432

Article

Vol. 11. No. 5

Eur. J. Clin. Mierobiol. Infect. Dis., May 1992, p. 432-437 0934-9723/92/05 0432-06 $ 3.00/0

Molecular Epidemiological Analysis of Pseudomonas aeruginosa Strains Causing Failure of Antibiotic Therapy in Cystic Fibrosis Patients

E. Bingen 1., E. D e n a m u r 2, B. Picard 4, E Goullet 4, N. L a m b e r t - Z e c h o v s k y 1, E F o u c a u d 3, J. N a v a r r o 3, J. Elion 2

A combination of esterase electrophoretic typing and analysis of the restriction fragment length polymorphism of ribosomal D N A regions (ribotyping) was used to compare 27 Pseudoraonas aeruginosa strains isolated before and after two-week courses of anti-pseudomonal treatment in seven cystic fibrosis patients. A total of 12 courses of therapy were studied in which ciprofloxacin, ceftazidime, azlocillin or imipenem were used alone or in combination with tobramycin. Isolates at a count of > 106 cfu/ml of sputum were collected when there was evidence of therapeutic failure on the basis of persistence of isolates whether or not they were resistant to the antibiotic used for therapy. Emergence of resistance was observed in ten cases and failure to eradicate sensitive strains in five cases. Among the 27 isolates, eight zymotypes and five ribotypes were identified. With this typing approach, resistant post-therapy isolates were found to be identical to pre-therapy isolates in all cases but one. However, in one case an additional resistant strain was isolated after therapy besides that initially present. In all five cases in which susceptibility was still observed after treatment, pretherapy and post-therapy isolates were indistinguishable. Using this molecular typing approach, all the strains were typable. Thus combination of esterase typing and ribotyping should improve the analysis of therapeutic failure in cystic fibrosis patients.

Chronic lung infection caused by Pseudomonas aeruginosa is the major cause of morbidity and mortality associated with cystic fibrosis (CF) (1, 2). Despite in vitro efficacy antibiotics are in fact rarely capable of eradicating the pathogens on therapy (3). Analysis of such therapeutic failure requires distinction between persistence of the same strain with the same susceptibility pattern, emergence of resistance in the same strain or reinfection with a new strain. Bacterial typing methods should thus establish unambiguously the identity or not of paired pre-therapy and posttherapy isolates. Unfortunately, most of the currently available phenotyping systems (O-serotyping, biotyping, antimicrobial susceptibility patterns) are not sufficiently sensitive or stable under the usual laboratory conditions for surveil1Laboratoire de Microbiologie;2Laboratoire de Biochimie G6nrtique, and 3Service de Gastroentrrologie, Hrpital Robert Debr6, 48 Bd. Sdrurier, 75935 Paris Cedex 19, France. 4 Laboratoire de Microbiotogie, Facult6 Xavier Bichat, 16 rue Henri Huchard, 75018 Paris, France.

lance of Pseudomonas aeruginosa strains in CF (4-9). Combination of phage typing and pyocin typing might meet these criteria, but only 39 % of the Pseudomonas aeruginosa strains in CF are susceptible to phage lysis (9). We recently described a molecular approach to the typing of Pseudomonas aeruginosa isolates in CF consisting of a combination of esterase electrophoretic typing and ribotyping (10). Using this approach we observed no case of failure to type a strain. In the present study we used this molecular typing method to compare pre-therapy and post-therapy isolates of Pseudomonas aeruginosa obtained from CF patients treated with ciprofloxacin, ceftazidime, azlocillin or imipenem.

Materials and Methods

Patients. Seven CF patients (designated A to H in Table 1) with chronic bronchopulmonary Pseudomonas aeruginosa infections treated with antibiotics in the gastro-

Vol. 11, 1992

433

Table 1: Resultsoftypingpairedpre-therapyandpost-therapyisolatesofPseudomonasaeruginosa. Course of treatment no.

