Study area
The cattle samples were obtained from animals that were bTB tested during a cross-sectional study that was conducted at 4 dip tanks (Nibela, Nkomo, Mpempe, Masakeni) in Big 5 False Bay Municipality, in uMkhanyakude district, northern KwaZulu-Natal province, South Africa. Bovine TB testing of cattle was previously carried out in September 2016 and March 2017 as part of a One Health investigation into the epidemiology of bTB at the wildlife/livestock/human interface (32). Additional cattle samples were collected from farms and other dip tanks in the province during bTB outbreaks. Wildlife samples were collected from the surrounding game reserves that include HiP, Munyawana and Mkhuze. The map of the study area has been previously described by (Sichewo et al., (2019) (32).
The study area is defined as a wildlife/livestock/human interface due to the communal farmland being bordered by game farms or reserves where M. bovis infection has been diagnosed in African buffalo and other wildlife species (42).The farmers in this community depend on subsistence agriculture and livestock as their main source of income (43). Uncontrolled movement of livestock into/near game reserves during the dry seasons or periods of extensive drought is a common practice in the area.
Sample collection and processing
Milk, nasal and tissue samples from cattle
In June and July 2017, 30 milk samples were collected from interferon gamma positive animals that belonged to the 4 dip tanks were bTB testing was previously carried out in September 2016 and March 2017 (32). An average of 25 ml of milk was collected into 50 ml centrifuge tubes from all lactating animals. A total of 99 nasal swabs were collected from the same bTB infected herds, (32) through random sampling using 50 cm handmade swabs that were placed into phosphate buffer as the transport medium. The milk samples were frozen, and the nasal swabs were placed at 4oC and transported to the University of Pretoria, Department of Veterinary of Tropical Diseases in a cold chain.
Routine tissue sample submission was carried out by the state veterinary officials between 2002 and 2013 from i) tuberculin skin test and gamma interferon test positive cattle from farms in KZN and ii) suspect lesions detected in cattle during routine slaughter at abattoirs. These were processed following standard operating procedures at the Onderstepoort Veterinary laboratories as outlined by Hlokwe (2014) (18). In 2017, following slaughter of three test positive cows, one cow from each of the three dip tanks Nibela, Nkomo and Mpempe, appropriate samples were collected. Either the entire lymph node or approximately 5 -10 g of tissue were collected from the head, thoracic and mediastinal lymph nodes and, where relevant, from bTB like lesions observed in other organs such as the lungs, liver, spleen, kidney and mammary glands.
Wildlife tissue samples
During the annual bTB management culling program in the HiP, tissue samples were collected from skin test or gamma interferon positive buffalo between 2002 and 2017. In addition, samples were collected from other wild animals such as the lion, baboon and warthog in game reserves in KZN province that included HiP, Munyawana and Mkhuze either during i) TB passive surveillance of all wild animals found dead or ii) culling of skin test or gamma interferon positive wildlife. The samples included submandibular, retropharyngeal, tracheobronchial, mediastinal, mesenteric lymph nodes as well as sections of tonsils, lungs and any tissues with granulomatous lesions.
All the tissue samples that were collected from cattle and wildlife were packaged into zip lock bags and transported frozen to the University of Pretoria-Department of Veterinary Tropical Diseases and Onderstepoort Veterinary Research BSL2+ laboratory for mycobacterial culture.
Mycobacterium bovis culture and identification
Raw milk was decontaminated using 1% cetylpyridinium chloride (CPC) and 2% NaCl as previously described by Michel, 2015 (43).The sediment from the centrifugation was inoculated onto Löwenstein-Jensen (LJ) media with pyruvate and an antibiotic cocktail consisting of polymyxin B (200 IU/ml), amphotericin B (10 µg/ml) carbenicillin (100 µg/ml) and trimethoprim (10 µg/ml) (NHLS, South Africa) and incubated at 37o C for 10 weeks. Decontamination of nasal swabs was done using the modified Petroff method (2% HCl), followed by centrifugation and the sediment treated with amphotericin B (50 µg/ml). The solution was inoculated onto LJ media with pyruvate and the above antibiotics and incubated at 37o C for 10 weeks with weekly monitoring.
