Pseudomonas aeruginosa is one of the most important nosocomial pathogens, especially in the intensive care unit (ICU) setting, where it is the leading cause of ventilator-associated pneumonia, with a high mortality rate (
15). Antimicrobial resistance in this microorganism is a problem of growing concern that frequently limits our therapeutic options (
5). Antibiotic resistance in
P. aeruginosa can be effectively acquired either by mutations in intrinsic chromosomal genes or by the horizontal transfer of resistance determinants (
14). Both mechanisms are frequently found in
P. aeruginosa from ICUs, where resistance rates are frequently higher than resistance rates in other hospital units. Furthermore, mutation-mediated antibiotic resistance development during antibiotic treatment is one of the most important factors contributing to therapy failure (
6,
22). Hypermutation is a common feature of
P. aeruginosa from chronically infected cystic fibrosis (CF) patients that has previously been found to be linked with antibiotic resistance development (
19). On the other hand, hypermutation appears to be very infrequent in
P. aeruginosa isolates from acutely infected patients (
19). The main objective of this work was to find out if hypermutation could also be an important factor for antibiotic resistance development outside the context of chronic infection, choosing for that purpose the
P. aeruginosa infections from ICU, where, despite the fact that it mostly involves acute processes, resistance development is a frequent outcome.
A total of 216
P. aeruginosa isolates were recovered from clinical samples of 103 patients admitted to the Hospital Son Dureta ICU from September 2002 to November 2003. Identification and initial susceptibility testing was performed with the WIDER system (
3). MICs of ceftazidime, cefepime, imipenem, meropenem, ciprofloxacin, and tobramycin were determined with the Etest (AB Biodisk, Solna, Sweden). The first
P. aeruginosa isolate obtained from each of the patients and type of clinical sample, as well as any subsequent isolate that developed resistance to any of the antibiotic tested, were further studied. Following this criteria, 160 isolates from the 103 patients were included. The epidemiological relatedness of the strains was studied by pulsed-field gel electrophoresis (PFGE). Bacterial DNA embedded in agarose plugs prepared as described previously (
11) was digested with SpeI. DNA separation was performed in a contour-clamped homogeneous electric field-DRIII apparatus (Bio-Rad, La Jolla, Calif.) with the following conditions: 6 V/cm
2 for 26 h with pulse times of 5 to 40 s. DNA macrorestriction patterns were interpreted according to the criteria established by Tenover et al. (
27).
Analysis of PFGE patterns showed that dissemination of P. aeruginosa strains in the ICU was an infrequent event, with most patients harboring a unique clone. A total of 82 clones were documented, of which 65 (79.2%) were found in single patients. The remaining 17 clones were found in 2 or more patients; 11, 4, 1, and 1 clones were found in 2, 3, 4, and 5 patients, respectively. Table shows the overall resistance percentages for the 216 P. aeruginosa isolates recovered. Additionally, primary resistance (number and percentage of patients from which a resistant strain was isolated in the first P. aeruginosa sample cultured) and secondary resistance (number and percentage of patients from which a resistant strain was isolated after a previous isolation of a susceptible strain in the first P. aeruginosa sample cultured) results are also shown in Table . Results from the primary resistance rates, together with the low interpatient dissemination of strains documented, demonstrates that acquisition of antibiotic-resistant strains, with the exception of strains resistant to imipenem (23.3% of primary resistance), is not a major problem in this ICU. On the contrary, sequential resistance development (secondary resistance) is a relatively frequent event. For instance, in 21 (20.6%) of the patients, secondary resistance was observed for at least one of the antimicrobials tested. As shown in Table , secondary resistance was higher for cefepime and ceftazidime (17.5 and 15.5%, respectively) and lower for tobramycin (1%). In 20 of these 21 patients, secondary resistance was consequence of mutant selection (same PFGE of susceptible and resistant isolate) and not of the replacement of the susceptible clone by a resistant one.
