Differentation of Brucella abortus and Yersinia enterocolitica serotype 09 infections in cattle: The use of specific lymphocyte transformation and brucellin skin tests

Summary

The lymphocyte transformation test (using an in vitro whole‐blood lymphocyte stimulation procedure) and the Brucellin skin test were applied to five heifers infected with virulent Brucella abortus strain 544, five cows inoculated with Yersinia enterocolitica serotype 09, and four non‐exposed cows. Lymphocytes from Brucella‐inoculated animals persistently gave very high blastogenic reactions indicative of active Brucella infection. The test was persistently negative in Yersinia‐infected and non‐exposed cattle. Four of thefive cowsinfected with Yersinia enterocolitica type 09 and allfour control cattle were persistently negative to the delayed hypersensitivity skin reaction with brucellin. All cattle infected with Yersinia enterocolitica type 09 were strongly positive to the Rose Bengal, Serum agglutination, Complement fixation and Antibovine globulin tests using Brucella abortus antigens. One lactating cow infected with Yersinia enterocolitica type 09 was positive to Brucella milk ring test. These results indicate that standard Brucella serological tests are unreliable in differentiating the two infections in cattle and that both the Lymphocyte transformation and brucellin skin tests could be used to differentiate bovine brucellosis from yersiniosis.     

Serologic response to Brucella abortus of cattle and guinea pigs experimentally inoculated with a Yersinia enterocolitica serotype from milk

Ten strains of Yersinia enterocolitica belonging to ten various serogroups isolated from raw milk were inoculated into groups of five guinea pigs and five calves. Y. enterocolitica serotype 0:16 was the only serotype tested that induced an antibody response to Brucella abortus in calves. No anti-Brucella response could be demonstrated serologically in guinea pigs. Activity of the anti-Y. enterocolitica 0:16 calf sera against B. abortus antigen was shown by the tube agglutination test, and by the complement fixation test. The early agglutinating antibody response was partly sensitive to reduction by 2-mercaptoethanol. This sensitivity decreased later in the response. This is the first report of anti-Brucella responses induced by a serotype of Y. enterocolitica other than 0:9; sera from a group of five calves inoculated with 0:9 were tested by the same serological techniques for comparison.     

Cell-mediated immune responses differentiate infections with Brucella suis from Yersinia enterocolitica serotype O:9 in pigs

Due to almost identical lipopolysaccharide (LPS) O-antigens, infections with Yersinia enterocolitica serotype O:9 (YeO:9) cause false positive serological reactions (FPSR) in tests for Brucella and thus cause problems in National Brucella surveillance programs. As LPS are strong inducers of antibody responses it was hypothesized that cell-mediated immune responses to non-LPS antigens of the two bacteria can be used to separate immune responses to these two biologically very different infections. Following subclinical experimental infections with Brucella suis biovar 2, high interferon-gamma (IFN-gamma) assay responses with a commercial Brucella melitensis antigen preparation (Brucellergene OCB) preceded the development of antibodies. High IFN-gamma responses in the seven B. suis inoculated pigs with serological evidence of infection were consistent throughout a 20-week post-inoculation observation period. In contrast, IFN-gamma responses in two B. suis inoculated pigs without bacteriological or serological evidence of infection were below a cut-point of 25pg/ml at all samplings. IFN-gamma responses in repeated samplings from 5 uninfected control pigs and 18 pigs experimentally infected with YeO:9 were all negative, except for solitary false positives in 3.7% of the samples from both the experimentally YeO:9 infected pigs and control pigs. Skin tests using the same commercial Brucella antigen confirmed the ability of cell-mediated immune responses to differentiate between the two infections. In addition, a field evaluation of the diagnostic use of cell-mediated immune responses by IFN-gamma assay and skin test to resolve serological suspicions of Brucella was conducted in an YeO:9 infected pig herd. Following a screening of 200 pigs 39 pigs were identified with false positive serological Brucellosis reactions. While 36 of the 39 FPSR pigs were also FPSR in a second test, none of the pigs were test positive in whole blood IFN-gamma assay or Brucellergene OCB skin test. In conclusion, use of IFN-gamma assay and skin test as measurements of cell-mediated immune responses to non-LPS Brucella antigens were specific and sensitive in discriminating subclinical experimental infections with B. suis from both natural and experimental infections with YeO:9.     