The animal tissue samples were processed according to the method previously described by Alexander et al. (2002) (44). Briefly, tissues samples were decontaminated using two methods with a final concentration of 2% NaOH and 1% HCl. These were inoculated on four slants of LJ media supplemented with pyruvate and incubated at 37o C for up to 10 weeks. The Ziehl-Neelsen (ZN) staining and culture characteristics (growth rate and colony morphology) were used to identify M. bovis isolates from tissue, nasal and milk cultures. DNA extraction was performed using the boiling method whereby the isolates were suspended in100 μl of sterile distilled water and heated to 95°C for 25 minutes using a heating block. Thereafter the DNA was stored at -20oC and used in all the subsequent PCR based reactions (39). M. bovis was confirmed using PCR that is performed using primers that target the Region of Difference (RD), RD4 and RD9 for deletion analysis as previously detailed by Warren, 2006 (45). The distribution of the isolates according to animal species, location and sample type are as shown in Table 1.
Genotyping of Mycobacterium bovis isolates
Spoligotyping
Spoligotyping was applied to isolates that were confirmed as M. bovis using culture for further differentiation according to the standard protocol described by Kamerbeek 1997 (46). Briefly, a PCR reaction was carried out to amplify the spacer sequence of the DR locus using primers DRa (biotinylated) and DRb. The spacer sequences were detected by hybridization of the biotin labelled amplified products using ECL against 43 spacer oligonucleotides covalently linked to a membrane. A commercially available kit was used, and the procedure was conducted according to the manufacturer’s instructions (SPOLIGOTB, Mapmygenome, India). A specific pattern of hybridization signals was generated that represented the absence or presence of the spacer sequences presented in a binary code that was translated into a spoligotype octal code according to the established algorithm. The spoligotyping pattern were named following the nomenclature of the international M. bovis spoligotype database (www.Mbovis.org).
Variable number of tandem repeats
The M. bovis isolates were also genotyped by PCR amplification using a panel of 13 VNTR markers previously established for South African isolates by Hlokwe, van Helden and Michel, 2013 (39). These include 4 ETRs loci (A, B, C, E), 3 MIRUs loci (16,23,26), QUBs loci (11a, b,18,26) and 2 M. tubs (12,21). The PCR reactions for each marker were carried out separately using specific primer sequences (Inqaba Biotechnical Industries, South Africa). The procedure was carried out as described by Le Flèche, 2002 (47). In brief, a 20 µl reaction was used for the PCR and it contained 2 µl DNA template, 10 µl Qiagen master mix, 0.5 µl of each of the two 10 µM primers and 7 µl of DNA free water. The PCR cycle of reactions were as follows: initial denaturation at 94 oC for 5 minutes, followed by 40 cycles of initial denaturation at 94 oC for one minute, annealing at 62 oC for 1 minute and elongation at 72 oC for 1.5 minutes and the final elongation at 72 oC for 10 minutes.
The PCR was carried out using the PCR thermocycler machine (2700, Applied Biosystems). The PCR products were separated electrophoretically using 3 % agarose gel at 85V for 3 hours. The band sizes were estimated against a 100 bp DNA ladder and converted into copy number that was saved as the digital VNTR profile in Excel (47). The data was exported into Bionumerics software package version 7 (Applied Maths, Sint-Martens-Latem) as a character data for analysis.
The unweighted pair-group method using average linkages algorithm (UPGMA) was used to construct the relevant dendogram and the categorical coefficient was used to calculate the similarity of the Multi locus VNTR analysis (MLVA) profiles. In this study, a cluster was defined as a group of isolates with 100% genetic similarity. Additional information on the year of sample collection, location of sample collection (dip tank/game reserve/farm), animal species and type of sample collected was used to define the links between specific clusters. The MIRU-VNTR profile was analysed together with its corresponding spoligotype from this study and combined with results from previous studies in the province (18).