Discussion
The anthrax PA, in addition to the oedema and lethal factors (EF and LF), make up the tripartite anthrax toxin complex (Schwartz 2009). These anthrax toxins are encoded by the anthrax-specific virulence plasmid, pXO1, with PA combining with LF to form the lethal toxin (LT). To date, no cross reacting antigens to anthrax PA are known and available data indicate anti-PA antibodies as being the most diagnostically stable (Leppla, Robbins, Schneerson, & Shiloach, 2002; Turnbull et al., 1992). However, PA is antigenic complex and can elicit protective and non-protective antibodies. Abboud & Casadevall (2008) investigated the relationship between PA structure and antigenicity but could only conclude the genetic background of the host determines the relative efficacy of the antitoxin response (Abboud & Casadevall, 2008) and , thus the toxin reutilizing antibodies were determine on a random sample to verify the use of anti-PA antibodies in this study.
Both ELISA and toxin neutralization (TNA) assays can be used to detect antibodies to PA antigen in sera of animals exposed to B. anthracis . We report for the first time the dual existence of anthrax PA-specific and lethal toxin neutralizing antibodies in naturally exposed African lions. The TNA is a functional assay that measures the neutralizing activity of sub-sets of PA-specific antibodies against the cytotoxic effect of the lethal toxin complex formed by PA in association with LF. Though technically demanding, TNA confirms previous exposure toB. anthracis in animals and is species-independent (Ngundi, Meade, Lin, Tang, & Burns, 2010; Omland et al., 2008; Savransky et al., 2017). On the other hand the ELISA assay detects antibodies to PA antigen in sera of infected individuals and is less demanding but well correlated with the TNA (Arciniega & DomĂ­nguez-Castillo, 2011; Ndumnego, Crafford, Beyer, & van Heerden, 2013; Parreiras, Sirota, Wagner, Menzies, & Arciniega, 2009). Consequently, the use of ELISA immunoassays with PA as the sole antigen in diagnostic tests has been validated in humans (Ghosh et al., 2015; Quinn et al ., 2002; Semenova et al ., 2012), horses (Caldwell, Hathcock, & Brock, 2017) and goats (Ndumnego et al ., 2013). In this study, we validated the use of the ELISA assay for detecting anthrax anti-PA antibodies in exposed African lions by using the TNA assay. Indeed, earlier researchers (Bagamian, Alexander, Hadfield, & Blackburn, 2013; Hampson et al ., 2011; Switzer et al ., 2016; Turnbullet al ., 1992), have used the ELISA-PA assay to monitor B. anthracis exposure in lions among other carnivore species. Hence, the ELISA-PA immunoassay for the serological survey of B. anthracisexposure in African lions was adopted for use in this study.
Further, on the use of ELISA-PA assay in goats, Ndumnego et al . (2013) compared and observed that the use of skimmed milk powder gives a lower background reading compared to the use of foetal calf serum. While there may be the risk of lion IgG detecting milk or any other ruminant proteins leading to false positives, the presence of lethal toxin neutralizing antibodies in sero-positive sera in toxin neutralization test negates this. The presence of anti-PA antibodies in animals indicates that non-lethal systemic infection may have taken place (Hugh-Jones & Blackburn, 2009; Ndumnego et al ., 2013; Turnbullet al ., 1992). Therefore based on anti-PA antigen antibodies serological assay in African lions, there were indications that B. anthracis pathogen was present with a wide distribution in Zimbabwe.
In the case of HNP where anthrax outbreaks have not been reported in over 20 years to 2018, this represents the first confirmation of the presence and wide exposure of B. anthracis to lions in the park. There is a possibility that some lions from HNP could have been exposed to anthrax from consumption of livestock carcasses as outbreaks of anthrax were reported in adjacent communal lands of Tsholotsho southeastern to HNP (Fig. 2) (Mukarati et al ., 2018).Indeed, predation of livestock by lions from the park is an established human-wildlife conflict issue in the area (Loveridge et al ., 2017). However, other seropositive lions were far removed from this community at more than 90 km from the nearest park boundary. This distance was well outside the lions’ home range of about 20 km radius (Loveridge et al ., 2009) indicating unlikely influence of diseases between the lions and livestock. It is most likely that such seropositive lions were exposed to other sources of B. anthracisin HNP. Due to the limitation on samples available, the distribution of the positive lions was biased towards the northern half of the park. A uniform distribution of positive lion samples would have provided more insightful trends.
