Original and introduced lineages co-driving the persistence of Brucella abortus circulating in West Africa

Introduction Brucellosis, a serious public health issue affecting animals and humans, is neglected in West Africa (WA). Methods In the present study, bio-typing, multi-locus sequence typing (MLST), multiple-locus variable-number tandem repeat analysis (MLVA), and whole genome sequencing single-nucleotide polymorphism (WGS-SNP) analysis were used to characterize the Brucella abortus (B. abortus) strains from WA. Results All of the 309 strains analyzed in this study were extracted and downloaded from the international MLVA bank and were from 10 hosts (cattle, humans, ovine, buffalo, dromedaries, horse, sheep, zebu, dog, and cat) distributed in 17 countries in WA. Based on the bio-typing, three biovars, dominated by B. abortus bv.3, were observed and reported across seven decades (1958–2019). With MLST, 129 B. abortus strains from the present study were sorted into 14 STs, with ST34 as the predicted founder. These 14 STs clustered into the global MLST data into three clone complexes (C I–C III) with the majority of strains clustering in C I, while C II forms an independent branch, and C III harbors three STs shared by different continents. These data revealed that most cases were caused by strains from native lineages. According to the MLVA-11 comparison, 309 strains were divided into 22 MLVA-11 genotypes, 15 of which were unique to WA and the remaining seven had a global distribution. MLVA-16 analysis showed that there were no epidemiological links among these strains. Based on the MLVA data, B. abortus strains from WA have high genetic diversity, and predominated genotypes were descended from a native lineage. While the MLVA-16 globally highlights that the dominant native and few introduced lineages (from Brazil, the USA, South Korea, Argentina, India, Italy, Portugal, the UK, Costa Rica, and China) co-driving the B. abortus ongoing prevalence in WA. The high-resolution SNP analysis implied the existence of introduced B. abortus lineages, which may be reasonably explained by the movement and trade of dominant hosts (cattle) and/or their products. Discussion Our results indicated that B. abortus strains in WA consist of native and introduced strains that necessitate control such as vaccination, testing, slaughtering, and movement control by the relevant country authorities to reduce brucellosis in livestock.


Introduction
Brucellosis, a widespread bacterial zoonotic disease that can cause considerable suffering in humans and massive economic losses in the animal industry, is caused by Gram-negative facultative intracellular bacteria of the genus Brucella (1-3). The genus includes 12 species, among which Brucella melitensis, B. abortus, and B. suis are the most important species and responsible for the majority of human and animal brucellosis cases (4). It has been reported that there are ∼500,000 new cases of human infection annually, especially in low-income tropical countries (5). Cattle are natural hosts of the intracellular pathogen B. abortus, which can also be isolated in other hosts, such as sheep, camels, and horses, and imposes a significant burden on the health and reproduction of these important livestock (6,7). Infected animals can have live offspring following the initial abortion, and they may continue to shed the bacteria (8). Transmission to human occurs through unprotected handling of tissues or body fluids from infected animals, consumption of unpasteurized milk and milk products, or inhalation of Brucella-contaminated aerosols (9, 10). Although B. abortus causes less severe infections than B. melitensis (11), back pain and arthralgia are common symptoms in patients infected with B. abortus in Turkey (12).
A previous study reported that the annual economic burden of bovine brucellosis in Latin America is ∼$600 million (13). In 2015, in Kazakhstan, around $21 million was spent on compensation for slaughtered animals with brucellosis, and an additional $24 million was spent on testing and screening of animals (14). Therefore, brucellosis can not only lead to substantial economic losses in the animal industry but also pose an ongoing threat to public health.
Countries declared brucellosis-free are located in Europe and Oceania, whereas the prevalence is high in enzootic countries in central and South America, Africa, and parts of Asia (15). In these developing countries, especially in Africa, livestock husbandry development is continuously challenged by brucellosis (16,17). Bovine brucellosis remains the most widespread disease in animals and the main concern in sub-Saharan countries (18). In African countries, B. abortus has been reported in Sudan, in both cattle and their handlers (19), as well as in Gambia, Mali, Niger, and more frequently in Nigeria, Senegal, and Ivory Coast (20). Therefore, a comprehensive analysis of the genetic diversity of B. abortus strains and the epidemiology of the disease in animals and humans in sub-Saharan countries is necessary.
Molecular characterization of the predominated circulating strains is critical to understanding B. abortus diversity and epidemiology in the country. Multi-locus sequence typing (MLST) is a reliable tool for the characterization of Brucella spp. populations and the determination of phylogenetic relationships (21,22); however, this method yields less detailed typing results than MLVA and WGS-SNP due to the lower-resolution power (21,23). Multiple locus variable-number tandem repeat analysis (MLVA) enables Brucella genotyping to infer genetic diversity and investigate the geographic clustering of the isolates (24)(25)(26). Moreover, whole genome sequencing single-nucleotide polymorphism (WGS-SNP) has higher discriminatory power to efficiently track the origin and spread of Brucella strains, leading to be targeted and effective control of disease spread (27,28). At present, MLVA has been used to explore the diversity of B. abortus biovar 3 isolated in West Africa (29) and investigate the epidemiological links of the B. melitensis in Egypt (30). Moreover, studies that used MLVA and WGS-SNP (31,32) indicated that WGS-SNP analysis allows a better differentiation than MLVA-16. WGS-SNP analysis has been used to type, discriminate and track outbreak strains (31)(32)(33). Therefore, the purpose of this study was to use a series of molecular typing assays, including MLST, MLVA, and WGS-SNP, to investigate the species population, genetic diversity, geographical origin, and molecular epidemiology of B. abortus strains from West Africa (WA), to provide insight into the comprehensive understanding of the B. abortus brucellosis epidemiology features in WA, and to facilitate surveillance and control strategy development.