Serological cross-reactivity between Brucella abortus and Yersinia enterocolitica 0:9:: IV. Evaluation of the M- and C-epitope antibody response for the specific detection of B. abortus infections

Smooth lipopolysaccharides (SLPS) from Brucella abortus contain A-epitopes against which the majority of serum antibodies are directed during infections. SLPS from Yersinia enterocolitica 0:9 possesses identical epitopes, which are the cause for serological cross-reactivity. All Brucella spp. possess M- and C-epitopes which are not present in Y. enterocolitica 0:9. In order to examine the usefulness of these M- and C-epitopes for discrimatory serological testing, a panel of sera were used in this study, comprising sera from Y. enterocolitica 0:9-infected heifers, sera from B. abortus-infected cattle of comparable strength in the serological brucellosis tests to the sera from Y. enterocolitica 0:9-infected heifers, sera from B. abortus-infected bovines with strong serological reactions and sera from animals free from B. abortus or Y. enterocolitica infections. These sera were tested in blocking ELISAs with seven M- and one C-epitope-specific monoclonal antibodies in combination with SLPS from B. melitensis M16 high in M-epitopes as antigen. Strong B. abortus sera inhibited most strongly, while negative sera showed no or little inhibition. Sera with weak or intermediate titres blocked to a lower extent. Unexpectedly, the sera from Y. enterocolitica 0:9-infected heifers showed inhibition behaviour virtually identical to the comparable sera from B. abortus infected animals. Absorbing out of the A-epitope specific serum antibodies with either Y. enterocolitica 0:9 SLPS or with Y. enterocolitica 0:9 bacteria, indicated the presence of M- or C-epitope-specific serum antibodies in some sera from B. abortus-infected cattle but not in the sera from Y. enterocolitica 0:9-infected animals. These results demonstrate that the M- or C-epitope-specific antibody response in sera from B. abortus infected cattle is only of limited value for the serological discrimination between B. abortus and Y. enterocolitica 0:9 infections.     

Experimental studies on the serological relationships between Yersinia enterocolitica and Brucella abortus.

Yersinia enterocolitica O9 was shown to provoke O, OH and H antibody response in calves. Brucella abortus failed to generate Yersinia H antibody response and generated Brucella O antibodies that cross-reacted with Yersinia O and OH antigens. The presence of Yersinia H agglutinins along with a higher titre of Yersinia OH antibody than Brucella O antibody is suggestive of subclinical infection with yersiniosis rather than brucellosis. Cross-absorption of sera from calves with established Yersinia infections indicated that absorption of sera with Brucella O antigens, although completely removing Brucella O antibodies, failed to remove completely Yersinia O, OH and H antibodies, and thus provides an additional method of distinguishing between the two infections.     

The serological response of cattle immunized with Yersinia enterocolitica O:9 or O:16 to Yersinia and Brucella abortus antigens in enzyme immunoassays

Sera from calves immunized with Yersinia enterocolitica serotypes O:9 or O:16 were tested by indirect enzyme linked immunosorbent assay (ELISA) using lipopolysaccharide (LPS) preparations from Brucella abortus or Y. enterocolitica O:9 or O:16 for their antibody content of the IgG1 or IgG2 subclasses. High IgG1 responses were present with the three antigens in both groups although some individual variations between animals were noted. The IgG2 responses were modest and in some cases not above background 'noise'. Thus IgG2 antibody was not measurable in sera from serotype O:9 injected calves when using serotype O:16 LPS or in serotype O:16 injected calves when using B. abortus or serotype O:9 LPSs. A competitive ELISA using B. abortus O-polysaccharide and a monoclonal antibody to B. abortus LPS (initially designed to differentiate the antibody responses of cattle naturally infected with B. abortus from those vaccinated with strain 19) was used on sera from both groups of calves. Using this test, no antibody was detected in the group immunized with serotype O:16 and except for one animal in the serotype O:9 immunized group, only low levels of antibody were transiently in evidence. One animal in this group responded with quite high levels of competing antibody which, however, declined towards the end of the test period. The competitive ELISA may prove a useful serological tool for differentiating vaccinal and field infection titers to B. abortus and also to eliminate cross-reactions observed with Y. enterocolitica serotypes.     

