Comparison of decontamination methods for primary isolation of Mycobacterium bovis in paucibacillary bovine tissues

Letters in Applied Microbiology ISSN 0266-8254 ORIGINAL ARTICLE Comparison of decontamination methods for primary isolation of Mycobacterium bovis in paucibacillary bovine tissues L. Medeiros 1, C.D. Marassi 1, R.S. Duarte 2, M.G. da Silva 2 and W. Lilenbaum 1 1 Universidade Federal Fluminense, Laboratório de Bacteriologia Veterinária, Rua Hernani Mello, Niterói-RJ, Brazil 2 Universidade Federal do Rio de Janeiro, Instituto de Microbiologia, Av. Carlos Chagas Filho, Cidade Universitária, Ilha do Fundão, Rio de Janeiro, Brazil Keywords decontamination, Mycobacterium bovis, paucibacillary sample, tuberculosis. Correspondence Luciana Medeiros, Universidade Federal Fluminense, Centro de Ciências Médicas, Instituto Biomédico, Rua Prof. Hernani Mello, 101 Centro 24210-030, Niteroi, RJ, Brazil. E- mail: lusmedeiros@yahoo.com.br 2011 0824: received 15 May 2011, revised 7 November 2011 and accepted 18 November 2011 doi:10.1111/j.1472-765x.2011.03185.x Abstract Aims: To compare three decontamination methods applied to paucibacillary samples for primary isolation of Mycobacterium bovis from suspect lesions. Tuberculosis caused by Myco. bovis is an important infectious disease of cattle in Brazil and also has zoonotic potential. Although a national campaign based on testing and slaughtering cattle has achieved good results, there is a strong need to develop better diagnostic methods to identify cattle with recent infections harbouring few bacilli. Methods and Results: A dairy herd (274 adult crossbred cows) located in the state of Rio de Janeiro was tested for tuberculosis with both single intradermal tuberculin test and comparative intradermal tuberculin test. Reactive cows (n = 27, 9Æ8%) were slaughtered and suspect lesions were collected (one sample per cow). Samples considered paucibacillary (based on microscopy) were decontaminated with 0Æ75% hexadecylpyridinium chloride (HPC), 4% sodium hydroxide (Petroff) or 6% sulphuric acid. Using these methods, 10, five and six, respectively, of the 27 samples yielded positive cultures. Overall, Myco. bovis was isolated from 14 of 24 cows. Although the HPC method resulted in isolation of more Myco. bovis strains than either Petroff or sulphuric acid methods (P =0Æ015), it did not result in the recovery of Myco. bovis from all samples. However, using both HPC and 6% sulphuric acid methods for decontamination was possible to identify 13 of 14 (92Æ9%) of infected cows. Conclusions: At least two methods should be used concurrently for primary isolation of Myco. bovis from bovine tissues, particularly for paucibacillary samples. Significance and Impact of the Study: Detection of low numbers of Myco. bovis in tissue is an important goal in optimizing the detection of bovine tuberculosis and should assist in identification of infected cattle, in particular, those with few Myco. bovis bacilli. This was apparently the first study comparing three decontamination methods for the detection of Myco. bovis in paucibacillary samples from naturally infected cattle. Introduction Mycobacterium bovis causes bovine tuberculosis, an important infectious disease of cattle in several countries (OIE 2008; Thoen et al. 2009). Although national campaigns have reduced the incidence of the infection worldwide, it remains an important public health concern because of its zoonotical potential and re-emergence in animals and humans (Thoen et al. 2009). The standard method for detection of bovine tuberculosis is the intradermal tuberculin skin test, based on delayed hypersensitivity to a purified protein derivate (PPD) extracted from 182 Letters in Applied Microbiology 54, 182 186 ª 2011 The Society for Applied Microbiology

L. Medeiros et al. Myco. bovis in paucibacillary tissues the agent (Monaghan et al. 1994; OIE 2008). Nevertheless, the definitive diagnosis that a herd is infected requires clear evidence of the agent, based on either bacteriological culture, molecular methods, or both (Corner 1994; Figueiredo et al. 2009). In cattle recently infected with Myco. bovis, most lesions are paucibacillary (Liebana et al. 2008), making it more difficult to culture the causative agent. Additionally, because samples used to cultivate Myco. bovis may be contaminated by other bacteria, sputum, milk and suspected lesions should be decontaminated prior to culture to minimize overgrowth. Quaternary ammonium compounds, sodium hydroxide, sulphuric acid, oxalic acid and hexadecylpyridiumchloride (HPC) have been used for decontamination of clinical specimens, (Corner 1994; OIE 2008); these approaches yielded variable results for sensitivity and specificity, depending on the bacterial load and