Use of PrimeStore® Molecular Transport Medium and Xpert MTB/RIF Ultra for Rapid Detection of Mycobacterium Bovis in African buffaloes (Syncerus Caffer).


 Mycobacterium bovis is the causative agent of bovine tuberculosis (bTB) in wildlife. Confirmation of M. bovis infection relies on mycobacterial culture, which is time-consuming. Collection and transportation of infectious material also pose a human health risk. PrimeStore® Molecular Transport Medium (MTM) has been shown to effectively inactivate infectious organisms, making it a safe method for handling infectious samples. This study investigated an in-field sampling technique for rapid, safe detection of M. bovis in buffalo tissues. Potentially infected tissues from bTB test-positive buffaloes were swabbed at post-mortem examination and stored in PrimeStore® MTM at ambient temperature until Xpert® MTB/RIF Ultra testing was performed. Additionally, tissue samples were frozen and transported before homogenisation for culture and Ultra testing. Oral swabs were collected from M. bovis-unexposed buffaloes as a negative control cohort. Mycobacterium tuberculosis complex (MTBC) DNA was detected by Ultra in 13/16 tissue swabs and 9/16 matched tissue homogenates from culture-confirmed M. bovis-positive buffalo tissues. MTBC DNA was not detected in swabs from M. bovis-unexposed animals, showing the potentially high specificity of Ultra with PrimeStore® swabs. PrimeStore® MTM sample processing, in combination with the Ultra assay, has the potential to provide a safe, rapid post-mortem screening test for M. bovis in buffaloes.

This highlights the potential disease risk and importance of its control 1,2 . Diagnosis of M. bovis infection in African buffaloes typically relies on the measurement of cell-mediated immune (CMI) responses to M. bovis antigens 3 . Buffalo TB tests include in vitro interferon-gamma (IFN-γ) release assays (IGRAs) and IFN-γ inducible protein 10 (IP-10) release assays (IPRA) that use a modi ed QuantiFERON®-TB Gold Plus (QFT) In-tube antigen stimulation platform, and the in vivo single comparative intradermal tuberculin test (SCITT) 4,5 . However, con rmation of M. bovis infection usually requires mycobacterial culture and genetic speciation of mycobacterial isolates cultured from tissue specimens collected during postmortem examination. Conventional culture methods, however, have suboptimal sensitivity and are timeconsuming, taking up to 56 days before results are available 6 . In addition, post-mortem examination of potentially M. bovis-infected buffaloes, and the collection, transport and storage of tissue specimens are considered a potential human health risk 7,8 . Infectious pathogens, other than M. bovis, may also be present during sample collection and transport, such as Foot-and-Mouth Disease (FMD) virus, which may lead to sample movement restrictions to laboratories due to the risk of spreading the disease. Movement restrictions placed on buffalo samples from locations with reportable diseases to laboratories may result in inadequate bTB surveillance in these areas 9 .
PrimeStore® Molecular Transport Medium (MTM), used in combination with PrimeStore® swabs, has been shown to effectively inactivate infectious organisms, while stabilising the nucleic acids, through lysis of the cell membranes, destruction of proteins and enzymes, and inactivation of nucleases 10 . PrimeStore® MTM therefore provides a safer option for collection and transportation of samples that potentially pose a risk to human or animal health. In addition, this method may facilitate sample transport from areas that may be restricted due to the presence of reportable diseases (such as FMD), since it inactivates all pathogens.
The rapid Xpert MTB/RIF Ultra (Ultra) assay is an automated cartridge-based semi-quantitative nested real-time PCR assay that detects Mycobacterium tuberculosis complex (MTBC) DNA and rifampicin resistance in clinical specimens 11 . The assay targets the multicopy sequences, IS6110 and IS1081, which greatly improves the detection of DNA in paucibacillary samples. IS1081 and IS6110 are exclusively found in MTBC members, which makes it ideal for samples collected from animals that may be infected with members of the MTBC 12,13 . Although the Ultra is typically used for detection of MTBC DNA in human samples, it has been shown to accurately detect M. bovis in wildlife specimens from white rhinoceros (Ceratotherium simum) and elephants (Loxodonta africana) 13 , and several other wildlife species 14,15 .
Therefore, the aim of this study was to investigate the combined use of the Ultra assay with PrimeStore® MTM swab samples as a new screening technique for the detection of MTBC DNA in buffalo tissue specimens, which could provide a rapid and safer method for sample collection and transport, or in-eld testing.