| TABLE 1.Overall, primary, and secondary resistance percentages of the P. aeruginosa strains |
This ICU setting of low interpatient strain transmission and relatively high mutational antibiotic resistance development seemed adequate for the accomplishment of the main objective of this work: to find out if hypermutation is an important factor for antibiotic resistance development in the context of acute infection but with high mutational antibiotic resistance. Rifampin (300 μg/ml) resistance mutation frequencies were determined as previously described (
19). Strains PAO1 and its
mutS-deficient hypermutable derivative PAOΔ
mutS (
20a) were used as controls. Distribution of the mutation frequencies of the studied strains is shown in Fig. . Only 2 (both from the same patient and belonging to the same clone) of the 160 isolates were found to be hypermutable (mutation frequencies were >20-fold greater than those for PAO1) (
19). Nevertheless, a relatively wide range of mutation frequencies was observed. Close to half (48%) of the strains (as well as the control strain PAO1) had mutation frequencies between 1 to 4 mutants/10
8 cells. Two isolates had mutation frequencies of 10 to 20 mutants/10
8 cells (10-fold higher than that for PAO1) and could be considered hypermutable at low levels. On other hand, a certain number of isolates had mutation frequencies significantly lower than that of PAO1: 4 isolates had mutation frequencies in the range of 0.1 to 0.2 mutants/10
8 cells (10-fold lower than that for PAO1), and 15 isolates had frequencies in the range of 0.2 to 04 (5-fold lower than that for PAO1). No significant differences in mutation frequencies were observed between antibiotic-resistant or -susceptible isolates for any of the antibiotics (data not shown). Although it is only anecdotal, the hypermutable strain was actually multiresistant (ceftazidime, cefepime, imipenem, meropenem, and ciprofloxacin resistant), which happened in less than 5% of the isolates, although the coselection of the hypermutable and antibiotic resistance phenotypes during the antibiotic treatment could not be documented (it was already hypermutable and multiresistant when first isolated).
Hypermutable (or mutator) strains are those that have an increased spontaneous mutation rate due to defects in genes involved in DNA repair or error avoidance systems (
9,
18). In CF
P. aeruginosa strains, as well as in other natural bacterial populations (
12,
16), the most frequently involved mechanism is the mismatch repair system, and
mutS is the most frequently affected gene (
20). Mutator strains have been described in natural populations of
Escherichia coli,
Salmonella spp.,
P. aeruginosa,
Helicobacter spp.,
Neisseria meningitidis, and
Staphylococcus aureus (
1,
12,
16,
19,
21,
23). In vitro and theoretical approaches suggest that the acquisition of a stable mutator phenotype may confer a selective advantage for bacteria, particularly in stressful and fluctuating environments (
10,
13,
25,
26). In vivo experiments have also shown a potential benefit of hypermutation for bacterial adaptation to new environments, although once adapted, this advantage disappears and the transmissibility of the mutator strains is considerably reduced (
7). In natural bacterial populations, the presence of hypermutable strains has been linked to resistance to host immunological defenses as mutation-dependent phase variation in
N. meningitidis (
23,
24) or resistance to antibiotics in
P. aeruginosa isolates from chronically infected CF patients, where the frequency of hypermutable variants is by far the highest ever found in natural populations (
19). Up to 37% of the CF patients were colonized by hypermutable strains in the mentioned study, whereas not a single hypermutable strain was found in 75 acutely infected patients (
19). Additionally, a high frequency of mutator variants of
S. aureus from CF patients has recently been found to be linked with resistance to macrolides (
21).
In the last 2 years the role of hypermutation in antibiotic resistance development has begun to be acknowledged as a potential problem of concern (
2,
4,
8). There are various possible explanations for the high prevalence of hypermutable strains in chronically infected CF patients but not in acutely infected patients. One of them is that it is a specific feature of the strains infecting those patients, but this seems not to be the case, because a high frequency of hypermutable strains is also found in other chronic infections, such as those of patients with bronchiectasis or chronic obstructive pulmonary disease (M. D. Maciá and A. Oliver, unpublished data). A second possibility is that the presence of hypermutable strains could mainly be a consequence of the coselection of the hypermutable variants, with the adaptive mutations required for long-term persistence in the lungs of patients. Finally, the presence of hypermutable strains could mainly be a consequence of the coselection of the hypermutable variants with antibiotic resistance mutations, given the strong association between hypermutation and antibiotic resistance found in CF patients. In the present work, using a large collection of
P. aeruginosa strains from ICU patients, we found that although mutational antibiotic resistance development is a frequent outcome, the prevalence of hypermutable strains is low (found in isolates from only 1 of the 103 patients). Additionally, no evidence was found of the coselection of the hypermutable and antibiotic resistance phenotypes. Future works will be necessary to finally understand whether the chronic infection process is the determining factor for the selection of hypermutable strains in natural
P. aeruginosa populations or whether, on the contrary, it is the acute infection process that limits the presence of hypermutable strains due, for instance, to a lower virulence of these variants, as has recently been found for
Listeria monocytogenes experimental models of infection (
17).
This article has been 
