Microbiological, Molecular Diagnosis and Antimicrobial Resistance of Bordetella bronchiseptica in the Equine Reproductive Tract in Brazil Authors: Samira Salim Mello Gallo, Maria de Lourdes Amaral Bernardino, Luiz Antonio Silveira da Motta Júnior, Francisco Carlos Rodrigues de Oliveira, Marcos Antonio Pessanha Barreto, Marcio Manhães Folly Published: 2026 ABSTRACT ABSTRACTBackgroundThe genus Bordetella comprises 16 species of aerobic Gram‐negative bacteria, with the most notable being B. pertussis, B. parapertussis and B. bronchiseptica. The latter stands out for its ability to infect various hosts, causing both respiratory and reproductive issues, such as endometritis and infertility in mares. These infections impair reproductive efficiency and facilitate the spread of the pathogen among susceptible animals.ObjectivesThe present study aimed to report the isolation and molecular diagnosis of B. bronchiseptica associated with endometritis and infertility in mares in Brazil, as well as to evaluate the antimicrobial susceptibility profile of the isolates.MethodsFor this purpose, uterine lavage was collected from three Mangalarga Marchador mares using an aseptic technique with a transcervical probe and Ringer's lactate solution. The samples were centrifuged, cultured on MacConkey agar, blood agar and Sabouraud agar, and identified by biochemical characteristics. PCR was performed to amplify the flanking region of the flaA gene of B. bronchiseptica, followed by sequencing and analysis.ResultsThe antibiogram was conducted using the agar diffusion method on Mueller–Hinton agar. Microbial culture identified B. bronchiseptica, confirmed by biochemical and molecular tests. The isolates showed resistance to multiple antibiotics, such as penicillins and cephalosporins.Main limitationsNo investigation into transmission routes, prevalence or risk factors in the broader equine population.ConclusionThis study describes, for the first time in Brazil, the detection of B. bronchiseptica causing uterine infection and infertility in mares, expanding knowledge about its pathogenic potential. The results highlight the need to consider this bacterium in the diagnosis of reproductive disorders, even in the absence of respiratory manifestations.Keywords: bacteria, endometritis, infertility, mares, uterine lavage Introduction The genus Bordetella belongs to the family Alcaligenaceae and comprises 16 species: B. pertussis, B. parapertussis, B. bronchiseptica, B. flabilis, B. sputigena, B. bronchialis, B. muralis, B. tumulicola, B. tumbae, B. petrii, B. trematum, B. holmesii, B. avium, B. hinzii, B. pseudohinzii and B. ansorpii (Soumana et al. 2017). Members of this genus are Gram‐negative, obligate aerobic bacteria involved in various respiratory and systemic infections in animals and humans. Bordetella pertussis, B. parapertussis and B. bronchiseptica are classified as the ‘classical Bordetella’ (Mattoo and Cherry 2005). Bordetella bronchiseptica has a broad host range, infecting animals from guinea pigs to dogs and humans, including pinnipeds (Register et al. 2015) and lagomorphs (Pathak et al. 2011). Among the classical Bordetella, B. bronchiseptica is the primary species clinically relevant to equines. Equine bordetellosis has been traditionally reported as a cause of respiratory disorders including pneumonia, sinusitis and guttural pouch lesions (Gallagher 1965; Koehne et al. 1981; Mather et al. 1973; Miller 1965; Mohan and Obwolo 1991; Saxegaard et al. 1971; Vaid et al. 2018). However, B. bronchiseptica is also the only species of the genus linked to reproductive tract pathology in mares, historically associated with sporadic cases of infertility and abortion (Mather et al. 1973; Mohan and Obwolo 1991). In the equine industry, endometritis is recognised as one of the main pathologies impairing reproductive performance. The inflammatory process of the endometrium—often infectious—compromises uterine function and is a leading cause of subfertility and infertility (Ravaioli et al. 2022). This underlying inflammation results in conception failure, early embryonic losses and shortened luteal phases, significantly impacting reproductive efficiency and herd productivity (Christoffersen and Troedsson 2017). Carrier animals continue to spread the pathogen, infecting susceptible individuals (Bemis et al. 1977; Coutts et al. 1996; Schulz et al. 2014). The present study aimed to report the isolation and molecular diagnosis of B. bronchiseptica associated with endometritis and infertility in mares in Brazil, as well as to evaluate the antimicrobial susceptibility profile of the isolates. Materials and Methods Sample