Aims: To determine the level of anthrax spore contamination in endemic regions of northern Canada between outbreaks.
Methods and Results: Bacterial endospores were extracted from specimens via flotation and cultured on selective PLET medium. Of 588 environmental specimens collected, 11 (1·9%) contained viable anthrax spores.
Conclusions: High environmental concentrations of anthrax spores in northern Canada appear limited to scavenger faeces and anthrax carcass sites. Burial and cremation appear equally effective at removing anthrax spores from the immediate environment, though cremation may be improved by re-burning cremation sites containing unburned animal hair.
Significance and Impact of the Study: This study describes an effective anthrax spore detection system. It provides the first bacteriological evidence that mammalian scavengers can disseminate anthrax spores in northern Canada, and its results may be compared with future environmental studies of untreated anthrax carcass sites to help improve government response plans.
Bacillus anthracis is the causative agent of anthrax, an infectious, often fatal disease of wild and domestic animals, and man. Bacillus anthracis is global in its distribution, with endemic regions on all continents except Antarctica. Central to the maintenance of the disease in an area is the ability of the bacterium to form metabolically-dormant endospores which are highly resistant to a number of environmental forces. Anthrax spores may remain viable and infectious in the environment for years before coming into contact with a susceptible host and initiating a new cycle of disease.
In Canada, a large epizootic of the disease occurred in the free-roaming bison (Bison bison) herds of the Hook Lake region of the North-west Territories (NT) during the summer of 1962 (Fig. 1). Over the next two summers, anthrax outbreaks continued in the Hook Lake region and spread, first across the Slave River into the Grand Detour and Park Central regions, and later, across the Peace River to the Lake One region. Between 1962 and 1991, eight epizootics of anthrax were recorded in these four regions of the NT and northern Alberta, resulting in the deaths of over 1200 bison and several moose (Alces alces) (Broughton 1992; Dragon and Elkin 2001). In 1993, a large epizootic occurred in the Mackenzie Bison Sanctuary (MBS) where 169 bison and three moose carcasses were found concentrated in five main areas: Slave Point, Boulogne Lake, Falaise Lake, Calais Lake and Mink Lake (Gates et al. 1995). Between July and August of 2000, a tenth outbreak occurred in the northern bison herds. The carcasses of 106 bison and three moose were found within the confines of Wood Buffalo National Park (WBNP) in the Lake One region, and along the shores of the Peace River in the south-western portion of the Park.
Research on anthrax in northern Canada has been limited to field observations during active epizootics and has focused mainly on the host. The environmental source and movement of B. anthracis between outbreaks is unknown. The bacteriological study described here was initiated in order to determine the level of spore contamination between outbreaks, and to assess the various methods used to dispose of infected animal carcasses during anthrax outbreaks.
MATERIALS AND METHODS
Between 1992 and 1997, field trips were undertaken to anthrax endemic regions in northern Canada to collect environmental specimens from carcass disposal sites and bison habitat. In August 1992 and July 1993, specimens were collected from anthrax mounds in the Hook Lake region and from burial sites along Parson’s Lake Road (PLR) in central WBNP. Samples collected from the Park were obtained under research permit 92-13 (WBNP, Environment Canada; Environmental assessment registry number WB 92-32). In July of 1994 and 1997, trips were made into the Falaise Lake region of the MBS to collect specimens from carcass disposal sites under NT wildlife research permits WL001018 and WL001547. Soil specimens were also collected randomly from bison wallows and open meadows not associated with known anthrax carcass sites in the Falaise Lake region.
At each carcass site, a central wooden peg was hammered into the ground and environmental specimens were collected from around the site, recording the distance and bearing of each relative to the wooden peg using a tape measure and compass. During surveying of disposal sites, the area was considered contaminated and appropriate safety precautions were taken. Personnel collecting specimens were vaccinated with the US anthrax vaccine preparation (Michigan Department of Public Health, Lansing, Michigan, USA), and wore protective clothing comprising rubber boots, disposal coveralls with hood and elastic cuffs, double-layered latex gloves and a full-face HEPA-filtered M-95 respirator.
Environmental specimens collected comprised a heterogeneous mixture which included soil, charred bone, bison hair, animal faeces, maggot casings, vegetation and ash. Samples were collected using disposable plastic spoons and were taken only to a maximal depth of 5 cm. The samples were carefully transferred to labelled bottles; approximately 15 ml (3–12 g) of sample were collected in each bottle.
