Detection and Characterization of Clinical Bordetella trematum Isolates from Chronic Wounds

Bordetella trematum is a relatively newly discovered and potentially frequently overlooked Bordetella species, mostly isolated from chronic wounds and predominantly in those of the lower extremities. Its susceptibility profile and clinical significance is still debated, given the limited amount of available data. We contribute providing a molecular and phenotypical analysis of three unique clinical B. trematum isolates detected between August 2019 and January 2020 to aid the matter. Cryo-conserved isolates were subcultured and re-identified using various routine means of identification. Bacterial genomes were fully Illumina-sequenced and phenotypical susceptibility was determined by broth microdilution and gradient-strip tests. All isolates displayed increased susceptibility to piperacillin–tazobactam (<2/4 mg/L), imipenem (<1 mg/L), and meropenem (<0.047 mg/L), whereas they displayed decreased susceptibility to all tested cephalosporins and fluoroquinolones (according to PK-PD, EUCAST 10.0 2020). One isolate carried a beta-lactamase (EC 3.5.2.6) and a sulfonamide resistance gene (sul2) and cells displayed resistance to ampicillin, ampicillin/sulbactam, and trimethoprim/sulfamethoxazole. All isolates carried genes conferring decreased susceptibility to aminoglycosides (aadA), fosfomycin (fosA) and fluoroquinolones (gyrB EC 5.99.1.3). Awareness that B. trematum can be resistant to trimethoprim/sulfamethoxazole is warranted.


Introduction
Bordetella trematum is a relatively newly discovered [1] Gram-negative, encapsulated, nonsporulating rod in the family of the Alcaligenaceae that belongs to the taxonomically heterogeneous group of the non-fermenters and as such does not rely on sugars for energy production. While its fastidious growth and barely visible colonies hamper its detection, once detected, biochemical identification systems frequently misidentify it, most commonly as Bordetella avium, Bordetella bronchiseptica, or Achromobacter denitrificans [2][3][4]. To date a reliable accurate discrimination in laboratory routine is only possible with matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry that should further be confirmed by 16s rRNA gene sequencing.
Nonfermenting Gram-negative bacilli have emerged as important healthcare-associated pathogens [5]. The most important ones in this regard are Pseudomonas aeruginosa, Acinetobacter baumannii, and Stenotrophomonas maltophilia. A recent literature review reveals the clinical significance of B. trematum to still be a matter of debate as the little information available on this pathogen left authors largely uncertain as to whether therapy was necessary or superfluous in the respective cases [2][3][4][6][7][8][9][10][11][12][13]. Previously documented cases have mainly been isolated from ulcers and soft tissue infections, even though occasional cases of bacteremia have been described [6,7]. Given the limited amount of documented cases, each new documented case adds evidence to the issue. We contribute to the matter by providing an in-depth analysis of three unique clinical B. trematum isolates, detected in wound swabs of three patients of the University Hospital Bonn between December 2019 and April 2020 and hope to aid the matter by being among the first to provide molecular susceptibility profiles of clinical B. trematum isolates that explain and supplement previously documented and current phenotypical susceptibility-patterns.

Origin of the Isolates
The three Bordetella trematum isolates were isolated from wound swabs of patients of the University Hospital Bonn between December 2019 and April 2020. Isolate one belonged to a 75-year-old female patient with chronic renal insufficiency, arterial hypertension, and hypothyroidism, and was detected in a wound swab of an ulcer on the right heel, together with Staphylococcus aureus and Enterobacter cloacae complex. Isolate two belonged to a 49-year-old obese male patient with hereditary factor V Leiden, post-thrombotic syndrome, chronic ulcerations, chronic obstructive pulmonary disease, and nicotine abuse, and was the only bacterial agent to be detected in a wound swab of an ulcer on the left foot. Isolate three belonged to an 81-year-old male patient with peripheral arterial disease and peripheral neuropathy and was detected in a tissue sample of the left distal tibia together with Staphylococcus capitis.

Susceptibility Testing
The results of the susceptibility testing by gradient strip and automated dilution are displayed in Table 1. B. trematum isolates one and three were very similar in terms of their antibiotic susceptibility profile. All three isolates were highly susceptible to piperacillintazobactam and all tested carbapenems. All three isolates tested resistant or decreasingly susceptible to most tested cephalosporins and fluoroquinolones. Isolate two displayed decreased susceptibility or resistance to ampicillin, ampicillin/sulbactam, and trimethoprim/sulfamethoxazole. Small colony variation was determined to be unlikely based on colony morphology.

