The Bacteroides fragilis Group: The Next Superbugs in Waiting?

Treatment of infections caused by anaerobic bacteria remains largely empiric, despite recent knowledge and reduced antimicrobial susceptibility to commonly used antibiotics. The objective of this study was to obtain current surveillance data for the Bacteroides fragilis group in Australia. We obtained 255 faecal isolates, 100 derived from the community, 98 derived from hospital high-dependency unit samples, and 57 derived from blood cultures and other sterile sites. Susceptibility testing was performed using both the BSAC disc diffusion method and the CLSI agar dilution method. Molecular markers of resistance (erm genes, nim genes, gyrA mutations) were tested to determine the presence of resistance to clindamycin, metronidazole and fluoroquinolones respectively, and to assess their potential diagnostic use. By comparison with previous Australian studies last conducted in 1992, overall resistance of the Bacteroides fragilis group in Australia has increased from 11% to 44% for clindamycin, 1% to 3% for the carbapenems, and from 1% to 9% for ticarcillin/clavulanic acid. No resistance to metronidazole was observed. At species level, B. fragilis isolates (n=82) showed low rates of resistance, whilst B. distasonis (n=47) displayed high resistance to clindamycin (83%) and ticarcillin/clavulanic acid (21%), B. uniformis (n=32) showed high resistance to moxifloxacin (53%) and meropenem (9%), and B. thetaiotaomicron (n=31) presented high resistance to clindamycin (58%) and ticarcillin/clavulanic acid (19%). Comparison of the two susceptibility methods used demonstrated a lack of correlation. Examples of all molecular markers of resistance tested were found within the study population, with 66/255 strains harbouring at least one erm gene, 2/255 strains harbouring nim genes, and 2/6 selected strains carrying a gyrA mutation. Authorship Helen Boyd, Helen Agus,, John Merlino, Thomas Gottlieb, Lee Thomas, Tom Olma , Jon Iredell, Steven Siarakas* School of Molecular and Microbial Biosciences, University of Sydney; Department of Microbiology and Infectious Diseases, Concord Repatriation General Hospital; and Institute of Clinical Pathology and Medical Research, Westmead Hospital, Sydney, Australia * Corresponding author: Mailing address: Department of Microbiology and Infectious Diseases, Concord Repatriation General Hospital, Hospital Road, Concord, 2139 Sydney, New South Wales, Australia. Phone: +61-2-9767-5459. Fax: +61-2-9767-7868. E-mail: [email protected] INTRODUCTION The culture, isolation and identification of anaerobic bacteria has long been considered more time-consuming, labour-intensive and expensive than that of aerobic bacteria, and is therefore considered to be too demanding for many diagnostic microbiology laboratories (7). As a result, the culture and isolation of anaerobic bacteria is often either minimal or omitted, lending a bias to the identification of aerobes or facultative anaerobes as the infective agent. However, when all the appropriate investigations are undertaken, anaerobes have been found to comprise a significant proportion of most infections. Overall, anaerobic bacteria are implicated in up to 20% of bacteraemias, 70% of bite wounds, 90% of community-acquired aspiration pneumonia, and 73% of appendicitis cases (18). Anaerobes are also known to be associated with many diverse infections including intra-abdominal infection, brain abscess, otogenic meningitis, chronic sinusitis, breast abscess, vaginitis, post-partum sepsis, endometritis, gas gangrene, clostridial myonecrosis, perirectal abscess and diabetic foot infections. The members of the Bacteroides fragilis group have been widely recognised as the most common anaerobic bacteria encountered in clinical specimens (13, 22). With the increasing recognition of the importance of anaerobic bacteria as pathogens, as well as growing suspicion and interest regarding changing susceptibility profiles, a large number of laboratories world-wide have undertaken investigations into their local anaerobic population (3, 4, 5, 6, 15, 25, 28, 34, 35, 36, 38, 39, 41, 42, 47, 50, 60, 63). In Europe, North America and Japan, this has included the formation of large study groups with a view to conducting regular surveillance studies (22, 30, 31, 32, 33, 40, 52, 53, 54, 56, 57, 58). In contrast, there is little recent Australian data on the antibiotic susceptibility patterns of the local anaerobic population, with the last such study having been conducted in 1992 (9). This has meant that the selection of antibiotic treatment is empiric, based on identification of the suspected pathogen and the anticipated susceptibility pattern. The antibiotics most commonly used to treat infections caused by the members of the Bacteroides fragilis group are currently metronidazole, clindamycin, the carbapenems and the β-lactam/β-lactamase-inhibitor combinations. Recent anaerobe research has also focused on investigation