Spore extraction and B. anthracis identification
Bacterial endospore extraction and screening was carried out under full Level III biosafety precautions in the anthrax diagnostic laboratory at the Animal Diseases Research Institute (ADRI), Canadian Food Inspection Agency, Lethbridge, Alberta. Prior research by our laboratory with spores from an avirulent strain of B. anthracis indicated that flotation extraction with a high specific gravity sucrose plus Triton X-100 solution may be more effective at isolating spores than standard water extraction in soil types similar to those of the surveyed endemic regions (Dragon and Rennie 2001). Research with avirulent spores also demonstrated that ethanol purification of spores from vegetative contaminants was as effective as standard heat treatment yet less stringent with regard to pre-incubation and exposure time. Flotation extraction and ethanol purification were used in conjunction with culture on PLET (polymyxin, lysozyme, ethylenediaminetetraacetic acid, thallium acetate) medium to isolate viable B. anthracis spores from the environmental specimens. The spore extraction methods employed were as reported in Dragon and Rennie (2001). PLET cultures were incubated at 37°C for 48 h to optimize recovery of anthrax spores.
The identity of colonies isolated with a morphology similar to B. anthracis on the PLET plates was confirmed by lack of haemolysis on blood agar plates, a zone of inhibition around a 10 U penicillin G disc and development of mucoid, encapsulated colonies on bicarbonate agar as determined by Giemsa stain (Carman et al. 1985). Final identification was achieved through multiple-locus variable-number tandem repeat analysis (MLVA) (Keim et al. 2000).
Screening of environmental specimens
During the first outbreaks in the Hook Lake area, carcasses were buried in deep pits with a mound of earth piled on top. Even decades after their construction, these anthrax mounds remain clearly distinguishable on the surrounding flat meadows. Environmental specimens (169) were collected from five anthrax mounds in the Hook Lake region. During the 1991 epizootic, carcasses found along PLR were buried but no mound was erected on the site. Six burial sites in areas of loose-packed sand amid old-growth pine and spruce forest were sampled along PLR, and 157 specimens were collected.
During the 1993 MBS epizootic, bison carcasses were incinerated with either coal or wood and the remains left unburied. One year after the epizootic, the cremation sites were covered in ash with the charred remains of logs about their perimeter. At each site, slightly offset from the centre, was a small bed of charred bones. At six of eight sites surveyed, there were thick mats of bison cape hair and large concentrations of maggot casings underneath the bone bed. Although sometimes singed, the hair and casings were intact and undamaged. A total of 195 specimens were collected from eight cremation sites in the Falaise Lake region. In addition, 23 samples from bison wallows and 44 soil samples were collected from areas of meadow around the lake not associated with carcass sites.
After spore extraction and purification, samples were spread on PLET agar. Although colonies were observed on the plates after overnight incubation, it was impossible to differentiate colonies morphologically until after 48 h of incubation. Of the 588 specimens screened with PLET medium, 174 (29·6%) exhibited domed, circular, white colonies, 4–8 mm in diameter, that were morphologically similar to B. anthracis (Table 1). The highest percentage of specimens with anthracis-like colonies were from Falaise Lake cremation sites and meadow samples, while the lowest percentage was recovered from Hook Lake anthrax mounds.
Table 1. Number of Bacillus anthracis isolates and contaminants with B. anthracis-like properties recovered from environmental specimens from anthrax endemic regions in northern Canada
Despite the selectivity of PLET medium, contamination of the plates was observed. The mean number of contaminants per plate, including B. anthracis-like colonies that were later discounted through confirmatory tests, was 191·2 ± 132·0, 12·7 ± 49·7, 160·0 ± 133·8 and 157·9 ± 118·8 for specimens from Hook Lake, Parson’s Lake and Falaise Lake cremation sites, and meadow samples, respectively. A one-way ANOVA test using pairwise multiple comparisons via the Student–Newman–Keuls method demonstrated that significantly fewer contaminants were isolated with PLET medium from PLR specimens than from samples of the other regions (P < 0·001). PLET cultures of PLR specimens were also less likely to show any microbial growth after 48 h of incubation. Eighty-six (54·8%) Parson’s Lake samples failed to yield any colonies, compared with three (1·8%) of Hook Lake samples, 11 (5·6%) of Falaise Lake carcass site specimens and none of the Falaise meadow samples.