specimens. Furthermore, chemical compounds may reduce mycobacterial viability (up to 3Æ1 log 10 loss of organisms in tissues; Reddacliff et al. 2003), thereby increasing false-negative results, particularly with paucibacillary samples. The detection of low numbers of Myco. bovis in tissue in final rounds of control programs is important in the control of bovine tuberculosis and removes the final pockets of Myco. bovis from an infected cohort. The objective of this study was to compare decontamination methods applied to paucibacillary samples for primary isolation of Myco. bovis from suspect lesions of skin testreactive cattle. Materials and methods Study design A dairy herd comprising 274 adult crossbred (Gyr Holstein) cows from Rio de Janeiro, Brazil, was studied from 2006 to 2007. The herd had been annually tested for bovine tuberculosis, with negative results, and the animals had not presented any clinical signs suggestive of infection. Approximately 6 months after introduction of a new batch of dairy cows, the herd was retested for Myco. bovis infection with a single intradermal tuberculin test (SITT), resulting in 23 positive animals plus nine inconclusive animals. After 90 days, positive and inconclusive animals were retested by comparative intradermal tuberculin test (CITT), confirming the infected status of the 32 animals. Unfortunately, five CITT reactive cows were pregnant and the owner would not give consent to have them slaughtered. The remaining CITT reactive cows (n = 27) were slaughtered, lymph nodes and suspect lung lesions were collected and submitted for microscopic examination and bacterial culture. Samples were decontaminated by three methods and proportions of positive samples were compared (Q Cochran test). Intradermal tests A SITT was performed by injecting 0Æ1 ml of bovine PPD (bovppd Myco. bovis strain AN5, 1 mg protein per ml; Instituto Biologico, Sao Paulo, SP, Brazil) in the cervical area of each cow. After 72 h, the site was measured with callipers and the cow considered reactive if a swelling more than 4Æ0 mm was present. The CITT consisted of the same procedure, plus inoculation of 0Æ1 ml avian PPD (Myco. avium strain D4, 0Æ5 mg protein per ml; Instituto Biologico) in the cervical area, approximately 20 cm from the bovppd inoculation. Cattle were considered reactive if the two sites differed in thickness by more than 4Æ0 mm. Decontamination of samples All reactive cows were slaughtered 30 days after the CITT. Portions of lung, with lesions suggestive of tuberculosis, plus subscapular lymph nodes (without regard to the presence of lesions) were collected. A composite sample, consisting of 15 g of tissue from each cow, was sliced into small pieces using sterile scissors and macerated with sterile sand. The macerated tissue was resuspended in 40 ml of sterile water and 15 ml of the supernatant was divided into three aliquots of 5 ml for decontamination. To the first aliquot (5 ml), 10 ml of 0Æ75% hexadecylpyridinium chloride (HPC) was added, and to the second, an equal volume of 6% sulphuric acid was added (Holanda et al. 2002). Samples decontaminated with HPC and sulphuric acid were held at room temperature for 15 min. Tissue suspensions were centrifuged for 15 min at 1000 g. The supernatant was discarded and precipitate washed twice with sterile distilled water and centrifuged again, as previously described. The pellet was resuspended in 2 ml of sterile distilled water. The third aliquot (5 ml) was mixed with an equal volume of 4% sodium hydroxide, the sample shaken by hand and allowed to settle for 10 min at room temperature. This process was repeated three times. The samples were then centrifuged for 25 min at 3000 g and then neutralized with a KH 2 PO 4 14% solution. The supernatant was discarded and pellet resuspended in 2 ml of sterile distilled water. Detection of acid fast bacilli Smears were prepared from the sediment of each tissue pool (OIE 2008) before and after decontamination, stained by Ziehl Nielsen for visualization of acid fast Letters in Applied Microbiology 54, 182 186 ª 2011 The Society for Applied Microbiology 183

