Is Borrelia burgdorferi Sensu Stricto in South America? First Molecular Evidence of Its Presence in Colombia

The genus Borrelia encompasses spirochetal species that are part of three well-defined groups. Two of these groups contain pathogens that affect humans: the group causing Lyme disease (LDG) and the relapsing fever group (RFG). Lyme disease is caused by Borrelia burgdorferi s.l., which is distributed in the Northern Hemisphere, and relapsing fevers are caused by Borrelia spp., which are found in temperate and tropical countries and are an emerging but neglected pathogens. In some departments of Colombia, there are records of the presence of Borrelia sp. in humans and bats. However, little is known about the impact and circulation of Borrelia spp. in the country, especially in wildlife, which can act as a reservoir and/or amplifying host. In this context, the objective of our research was to detect and identify the Borrelia species present in wild mammals in the departments of Caldas and Risaralda in Colombia. For morphological detection, blood smears and organ imprints were performed, and molecular identification was carried out through a nested PCR directed on the flagellin B (flaB) gene. A total of 105 mammals belonging to three orders (Chiroptera, Didelphimorphia and Rodentia) were analyzed, of which 15.24% (n = 16) were positive for Borrelia. Molecularly, the presence of Borrelia burgdorferi s.s. in lung tissues of Thomasomys aureus and blood of Mus musculus (Rodentia) was detected, with 99.64 and 100% identity, respectively. Borrelia sp. genospecies from a clade branch of a bat-associated LDG sister group were identified in seven individuals of bat species, such as Artibeus lituratus, Carollia brevicauda, Sturnira erythromos, and Glossophaga soricina. Furthermore, two Borrelia genospecies from the RFG in seven individuals of bats (A. lituratus, Artibeus jamaicensis, Platyrrhinus helleri, Mesophylla macconnelli, Rhynchonycteris naso) and rodents (Coendou rufescens, Microryzomys altissimus) were documented. Additionally, the presence of a spirochete was detected by microscopy in the liver of a Sturnira erythromos bat specimen. These results contain the first molecular evidence of the presence of B. burgdorferi s.s. in South America, which merits the need for comprehensive studies involving arthropods and vertebrates (including humans) in other departments of Colombia, as well as neighboring countries, to understand the current status of the circulation of Borrelia spp. in South America.


Capture and Sampling of Mammals
Mammal capture was carried out using between two and six mist nets, 50-60 Sherman, and 10-12 Tomahawk traps per sampling site, using directed sampling without standardized efforts. One rodent (Coendou rufescens) was found dead in the study area. Each locality was sampled once for three to four days between March 2021 and April 2022. The Sherman and Tomahawk traps were installed on the first day of sampling between 10:00 and 12:00 h and were removed on the last sampling day between 6:00 and 8:00 h. The mist nets were installed and opened between 17:30 and 20:00 h. A capillary blood sample for blood smear was taken from each of the captured animals. To corroborate taxonomic identification, individuals were collected and euthanized following animal care recommendations [58,59]. Identification of the collected specimens was based on taxonomic keys [60,61] and subsequently deposited in the Mammal Collection of the Museum

Capture and Sampling of Mammals
Mammal capture was carried out using between two and six mist nets, 50-60 Sherman, and 10-12 Tomahawk traps per sampling site, using directed sampling without standardized efforts. One rodent (Coendou rufescens) was found dead in the study area. Each locality was sampled once for three to four days between March 2021 and April 2022. The Sherman and Tomahawk traps were installed on the first day of sampling between 10:00 and 12:00 h and were removed on the last sampling day between 6:00 and 8:00 h. The mist nets were installed and opened between 17:30 and 20:00 h. A capillary blood sample for blood smear was taken from each of the captured animals. To corroborate taxonomic identification, individuals were collected and euthanized following animal care recommendations [58,59]. Identification of the collected specimens was based on taxonomic keys [60,61] and subsequently deposited in the Mammal Collection of the Museum of Natural History of the University of Caldas (MHN-UCa). Wild mammal capture and collection were conducted with the approval of the Comité de Bioética de la Facultad de Ciencias Exactas y Naturales of the Universidad de Caldas (20 September 2019).

