Diagnostic Strategies and Strain Typing for Johne's Disease in Wood Bison (Bison bison athabascae) Open Access

Wood bison (Bison bison athabascae) are found in the boreal forest across Northern Canada and Alaska, US. After a near-extinction event in the late 1800s (Sanderson et al. 2008), re-introduction and conservation efforts have enabled the wood bison population in Canada to grow to approximately 8,600 animals (Environment and Climate Change Canada, 2018). Nevertheless, wood bison are still listed as (sub-)Special Concern under the Species at Risk Act (SARA) in Canada (COSEWIC 2013). Cattle diseases are considered a major threat to wood bison conservation, as currently half of the free-ranging wood bison reside in populations affected by bovine tuberculosis (bTB) and/or brucellosis (Environment and Climate Change Canada 2018).

Johne’s disease (JD) is a chronic infectious enteritis affecting domestic and wild ruminants worldwide and could also have negative impact on bison health; it is caused by Mycobacterium avium subsp. paratuberculosis (Map), a slow-growing mycobacterium (Whittington et al. 2004; Carta et al. 2013). Furthermore, control of JD poses challenges due to probable cross-reactivity with bTB diagnostic tests (Picasso-Risso et al. 2019). In North America, JD primarily affects cattle, manifesting as progressive weight loss and diarrhea. Clinical signs typically appear 2–5 y postinfection, posing challenges for early detection and complicating JD control initiatives (Sweeney 2011; Eisenberg et al. 2012). In Canada, Map has been detected in wild ruminants, including elk (Cervus canadensis), bison (Bison bison athabascae), caribou (Rangifer tarandus), and bighorn sheep (Ovis canadensis; Forde et al. 2012a, b, 2013; Pruvot et al. 2013). In fecal samples from clinically healthy wood bison, Map DNA has been detected using a direct nested PCR targeting gene 251 (Sibley et al. 2007). More recently, Forde et al. (2013) detected Map DNA in fecal samples from nine free-ranging wood bison herds across northern Canada using IS900, ISMAP02, and F57 as target genes, but were unsuccessful in obtaining a Map isolate. Map isolates were initially classified according to the host species (cattle strain, sheep strain, bison strain). However, as strain typing became more widely applied and these two main types were isolated from a broad range of host species, it was proposed that these strain types be referred to as type I/III (sheep) strain and type II (cattle) strain (Stevenson, 2015). Amongst other important strains that have been described in Map lies the “type B” (bison) strain, a sublineage of type II (cattle) strains first isolated from bison in Montana using IS1311 restriction fragment length polymorphism (Whittington et al. 2001). Further molecular investigation revealed that type B isolates from animals in India differed from those isolated from USA; these were named the “Indian bison type” strain (Sohal et al. 2013).

Clinical manifestation and pathology of JD in wood bison has not been described. Nevertheless, captive and free-ranging plains bison (Bison bison bison) have been described with lesions similar to those seen in cattle, including granulomatous inflammation in the distal small intestine and associated mesenteric lymph nodes (LNs; Buergelt and Ginn 2000; Stabel et al. 2003; Huntley et al. 2005). Bison with histopathological changes consistent with JD had negative results in fecal and tissue culture, suggesting that diagnostic methods validated for cattle may not be reliable for bison (Buergelt and Ginn 2000; Huntley et al. 2005).

We describe the clinical presentation, pathology, and related diagnostics of JD in six wood bison belonging to a captive herd that had a laboratory-confirmed clinical case of JD in 2019. Our objectives were 1) to detect Map in clinically suspected bison by histopathologic examination, quantitative PCR (qPCR), and culture; 2) to identify optimal tissues for Map detection; and 3) to identify the strain isolated.

This research was approved by the Veterinary Sciences Animal Care Committee at the University of Calgary, Calgary Alberta, Canada (AC21-0187). The study focused on six wood bison from a captive herd with 234 individuals in 2020 at the start of the study, located in northeastern Alberta. In this herd, a clinical JD case had been confirmed in 2019 through direct fecal qPCR by a regional commercial laboratory (Prairie Diagnostic Services Inc., Saskatoon, Saskatchewan, Canada). Bison were culled during 2021–2022 based on clinical signs of weight loss and diarrhea, plus qPCR results from fecal samples. Body weights were measured yearly in the fall from 2020 to 2022, using a scale attached to a hydraulic bison squeeze (Berlinic Manufacturing, Saskatoon, Saskatchewan, Canada) as part of routine herd health management interventions. Before euthanasia, as a part of the Map control strategy in the herd, 2–3 fecal samples were obtained; these samples were maintained at 4 C and sent to the Faculty of Veterinary Medicine, University of Calgary, arriving between 1 and 3 d after collection, then stored in the laboratory at 4 C until analyzed within 5 d of arrival at the university.

