Pulmonary Tuberculosis in Asian Elephants ( Elephas maximus )

Abstract

Although Mycobacterium tuberculosis infection is an important health concern for Asian elephants (Elephas maximus), no studies have evaluated the associated local immune responses or histologic lesions. In primates including humans, latent tuberculosis is distinguished by well-organized granulomas with T_H1 cytokine expression, whereas active disease is characterized by poorly organized inflammation and local imbalance in T_H1/T_H2 cytokines. This study examined archival, formalin-fixed, paraffin-embedded lung samples from 5 tuberculosis-negative and 9 tuberculosis-positive Asian elephants. Lesions were assessed by light microscopy, and lymphoid infiltrates were characterized by CD3 and CD20 immunolabeling. Expression of T_H1 (interferon [IFN]–γ, tumor necrosis factor [TNF]–α) and T_H2 (interleukin [IL]–4, IL-10, transforming growth factor [TGF]-β) cytokines was determined using in situ hybridization. In 6 of 9 samples, inflammation was similar to the pattern of primate active disease with low to moderate numbers of lymphocytes, most of which were CD20 positive. In 1 sample, inflammation was most similar to latent tuberculosis in primates with numerous CD3-positive lymphocytes. Expression of IFN-γ was detected in 3 of 8 tuberculosis-positive samples. Expression of TNF-α was detected in 3 of 8 positive samples, including the one with latent morphology. Low-level expression of IL-4 was present in 4 of 8 positive samples. Only single positive samples displayed expression of IL-10 and TGF-β. Tuberculosis-negative samples generally lacked cytokine expression. Results showed heterogeneity in lesions of elephant tuberculosis similar to those of latent and active disease in primates, with variable expression of both T_H1 and T_H2 cytokines.

The Asian elephant (Elephas maximus) is a CITES, Appendix I endangered species. In addition to factors such as habitat loss and human-animal conflict, infectious disease contributes to the endangered status of the Asian elephant. One of the most important infectious diseases currently affecting elephant health, management, and conservation is tuberculosis. Since 1994, more than 50 culture-confirmed cases of elephant tuberculosis have been diagnosed in the US population. ^{30 –33} Most cases have been due to Mycobacterium tuberculosis, the cause of human tuberculosis. ^{30 –33,37} In addition to individual animal and herd health concerns, elephant mycobacterial infection has significant public health implications. Throughout the world, captive elephants interact closely with human handlers for work and exhibition, providing opportunity for exposure and potential zoonotic and/or anthropozoonotic transmission of infection. ^{20,29,30,32,34} The threat of disease also extends to wild populations of elephants. Although infection has not been documented in wild elephants to date, captive working elephants in native countries frequently mingle with free-ranging elephants, providing opportunity for disease transmission. ^30,35 Public health vigilance and effective conservation of this endangered species require a better understanding of the pathogenesis of tuberculosis.

Tuberculosis in elephants is typically a subclinical and chronic disease, and the cause of elephant susceptibility to M. tuberculosis infection is currently unknown. ^{20,26,28,30 –32,37,42} In humans and other studied species, the character of the associated immune response is believed to be instrumental in determining infection outcome. Because M. tuberculosis is an intracellular pathogen, an appropriate, cell-mediated response driven by T_H1 cytokines is required to control infection, and inadequate cell-mediated immunity is associated with active disease. ^{1,6,8,12,13,23,40,41,44,46 –48,52 *} The immune response to tuberculosis consists of both local and systemic components. Local innate immunity is the first line of defense, and the character of the local immune response will determine disease progression. ^{9,10,18,24,38} In humans and nonhuman primates, latent and active tuberculosis can be distinguished by unique local microanatomic morphologic and immunologic features. ^5,19,38,45 Infection of resistant individuals most commonly results in latent disease characterized by well-organized granulomas containing low numbers of organisms. Susceptible individuals more often succumb to progressive disease characterized by poorly organized, multibacillary granulomas. Latent granulomas have richly vascularized walls and are typically solid, with large numbers of activated, epithelioid macrophages. Progressive granulomas of active disease are poorly vascularized and often contain large central regions of necrosis with few activated macrophages. ^19,49,50 In addition, progressive granulomas are associated with decreased expression of T_H1 cytokines (interferon [IFN]–γ, tumor necrosis factor [TNF]–α, and interleukin [IL]–12) compared with latent granulomas. ^{2,16,17,19,43,50} Importantly, local immune responses of active tuberculosis patients are dependent on stage of disease and may not correlate with systemic immune responses. ^3,4,7,21 Therefore, assessment of the local immune response at the site of infection may provide a more accurate representation of disease stage and ultimate infection outcome.

