histopathologic examination of all 6 animals revealed granulomatous inflammation consisting predominantly of histiocytes, with variable numbers of degenerate neutrophils, lymphocytes, and multinucleated giant cells surrounding central caseous necrosis. Ziehl–Neelsen staining revealed abundant acid-fast bacilli in granulomatous lesions involving various tissues in all 6 coatis (Figure 2; Appendix Table 1).

For each coati, we individually homogenized available skin, lung, and liver tissue specimens in sterile saline by using a sterile tissue grinder. We decontaminated tissue homogenates by using the modified Petroff method (7) and performed DNA extraction and mycobacterial culture on the resulting sediments. We extracted total DNA by using the RSC Viral Total Nucleic Acid Purification Kit (Promega, https://www.promega.com) with the Maxwell instrument, following the manufacturer’s protocol. We inoculated all samples into Löwenstein-Jensen solid medium and MGIT liquid medium (Becton, Dickinson and Company, https://www.bd.com). We incubated Löwenstein-Jensen slants at 35°C–37°C, monitored them weekly for up to 56 days, and classified cultures with no visible growth on day 56 as negative. We incubated MGIT tubes in the BACTEC MGIT 960 system (Becton, Dickinson and Company) at 35°C–37°C for up to 42 days, until the instrument flagged them as positive or negative. We classified cultures with no detectable growth by day 42 as negative.

We detected MTBC DNA by using either the Anyplex MTB/NTM Real-Time Detection Kit (Seegene, https://www.seegene.com) or the Xpert MTB/RIF Ultra cartridge (Cepheid, https://www.cepheid.com), depending on reagent availability, according to manufacturers’ instructions (8). We speciated and typed MTBC-positive samples by using the GenoType MTBC assay (Hain Lifescience, https://www.hain-lifescience.de) and screened for resistance mutations by using GenoType MTBDRplus and MTBDRsl (Hain Lifescience) (9). We included negative extraction and amplification controls in all molecular runs, which yielded no detectable MTBC signal.

We performed tNGS on a lung tissue sample from coati D7264-25 to further characterize MTBC lineage assignment by using the Deeplex Myc-TB assay (Genoscreen, https://www.genoscreen.fr). We quantified, normalized, and sequenced libraries by using paired-end sequencing on the MiSeq platform (Illumina, https://www.illumina.com) and analyzed sequence data by using Deeplex software.

We independently confirmed MTBC DNA detection, by using tissue samples from 4 coatis (D4413-25, D5414-25, D0299-26, and D1753-26). We processed tissue samples separately and mechanically homogenized them by using a BeadBug-6 homogenizer (Benchmark Scientific, https://www.benchmarkscientific.com) with silica glass beads (0.5 mm and 1.0 mm) in Tris-EDTA buffer. We extracted DNA by using the DNeasy Blood and Tissue Kit (QIAGEN, https://www.qiagen.com) after overnight digestion with proteinase K at 56°C before purification. For MTBC detection, we used 2 independent real-time PCRs targeting insertion sequences IS1081 and IS6110. IS1081 primers and probe were from the US Department of Agriculture, and IS6110 used the ID GENE Mycobacterium tuberculosis Complex Duplex kit (IDvet, https://www.innovative-diagnostics.com). We included M. africanum ATCC 35711 and Mycobacterium bovis NCTC 10772 as positive controls. We performed additional characterization by using a multiplex PCR assay targeting regions of difference (RDs; RD1, RD4, RD9, RD12, and RD1mic) (10). We analyzed amplified fragments by using capillary electrophoresis (QIAxcel Advanced System, Qiagen).

Figure 3

Figure 3. Molecular characterization of Mycobacterium africanum lineage 6 in free-ranging, white-nosed coatis, Costa Rica. A) Targeted next-generation sequencing result showing assignment to lineage 6 (M. africanum) within...

Figure 4

Figure 4. Real-time PCR amplification curves of Mycobacterium tuberculosis complex DNA in representative tissue samples in study of Mycobacterium africanumlineage 6 in free-ranging, white-nosed coatis, Costa Rica. D4413-25,...

We did not recover viable isolates, despite repeated culture attempts. We detected MTBC DNA from all 8 coati samples. GenoType MTBC line-probe assays revealed hybridization patterns consistent with M. africanum, and GenoType MTBDRplus and MTBDRsl did not detect mutations associated with first- or second-line antimicrobial drugs. TNGS of the lung tissue sample from coati D7264-25 assigned the sample to M. africanum lineage 6 (Figure 3, panel A). We confirmed MTBC DNA in tissue specimens from 4 coatis. Multiplex PCR analysis of RDs detected RD1, RD4, RD12, and RD1mic but not RD9, a profile compatible with M. africanum (Figure 3, panel B; Figure 4, panel B).

We report M. africanum infection in 8 free-ranging, white-nosed coatis from northwestern Costa Rica from 2024–2026. Affected animals had disseminated granulomatous disease, involving skin and multiple internal organs, with abundant acid-fast bacilli. GenoType MTBC assays, NGS, and RD analysis identified M. africanum lineage 6. M. africanum is primarily associated with human TB and remains largely restricted to West Africa (3). Although sporadic infections have been reported in domestic animals and a few wildlife species, including hyraxes, no definitive animal reservoir has been identified (3–5). Our findings expand the recognized host range of M. africanum in free-ranging wildlife in the Americas (1,3).

Because no viable isolates were recovered for whole-genome sequencing, we cannot exclude an undescribed or highly divergent MTBC variant (6,11). Isolate recovery and whole-genome sequencing will help resolve phylogenetic position and support future investigations of transmission pathways and infection sources.

Cases occurred in areas with mixed agricultural and peridomestic land use, suggesting opportunities for contact among wildlife, livestock, and humans. The source of infection remains uncertain; exposure to infected humans, domestic animals, or contaminated fomites cannot be excluded (3). Whether a sylvatic transmission cycle exists within local coati populations is unknown. In Costa Rica, M. africanum infections in humans have only been documented sporadically (2).

Because bovine TB surveillance in Costa Rica is primarily focused on detection of M. bovis, a member of the MTBC, other MTBC members, including M. africanum, are not routinely investigated in productive animals (12). Consequently, infections caused by less-recognized MTBC lineages could remain undetected in domestic animal populations. Our findings highlight the need to include a broader range of MTBC members in national TB surveillance.

Our findings triggered a coordinated One Health response involving wildlife, veterinary, public health, and academic institutions. Activities included evaluating persons in close contact with affected coatis, informing wildlife professionals and facility personnel, training staff to recognize and report compatible lesions, and reinforcing passive wildlife disease surveillance. The response also strengthened interinstitutional coordination, MTBC diagnostic preparedness, and risk communication without causing unnecessary alarm. The investigation found no evidence of epidemiologic linkage to human TB cases. Our findings illustrate the value of wildlife disease surveillance for detecting unusual health events and prompting interinstitutional coordination and highlight the need for more comprehensive One Health investigations integrating wildlife, livestock, environmental, and human health sectors (13,14).