Mechanisms of-lactam resistance amongst Pseudomonas aeruginosa ...

JAC

Journal of Antimicrobial Chemotherapy (1998) 42, 697–702

Mechanisms of -lactam resistance amongst Pseudomonas aeruginosa isolated in an Italian survey Giovanni Bonfiglioa*, Younes Laksaia, Luigi Franchinob, Gianfranco Amicosantec and Giuseppe Nicolettia a

Istituto di Microbiologia, Università di Catania, via Androne 81, 95124 Catania; bDirezione Medica, Zeneca S.p.A., Palazzo Volta, 20080 Basiglio, Milano; cDipartimento di Scienze Biomediche, Università dell’Aquila, via Vetoio, Coppito, L’Aquila, Italy The mechanisms of resistance to -lactam antibiotics in 325 isolates of Pseudomonas aeruginosa were examined. These isolates were selected because of their resistance to meropenem and imipenem (breakpoint, >4 mg/L), carbenicillin (>128 mg/L), ceftazidime (>8 mg/L), piperacillin and ticarcillin/clavulanate (>64 mg/L). The most frequent mechanism of resistance was -lactamase-independent, so called ‘intrinsic resistance’, which was found in 183 isolates and was probably due to impermeability and/or efflux mechanisms. -Lactamase-mediated resistance was demonstrated in 111 strains (11.1%). Derepression of Ambler Class C chromosomal -lactamase was detected in 64 isolates, most of which were resistant to ceftazidime and piperacillin but susceptible to meropenem, whereas secondary plasmid-encoded -lactamases were found in 34 isolates, all of them resistant to carboxypenicillins and ureidopenicillins and susceptible to carbapenems. Twelve strains showed more than one plasmidencoded -lactamase plus derepression of chromosomal Class C enzyme. Resistance to carbapenems was independent of resistance to other -lactam antibiotics, indicating a different mechanism of resistance, probably due to the loss of the D2 porin. In total, 32 strains were resistant to carbapenems: 24 only to imipenem and eight to both imipenem and meropenem.

Introduction Pseudomonas aeruginosa is intrinsically resistant to many antibiotics, or can develop resistance during treatment with consequent high mortality, and is, increasingly, a cause of infection in immunocompromised patients. In 1994, it accounted for 13.9% of 25,266 consecutive isolates of aerobic bacteria and for 20.9% of 16,863 isolates of Gram-negative bacteria examined in an Italian survey of clinically significant isolates from in- and out-patients.1 During 1995, in order to assess the level of resistance to widely used antipseudomonal antibiotics in clinical isolates of P. aeruginosa, a national survey was undertaken.2 One thousand and five P. aeruginosa strains were isolated and their susceptibility to antipseudomonal antibiotics was analysed. Percentages of resistance to -lactam antibiotics tested, using NCCLS breakpoints (shown in parentheses) were as follows: meropenem

*Tel:

95-311-352; Fax:

( 4 mg/L), 9.1%; imipenem ( 4 mg/L), 19.2%; ceftazidime ( 8 mg/L), 13.4%; carbenicillin ( 128 mg/L), 27.3%; piperacillin ( 64 mg/L), 12.0%; ticarcillin/ clavulanate ( 64 mg/L), 22.8%; amikacin ( 16 mg/L), 10.6%; and ciprofloxacin ( 1 mg/L), 31.9%. In the present study the mechanisms of resistance to -lactam antibiotics in these P. aeruginosa isolates were investigated.

Materials and methods Bacterial strains During 1995, in a survey of antibiotic resistance in P. aeruginosa in Italy, 1005 strains were isolated from different in-patients in 15 hospitals.2 Three-hundred and twenty-five isolates were selected because of their

95-325-032; E-mail: [email protected]

697 © 1998 The British Society for Antimicrobial Chemotherapy

G. Bonfiglio et al. -lactam resistance. Reference P. aeruginosa strains were also included: (i) strain ATCC 27853 as the control for the susceptibility test; (ii) strain R20 as a control with known inducibility of Ambler Class C chromosomal -lactamase;4 and (iii) 2297-con, with derepressed Class C chromosomal -lactamase expression.4

Antibiotics The following antibiotics were used in this study: meropenem (Zeneca, Milan, Italy); ticarcillin and clavulanate (SmithKline Beecham, Milan, Italy); piperacillin (Lederle, Catania, Italy); imipenem (Merck Sharp and Dohme, Italy); nitrocefin (Unipath, Milan, Italy); the remaining antibiotics and reagents were from commercial sources or from Sigma (Milan, Italy). Susceptibility testing of these antibiotics has already been described.2 NCCLS3 breakpoints were used as follows: meropenem and imipenem, 4 mg/L; carbenicillin, 128 mg/L; ceftazidime, 8 mg/L; piperacillin and ticarcillin/clavulanate, 64 mg/L.