Collection Uterine lavage samples were collected from three Mangalarga Marchador mares from a rural property in the municipality of Campos dos Goytacazes, Rio de Janeiro State, and transported to the Bacteriology Laboratory of the Universidade Estadual do Norte Fluminense (UENF). All animals selected for this study were multiparous Mangalarga Marchador mares, aged between 8 and 14 years, with a clinical history of recurrent endometritis and subfertility. Recurrent endometritis was defined by persistent intraluminal uterine fluid accumulation detected via ultrasound across consecutive examinations, combined with a history of failure to conceive after three or more consecutive breedings. These animals had previously undergone conventional uterine therapies (lavages and antibiotic infusions) with transient success. Sampling was performed strictly during the oestrus phase of the cycle, confirmed via transrectal ultrasonography by the presence of an endometrial oedema pattern, a dominant pre‐ovulatory follicle (> 35 mm), and the absence of a functional corpus luteum during the 2024–2025 breeding season. Ultrasonography Transrectal ultrasonographic evaluation of the mares' reproductive tract was performed using a high‐frequency linear transducer (Mindray DP‐2200vet, China). Uterine condition was characterised by evaluating the volume and echogenicity of intraluminal fluid accumulation. Endometrial oedema severity was classified using a standard 0–4 scoring system (0 = no oedema/dioestrus, 1 = slight oedema, 2 = moderate oedema, 3 = severe/oestrus peak oedema; 4 = physiological exaggeration/pathological oedema). All mares in this study exhibited marked uterine fluid accumulation and advanced endometrial oedema scores (Grades 3–4). Uterine Lavage For transcervical uterine lavage, a silicone probe equipped with an inflatable balloon with a capacity of 100 mL was used. Initially, the tail was hygienically restrained with a bandage, followed by careful antisepsis of the perineal and vulvar region with a 0.5% chlorhexidine digluconate solution. To mitigate zoonotic risks and prevent contamination, the veterinary personnel wore full personal protective equipment (PPE), including sterile disposable gowns, face masks, and sterile gloves throughout the procedure. To prevent vaginal contamination, a guarded technique was employed: the catheter tip was protected during vaginal passage and only exposed upon reaching the cervical os. The tubing system was fully primed with fluid prior to insertion to prevent air from entering the uterus. Once positioned, the balloon was inflated with 40 mL of air, promoting adequate sealing of the cervical lumen and allowing retrograde fluid infusion. Subsequently, 1 L of sterile Ringer's lactate solution was administered into the uterus and left inside the uterine lumen for a retention period of 5 min to optimise sample recovery. Following this, the uterine contents were aseptically recovered through a closed gravity‐flow system established between the catheter and the collection bag, ensuring the microbiological integrity of the sample for subsequent analyses. Bacterial Culture For concentration of cellular elements in the sediment, the collected uterine material was transported at 4°C and processed within 2 h of collection. Representative aliquots of the collected uterine material were transferred to 15 mL Falcon tubes and subjected to centrifugation at 5000 rpm for a period of 10 min. After this process, the supernatant was carefully discarded, and the resulting pellets were pooled and aseptically inoculated, using platinum loops, onto Petri dishes containing MacConkey agar, blood agar and Sabouraud agar culture media, through the exhaustion seeding technique. The samples were incubated in a bacteriological incubator at 36°C under aerobic conditions for an initial period of 24–48 h for subsequent analysis of microbial growth. All microbiological procedures were strictly performed inside a Class II Biosafety Cabinet using PPE including lab coats, gloves and masks to mitigate zoonotic risks to laboratory personnel. Identification of Isolates For the identification of the bacteria, typical cultural characteristics of the colonies were observed after 24–48 h of incubation, such as size, consistency, shape, pigmentation and haemolysis (Hernández‐Durán et al. 2017; Khine et al. 2023). All visible colonies representing predominant growth phenotypes were selected for identification. Subsequently, each isolate was subjected to biochemical tests, including catalase and oxidase tests, nitrate reduction and urease tests. Other biochemical characteristics were obtained