Source of MLST and MLVA genotyping data
The MLST genotyping data of 129 B. abortus strains (

Analysis and visualization of MLST and MLVA data
All data in this study were cleaned and processed using Excel 2016 software (Microsoft Corporation, Redmond, WA, USA). For MLST analysis, genetic similarities between the STs of 129 strains from WA and 631 strains from other continents (PUBMLST) (Supplementary Table S1) were investigated using eBURST software 2.0. Finally, the genetic relationships between the isolates and allelic profiles of MLST data were analyzed with the software PHYLOViZ version 2.0 using the goeBURST Full Minimum Spanning Tree algorithms (34). Subsequently, MLVA data were processed according to genotype information, species/biovars, hosts, location, and years isolated for all B. abortus strains. A minimum spanning tree (MST) of B. abortus strains was constructed using BioNumerics 8.0 software based on the 309 B. abortus strains' MLVA-11 data (Supplementary Table S2) to investigate the host lineage, genotype distribution, and the geographical origin of the B. abortus strains. At the global level, an MST was constructed using the MLVA-11 data of 1,746 strains (Supplementary Table S2) to explore the geographical origin feature of genotypes from this study at the global level. Moreover, the MST based on the MLVA-16 data of 1,746 B. abortus strains (Supplementary Table S2) was used to determinate the relationship among strains at a global scale. In this study, the MST trees were weighted according to MLVA panels, as described by previous reports (35).

Phylogenetic analysis based on the SNPs of strains at the global level
WGS-SNP analysis was performed as previously described (36). Briefly, genomic alignment between the sample genome and reference genome (B. abortus 2308; Assembly ID: GCA_000054005.1) was performed using the MUMmer (37) and LASTZ (38)   (PHYML) algorithm with 1,000 bootstrap replicates to investigate the genetic relationships of B. abortus strains.

Molecular epidemiological links of B. abortus in West Africa and on a global scale
Based on the MST of MLVA-16 data on 309 strains from 17 countries, the strains were divided into two branches (I and II). Branch I comprised strains from Egypt, Zimbabwe, and Uganda ( Figure 5), and branch II contained strains from the other 14 countries ( Figure 5). Remarkably, only one MVA-16 genotype was shared among two strains from two different countries (Senegal and Nigeria) ( Figure 5). Based on the MST of 1,746 strains on a global scale, the majority of strains from this study formed  independent clades, and only few strains have shared the same MLVA-16 genotype with strains from many countries, including Brazil, the USA, South Korea, Argentina, India, Italia, Portugal, the UK, Costa Rica, and China ( Figure 6). Additionally, the MST based on our WGS-SNP matrix analysis showed that the 18 strains from the present study formed 17 SNP genotypes, and which one was shared by two strains from Mozambique and Zimbabwe (Figure 7). Furthermore, five out of 17 SNP genotypes contained only strains from this study (Figure 7), the remaining 12 were shared SNP genotypes that comprised strains from this study and strains from eight different countries, including the USA (Nigeria and Zimbabwe), the UK (Uganda and Senegal), Germany (Chad), France (Egypt), Poland (Sudan), Spain (Chad), New Zealand (Zimbabwe), Bolivia (Mozambique and Zimbabwe), and China (Sudan) (Figure 7).