-Lactamase production in strains with derepressed Class C enzyme For enzymes that were strongly inhibited by cloxacillin and not by clavulanate, the activity of crude enzyme extracts was studied in order to assess the inducibility of Class C chromosomal -lactamase production by cefoxitin. Micro-organisms were grown overnight in 10 mL of nutrient broth at 37°C with continuous shaking, then a tenfold dilution of this culture was used to inoculate, in duplicate, fresh broth warmed to 37°C. After incubation for 90 min, cefoxitin, a strong inducer of Class C chromosomal -lactamases, was added to one broth (‘induced culture’), to a final concentration of 500 mg/L. No addition was made to the second broth (‘uninduced culture’). Incubation was continued for a further 4 h at 37°C. Cells were harvested by centrifugation at 6000g, washed in 100 mM phosphate buffer, pH 7.0 and ultrasonically disrupted. The -lactamase content of the extract was determined spectrophotometrically with 0.1 mM nitrocefin in 0.1 mM phosphate buffer, pH 7.0.6 -Lactamase yields were standardized against protein concentrations determined by the method of Lowry et al.7

-Lactamase detection and typing -Lactamase extracts were prepared as previously described.5 Briefly, cultures were grown overnight at 37°C in 25 mL volumes of nutrient broth with continuous shaking and diluted 1 in 10 (250 mL) in fresh warm broth. Incubation was then continued for 4 h. Cells were harvested by centrifugation at 6000g, washed twice, resuspended in 2 mL of 100 mM phosphate buffer, pH 7.0 in Eppendorf tubes and sonicated twice (for 30 s) with intermediate cooling. The supernatant was collected by centrifugation at 12,000g for 20 min at 4°C. The crude cell extracts were assayed against 0.1 mM nitrocefin in 100 mM phosphate buffer, pH 7.0, and enzyme activity was measured spectrophotometrically at 482 nm. For inhibition studies, the same extracts were mixed with an equal volume of 100 mM cloxacillin or clavulanate. After 5 min incubation, 0.1 mM nitrocefin was added to the mixture in the cuvette and hydrolysis was recorded.

-Lactam hydrolysis assay For enzymes that were strongly inhibited by clavulanate and not by cloxacillin, their hydrolysis of -lactam antibiotics was assayed. Briefly, each enzyme was added to 0.5 mM -lactam solutions in 100 mM phosphate buffer, pH 7.0, and reactions were followed by UV spectrophotometry in cuvettes with a 1 cm light-path cuvettes, with readings taken every 10 s at the wavelength of optimal absorbance for the -lactam ring of each antibiotic as follows: benzylpenicillin and carbenicillin, 235 nm; cephaloridine, 295 nm; oxacillin, 265 nm.

Isoelectric focusing of -lactamases -Lactamases for isoelectric focusing (IEF) were extracted as described above, with the only exception that 10 mM rather than 100 mM phosphate buffer was used. After centrifugation for 2 min at 12,000g, the supernatants were subjected to IEF, at 15 W constant power, performed on an Ampholine PAGplate (Pharmacia Biotech AB, Uppsala, Sweden). In the preliminary studies a pH gradient from pH 3.5 to pH 9.5 was used in a Pharmacia apparatus. Enzymes that focused between pH 5 and pH 6 were retested in a pH gradient of pH 4–6.5, while those that focused between pH 6 and pH 8 were retested in a gradient of pH 6.5–8.5. -Lactamase activities were visualized by flooding the IEF plates with 0.1 mM nitrocefin.8 The isoelectric points (pI) of the extracts were determined by comparison with reference enzymes placed on the same gel. The correct focusing of the reference -lactamases used and of our crude enzyme extracts was further verified by measuring the pH gradient in the gel.