Patient and isolate designation

Relation to time of therapy

Antimierobial agent and susceptibility

O-serotype

Zymotype

Ribotype

1

A1 A2

before after

ciprofloxacin S ciprofloxacin R

NT NT

I I

a a

2

B1 B2 B3 B4

before before after after

ciprofloxacin ciprofloxacin ciprofloxacin ciprofloxacin

S S S R

AA PI NT NT

II II II II

b b b b

3

C1 C2 C3

before after after

ciprofloxacin S ciprofloxacin R ciprofloxacin R

P1 NT AA

III IV III

c c c

4

C3 C4

before after

imipenem imipenem

S R

AA NT

III III

c c

5

D1 D2

before after

ciprofloxacin S ciprofloxacin S

NT NT

V V

c c

6

F1 F2 F3

before after after

ciprofloxacin S ciprofloxacin R ciprofloxacin S

AA NT NT

VI VI VI

d d d

7

F3 F4

before after

ciprofloxacin S eiprofloxacin R

NT P4

VI VI

d d

8

G1 G2

before after

ceftazidime ceftazidime

S R

PA NT

VII VII

c c

9

G3 G4

before after

ceftazidime ceftazidime

S S

NT NT

VII VII

c c

10

G5 G6

before after

ciprofloxacin S ciprofloxacin R

NT AA

VII VII

c c

11

G7 G8 G9

before after after

imipenem imipenem imipenem

S S R

NT NT NT

VII VII VII

c c c

12

H1 H2

before after

azlocillin azlocillin

S R

NT NT

VIII Vll

e e

S: Susceptible, R: Resistant, NT: Non Typable, AA: Autoagglutinable, PA: Polyagglutinable.

enterology unit of the Hfpital Robert Debr6, Paris, France, were entered in the study. Siblings were excluded. A diagnosis of chronic Pseudomonas aeruginosa infection Was based upon the monthly finding of Pseudomonas aeruginosa at a count of > 106 cfu/ml in bronchial secretions for a period of six consecutive months, and/or two or more precipitin antibodies against Pseudomonas aerugh~osa (normal: 0-1 precipitin) (4). Patients received Clprofloxacin (30-35 mg/kg/day), ceftazidime (200-250 ~g/kg/day), azlocillin (350 mg/kg/day) or imipenem (75 _nag/kg/day) alone or in combination with tobramycin (1020 mg/kg/day) in 14-day courses. A total of 12 courses of anti-pseudomonal chemotherapy were considered (Table 1). Four patients (A, B, D and H) received one COurse of treatment. Patient F was given two courses of ciprofloxacin, patient C was treated with ciprofloxacin and imipenem successively and patient G received se-

quentially two courses of ceftazidime, one of ciprofloxaein and one of imipenem. Some of these courses of antibiotic were administered concurrently. This was the case for courses 9 and 12 (patients G and H), courses 3 and 6 (patients C and F) and courses 1 and 2 (patients A and B).

Bacterial Isolates. Bacteriological failure of therapy was evidenced by the persistence of strains with unchanged antibiotic susceptibility patterns or by the appearance of isolates resistant to ciprofloxacin, ceftazidime, azlocillin or imipenem. Altogether, 27 Pseudomonas aeruginosa isolates (designated AI to H2 in Table 1) were collected from the seven patients. Single pairs of pro-therapy and post-therapy isolates were analyzed (courses I, 4, 5, 7-10 and 12), except when isolates with different colony morphology were observed either before or after treatment.

434

For instance, one pre-therapy and two post-therapy isolates were analyzed for courses 3, 6 and 11, and two pre-therapy and two post-therapy isolates were analyzed for course 2 (Table 1). All the isolates, both sensitive and resistant, were obtained from quantitative sputum cultures before antibiotic administration and again on day 14 of therapy or at the first follow-up visit. Isolates were identified as Pseudomonas aeruginosa on the basis of typical morphology, positive oxidase reaction, ability to produce pigments, growth at 42"C and biochemical test results in the API-20 NE system (API, France). Serotyping was performed with O-antisera (Diagnostic Pasteur, France). Antimicrobial susceptibility testing of all isolates was done by the standardized disk-diffusion method (11). Minimal inhibitory concentrations (MICs) were determined by inoculating 5 x 105 organisms into cation-supplemented Mucller-Hinton broth (12).

