The antibiotic susceptibility of water-based bacteria Ralstonia pickettii and Ralstonia insidiosa

Ralstonia pickettii and Ralstonia insidiosa are waterborne bacteria that can survive and grow in various water sources, and that are emerging pathogens in hospital settings. Not much is known about the antibiotic resistance of these bacteria. Previous reports of antimicrobial susceptibility have been largely limited to a few clinical strains with no accounting for genotypic or phenotypic diversity or that these species could vary from the set breakpoints. Etests and disc diffusion tests were carried out to compare the antimicrobial susceptibilities to 12 different antibiotics of 68 different isolates of R. pickettii (53) and R. insidiosa (15) from varying environments, which have previously been well characterized both phenotypically and genetically. The majority of the R. pickettii and R. insidiosa isolates showed susceptibility to most of the antibiotics tested in this study. The most effective were found to be the quinolones and sulfamethoxazole/trimethoprim. Antibiotic susceptibility was also found not to vary between environmental niches for R. pickettii and R. insidiosa isolates.


INTRODUCTION
Ralstonia pickettii is abundant in the environment having been isolated from a wide array of environmental sources (Ryan et al., 2011c).The bacterium has been recovered from numerous water sources, including municipal drinking water supplies (Lee et al., 2010), bottled water (Falcone-Dias et al., 2012), dental water supplies (Szyman ´ska 2006), hospital water supplies (Kendirli et al., 2004;Ryan et al., 2006), space shuttle water systems (Koenig & Pierson 1997), standard purified water (Penna et al., 2002), laboratorybased high-purity water systems (Adley et al., 2005) and industrial ultra-pure/high purity water (Kulakov et al., 2002;Adley & Saieb 2005;Bohus et al., 2010).R. pickettii has also been identified as forming and maintaining biofilm in plastic industrial water piping (Anderson et al., 1990;Adley & Saieb 2005).In addition, the bacterium has been found in a wide variety of clinical environments worldwide and has become recognized as a nosocomial pathogen that is particularly associated with patients who are immunosuppressed or are in some other way debilitated (Ryan et al., 2006).Conditions associated with R. pickettii range from range from minor infections to more severe invasive infections such as sepsis or meningitis.Minor respiratory illnesses were found in 34 patients in an outbreak due to contaminated solutions (Labarca et al., 1999).More invasive infections, such as osteomyelitis (Wertheim & Markovitz 1992), have been reported in a 71-year-old man suffering from chronic renal failure, and in a 75-year-old women in Bulgaria who was also suffering with renal failure and was reported to have a case of R. pickettii-related renal failure sepsis (Strateva et al., 2012).Cases of meningitis have also been reported associated with R. pickettii, including that of a 54-year-old male who was otherwise healthy (Heagney, 1998).These infections have been recorded in association with contamination of hospital water supplies such as respiratory solutions (water based) and water for injection (Gardner & Shulman, 1984;McNeil et al., 1985;Roberts et al., 1990;Maki et al., 1991;Raveh et al., 1993;Labarca et al., 1999).The majority of tested isolates of R. pickettii display multiresistance to common antibiotics (Zellweger et al., 2004).The bacterium has shown itself to be very resilient to treatment in water supplies, with resistance to disinfectants such as chlorhexidine, and under certain circumstances R. pickettii has been shown to penetrate 0.2 micron filters (Sundaram et al., 1999(Sundaram et al., , 2002;;Adley et al., 