Results Based on their reactions with nitrocefin and inhibitors, and on their resistance to carbapenems, we divided the 325 resistant isolates into five groups, as summarized in Table I and explained below. The first group included 183 isolates and reacted very weakly with nitrocefin as judged spectrophotometrically. This reaction was inhibited completely by cloxacillin but not by clavulanate, and probably reflects the production of Class C chromosomal -lactamase.9 This enzyme is

698

Mechanisms of -lactam resistance in P. aeruginosa inducible and is produced in P. aeruginosa strains at low levels. One hundred and forty-nine of these strains were resistant to carbenicillin and showed reduced susceptibility to ceftazidime, penicillins, ciprofloxacin and amikacin. This could be considered the typical pattern of ‘intrinsic resistance’ and was probably caused by reduced permeability and/or multi-drug efflux.10,11 The remaining 34 strains were susceptible to carbenicillin yet resistant to ceftazidime and the penicillins. This atypical pattern of resistance seems most likely to entail some combination of impermeability and efflux. The second group, which included 64 strains, rapidly hydrolysed nitrocefin and this was completely inhibited by cloxacillin but not by clavulanate. This pattern was also observed for the derepressed control 2297-con. This pattern is typical of hyperproduction of Class C chromosomal -lactamase.12–16 IEF showed that all the isolates had -lactamases with pI values of 7.5. These isolates produced at least five-fold higher uninduced levels of -lactamase than the R20 strain, confirming derepression of Class C chromosomal -lactamase as shown in Table II. Induction by cefoxitin showed that in 56 of the 64 isolates -lactamase expression remained inducible. This behaviour indicates that in these strains the derepression of Class C chromosomal -lactamase was partial but still efficient enough to give resistance to penicillins and cephalosporins. The remaining eight isolates from this group of 64 showed high -lactamase specific activity; they synthesized almost equal amounts of enzyme in the presence or absence of cefoxitin, and were totally derepressed for Class C chromosomal -lactamase, as shown in Table II. Almost all of the strains in this group were resistant to ceftazidime and, to some extent, to piperacillin. The level of resistance to meropenem was unrelated to the amount of -lactamase produced without induction, whereas resistance to penicillins, ceftazidime and imipenem was strongly related to enzyme production (Table II). For strains in the third group, crude cell extracts displayed strong hydrolysis of nitrocefin and were inhibited by clavulanate but not cloxacillin. This inhibition profile and high resistance to carboxypenicillins and ureidopenicillins suggested that these strains had secondary plasmid-mediated -lactamases.14–16 The types of plasmid- or transposon-mediated -lactamases detected in P. aeruginosa isolates are listed in Table III. The PSE-1type enzyme was found in 17 isolates and was the most common plasmid -lactamase detected, followed by the PSE-4 type (eight isolates). Two isolates from the same centre carried the same enzyme, which had a pI of 4.3 and co-focused with the CARB-4-type enzyme. Three other strains with PSE-2-type -lactamase activity all came from the same centre and probably were carrying the same type of enzyme. The remaining four isolates had OXA-type enzymes, as characterized by their stronger activity against oxacillin than carbenicillin. Isolates with PSE-1-, PSE-2699

G. Bonfiglio et al. or PSE-4-type enzymes were more resistant to carbenicillin, piperacillin and ticarcillin/clavulanate than strains with OXA-type enzymes, but were not more resistant to meropenem, imipenem or ceftazidime (Table III). The fourth group comprised 12 isolates that gave a strong reaction to nitrocefin and were not inhibited by cloxacillin or clavulanate. IEF and hydrolysis assays were performed to characterize the enzymes. All of them had one plasmid-encoded -lactamase in association with a band of pI 7.5. As judged by IEF and hydrolysis of the crude enzyme extract, eight strains had PSE-1-type enzymes and four PSE-4-type enzymes. Quantitative nitrocefin assays confirmed that all the strains had the highest -lactamase specific activities and were partially derepressed for Class C chromosomal -lactamase. These microrganisms were characterized by a strong activity against penicillins and ceftazidime. Isolates resistant to carbapenems were placed in a separate group since imipenem resistance in P. aeruginosa seems to be independent of resistance to penicillins and cephalosporins. Thirty-two strains were resistant to either imipenem or meropenem. Of these 24 were resistant to imipenem and only eight to both imipenem and meropenem. Thus the majority of imipenem-resistant P. aeru ginosa remained susceptible to meropenem. However, the activity of meropenem against these isolates was moderate because the MIC values were higher than those for the fully susceptible strains, as shown in Table I.