After humane euthanasia on site, carcasses were transported within approximately 8 h to the Diagnostic Services Unit (DSU), Faculty of Veterinary Medicine, University of Calgary, with necropsies performed within 48 h after euthanasia. To avoid cross-contamination, no other ruminants were allowed in the pathology laboratory during necropsies, and for each tissue sample a new set of disinfected instruments and clean gloves were used. All carcasses were weighed prior to necropsy and gross necropsy findings recorded.

A total of 24 tissue samples were collected from each bison, as described in detail (Mortier et al. 2013). Intestinal tissues collected included duodenum; mid- and distal jejunum; proximal, mid- and distal ileum; ileocecal valve; cecum; spiral colon; transverse colon; and rectum. Lymph nodes were sampled at locations corresponding with intestinal segments, except for the spiral colon, transverse colon, and rectum. The hepatic LN, tonsil, retropharyngeal LN (RP LN), and superficial inguinal LN were also sampled. Samples were also collected from the kidney, liver, and spleen. Tissue samples were collected in duplicate; one set was stored for maximum of 2 d at 4 C in a sterile Whirl-Pak Bag (Nasco, Fort Atkinson, Wisconsin, USA) for Map detection and isolation; the other set of samples was placed in 10% neutral buffered formalin solution for 24 h, processed routinely for histologic examination, and stained with both H&E and acid-fast Ziehl-Neelsen (ZN) stains.

For each tissue, H&E and ZN slides were examined in tandem by a board-certified veterinary anatomic pathologist. All H&E slides were then examined on high power (400×) for the presence of granulomatous inflammation characterized by epithelioid macrophages and Langhans-type multinucleated giant cells in the intestinal mucosa or lymph nodes. All ZN slides were scrutinized on high power (400×) for intracellular acid-fast bacilli typical of mycobacteria. A ZN was determined to be negative if no acid-fast bacilli were identified after 20 min of examination. Tissues were considered 1) positive, if granulomatous inflammation and at least one intracellular, acid-fast bacillus were identified; 2) suspicious, if granulomatous inflammation was present but acid-fast bacilli were absent; or 3) negative, if neither granulomatous inflammation nor acid-fast bacilli were present. Because of the degree of autolysis, the severity of the inflammation and the semiquantification of identifiable acid-fast bacilli was not attempted, as has been previously reported by Buergelt et al. (2000). Tissue sections were examined by the pathologist blinded to the culture and qPCR results.

Fecal samples were processed individually. The MagMAX™ Total Nucleic Acid Isolation Kit (Applied Biosystems by Thermo Fisher Scientific, Vilnius, Lithuania) was used following the manufacturer’s protocol. Briefly, 0.3±0.1 g of feces were transferred to tubes prefilled with phosphate-buffered saline (PBS) and mixed in a thermomixer at 13,000 × G for 3 min. The supernatant was combined with 235 μL binding solution in prefilled tubes containing 0.1-mm zirconium beads, followed by a bead beating step, twice for 5 min, with a 5-min pause between. The lysate was transferred to a 96-well plate after centrifugation at 16,000 × G for 3 min. Isopropanol was added to each well, and wells shaken at 550 rpm for 5 min, followed by the addition of magnetic beads, which were shaken at 550 rpm for 5 min. The supernatant was collected and discarded using a 96-well magnetic stand (Applied Biosystems). The beads were rinsed twice with wash solutions 1 and 2, and the supernatant was discarded after each wash step. Next, the plate was shaken at 550 rpm without the lid for 3 min to dry the beads. UltraPure DNase/RNase-Free Distilled Water (Gibco, Carlsbad, California, USA) was added and shaken at 550 rpm for 3 min, and the DNA-containing UltraPure DNase/RNase-Free Distilled Water was obtained using the magnetic stand. The purified DNA was stored at −20 C.

