Brucella is a genus of Gram-negative facultative intracellular pathogens that infect humans and a variety of animal species (Shevtsov et al., 2015). Brucellosis is transmitted to humans following direct contact with infected animals or indirectly through the consumption of unpasteurized animal products (Deshmukh et al., 2015). The disease causes severe morbidity in humans and results in many medical challenges globally, especially in poor regions (Corbel, 1997; Pappas et al., 2006; Ducrotoy et al., 2015). Furthermore, animal brucellosis causes abortion and infertility in livestock, resulting in serious economic losses (Araj, 1999).

Since 2000, the incidence of brucellosis has increased rapidly in many regions in China (Deqiu et al., 2002; Jiang et al., 2011). The incidence of brucellosis in Inner Mongolia ranks in the top three incidence rates in Chinese provinces. The incidence of brucellosis in Ulanqab was 177.1/100,000 cases in 2011, 42.2/100,000 cases in 2012, 17.2/100,000 cases in 2013, 24.1/100,000 cases in 2014, and 26.5/100,000 cases in 2015. Ulanqab is a relatively small region in Inner Mongolia, covering a geographical area of 545,000 km², and comprises 11 regions that are extensively pastoral and semi-pastoral. The people that inhabit these regions are economically dependent on small ruminant (sheep) livestock (Jiang et al., 2013). Livestock exchange occurs frequently in this region, and includes import from other areas, internal exchange, livestock foster care, and shifting field grazing. Due to excessive poverty and limitations in relation to control measures, brucellosis is endemic among sheep and humans in these regions. Technical difficulties have meant that previous studies only concentrated on Brucella infection surveillance in this region, and adequate attention was not paid to the molecular typing of Brucella species. However, rapid and accurate identification and typing procedures are important for epidemiologic surveillance, investigation of outbreaks, and control program follow-up (Whatmore et al., 2006; Al Dahouk et al., 2007; García-Yoldi et al., 2007). Previous studies have confirmed that MLVA (Multiple-locus variable-number tandem repeat analysis) is a useful tool for identifying and genotyping Brucella isolates and the resultant data can be used for genetic diversity analysis and epidemiology trace-back investigations (Whatmore et al., 2006; Kattar et al., 2008; Ferreira et al., 2012; Garofolo et al., 2013). In this study, the MLVA-16 scheme was used to type samples and determine the genotyping characteristics and epidemiological links of associated strains. The results of this study contribute to our understanding of the main transmission routes associated with this pathogen as well as effectively informing brucellosis control policies in Ulanqab.

In the present study, we used MLVA methods to genotype Brucella isolates from Ulanqab. A total of 116 Brucella isolates were examined by classical biotyping methods and MLVA typing. All isolates were B. melitensis (bv. 1 and bv. 3), and B. melitensis bv. 3 was the predominant biovar in this region. These data were in accordance with previous studies where human brucellosis was almost exclusively caused by B. melitensis bv. 3 in China (Deqiu et al., 2002). Furthermore, nearly all of the associated patients had a history of exposure to sheep. The results reveal that human brucellosis in Ulanqab is more related to infections in sheep than other livestock. Contact with infected domestic animals (sheep) has been identified in another study as the most likely source of infection (Hasanjani Roushan and Ebrahimpour, 2015). Since 2011, S2 vaccine immunization programs have been implemented for sheep for 5 consecutive years in Inner Mongolia. However, human brucellosis continues to occur with a considerable amount of B. melitensis infections diagnosed. Consequently, stringent animal quarantine measures could be a promising control strategy for human brucellosis in this region.

A total of 76 (65.5%) B. melitensis strains were isolated from the Qianshan area of Ulanqab. However, the incidence of brucellosis in Houshan, the region directly to the north of Ulanqab, is significantly higher than that in Qianshan districts. This may be related to the underreporting of cases in low-incidence areas. However, the most important reason relates to the fact that the Qianshan region is close to our hospital, and it is convenient for these individuals to attain a diagnosis and related treatment regimens.

Based on the MLVA-16 marker, our results clearly show that B. melitensis, despite the high genetic homogeneity within the genus, is highly polymorphic at the microsatellite level. The complete MLVA-16 profile offers excellent discriminatory power; the Hunter-Gaston index was 0.985. Panel 2B markers displayed very high discriminatory powers. This result is consistent with previous reports (Jiang et al., 2011; Xiao et al., 2015). With regard to MLVA-8 genotypes, the most frequently observed genotype in this study, genotype 42, was also observed in other regions of China (Jiang et al., 2011; Ma et al., 2016; Sun et al., 2016). MLVA-8 and MLVA-11 yielded three and eight genotypes, respectively; whereas the addition of panel 2B increased the number of genotypes to 69. These findings show that the genotypic variation of Ulanqab isolates was predominantly associated with the highly variable panel 2B loci and to a much lesser extent panel 2A (locus Bruce19 and 21) and panel 1 (loci Bruce 43) loci. In agreement with a previous report (Kilic et al., 2011), this may reflect microevolution through a stepwise mutational event of the most variable loci from a very limited number of ancestors. In addition, there was no clear relationship between biovar and genotype within this B. melitensis population. This finding is in agreement with previous studies that used multiple VNTR typing approaches to show that there was no discernible correlation between genotype and biovars (Whatmore et al., 2006). This suggests either that B. melitensis biovars do not adhere to specific genetic groups, or that biotyping is so subjective that field strains are often incorrectly assigned masking any potential genetic relationship (Whatmore et al., 2007).