2005).R. pickettii has been shown to survive in low nutrient (oligotrophic) conditions (McAlister et al., 2002).In addition, the bacterium has been shown to possess a wide range of biodegradative abilities that could be used for commercial applications (pollution clear ups) and that could assist in the survival and adaptation of the organism to low nutrient environments (Ryan et al., 2007).Integrating conjugative elements (ICE) (such as Tn4371) have been found in various isolates of this bacterium, signifying a degree of plasticity in their genomes (Ryan et al., 2009;Van Houdt et al., 2012).
Abbreviations: CLSI, Clinical and Laboratory Standards Institute; MIC 50 , the minimal concentration of antibiotic capable of inhibiting 50 % of the isolates tested; MIC 90 , the minimal concentration of antibiotic capable of inhibiting 90 % of the isolates tested.Ralstonia insidiosa is the most closely related bacterium to R. pickettii (Ryan et al., 2011a) and has also been isolated from natural water sources such as ponds and rivers, soil, activated sludge (Coenye et al. 2003), laboratory-purified water systems (Ryan et al., 2011b), industrial ultra-pure/ high purity water (Ryan et al., 2011b) and water distribution systems (Hoefel et al., 2005).It has been reported to be the causative agent of hospital-based infections in two immunocompromised individuals (Van der Beek et al., 2005) and was found to be the cause of sepsis in eight haemodialysis patients in the Czech Republic (Orlı ´kova ´et al., 2011) due to contaminated haemodialysis solutions.Both bacteria have been found in the lungs of cystic fibrosis sufferers (Coenye et al., 2003).
The treatment of R. pickettii infections is often challenging as this bacterium has been reported as being intrinsically resistant to many antimicrobial agents (Zellweger et al., 2004;Ryan et al., 2006), which could be due to the presence of mobile genetic elements (Ryan et al., 2009).The antimicrobial resistance mechanisms in Ralstonia species are generally unknown; however, two specific mechanisms of resistance to some b-lactam antibiotics have been identified.Two different types of class D b-lactamases have been identified: bla OXA-22 and bla OXA-60 and these genes have contributed to the multidrug-resistant phenotype displayed by R. pickettii.bla OXA-22 has activity against benzylpenicillin, cloxacillin and restricted-spectrum cephalosporins, and bla OXA-60 has activity against imipenem (Nordmann et al., 2000;Girlich et al., 2004).R. pickettii and R. insidiosa are a growing threat in clinical situations; this is mainly due to their presence in water and their ability to survive in purified/distilled water used to make-up medicinal products in hospitals for patient treatment, their ability to survive in nutrient-limited environments and their ability to survive treatment with common clinical antimicrobial agents such as chlorhexidine (Adley et al., 2005;Ryan et al., 2006Ryan et al., , 2011c)).It is thought that infection with Ralstonia spp. is mainly due to environmental sources such as contaminated water supplies (Ryan et al., 2006), similar to Brevundimonas diminuta (Han & Andrade, 2005).
There is limited information on the surveillance and monitoring of the antibiotic resistance of these bacteria, and even less information on any possible differences in antibiotic resistance between clinical and environmental isolates.This study, which used a unique collection of R. pickettii and R. insidiosa (68 isolates in total) that included clinical and environmental isolates, reports the antimicrobial susceptibility profile of R. pickettii and R. insidiosa isolates, as well as comparing the Etest and disc diffusion methodologies for determining the antimicrobial susceptibility of these bacteria.