Tissue samples were processed for culture as described (Corbett et al. 2017). Briefly, the mucosal layer of the intestinal tissue was removed with a microscope slide. Perinodal fat was removed from LNs, which were cut into smaller pieces (≤0.5 cm). Subsequently, 2.5 g of each sample was placed in 10 mL of a 0.5% Triton X-100 PBS solution in dissociation tubes (gentleMACS M tubes, Miltenyi Biotech Inc, San Diego, California, USA). For dissociation, a Miltenyi dissociator was used, settings were: 2,753 rounds per run for 53 s each, three times per sample. Samples were transferred to a 50-mL Falcon tube and centrifuged at 4,700 × G for 20 min. Thereafter, the pellet was resuspended in 25 mL of 0.75% hexadecylpyridinium chloride (Thermo Fisher Scientific, Waltham, Massachusetts, USA) and half-strength brain heart infusion (BD Diagnostics, Franklin Lakes, New Jersey, USA) with 4-mm sterile glass beads (n=10) and vortexed vigorously for 2–3 min. Then samples were incubated at 37 C for 3 h before centrifugation at 4,700 × G for 15 min. The pellet was resuspended in 3 mL of antibiotic brew (1 mL paraJEM® AS (Thermo Fisher Scientific), 1 mL full-strength brain heart infusion, and 1 mL double distilled water and incubated overnight at 37 C. Finally, 1 mL of antibiotic brew and 2 mL of enrichment media were added (1 mL para-JEM GS and 1 mL Para-JEM EYS® Thermo Scientific) to culture bottles (para-JEM, Thermo Scientific) and the culture bottles incubated at 37 C for 49 d.

We extracted DNA as described (Forde et al. 2012a). Briefly, after the 49-d incubation, to obtain DNA from the cultures 200 μL of culture broth was added to 800 μL absolute ethanol and centrifuged at 4,200 × G for 9 min. The pellet was washed twice with 900 μL sterile Dulbecco’s PBS (Gibco), centrifuged at 4,200 × G for 9 min, then boiled in 100 μL of sterile UltraPure DNase/RNase-Free Distilled Water (Gibco) at 100 C for 30 min. A final centrifugation step at 4,200 × G for 2.5 min was performed. Finally, the supernatant was transferred to a new sterile 1.5-mL microcentrifuge tube for qPCR. Negative controls for DNA extraction were also included.

For extraction of DNA, 100 mg of macerated tissue samples were homogenized in a 2-mL conical bead-beating tube containing 40 mg of 0.5-mm beads and 400 μL Tris-EDTA buffer. This mixture was bead-beaten three times for 30 s each in a mini-Bead beater (BioSpec Product Inc, Oklahoma, USA), followed by lysis and enzymatic digestion with lysis buffer (buffer ATL) and 30 μL of proteinase K (both Qiagen). Thereafter, we followed the spin column protocol from DNeasy Blood & Tissue kit (Qiagen). Negative controls were included for each extraction.

The JD diagnostic testing of fecal samples of our laboratory was certified through the US Department of Agriculture National Veterinary Services Laboratories (NVSL) by successfully completing the annual Johne’s disease direct PCR proficiency panel. All extracted DNA from feces, tissue, and tissue culture were analyzed by qPCR targeting F57 and IS900 genes and an internal amplification control (IAC), as described (Slana et al. 2008; Forde et al. 2013). Briefly, each reaction contained 10 μL of TaqMan Fast Advanced Master Mix (Applied Biosystems), 10 pmol of each primer, 10 pmol of probes for IS900, F57 and IAC, 500 copies of the internal control plasmid, and 2 μL of DNA template. For qPCR, the thermocycler (CFX96 Thermal Cycler, Bio-Rad, Hercules, California, USA) conditions were 50 C for 2 min; 95 C for 20 s for initial denaturation; followed by 42 cycles of 95 C for 3 s and 60 C for 30 s; then finally, 72 C for 5 min. Negative controls were included in each reaction. Samples were run in duplicate, and mean quantification cycle Cq was recorded, with Cq values <37 considered positive. Results of fecal qPCR were classified according to the Cq value of the IS900 gene, as low (Cq>35), moderate (Cq 26–35), and high shedders (Cq<26; Russo et al. 2022).

To enhance quality of Map isolates for strain typing, culture broth from positive tissue samples was transferred to Middlebrook 7H11 agar slants supplemented with 2 mg/L of mycobactin J (Allied Monitors, Montreal, Quebec, Canada) and growth supplement (OADC, Thermo Fisher Scientific) and incubated for 4–8 wk at 37 C, as described (Ahlstrom et al. 2014). Thereafter, DNA was extracted by dissolving multiple colonies in distilled water and boiling at 100 C for 20 min.