Esterase Electrophoresis. Bacteria were grown at 30 *C in L-broth under constant agitation for 18 h and collected by centrifugation. The pellets were washed with 60 mM Tris-glycine buffer, pH 8.7, resuspended in the same buffer and disrupted by sonication for 18 min at 4 °C. Crude extract supernatants, containing 40-60 mg of protein per ml, were stored at -20 °C until used, Horizontal slab gel etectrophoresis was performed according to the method of Uriel (13) in a composite polyacrylamide agarose gel (7 and 1.4 %, respectively) in a discontinuous Tris-glycine buffer, pH 8.7, at a constant voltage (7 V/cm) until the bromophenol blue marker had run 13 cm. The relative mobility (Mr: value) was defined as the distance moved by the esterase band as a percentage of the distance moved by the dye front (14, 15). The esterases were stained on the gel using the following specific substrates: et~naphthyl acetate, 13-naphthyl acetate, ~-naphthyl butyratc, 13-naphthyl butyrate, and indoxyl acetate (Sigma, USA). The sensitivity of the esterases to diisopropyl-fluorophosphate (DFP) (10 .3 M) was also tested (15). Reproducibility of the results was confirmed in several ways. All Mv values were tested by comparing side by side each new strain with a reference strain in all electrophoretic runs. Independent extracts from different cultures of a strain of a patient were compared under the same conditions to assess stability of the results.

Analysis of Restriction Fragment Length Polymorphism (RFLP) of Ribosomal DNA Regions (Ribotyping). rDNA-RFLP profiles were obtained after digestion of bacterial DNA with HhzdlII, EcoRl and BclI restriction enzymes. Restriction fragments were size-separated by agarose gel electrophoresis, Southern-transferred to membranes, and probed with 32p-labelled reversetranscribed 16 + 23 S Escherichia coli RNA. Bacterial DNA preparation, digestion by the restriction enzymes, submarine agarose gel clectrophoresis, transfer to nylon membrane, labelling of the probe, hybridization and autoradiographic procedures were as described previously (10, 16). Each specific combination of the RFLP patterns obtained with each of the three endonucleases defined a ribotype. Reproducibility of the results was established in multiple runs, comparing size-separated DNA restriction fragments side by side on the same gel. Absolute stability of the RFLP profiles was observed for the type strain of the species ATCC 10145 after 10, 20 and 30 subcultures.

Eur. J. Clin. Microbiol. Infect. Dis.

Results

Bacteriological Results. T h e c o u n t s o f Pseudomonas aeruginosa isolates present in sputum varied f r o m 106 to 108 cfu/ml and did not d e c r e a s e significantly during the t w o - w e e k t r e a t m e n t courses. Failure to eradicate Pseudomonas aeruginosa was o b s e r v e d in seven cases after ciprofloxacin t h e r a p y (courses 1-3, 5 - 7 and 10, Table 1), in two cases after ceftazidime plus t o b r a m y c i n t h e r a p y (courses 8 and 9), in two cases after i m i p e n e m plus t o b r a m y c i n t h e r a p y (courses 4 and 11) and in o n e case after azlocillin plus t o b r a m y c i n t h e r a p y (course 12). In ten cases at least o n e o f the p o s t - t h e r a p y isolates was f o u n d to b e resistant to the antimicrobial a g e n t used f o r t h e r a p y (courses 1--4, 6 8 a n d 10-12, Tables 1 and 2). N o n e of the isolates exhibited t o b r a m y c i n resistance. A t least o n e o f the p o s t - t h e r a p y isolates was susceptible to the antimicrobial agent used for t h e r a p y in five cases (courses 2, 5, 6, 9 and 11, Table 1), It was the only p o s t - t h e r a p y isolate o b s e r v e d in c o u r s e s 5 a n d 9.

O-Serotyping. All isolates b u t two w e r e nontypable, a u t o a g g l u t i n a b l e o r polyagglutinable. T h u s no c o m p a r i s o n b e t w e e n p r e - t h e r a p y and p o s t - t h e r a p y isolates was possible o n the basis o f O - s e r o t y p i n g (Table 1). Molecular Typing. A m o n g t h e 27 Pseudomonas aeruginosa strains o b t a i n e d f r o m the seven C F patients, eight z y m o t y p e s (I to V I I I ) and five ribotypes (a to e) w e r e identified b y esterase e l e c t r o p h o r e s i s and analysis of the R F L P of r D N A regions, respectively (Figures 1 and 2, Table 1). U s i n g these m e t h o d s all the strains w e r e typable. W h e n d e v e l o p m e n t of resistance was o b s e r v e d , p r e - t h e r a p y and p o s t - t h e r a p y isolates were identical on m o l e c u l a r typing in nine o u t o f ten cases (courses 1-4, 6-8, 10 and 11). In o n e o f these cases (course 3, patient C), however, a distinct additionTable 2: Susceptibility patterns of pre-therapy and posttherapy isolates with emergence of resistance. MICs (mg/I) Drug Pre-treatment Ciprofloxacin Ceftazidime Imipenem Azlocillin