METHODS
R. pickettii and R. insidiosa strains.Sixty-eight isolates of a unique culture collection including 53 R. pickettii and 15 R. insidiosa were examined; 32 isolates came from industrial purified water, 11 from various laboratory purified water sources, eight from various clinical sources, six were isolated from washing water from an endoscopy unit, two (whole genome sequenced isolates) from a heavy metal-contaminated lake, two were purchased soil strains, five were purchased strains of R. pickettii and two were purchased strains of R. insidiosa.They came from various sources, including the BCCM/LMG Bacteria Collection (Ghent, Belgium); the Japan Collection of Microorganisms (Hirosawa, Wako-shi, Japan); the National Collection of Type Cultures (London, UK); the Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (Braunschweig, Germany); the Collection Bacte `rienne de l'Institut Pasteur (Paris, France), the American type Culture Collection (Manassas, VA, USA); and the Culture Collection, University of Go ¨teborg (Go ¨teborg, Sweden).The full list can be seen in Table 1.The isolation, identification and genotyping of all these isolates is dealt with in a previous paper (Ryan et al., 2011b).Briefly, these isolates were identified using a combination of biochemical and PCRbased methods of identification.Genotyping was carried out using a whole genome approach with RAPD (random amplification of polymorphic DNA) and BOX PCR (Ryan et al., 2011b.).All isolates were stored at 280 uC in nutrient broth (Oxoid) with 50 % glycerol.Isolates were grown aerobically on nutrient agar (Oxoid) and incubated overnight at 30 uC.
Antimicrobial susceptibility disc diffusion tests.The antibiotic discs were purchased from Oxoid.They are listed in Table 2.All tests were carried out on Mu ¨ller-Hinton (MH) agar (Oxoid) according to the Clinical and Laboratory Standards Institute (CLSI) standards (CLSI, 2013).As there are no CLSI susceptibility breakpoints available for R. pickettii or R. insidiosa, the antibiotic susceptibility results were interpreted using the CLSI criteria for Pseudomonas sp., Burkholderia cepacia and Acinetobacter spp.Zone diameters for the disc susceptibility tests were measured with vernier callipers.All testing was carried out in triplicate and all discs were within their use by date.Pseudomonas aeruginosa ATCC 27853 was included with each testing session.All results were found to be within recommended limits, demonstrating the validity of the testing procedures used.
MIC determination using Etests.The MICs of the 68 isolates of R. pickettii and R. insidiosa to the 12 antibiotics listed in Table 3 were determined using Etests (Biodisk) on MH agar.The MH agar plates were inoculated using a sterile cotton swab with suspensions of the cultures (made to 0.5 McFarlane standard) diluted with sterile tryptone soya broth (Oxoid) and resulted in confluent growth.The MICs were read after incubation overnight at 35 uC.The MICs for the bactericidal drugs (aztreonam, ceftazidime, ciprofloxacin, cefotaxime, gentamicin, meropenem, ofloxacin and piperacillin) were defined as the point of intersection between the ellipse edge and the Etest strip where there was complete inhibition of all growth.The MIC for the bacteriostatic drugs (tetracycline, minocycline and sulfamethoxazole/ trimethoprim) was read at 80 % inhibition.There are no CLSI susceptibility breakpoints for R. pickettii or R. insidiosa, so the antibiotic susceptibility testing results were interpreted using the CLSI criteria for Pseudomonas sp., Burkholderia cepacia and Acinetobacter spp.(CLSI, 2013).Pseudomonas aeruginosa ATCC 27853 was used as a control.
Statistical analysis.Simple linear regression analysis was applied to define linear functions correlating the zone of inhibition (mm) with MICs obtained by disc diffusion and Etest (mg l 21 ).The strength of the linear association between pairs of variables was determined by coefficients of determination (R-square): R-square ¢50 %, strong correlation; R-square 25-49 %, moderate correlation; and R-square ,25 %, weak correlation.The disc diffusion method was accepted as the reference method.Categorical agreement was defined if the tests results were within the same susceptibility category, and errors of disc IP: 54.70.40.11 On: Sun, 02 Dec 2018 02:03:56 diffusion and Etest methods were determined as follows: very major error, resistant by disc diffusion method, susceptible by Etest method; major error, susceptible by disc diffusion method, resistant by Etest method; and minor error, intermediate result obtained by one method but not the other (Tatman-Otkun et al., 2005).Percentage errors were calculated based on the total number of isolates that were tested.A good agreement was defined as complete category agreement over 90 %, and the total of very major and major errors below 5 %.The Pearson correlation was calculated using SPSS v20.