No previous studies have described the histopathology or local immune responses in Mycobacterium spp–infected elephants. Given what is known about latent and active tuberculosis in primates, evaluation of lesion morphology in elephants and analysis of local T_H1 and T_H2 cytokine expression may provide insight into disease susceptibility and progression. This study sought to describe and characterize pulmonary lesions in Asian elephants infected with M. tuberculosis and to assess local immune responses by examining lymphocyte populations and measuring cytokine levels at the site of disease.

Materials and Methods

Cases/Sample Collection

Cases were from 14 deceased captive Asian elephants. Tuberculosis-positive cases (n = 9) were previously diagnosed at contributing facilities based on histologic lesions, acid-fast staining, and postmortem mycobacterial culture of affected lung tissue. Tuberculosis-negative cases (n = 5) included animals that died or were euthanized due to a variety of causes and exhibited no postmortem evidence of mycobacterial infection. For each case, a single representative sample of archival, formalin-fixed, paraffin-embedded lung collected during routine postmortem examination was used for all analyses.

Light Microscopy and Immunohistochemistry

Formalin-fixed, paraffin-embedded tissues were sectioned at 5 μm and stained with hematoxylin and eosin. Sections from tuberculosis-positive cases were also processed routinely for immunohistochemistry (IHC) using polyclonal rabbit anti-CD3 and anti-CD20 antibodies (Biocare Medical, Concord, CA [CD3]; Thermo-Scientific, Waltham, MA [CD20]). Sections of normal Asian elephant lymph node served as a positive control. Relative numbers of CD3-positive and CD20-positive lymphocytes were determined by visual evaluation of ten 400× fields containing lymphocytes.

In Situ Hybridization

Riboprobes were designed using previously determined Asian elephant cytokine sequence information. ²⁵ Elephant template DNA lengths varied from 250 to 600 bps. Riboprobes were synthesized by in vitro transcription of linearized pGEM-T (pGEM-T Vector; Promega, Madison, WI) plasmid templates containing the appropriate insert to generate the antisense or corresponding sense probes using T7 or SP6 RNA polymerase. In vitro transcription was performed using the Maxiscript SP6/T7 kit (Ambion, Austin, TX), with plasmid DNA (0.2 μg) and ³⁵ S-UTP (0.06 mCi; Perkin Elmer, Waltham, MA). Transcription proceeded at 37°C for 3 hours prior to addition of DNAse for an additional 20 minutes. Probes were then heated to 65°C for 5 minutes prior to addition of 15 μl RNAse-free water. Newly synthesized probes were centrifuged for 4 minutes at 1000 rcf in RNA Mini Quick Spin Columns (Roche Diagnostics, Indianapolis, IN) to remove unincorporated nucleotides and counted on a scintillation counter (Packard Tri-Carb 1500 Liquid Scintillation Analyzer; Global Medical Instrumentation, Ramsey, MN). For use in hybridization experiments, all probes were diluted with RNAse-free water to a final concentration of 500 000 cpm/μl.