The phylogeographic relationships of the 116 B. melitensis isolates from Ulanqab were compared to 528 MLVA profiles from the international database using MLVA-11. A minimum spanning tree analysis revealed the presence of eight different genotypes; all of the strains belonged to the East Mediterranean lineage. Distribution of the MLVA-11 genotypes was consistent in most of the provinces in China. Genotype 116 is the predominant genotype and is likely to have been historically present in this region. Moreover, this genotype was also responsible for the vast majority of Brucella infections in China. The four novel genotypes (342–345) represent single strains. Upon comparison with the MLVA international database we observed that the four novel MLVA-11 genotypes (342–345) had been observed for the first time in Ulanqab. No isolates from America and West Mediterranean lineages were observed in this study.

The MLVA-16 genotyping clustering rate was 60.3%, with most isolates differing in one to three loci in panel 2B. The proportion of strains being in clusters suggest that a significant proportion of brucellosis in Ulanqab is due to multiple contaminations from a single source. Four large clusters (genotypes 6, 18, 43, and 45) including strains from different regions. These data show that ongoing transmission of human brucellosis not only in a specific region but also among the regions in Ulanqab. Additionally, the determination that 46 isolates showed distinct genotypes reflected that more than 39.7% (46/116) of the brucellosis cases in Ulanqab had epidemiologically unrelated sporadic characteristics.

Previous studies have confirmed that MLVA-16 genotyping results show good correlation with epidemiological data with epidemiologically related isolates displaying identical or very closely related genotypes (Al Dahouk et al., 2007; Nöckler et al., 2009; Kilic et al., 2011; Allen et al., 2015) The current study also confirms the occurrence of multipoint outbreak epidemics and intrafamilial brucellosis. The 11 shared genotypes observed during this analysis pertained to 33 Brucella strains obtained from the same area. This suggested the occurrence of a multipoint outbreak epidemic from multiple common sources. Meantime, MLVA genotyping also confirmed intrafamilial brucellosis outbreaks in two cases (genotypes 22 and 26), where brucellosis most probably resulted from contact with infected sheep. Other nine shared genotypes confirms the occurrence of multipoint outbreak epidemics. This source of infection also resulted in a higher incidence of brucellosis in other regions of China (Xiao et al., 2015; Ma et al., 2016). In accordance with the previous studies (Kilic et al., 2011), two isolates (genotype 7, ws049; genotype 16, ws050) with different MLVA-16 genotypes were obtained from members of the same family are likely to have resulted from either persistent circulating strains causing sporadic infections or mutation events during the course of the outbreak.

Two isolates from human blood exhibiting identical genotypes were obtained 13 days apart from the same patient, suggesting the occurrence of a singular B. melitensis infection. This occurrence is consistent with previous research findings (Kasymbekov et al., 2013). These data confirmed that brucellosis bacteremia persist for at least 13 days (Memish et al., 2000; Pappas et al., 2006; Vrioni et al., 2008). The previous investigations confirmed that strains of low virulence (e.g., Brucella abortus S19) are almost always cleared from the bloodstream within a week, whereas virulent strains produce a bacteremia that regularly persists for 4 weeks and may last much longer (Cruickshank, 1957).

Eighteen shared genotypes (5, 6, 7, 10, 14, 15, 17, 18, 24, 25, 27, 35, 36, 43, 45, 53, 61, and 67) including the 58 clustered strains were obtained from two or more regions and displayed identical MLVA-16 profiles. MLVA-16 genotyping confirmed that more than half of the human brucellosis cases in this region resulted from either very close cross-transmission in a singular location or ongoing transmission among different regions. This is likely to have occurred as a result of the frequent trade and exchange of sheep in the associated regions (Garofolo et al., 2013; Shevtsov et al., 2015). In future, a more precise investigation of the strains circulating in animal reservoirs in these regions will be required to confirm this result.

When the strains observed during this analysis were compared with those from neighboring provinces of Inner Mongolia, 10 identical MLVA-16 genotypes were recorded, suggesting possible epidemiological links. Additionally, one strain was isolated from a microbiology laboratory technician in Beijing hospital in 2011 (Jiang et al., 2011). This isolate harbored an identical MLVA-16 profile with two isolates (ws031 and ws037) from Ulanqab. This result suggests that Ulanqab might be the source of infection. It should be noted that isolate ws097 from Ulanqab displayed an identical MLVA-16 profile with strains from four different provinces in China, Inner Mongolia, Shanxi, Hebei, and Guangdong, respectively. This result may be explained by the fact that Inner Mongolia is adjacent to northern Shanxi and Hebei province, and frequent livestock exchange occurs between these areas. These findings reflect either a lack of control in the movement of infected animals between regions or the circulation of infected animal products in the market. This suggests that inspection mechanisms and measures to ensure the quarantine of livestock exported from endemic regions should be more stringent.

In the present study, 116 B. melitensis isolates were genotyped and a comprehensive global genetic analysis was performed. All of the strains belong to the East Mediterranean lineage. We also observed that there was a broad genetic relationship among the analyzed population. Thus, we conclude that the MLVA-16 assay appears to be a very practical and important molecular genotyping tool that is capable of confirming epidemiological linkages in outbreak and trace-back investigations. This study will help to improve the effectiveness of brucellosis control programs.

ZL performed most of the strain isolations and MLVA typing. ZL also drafted the manuscript; DD performed the MLVA cluster analysis and manuscript revision; MW and RL were in charge of epidemiological investigations and data analysis; HZ, GT, and DP prepared the DNA samples; HJ participated in the design of the study and critically reviewed the manuscript; BC, WF, and XX participated in the design of the study and also managed the project. All authors read and approved the final manuscript.

This study was supported by the Medical and Hygiene Research Projects pertaining to the Inner Mongolia Health and Family Planning Commission (201301094), the National Nature Science Foundation (81271900), the Science and Technology Basic Work Program (2012FY111000), and the public welfare research special projects of Health and Family Planning Commission of China (201302006). The funders contributed to the study design and data collection.