MIC determination using Etests
The Etest MIC results are presented in the Table 3 and used the CLSI breakpoints for P. aeruginosa, Acintobacter and Burkholderia cepacia for interpretation (CLSI, 2013).Nearly all the isolates of both bacteria were highly resistant to the aminoglycoside gentamicin (.256 mg ml -1 , 51 isolates of R. pickettii and 15 isolates of R. insidiosa) and the b-lactam antibiotic aztreonam (.256 mg ml -1 , 52 isolates of R. pickettii and 15 isolates of R. insidiosa) and variably resistant to the ticarcillin/clavulanic acid mix (.256 mg ml -1 , 31 isolates of R. pickettii and 14 isolates of R. insidiosa).For the carbapenem antibiotic, meropenem, six isolates of R. pickettii and one isolate of R. insidiosa were found to be resistant and the rest susceptible.All isolates were susceptible to the quinolones (ciprofloxacin and ofloxacin), the tetracyclines (tetracycline and minocycline), the cephalosporins (cefotaxime and ceftazidime), the folate pathway inhibitor (trimethoprim/sulfamethoxazole) and the extended spectrum b-lactam antibiotic of the ureidopenicillin class (piperacillin).

Similar variation between MIC and disc diffusion results
was found in a comparable study that was carried out on Stenotrophomonas maltophilia (Nicodemo et al., 2004, Tatman-Otkun et al., 2005).
No major differences were observed between the MICs of different R. pickettii isolates from differing environmental niches except for the ticarcillin/clavulanic acid mix where the isolates from the industrial purified water supplies and the purchased isolates were all resistant, whereas the   clinical isolates, and hospital and laboratory purified water isolates were susceptible.This is interesting as it suggests the possibility of a mechanism of resistance in environmental isolates that is not present in clinical isolates.
The MICs of some antibiotics found in the present study were broadly similar to those reported previously for R. pickettii.For example Sader & Jones (2005), Gales et al. (2005) and Fung-Tomc et al. (1997) reported the MIC 50 (the minimal concentration of antibiotic capable of inhibiting 50 % of the isolates tested) and the MIC 90 (the minimal concentration of antibiotic capable of inhibiting 90 % of the isolates tested) of ceftazidime as .16mg ml -1 , whereas our isolates had an MIC 50 of 4 mg ml -1 and an MIC 90 of 12 mg ml -1 .This was also true for meropenem, tetracycline and ticarcillin/clavulanic acid, where similar levels of resistance were found to those in our study (Fung-Tomc et al., 1997;Sader & Jones 2005;Gales et al., 2005).However, for trimethoprim/sulfamethoxazole (an MIC 50 of 0.012 mg ml -1 and an MIC 90 of 0.023 mg ml -1 in our study and an MIC 50 of ¡0.5 and an MIC 90 of 1/.2 mg ml -1 in published studies) and ciprofloxacin (an MIC 50 of 0.06 mg ml 21 and an MIC 90 of 0.12 mg ml -1 in our study and an MIC 50 of 0.25/0.5/2mg ml -1 and an MIC 90 of 2/4 mg ml -1 in published studies) higher levels of resistance were found than in our study.Also, in our study much higher levels of resistance were found to aztreonam (an MIC 50 and an MIC 90 of .256mg ml -1 and an MIC 50 of .8 and an MIC 90 of 16/64 mg ml -1 in the previously published studies) and gentamicin (MIC 50 and an MIC 90 of .256mg ml -1 and an MIC 50 and an MIC 90 of .8mg ml -1 in the previously published studies) than in the published studies.These studies were, however, carried out with a limited number of clinical isolates (38, 10 and 14, respectively) compared with our study's 53 environmental and clinical isolates.
Resistance to the ticarcillin/clavulanic acid mix in some strains could possibly be due to the presence of the bla OXA-22 and bla OXA-60 genes, which have been identified in both R. pickettii 12J (Rpic_3817 and Rpic_3962) and 12D (Rpic12D_3930 and Rpic12D_4075) genomes.Mutational studies have shown that these genes confer resistance to ticarcillin and to a ticarcillin/clavulanic acid mix.These bla OXA-22 and bla OXA-60 gene products have also been shown to raise MIC values for carbapenems, such as meropenem and imipenem, and for cephalosporins, such as cefepime and cefotaxime (Nordmann et al., 2000;Girlich et al., 2004).
Resistance to aztreonam in Gram-negative bacteria (like R. pickettii and R. insidiosa) is usually due to extendedspectrum b-lactamases (Franceschini et al., 1998).Several genes for these extended-spectrum b-lactamases can be found in the genomes of R. pickettii 12J (21 proteins) and 12D (20 proteins).
These results indicate that sulfamethoxazole/trimethoprim and the fluoroquinolone ciprofloxacin are the best antibiotics with which to treat infections with R. pickettii and R. insidiosa.This is shown by the agreement of the statistical analysis with the experimental data shown.This agreed with the literature data, which showed that ciprofloxacin worked to treat R. pickettii infections after the failure of other treatments (Kendirli et al., 2004;Woo et al., 2002).However, issues such as pharmacokinetics and pharmacodynamics and clinical experience should be taken into consideration when choosing the best antibiotic to treat infections.This is believed to be the first in-depth study on the antibiotic resistance of a significant number of R. pickettii isolates, which is vital in determining the appropriate treatment in cases of infection.Our results suggest that infection with R. pickettii and/or R. insidiosa could be treated orally with quinolones or trimethoprim/sulfamethoxazole, which would reduce the invasiveness of treatment (Table 3).In-depth clinical studies are essential to confirm the in vivo effectiveness of these antibiotics for the treatment of R. pickettii/R.insidiosa infections, and to assess the correlation between the susceptibility testing results and the clinical outcomes of treatment.These results indicate that there is little difference in antibiotic resistance between clinical and environmental isolates of the two bacteria; this is of interest as most clinical infections with R. pickettii are thought to come from environmental sources such as purified water supplies.

Table 1 .
Ralstonia strains used in this work

Table 2 .
Antibiotic resistance profile of 53 R. pickettii and 15 R. insidiosa isolates using disc diffusion testing CLSI breakpoints for resistance for P. aeruginosa (table 2B-1) were used for the antibiotics, except for ceftazidime, meropenem, minocycline and trimethoprim/sulfamethoxazole where breakpoints for resistance for Burkholderia cepacia (table