Discussion The number of antibiotics that can be used to treat P. aeruginosa infections is limited, because this species is intrinsically resistant to many of the antibiotics that are successfully used to treat infections due to other species. Progress has been made in increasing the range of -lactams available for treating P. aeruginosa infections, but resistance to these antibiotics can arise by various mechanisms, including mutational derepression of Class C chromosomal -lactamases,13 acquisition of secondary plasmid- or transposon-mediated -lactamases,17 and loss of D2 porin (resulting in imipenem resistance).18 Multidrug resistance has also been associated with multi-drug efflux.10,11 Three hundred and twenty-five P. aeruginosa isolates resistant to one or more -lactams were isolated during a multicentre Italian survey,2 and the mechanisms underlying their resistance were studied. On the basis of their antibiotic susceptibility profiles, hydrolysis of nitrocefin and inhibition of cloxacillin or clavulanate, we divided the isolates into five major groups. The largest group of -lactam-resistant isolates were those with poor -lactamase activity. The 183 isolates in this group showed inter-related resistance to penicillins, ceftazidime, carbapenems and ciprofloxacin. This pattern 700

Mechanisms of -lactam resistance in P. aeruginosa of resistance, also called ‘intrinsic resistance’, is typical of P. aeruginosa and was assumed to be result from impermeability or multi-drug efflux. 10,11 One hundred and fortynine of the 183 were resistant to carbenicillin while the remaining 34 were carbenicillin-susceptible but resistant to ceftazidime and other penicillins. A similar result was obtained in a recent survey in the UK.12 -Lactamase-mediated resistance was quite frequent, being found in 11.2% of the total number of P. aeruginosa strains isolated and accounting for 34.8% of the isolates resistant to -lactam antibiotics. Plasmid-encoded lactamases were found in 34 (3.4%) of the 1005 isolates. It must be borne in mind, however, that a study of this type can be distorted by outbreak strains. In fact, one hospital sent three isolates with the PSE-2-type enzyme, and another hospital sent two isolates with the CARB-4-type enzyme. Although similar work has never been performed in Italy, our data are not in agreement with results of surveys carried out in France19 and Spain,20 where such enzymes occurred in 7–12% of isolates, or in the UK,12 where these enzymes were detected in 1% of the isolates studied. However, our findings confirm that the PSE-1type enzyme is the most common plasmid-mediated -lactamase isolated in P. aeruginosa, followed by the PSE-4 type. OXA-type enzymes were found in only four isolates. As expected, isolates with a plasmid-mediated -lactamase were highly resistant to carboxypenicillins and ureidopenicillins, but susceptible to meropenem, imipenem and ceftazidime. The final mechanism studied was that affecting carbapenems. Twenty-four strains were resistant only to imipenem and only eight to both imipenem and meropenem. The difference in the MIC values of imipenem and meropenem for the imipenem-resistant strains has been attributed to (i) the poorer -lactamase-inducing ability of meropenem, (ii) improved -lactamase stability, (iii) high affinity for penicillin-binding protein 3 (PBP 3),21 or (iv) the possibility that meropenem diffuses through the outer membrane by non-specific and D2 imipenem-specific pathways.22 One or more of these mechanisms could result in higher steady-state periplasmic concentrations for meropenem and thus in a reduced MIC. Although the imipenem-resistant strains have reduced susceptibility also to meropenem, the MICs of the latter remained in the clinical range. Our findings indicate that meropenem retains activity against P. aeruginosa with the commoner type of resistance mechanisms known to affect other -lactams, including imipenem, ceftazidime and piperacillin. In conclusion, our study has considered the mechanisms whereby resistance to -lactam antibiotics can arise in clinical isolates of P. aeruginosa in Italy. The principal mechanism of resistance in Italian clinical isolates of P. aeruginosa was impermeability, followed by -lactamase production and imipenem-specific resistance. Furthermore, as Chen et al. have suggested,12 it should be 701

G. Bonfiglio et al. emphasized that the occurrence of a mechanism does not prove it to be the only cause of resistance. It is possible that some isolates may possess further, undetected mechanisms.

Acknowledgements This work was supported, in part, by a grant from Zeneca S.p.A. Italy and in part from ‘Consiglio Nazionale delle Ricerche’. The authors would like to thank Dr D. M. Livermore for the generous provision of the strains with reference -lactamases, and A. Bridgewood for invaluable assistance in preparing the manuscript.

10. Poole, K., Krebes, K., McNally, C. & Neshat, S. (1993). Multiple antibiotic resistance in Pseudomonas aeruginosa: evidence for involvement of an efflux operon. Journal of Bacteriology 175, 7363–72. 11. Li, X.-Z., Ma, D., Livermore, D. M. & Nikaido, H. (1994). Role of efflux pump(s) in intrinsic resistance of Pseudomonas aeruginosa: active efflux as a contributing factor to -lactam resistance. Antimicrobial Agents and Chemotherapy 38, 1742–52. 12. Chen, H. Y., Yuan, M. & Livermore, D. M. (1995). Mechanisms of resistance to -lactam antibiotics amongst Pseudomonas aeruginosa isolates collected in the UK in 1993. Journal of Medical Microbiology 43, 300–9. 13. Curtis, N. A., Eisenstadt, R. L., Rudd, C. & White. A. J. (1986). Inducible type-I -lactamase of Gram-negative bacteria and resistance to -lactam antibiotics. Journal of Antimicrobial Chemotherapy 17, 51–61.

