Molecular detection of Mycobacterium bovis in cattle herds of the state of Pernambuco, Brazil

Background The present study aimed to direct detect Mycobacterium bovis in milk (n = 401) and blood (n = 401) samples collected from 401 dairy cows of 20 properties located in the state of Pernambuco, Brazil, by real-time quantitative PCR (qPCR) targeting the region of difference 4 (RD4). Risk factors possibly associated with bovine tuberculosis (BTB) were also evaluated. Results Of the 802 samples analyzed, one milk (0.25 %) and eight blood (2 %) samples were positive for M. bovis in the qPCR and their identities were confirmed by sequencing. Animals positive for M. bovis were found in six (30 %) of the 20 properties visited. None of the risk factors evaluated were statistically associated with BTB. Conclusions M. bovis DNA was detected in one milk sample what may pose a risk to public health because raw milk is commonly consumed in Brazil.


Background
Bovine tuberculosis (BTB) is caused by Mycobacterium bovis, a member of the Mycobacterium tuberculosis complex that affects mammals, including humans [1].
M. bovis has been isolated from milk and colostrum samples what can be important to perpetuate BTB infection in a herd through the digestive route [2]. Raw milk is commonly consumed in Brazil [3] and clandestine milk is an important public health issue in the country [4,5].
Despite the fact that Brazil has a National Program for Control and Eradication of Tuberculosis and Brucellosis (Programa Nacional de Controle e Erradicação da Tuberculose e da Brucelose -PNCETB) supervised by a public agency, its implementation is not mandatory [6]. The tuberculin skin test is the diagnostic method recommended by the PNCETB and it must be followed by bacterial isolation for result confirmation [7]. Efforts to reduce the risk of M. bovis infection must include sanitary measures to ensure a healthy cattle herd.
Molecular techniques such as Polymerase Chain Reaction (PCR) have been used for BTB diagnosis in several clinical samples such as blood, milk and nasal exudates [2]. Standardization of direct methods for detection of M. bovis in clinical samples will enable a more accurate BTB diagnosis and facilitate epidemiological studies on M. bovis prevalence [1,8].
In the present study, we used qPCR for direct detection of Mycobacterium bovis in milk and blood samples of cattle from the state of Pernambuco, Brazil.

Sampling
The sample size was calculated as recommended by Thrusfield [9] using the following parameters: bovine population of 336,221 animals in the micro region of Garanhuns, state of Pernambuco, Brazil [10], 95 % confidence interval and 5 % sampling error margin using a prevalence of 50 %, since there is no official data on BTB prevalence in the studied region. According to the calculation, the minimum sample size should be 385 dairy cattle.
Written informed consent was obtained from the farmers to take samples from the cattle. The blood samples (n = 401) were collected by caudal venipuncture, stored in tubes containing citrate, properly identified and sent to the Garanhuns Laboratories Center (Central de Laboratórios de Garanhuns -CENLAG), located in the Garanhuns Academic Unit (Unidade Acadêmica de Garanhuns -UAG) of the Federal Rural University of Pernambuco (Universidade Federal Rural de Pernambuco -UFRPE), Brazil.
The milk samples (n = 401), which consisted of 50 ml of a pool of milk from the four quarters of each cow, were collected during milking after the udder disinfection with 70 % alcohol and the first jets of milk were discarded. Then the samples were stored in sterile bottles, cooled and sent to the CENLAG (UAG-UFRPE).
An epidemiological questionnaire containing multiplechoice questions concerning animal production characteristics, hygiene and sanitary aspects of the herd, and reproductive management was applied in each property. The questionnaire comprised 11 possible risk factors for M. bovis infection, as follows: herd size (less than 50 animals, 51-100 animals, 101-200 animals, more than 201 animals), rearing system (intensive, extensive, semiintensive), origin of replacement animals (farm's own herd, another farm, both), conducting quarantine after animal's purchase, performance of BTB diagnostic tests upon animals' acquisition, water source (stagnant or running), milking procedure (manual or mechanic), frequency of cleaning the farm facilities, udder disinfection, feeding colostrum to calves and history of BTB in the herd.

DNA Extraction
M. bovis DNA was extracted from milk samples using the QIamp® kit (Qiagen Inc.) following the manufacturer's instructions. Leukocyte DNA was isolated by a modified phenol-chloroform extraction method [11], 100 μl of white blood cells were used, ammonium acetate and phenol-chloroform for DNA extraction from the blood samples.

