In South Africa, the largest proportion of the African wild dog (Lycaon pictus) population resides in regions where buffaloes have a high prevalence of Mycobacterium bovis, the causative agent of bovine tuberculosis (bTB). Recent reports of deaths of wild dogs associated with bTB have raised concerns regarding the threat this disease might pose for this species. In order to understand the potential impact of the disease in wild dogs, diagnostic tools are required to identify infected individuals. The interferon gamma (IFN-γ) release assay (IGRA) is commonly used for tuberculosis (TB) screening of humans, cattle, and other species, and the aim of this study was to develop an IGRA for wild dogs to detect immune sensitization. Blood was collected from immobilized wild dogs from the Ann van Dyk Cheetah Centre (AvDCC; n=9) and Kruger National Park (KNP; n=31). Heparinized whole blood was incubated overnight in QuantiFERON®-TB Gold (QFT) blood collection tubes and with selected mitogens, after which the plasma fraction was harvested. Three canine IFN-γ enzymelinked immunosorbent assays (ELISAs) were compared for detection of wild dog IFN-γ in plasma and the R&D Quantikine canine IFN-γ ELISA was selected for measurement of M. bovis-specific IFN-γ release in plasma samples. An IGRA result was calculated as the concentration in plasma derived from the QFT TB Antigen tubes minus that in the QFT Nil tube. An IGRA cut-off value was calculated using the IGRA results of M. bovis-unexposed individuals from AvDCC. Using this cut-off value, 74% (23/31) of M. bovis-exposed KNP wild dogs were IGRA positive, indicating immune sensitization to TB antigens in these animals. Three M. bovis culture-positive wild dogs from KNP had IFN-γ concentrations between 758 and 1,445 pg/mL, supporting this interpretation. This warrants further investigation into the prevalence of M. bovis infection in the KNP population.
The African wild dog (Lycaon pictus) is listed as endangered and declining on the International Union for Conservation of Nature Red List (Woodroffe and Sillero-Zubiri 2012), and is South Africa's most endangered carnivore (Davies-Mostert et al. 2016). Globally, wild dog populations are small and fragmented, and are susceptible to local extinctions due to infectious diseases and from other threats that this species faces (Woodroffe and Ginsburg 1997). In South Africa, the largest wild dog population occurs in Kruger National Park (KNP; Davies-Mostert et al. 2016) where bovine tuberculosis (bTB) is endemic (Renwick et al. 2007). Bovine tuberculosis is a chronic disease, caused by infection with Mycobacterium bovis, a member of the pathogenic M. tuberculosis complex (MTBC). Mycobacterium bovis causes morbidities and mortalities in numerous wildlife species, including African buffalo (Syncerus caffer), greater kudu (Trag-elaphus strepsiceros), and African lions (Panthera leo; Renwick et al. 2007) and recently, cases have been identified in wild dogs in South Africa (M.A.M. unpubl. data).
Although wild dogs are susceptible to infection with M. bovis, and mortalities associated with bTB have been observed in KNP, Hluhluwe-iMfolozi Park, and Mkuze Game Reserve, little is known about the extent of infection in exposed populations. In order to improve our understanding of the epidemiology of bTB in wild dogs, and to mitigate the spread of the disease through translocation of infected animals, diagnostic tools are required to accurately identify infected individuals. Currently, there are no validated tests for diagnosis of bTB in this species; however, assays used for other species might be adapted for wild dogs.
Infection with M. bovis is most commonly diagnosed by detecting a host's cell-mediated immunity toward pathogen-specific antigens (Welsh et al. 2005). This can be done using an interferon-gamma (IFN-γ) release assay (IGRA), which includes antigenic stimulation of blood lymphocytes, followed by measurement of the release of IFN-γ. Interferon-γ is a commonly used cytokine biomarker of cellmediated immunity for the diagnosis of tuberculosis (TB) and is considered vital to a host's protection against intracellular pathogens such as M. bovis (Ottenhoff et al. 2005).