< 0.125 - 0.25 1 - 2 0.5 - 1 1

Post-treatment > 32 > 256 > 64 > 256

Vol. 11, 1992

435

type. Pre-therapy and post-therapy isolates with persistence of susceptibility were identical on molecular typing in all cases (courses 2, 5, 6, 9 and 11) (Table 1). E1 t=2 ,,. E3~ i

~E2 E3"

1

2

Discussion

~E 3

3

4

Figure 1: Esterase electrophoretic patterns of four Pseudomonas aeruginosa isolates revealed by 13-naphthyl acetate staining. Strains F3 (1) and F4 (2) exhibited the ~nae zymotype whereas strains H1 (3) and H2 (4) were utstinguished by two different zymotypes. El, esterase El; E2, esterase E2; E3, esterase E3. al resistant strain (C2) was isolated after therapy besides that initially present. This strain had the Same ribotype as the other C strains, but differed in its zymotype. In one case only (course 12), the resistant post-therapy isolate (H2) was different from the pre-therapy isolate (H1) (Table 1). H2 and H I differed both in ribotype and in zymo-

1

2

3

4

1

2

To be of epidemiological value a typing system requires markers that are discriminative, stable and reproducible. Furthermore, it should be possible to type all strains to be studied with the system. Esterase typing and ribotyping have been shown in a number of bacterial species to fulfill these conditions (17-19). In Pseudomonas aeruginosa we found r D N A R F L P profiles to be stable after 30 passages in vitro. Stability of the R F L P of the exotoxin A and of the pilin gene regions in this species has been established (20, 21). Recently, Hjelrn et al. (22) observed the generation of novel RFLPs in Pseudomonas aeruginosa CF isolates serially passaged from a single colony. However, the probe used was in a region that flanks a hot spot for Tn7 insertion and thus R F L P instability is not surprising. In the present study we used both esterase typing and ribotyping to analyze bacteriologic failure in 12 courses of antibiotic therapy in seven CF patients. All the strains were typable using these

3

4

1

2

3

4

5.6-

9.0- ~

~

-4.3-

9.0.

, , , m a l flip J

4.3-[

h

,

,?-,

.

1.2-

a

HindIII

digestion

b

EcoRI

digestion

c

Bc_~iI d i g e s t i o n

Figure 2: rDNA RFLP patterns of four Pseudomonas aeruginosa isolates, a: HindIII digestion; b: EcoRI digestion; c: Bc/I digestion, Strains C3(1) and C4 (2) exhibited the same ribotype whereas strains HI (3) and H2 (4) exhibited distinct ribotypes.

436

molecular methods. T h e antibiotics administered consisted of ciprofloxacin, ceftazidime, azlocillin and imipenem alone or in combination with tobramycin. A p p e a r a n c e of resistance to the antibiotic was observed in ten cases, but in three of them together with persistence of a sensitive strain. D e v e l o p m e n t of resistance during treatment of CF patients with ciprofloxacin, aztreonam and ceftazidime has already b e e n reported (23, 24). The relationship between pretherapy and post-therapy isolates has been investigated by establishing antibiotic susceptibility patterns (23) and by phage and pyocin typing (24). Using our combination of methods, we found that all but two of the 11 resistant posttherapy isolates were indistinguishable from the corresponding pre-therapy isolate. Ogle et al. (25), using R F L P to study the exotoxine A gene, observed the development of resistance to various antibiotics in Pseudomonas aeruginosa isolates from n o n - C F patients. In two cases only in our study a resistant strain was observed that was not related to the initial sensitive strain: in course 3 an additional resistant strain appeared and in course 12 reinfection with a new resistant strain occurred parallel to eradication of the pretherapy sensitive strain. Resistance to ciprofloxacin, ceftazidime and imipenem corresponds to chromosomic mutations occurring with a frequency in the order of 10-7 (26-28). Such differences in the bacterial g e n o m e are not detected by our molecular typing method. Resistance as seen in our study might represent the selection of preexisting resistant strains that were not identified at the outset of therapy since our sensitivity threshold was 105 cfu/ml. Failure to eradicate sensitive strains during therapy was observed in five cases (courses 2, 5, 6, 9 and 11). This further illustrates the discrepancies reported between the in vitro susceptibility and in vivo efficacy of antibiotics particularly in CF patients (29). Using electrophoretic esterase typing and ribotyping, we were able to distinguish strains that were otherwise indistinguishable by a combination of conventional criteria including morphology, serotype and antibiotic susceptibility pattern. Our findings provided compelling evidence that all the isolates from patient F, on the one hand, (courses 6 and 7), and from patient G, on the other, (courses 8, 9, 10 and 11) were indistinguishable with our molecular typing system. This suggests that each patient was colonized by a single strain o f Pseudomonas aeruginosa, even after two or four courses of antibiotic therapy.