Genome Analysis
The ResFinder of the CGE Server [14,15] detected one acquired antimicrobial resistance gene in isolate two, namely sul2, with 100% identity, conferring resistance to sulphonamides. No other acquired resistance genes present in the database were detected. The gene annotation server RAST [16][17][18] identified genes conferring decreased susceptibility to aminoglycosides (aadA), fosfomycin (fosA), and fluoroquinolones (gyrB EC 5.99.1.3) on the genomes of all three isolates. On the genome of isolate two additionally a gene encoding for a beta-lactamase (EC 3.5.2.6) was detected. The latter was not encoded on the same contig as the sul2 gene. The number of annotated genes per subsystem for each isolate is reported in Table 2.

Discussion
In 1996 Bordetella trematum was described as being the only oxidase-negative Bordetella species [1]. Later, however, the emergence of oxidase-positive isolates [8] was reported, questioning reliability of this tool for biochemical differentiation. Reasons for this incongruity were hypothesized to lie in the different derivatives used for the oxidase reaction in the two studies [2], N,N -dimethyl-p-phenylene monochloride in the earlier and N,N,N ,Ntetramethyl-p-phenylenediamine dihydrochloride in the latter study. Our isolates tested positive with the latter reagent.
Colony morphology of B. trematum can result in misidentification as coagulasenegative staphylococci. Its most frequent location in clinical practice being chronic wounds, resembling skin microbiota, a rather marginally significant finding in this context, which makes it prone to overlooking. Given that B. trematum is then also frequently misidentified when identification is performed by means other than MALDI-TOF MS or sequencing, a significant number of isolates might currently be missed. Biochemical identification systems frequently misidentify it as Bordetella avium, Bordetella bronchiseptica, or Achromobacter denitrificans [2][3][4]. In our study, biochemical identification systems misidentified B. trematum as Bordetella hinzii, Acinetobacter lwoffii, and Alcaligenes faecalis. Hence, we postulate this organism to be more prevalent than currently assumed and it to be more frequently detected in the future, if biochemical identification means will be largely replaced by mass spectrometry and DNA sequencing.
B. trematum is generally isolated from necrotic lesions or sites with impaired tissue perfusion such as chronic ulcers, even though multiple reports have emerged highlighting the organism's capability of invading the bloodstream, leading to bacteremia and even sepsis [6][7][8][9]. None of our three clinical B. trematum isolates caused invasive disease. While surgical debridement surely represents the treatment of choice for most chronic ulcers, systemic B. trematum infection requires antimicrobial therapy. Generally, poly-microbial infections of chronic wounds are always best tackled by an interdisciplinary approach, involving surgeons as well as microbiologists.
Chronic, non-healing wounds are frequently hypoxic [19]. Cell death, tissue necrosis, and an impaired immune response, all linked to decreased tissue perfusion, create ideal growth conditions for a large number of microorganisms, thus explaining the polymicrobial nature of such infections [20]. Since microorganisms from a variety of sources are presented with an opportunity to colonize this unnatural habitat, microbial interactions unique to this particular environment may significantly influence wound pathogenesis and healing [21]. Antimicrobial treatment that covers potentially synergistic microorganisms such as B. trematum in addition to known pathogens should be considered in cases in which improvement stays out after appropriate antimicrobial coverage of the latter. B. trematum is frequently isolated along with other agents, most commonly S. aureus, Enterobacterales and other non-fermenters [2,3,8,10,11]. Nevertheless, as was the case with one of our isolates, cases in which B. trematum is the only agent isolated from a clinical specimen do occur [7,13]. Given the higher likelihood for chronic wounds in patients with diabetes and peripheral arterial and/or cardiovascular disease, these patients have been prominently featured, among those detected with B. trematum. All three of our isolates belonged to patients with underlying cardiovascular disease; however, none had a diagnosis of diabetes.
As regards documented B. trematum antibiotic susceptibility profiles we see some similarities. Except in one case [13], in which identification occurred by unstandardized means and the susceptibility profile rendered the results even more questionable, all B. trematum isolates described in the literature were susceptible to tested carbapenems and piperacillin-tazobactam and showed markedly decreased susceptibility to 2nd and 3rd generation cephalosporins and fluoroquinolones. High MIC-values for aminoglycosides and aminopenicillins [2,6,10,13] were not uncommon. That we identified genes conferring decreased susceptibility to aminoglycosides, fosfomycin, and fluoroquinolones on the genomes of all three isolates, together with the observation that most if not all isolates described in the literature showed decreased susceptibility to these substances, suggests that such genes are common among B. trematum strains. To our knowledge, we describe the first documented B. trematum isolate phenotypically resistant to trimethoprim/sulfamethoxazole and detected the sul-2 gene on its genome, likely responsible for this resistance encoding forms of dihydropteroate synthase that are not inhibited by the substance, and thus invite to rethink the assumption that B. trematum can safely be considered as susceptible to this substance [12].
We also detected the same isolate to carry a beta-lactamase-encoding gene. The capability of B. trematum to carry and express such genes could explain why some described isolates display high MIC-values for aminopenicillins.
This study has several limitations. It is a single center study with three isolates. The two susceptibility-testing systems were commercially available test systems and not the recommended in-house systems; accuracy of these certified test systems should, however, be comparable to in-house systems. No specific breakpoints exist for the interpretation of the minimal inhibitory concentration, hence interpretation followed EUCAST PK/PD breakpoints.