The use of MTBC-specific antigens, such as 6 kDa early secretory antigen (ESAT-6) or 10 kDa culture filtrate antigen (CFP-10) can be used to improve test specificity because these antigens are absent from the majority of nontuberculous mycobacteria (Cellestis 2006; Parsons et al. 2011). The QuantiFERON®-TB Gold (QFT) stimulation platform, which incorporates these specific peptides into commercially available tubes, is commonly used for detection of MTBC infection in humans (Tsiouris et al. 2006), and has been adapted for use in spotted hyena (Crocuta crocuta), African lion, and African buffalo (Parsons et al. 2011; Higgitt et al. 2017; Olivier et al. 2017). Following stimulation, the release of antigen-specific IFN-γ by lymphocytes can be measured using an enzyme-linked immunosorbent assay (ELISA).
Interferon-γ release assays have been developed for diagnosis of MTBC infections in numerous wildlife species, including European badgers (Meles meles), nonhuman primates, white rhinoceros (Ceratotherium simum), and African buffaloes (Garcia et al. 2004; Dalley et al. 2008; Parsons et al. 2011; Morar et al. 2013), and are routinely used for diagnosis in humans and cattle (Cellestis 2006; Gormley et al. 2006). We aimed to develop an IGRA for the detection of M. bovis infection in wild dogs, using the QFT stimulation platform and a commercial, canine-specific IFN-γ ELISA.
MATERIALS AND METHODS
Captive wild dogs from the Ann van Dyk Cheetah Centre (AvDCC), Brits, South Africa (25°40′25″S, 27°55′25″E) and free-ranging wild dogs from the KNP, South Africa (23°59′S, 31°33′E) were opportunistically sampled for this study. Animals from AvDCC were presumed M. bovis-unexposed and nine serologically negative animals (DPP® VetTB serological test, Chembio Diagnostic Systems, Inc., Medford, New York, USA) were used to calculate the cut-off value of the IGRA. Animals from KNP (n=31) were considered to be at risk of infection with M. bovis because bTB is endemic in this region. Wild dogs were chemically immobilized (Kock and Burroughs 2012) and whole blood was collected by venipuncture of the jugular or femoral veins into a sodium heparin collection tube (Becton Dickinson, Franklin Lakes, New Jersey, USA). Oropharyngeal swabs (n=29) and bronchoalveolar lavage (BAL) fluids (n=7) were collected antemortem from selected individuals. These samples were processed for mycobacterial culture in the Biosafety Level 3 Facility at the Department of Biomedical Sciences, Stellenbosch University (Cape Town, South Africa) using the BACTECTM Mycobacterial Growth Indicator Tube (MGITTM) system (Becton Dickinson) as previously described (Goosen et al. 2014). Oropharyngeal swabs and BAL fluid were not homogenized prior to decontamination with 5 mL MycoPrepTM solution (Becton Dickinson). Ethical approval was granted by the University of Stellenbosch Animal Care and Use Committee (SU-ACUD16-00076) and Section 20 approvals were granted by the Department of Agriculture, Forestry and Fisheries, South Africa (12/11/1/7/2).
Whole blood stimulation
One milliliter of heparinized whole blood was transferred into each of the QFT Nil and TB Antigen tubes (Qiagen, Venlo, the Netherlands). A mitogen tube was included as a positive control for the assay. Initially, 1 mL of heparinized whole blood was stimulated with pokeweed mitogen (PWM; n=29) at a final concentration of 10 µg/ mL. During the study, it was observed that the mitogen tube supplied with the second generation QFT assay (QuantiFERON®-TB Gold Plus; QFT Plus) provided a satisfactory IFN-γ response, and theQFT PlusMitogen tube was used in place of the PWMtube for some animals (n=11). All tubes were incubated at 37 C and under 5% carbon dioxide for 24 h, after which samples were centrifuged at 3,000 × G for 6 min. The plasma fraction was harvested and stored at_80 C until analyzed.
