Smooth and Rough Brucella Infections in Dogs

Dog brucellosis caused by Brucella canis was first described by Carmicheal in greyhound colonies in United States (USA) in 1966 (Carmichael and Joubert, 1987). It is an important cause of abortion and infertility in dogs worldwide. B. canis infections have been reported in many countries of the world (Carmichael et al., 1970; Carmichael, 1990; Lucero et al., 2002; Wanke, 2004). Dog brucellosis is also caused by other brucella species such as B. melitensis and B. abortus, which are both smooth strains.  B. canis, which is a rough strain, is also an infectious agent for humans and  it is transmitted to humans by various ways including laboratory accidents and contact with sick dogs. Smooth strains of Brucellae have lipopolysaccharide (LPS) in their cell wall while rough strains devoid of this molecule but rough LPS. Therefore, antibodies against B.canis do not react with antibodies produced against smooth strains. For this reason we aimed to detect seropositivity of dog brucellosis caused by smooth and rough species. The cause of under-reporting of B.canis incidence in human is due to absence of standardized tests (Currier et al., 1982; Lopez et al., 2005; Lucero et al., 2002; Lucero et al., 2005). Previous evidence of B. canis infections in Turkey was obtained from the couple of reports regarding to serological evidence of B. canis infections in dogs in Turkey. In a study carried out in 1983, 2-mercaptoethanol tube agglutination test (ME-TAT) was used to test 134 healthy dog serum samples in Ankara and a 6.7% seropositivity rate was detected (Diker et al., 1983) and a similar study of 1987 on22 canine serum samples, researchers obtained similar results (Diker et al., 1987). In another study, seropositivity rate of 12.7%, 7.73% and 7.45% was obtained by TAT, METAT and ELISA, respectively on 362 serum samples in Izmir and Istanbul provinces (Oncel et al., 2005). However, studies on this subject are very few and there is not enough data to reveal the current status of the disease in dogs in our country. In addition, there is no report for dog brucellosis survey caused by smooth Brucella species in last decade. Diagnosis of the disease is based on bacteriological and serological tests. The isolation of the agent is time consuming and required skilled personel. Therefore, the diagnosis of canine brucellosis is largely based on serological tests (Badakhsh et al., 1982; Carmichael, 1990; Carmichael and Joubert, 1987; Lopez et al., 2005; Lucero et al., 2002). Tests used in the serological diagnosis of the disease are rapid slide agglutination test (RSAT; RBPT) (George and Carmichael, 1974; George and Carmichael, 1978), Tube agglutination test (TAT), 2- mercaptoethanol TAT (2ME-TAT), agar gelimmunodiffusion test (AGID), microplate agglutination (MAT) (Kimura et al., 2008) fluorescent antibody test (IFA) and ELISA (Alton et al., 1988; Carmichael, 1990; Currier et al., 1982). Although AGID test is highly sensitive, detection of false positives, the inability to detect early cases and the difficult interpretation of the formed precipitate lines are among the disadvantages of the test (Alton et al., 1988). RSAT is the most widely used screening test among veterinary clinics in the USA (Carmichael, 1990).  Although it is common and practical test, large number of false positive cases are detected by RSAT (Alton et al., 1988; Carmichael, 1990; Mateu-De-Antonio et al., 1994).  In this study, an “in house” ELISA was used to detect anti brucellae antibodies in serum samples taken from the dogs in Diyarbakýr Metropolitan Municipality Animal Care and Rehabilitation Center. Smooth and rough lipopolysaccharide antigens were used as antigens in ELISA.

Bacterial strains
 
B.canis M (-) and B.abortus S99 strains and positive  antisera for Smooth and rough strains were obtained from Harran University, Faculty of Veterinary Medicine, Department of Microbiology.
 
Reference and test sera
 
Serum samples were collected by using stratified random sampling method from the Diyarbakir Metropolitan Municipality Animal Care and Rehabilitation Center, where approximately 600 dogs were located.  Positive and negative control sera for B.abortus and B.canis were obtained from OIE Brucellosis Reference Laboratory. Weybridge, England. The equation used to collect samples was
   
Here
N=  Sample size.
Z=   Confidence coefficient (this coefficient is taken 1.96 for 95% confidence).
N=  Main mass size.
P=   The probability that the feature we want to measure is in the main mass (This ratio is taken as 50% because the study is multi-purpose).
Q=  1-P.
D=  Accepted sampling error (a 10% sampling error was received for the study). According to this formula, the number of samples to be collected was determined as 83.
 
Antigen production
 
Raw LPS isolation was made from  the test strains according to the methods described earlier (Yi and Hackett, 2000) with minor modifications. Briefly, B. canis M (-) and B. abortus S99 strains were cultured in tryptic soy agar supplemented with 5% foetal calf serum, harvested after 72 hours of incubation and inactivated by heat (one hour, 56°C). Inactivated bacterial suspensions were washed two times by PBS and centrifuged at 3500xg for 10 minutes. Resulted pellets were mixed with tri-reagent and it was left at room temperature for 10-15 minutes for a complete homogenization. Then chloroform was added to the mixture to create phase separation. The suspension was incubated for a further 10 minutes with rapid stirring in the vortex and then centrifuged at 12,000 g for 10 minutes. Thus, water and organic phases were separated. The water phases were collected and stored at -20°C for later use as the ELISA solid phase antigen.
 