Using colonies grown with 7H11 and amplified IS1311, we performed IS1311 PCR–restriction enzyme analysis (PCR-REA) as described (Marsh et al. 1999; Sevilla et al. 2005), adding 10 pmol of M56 and M119 primers, 40 mL of Taq PCR Master Mix (Qiagen) containing 2 U of Taq polymerase, and 5 mL of DNA from each sample, using a Veriti thermocycler (Applied Biosystems). To distinguish between strain type I/III, type II, and bison type, the REA was performed using 2 U of each endonuclease (Hinfl and MseI), supplemented with buffers from the manufacturer (New England Biolabs, Inc, Ipswich, Massachusetts, USA), separated on a 3.5% agarose gel including SYBR SafeDNA gel stain (Thermo Fisher Scientific) and viewed on a Chemi Doc MP imaging system (Bio-Rad). Banding patterns were interpreted as described (Whittington et al. 2001).

We conducted strain typing using single-nucleotide polymorphism-PCR (SNP-PCR) as described (Ahlstrom et al. 2016), enabling identification of the five predominant Map strains in circulation across Canada. Two regions (regions A and B) in the Map genome were amplified and sequenced to identified differences in five SNPs, which between them enable differentiation between type I/III (sheep) strain, bison strain, type II (cattle) identified as “dominant strain,” a subtype of type II (cattle) designated as the “secondary clade,” and another subtype of type II (cattle) designated “other clade.” The SNP-PCR was performed using 50 mL of Taq PCR Master Mix (Qiagen) containing 2.5 U of Taq polymerase, 10 pmol of RegA-F, 10 pmol of RegB-R, 5 pmol of RegA-R, and 5 pmol of RegB-F, with 1 mL of DNA added. Conditions for qPCR were as follows: an initial denaturation step of 95 C for 5 min was followed by 40 cycles of 95 C for 30 s, 60 C for 30 s, and 72 C for 45 s, with a final elongation step of 72 C for 7 min. A no-template control was included in each PCR run.

After PCR, samples were prepared for Sanger sequencing as described (Ahlstrom et al. 2016), then submitted to the University of Calgary Core DNA Services Centre. The complete genome of Map strain K10 (accession number AE016958.1) was used as a reference to identify regions amplified in the PCR and SNPs. Additionally, bison strain and strain type I/III were used as a control. Chromatograms were aligned and visualized in Geneious Prime® ver 2023.1.2 (Biomatters Ltd) to assign alleles at each locus.

The complete genome of Map strain K10 (accession number AE016958.1) classified as dominant strain according to Ahlstrom et al. (2016), was used as a reference to identify the five SNPs. The coordinates for the SNPs were as follows: In region A, the coordinates are 2833167 for SNP1 and 2833233 for SNP2. In region B, the coordinates are 3552999 for SNP3, 3553287 for SNP4, and 3553309 for SNP5. The differences in nucleotides in the SNPs were identified as described by Ahlstrom et al. (2016).

All six wood bison were females between 2 and 10 y old (Table 1). All individuals presented weight loss, ranging from 13 to 42% (bison 5 and 2, respectively; Fig 1).

Fecal samples from bison 1 were not available before necropsy. Bison 2, 3, and 4 were each positive by fecal qPCR at 3/3 sampling times: October 2021 and February and September 2022; October 2021 and September and October 2022; and October 2021 and February and October 2022, respectively. Fecal samples from bison 5 and 6 were each positive in 2/3 time points of sampling (Table 1).

There was moderate to severe autolysis in all animals examined. In some instances, autolysis was so severe that some tissue samples could not be identified and obtained. Diarrhea and signs of dehydration (sunken eye, tightening of skin) were noted in bison 1, 2, and 6. No evidence of Johne’s disease characteristic thickening or corrugation of intestinal mucosa was observed in any bison. Bison 1, 2, and 6 had marked enlargement of the mesenteric LNs.