Bacillus anthracis-like colonies from the PLET were picked and tested with blood agar plates and penicillin discs. Only 25 of the specimens yielded isolates that, like B. anthracis, were non-haemolytic and were inhibited to some degree by penicillin (range 7–40 mm zones of inhibition). None of the Hook Lake samples yielded non-haemolytic isolates that were inhibited by penicillin, while only three specimens each from PLR and the Falaise meadow, and 19 specimens from the Falaise cremation sites, exhibited colonies of this phenotype. None of the isolates from the three remaining Falaise meadow specimens formed mucoid colonies on bicarbonate plates after overnight incubation in 10% CO2. Isolates from the three remaining PLR specimens, and 11 of the 19 remaining Falaise cremation samples, developed heavily mucoid colonies on the bicarbonate plates that were consistent with B. anthracis. Giemsa-stained smears prepared from the bicarbonate cultures of 11 of the mucoid samples displayed rods with thick, enveloping, purple capsules compatible with B. anthracis. All of these 11 isolates were later confirmed to be B. anthracis via MLVA.
Location of positive B. anthracis isolates
The 11 B. anthracis positive specimens came from seven carcass disposal sites; three burial sites along PLR (50% of the sites in the region surveyed) and four (50%) cremation sites around Falaise Lake (Table 2). No positive samples were isolated from Hook Lake anthrax mounds or Falaise meadow specimens. However, the sample size of meadow specimens was small and they were collected from an area roughly 1 km away from the closest positive disposal site.
Table 2. Source and anthrax spore concentration of Bacillus anthracis positive environmental specimens
In six (54·5%) of the 11 positive specimens, B. anthracis was detected at the sensitivity limit of the extraction/PLET culture procedure, 2 spores g–1, represented by a single colony on the PLET medium. At five (71·4%) of the seven positive carcass disposal sites, B. anthracis was isolated from a single environmental specimen. High levels of anthrax spores were associated with a red fox (Vulpes vulpes) scat and soil specimens collected from within the bone bed of cremation sites.
Site 16 was right alongside PLR and was sampled in both 1992 and 1993 for a total of 47 specimens. The sole positive sample from this site, containing 800 anthrax spores g–1, was a red fox scat collected in 1992. The site was intact and no bison or other animal remains were observed in its immediate vicinity. Sites 22 and 23 were both located away from PLR, under the forest canopy. From 22 and 25 specimens collected from sites 22 and 23, respectively, only a single B. anthracis colony was isolated. The positive specimen from site 22 was collected at the very edge of the disturbed sand, while the positive specimen from site 23 was obtained from near the centre of the disturbed area.
At three of the four positive cremation sites at Falaise Lake, anthrax spores were closely associated with the charred bone bed. Sites 37 and 89 each yielded one positive sample, with approximately 70 and 1200 viable anthrax spores g–1, respectively, from within their bone beds. Site 41 yielded the most positive specimens with four, all of which were associated with the bone bed. As depicted in Fig. 2, the highest concentration of anthrax spores was found near the centre of the bone bed, and the concentration of viable anthrax spores appeared to decrease towards the periphery of the bed. Two positive samples representing 2 spores g–1 each were found at site 142. Both were located on the south side of the site, well away from the bone bed.
The theoretical sensitivity limit of the extraction and PLET culture procedure based on sample dilution was 2 spores g–1. However, this assumes that the extraction procedure was able to isolate every anthrax spore in a given sample and that the spores were completely germinable on the PLET medium. Both assumptions are unlikely. Previous studies have demonstrated that anthrax spores can form strong adhesions to soil particles and other solid matrices (Cole et al. 1984; Doyle et al. 1984; Bowen et al. 1996). While the flotation extraction method employed in this study was shown to be more effective at isolating anthrax spores than water from spiked soil samples, it never recovered 100% of the inoculated spores and recovery rates between soil types were variable (Dragon and Rennie 2001). Compared with culture on enriched media, spores of several B. anthracis strains have been found to have reduced germinability on PLET medium (McGetrick et al. 1982; Dragon and Rennie 2001). Therefore, the anthrax spore concentrations listed in Table 2 for positive environmental specimens can only be considered rough estimates of the true concentrations. However, relative to each other, the concentrations can still be used to identify high and low levels of anthrax spores in the environment.
Of the 588 environmental specimens collected, 11 (1·9%) were shown to contain viable B. anthracis spores. All 11 were directly associated with the disturbed area of carcass disposal sites in the PLR and Falaise Lake regions. Hook Lake mounds were between 14 and 30 years old when sampled, while PLR burial and Falaise cremation sites were less than two years old when samples were collected. The discrepancy in age allows a much greater opportunity for spores at the Hook Lake sites to have been either inactivated or dispersed by environmental forces.
Bacillus anthracis spores have been recovered from the faeces of black-backed jackals (Canis mesomelas) and spotted hyenas (Crocuta crocuta) collected from around carcass sites during active anthrax epidemics in Etosha National Park, Namibia, and Luangwa Valley, Zambia (Turnbull et al. 1989, 1991; Lindeque and Turnbull 1994). While carnivorous mammals have been suspected of intestinal carriage and dissemination of anthrax spores in northern Canada, the successful isolation of B. anthracis from the red fox scat is the first time it has been demonstrated bacteriologically.