Although this was not strictly a cross-sectional survey in that the lions were sampled over 20 years, it nonetheless provides useful information. That lions sampled over this period were positive forB. anthracis in the absence of reported disease outbreaks suggests either occurrence of undetected outbreaks of the disease or sublethal infections in wildlife in HNP (Hugh-Jones and Blackburn, 2009; Cizauskas et al., 2014). Thus, the hitherto general belief that HNP was free of anthrax based on no reported cases is uncertain. The presence of antibodies to anthrax in lions in Mana Pools NP, Save Valley Conservancy and Malilangwe Wildlife Reserve indicates that B. anthracis was already circulating in the area long before the disease outbreaks occurred (Mukarati et al., 2018; Clegg et al.,2007; OIE, 1997). However, there was no correlation between anthrax outbreaks and seropositivity in lions from this dataset. Improved surveillance and additional studies on possible environmental and soil geochemical factors possibly influencing B. anthracisdistribution in protected areas are necessary (Hugh-Jones and Blackburn, 2009; Griffin et al., 2014).
There were no known anthrax outbreaks that could account for exposure and seroconversion in captive lions in this study. According to de Vos and Turnbull (2004), anthrax outbreaks affecting wildlife in captivity are limited to consumption of infected meat. Carnivores on small game parks are managed essentially as in zoological gardens and thus are presumably similarly exposed to anthrax. An anthrax outbreak occurred in lions and cheetahs in 1997 at Lion and Cheetah Park, Harare, when they were fed infected cattle meat donated by a farm (OIE, 1997) . On the other hand, wild carnivores in captivity maybe exposed to B. anthracis from infected meat without necessarily developing clinical disease either due to their relative resistance or exposure to sublethal anthrax doses (Beyer and Turnbull, 2009; Hugh-Jones and Blackburn, 2009; Cizauskas et al ., 2014).
A pertinent question which arises is the specificity of the serological assay used in this survey, given that there are few reports of atypicalB. cereus strains causing similar disease in humans and wild primates (Marston et al., 2016). Rare cases of anthrax-like illness in humans and wild chimpanzees caused by a B. cereusstrain possessing the anthrax toxin genes have been reported in the US and West Africa (Antonation et al ., 2016; Hoffmaster et al ., 2006; Klee et al ., 2006). However, to date there are no documented cases of outbreaks caused by this strain of B. cereusin wild or domestic ruminants, or the carnivores that predate on these animals. While not ruling out the possible exposure to environmentalB. cereus in grazing animals, the presence of the rare toxin-producing B. cereus strains have not been reported in Southern Africa (Romero-alvarez et al., 2020). Thus, the seropositivity of lions in this study is assumed to be attributable to exposure toB. anthracis strains.
The finding of no significant difference in seropositivity between lions located in currently recognized high- and low-risk zones for anthrax (p = 0.852, Table 1) tally with earlier findings in domestic dogs (Mukaratiet al., 2018). A much wider range of anthrax endemic areas in Zimbabwe can be hypothesized similar to other endemic regions of the world. This suggests that the categorization of areas in Zimbabwe into high- and low-risk zones may not represent the true status of anthrax presence across the country. This needs review based on improved surveillance and epizootiological investigations. Anthrax serology in resident wild and/or domestic carnivores could serve as sentinel and indicator of anthrax circulation in given areas and thus can be useful epidemiological tools.
The widespread presence of anthrax antibodies in lions in protected areas, irrespective of absence of reported disease outbreaks in wild or domestic ungulates or humans, confirms a much larger circulation ofB. anthracis in Zimbabwe. These results raise new questions on the epidemiology of anthrax in endemic regions. There is need to investigate local factors that could be associated with anthrax outbreaks apart from the presence of the pathogen. On the other hand, there could be smaller outbreaks of anthrax occurring but going unnoticed in HNP as has been noted elsewhere (Cizauskas, Bellan, Turner, Vance, & Getz, 2014). Overall, improved surveillance of anthrax in all animals could shed more light on whether outbreaks were indeed taking place but being missed because of inadequate surveillance and may also give pointers on risk factors. Further, the spatial distribution of lions tested for anthrax anti-PA antibodies in this study in the northern half of HNP meant that insightful trends in the whole of the park could not be determined. A more uniform distribution of lion samples tested for anthrax anti-PA antibodies across the park would be helpful in establishing likely risk areas for the disease outbreaks.