Eur. J. Clin. Microbiol. Infect. Dis,

These results are in agreement with those of Paso loke et al. (30) and Wolz et al. (31) who analyzed the R F L P of the pilin and the exotoxin A genes respectively. Availability of precise methods for strain identification should improve bacteriological monitoring and m a n a g e m e n t of antibiotic therapy in CF patients.

References 1. Wood RE, Boat TF, Doershuk CF: State of the art:

cystic fibrosis. American Review of Respiratory Disease 1976, 113: 833-878. 2. Huiby N, Friis B, Jensen K, Koch C, Moeiler NE, Stovring S, Szaff M: Antimicrobial chemotherapy in

cystic fibrosis patients. Acta Paediatrica Scandinavica 1982, 71, Supplement 301: 75-100. 3. Haihy N, Fiensborg EW, Beck B, Friis B, Jacobsen

SV, Jacobsen L: Pseudomonas aeruginosa infection in cystic fibrosis. Diagnostic and prognostic significance of Pseudomonas aeruginosa precipitins determined by means of crossed immunoelectrophoresis. Scandinavian Journal of Respiratory Disease 1977, 58: 65-79. 4. Brokopp CD, Farmer J J: Typing methods for Pseudomonas aerugh~osa. In: Doggett RG (ed): Pseudom o n a s aeruginosa: clinical manifestations of infections and current therapy. Academic Press, New York, 1979, p. 89-133. 5. Liu PV: Changes in somatic antigens of Pseudomonas aeruginosa induced by bacteriophages. Journal of Infectious Diseases 1969, 119: 237-246. 6. Lanyi B, Lantos J: Antigenic changes in Pseudomonas aeruginosa in vivo and after lysogenization in vitro. Acta Microbiologica Hungarica 1976, 23: 337-351. 7. Hancock REW, Mutharia LM, Chan L, Darveau RP, Speert DP, Pier GB: Pseudomonas aeruginosa isolates

from patients with cystic fibrosis: a class of serumsensitive, non-typabl¢ strains deficient in lipopolysaecharid¢ O side chains. Infection and Immunity 1983, 42: 170-177. 8. Ojeniyi B, Back L, Holby N: Polyagglutinability due to loss of O-antigenic determinants in Pseudomonas aeruginosa strains isolated from cystic fibrosis patients. Acta Pathologica Microbiologica Immunologica Scandinavica B 1985, 93: 7--13. 9. Pill TL: Epidemiological typing of Pseudomonas aeruginosa. European Journal of Clinical Microbiology & Infectious Diseases 1988, 7: 238-247. 10. Denamur E, Picard B, Goullel Ph, Bingen E, Lambert

N, Elion J: Complexity of Pseudootonas aeruginosa infection in cystic fibrosis: combined results from esterase electrophoresis and rDNA restriction fragment length polymorphism analysis. Epidemiology and Infection 1991, 106: 531-539. 11. National Committee for Clinical Laboratory Standards: Performance standards for antimicrobial disk susceptibility tests. Approved standard, M2-A3. NCCLS, Villanova, PA, 1984. 12. National Committee for Clinical Laboratory Standards: Methods for dilution antimicrobial susceptibility

tests for bacteria that grow aerobically. Approved standard, M7-A. NCCLS, Villanova, PA, 1985.