From each sample, 5-μm-thick serial sections were cut and mounted on SUPERFROST PLUS-Adhesion slides (Fisher Scientific, Pittsburgh, PA). For each sample, sense and antisense probes were hybridized in parallel on serial sections of lung. Detection of Asian elephant β-actin and CXCL9 served as positive controls for in situ hybridization (ISH) experiments. To serve as a general protocol control, fresh-frozen sections of lung from a M. tuberculosis–infected rhesus macaque (Macaca mulatta) were hybridized with a macaque RANTES probe in parallel with the elephant samples. The macaque tissue and probe had been previously validated with the ISH protocol during prior studies; all macaque tissues displayed antisense probe-specific signal in the vicinity of tuberculosis lesions as expected. ¹⁹ Following deparaffinization and rehydration of elephant samples, microwave pretreatment to expose messenger RNA (mRNA) targets for hybridization was performed. ISH was then performed as previously described. ^15,39 Hybridization solution was composed of 10× HYB buffer (20% 1 M HEPES, pH 7.2 [Fisher Scientific]; 2% 0.5 M EDTA, pH 8 [Fisher Scientific]; 20% 50× Denhart’s Solution [Sigma-Aldrich, St Louis, MO]; 5% 20 mg/ml PolyA [Roche Diagnostics]; 53% ultrapurified water), 5 M NaCl, 50% dextran sulfate formamide (Sigma-Aldrich), 100 μg/ml yeast tRNA (Applied Biosystems, Foster City, CA), 50 000 cpm/μl probe, 0.1 M dithiothreitol (DTT; Roche Diagnostics), and ultrapurified water to final volume of 20 μl per sample. Following hybridization, samples were dehydrated in graded ethanols containing 0.3 M ammonium acetate, air dried, and coated with NTB-2 radiography emulsion (Kodak Eastman, Rochester, NY). Samples were packaged into light-tight containers with desiccant and stored at 4°C for exposure. Exposure time was probe specific: RANTES, 7 days; β-actin, 5 days; CXCL9, 20 days; IL-2, 35 days; IL-4, 28 days; IL-10, 39 days; IL-12, 35 days; transforming growth factor (TGF)–β, 36 days; TNF-α, 38 days; and IFN-γ, 39 days. Following exposure, samples were developed in D19 developer (Sigma-Aldrich), briefly rinsed in water, fixed in rapid fixer (Sigma-Aldrich), and then counterstained with hematoxylin. Samples were assessed for overall level of sense background signal and distribution and amount of antisense probe-specific signal. Background was subjectively assessed as low, medium, or high. Antisense probe-specific signal was defined as aggregates of silver granules distinct from background and unique to antisense probe-labeled sections. Microscopic distribution of any antisense probe-specific signal was determined. The signal intensity was graded as follows: grade 0, no antisense probe-specific signal; grade 1, <10 antisense probe-specific signals/section; grade 2, 10 to 25 antisense probe-specific signals/section; grade 3, 25 to 50 antisense probe-specific signals/section; and grade 4, >50 antisense probe-specific signals/section.

Results

Case Nos. 1 to 5 were from tuberculosis-negative animals that died or were euthanized due a variety of other causes, including elephant endotheliotropic herpesvirus infection and chronic pododermatitis and osteoarthritis. Case Nos. 6 to 14 were from M. tuberculosis–positive animals. Infection had been previously confirmed by the presence of granulomatous pneumonia containing acid-fast bacilli and mycobacterial culture of postmortem lung samples. In all cases, rare to low numbers of acid-fast bacilli were detected in pulmonary lesions, and M. tuberculosis was isolated from affected lung by national and/or regional microbiology reference laboratories. Within individual positive animals, lesion morphology was uniform throughout affected lung, and archival samples used for the study were considered representative.

A summary of histologic lesions and immunolabeling results is provided in Supplemental Table S1. Histologically, sections from 6 of the 9 tuberculosis-positive cases (Nos. 6–10, 12) shared similar morphology. In these samples, the pulmonary interstitium was expanded by multiple and often coalescing, nodular accumulations of foamy to epithelioid macrophages and few multinucleated giant cells that distorted alveoli and typically obscured greater than 50% of the normal parenchyma. In many areas, macrophages surrounded variably sized but often expansive, central accumulations of hypereosinophilic, necrotic cellular debris and few admixed neutrophils. Necrotic debris was often partially mineralized. The number of lymphocytes and plasma cells surrounding areas of granulomatous inflammation was very few (case Nos. 6, 12), low (case No. 9), or moderate (case Nos. 7, 8). The percentage of lymphocytes that were CD3 positive was <10% (case Nos. 7, 8, 12) or 20% to 30% (case Nos. 6, 9). The percentage of lymphocytes that were CD20 positive was approximately 75% (case Nos. 6, 8), 50% (case Nos. 7, 9), and less than 25% (case No. 12). No specific immunolabeling for CD3 or CD20 was detected in case No. 10. In some areas, scant moderately cellular, collagen-rich fibrous connective tissue surrounded foci of inflammation and minimally extended in the surrounding interstitium. Occasionally, inflammation was centered on and obscured entire small bronchioles. Alveoli adjacent to affected regions had segmental type II pneumocyte hyperplasia. Regionally, alveoli were filled with proteinic material and contained few foamy macrophages (Figs. 1, 2).