Positive control
The Mycobacterium bovis ATCC 19274 strain was gently provided by the Oswaldo Cruz Foundation (Fundação Osvaldo Cruz -FIOCRUZ, Rio de Janeiro, Brazil) and was used for construction of a plasmid harboring the target sequence, which was the positive control in the molecular assays. M. bovis genomic DNA was extracted and the fragment corresponding to the region of difference 4 (RD4) was amplified with the specific primers reported by Sales et al. [12]. The target fragment was cloned using Escherichia coli XL1 blue strain and TA cloning kit® (Invitrogen) according to the manufacturer's instructions. The recombinant plasmid pRD4-TA was sequenced by the Sanger method using an ABI3100 Genetic Analyzer (Applied Biosystems).

Real-time PCR
The molecular detection of M. bovis DNA in the milk and blood samples was performed in the Laboratory of Immunogenetics (Laboratório de Imunogenética) of FIOCRUZ, state of Pernambuco, Brazil. Quantitative real time PCR (qPCR) was performed using the same primer set used for amplification of the RD4 fragment of the positive control and a fluorescent probe that discriminates M. bovis from other M. tuberculosis complex members since it hybridizes with both the 5′ and 3′ RD4 deletion flanking sequences, which only occur directly adjacent to each other in M. bovis. The probe was designed using the software Primer Express® Software and targeted a region in the amplicon in between the primer pair. The probe showed 100 % homology to M. bovis in BLAST/ncbi. Probe sequence: 5'-/56-FAM/AGCCG-TAGTCGTGCAGAAGCGCA/3BHQ_1/-3'. The total reaction volume was 25 μL comprising 2.0 μL of DNA, 12.5 μL of TaqMan®Universal PCR Master Mix, 1.0 μL each primer (5 pmol), 0.5 μL of probe (5 pmol) and water (8 μL). The amplification conditions were 95°C for 15 min (denaturation) followed by 40 cycles of 94°C for 15 s and 60°C for 60 s. In all PCR runs, standard curves were obtained using the positive control, plasmid DNA encompassing the mycobacteria RD4 sequence, which was prepared in triplicate by serial dilution of 10x plasmid DNA from 200 ng (Quantification cycle -Cq = 11.8) to 0.0002 ng (Cq = 32.2). The qPCR was performed in an ABI 7500 Real-Time PCR system set for absolute quantification The slope of the standard curve was −3.40 and R = 0.999, with 97 % efficiency.
Spiked samples were not used. The standard curve and the detection limit were determined using a serial 10X dilution from 100 to 10 −10 ng/ul in triplicate and the positive control was detected in samples with up to 10 −6 ng/ul. The reaction was repeated four times in different days and the same results were obtained in each day.
To verify the presence of inhibitors in the samples, a few blood samples of 1000 ng/μl DNA were randomly selected and diluted them to 800, 600, 400, 200 and 50 ng/μl of DNA. Then 20 ng were added of positive control to the diluted samples and performed a qPCR. All the dilutions had the same Cq in the qPCR; therefore, there were no inhibitors in the samples. The commercial kit ABI PRISM BigDye Terminator Cycle Sequencing Ready Reaction v3.1 (Applied Biosys-tems®) was used for DNA sequencing following the manufacturer's recommendations. The RD4 fragments were sequenced by the Sanger method and the reaction products were analyzed in the ABI 3500xL Genetic Analyzer (Applied Biosystems).

Sequencing
All the sequences obtained in the present study were compared with the RD4 fragment of the reference genome (88 bp) (GenBank Access number BX248339.1) using the software Blast-N (http://www.ncbi.nlm.nih.gov) and MEGA6 [13].

Ethical considerations
The Ethics Committee on Animal Use (Comissão de Ética no Uso de Animais -CEUA) of UFRPE provided scientific and ethical clearance for the present study (reference number 23082.004671/2013, license number 028/2013).

Statistical analysis
The absolute and relative prevalence of M. bovis in the milk and blood samples were determined by descriptive analysis. Univariate analysis using chi-square test, Pearson's test or Fisher's exact test were used to evaluate the possible risk factors associated with BTB. All statistical analyzes were performed in the Epi Info 3.5.1 software.

Results
Of the 802 samples analyzed, one milk (0.25 %) and eight blood (2 %) samples were positive for M. bovis in the qPCR and their identities were confirmed by sequencing. All positive samples were from different animals ( Table 1).
Six (30 %) of the 20 properties visited had animals positive for M. bovis ( Table 1). None of the risk factors evaluated in the present study were statistically associated with BTB as shown in Table 2.