Myco. bovis in paucibacillary tissues L. Medeiros et al. bacilli and examined using a light microscope (1000 magnification) by two technicians (in a blinded analysis). 4% Sodium hydroxide (6) Bacterial culture Each pellet was resuspended in 2 ml of sterile distilled water, and 0Æ2 ml of the solution was inoculated onto two slopes of a solid, egg-based media (Lowenstein- Jensen, with 0Æ5% pyruvate). The remaining suspension was used to prepare the smears and stored for further studies. Cultures were incubated at 37 C and observed weekly for 12 weeks. 1 3 0 2 5 0 3 Identification of isolates Smooth and off-white colonies, suggestive of Myco. bovis, were subjected to molecular characterization. The mycobacterial DNA was extracted as described by Meikle et al. (2007), and a multiplex PCR targeting the RvD1Rv2031c genomic sequence was employed (Figueiredo et al. 2009) to identify specific Myco. bovis species. Statistics The Q Cochran test was used to compare decontamination methods. Results Pools of tissues were created for each of the 27 reactive slaughtered cows (9Æ8% of the herd). Fifteen of these cows were resident on the farm, whereas the remaining 12 had been acquired 6 months before testing. Acid fast bacilli were not detected in any smear, neither before nor after decontamination methods. As the cattle were tuberculin reactive and had characteristic lung lesions, tissues were regarded as paucibacillary. Mycobacterium bovis was recovered from 14 of 27 samples (51Æ8%) and all isolates were confirmed by molecular characterization. Decontamination with 0Æ75% HPC yielded Myco. bovis from 10 samples, whereas 4% sodium hydroxide and 6% sulphuric acid each yielded Myco. bovis from six and five samples, respectively (Fig. 1). The proportion of positive samples was higher for HPC than for each of the other two methods (P =0Æ015). Using both HPC and 6% sulphuric acid methods for decontamination identified 13 of 14 (92Æ9%) samples of infected cows. Discussion 0 75% Hexadecylpyridinium (10) 6% Sulphuric acid (5) Figure 1 Venn diagram of performance of three methods for decontamination of paucibacillary samples, for detecting Mycobacterium bovis from pools of suspect lesions in bovine tissues. The total number detected by each method is listed in parentheses. The studied herd was regularly tested once a year by SITT, with negative results for the previous 2 years. However, 32 cows were purchased from other herds and introduced apparently without testing. Unfortunately, the origin of the animals and their sanitary status could not be determined. Six months later (August 2006), 23 cows were reactive by SITT, with another nine cows regarded as inconclusive. After 90 days, the status of the 32 cattle was confirmed by CITT, including 13 recent acquisitions and 19 that were long-term members of the herd. Unfortunately, four supposedly recently infected cows and one recent acquisition could not be slaughtered. We inferred that these 19 CITT reactive cows were recently infected, leading to paucibacillary lesions, and furthermore, that the most probable source of infection was the recently acquired cows. The definition of paucibacillary samples for humans infected with Myco. leprae is based on the presence of a lesion with no bacilli visible with microscopic examination (WHO 2011). However, classification of paucibacillary lesions caused by Myco. bovis in bovine tissues is not well defined. This type of classification is of considerable relevance, especially in the final stages of bovine tuberculosis eradication programs, when cattle often have lesions with few or no bacilli detected microscopically. In the absence of a standard definition of paucibacillary for tissue samples of bovine origin, the authors applied the same guidelines used for human leprosy: there were no visible acid fast bacilli, but the cattle had suspect lesions and were tuberculin reactive. It is well known that recently infected cattle often have lesions with low concentrations of bacilli (Liebana et al. 2008). Furthermore, decontamination procedures may damage the mycobacterial cells in the sample (Reddacliff et al. 2003; Giacomelli et al. 2005), leading to false-negative results at microscopic examination. Therefore, it was 184 Letters in Applied Microbiology 54, 182 186 ª 2011 The Society for Applied Microbiology

L. Medeiros et al. Myco. bovis in paucibacillary tissues not surprising that samples had negative results at microscopic examinations both before and after decontamination. In addition to being recently infected, it was noteworthy that these cattle were approximately 50% Bos indicus and 50% Holstein. In that regard, Bos indicus cattle have been reported to develop less severe lesions than Holstein cows (Ameni et al. 2007). Mycobacterium bovis was isolated from 14 of 27 (51Æ8%) of samples, consistent with a low index of recovery of mycobacteria from bovine lesions following decontamination (Corner 1994; Fraguás et al. 2006; Ambrosio et al. 2008). Contamination by other micro-organisms may have influenced isolation of Myco. bovis cells, but unfortunately, because of technical limitations, it was not possible to confirm this. Furthermore, as the prevalence of the infection diminishes because of successful eradication programs, primary isolation of Myco. bovis becomes less frequent (Corner and Trajstman 1988; Corner 1994). The decontamination methods compared in this study have been