Canine brucellosis is a neglected disease mainly caused by B. canis. Since its first identification in 1960s, human brucellosis case with B. canis infection was intermittently reported, arousing increased concern on its zoonotic feature [20–22]. In China, canine brucellosis was first reportedly in 1984 and endemic in 20 provinces before 1990s [23]. Canine brucellosis reemerged again and showed a trend of spread in the past decade. In 2011, a brucellosis outbreak was reported in a beagle dog breeding farm in Beijing and B. canis infection was confirmed by bacteriological isolation [24]. In the same year, a 45-year-old woman was diagnosed as B. canis infection in Zhejiang provinces [9]. These cases revealed the human populations at high risk for exposure to B. canis, which are kennel workers and pet owners, especially immunosuppressed patients, children and pregnant women. As data displayed in this study, prevalence of canine brucellosis in some cities could be at least 2.78% (34/1, 220). But for most areas of China, the prevalence of canine brucellosis is still unknown.
The unknown to prevalence of canine brucellosis, in a great extent, is due to the lack of accurate and commercially available testing reagents. The RSAT and 2ME-RSAT using rough Brucella as antigen was the first method to detect canine brucellosis [25, 26], and it is still popularly used in some countries [27, 28]. But in China, there is no officially authorized RSAT antigen and homemade reagents was only limitedly available, making large scale screening test for canine brucellosis impossible. Even though RSAT can be easily accessible, the result with this method should be carefully interpreted. Our data showed that the RSAT has a strong cross reaction to rabbit serum infected with Vibrio Parahaemolyticus. In China, Vibrio Parahaemolyticus is a major food borne pathogen to human [29], especially in coastal cities, where pet dog can be infected with this pathogen in household and misdiagnosed as brucellosis. We once isolated Vibrio Parahaemolyticus from dog blood collected in coastal city of Tianjin. Moreover, cross reaction to Bordetella, Pseudomonas and Moraxella was also reported in RAST using rough Brucella cell antigen [10]. Thus, the RAST positive dog should be further diagnosed with more specific method.
PCR and real time PCR have high sensitivity in detecting pathogen specific nucleic acid fragment. Primer sets have been proposed to detect B. canis using blood, swab and tissue samples, including ITS66 and ITS279 directed to 16S-23S rRNA interspace region and BcSS-PCR primers for B. canis species-specific fragment [30–32]. We also used these two sets of primers to determine the truly Brucella infected sera. As shedding of B. canis is undulated, PCR testing using blood and swab samples may lead to false negative result. So, the actual prevalence of brucellosis may be higher than 2.78% as displayed in this study. In our practical application, the animal host DNA can be easily amplified with ITS66 and ITS279 primers, and the size of product is equal to the targeting Brucella fragment in agarose gel. To avoid false positive result, PCR product by this primer set has to be sequenced and compared to Brucella reference sequence. The Real time PCR targeting for Brucella genus was also used in this study [33], which showed high sensitivity and specificity in detecting blood and swab samples and can be a useful tool to replace PCR methods.
The oligo-polysaccharide (OPS) antigen is not applicable in diagnosing rough B. canis caused brucellosis, so a quite number of protein antigens have been assessed for their potential use in detection for canine brucellosis [13, 15, 16]. These protein antigens demonstrated considerable efficiency in diagnosing dog sera, but their cross reaction to other pathogen infected sera were not evaluated, especially the pathogens often isolated from dog like Escherichia coli, Salmonella and Vibrio parahaemolyticus. Previous study showed that BP26 and Omp31 have strong reactivity to sera of animals infected by both smooth and rough Brucella [19, 34, 35], in consistent with the data in this paper. However, the cross reaction of BP26 to Listeria monocytogenes positive serum and Omp31 to Vibrio parahaemolyticus positive serum were prominent [19]. Multi-epitope based fusion protein showed a bit less sensitivity than BP26 and Omp31, but it had the least cross reaction to sera infected with other pathogens [17]. Based on all these results, the fusion protein is the best candidate antigen for serological diagnosis for canine brucellosis.
Canine brucellosis is principally caused by B. canis but it can be infected with B. melitensis or B. abortus in rural and pastoral area where dogs were closely contact with small ruminants and cattle. We cannot guarantee that all of 34 brucellosis positive sera used in this study were solely from B. canis infected dogs, as the primers used for PCR testing were not B. canis specific. The ITS66 and ITS279 primers can amplify the genomic DNA of B. canis, B. melitensis, B. abortus and B. suis, while BcSS primers can detected B. canis and B. melitensis strains. We successfully isolated one B. canis strain from these brucellosis positive samples, and subsequent MLVA16 typing revealed a MLVA profile as 2-3-9-11-3-1-5-2-5-40-9-6-7-9-4-3. This MLVA16 genotype was closely related to a B. canis strain isolated from Beijing in 2011 [9, 24]. Bacteriological isolation suggested that most of dogs raised in cities were B. canis infected. Even if some dogs were infected with smooth Brucella, they can still be detected by this fusion protein, as the epitopes contained in this antigen are shared almost by all Brucella species.