Morphological and Molecular Detection of Borrelia
To morphologically detect spirochetes, fine-drop blood smears (four plates per individual) and organ impressions on plates (two for each organ) were performed. Blood samples were obtained by puncture in the brachial vein (in bats) or by making a small cut at the end of the tail (in non-flying mammals), as well as through cardiac puncture after euthanasia. The organs (kidneys, liver, lungs, heart, and in some cases, brain and/or marrow), were arranged individually in sterile Petri dishes and washed with phosphate buffered saline (PBS) [62]. Each organ was then taken separately, washed again with PBS, cut longitudinally or transversely, excess blood was removed with sterile WypAll towels and pressed several times on the slide, and when necessary, several cuts were made. Additionally, and individually, each tissue was stored in absolute ethanol at 4 • C for subsequent molecular analysis (in paired organs, such as kidneys and lungs, a sample was taken from each one). All the materials used in the manipulation of the organs (forceps, scissors, Petri dishes) were washed with iodopovidone and distilled water and sterilized in a portable disinfection box (UV sterilization box 99% Obecilc I-lmh200317), including the WypAll towels. This process was performed each time a different organ was handled. Finally, the plates were fixed with absolute methanol for three minutes for blood smears and five minutes for organ impressions, then stained with 4% Giemsa solution for 40 min for blood and 45 min for organs. Each plate was reviewed using an OLYMPUS BX43 microscope and brightfield at the Laboratory of Molecular Biology of the Universidad de Caldas.
For the detection and molecular identification of Borrelia spp. extractions were performed for each individual and tissue, using the Wizard ® Genomic DNA Purification Kit (according to the standard protocol suggested by the manufacturer). In the DNA extraction, ultrapure water was used as a negative control, and in the PCR amplification, a reaction control was used in each reaction with Borrelia anserina and Borrelia venezuelensis as positive controls, which were donated by the Laboratório de Doenças Parasitárias under the direction of Dr. Marcelo Bahia Labruna of the Departamento de Medicina Veterinária Preventiva e Saúde Animal da Faculdade de Medicina Veterinária e Zootecnia, Universidade de São Paulo (USP-Brazil). Extracted DNA was subjected to nested PCR (nPCR) to amplify the flagellin B (flaB) gene: the first reaction mix was performed at a final volume of 20 µL with the external primer pair FLA LL 5 -ACATATTCAGATGCAGACAGAGAGGT-3 and FLA RL 5 -ACATCATAGCCATTGCAGCAGACAGAGGT-3 , which amplified a 664 bp fragment. The mixture for the nested reaction was made at a final volume of 30 µL with the primers FLA LS 5 -AACAGCTGAAGAGAGCTTGGAATG-3 and FLA RS 5 -CTTTGATCACTTATCATTCATTCTAATAGC-3 , which amplified a fragment of approximately 330 bp [63]. PCR products were evaluated by horizontal electrophoresis on 1% agarose gels stained with ethidium bromide, visualized on a GelDoc-It ® 2310 Image photodocumenter (UVP) (Thermo Fisher Scientific, Waltham, MA, USA), and sent to Macrogen (South Korea) for purification and sequencing.

Phylogenetic Analysis
Partial sequences of the flaB gene were evaluated and edited in Geneious Prime ® 2022.2.2 software [64]. Species identification and analysis considered similarity estimates with 50 public Borrelia sequences available in GenBank, including four sequences from Colombia and Borrelia turcica (associated with reptiles and monotremes) as outgroup. Sequence alignment was run in MAFFT [65] included in Geneious, and the best evolutionary model was selected in ModelFinder using AIC criterion [66]. A maximum likelihood (ML) phylogenetic analysis was performed in IQ-TREE [67], under the TIM3 + G4 + F model for 5000 ultrafast [68] bootstraps, as well as the Shimodaira-Hasegawa-like approximate likelihood ratio test (SH-like aLRT branch test) with 1000 replicates [69], included in the PhyloSuite platform [70], and the phylogenetic tree viewer FigTree v.1.4.3 [71] (Rambaut 2007) was used. Sequences obtained in this study were deposited in GenBank, and allelic matches were identified in Borrelia typing database at http://pubmlst.org/borrelia/ (accessed on 14 October 2022) [72]. Finally, genetic distances were estimated using the p-distance method with the MEGA 11 program [73]. Additionally, association networks were performed between Borrelia species and wild mammalian species, as well as between Borrelia species and the tissues evaluated; for this, R studio version 4.2.1 was used [74].