Brucella titers in swine caused by Yersinia enterocolitica, serotype 0:9

Antibody to Brucella abortus developed in two thirds of all gilts kept on a pig breeding station. Systematic tests taken for the purpose of detecting clinical symptoms and of isolating Brucella were negative. however, Yersinia (Y.) enterocolitica, Serotype 0:9, was cultured from rectal swab samples which had been obtained from 31 to 78 gilts. The clinical, bacteriological, and serological tests gave rise to the assumption that the Brucella titres have been caused by Yersinia enterocolitica infection. Such conclusion, however, could be drawn only as a result of a complex investigation. Detection of Yersinia enterocolitica alone is not sufficient a proof by which to rule out the possibility of concomitant Brucella infection. The question is discussed to what extent swine may be considered to be a potential source of infection of man.     

Levamisole potentiation of antigen specific lymphocyte blastogenic response in Brucella abortus exposed but nonresponsive cattle

Incubation of Brucella abortus (field strain) infected and strain 19 vaccinated bovine peripheral blood lymphocytes with B. abortus antigen and levamisole caused a consistently significant increase in [3H] thymidine uptake when compared to cultures without levamisole. Levamisole did not potentiate B. abortus-induced blastogenic response of lymphocytes from non-exposed cattle. A dose response study showed that 10 micrograms/culture induced maximum potentiation of B. abortus-induced lymphocyte stimulation. Using the 10 micrograms/well concentration of levamisole, further studies were conducted to determine the net potentiation of the blastogenic responses in lymphocytes from B. abortus (field strain) infected cattle. B. abortus strain 19 vaccinated but nonresponsive and non-exposed cattle. Levamisole significantly potentiated the B. abortus-induced lymphocyte blastogenesis in lymphocytes from unresponsive cattle.     

Specific lymphocyte stimulation in cattle naturally infected with strains of Brucella abortus and cattle vaccinated with Brucella abortus strain 19.

Cell-mediated immune responses in cattle naturally infected with strains of Brucella abortus and in cattle vaccinated with B abortus strain 19 during calfhood were studied by an in vitro lymphocyte-stimulation procedure. Lymphocytes were prepared from peripheral bovine blood by the Ficoll-diatrizoate technique, suspended in RPMI-1640 medium (1.5 X 10(6) lymphocytes/ml), cultured with B abortus-soluble antigen or phytohemagglutinin, and incubated for 6 days. Sixteen hours prior to termination of incubation, cultures were labeled with 1 muCi of [3H]thymidine (3HdT) and, after harvesting, assayed for 3HdT incorporation into DNA by liquid scintillation spectrometry. Lymphocytes from cattle with bacteriologically confirmed isolation of B abortus underwent a significantly higher lymphocyte stimulation with B abortus-soluble antigen than did cattle vaccinated with B abortus strain 19 during calfhood (P less than 0.005). Standard seroagglutination tests were conducted simultaneously with lymphocyte-stimulation tests, but there was no apparent correlation between levels of humoral antibodies and the cell-mediated immune responses as measured by in vitro specific lymphocyte stimulation.     

Serological relationship between cattle exposed to Brucella abortus, Yersinia enterocolitica O:9 and Escherichia coli O157:H7