Figures 1–6.

Lung; Asian elephant, tuberculosis. Figure 1. Case No. 6. Pulmonary parenchyma is obscured by multiple nodular areas of poorly organized granulomatous inflammation with necrosis, consistent with an “active” lesion. Hematoxylin and eosin (HE). Figure 2. Case No. 6. Inflammation is characterized by central hypereosinophilic and smudged necrotic debris surrounded by peripheral accumulations of plump, epithelioid macrophages and fewer lymphocytes (inset). The lesion lacks significant fibrosis. HE. Figure 3. Case No. 11. The lung contains discrete granulomas with mineralized necrotic centers surrounded by abundant mature fibrous connective tissue, consistent with a “latent” lesion. HE. Figure 4. Case No. 11. Inflammation is characterized by central mineralized debris surrounded by several macrophages, few multinucleated giant cells, lymphocytes (inset), and large peripheral accumulations of hypocellular, collagen-rich, dense fibrous connective tissue. HE. Figure 5. Case No. 6. Lung section hybridized with probe for Asian elephant interferon-γ Antisense probe-specific signal is evident as multiple, distinct aggregates of black silver granules in the granuloma wall (inset). In situ hybridization. Figure 6. Case No. 12. Lung section hybridized with probe for Asian elephant tumor necrosis factor–α Antisense probe-specific signal is evident as multiple, distinct aggregates of black silver granules in the granuloma wall (inset). In situ hybridization.

Lesions in case Nos. 11 and 13 were both discrete and well demarcated but differed in the amount of necrosis and degree of inflammation. In sections from case No. 11, the pulmonary interstitium contained a few discrete nodules characterized by mineralized, central necrotic debris surrounded by several epithelioid macrophages, fewer multinucleated giant cells, numerous lymphocytes, and large concentric accumulations of hypocellular, collagen-rich, dense fibrous connective tissue. Intervening pulmonary parenchyma was histologically normal (Figs. 3, 4). Greater than 50% of noted lymphocytes were CD3 positive, and approximately 50% were positive for CD20. In sections from case No. 13, several large, discrete, nodular regions of the pulmonary parenchyma were hypereosinophilic and smudged with loss of individual cellular detail and karyolysis (necrosis). Necrotic areas were surrounded by moderate numbers of foamy to epithelioid macrophages, scattered neutrophils, multinucleated giant cells, few plasma cells, and low to moderate numbers of lymphocytes, 75% of which had positive immunolabeling for CD20. Approximately 20% of the lymphocytes were positive for CD3. The surrounding interstitium was expanded by moderate accumulations of hypercellular, well-vascularized, collagen-rich, and moderately dense fibrous connective tissue. Intervening pulmonary parenchyma was histologically normal.

The section from case No. 14 was centered on a large bronchus and included a few adjacent bronchioles and only small amounts of intervening alveolar parenchyma. Consequently, noted lesions were distinct from other samples. The bronchus-associated lymphoid tissue (BALT) was hyperplastic. Most lymphocytes were CD20 positive; less than 10% had CD3-positive immunolabeling. Multifocally, epithelium was also elevated by nodular to coalescing aggregates of epithelioid macrophages. Inflammation multifocally extended into the peribronchial interstitium and displaced adjacent alveoli. Affected alveoli had segmental type II pneumocyte hyperplasia, and some contained few macrophages and/or proteinic material.

Results of all ISH experiments are summarized in Supplemental Table S1. Neither β-actin nor CXCL9 antisense probe-specific signals were detected in 2 negative samples (case Nos. 3, 4) and 1 positive sample (case No. 10). Lack of positive control probe signal indicated these samples were unsuitable for ISH, necessitating their exclusion from final analyses.

The patterns of expression for TNF-α and IFN-γ were similar. There was no TNF-α antisense probe-specific signal in any of the tuberculosis-negative samples. Grade 1 IFN-γ–specific signal was present in 1 tuberculosis-negative sample (case No. 5). Of the tuberculosis-positive samples, 3 of 8 contained antisense probe-specific signal for TNF-α (case Nos. 6, 11, 12), and 3 of 8 were positive for IFN-γ (case Nos. 6, 9, 12). Signal for both probes was located within viable portions of granuloma walls and adjacent to areas of necrosis (Figs. 5, 6). No signal was detected within necrotic tissue or portions of normal lung distant from areas of inflammation.