using the VITEK 2C system (BioMérieux, Inc., Durham, NC, USA). The isolates were tested for ala‐phe‐pro arylamidase (APPA), adonitol (ADO), L‐pyrrolidonyl‐arylamidase (PyrA), L‐arabitol (lARL), D‐cellobiose (dCEL), β‐galactosidase (BGAL), sulfeto de hidrogênio (H2S), β‐N‐acetilglucosaminidase (BNAG), glutamyl arylamidase pNA (AGLTp), D‐glucose (dGLU), gama‐glutamil transferase (GGT), fermentation glucose (OFF), beta‐glucuronidase (BGLU), D‐maltose (dMAL), d‐manitol (dMAN), D‐mannose (dMNE), B‐xylose (BXYL), beta‐alanine arylamidase (BAlap), L‐prolin arylamidase (ProA), lipase (LIP), palatinose (PLE), tyrosine arylamidase (TyrA), urease (URE), D‐sorbitol (dSOR), saccharose/sucrose (SAC), D‐tagatose (dTAG), D‐trehalose (dTRE), citrate/sodium (CIT), malonate (MNT), 5‐keto‐D‐gloconate (5KG), L‐lactate alkalinisation (lLATk), alpha‐glucosidase (AGLU), succinate alkalinisation (SUCT), beta‐N‐acetyl (NAGA), alpha galactopyranusidase (AGAL), phosphatase (PHOS), glycine arylamidase (GlyA), ornithine decarboxylase (ODC), lysine decarboxylase (LDC), histidine assimilation (lHISa), coumarate (CMT), beta‐glucuronidase (BGUR), 0/129 resistance (O129R), glu‐gly‐arg‐arylamidase (GGAA), L‐malate assimilation (IMLTa), ellman (ELLM) and L‐lactate assimilation (ILATa). The test card was filled with a bacterial suspension according to the manufacturer's recommendations (BioMérieux), sealed, and incubated for 3 h. During this period, the card was read by kinetic fluorescence every 15 min. The VITEK 2C system software first analysed the data and then automatically reported the results. Antibiogram The isolated bacteria were inoculated into sterile saline solution, placed in glass test tubes, until reaching turbidity equivalent to the 0.5 McFarland standard, verified using a turbidimeter (Densimat, bioMérieux, France). Then, using a sterile swab, the bacterial suspension was streaked onto Petri dishes containing Mueller–Hinton agar; for the Streptococcus genus, blood agar or Mueller–Hinton agar supplemented with blood was used. After a 5‐min period, sufficient for the inoculum to be absorbed by the culture medium, paper disks impregnated with the antibiotics to be tested were applied, following the diffusion method described by Bauer et al. (1966). The antibiotic panel used consisted of amikacin (30 µg), amoxicillin (20 µg), amoxicillin with clavulanate (20/10 µg), tetracycline (30 µg), cephalothin (30 µg), cephalexin (30 µg), cefotaxime (30 µg), penicillin (10 UI), neomycin (30 µg), sulfamethoxazole–trimethoprim (23.75/1.25 µg), doxycycline (30 µg), gentamicin (10 µg), levofloxacin (5 µg), norfloxacin (10 µg), enrofloxacin (5 µg) and ciprofloxacin (5 µg) (CECON, Brazil). Polymerase Chain Reaction (PCR) The 24‐h bacterial cultures were used for DNA extraction with the DNeasy kit (Qiagen) following the manufacturer's instructions. Bacterial colonies were harvested from the agar surface using a sterile inoculation loop and suspended in lysis buffer. PCR was performed with specific primers for B. bronchiseptica that amplified a sequence upstream of the flagellin (fla) gene to obtain a 237 bp fragment (Hozbor et al. 1999). For the reaction, 12.5 µL of GoTaq Hot Start Green Master Mix (Promega, USA), 0.5 µL of each primer and 1 µL of sample were used. After initial denaturation (94°C/3 min), PCR amplifications were carried out for 35 cycles in a TGradient thermocycler (Analytic Jena, Germany), with a denaturation temperature of 94°C for 1 min, annealing at 58°C for 1 min and extension at 72°C for 1 min, followed by a final extension at 72°C for 7 min. The PCR products were evaluated by electrophoresis on a 2% agarose gel stained with Unisafe Dye (Uniscience, USA) and visualised in a photodocumentation system (Gel Logic 200 Imaging System, USA) and Carestream Molecular Imaging software after electrophoretic run. Sequencing The amplified fragment (∼237 bp) resulting from PCR was purified using the GFX PCR DNA Band Purification Kit (GE Health Sciences, Champaign, IL, USA) and sequenced with the DYEnamic ET Kit Cycle Sequencing Terminator Dye (GE Health Sciences) in a MegaBACE sequencer (GE Health Sciences). Phylogenetic Analysis The consensus sequence was analysed using CodonCode Aligner v.2.0.4 software (CodonCode Corp., Dedham, MA) and aligned with reference sequences of B. bronchiseptica published in GenBank (https://blast.ncbi.nlm.nih.gov/Blast.cgi) using MEGA X software (Kumar et al. 2018) by the maximum likelihood method and Hasegawa–Kishino–Yano model (Hasegawa et al. 1985). The analysis involved 27 nucleotide sequences (3 sequences from this study and 24 reference sequences from GenBank). Missing data and