Results
A total of 105 mammals across 40 species, 10 families, and three orders (Didelphimorphia, Chiroptera, and Rodentia) were captured and analyzed (Table S1). A total of eight species of the order Chiroptera and four of the order Rodentia were found to be infected by Borrelia spp. with a prevalence of infection of 15.2% (n = 16). The families Phyllostomidae (Chiroptera) and Cricetidae (Rodentia) presented the highest number of infected species with seven and two, respectively (Table 1).
A total of 470 tissue and organ samples (blood, liver, lung, heart, kidney, and brain) were examined. The prevalence of Borrelia spp. infection in samples was 5.1% (n = 24). In the liver, 25% (6/24) of the cases of infection were recorded, and in one bat specimen of Artibeus lituratus (Phyllostomidae), Borrelia sp. infection was present in all five tissues evaluated (except brain, which was not examined) (Table 1, Figure 2a). An individual of Microryzomys altissimus (Rodentia: Cricetidae) presented an exclusive infection in the brain with an identity of 99.61% with Borrelia venezuelensis [MG651650] from the RFG (Table 1) Table 1).
The phylogenetic reconstruction obtained by ML using flaB sequences showed that two of the sequences of this study form a monophyletic clade with B. burgdorferi s.s. with a statistical support greater than 92% and a group with the sequences of the same species that were reported in the USA and Canada, with divergences between 0-1.1% (Figure 3; Table  S2). Another monophyletic group was formed by 15 sequences of Borrelia sp. genospecies isolated from bats in Colombia (statistical support of 98.8%), which branches as a sister group of the LDG borreliae and are related to the sequences recorded in Macaregua Cave, Colombia ( Figure 3; Table S2). Finally, the remaining seven sequences had genetic distances between them of 0-0.4%, are associated with the RFG, and are more closely related to B. venezuelensis (Brazil) with a statistical support of 83.5% in the ML tree and with genetic distances between 0-1.1% (Figure 3; Table S2). Finally, 13 associations between Borrelia and wild mammals were reported, invol Borrelia genospecies and 12 mammal species. Borrelia from the RFG presented the hig number of associations with several species of bats and rodents. In contrast, Borrelia g species branching as a sister group of the LDG were associated only with bats, an Finally, 13 associations between Borrelia and wild mammals were reported, involving Borrelia genospecies and 12 mammal species. Borrelia from the RFG presented the highest number of associations with several species of bats and rodents. In contrast, Borrelia genospecies branching as a sister group of the LDG were associated only with bats, and B. burgdorferi s.s. found in this study was associated with two rodent species: Thomasomys aureus (Rodentia: Cricetidae) and Mus musculus (Rodentia: Muridae) (Figure 4). Overall, blood was the tissue with the highest richness of Borrelia genospecies reported in this study (Figure 4g). Figure 3. Phylogenetic tree of the partial sequences of the flaB gene of Borrelia species in the present study (in bold) and of the sequences in GenBank (accession numbers in parentheses), using the maximum likelihood (ML) method and the GTR + G4 + F model. Numbers at nodes are selected branch support analyses; from left to right: ultrafast bootstrap values, and Shimodaira-Hasegawalike approximate likelihood ratio test (SH-like aLRT). The Borrelia turcica sequence was used as an outgroup.
Finally, 13 associations between Borrelia and wild mammals were reported, involving Borrelia genospecies and 12 mammal species. Borrelia from the RFG presented the highest number of associations with several species of bats and rodents. In contrast, Borrelia genospecies branching as a sister group of the LDG were associated only with bats, and B. burgdorferi s.s. found in this study was associated with two rodent species: Thomasomys aureus (Rodentia: Cricetidae) and Mus musculus (Rodentia: Muridae) (Figure 4). Overall, blood was the tissue with the highest richness of Borrelia genospecies reported in this study (Figure 4g).