For IL-4, antisense probe-specific signal was noted in 4 of 8 tuberculosis-positive samples (case Nos. 6, 7, 9, 13) but not in tuberculosis-negative samples. Signal was scant when present and distributed multifocally in areas surrounding some granulomas. For both IL-10 and TGF-β, antisense probe-specific signal was noted only in a single tuberculosis-positive sample each (Suppl. Table S1). Signal for IL-10 was noted regionally in a portion of unaffected lung adjacent to but distinct from a region of granulomatous inflammation. Signal for TGF-β was distributed multifocally in areas surrounding a few granulomas.

Discussion

In this study, pulmonary lesions of tuberculosis in elephants were histologically similar to those noted in mycobacterial infection of other species. ^{19,46,47,49,50} Distinct histologic and immunologic differences were present that paralleled features noted in active and latent pulmonary tuberculosis in primates, including humans. Inflammation was predominantly granulomatous and accompanied by variable amounts of necrosis, mineralization, and fibrosis. In 6 of the 9 tuberculosis-positive samples, lesions were morphologically similar and histologically compatible with active tuberculosis, as evidenced by the multifocal to coalescing distribution of noted lesions, lack of organization, extensive necrosis, and minimal fibrosis. In contrast, lesions in case No. 13 were more discrete and organized, potentially suggesting infection was more chronic or that the local immune response was more effective in this individual. Case No. 11 had only quiescent granulomas typical of chronic, largely resolved infection, reminiscent of latent disease in primates.

Lesions from case Nos. 6, 9, 10, and 12 appeared morphologically active and contained very few lymphocytes. Case Nos. 7 and 8, also morphologically active lesions, had higher overall numbers of lymphocytes, but in all 6 of these samples, only a low percentage of the noted lymphocytes were CD3 positive. Most lymphocytes in most active lesions were CD20 positive. Results indicated that poorly organized, active lesions of elephant tuberculosis are characterized by minimal to mild lymphocytic infiltration comprised predominantly of B lymphocytes. Thus, a reduced/inadequate lymphocytic (particularly T-lymphocyte) response may be a feature of tuberculosis immunopathogenesis in elephants, contributing to active disease and progression. In contrast, case No. 11, which was morphologically characterized by quiescent granulomas compatible with latent disease, contained numerous lymphocytes, greater than 50% of which were CD3 positive. Findings in this sample suggested that local T-lymphocyte recruitment may be associated with improved inflammatory organization and containment of disease in elephants. It was important to note, however, that assessment of T- and B-lymphocyte numbers was limited by the low numbers of animals and samples examined. In addition, lymphocyte distribution and density in examined sections may not have been representative of each lesion as a whole.

For this study, ISH was selected as a technique that could provide information about location and distribution of cytokine expression within elephant tuberculosis lesions and allow for correlation between cytokine expression and lesion morphology. Because precise fixation and processing history was unavailable for most of the archival samples, hybridization with β-actin and CXCL9 probes provided a way to assess overall tissue quality for detection of T_H1 and T_H2 cytokines of interest. The housekeeping gene, β-actin, was chosen for its ubiquitous expression, unaffected by exogenous stimuli. ³⁶ Within viable tissues, β-actin hybridization resulted in a low-level, diffuse, non-cell-specific signal. Hybridization for the IFN-γ inducible chemokine, CXCL9, was selected as an additional control based on previous ISH studies in nonhuman primates that demonstrated consistent expression of CXCL9 in association with granulomas of pulmonary tuberculosis. ¹⁹