widely used. Previous studies of our research group demonstrated a low frequency of Myco. bovis recovery from bovine samples using only the Petroff method (Fraguás et al. 2006). Consequently, we decided to improve the isolation rate and reduce the interval needed to obtain positive cultures by concurrently comparing three internationally recognized decontamination methods: the Petroff technique (OIE 2008), HPC (Corner 1994) and sulphuric acid (Holanda et al. 2002). The number of samples positive for only a single decontamination method were 5, 3 and 1 for 0Æ75% HPC, 6% sulphuric acid and 4% sodium hydroxide, respectively, whereas only two of 27 samples were positive following decontamination by any of the three methods (Fig. 1). The HPC was more effective that either Petroff or sulphuric acid methods (P =0Æ015). Ambrosio et al. (2008) compared decontamination methods for bovine tissue samples; the HPC method had the best results, with a significant advantage compared to no treatment or treatment with sodium hydroxide or sulphuric acid. Although Ambrosio et al. (2008) used different concentrations of decontaminants, their results were consistent with those of the present study. These authors suggested the use of HPC as an alternative to the Petroff method for isolation of Myco. bovis from naturally infected bovine samples. However, in the present study, despite the efficiency of 0Æ75% HPC method, when compared to sulphuric acid and Petroff methods, it could not recover Myco. bovis from all infected samples. However, using both HPC and sulphuric acid decontamination identified 13 of 14 positive samples (92Æ9%) as positive. The use of 0Æ75% HPC was superior to other decontaminants for other samples (e.g. milk; (Dundee et al. 2001), thereby extending the use of HPC decontamination to mycobacteria other than Myco. bovis and to additional types of samples. The advantages of using the HPC method have been discussed (Corner and Trajstman 1988; Corner 1994). It is noteworthy that each method has a different method of action; HPC is a detergent, whereas sulphuric acid and sodium hydroxide are an acid and a base, respectively (OIE 2008). Based on these differences, the present results supported the view that decontamination protocols should include chemicals of different classes. Genetic composition of mycobacteria could affect the success of decontamination, because antibiotic-resistant strains have been described as more susceptible to alkaline agents (Yesilkaya et al. 2004). Because this herd was naturally infected, perhaps multiple strains were present, which could have contributed to differences among methods in the number of positive samples. Despite the existence of antibiotic-resistant mycobacteria strains, the use of antibiotics in culture media is required to reduce contamination with nontarget micro-organisms without excessive suppression of nontuberculous mycobacteria (Radomski et al. 2010). In that study, chemical decontamination and addition of PANTA to culture media were more effective together than separately for decreasing numbers of nontarget micro-organisms. Presumably, in accordance with Radomski et al. (2010), using two methods in parallel is an effective strategy to maximize the probability of positive cultures. The use of antibiotics in culture media in association with less harsh decontamination methods, for example HPC, could also improve the isolation of Myco. bovis, particularly when bovine tissues are paucibacillary and potentially contaminated by other bacteria. The statement that one, two or three combined methods should maximize sensitivity may not be valid, as division of samples into smaller parts may give the same results as applying a single method to a larger sample. Nevertheless, division of samples into smaller parts is a strategy to take into consideration when multiple strains are present in a naturally infected herd. Multiple strains may vary the recovery index of the decontamination methods individually and diminish the recovery of Myco. bovis from tissue samples. Further studies are required to determine whether genetic differences among isolates affect their resistance to decontamination methods. The authors also recognize that the small number of samples (27) of this study did not allow strong conclusions to be drawn with regard to the efficacy of each method, or in a comparison of methods. Notwithstanding, the herd studied was naturally infected with Myco. bovis, and the low number of samples was a result of prompt veterinary intervention. We expect that similar investigations could be very useful in countries implementing bovine tuberculosis control programs. Letters in Applied Microbiology 54, 182 186 ª 2011 The Society for Applied Microbiology 185