Because of the moderate to severe autolytic changes, only the presence or absence of granulomatous lesions (epithelioid macrophages and multinucleated giant cells) with acid-fast bacilli were reported (Table 2). Two samples, one each from bison 3 and 6, were classified as suspicious and merged with the positive category. In less severely autolyzed tissues lesions were typical of JD, with granulomatous enteritis characterized by epithelioid macrophages and Langhans-type multinucleated giant cells within the lamina propria and occasionally extending into the submucosa; granulomatous lymphangitis; and granulomatous lymphadenitis characterized by multifocal, noncaseating granulomas in the paracortex of mesenteric lymph nodes. Variable numbers of 1-μm, intracellular acid-fast bacilli were noted in epithelioid macrophages and multinucleated giant cells (Fig. 2). In severely autolyzed tissues, lesions not visible on H&E could still be visualized on ZN. The microscopic findings for each sample are summarized in Table 2. In bison 1, 42% (10/24) of tissue samples were positive, from midjejunum to distal ileum, associated LNs, and ileocecal valve. In bison 2, 36% (4/11) of samples were positive, including midjejunal LN, proximal ileum, ileal LN, and ileocecal valve. Bison 3 had 13% (3/24) of tissue samples positive, including midjejunal LN, distal jejunal LN, and ileal LN. In bison 4, 33% (8/24) of tissue samples were positive, including duodenal LN, midjejunum, distal jejunum, and distal ileum and associated LN. Bison 5 was positive in 29% (7/24) of tissue samples, including midjejunum, distal jejunum, proximal ileum, midileum and associated LN. Bison 6 was positive in 25% (6/24) of samples, including distal jejunum and associated LN, proximal, mid- and distal ileum, and ileocecal valve.

In culture, 62.5% (90/144) of samples were positive. In bison 1, 71% (17/24) of tissue samples were culture positive, including all intestinal sections from midjejunum to rectum, plus tonsil and hepatic LN. In bison 2, 75% (18/24) samples were positive, including inguinal LN and intestinal sections from duodenum to rectum. In bison 3, 46% (11/24) of samples were positive, including RP LN, jejunum and associated LN, midileum, ileal LN, ileocecal valve, cecum, spiral colon, and rectum. In bison 4, 63% (15/24) of samples were positive, including all intestinal sections from duodenal LN to rectum except for cecal LN. In bison 5, 54% (13/24) of samples were positive, from midjejunum to transverse colon, except for spiral colon. In bison 6, 67% (16/24) of samples were positive, including hepatic LN and all intestinal sections from midjejunum to rectum (Table 2).

According to positive results in both genes IS900 and F57, bison 1 was positive in 36% (5/14) of samples; four were LN (hepatic, midjejunal, distal jejunal, and ileal LN) and ileocecal valve (Table 2). Bison 2 was positive in 50% (12/24) of samples, from midjejunum to distal ileum, except for ileal LN. Bison 3 was positive in 18% (4/22) of samples, midjejunal LN, distal jejunal LN and ileal LN, and midileum. In bison 4, 38% (9/24) of samples were positive, inguinal and duodenal LN, all jejunal sections, midileum, distal ileum, and rectum. In bison 5, 50% (11/22) of samples were positive, from midjejunum to distal ileum, ileocecal valve, cecum, transverse colon, and rectum. Finally, in bison 6, 63% (15/24) of samples were positive, including hepatic LN and intestinal sections from midjejunum to transverse colon.

The IS1311 REA was performed for Map isolates from midjejunum, midjejunum LN, midileum, ileocecal valve, and cecum tissue samples in all six bison. The SNP-PCR was performed from the same tissue samples except for cecum and ileocecal valve. All samples produced a banding pattern consistent with a type II (cattle) strain (Fig. 3). Additionally, the SNP 2 from region A confirmed a type II (cattle) strain, specifically secondary clade, in the isolates.

Mycobacterial culture from tissues was the diagnostic test with the highest number of positive samples, 62.5% (90/144), followed by direct qPCR with 43.1% (56/130), and histopathology with 29% (38/131; Table 3). Considering the three diagnostic techniques together, the jejunum and associated lymph nodes (from mid- and distal sections) were the most reliable for detecting Map, with 100% (17/17) positive samples in distal jejunal LN, 89% (16/18) positive samples in midjejunal LN, 88% (14/16) positive samples from distal jejunum; there were 83% (15/18) positive samples from ileal LN (Fig. 4).

In this study we described the pathology findings and diagnostic test results for JD using qPCR IS900/F57 and culture from tissue samples in six clinically affected captive wood bison. Our study complemented the findings of Sibley et al. (2007), and Forde et al. (2013), who reported Map DNA-positive fecal samples in healthy free-ranging wood bison, as evidence of Map presence in this species.