Results
Sixteen tissue samples collected from 13 buffaloes were con rmed to be M. bovis positive by mycobacterial culture and PCR-based speciation ( Table 1). Each of these 16 culture-con rmed M. bovis positive tissue samples had a matching PrimeStore® swab sample collected during post-mortem examination and stabilised in PrimeStore® MTM, and a pre-culture tissue homogenate aliquot. Both the swab and tissue homogenate were tested with the Ultra assay prior to mycobacterial culture. The Ultra assay detected MTBC DNA in 13 of 16 (81.3%) tissue swab samples, while MTBC DNA was detected directly in 9 of 16 (56.3%) tissue homogenates from culture-con rmed M. bovis-positive specimens. The difference between the proportion of positive Ultra assay results from tissue homogenates compared to tissue swabs was not signi cant using the McNemar's test (p > 0.05). The agreement between these tests were regarded as "fair" on the kappa agreement scale, with κ = 0.259 (95% con dence interval (CI) -0.2 to 0.72 and standard error (SE) of 0.23).
Additionally, Ultra assay did not detect MTBC DNA in any oral swabs collected from the 12 randomly selected M. bovis unexposed buffaloes from historically bTB free buffalo herds (data not shown). Studies performed with human specimens have shown that Ultra results provide a measure of the bacillary load in a specimen, with a positive correlation between the semi-quantitative Ultra results and detection of acid-fast bacilli with smear-microscopy 17 . The "MTB trace detected" and "MTB not detected" Ultra results have been shown to be correlated with no acid-fast bacilli detected, whereas Ultra results "MTB detected high" and "medium" typically correlate with positive smear-microscopy results 17 . The Ultra test results in this study included "MTB detected low", "MTB detected very low", "MTB trace detected" and "MTB not detected". Since the "MTB trace detected" Ultra results from tissue swabs and homogenates were seen in culture-con rmed M. bovis positive specimens, it may indicate greater sensitivity of this method compared to smear microscopy. However, varying copy numbers of IS6110 and IS1081 between MTBC members and between strains have also been shown to in uence the sensitivity and may not entirely relate to the performance of the Ultra assay 11,13 . The bacillary load in samples in which no MTBC DNA were detected ("MTB not detected") was likely below the threshold for detection by the Ultra assay, although the homogenates from the same tissue sample had a positive M. bovis result by mycobacterial culture and PCR speciation. This could be due to a low but viable bacillary load that would have increased to a detectable level during the 56 days of mycobacterial culture, which could further explain the discordant results between mycobacterial culture (all M. bovis positive) and Ultra results from tissue homogenates. Difference in sample volume between mycobacterial culture and Ultra from tissue swabs may also contribute to discordant results between these two methods.
The oral swabs from the M. bovis-unexposed buffaloes were all negative on the Ultra assay. These results are consistent with the reported high speci city of the Ultra assay 11 . Interestingly, although these swabs were used as negative controls, the swabs would have had a greater likelihood of contamination than those collected post-mortem, since these animals were sampled in a dusty, natural environment, where they were likely exposed to various environmental organisms.
The study had several limitations. Firstly, there was a small sample size of culture-con rmed M. bovispositive buffaloes and oral swab samples available from the M. bovis-unexposed buffaloes, which did not allow for direct comparison of the same sample type tested in the Ultra. Secondly, since the Ultra assay was only performed in the laboratory, we were unable to compare laboratory and in-eld testing with the Ultra assay and PrimeStore® sampling platform. Finally, since tissue samples for Ultra were speci cally selected based on visible lesions, it is unknown how the Ultra would perform using M. bovis culture-positive tissue samples with no visible lesions and therefore presumed low bacillary load. Future studies should include a larger sample size with buffaloes at various stages of infection. The application of eld testing should also be investigated, which would allow more rapid sample processing and the use of fresh samples.
The Ultra assay with samples stored in PrimeStore® MTM shows promise as a rapid post-mortem screening test, which could be used to identify MTBC DNA in tissue samples and ultimately identify infected herds more rapidly than mycobacterial culture. This test can also be used to decrease the number of samples sent for mycobacterial culture by identifying those that contain mycobacterial DNA. Pooling swab samples from each animal will further decrease the number of samples tested. In addition, the use of PrimeStore® MTM to safely store and transport infectious samples will safeguard human and animal health as well as permit greater use of samples that may otherwise be restricted from being transported to diagnostic laboratories.