Myco. bovis in paucibacillary tissues L. Medeiros et al. In conclusion, Myco. bovis was isolated from approximately 50% of pooled tissue samples from reactive cows. Using both HPC and 6% sulphuric acid methods for decontamination identified 13 of the 14 positive samples (92Æ9%). Although it would certainly be more convenient to use a single method, staff in a diagnostic laboratory or research institutes can easily use at the least two of these methods. This routine should be considered for primary isolation of Myco. bovis from bovine tissues, particularly for paucibacillary samples. Furthermore, the present study provided an impetus for establishment of guidelines regarding characterization of paucibacillary tissue samples from field investigations. Acknowledgements This study was supported by FAPERJ, CNPq and CAPES. W.L. is a research fellow of CNPq. We thank John Kastelic (Lethbridge Research Centre, Canada) for technical assistance. References Ambrosio, S.R., Oliveira de Deus, E.M., Rodriguez, C.A.R., Ferreira Neto, J.S. and Amaku, M. (2008) Comparison of three decontamination methods for Mycobacterium bovis isolation Comparação de três métodos de descontaminação para isolamento de Mycobacterium bovis. Braz J Microbiol 39, 241 244. Ameni, G., Aseffa, A., Engers, H., Young, D., Gordon, S., Hewinson, G. and Vordermier, S. (2007) High prevalence and increased severity of pathology of bovine tuberculosis in Holsteins compared to Zebu breeds under field cattle husbandry in Central Ethiopia. Clin Vaccine Immunol 14, 1356 1361. Corner, L.A. (1994) Post mortem diagnosis of Mycobacterium bovis infection in cattle. Vet Microbiol 40, 53 63. Corner, L.A. and Trajstman, A.C. (1988) An evaluation of 1-hexadecylpyridinium chloride as a decontaminant in the primary isolation of Mycobacterium bovis from bovine lesions. Vet Microbiol 18, 127 134. Dundee, L., Grant, I.R., Ball, H.J. and Rowe, M.T. (2001) Comparative evaluation of four decontamination protocols for the isolation of Mycobacterium avium subsp. paratuberculosis from milk. Lett Appl Microbiol 33, 173 177. Figueiredo, E., Silvestre, F.G., Campos, W.N., Furlanetto, L.V., Medeiros, L., Lilenbaum, W., Fonseca, L.S., Silva, J.T. et al. (2009) Identification of Mycobacterium bovis isolates by a multiplex PCR. Braz J Microbiol 40, 231 233. Fraguás, S.A., Cunha-abreu, M.S., Lilenbaum, W., Marassi, C.D., Oelemann, W.M.R. and Fonseca, L.S. (2006) Use of ELISA as a confirmatory test for bovine tuberculosis at slaughter. Prev Vet Med 77, 304 305. Giacomelli, L.R.B., Hebel, C., Ogassawara, R.L.N., Barreto, A.M.W., Martins, F.M., Cardoso, C.L. and Leite, C.Q.F. (2005) Improved laboratory safety by decontamination of unstained sputum smears for acid-fast microscopy. J Clin Microbiol 43, 4245 4248. Holanda, E.D., Lobato, F.C., Mota, P.M. and Abreu, V.L. (2002) Avaliação de Métodos de descontaminação para isolamento de Mycobacterium bovis. R Bras Med Vet 24, 54 57. Liebana, E., Johnson, J., Gough, J., Durr, P., Jahans, K., Clifton-Hadley, R., Spencer, Y., Hewinson, R.G. et al. (2008) Pathology of naturally occurring bovine tuberculosis in England and Wales. Vet J 176, 354 360. Meikle, V., Scneider, M., Azenzo, G., Zumárraga, M., Magnano, G. and Cataldi, A. (2007) Individual animals of cattle herd infected with the same Mycobacterium bovis genotipe shows important variations in bacteriological, histopathological and immune response parameters. Zoonoses Public Health 54, 86 93. Monaghan, M.L., Doherty, M.L., Collins, J.D., Kazda, J.F. and Quinn, P.J. (1994) The tuberculin test. Vet Microbiol 40, 111 124. Radomski, N., Cambau, E., Moulin, L., Haenn, S., Moilleron, R. and Lucas, F.S. (2010) Comparison of culture methods for isolation of nontuberculous mycobacteria from surface waters. Appl Environ Microbiol 76, 3514 3520. Reddacliff, L.A., Vadali, A. and Whittington, R.J. (2003) The effect of econtamination protocols on the numbers of sheep strain Mycobacterium avium subsp. paratuberculosis isolated from tissues and faeces. Vet Microbiol 95, 271 283. Thoen, C.O., LoBlue, P.A., Enarson, D.A., Kaneene, J.B. and de Kantor, I.N. (2009) Tuberculosis: a re-emerging disease in animals and humans. Vet Ital 45, 135 181. World Health Organization, WHO (2011). Classification of Leprosy. http://www.who.int/lep/classification/en/. (last accessed 04 April 2011). World Organization for Animal Health OIE (2008). Manual Terrestre. http://www.oie.int/eng/normes/mmanual/2008/ pdf/2.04.07_bovine_tb.pdf. (last accessed 10 September 2010). Yesilkaya, H., Barer, M.R. and Andrew, P.W. (2004) Antibiotic resistance may affect alkali decontamination of specimens containing mycobacteria. Diagn Microbiol Infect Dis 50, 153 155. 186 Letters in Applied Microbiology 54, 182 186 ª 2011 The Society for Applied Microbiology