Materials and Methods
After completion of the routine diagnostic procedures the isolates were cryo-conserved at −78 • C. For this study the isolates were thawed and sub-cultured twice on Columbia 5% sheep blood agar (Becton Dickinson, Heidelberg, Germany) prior to testing.

Susceptibility Testing
Susceptibility of all isolates was determined thrice by broth microdilution and gradient strip. Susceptibility tests were employed strictly according to the manufacturer's instruction. From each isolate, a bacterial suspension in 0.9% saline solution was prepared. The suspension was adjusted to a McFarland value of between 0.48 and 0.52 using a Den-siCHEK plus photometer (bioMerieux, Marcy-l'Etoile, France). For the gradient strip test, MIC Test Strips (Liofilchem, Roseto degli Abruzzi, Italy) were used (except for piperacillintazobactam (bioMerieux, Marcy-l'Etoile, France)) and were carried out on Mueller-Hinton agar plates (Becton Dickinson, Heidelberg, Germany). MIC was visually read from the gradient strip and was rounded up to the next standard upper two-fold serial dilution value. For the broth microdilution, MIC-Plates Micronaut-S MDR MRGN-Screening and Micronaut Sepsis GN (Merlin, Bornheim, Germany) were taken. Tests were performed with Mueller-Hinton broth (Merlin, Bornheim, Germany) and were visually read. All measurements were continuously supervised head-on by two medical doctors. Each sample was visually and independently read by three trained observers. MICs were interpreted according to EUCAST PK-PD breakpoints (version 10.0, 2020). Small colony variation was evaluated by daily subcultivating cultures over a two-week-long period.

Sequencing and Analysis
Highly purified DNA was extracted from all strains using the column-based DNeasy UltraClean Microbial Kit (Qiagen GmbH, Hilden, Germany). The isolation was performed according to the manufacturer's instructions with the exception that at the end of the extraction process the DNA was eluted to 100 µL volume. The obtained DNA was qualitatively and quantitatively evaluated using the NanoDrop OneC from Thermo Fisher Scientific Inc. (Waltham, MA, USA) Dual-indexed Illumina sequencing libraries were constructed from each sample using the NexteraXT kit (Illumina, San Diego, CA, USA), pooled, and sequenced on the Illumina MiSeq platform (Illumina, San Diego, CA, USA). Raw reads have been uploaded to the Sequence Read Archive (SRA); accession PRJNA747254. For de novo assembly, paired-end reads were trimmed and filtered with BBDuk Trimmer, a Q value of 20 was chosen. De novo assembly was then performed using Geneious Prime (software version 2020.1 Biomatters, Auckland, New Zealand) and multi-contig draft genomes (<150 contigs per isolate; N50 varied between 78,865 and 113,381) were generated for each isolate, by Mauve-aligning the contigs to the reference genome sequence NZ CP016340.1 of Bordetella trematum strain F581 and subsequent concatenation. Genome analysis was performed with software tools of the CGE Server (Update 8 June 2020, Center for Genomic Epidemiology, DTU, Denmark). Complete bacterial genome annotation was performed using RAST (Annotation scheme RASTtk).

Conclusions
Thanks to technological progress in the identification of microorganisms Bordetella trematum, mainly isolated from chronic open ulcers and wounds on the lower extremities, is being increasingly identified. As its role in the pathogenesis of these conditions does not appear to always be causative by itself, antibiotic treatment might not be necessary in every patient, but can improve the outcome in some. In synopsis with previously documented cases our study shows that identifications of B. trematum are increasing. In cases in which antibiotic treatment is necessary, awareness that B. trematum can be resistant to trimethoprim/sulfamethoxazole is warranted.  Institutional Review Board Statement: Patients were not exposed to additional sampling or interviewing, all the processes were according to the standards of care.

Informed Consent Statement: Not applicable.
Data Availability Statement: Raw reads have been uploaded to the Sequence Read Archive (SRA); accession PRJNA747254.

Conflicts of Interest:
The authors declare no conflict of interest.