Discussion and Conclusions
Our results represent the first molecular evidence of the presence of B. burgdorferi s.s. in South America, a species that was considered absent in this region of the continent, where the presence of its known vectors is not recorded [16,75]. Nevertheless, the results obtained do not rule out the possibility that other ectoparasite species may be acting as vectors of these spirochetes, as recorded by other authors [38,76,77]. For South America, several Lyme disease group genospecies have previously been documented in both wild mammals and ticks [8,38,42,[78][79][80]. In countries such as Brazil, the Baggio-Yoshinari syndrome has occurred, which is clinically similar to Lyme disease [81]; however, the causative agent could not be confirmed to be B. burgdorferi s.s. [82].
The ML reconstruction is congruent with previous reported phylogenetic inferences [37,40,52,80,83]. In particular, the Borrelia sequences recorded from bats in the present study are closely related to genospecies isolated from Carollia perspicillata in the Department of Santander, Colombia [40]. This relationship may indicate that the genospecies reported in our study and those recorded by Muñoz-Leal et al. (2021) [40] and Jorge et al. (2022) [80] are Borrelia restricted to bats, but further studies are needed to help support this hypothesis.
Furthermore, the presence of B. burgdorferi s.s. in rodents (i.e., T. aureus and M. musculus) indicates their potential role as reservoirs for Lyme disease in the country, as has been observed in other rodents in Eurasia and North America [84,85]. Cricetid rodents (e.g., Peromyscus leucopus) have been evidenced to be competent reservoirs for B. burgdorferi s.l. in North America [84][85][86][87] and South America [8,42]. A high prevalence of infection by B. burgdorferi s.l. and other Borrelia species has been reported in other rodents such as murids [88,89]. Muridae have acquired great importance in zoonotic diseases, since their populations usually reach a large number of individuals that can act as hosts of ectoparasites and reservoirs of pathogens [89]. Similarly, some murids have colonized urban or rural environments, where they have close contact with humans and domestic animals, increasing the risk of infection [89].
The presence of Borrelia in the brain of M. altissimus coincides with the neurotropic characteristic of several Borrelia species [90,91], as is the case with Treponema pallidum [91]. Borrelia species are thought to infect the brain to evade the host immune response and may occasionally invade the blood to facilitate bacterial transmission [92]. This is likely the case for B. duttonii and B. turicatae, which have detected in the brains of rodents [91][92][93], and B. miyamotoi and B. burgdorferi, which were detected in the brain of rodents and a shrew (Sorex maritimensis) [94]. Due to the complexity of brain extraction, this organ was studied only in one specimen; therefore, we recommend including the brain in Lyme disease studies, as the ability of an infected host to potentially serve as a reservoir could be underestimated [95]. Additionally, the high prevalence of infection in the liver (25%) can be explained through one of the functions of this organ, which by filtering the blood, is likely to be one of the first tissues to be infected when spirochetes leave the circulatory system [94]. Furthermore, species of the B. burgdorferi complex have been shown to use the liver as a refuge to evade the immune response in long-term infections [96][97][98].
Finally, this research provides molecular evidence of the presence of B. burgdorferi s.s. in South America. Likewise, new associations between mammals and Borrelia sp. are presented for the American continent. These findings suggest that the presence of species of the genus Borrelia could be widely distributed in the wildlife of Colombia. However, complementary studies should be carried out to determine the real status of Borrelia spp. in South America and to establish the role played by wild mammals in the maintenance of infections by these spirochetes, as well as their participation in the enzootic and zoonotic cycles of Borrelia sp.
Supplementary Materials: The following supporting information can be downloaded at https://www. mdpi.com/article/10.3390/tropicalmed7120428/s1: Table S1: Wild mammals sampled (Departments of Caldas and Risaralda). Individuals that were positive for Borrelia are highlighted in bold.-indicates that the information is not available; Table S2: Intraspecific and interspecific distances based on the p-distance method for the flagellin gene fragment (flaB) with the MEGA v. 11 program. In bold are the sequences obtained in this study and the GenBank accession numbers are provided in parentheses.