Results reported here supported a role for certain T_H1 cytokines (TNF-α and IFN-γ) in local immunity to tuberculosis in elephants. Both of these cytokines were detected in the viable walls of some pulmonary granulomas. Expression was lacking in necrotic lesions and in distant unaffected portions of lung. The samples with IFN-γ expression were all characterized by widespread, poorly organized, granulomatous inflammation. IFN-γ is the prototypical T_H1 cytokine that is essential for effective defense against mycobacterial infection. ^18,24,38,53 It is secreted by T lymphocytes and results in activation of infected macrophages. Macrophage activation is required for killing of intracellular mycobacteria and promotion of the cell-mediated immune response. ³⁸ In the case of mycobacterial infection, TNF-α is synergistic with IFN-γ in increasing phagocytic and mycobactericidal activities of macrophages within infected tissues. Effective organization and maintenance of granulomas responsible for control of mycobacterial disease also requires TNF-α. ^10,14 Case No. 11, morphologically characterized by the well-demarcated, quiescent granulomas, was one of the samples with TNF-α expression. Expression of this cytokine could have contributed to containment of disease in this case. However, TNF-α expression was lacking in the other case with more discrete lesion morphology (case No. 13) and was present in 2 cases with disseminated disease (case Nos. 6, 12). Interestingly, case No. 11 also lacked CXCL9 expression. Because this chemokine has been shown to be highly expressed in active lesions of tuberculosis in other species, absence of expression provided additional evidence of chronic, “latent” type local disease in this case. ¹⁹ Again, however, CXCL9 expression was present in case No. 13, the other case with more latent morphology. Stage of disease/duration of infection was likely a contributing factor to these discrepancies in ISH results. Due to variation in expression of TNF-α and CXCL9 among positive samples with different lesion morphology, however, definitive conclusions could not be drawn.

In addition to TNF-α and IFN-γ, a proportion of the positive elephant samples displayed low levels of IL-4 expression. In humans, local expression of IL-4 has been reported only with progressive granulomas of active disease and often correlates with greater degrees of necrosis. ^11,16 In the elephants, IL-4 signal was scant and distributed multifocally in areas surrounding some granulomas. Expression was not specifically correlated with central regions of necrosis. No IL-4 expression was detected in case No. 11, characterized by the well-demarcated, quiescent granulomas. Although these finding suggest IL-4 may contribute to the local immunopathogenesis of tuberculosis in elephants, additional information regarding the initiation and progression of the local pulmonary inflammatory response following mycobacterial infection is required prior to any definitive assessment of significance. For both IL-10 and TGF-β, antisense probe-specific signal was noted only in a single tuberculosis-positive sample each. Limited data precluded meaningful interpretation of potential contributions of these cytokines to tuberculosis local immunopathogenesis in elephant lung.

Overall low expression of cytokines evaluated in this study was not unexpected. Experience with ISH detection of cytokine mRNA in other species suggests low expression of these targets is the norm, especially compared with other targets such as chemokines. ^19,51 In addition, sample quality due to the archival nature of available study material likely affected the success of ISH and IHC experiments. When available, use of fresh-frozen sections can facilitate improved ISH signals. ²² When frozen samples are not available and fixed samples must be used, as was the case for this study, target exposure as well as viability of mRNA targets within tissues must be considered. With ISH procedures, target exposure within samples is essential for effective probe hybridization. Conformational changes resulting from fixation can mask or otherwise alter targets within samples. In addition, longer fixation times prior to tissue processing result in greater alteration of targets. ²² Because the study samples were archival and contributed by several different facilities, formalin fixation times varied, and in many cases, precise times were unknown.

In addition, some degree of RNA degradation will accompany routine tissue fixation and processing. ²² To increase the likelihood of cytokine detection within the archival elephant tissues, ³⁵ S-labeled riboprobes were used in this study. Radiolabeled probes have superior sensitivity to probes labeled with digoxigenin or fluorophores. In addition, for the detection of RNA targets, riboprobes take advantage of RNA-RNA bonds that have higher affinity than DNA-RNA or DNA-DNA bonds. ²⁷ Despite all efforts to optimize hybridization and detection, it was evident that sample quality remained a factor affecting the success of the ISH procedures in this study. Three of the study samples were excluded based on lack of hybridization with both the β-actin and CXCL9-positive control probes. It is also possible that sample quality affected the intensity of hybridization in the other samples.

Results of this study suggest that histologic changes compatible with active and latent manifestations of human pulmonary tuberculosis occur in elephants, but further research is needed to correlate lesion morphology with clinical disease course. Findings also showed that active lesions are associated with reduced lymphocytic infiltration, a general paucity of CD3-positive T lymphocytes, and a predominance of CD20-positive B lymphocytes. Thus, altered/impaired local lymphocyte recruitment could be factor in immunopathogenesis of active disease in elephants. Important information could be gained prospectively through more systematic examination of postmortem lung samples collected from positive elephants. Continued study will be key to advancing current knowledge and further elucidating contributions of elephant local immunity to tuberculosis susceptibility.