Selection of an optimal canine IFN-c ELISA
Three commercial canine IFN-γ ELISA kits were screened as candidate assays for detection of wild dog IFN-γ: Quantikine Canine IFN-γ ELISA (R&D Systems, Inc., Minneapolis, Minnesota, USA), Canine IFNG/Interferon Gamma ELISA kit (Sigma Aldrich, St. Louis, Missouri, USA), and RayBio® Canine IFN-gamma ELISA kit (RayBiotech, Inc., Norcross, Georgia, USA). Plasma from PWM-stimulated whole blood from four randomly selected wild dogs was pooled, and for each ELISA an aliquot was diluted 1:5 in kitspecific dilution buffer. Two-fold serial dilutions of this pooled sample were assayed in duplicate wells. Additionally, the respective sample dilution buffer for each assay was assayed in duplicate as a negative control. The ELISAs were performed according to each manufacturer's instructions. The optical density (OD) of each well was measured at wavelengths of 450 nm (OD450) and 630 nm (OD630) using an LT-4000 Microplate Reader (Labtech International Ltd., Heathfield, UK). Sample values were calculated as OD450–OD630 to account for light absorbance of other materials, such as polystyrene of the plate. The OD result was calculated as the mean sample value for duplicate samples minus the mean sample value of the negative control. The optimal kit was selected as the ELISA that returned the highest OD result over the dilution range of PWM-stimulated plasma.
Detection of wild dog M. bovis-specific IFN-c Release
The ELISA kit selected as the most sensitive for detection of wild dog IFN-γ was used to test plasma from the QFT Nil, QFT TB Antigen, and mitogen tubes. Briefly, each ELISA included duplicates of serial dilutions of recombinant canine IFN-γ (rIFN-γ), ranging from 15.6–4,000 pg/mL for the standard curve; negative control wells of assay diluent to which no sample was added; and a canine IFN-γ positive control (rIFN-γ of known concentration), which was supplied with the kit. Test plasma harvested from QFT Nil and QFT TB Antigen tubes were assayed in duplicate for each wild dog, and a single replicate of mitogenstimulated plasma was included as a positive control for each stimulation. Absorbance was read as OD at wavelengths of 450 nm and 650 nm (OD650) using an iMarkTM Microplate Absorbance Reader (Bio-Rad) and OD650 was subtracted from OD450 to obtain the sample value. The OD result was calculated as previously described, and the IFN-γ concentration (pg/mL) for each sample was derived from the standard curve. An IGRA result was calculated as the IFN-γ concentration measured in plasma harvested from the QFT TB Antigen tube minus that in plasma harvested from QFT Nil tube.
To identify the linear range of the optimal ELISA, the mean OD results over the standard dilution from 11 ELISAs were plotted to identify the dilution range with a linear regression (r2) nearing 1.00.Wild dog IFN-γ concentrations in the QFT Nil, QFT TB Antigen and mitogen-stimulated plasma from M. bovis-unexposed and -exposed cohorts were compared using a Kruskal-Wallis test with Dunn's Multiple Comparisons test. Comparison of IFN-γ release in response to the two mitogens used was made using a Kruskal-Wallis test and Dunn'sMultiple Comparisons test.Median IGRA results were calculated, along with the interquartile range (IQR), and results compared between M. bovis-unexposed and -exposed cohorts using a one-tailed Mann-Whitney U-test. Using IGRA results from the unexposed cohort, a diagnostic cut-off value was calculated as the mean assay result+2 SDs, as previously described (Shin et al. 2013). The IGRA results of exposed wild dogs that were greater than the cut-off value were considered IGRA positive. Interferon-γ concentrations and IGRA results were log-transformed. All data analyses were done using GraphPad Prism 5 software (GraphPad Software, Inc., San Diego, California, USA) and results with P< 0.05 were considered statistically significant.
Wild dog IFN-γ was detected using the three commercial canine IFN-γ ELISAs. Of the ELISAs screened, the R&D Quantikine canine IFN-γ ELISA returned the highest ODs for all dilutions of pooled PWM-stimulated plasma (Fig. 1). Hereafter, the IFN-γ concentrations in plasma harvested from QFT Nil, QFT TB Antigen, and mitogen tubes from 9 AvDCC DPP-negative and 31 randomly selected KNP wild dogs were measured with the R&D Quantikine canine IFN-γ ELISA. The linear range of this ELISA was between 15.6 and 1,000 pg/mL (r2=0.99).