Discussion
In this study, three molecular methods, including MLST, MLVA, and WGS-SNP, were applied to analyze the genetic diversity, population structure, and molecular relationship among 309 B. abortus strains from WA with global strains. Our analysis highlights that B. abortus strains have a high species/biovars, host diversity, as well as wide geographic distribution in WA. In the present study, more than 90% (280/309) of B. abortus strains were isolated from cattle, and the remaining strains were obtained from 10 other hosts, including wildlife. B. abortus predominantly infect cattle but also infect other host species and seems to be a spill-over from the dominant host species to other species due to farming and grazing (40), which is also evident from in the present study. A previous study showed that 273 Brucella strains from Africa were identified and typed since 1976, among which 272 strains were from cattle and ∼95% (260/270) were from native animals showing hygromas (41). Similar when focusing on WA, B. abortus strains were historically prevalent mainly in cattle (42)(43)(44). Biotyping identified three biovars in WA, and B. abortus bv.3 was the predominant biovar (45). MLVA11 identified B. abortus bv 3 isolated from cattle as new major clusters from WA. In contrast to WA where B. abortus biovar 3 strains occur in cattle, B. abortus biovar 1 is the dominate biovar in South Africa (46). These data suggest that the predominated B. abortus biovars are unique in WA compared to southern Africa, but further investigation is needed to better understand the epidemiological trait of B. abortus in Africa.
Cattle production plays a crucial role in Africa including WA, as cattle are traded for status and serve as a "savings account" in nomadic systems in Nigeria (47). It has been reported that 20% of cattle are imported in Nigeria, mostly from Chad and Niger (48). Moreover, according to literature, brucellosis is one of the major transboundary animal diseases in North African countries, while the illegal animal movement was identified as one of the major constraining factors (49).
In cattle, B. abortus mainly invades the reproductive organs and causes symptoms such as abortion and a decline in milk productivity, resulting in substantial economic losses (50). Therefore, strengthening disease surveillance in cattle is necessary to map the epidemiology of B. abortus. However, strains were found in 10 wildlife species in WA, indicating that the B. abortus strains circulate among animal species that are not the preferential hosts. These reservoirs pose a crucial risk to domestic animals . /fpubh. . and humans. A previous study showed that exploring the wildlife reservoir of brucellosis may represent a new challenge to be faced by the medical and veterinarian community in the twenty-first century (51). Brucellosis in wildlife in this region cannot be neglected, detection and surveillance should be implemented to block the transmission chain. Although MLVA has a higher resolution power for Brucella strains than MLST, both the MLST and MLVA are crucial to epidemiologic surveillance and investigation into the geographical distribution of B. abortus strains. MLST analysis highlighted that the predominant circulating B. abortus population of WA was native lineages, while three STs were shared by many continents (Figure 2). However, these shared populations were predominant in Europe (ST1 and 2) and America (ST1 and 5), implying that the introduced B. abortus lineage was from outside WA, but a further survey into the trade activities of cattle is needed.
The MLVA-11 highlights the high genetic diversity and multiple geographic original of strains from the present study as reflected by the 37 MLVA-11 genotypes were identified among 309 strains that include 22   The global analysis of B. abortus strains base on MLVA-16 highlights that the native dominant strains in WA are driving the ongoing prevalence of B. abortus in this region. This was similar to findings by Liu et al. (58) suggesting that B. abortus strains were transmitted within the national borders of a country despite common geographical origins of the strain in countries along the silk road. B. abortus bv 3 strains from Algeria grouped in two separate clusters using MLVA11 which most strains clustering with European isolates from France and Spain while a few strains cluster in African lineage (59). These studies thus indicate presence of introduced strains as well as native strains in WA. Therefore, further extensive bacteriological and molecular investigations are necessary for a comprehensive understanding of the epidemiology of cattle brucellosis in African countries.
WGS-SNP analysis identified B. abortus lineages circulating in WA was introduced through the import of animals and/or animal products from multiple countries as 12 of 17 SNP genotypes were from nine countries (regions). The MLST, MLVA, and SNP analyses demonstrated that native strains and introduced lineages are co-driving with the persistence of B. abortus circulating in WA. Similarly, WGS-SNP analysis demonstrated that brucellosis in South Africa spreads within the herd on some farms, whereas the introduction of infected animals is the mode of transmission on other farms (33). WGS phylodynamics analysis identified the main B. abortus lineage circulate in Costa Rica are widespread while new introductions seem to be more geographically restricted, which might be similar in various other middle-and low-income countries where brucellosis is endemic with similar farming practices and lack of control (60).
Therefore, a control program involving improved surveillance, animal movement restrictions, public health education is suggested. Moreover, there is a strong need for more sustainable molecular data on prevailing Brucella strains in WA, and on all susceptible species, including humans, to comprehensively analyze the relatedness between field strains of Brucella and the epidemiology of brucellosis within WA countries (29).

Conclusions
The present study revealed the constant circulation of B. abortus strains in cattle throughout WA and neighboring countries. These strains exhibited high species/biovars, host spectrum, and genetic diversity as well as multiple geographic origins. Moreover, most cases were caused by native strains, and few cases resulted from introduced lineages. The surveillance and control of B. abortus in WA should be made a priority which can be enhanced by improving molecular databases that can elucidate epidemiological relationships between strains. Regulations should be strictly implemented when introducing animals to prevent the spread of this species (48).

Data availability statement
The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding authors.

Author contributions
LG and ZLiu performed the data collection and data analysis and drafted the manuscript. MW and QS conducted the data check and critically reviewed the manuscript. ZLiu, ZLi, and XD participated in the design of the study and managed the project. All authors read and approved the final manuscript.