References 1. Stefani, S., Bonfiglio, G., Urzì, S. & Nicoletti, G. (1998). Susceptibility survey of piperacillin/tazobactam and some -lactam comparators in Italy. Drugs Under Experimental and Clinical Research 24, 105–113. 2. Bonfiglio, G., Carciotto, V., Russo, G., Stefani, S., Debbia, E., Schito, G. C. et al. (1998). Antibiotic resistance in Pseudomonas aeruginosa: an Italian survey. Journal of Antimicrobial Chemotherapy 41, 307–10. 3. National Committee for Clinical Laboratory Standards. (1990). Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically—Second Edition: Approved Standard M7A2. NCCLS, Villanova, PA. 4. Williams, R. J., Livermore, D. M., Lindridge, M. A., Said, A. A. & Williams, J. D. (1984). Mechanisms of -lactam resistance in British isolates of Pseudomonas aeruginosa. Journal of Medical Microbiology 17, 283–93. 5. Akova, M., Bonfiglio, G. & Livermore, D. M. (1991). Susceptibility to -lactam antibiotics of mutant strains of Xanthomonas maltophilia with high- and low-level constitutive expression of L-1 and L-2 -lactamases. Journal of Medical Microbiology 35, 208–13. 6. O’Callaghan, C. H., Morris, A., Kirby, S. M. & Shingler, A. H. (1972). Novel method for detection of -lactamases by using a chromogenic cephalosporin substrate. Antimicrobial Agents and Chemotherapy 1, 283–8. 7. Lowry, O. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951). Protein measurement with the folin phenol reagent. Journal of Biological Chemistry 193, 265–75. 8. Mathew, A., Harris, A. M., Marshall, M. J. & Ross, G. W. (1975). The use of analytical isoelectric focusing for detection and identification of -lactamases. Journal of General Microbiology 88, 169–78. 9. Ambler, R. P. (1980). The structure of

sophical Transactions of the Royal Society of London, Series B: Biological Sciences 289, 321–31.

-lactamases. Philo-

14. Bush, K. (1989). Characterisation of -lactamases. Antimicrobial Agents and Chemotherapy 33, 259–63. 15. Bush, K. (1989). Classification of -lactamases: group 1, 2a, 2b, and 2b’. Antimicrobial Agents and Chemotherapy 33, 264–70. 16. Bush, K. (1989). Classification of -lactamases: groups 2c, 2d, 2e, 3 and 4. Antimicrobial Agents and Chemotherapy 33, 271–6. 17. Jacoby, G. A. & Sutton, L. (1979). Activity of -lactam antibiotics against Pseudomonas aeruginosa carrying R-plasmids determining different -lactamases. Antimicrobial Agents and Chemotherapy 16, 243–5. 18. Buscher, K. H., Cullmann, W., Dick, W. & Opferkuch, W. (1987). Imipenem resistance in Pseudomonas aeruginosa resulting from diminished expression of an outer membrane protein. Antimicrobial Agents and Chemotherapy 31, 703–8. 19. Thabaut, A., Philippon, A. & Meyran, M. (1985). -Lactamases of Pseudomonas aeruginosa and susceptibility against -lactam antibiotics. Chemioterapia 4, 36–42. 20. Tirado, M., Roy, C., Segura, C., Reig, R., Hermida, M. & Foz, A. (1986). Incidence of strains producing plasmid determined -lactamases among carbenicillin resistant Pseudomonas aeruginosa. Journal of Antimicrobial Chemotherapy 18, 453–8. 21. Sumita, Y., Fukasawa, M. & Okuda, T. (1987). SM-7338 a new carbapenem antibiotic: affinities for PBPs and morphological changes. In Program and Abstracts of the Twenty-Seventh Interscience Conference on Antimicrobial Agents and Chemotherapy, New York, 1987. Abstract 756, p. 228. American Society for Microbiology, Washington, DC. 22. Sumita, Y. & Fukasawa, M. (1996). Meropenem resistance in Pseudomonas aeruginosa. Chemotherapy 42, 47–56. Received 4 March 1998; returned 7 May 1998; revised 29 May 1998; accepted 29 June 1998

702