All six bison had shown typical clinical signs of JD, that is, weight loss or diarrhea, similar to those described in domestic cattle and plains bison (Buergelt et al. 2000; Manning and Collins 2001). Thickening and corrugation of the intestinal mucosa typical of JD was not observed in any of the bison analyzed. This was possibly because of advanced autolysis, as bison were necropsied 24–48 h after euthanasia; however, macroscopic lesions may not be observed in all animals with JD. For example, in a previous study in plains bison, corrugation was presented only in 11% of animals with JD (Buergelt et al. 2000). Despite severe autolysis and the absence of macroscopic lesions, histopathologic lesions of granulomatous inflammation of the mesenteric lymph nodes and distal small intestine were observed (often better seen with ZN stain than H&E) in all six bison. This suggests that routine microscopic examination of these tissues including ZN staining could play an important role in passive surveillance of JD in this species. The degree of autolysis that we worked with is commonly encountered, especially in wildlife species, but still enabled JD diagnosis, again supporting the routine histopathological examination of a wide array of intestinal and associated lymph node samples for JD diagnostics in passive surveillance. Buergelt and Ginn (2000) found that in most bison analyzed, granulomatous lesions were more consistent in mid- and upper jejunum, followed by ileocecal LN, whereas Huntley et al. (2005) reported middle ileum as the tissue location most frequently found positive in 14 plains bison. In our study, the more frequent tissue locations with granulomatous lesions were midjejunum and distal jejunum LN, followed by ileal LN (Table 2); suspected histopathology was classified as positive, as the lack of confirmation was probably a result of autolysis. These findings, although not accompanied by visible thickening or corrugation, may explain disturbed nutrient absorption and weight loss.

Mycobacterial culture from tissues, with 62.5% (90/144) of positive samples, was the diagnostic test with the most positive samples, followed by direct qPCR with 43.1% (56/130), and histopathology with 29% (38/131; Table 3). Although qPCR might have higher sensitivity compared to conventional culture, other studies have reported similar results using qPCR and liquid culture, concluding that conducting both tests can be beneficial (Acharya et al. 2017). Several factors can affect qPCR results, including the amount of tissue used in the DNA extraction protocol, as this can affect bacterial load in samples (Lorente-Leal et al. 2019). We used 2.5 g of intestinal mucosa for culture, but only 0.1 in the DNA extraction for qPCR; this might have reduced availability of bacterial DNA for qPCR, making it more challenging to detect, especially when no macroscopic lesions were observed. Our finding that distal jejunum and associated lymph nodes were the most reliable tissue samples for detecting Map with either qPCR or bacterial culture, regardless of tissue autolysis or the absence of visible lesions, has practical relevance for field necropsies because it can help with selection of preferred sampling sites to detect Map in wood bison, as well as selection of diagnostic technique.

The type II (cattle) strain identified in a selected set of tissues using IS1311 REA and SNP-PCR is well known for being a common and relatively quick and easy strain to culture (between 4 and 6 wk) compared to other strains (Dimareli-Malli et al. 2013; Stevenson 2015). This may explain our high success rate at culturing, compared to previous studies in bison, in which only DNA from fecal samples was detected.

Strain typing is important to understand epidemiology and phylogeny of isolates. Techniques such as whole-genome sequencing and SNP detection can help to understand within-herd and within-animal strain diversity, plus transmission patterns (Fawzy et al. 2018; Nigsch et al. 2021). The type II (cattle) strain designated the secondary clade, which we identified using SNP-PCR, has been reported in dairy herds from Alberta, Saskatchewan and Quebec, Canada, but has not yet been observed among other globally reported Map isolates (Ahlstrom et al. 2016).

In previous studies in wood bison, obtaining samples from clinical cases and isolating Map has been challenging, primarily because studies were conducted on free-ranging wood bison, with samples collected opportunistically (Sibley et al. 2007; Forde et al. 2013). In contrast, our study used animals from a captive herd with management and density characteristics more similar to those of domestic cattle. Such conditions might increase the probability of Map transmission and observation of clinical cases (Manning 2011; Krieger et al. 2023). Thus, information presented in this study is most useful to inform management decisions for controlling and detecting Map in wood bison maintained under similar conditions to the herd studied.

We thank the many individuals involved in the fecal and tissue sample processing, including Samita Shrestha, Zhuohan Miao, Karina Cirone. We also express our appreciation to the necropsy staff at the DSU of the University of Calgary. Special thanks to the personnel at the Beaver Creek Ranch and DVMs Roy Lewis and Bob Gilbert. We acknowledge Samantha Tavener for her support during this research. Funding for this project was provided by Syncrude Canada Ltd.