The concentrations of IFN-γ in whole blood samples that had been stimulated with PWM or in the QFT Plus Mitogen tubes were compared (Fig. 2). The median IFN-γ concentrations were 355 pg/mL (IQR: 178–771 pg/mL, n=29) in PWM-stimulated samples and 414 pg/mL (IQR: 235–1,119 pg/mL, n=11) in QFT Plus Mitogen-stimulated samples. There was no significant difference (P=0.332) between PWM and QFT Plus Mitogen-stimulated samples.
The concentrations of IFN-γ in plasma harvested from the QFT Nil, QFT TB Antigen, and mitogen tubes were determined for all wild dogs (Fig. 3). For the unexposed cohort, median IFN-γ concentrations were 34 pg/mL (IQR: 29–53 pg/mL) in unstimulated samples, 40 pg/mL (IQR: 26–66 pg/mL) in QFT TB Antigen-stimulated samples, and 398 pg/mL (IQR: 209–926 pg/mL) in mitogenstimulated samples. The median IFN-γ concentration in mitogen-stimulated samples was significantly (P<0.001) greater than median concentrations in the unstimulated and QFT TB Antigen-stimulated samples. No significant difference (P>0.05) was found between median concentrations in the unstimulated and QFT TB Antigen-stimulated samples. For the exposed cohort, median IFN-γ concentrations were 25 pg/mL (IQR: 6–59 pg/mL) in unstimulated samples, 265 pg/mL (IQR: 134–1,026 pg/mL) in QFT TB Antigen-stimulated samples, and 444 pg/mL (IQR: 187–799 pg/ mL) in mitogen-stimulated samples. Significantly (P<0.001) greater median IFN-γ concentrations were found in QFT TB Antigenstimulated and mitogen-stimulated samples compared to unstimulated samples. The median IFN-γ concentration of QFT TB Antigen-stimulated samples from the unexposed cohorts was significantly (P<0.001) lower than that of the exposed cohort.
The antigen-specific IGRA results were determined for the M. bovis-unexposed and -exposed groups (Fig. 4). Using the unexposed cohort (median: 17 pg/mL; IQR: 0–33 pg/ mL), a cut-off value of 51 pg/mL was calculated and, based on this cut-off value, 74% (23/31) of exposed KNP wild dogs (median: 260 pg/mL; IQR: 40–985 pg/mL) were IGRA positive. Of the 23 IGRA-positive wild dogs, three animals were confirmed to be positive for M. bovis infection by antemortem mycobacterial culture of oropharyngeal swabs or BAL samples (IGRA results: 758 pg/mL, 911 pg/mL, and 1,445 pg/mL).
In our study, wild dog IFN-γ release in whole blood was detected using the QFT stimulation platform and a commercial canine-specific IFN-γ ELISA. Stimulation of wild dog whole blood with PWM or in the QFT Plus Mitogen tube was suitable as a positive control for this assay. Significantly higher IFN-γ concentrations were detected in QFT TB Antigen-stimulated samples from M. bovis-exposed wild dogs compared to -unexposed animals. Using the IGRA cut-off value calculated from the unexposed cohort, the majority of exposed wild dogs (74%, 23/31) were considered IGRA positive. Additionally, M. bovis infection was confirmed in the exposed population by antemortem culture.
The R&D Quantikine canine IFN-γ ELISA showed the greatest sensitivity for detecting wild dog IFN-γ in mitogen-stimulated samples. This ELISA has previously been used to measure IFN-γ responses to ESAT-6/CFP-10 in domestic dogs in an IGRA that can distinguish M. tuberculosis-exposed from -unexposed individuals (Parsons et al. 2012). This study further confirms the utility of this ELISA for use in canid species in the context of TB diagnosis.