ambiguous bases were allowed at any position (partial deletion option). Group confidence was assessed by bootstrap values using 1000 replicates. Results Ultrasonographic examination revealed an extensive accumulation of hyperechoic intraluminal fluid, with a fluid depth exceeding 20 mm in all evaluated animals. The hyperechogenicity was classified as marked due to the high density of floating cellular debris, suggesting a severe purulent exudative process (Figure 1) in the three examined mares. Ultrasonographic evaluation of the uterus of a mare with endometritis. In (a) the uterus with oedema and absence of intraluminal fluid, in (b) a small amount of fluid distributed in the lumen and in (c) a large amount of intraluminal fluid. Bars: 10 mm. In culture, B. bronchiseptica showed predominant growth and its morphology varied from smooth to rough when cultivated on agar plates. On blood agar medium, it exhibited bright β‐hemolytic colonies and developed an average diameter of 2.0 mm within 1–2 days. The colonies grew equally well on MacConkey and Sabouraud agar. In manually performed biochemical tests, the isolates were positive for oxidase, catalase, nitrate, and urease tests. Fifty‐seven biochemical tests were performed using a Gram‐negative bacterial identification card. The three isolates were characterized as positive for L‐proline arylamidase (ProA), L‐lactate alkalinisation (ILATk), L‐pyrrolidonyl‐arylamidase (PyrA), succinate alkalinisation (SUCT) and urease (URE), and were identified by the VITEK 2 system as Bordetella bronchiseptica. The antimicrobial susceptibility profile revealed that all isolates were resistant to amoxicillin, amoxicillin with clavulanate, cephalexin, cefotaxime, cephalothin and penicillin. Initially, the animals underwent a combined acute therapeutic protocol. This consisted of the intrauterine infusion of 100 mL of a diluted 1% povidone–iodine solution (prepared from a 10% stock solution in sterile 0.9% NaCl), administered transcervically once daily for 3 consecutive days, and completely syphoned out after a 15‐min retention period. Concurrently, systemic antibiotic therapy was administered using Ceftiofur sodium (2.2 mg/kg, IM, once daily for 7 days). This initial management effectively controlled the bacterial infection in its acute phase, which was confirmed 48 h post‐treatment by follow‐up clinical and ultrasonographic evaluations demonstrating the complete resolution of intraluminal fluid and abnormal endometrial oedema. However, clinical recurrence was observed within 14–21 days post‐treatment, characterised by the return of purulent vulvar discharge and intraluminal fluid accumulation, suggesting potential persistence or intracellular survival of the etiological agent. As the final stage of therapeutic management, approximately 30 days after recurrence, autologous vaccines (formulated with formaldehyde‐inactivated local isolates at 109 CFU/mL and aluminum hydroxide adjuvant, administered in a two‐dose protocol 21 days apart) were administered with the aim of inducing a specific and targeted immune response against the pathogen. This immunotherapeutic approach allowed the production of viable embryos for donation, demonstrating a protective window of transient efficacy in controlling the infection before mild clinical relapses occurred at the end of the breeding season. All samples tested positive by PCR (Figure 2), confirming the presence of the species. The sequences were deposited in GenBank under accession numbers PV836588, PV836589 and PV836590. Electrophoresis of the PCR product for B. bronchiseptica detection in uterine lavage from mares. L, 100 bp molecular weight marker; Neg, negative control; Pos, positive control; and 1, 2 and 3, microbiological culture samples from uterine lavage of mares. The results of the phylogenetic analysis of our own isolate sequences and the sequences available in GenBank are shown in Figure 3. It can be observed that the sequences from the samples in this study (PV836588, PV836589 and PV836590) formed a well‐supported clade (bootstrap ≥ 95) with sequences EU327790, EU327789 and AJ012319, all derived from the flanking region of the flaA gene, while the other clade grouped sequences amplified from the flaA gene itself. Phylogenetic tree based on partial nucleotide sequences (21 Bordetella bronchiseptica strains and upstream sequences (6 B. bronchiseptica strains) of flaA gene. The tree was constructed using the maximum likelihood method and Hasegawa–Kishino–Yano model, scaled or better visualization. The optimal tree with the highest log likelihood (−392,324) is shown. The percentage of replicate trees in which the associated strains clustered together in the bootstrap test (1000 replicates) are shown next to the branches. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. There were a total of 1051 positions in the final dataset. Discussion The isolation of B. bronchiseptica from the uterine lavage of mares in the present study indicates that this pathogen can be associated with the development of chronic reproductive disorders in equines. This finding aligns with previous evidence describing the presence of B. bronchiseptica in pure cultures of stomach contents, liver tissue and lung tissue from aborted equine foetuses (Mohan and Obwolo 1991), as well as its historical association with infertility in mares (Mather et al. 1973). The detection of this microorganism in the uterine cavity, traditionally considered a low‐microbial abundance niche rather than strictly sterile, highlights its capability to survive and proliferate under inflammatory conditions, potentially causing severe endometritis. Although Bordetella bronchiseptica is classically recognised as a respiratory pathogen (Diavatopoulos et al. 2005; Gerlach et al. 2001), the animals evaluated exhibited an atypical clinical presentation characterised exclusively by reproductive failure, without any concomitant respiratory manifestations. This unusual manifestation may be related to the specific microbiological characteristics of this case, where B. bronchiseptica was recovered as a predominant growth in the absence of other classical uterine pathogens. While B. bronchiseptica is often described as a slow‐growing organism compared to rapid environmental contaminants, requiring up to 48 h of incubation for full colonial development, its established pure growth in these chronic cases suggests a lack of active bacterial competition within the altered uterine environment of these specific mares. The absence of respiratory signs associated with reproductive manifestations raises important questions about the underlying pathogenicity mechanisms. It is possible that strains with distinct tissue tropisms are circulating in the equine population, or that host factors, such as immune status or endometrial mucosal barrier integrity, may influence the infection site. This unusual manifestation may be related to the specific microbiological characteristics of this case, where B. bronchiseptica was recovered as a predominant growth in the absence of other classical equine uterine pathogens. While B. bronchiseptica is often described as a slow‐growing organism compared to rapid environmental contaminants (Bemis and Fenwick 2004), requiring 24–48 h of incubation for full colonial development, its established pure growth in these chronic cases suggests a lack of active bacterial competition within the altered uterine environment of these specific mares. This particularity may explain its ability to establish persistent infections in unconventional niches, such as the reproductive tract, where it finds favourable conditions for proliferation. One potential hypothesis for the introduction of this microorganism into the reproductive tract involves lapses in aseptic protocols during reproductive biotechnology procedures such as artificial insemination. Improper handling of fomites or even the administration of semen under non‐sterile conditions could facilitate the inoculation of opportunistic pathogens into the uterine lumen (Liu and Troedsson 2008). The detection of B. bronchiseptica in the normally sterile uterine environment supports this possibility, suggesting plausible iatrogenic contamination, particularly in scenarios involving intensive reproductive management practices. Recent studies demonstrate that B. bronchiseptica possesses multiple virulence factors, including adenylate cyclase toxin and sophisticated immune evasion mechanisms, which not only facilitate tissue colonisation but also enable persistent infection in apparently healthy hosts for extended periods (Van Crombrugge et al. 2022; Chamorro et al. 2023). Although the specific pathogenic role of Bordetella spp. in the equine uterus has not yet been fully established, its detection in endometritis cases warrants further investigation into its potential involvement in uterine inflammatory processes. Given this capacity for bacterial persistence, the use of sensitive diagnostic methods such as PCR is valuable—proving critical both for accurate pathogen detection and for epidemiological surveillance to control bacterial spread within equine populations. It is important to note that while targeted PCR confirms the presence of B. bronchiseptica DNA, it does not automatically