V.~01.11,1992

13. Uriel, J: M6thode d'61ectrophor~se dans des gels d~acrylamide-agarose. Bulletin de la Soci6t6 de Chimie Biologique 1966, 48: 969-982. 14. Goullet p, Pieard B: A two-dimensional electrophoretic profile for bacterial esterases. Electrophoresis 1985, 6: 132-135. 15. Goullet P, Picard B: Pseudomonas aeruginosa isolate typing by esterase electrophoresis. FEMS Microbiology Letters 1991, 78: 195-200. 16. Picard-Pasquier N, Ouagued M, Picard B, G0ullet P, Krishnamoorihy R: A simple sensitive method of analyzing bacterial ribosomal DNA polymorphism. Electrophoresis 1989, 10: 186--189. 17. Picard B, Goullet P: Epidemiological complexity ot hospital Aeromonas infections revealed by electrophoretic typing of esterases. Epidemiology and Infection 1987, 98: 5-14. 18. Lipuma J J, Fisher MC, Dasen SE, Mortensen JE, Slull TL: Ribotype stability of serial pulmonary isolates of Pseudomona.s cepacia.. Journal of Infectious Diseases 1991, t64: 133-136. 19. Stall TL, Lipuma JJ, Edlind TD: A broad spectrum probe for molecular epidemiology of bacteria: ribosomal RNA. Journal of Infectious Diseases 1988, 157: 280-286. 20. Speert DID,Gampbell ME, Farmer SW, Volpel K, Joffe AM, Paranchy W: Use of a pilin gene probe to study molecular cpidemiology of Pseudomonas aeruginosa. Journal of Clinical Microbiology 1989, 27: 2589-2593. 21. Ogle JW, Janda JM, Woods DE, Vasil ML,: Characterization and use of a DNA probe as an epidemiological marker for Pseudomonas aeruginosa. Journal of Infectious Diseases 1987, 155: 119-126. 22. Hjelm LN, Branslrom AA, Warren RL: Detection of restriction fragment length polymorphisms in clinical isolates and serially passaged Pseudomonas aeruginosa strains. Journal of Clinical Microbiology 1990, 28: 2178-2182.

437

23. Bosso JA, Allen JE, Malsen JM: Changing susceptibility of Pseudomonas aeruginosa isolates from cystic fibrosis patients with the clinical use of newer antibiotics. Antimicrobial Agents and Chemotherapy 1989, 33: 526--528. 24. Giwercman B, Lambert PA, Rosdahi V, Shand GH, H~iby N: Rapid emergence of resistance in Pseudomonas aeruginosa in cystic fibrosis patients due to in-vivo selection of stable partially derepressed 13-1actamase producing strains. Journal of Antimicrobial Chemotherapy 1990, 26: 247-259. 25. Ogle JW, Relier LB, Vasil ML: Development of resistance in Pseudomonas aeruginosa to imipenem, norfloxacin and ciprofloxacin during therapy: proof provided by typing with a DNA probe. Journal of Infectious Diseases 1988, 157: 743-748. 26. Kaalz GW, Seo SM: Mechanism of ciprofloxaein resistance in Pseudomonas aeruginosa, Journal of Infectious Diseases 1988, 158: 537-541. 27. Neu H e : Bacterial resistance to tluoroquinolones. Reviews of Infectious Diseases 1988, 10, Supplement 1: 57-63. 28. Sanders CC, Sanders WE: Microbial resistance to newer generation 15-1actam antibiotics: clinical and laboratory implications. Journal of Infectious Diseases 1985, 151: 399--406. 29. Levy J: Antibiotic activity in sputum. Journal of Pediatrics 1986, 108: 841--845. 30. Pasloke BL, Joffre AM, Sun Q, Volpel K, Paranchych W, Eflekhar F, Speert DP: Serial isolates of Pseudomonas aeruginosa from a cystic fibrosis patient have identical pilin sequences. Infection and Immunity 1988, 56: 665-672. 31. Wolz C, Kiosz D, Ogle JW, Vasil ML, Schaad U, Bolzenhart K, Diiring G: Pseudomonas aeruginosa crosscolonization and persistence in patients with cystic fibrosis. Use of a DNA probe. Epidemiology and Infection 1989, 102: 205-214.