The other two positive burial sites encountered in the PLR region had barely detectable levels of anthrax spores. It is difficult to interpret these findings. In several parts of the world, recurrent anthrax outbreaks have been associated with low-lying, alkaline soils rich in organic matter (Van Ness 1971; Hugh-Jones and Hussaini 1975). It was originally believed that these soil conditions influenced vegetative anthrax bacilli and allowed for cycles of germination, growth and resporulation resulting in an overall increase in spore concentration (Van Ness 1971). However, vegetative B. anthracis have very specific nutrient and physiological requirements and are unlikely to survive outside a host. Instead, the specific soil factors linked to endemic areas may reflect environmental conditions that aid in maintaining anthrax spores at the site and prolonging their viability (Dragon and Rennie 1995).
Burial site specimens from PLR contained significantly fewer endospores from Bacillus species able to grow on PLET medium than the other two regions surveyed. It is possible that the sandy soil was unsuitable for holding spores, and any present are quickly removed via water action. The acidity of the soil may have reduced the long-term viability of spores formed at the site (Dragon and Rennie 1995). It is also possible that the PLR habitat supported a large number of Bacillus species whose spores were unable to germinate and grow on PLET medium, and that the perceived difference between the regions is an artifact introduced by the isolation protocol.
The majority of positive B. anthracis spore isolates from Falaise Lake were associated with the bone beds of cremation sites. The beds comprised charred shards of vertebrae and ribs, suggesting that the bulk of the carcass had rested above these beds prior to burning. The beds often contained thick mats of bison cape hair and masses of maggot casings under the bones. For cremation, wood and coal were piled on top and beside the body, dosed with aviation fuel and lit. The bulk of the body apparently insulated the ground underneath from the intense heat and flame. The higher concentration of anthrax spores in the centre of the bone bed at site 41 compared with the edges is most likely the result of decreased protection of spores from the destructive effects of fire nearer the edges of the bed. These results indicate the need, when disposing of anthrax-infected carcasses by cremation, to return to the sites and re-burn any bone beds containing unburned hair.
Overall, the level of anthrax spore contamination in the three endemic regions surveyed appeared low. All positive samples were obtained in the immediate vicinity of carcass disposal sites within the area disturbed by clean-up operations. Both burial and cremation appear equally effective at removing anthrax spores from the immediate environment. A study in the Falaise Lake region of wallow and meadow samples not associated with known carcass locations failed to detect any B. anthracis spores, although these areas were not surveyed as intensively as the disposal sites and the results are not conclusive. However, high concentrations of anthrax spores in the environment of northern Canada appear limited to scavenger faecal matter and anthrax carcass sites. Ecological studies of anthrax spores with PLET medium in African endemic regions have also demonstrated an association of high spore concentrations with scavenger faeces and soil around carcass sites, and the bacterium was only rarely found in environmental specimens not associated with carcasses (Turnbull et al. 1989, 1991; Lindeque and Turnbull 1994).
While sporadic outbreaks have shown that concentrations of anthrax spores high enough to cause disease in bison are obviously present somewhere in the endemic regions, it is unknown if the levels of spores found here represent a sufficient dosage to cause disease. Neither the infectious dosage of anthrax spores for bison, nor how the animal contracts the disease, is known. Bison may acquire the disease through inhalation of aerosolized spores during wallowing, or may ingest lethal levels of spores while grazing. Further complicating the matter is the possibility of seasonal transient stresses, such as the rut, which may compromise bison immunologically, thereby reducing the infectious dosage required (Gainer and Saunders 1989).
Because of the remote area involved and associated logistic constraints, the carcasses of bison dying during the 2000 WBNP anthrax outbreak were inventoried but not treated. The anthrax spore detection system described here will be used to survey these sites, and both studies will be compared to assess the efficacy of carcass disposal methods and to aid federal and territorial governments in improving their response plans to the disease.
This research was supported in part by the Department of Education, Culture and Employment, Government of the NT. The authors thank Drs C.C. Gates and J. Nishi, Troy Ellsworth and Dean Robertson of the Department of Resources, Wildlife and Economic Development, Government of the NT, George Mercer of WBNP, Daniel Gates of Fort Smith, Dr S. Tessaro and Greg Tiffin of ADRI, Lethbridge, Drs M. Hugh-Jones and P. Corker of the Department of Epidemiology and Community Health, School of Veterinary Medicine, Louisiana State University and Dr P. Keim of the Department of Biology, Northern Arizona University.