Sera from cattle naturally infected with Brucella abortus (n = 160), vaccinated with B. abortus S19 (n = 88) or immunized with Yersinia enterocolitica O:9 (n = 25) or Escherichia coli O157:H7 (n = 80) were collected. The sera were compared for antibody content to the same bacteria by indirect enzyme immunoassay (IELISA), fluorescence polarization assay (FPA) and competitive enzyme immunoassay (CELISA). Cattle sera (n = 523) collected randomly from across Canada were tested in the same tests. Sera from the B. abortus infected group reacted positively in the brucellosis IELISA (IELISA(Br)), CELISA and FPA (FPA(Br)) and the Y. enterocolitica IELISA (IELISA(Ye)) while the Y. enterocolitica FPA (FPA(Ye)) detected antibody in 93.8% and the E. coli IELISA (IELISA(Ec)) 86.9% and the E. coli FPA (FPA(Ec)) 48.1%. About 70% of the sera from B. abortus S19 vaccinated animals reacted in the three IELISAs, 45% in the CELISA, and 37.7% in the FPA(Ec), 21.6% in the FPA(Br) and 5.7% in the FPA(Ye). Sera from E. coli O:157 exposed cattle reacted mainly in the IELISA(Ec) and FPA(Ec) although surprisingly 87.5% reacted in the IELISA(Ye) and only 3.8% in the IELISA(Br). No reactions were observed with these sera in the FPA(Br) and FPA(Ye) but one serum gave a low positive reaction in the CELISA. All sera from Y. enterocolitica O:9 exposed cattle reacted in the IELISA(Br) and IELISA(Ye) and 80% in the IELISA(Ec). In the CELISA, 44% gave a positive reaction and 64% were positive in the FPA(Br), 28% in the FPA(Ye) and 12% in the FPA(Ec). Of the 523 Canadian sera, about 50% reacted in the E. coli tests with only minor reactions in the Y. enterocolitica O:9 and B. abortus assays. From the data, the cross reaction between E. coli O157:H7, Y. enterocilitica O:9 and B. abortus is dependent on the test used. Thus, extensive cross reaction was observed with the IELISA with much less reactivity in the FPA and the CELISA.     

Studies on the relationship between Yersinia enterocolitica IX and Brucella abortus agglutinins in naturally infected animals

Attempts were made to define the Brucella abortus and Yersinia enterocolitica IX infection status of animal populations by means of selected agglutination tests. The Brucella abortus O, the Yersinia enterocolitica IX OH and the Y enterocolitica IX H agglutinin titres were measured in a large number of cattle, goat and pig sera In the goats and, to a much lesser extent, the pigs, the relationships between these titres suggested that Yersinia infection was common. In contrast, the results from the cattle sera were complex and tended to indicate the presence of both Yersinia infection and brucellosis.     

Serological crossreactivity between Brucella abortus and Yersinia enterocolitica 0:9 III. Specificity of the in vitro antigen-specific gamma interferon test for bovine brucellosis diagnosis in experimentally Yersinia enterocolitica 0:9-infected cattle

The course of immunological reaction in 10 Yersinia enterocolitica 0:9 experimentally-infected heifers was followed using the conventional brucellosis tests complement fixation test (CFT), serum agglutination test (SAT) and brucella card test (BCT), and a recently developed Brucella antigen-specific gamma interferon (IFN-γ) test. Initially, the animals were exposed orally to 10¹⁰ colony-forming units (CFU) of Y. enterocolitica 0:9. Four weeks later, they were inoculated intravenously with 10⁸ CFU of Y. enterocolitica 0:9 cells. After oral inoculation, the response in the conventional brucellosis tests was minimal. Only after intravenous inoculation were CFT and SAT titres and BCT reactions comparable to natural, false positive brucellosis reactors. After oral exposure the Brucellergen-stimulated release of IFN-γ peaked at values above the cut-off stimulation index of 2.5 in 80% of the heifers. After intravenous inoculation, stimulation indices above 2.5 were present in only 10% of the animals. Two B. abortus infected control cattle showed stimulation indices of 3.1 and 3.4, and a negative control animal exhibited a stimulation index of 1.0. These findings show, in contrast to a previous study, that the Brucellergen-specific IFN-γ assay cannot be used as a specific and discriminatory test for B. abortus infections.     

