Background

Brucellosis in cattle, which can be caused by B.abortus, B. melitensis and B. suis [1], can significantly impact productivity on beef and dairy farms, and it poses zoonotic risk to humans, in whom infection can cause severe illness. The last reported B. abortus infections in cattle in Croatia occurred in 1964, while B. melitensis infections in cattle were reported in 2008 in herds kept with infected sheep [2] and in 2019 in herds kept with infected goats [3]. Since 2011, Croatia has conducted a brucellosis eradication programme in cattle according to European Directive 64/432/EC. All animals older than 12 months are tested annually using the Rose Bengal test (RBT), and positive animals are further tested using complement fixation (CFT), as well as indirect and competitive ELISAs. Depending on the epidemiological situation, seropositive animals are tested using a brucellin skin test, tissues and organs (head, mammary and genital lymph nodes, uterus, spleen, udder, fetal membranes) and stomach contents, spleen and lung from foetuses are collected at slaughterhouse for bacteriological testing.

Within this eradication programme, potential Brucella isolates are identified to genus and species levels using a combination of classical biotyping, multiplex PCR and molecular genotyping methods, which can include Multi-Locus Variable number of tandem repeats Analysis (MLVA), Multi-Locus Sequence Typing (MLST) [1, 4,5,6,7] and Whole-Genome Sequencing (WGS).

Here we describe the identification of a dairy cow herd with brucellosis within the framework of the Croatian eradication programme. The disease was attributed to infection with B. melitensis bv. 3, except in one cow infected with a B. melitensis variant difficult to identify using standard classical and molecular methods. The emergence of this novel strain points to ongoing Brucella evolution in the western Balkans area, which may be due to the appearance of new reservoirs or vectors forced strain mutation and poor efficacy of eradication measures. Future studies should explore new reservoirs and zoonotic significance for this and other potential new Brucella variants in the region.

Case presentation

During routine annual testing within the national eradication programme, a dairy cow herd (12 cows, 7 heifers and 4 calves) with brucellosis was identified in October 2018 on a farm in the Croatian village of Katinovac (45°14′30.6″N 15°55′31.8″E), close to the northern border with Bosnia and Herzegovina. At the time of testing, animals were healthy and showed no clinical signs of brucellosis. Other animal species were not present on farm. A total of 19 animals aged 1 year and older were tested using the RBT. A total of 10 animals were identified positive by RBT and indirect ELISA, and 7 of these animals were also positive by the CFT and competitive ELISA. In November 2018, 19 animals older than 1 year were tested on the brucellin skin test. Positive reaction on brucellin was found in 5 previously seropositive cows, which were sent to the slaughterhouse, where samples were collected and analysed bacteriologically. During the post mortem inspection visible lesions were not recorded.

Epidemiological investigation showed that the Croatian herd had been brucellosis-free since 2013, and no new animals had been introduced since 2008. Since 2016, the farm practised artificial insemination with no recent history of abortions. The farm owner denied contact with other sheep or cattle herds and indicated that the herd was kept on pastures bordering Bosnia and Herzegovina. Their water source was the river Glina, a natural borderline in this area.

Brucella strains were isolated from milk and various tissues from three cows that were found without any clinical symptoms. The strains were classically biotyped as described [4] based on CO2 requirement, H2S production, oxidase and urease activity, growth on dyes, lysis by phages and agglutination with monospecific sera (Tables 1 and 2). This biotyping was performed at the Croatian National Reference Laboratory (Zagreb) and European Reference Laboratory (Maisons-Alfort, France). All three isolates could grow without CO2, they produced H2S and expressed oxidase, and they hydrolysed urea. They did not grow on basic fuchsin medium, and they triggered agglutination of anti-A and anti-M sera. These findings are consistent with B. melitensis bv.3. However, isolate 7 was lysed by Tbilisi phages at 104 routine test dilution (RTD).

Table 1 Results of classical biotyping tests of Brucella isolates
Full size table
Table 2 Molecular identification and genotyping of Brucella isolates
Full size table

Species was determined using multiplex PCR based on the Bruce ladder [8] and the “AMOS” method [9], followed by the Suis ladder [10] and another PCR based on detection of the omp31 gene [11]. Isolates 6 and 11 gave results consistent with B.melitensis, but isolate 7 lacked omp31 gene sequences tested in the Bruce ladder, suggesting that it was B.abortus.