of molecular markers of anaerobic resistance as potential screening methods. The implicated molecular marker has been identified for the main antimicrobials used in the treatment of infections caused by members of the Bacteroides fragilis group. Research to date has been primarily focused on the erm genes associated with clindamycin resistance (21, 25, 41, 55), the nim genes associated with resistance to metronidazole (14, 16, 20, 24, 29, 50, 62), and mutations in the quinolone-resistancedetermining region (QRDR) of the gyrA gene that has been related to fluoroquinolone resistance (37, 43, 45). This study presents the first Australian epidemiological data into the susceptibility patterns of the Bacteroides fragilis group for over 14 years. In addition, two of the most commonly used methods of susceptibility testing were compared for reproducibility and correlation of results. Finally, we also present the first Australian data looking at the prevalence of some of the more common molecular markers of resistance found in Bacteroides species: erm genes (ermF, ermG and ermB), nim genes and gyrA mutations. MATERIALS AND METHODS Bacterial Strains. Faecal samples were collected from 125 community and 256 hospital high dependency unit (HDU) patients during the six-month period between February and July 2004. An HDU patient was considered to be one currently admitted to an intensive care unit (ICU), high dependency unit (HDU) or burns unit who had been hospitalized for no less than 48 hours. Following collection, each sample was plated directly onto selective Bacteroides Bile Esculin (BBE) agar (26). Isolates were identified to genus level by anaerobic growth, presentation of aesculin hydrolysis and Gram morphology. This resulted in 100 community and 98 HDU isolates. Identification to species level was conducted using RapIDTM ANA II test panels (Remel INC, USA). An additional clinical group of 57 previously identified isolates from blood culture and other sterile site infections was also included. Control strains used in susceptibility testing were Bacteroides fragilis ATCC 25285 and Bacteroides thetaiotaomicron ATCC 29741. The positive control strain used in screening for nim genes was strain BF8 (19). The erm gene screening positive controls were strain BFV479 containing the pBF4 plasmid known to harbour the ermF gene (64) and a wild-type Bacteroides fragilis determined during the course of this study to harbour both the ermB and ermG genes. Antimicrobial Agents. The following antimicrobial agents were investigated in this study: penicillin G and metronidazole (Sigma-Aldrich); moxifloxacin (Bayer Pharmaceuticals); meropenem (Astra Zeneca); ticarcillin and clavulanic acid (Glaxo-Smith Kline); and clindamycin (Pfizer). All powders were stored at –20°C prior to use. Susceptibility Testing. MIC determination was conducted on all 259 isolates using the CLSI agar dilution method. The antimicrobial agents were suspended, diluted and incorporated into brucella agar supplemented with 5% laked sheep blood according to the CLSI published standard (10). The inoculum of each isolate was prepared by suspending colonies from a 48hour-old anaerobic blood agar plate culture directly into 3mL of brucella broth to a density equivalent to that of a 0.5 McFarland turbidity standard. A Steers replicator was used to inoculate the plates, delivering approximately 10 CFU per spot. Plates were incubated anaerobically for 48 hours at 35°C in a Modular Atmosphere Controlled System anaerobic chamber (Don Whitley Scientific, UK). The MIC was defined as the lowest concentration of antibiotic that caused a marked reduction in the appearance of growth compared to the control plate. Susceptibility testing was also performed on the community and HDU isolates using the disc diffusion method outlined by the British Society for Antimicrobial Chemotherapy (BSAC) (8). The inoculum of each isolate was prepared by suspending colonies from a 48-hour-old anaerobic blood agar plate culture directly into 5mL of sterile distilled water to a density equivalent to that of a 0.5 McFarland turbidity standard. Each inoculum was swabbed evenly over the surface of a brucella agar plate supplemented with 5% horse blood and 1mg/L vitamin K. The following antimicrobial susceptibility testing discs (Oxoid, Australia) were placed evenly over the surface of the inoculated plate: clindamycin (2μg), meropenem (5μg), metronidazole (5μg), moxifloxacin (5μg), penicillin G (2units), and ticarcillin/calvulanic acid (7:5.1; 85μg). Plates were incubated anaerobically for 48 hours at 35°C. DNA Extraction. The target DNA was extracted using a previously described rapid extraction procedure (61). Bacterial colonies were picked from agar and suspended in 100μL of 1x TE buffer, pH 7.5. The suspension was heated for 10 minutes at 95°C and centrifuged at 13,000rpm for 5 minutes. 