A significant release of IFN-γ was measured in plasma from wild dog whole blood stimulated with PWM and in QFT Plus Mitogen tubes. The inclusion of a mitogenstimulated positive control is essential to ensure viability and function of leukocytes in whole blood during in vitro stimulation. Pokeweed mitogen has been shown to elicit IFN-γ release in a variety of species, including elephants (Angkawanish et al. 2013) and domestic cattle (Bos taurus; Aranaz et al. 2006), and is included as a positive control for the Bovigam assay (Bass et al. 2013). Phytohaemagglutinin, the positive control mitogen of the second generation QFT Plus system, is reported to be a strong stimulant of lymphocytes in horses (Equus caballus) and pigs (Sus scrofa; Peters and Veerkamp 1982); however, poor responses to phytohaemagglutinin have been reported for chacma baboons (Papio ursinus; Parsons et al. 2009) and African buffaloes (Parsons et al. 2011). In wild dogs, a strong IFN-γ response was observed following incubation in QFT Plus Mitogen tube. Use of either PWM or the second generation QFT Plus Mitogen tube should be used as a positive control in future studies using wild dog whole blood.
The QFT system was successfully used for in vitro stimulation of wild dog whole blood in this study. This platform has previously been used for detection of MTBC-specific immune responses in spotted hyenas, African lions, and African buffaloes (Parsons et al. 2010; Higgitt et al. 2017; Olivier et al. 2017). Advantages of the QFT stimulation platform include relative ease of use (including self-contained sterile incubation), suitability for in-field application, and room temperature storage for short durations. This system also utilizes qualitycontrolled reagents, which improves standardization for diagnostic use. In addition, the use of ESAT-6 and CFP-10 antigens has been shown to improve the specificity of the IGRA in cattle, compared to PPD-based assays (Schiller et al. 2011).
An assay-specific cut-off value was calculated to further evaluate the potential use of the wild dog IGRA as a diagnostic test for M. bovis immune sensitization. An IGRA cut-off value of 51 pg/mL was determined using the test results from the unexposed cohort. Although none of the unexposed wild dogs had an IFN-γ test result greater than the cutoff and the median value was low (17 pg/mL), the median response in the exposed wild dogs was higher (260 pg/mL), supporting the hypothesis that antigen-specific IFN-γ was a good biomarker for M. bovis infection in this species.
Evidence of infection with M. bovis in the KNP wild dogs by antemortem culture confirmed the exposure of this population. Coupled with the high IGRA results reported for the three culture-confirmed animals, antemortem detection of M. bovis indicated that an antigen-specific immune response could be detected in infected animals. This result further highlighted the potential threat M. bovis can pose in wild dog populations, particularly in high-burden settings such as KNP.
Use of a blood-based assay, such as the QFT IGRA, offers numerous advantages for the diagnosis of M. bovis infection in an endangered species like the wild dog. The collection of a whole-blood sample is less invasive than methods that require repeated immobilization, such as the tuberculin skin test. In vitro assays allow repeated testing of an individual without potentially eliciting a boosting effect, as can occur with the tuberculin skin test (Coad et al. 2010). However, the IGRA is an indirect test and, although immune sensitization can be detected, further research is needed to determine test performance in wild dogs with known infection status.
We showed that M. bovis antigen-specific wild dog IFN-γ can be detected using the QFT stimulation platform and the R&D Quantikine canine IFN-γ ELISA. This IGRA could prove useful for the detection of M. bovis infection in this species, using convenient and commercially-available components. The identification of IGRA-positive wild dogs from the KNP, a bTB endemic area, warrants further investigation into the prevalence in this population. It would be beneficial if future studies included a greater number of known unexposed wild dogs to increase confidence in the calculated IGRA cut-off value, as well as wild dogs of known infection status to evaluate test performance.
We thank the Ann van Dyk Cheetah Centre, the veterinary technicians of the Kruger National Park State Veterinary Office in Skukuza, and the veterinary technicians of the South African National Parks Veterinary Wildlife Services for assistance with wild dog sampling. This work was supported by the South African Medical Research Council and the National Research Foundation of South Africa (86949). We declare that the content of this publication is the sole responsibility of the authors and that the funding agencies were not involved in any of the research design, implementation, analyses, or writing. Therefore, the publication does not necessarily represent the views of the funders.