establish exclusive causation or rule out the presence of other untargeted microorganisms. However, in this study, the clinical role of B. bronchiseptica is strongly supported by the microbiological culture, which consistently yielded heavy and predominant growth of this bacterium in the absence of classical equine uterine pathogens, such as Streptococcus zooepidemicus or Escherichia coli. Future metagenomic studies could provide a broader view of the uterine microbiome in these atypical chronic cases. The increasing reduction in antimicrobial susceptibility among Bordetella bronchiseptica isolates, particularly to β‐lactams such as penicillins and cephalosporins, has been widely reported worldwide (Kadlec et al. 2007; Vaid et al. 2018), a phenomenon corroborated by the results of this study which demonstrated significant resistance to antibiotics including penicillin, cefotaxime, cephalexin, cephalothin and amoxicillin. The isolates exhibited resistance to cephalexin, which is a well‐documented intrinsic characteristic of B. bronchiseptica. This biological resistance explains why cephalexin is frequently utilised as a selective agent in culture media to isolate this pathogen from contaminated samples, further corroborating the phenotypic identity of our strains. These findings reinforce concerns about the indiscriminate use of antimicrobials in veterinary practice, which may be contributing to the selection of resistant strains (Canisso et al. 2020; Rathbone et al. 2023). In this context, the adoption of strategies for rational antimicrobial use becomes essential, including the careful performance of antibiograms to guide precise therapy, minimising treatment failures and the spread of resistance. Continuous surveillance of B. bronchiseptica susceptibility profiles is crucial to ensure effective treatments and preserve the efficacy of available antimicrobial options. Given that B. bronchiseptica is an opportunistic pathogen capable of infecting multiple domestic and livestock animal species that have frequent contact with humans (Garcia‐Cantu et al. 2000), there is a potential risk of zoonotic transmission, particularly to immunocompromised individuals (Yacoub et al. 2014). The phylogenetic similarity between human strains and those from companion animals like dogs and rabbits highlights the need to monitor B. bronchiseptica in animal populations with close human contact, especially in rural or equine production settings. Although reports of human infections associated with horses remain limited in the literature, the epidemiological potential persists, particularly in cases of close contact or handling of infected animals. Therefore, to prevent cross‐species transmission, it is crucial to expand studies on the molecular epidemiology of circulating strains in horses and to develop educational strategies for professionals working at the human–animal interface. Conclusion This study represents the first report of B. bronchiseptica detection in the reproductive tract of mares in Brazil, expanding our understanding of this bacterium's pathogenic spectrum. Given the economic importance of reproductive losses in equine breeding, these findings highlight the need to include B. bronchiseptica in the differential diagnosis of reproductive failures in mares, even in the absence of respiratory signs. This underscores the importance of including B. bronchiseptica in the differential diagnosis of persistent equine reproductive disorders and emphasises the value of continuous epidemiological surveillance and rational antimicrobial stewardship. Author Contributions Samira Salim Mello Gallo: conceptualization, writing – original draft, review and editing; investigation. Maria de Lourdes Amaral Bernardino: investigation. Luiz Antonio Silveira da Motta Júnior: conceptualization, writing – review and editing. Francisco Carlos Rodrigues de Oliveira: writing – review and editing. Marcos Antonio Pessanha Barreto: writing – review and editing. Márcio Manhães Folly: conceptualization, writing – review and editing. Funding The authors have nothing to report. Ethics Statement The authors have nothing to report. Consent The owners gave consent for their horses' inclusion in the study. Conflicts of Interest The authors declare no conflicts of interest. Data Availability Statement The datasets generated during and/or analysed during the current study are available from the corresponding author upon reasonable request. References Associated Data Data Availability Statement The datasets generated during and/or analysed during the current study are available from the corresponding author upon reasonable request.