Non-specific seroreactions against Brucella abortus in ruminants in New Zealand and the presence of Yersinia enterocolitica 0:9

The level of non-specific reactions found in the brucellosis serology of ruminants in New Zealand was very low until July 1992. This changed when, in an export consignment of 1071 deer, 35% reacted in the Brucella abortus tube agglutination test with titres varying from 50 to 200 IU. The reactors were also positive in the Rose-Bengal agglutination test and most of them reacted in the complement fixation test with titres varying from 10 to 80 IU. Yersinia enterocolitica 0:9 was later isolated from one deer of this consignment. It was the first isolate of this serotype recovered in New Zealand from an animal. Shortly after, false reactors occurred more frequently than before in sera from Brucella abortus accredited free cattle herds. As the involvement of Yersinia enterocolitica 0:9 was suspected in these cases, faecal samples from reactors and in-contact animals were cultured for this organism. Yersinia enterocolitica 0:9 was isolated from nine of 19 herds showing one or more false Brucella abortus seroreactions.

Prior to 1990, Yersinia enterocolitica serotype 0:9 had not been isolated in New Zealand, despite the recovery of a number of other bio— or serotypes of the organism from humans and animals. From 1990 onward, serotype 0:9 began to be isolated from human faecal samples with increasing frequency. Since the first isolations from deer and cattle in 1992, it has now also been recovered from a cat and an alpaca and from cattle without any association with false positive Brucella abortus reactions. All serotype 0:9 isolates were of biotype 2.     

Studies on the use of Yersinia enterocolitica 09 antigens in rapid quantitative plate tests for the differentiation of Brucella and Yersinia infections in cattle and the diagnosis of bovine brucellosis

The use of the rapid quantitative plate agglutination test (QPAT) utilizing triphenyltetrazolium chloride (TTC) stained antigen for the differentiation of bovine brucellosis and yersiniosis is described. It was found with experimental laboratory animals and cattle, that titres against homologous antigen tended to exceed titres against heterologous antigen and that this reaction together with measurement of the Yersinia OH agglutinin titre could be used for purposes of differentiation. It was, however, not possible to differentiate these infections in serum from cattle with naturally-occurring antibodies. It was also shown that Brucella 0 agglutination in the QPAT was very weak in some sera with high titred anti-Brucella antibodies by serum agglutination, while Yersinia 0 agglutination was not. The QPAT, using Yersinia 0 antigen, therefore correlated well with the results obtained using Brucella 0 antigen in serum agglutination tests. It is suggested that Yersinia 0 antigen may be a suitable antigen for use in rapid agglutination tests because of this relative insusceptibility to false negative results.     

Non-specific seroreactions against Brucella abortus in ruminants in New Zealand and the presence of Yersinia enterocolitica 0:9

The level of non-specific reactions found in the brucellosis serology of ruminants in New Zealand was very low until July 1992. This changed when, in an export consignment of 1071 deer, 35% reacted in the Brucella abortus tube agglutination test with titres varying from 50 to 200 IU. The reactors were also positive in the Rose-Bengal agglutination test and most of them reacted in the complement fixation test with titres varying from 10 to 80 IU. Yersinia enterocolitica 0:9 was later isolated from one deer of this consignment. It was the first isolate of this serotype recovered in New Zealand from an animal. Shortly after, false reactors occurred more frequently than before in sera from Brucella abortus accredited free cattle herds. As the involvement of Yersinia enterocolitica 0:9 was suspected in these cases, faecal samples from reactors and in-contact animals were cultured for this organism. Yersinia enterocolitica 0:9 was isolated from nine of 19 herds showing one or more false Brucella abortus seroreactions. Prior to 1990, Yersinia enterocolitica serotype 0:9 had not been isolated in New Zealand, despite the recovery of a number of other bio- or serotypes of the organism from humans and animals. From 1990 onward, serotype 0:9 began to be isolated from human faecal samples with increasing frequency. Since the first isolations from deer and cattle in 1992, it has now also been recovered from a cat and an alpaca and from cattle without any association with false positive Brucella abortus reactions. All serotype 0:9 isolates were of biotype 2.     

Brucella clearance as a sensitive method for the detection of cross reactions of Brucella abortus with Yersinia enterocolitica 03, 06, 09 and Salmonella urbana and Salmonella abony

Determination of the Brucella clearance rate has proved to enable assessment of Brucella immune reaction in rat, even after vaccination with Yersiniae and Salmonellae. Vaccination with Yersinia (Y.) enterocolitica O6 and O9 produced 95 per cent of "high responders", whereas 65 per cent of "high responders" and 25 per cent of "non-responders" were recorded in the wake of O3. Salmonella (S.) urbana vaccination gave 50 per cent of "high responders" and 27 per cent of "non-responders", while 100 per cent "non-responders" resulted from S. dublin. Vaccination, using Brucella abortus Buck 19, gave 100 per cent "high responders". The differentiated nature of immune reactions to Y. enterocolitica O3, S. urbana, and S. abony has been attributed to an individual genetic capability of reaction to the cross-reactive antigen.     