To confirm the identification of isolates 6 and 11 as well as to complete the identification of isolate 7, we performed MLVA based on 16 loci [5, 6] in the following order: panel 1, Bruce06 - Bruce08 - Bruce11 - Bruce12 - Bruce42 - Bruce43- Bruce45 - Bruce55; panel 2a, Bruce18 - Bruce19 - Bruce21; and panel 2b, Bruce04 - Bruce07 - Bruce09 - Bruce16 - Bruce30. B. melitensis 16M was used as the reference strain for comparison and verification of test quality. In addition, MLST was performed based on 9 loci [7]: gap - aroA - glk - dnaK - gyrB - trpE - cobQ - int-hyp (orf1)-omp25.

Moreover, isolates 6 and 7 were subjected to whole-genome shotgun sequencing using the Illumina NexteraXT system (protocol 150,319,425,031,942, revision C), which has been deposited in DDBJ/ENA/GenBank under accession numbers CVI_6 ChI CP058599/CVI_6 ChII CP058600 and CVI_7 ChI CP058597/CVI_7 ChII CP058598. A phylogenetic tree was generated using Bioumerics 7.6.3 (Applied Maths, BioMérieux). A set of B. melitensis genomes was retrieved from public databases (NCBI and PATRIC) and numbered in table (see Additional file 1). Sequencing reads were simulated for each genome using ART and all reads were mapped against a chimeric genome of B. melitensis 16M genome. Eight B. abortus genomes were used as outgroup. The SNPs obtained were then filtered (20X of absolute coverage, 10 bp inter-SNP distance, ambiguous and unreliable bases were removed, repeated elements removed) and a maximum parsimony tree was generated from these SNPs. The tree is represented with a logarithmic scale (see Additional files 2 and 3). This sequencing also revealed that isolate 7 contained a 609-bp omp31 sequence also present in isolate 6, but not the 723-bp omp31 sequence present in isolate 6. Based on classical and molecular methods, we assigned Brucella isolates 6 and 11 as B. melitensis bv. 3, while isolate 7 appeared to be a novel B. melitensis variant.

Discussion and conclusions

Human brucellosis, considered one of the most dangerous zoonoses, is most often caused by B. melitensis and less often by B. abortus or B. suis. The disease is endemic to the Mediterranean in general and the Balkan peninsula in particular [12,13,14,15,16]. Nevertheless, bovine brucellosis cases are sporadic and infrequent in Croatia, with the only recent reports limited to instances of transmission from sheep and goats on the same farms [2, 3]. The present case is the first recent report of brucellosis in Croatia that cannot definitively be attributed to contacts with other infected animal species.

This work highlights the need for continuing vigilance and research into potential Brucella reservoirs and spreading pathways. The disease in the present report may easily have come from Bosnia and Herzegovina, because herds on both sides of the border often share pastures, and illegal migrations are common which is documented throughout complete border line between countries [2, 3, 17]. Bosnia and Herzegovina has conducted a vaccination programme to control brucellosis in small ruminants since 2009, yet incidence of the disease remains high in animals and humans [18], and has even been increasing since 2012 [3]. This lack of efficacy is likely due largely to non-compliance with vaccination programmes [18], which can also foster the emergence of new Brucella strains [19].

We were unable to identify isolate 7 using classical microbiological methods [4] which are based on phenotype. This suggests that classical methods may not be well suited for characterising new B. melitensis strains in brucellosis-endemic regions. In fact, we were able to unambiguously identify the three isolates only by combining MLVA, MLST and whole-genome sequencing. These techniques showed our strains to be phylogenetically related to strains circulating in Croatia as well as Bosnia and Herzegovina [15, 17]. In particular, MLVA typing allowed us to assign a unique 16-digit code to the novel isolate 7, based on differences from the Bruce42 locus. Isolates 6 and 7 were assigned to the previously reported sequence type 8, related to B. melitensis strains circulating in Turkey, Kosovo and Macedonia. The two strains CVI_6 and CVI_7 clustered together in a subclade comprising 4 others strains (F9/05 from Turkey, BwIM_XXX_12 from unknown origin, F8/01–155 from Kosovo and BwIM_ALB_46 from Albania). Interestingly, this subclade contains two strains from the Balkan and one from Turkey. Moreover, in a recent paper [20], a strain from Serbia is clustered with the strain of Albania. The 2 strains isolated in this study seem to belong to a clade composed by strains that circulate in the Balkan area (see Additional files 2 and 3). Cross-contamination at the borders with animals can be a reason.

Our findings highlight the need for continuing, even enhanced, efforts to surveillance brucellosis in domestic animals and to research potential Brucella reservoirs and transmission pathways to ensure timely detection of zoonotic threats.