Animals and post-mortem sample collection
The Hluhluwe-iMfolozi Park (HiP) in KwaZulu-Natal, South Africa, is a game park that is endemic for M. bovis-infection in the wildlife population. During the 2019 annual test-and-slaughter bTB control program, African buffaloes were captured, immobilised, sampled and tested for bTB, using the testing regime previously described 18 . Bovine TB test-positive buffaloes (n = 22) underwent post-mortem examination as previously described 18 . Various tissues with gross pathological lesions consistent with bTB 18 were dissected and swabbed during post-mortem examination using PrimeStore® swabs (Longhorn Vaccines and Diagnostics, San Antonio, Texas, USA) and stored in PrimeStore® MTM (Longhorn Vaccines and Diagnostics) at ambient temperature (which may reach temperatures up to 25°C during winter in HiP) until further processing. Tissues included retropharyngeal lymph nodes, tonsils, lung, mediastinal lymph nodes, abdominal serosa, and tracheobronchial lymph nodes. In addition, the scalpel blade used to cut tissues for a particular animal was also sampled on some occasions (Fig. 1). Samples with visible lesions were stored in separate containers and frozen at -20°C until processed for mycobacterial culture. Since tissue samples were not available from M. bovis-unexposed buffaloes, oropharyngeal swabs were collected from 12 randomly selected buffaloes that were tested for bTB in historically bTB-negative private game parks, as previously described 18 . PrimeStore® swab samples were stored in PrimeStore® MTM at ambient temperature until further processing. All samples were transported to Stellenbosch University for further analyses.
Ethical approval for this study was granted by Stellenbosch University Animal Care and Use Committee laboratory at Stellenbosch University, as previously described 3 . A 1 ml aliquot of the liquid fraction of each tissue homogenate was used for quantitative polymerase chain reaction (qPCR) analysis by the Xpert MTB/RIF Ultra (Ultra) assay (Cepheid Inc., Sunnyvale, California, United States). Tissue homogenates were then further processed for mycobacterial culture using the standard mycobacterial culture protocol with the BBL® MycoPrep™ specimen decontamination kit (Becton Dickinson) and BACTEC™ MGIT™ 960 Mycobacterial Detection System (Becton Dickinson), as previously described 3 . Speciation by region of difference PCR was performed on all culture positive samples to con rm the presence of M. bovis 19 (Fig. 1).

Xpert MTB/RIF Ultra assay
The Ultra assay (Cepheid) was performed using the oropharyngeal swabs collected from M. bovisunexposed buffaloes and tissue swab samples from HiP buffaloes, according to the Ultra alternative sample assay instructions, as previously described 13 . Brie y, Xpert MTB/RIF sample lysis reagent (Cepheid) was mixed in a 1:1 ratio with each PrimeStore® MTM stabilised swab sample and incubated for 10 min at room temperature (RT). Hereafter it was vortexed for 10 s before incubation at RT for 5 min followed by a 5 s vortex step. Approximately 2 ml of the prepared sample was dispensed into the sample chamber of the Ultra cartridge for automated processing. Aliquots of tissue homogenate processed for mycobacterial culture were also tested with the Ultra assay, with sample preparation as described above. The Ultra assay classi es MTBC DNA detection as high, medium, low, very low or trace, based on the quantity of MTBC DNA that was detected, and all of these classi cations were considered positive in this study. When the Ultra could not detect MTBC DNA ("MTB not detected"), it was considered a negative result. Rifampicin resistance was also reported but is irrelevant in the context of this study (Fig. 1).