Indirect ELISA
 
Each ELISA solid phase antigen to be used in the study was diluted in a  ratio  determined by checkerboard analysis in antigen-coated buffer solution (0.05M sodium carbonate (pH 9.6) and distributed to  96 well flat bottom maxisorp polystyrene plates (NUNC 692620) as100 µl except H11 and H12 wells, which are assigned as blanks. The coated plates were incubated at 4°C for 18-24 hours. The plates were then blocked for 2 hours with PBS solution containing 5% skim milk. The plates were washed 3 times with PBS containing 0.05% Tween 20 (PBST). Test and control sera were added at 1:100 dilution in PBS/T as primary antibody. The plates were shaken in the orbital shaker for 1 hour at room temperature. A / G recombinant protein conjugated with horseradishperoxidase (HRPO) was diluted in PBST containing 1% skim milk at the determined dilution and added to all wells as 100 µl after the washing steps. Following 1 hour incubation at room temperature, the plates were washed again 3 times with PBST and 100 µl of chromogenic substrate (2 µg ortho-phenylenediamine ve %0.03 H₂O₂  in 0.1 M citrate bufferpH 5.5)  were added.  After the plates were kept at room temperature for 15 minutes, 100 µl of 4 N H₂SO₄ was added to each well to stop the reaction and the absorbance values   were read at 490 nm with the plates automatic ELISA reader (VERSAmax 3.13/B2573). For each ELISA cut off value was calculated  as the mean of the OD values of the negatives plus 2 standard deviations. Sensitivity and specificity of newly developed ELISAs were determined following formula:
 

A newly developed indirect ELISAs for rough and smooth strains were  used to test dog brucellosis caused by smooth strains and B.canis. Antibodies against B.canis were observed in 9.6%  and 26.5% serum samples by RSAT and ELISA, respectively. While 6% of the samples were positive by B.abortus S99 RSAT, 9.6% of the serum samples were found as positive by B.abortus ELISA (Table 1). A wide variety of ELISA procedures have been described that use various types of antigens of B.canis (Barouin-Melo et al., 2007; Lucero et al., 2002; Nielsen et al., 2004; Nielsen et al., 2007; Oliviera et al., 2011; Oncel et al., 2005). Nielsen et al., (2007) used rough lipopolysaccharides (RLPS) isolated from  B. abortus RB 51 strain as antigen when they tested B.canis-infected dog sera and they found the sensitivity and specificity of the test as 95.8% and 100%, respectively (Nielsen et al., 2004). Researchers used horse radish peroxidase (HRPO) labeled A/G protein as conjugate in their test.In another study, HRP labeled protein A/G conjugate was reported to increase sensitivity and specificity of the test (Nielsen et al., 2007). Barrouin-Melo et al., 2007) used heat-soluble bacterial extract (HSE) as antigen in their ELISA and they  found sensitivity and specificity as 95% and 91%, respectively. In our study, sensitivity and specificity of B.abortus S99  ELISA were 98% and 95.5%, respectively. On the other hand, sensitivity and specificity of B.canis M(-) ELISA were found as 96.6% and 94%, respectively. Similarities and differences in specificity of ELISAs are largely depend on degree of purity of the antigen used while it seems that this factor do not play major role in sensitivity of the test. The humoral response to smooth strains and rough strains, except for only 2 serum samples, did not overlap with each other. Only 1 serum from paddock C and 1 serum from paddock B responded positively with both ELISA. The central (core) oligosaccharides in the LPS layer are similar in Rough and smooth LPS structures and therefore, it is considered usual to see such cross reactions in ELISA (Alton et al., 1988; Carmichael, 1990). It has been determined that brucellosis caused by B.canis is more common than smooth brucella species. Dogs may be infected with smooth brucella species by being exposed to aborted cattle, sheep and goat fetuses in farms. Unfortunately, to feed dogs and cats with aborted fetuses in the farm is a common practice among farmers. Because the dogs in shelters are stray dogs living in the streets of Diyarbakýr and its districts, it could be said that they have more B.canis infection compare to smooth species, which are more likely to be transmitted in rural areas. As a result, it was concluded that B.canis infection is common in dogs and this should be considered as concern for public health. Unfortunately, there is no standardised commercial serological test for diagnosis B. canis and other rough Brucella species currently. This is the main reason that rough brucella infections overlooked and undiagnosed most of the cases. For this reason it is utmost important that these tests be used routinely in the hospitals and veterinary clinics in the serologic diagnosis of brucellosis.
 

Activities for the control of disease should be followed strictly, in terms of testing quarantine and testing. Elimination of the infected by separating serological examinations at certain intervals and trying to prevent the transmission of the disease by taking the necessary hygienic measures are the main strategies in the control of the disease (Pickerill, 1970). In this regard, our newly developed in house ELISAs for both smooth and rough strains (mainly B.canis) showed relatively good sensitivity and specificity. It was concluded that  ELISA together with RSAT can be used as easy and practical tests for serological diagnosis of dog brucellosis. Also dog owners should be awere of the possibility of transmission of the B.canis from infected dogs. For this reason owners should be told that their dogs should be tested regularly for this disease.