Discussion
This is the first report of direct detection of M. bovis DNA in milk and blood samples from cattle of the region of Garanhuns, state of Pernambuco, Brazil.
The prevalence of M. bovis DNA in milk samples ranges from 2 to 87 % according to different studies [14][15][16][17][18][19][20], which evaluated the mycobacteria presence by PCR. The different prevalence rates observed by them may be related to management characteristics [21], sampling methods [22] and disease-control measures adopted in each location [23].
The presence of M. bovis in milk may pose a risk to public health, because humans can become infected by M. bovis through exposure to infected animals, consumption of infected raw milk and dairy products [24,25]. The presence of M. bovis in milk samples is a concern because it is estimated that 41 % of all milk consumed in Brazil is not pasteurized [12] being a source of infection of human TB. Since the clinical symptoms of human TB caused by M. bovis are indistinguishable from those caused by M. tuberculosis [23,26], a detailed epidemiological investigation considering the patients eating habits and professional occupation must be performed by the health surveillance service to determine the TB causal agent. In Brazil, a study performed by Silva et al. [27] with 189 TB patients identified coinfection with M. bovis in three patients. In two of these patients, consumption of cheese made with raw milk was the probable cause of infection. The other patient used to work in a slaughterhouse, so the infection was related to labor risk. Other study conducted in Brazil also identified M. bovis in humans, although with lower prevalence. It is believed that M. bovis prevalence in Brazil is underestimated [28].
In the present study, the milk sample positive for M. bovis did not belong to any of the eight animals that had positive blood samples what can be explained by various reasons: collection of only one milk sample per dairy cow, interaction between the bacillus and the bovine immune system cells, which may have decreased the amount of bacillus in milk [2,20,29], and presence of Refers to the total number of animals evaluated in the study; it is not a sum of each column milk proteins and fat that may have impaired the extraction of M. bovis DNA from the milk samples [14,30]. Despite these limitations, studies using experimentally contaminated milk have demonstrated that PCR can detect the mycobacteria in milk samples with much lower concentrations of M. bovis DNA than the concentration  [2,20,29,31]. The mycobacteria has been more frequently identified in blood than in milk samples [20,32,33]. Another factor that can hinder the detection of M. bovis in milk is its intermittent release during a short period post-infection [20,29]. Pardo et al. [34] evaluated the mycobacteria secretion pattern in 780 milk samples collected from 52 animals for 15 consecutive days and M. bovis showed an intermittent and irregular release pattern in 26.5 % of the samples [34].
In the present study, of the six farms that had animals positive for M. bovis, only one have a history of performing tuberculin tests upon acquisition of new animals. According to the farmer, all the animals were negative for M. bovis in the tuberculin tests. This data shows the importance of using tests that are more sensitive in enzootic areas, as the state of Pernambuco, Brazil, where cases of BTB have already been identified using tuberculin skin test [35,36].
Although there was no statistical association between herd size and M. bovis positivity, the mycobacteria was more prevalent in larger herds (101-200 animals). Herd size can influence BTB epidemiology because a high population density favors a more frequent contact between animals, facilitating the mycobacteria dissemination [37,38].
According to Skuce et al. [39], M. bovis can survive in water what favors its dissemination. M. bovis DNA was identified in water samples experimentally contaminated even 11 months after contamination [40]. In Uganda, Africa, where cattle commonly drinks running water from rivers or streams, a study evaluated the risk factors associated with BTB and concluded that the water source was statistically associated with the disease [41]. However, in the present study, no correlation was found between water source and presence of M. bovis in milk and blood samples ( Table 1). The fact that water sources could be implicated with BTB transmission may be a concern to health authorities, because control measures would also have to consider this contamination source besides slaughter of positive animals.
Despite the higher number of animals positive for M. bovis in herds with low frequency of cleaning the farm facilities and lack of udder disinfection before milking, no positive correlation was found between pathogen presence and these risk factors. Roxo [42] reported that cleaning, disinfection and hygiene are risk factors for TB. Waste management and treatment of organic matter can influence M. bovis prevalence in areas with previous cases of TB [43].

Conclusion
M. bovis DNA was detected in one milk sample what may pose a risk to public health. We suggest that environmental control measures should be implemented in farms at high risk of TB transmission because environmental factors contribute to bacteria perpetuation and dissemination.