90μL of the DNA-containing supernatant was collected and stored at –20°C until use. ermF gene amplification. PCR amplification was performed using the previously described primers ERMF1 (5’-CGG GTC AGC ACT TTA CTA TTG-3’) and ERMF2 (5’-GGA CCT ACC TCA TAG ACA AG-3’) (46) synthesised by Invitrogen, Australia. Each assay was carried out in a final volume of 25μL containing 12.5μL BioMix ready-to-go reaction mixture (Bioline, London, UK), each primer at a concentration of 25pM, 5μL of template DNA and sterile distilled water to volume. An initial denaturation step of 94°C for 10 minutes was followed by 32 cycles of amplification consisting of denaturation at 94°C for 30 seconds, annealing at 62°C for 1 minute and extension at 72°C for 1 minute, and a final extension step of 72°C for 10 minutes. 10μL of each PCR product was analysed by electrophoresis in TBE buffer on a 1.5% (w/v) agarose gel containing ethidium bromide. Template DNA of strain BFV479 was included in each PCR run as a positive control and DNA fragments of approximately 466-bp were considered indicative of the presence of an ermF gene. ermB and ermG gene amplification. Screening of isolates for the ermB and ermG genes was conducted as a multiplex PCR reaction. PCR amplification of the ermB gene was carried out using primers ERMB1 (5’-GAA AAG GTA CTC AAC CAA ATA-3’) and ERMB2 (5’-AGT AAC GGT ACT TAA ATT GTT TAC-3’). Amplification of the ermG gene was carried out using primers ERMGfor1 (5’-ACA TTT CCT AGC CAC AAT C-3’) and ERMGrev1 (5’-CGC TAT GTT TAA CAA GC-3’). Both primer pairs were previously described (55) and were synthesised by Invitrogen, Australia. Each assay was carried out in a final volume of 25μL containing 12.5μL BioMix ready-to-go reaction mixture (Bioline, London, UK), each primer at a concentration of 25pM, 5μL of template DNA and sterile distilled water to volume. PCR conditions were as previously described (55). An initial denaturation step of 95°C for 5 minutes was followed by 30 cycles of amplification consisting of denaturation at 95°C for 1 minute, annealing at 52°C for 1 minute and extension at 72°C for 2 minutes, and a final extension step of 72°C for 5 minutes. 10μL of each PCR product was analysed by electrophoresis in TBE buffer on a 1.5% (w/v) agarose gel containing ethidium bromide. Template DNA from a wild type B. fragilis strain found to possess both the ermB and ermG genes was included in each run as a positive control. DNA fragments of approximately 639-bp were considered to be indicative of the ermB gene, whilst DNA fragments of approximately 442-bp were considered to be indicative of the ermG gene. nim gene amplification. Previously described primers NIM-3 (5’-ATG TTC AGA GAA ATG CGG CGT AAG CG-3’) and NIM-5 (5’-GCT TCC TTG CCT GTC ATG TGC TC-3’) were used for the universal PCR amplification of the six members of the nim gene family (61). Primers were synthesised by Invitrogen, Australia. Each PCR assay was carried out in a final volume of 25μL containing 12.5μL of BioMix ready-to-go reaction mixture (Bioline, London, UK), each primer at a concentration of 25pM, 5μL of template DNA and sterile distilled water to volume. PCR conditions were as previously described (61). An initial denaturation step of 94°C for 10 minutes was followed by 32 cycles of amplification consisting of denaturation at 94°C for 30 seconds, annealing at 62°C for 1 minute and extension at 72°C for 1 minute, and a final extension step of 72°C for 10 minutes. 10μL of each PCR product was analysed by electrophoresis in TBE buffer on a 1.3% (w/v) agarose gel containing ethidium bromide. Template DNA from strain BF8 was included in each run as a positive control and DNA fragments of approximately 458-bp were considered to be indicative of a nim gene. gyrA amplification. PCR amplification of the gyrA QRDR of selected strains was carried out with previously described primers Pr-BFGBA03 (5’-ATG CTT GAA CAA GAC AGA ATT ATA AAG-3’) and Pr-BFGA02 (5’-GAC TGT CGC CGT CTA CAG AAC CG-3’) (37) synthesised by Invitrogen, Australia. Each assay was carried out in a final volume of 100μL containing 50μL of BioMix ready-to-go reaction mixture (Bioline, London, UK), each primer to a concentration of 25pM, 5μL of template DNA and sterile distilled water to volume. PCR conditions were as previously described (37): 25 cycles of denaturation at 94°C for 30 seconds, annealing at 62°C for 5 minutes, and extension at 72°C for 1 minute. An initial denaturation step of 94°C for 5 minutes and a final extension step of 72°C for 5 minutes were added to the protocol. Free primers and nucleotides were removed from the PCR products by the JETQUICK PCR Purification Spin Kit (Genomed, Germany). 10μL of purified PCR product was analysed by electrophoresis in TBE buffer on a 1% agarose gel containing ethidium bromide. DNA fragments of approximately 282 to 296-bp were obtained as expected. PCR products were sent to the Sydney University Prince Alfred Macromolecular Analysis Centre (Sydney, Australia; http://www.supamac.com) for genetic sequencing using ABI PRISM dye terminator cycle sequencing.