Serological cross-reactions between different brucella species and Yersinia enterocolitica. Agglutinating activity of 19 S and 7 S antibodies aginst somatic antigen of Brucella abortus and Yersinia enterocolitica type IX

Antisera from rabbits immunized with acetone-killed whole cells and with lipopolysaccharides from Brucella abortus (B.a.) and Yersinia enterocolitica type IX (Y.e.) were gel-filtered on Sephadex G-200 columns.All the antisera had serologically cross-reacting agglutinins against B.a. and Y.e. both in the 19 S and in the 7 S antibody fractions.The homologous and the heterologous agglutinating activity of 19 S and of 7 S antibodies was tested before and after treatment with dithiothreitol (DTT) and subsequent alkylation. Unlike 19 S antibodies, 7 S antibodies were resistant to the DTT reduction.The results from heterologous absorptions of the respective 19 S and 7 S antibody fractions indicated that 19 S antibodies both to B.a. and to Y.e. had a greater tendency towards being absorbed to the cross-reacting antigenic determinants than had the corresponding 7 S antibodies. The agglutination titres for 19 as well as 7 S antibody fractions derived from the immunizations with B.a. (both whole cells and LPS) fell more markedly after heterologous absorptions than did the analogous titres for corresponding Y.e. antibody fractions. Possible explanations of these differences are discussed.     

Competitive exclusion of Yersinia enterocolitica biotype 4, serotype O:3 by Yersinia enterocolitica biotype 1A,serotype O:6,30 in tissue culture and in pigs

AIMS: To study the adhesion properties of a biotype 4, serotype O:3 (human pathogenic) strain of Yersinia enterocolitica and to determine if adhesion in vitro and colonisation in vivo can be prevented by competition with a biotype 1A, serotype O:6,30 (non-pathogenic) strain. To study interaction between Y. enterocolitica biotype 4, serotype O:3 and cultured epithelial cells using the synthetic tripeptide arginine-glycine-aspartic acid (RGD).

METHODS: The human intestinal epithelial (HEp-2) cell line was used for in vitro studies. Inocula of Y. enterocolitica biotype 4, serotype O:3 radiolabelled using tritium were incubated with HEp-2 cells and RGD tripeptide, or with Y. enterocolitica biotype 1A, serotype O:6,30 sequentially or concurrently, then washed and lysed, and radioactivity measured to determine the effect of RGD on adhesion, and competitive exclusion of pathogenic by non-pathogenic bacteria. For in vivo studies, two groups of 5-week-old piglets (n=5/group) were sequentially inoculated orally with 5x10⁹ colony forming units (cfu) of either a non-pathogenic biotype 1A, serotype O:6,30 strain of Y. enterocolitica followed by a pathogenic biotype 4, serotype O:3 strain, or vice versa. Pigs were monitored for carriage of strains using bacterial culture and a multiplex polymerase chain reaction (PCR).

RESULTS: The RGD tripeptide significantly inhibited adherence of the pathogenic Y. enterocolitica strain to cultured epithelial cells, suggesting that adhesion involved the RGD tripeptide sequence. The non-pathogenic biotype 1A, serotype O:6,30 strain of Y. enterocolitica prevented adhesion of the pathogenic strain to cells in vitro when allowed to adhere first. Pathogenic Y. enterocolitica was consistently isolated from rectal swabs from 80-100% of pigs on all sampling occasions but not from oral swabs after 14 days in pigs first inoculated with the non- pathogenic strain or at 26 days in pigs first inoculated with the pathogenic strain.

CONCLUSIONS: A non-pathogenic strain of Y. enterocolitica reduced adhesion of a human pathogenic strain in vitro but not in vivo.