Pseudomonas aeruginosa is a saprophytic organism widespread in nature, particularly in moist environments (water, soil, plants, and sewage). This bacterial species is endowed with only a weak pathogenic potential in immunocompetent persons (
2,
17,
21,
23) but can cause severe and even fatal infections in patients with impaired specific or nonspecific defense systems (
40). Thus,
P. aeruginosa is rarely involved in community-acquired infections, while it is responsible for a wide range of hospital-acquired infections, such as pneumonia, urinary tract infections (UTIs), and bacteremia. Moreover, a series of outbreaks due to this important and frequent nosocomial pathogen has been reported in hospital intensive care, burn wound, and cancer units (
6,
13,
42). In contrast, little has been reported on
P. aeruginosa-induced infections in long-term-care facilities, including nursing homes; and most of the studies that have reported such infections have described sporadic cases rather than outbreaks (
16,
22,
44).
Linked to its nosocomial status,
P. aeruginosa is intrinsically resistant to most antimicrobials. In addition, resistance to the main active agents, i.e., β-lactams and aminoglycosides, by gene acquisition is common. However, in this bacterial species, extended-spectrum β-lactamases (ESBLs) are rare and are generally of the OXA type, which are poorly inhibited by clavulanic acid, and belong to class D of the Ambler classification scheme (
1). Clavulanic acid-susceptible ESBLs of the TEM and SHV families of class A, which are widespread among enterobacteria (
35), are highly infrequent in
P. aeruginosa. Nevertheless, some of these enzymes, i.e., TEM-4, TEM-21, TEM-24, TEM-42, SHV-2a, SHV-5, and SHV-12, as well as unaffiliated class A β-lactamases such as PER-1, VEB-1, VEB-2, GES-1, GES-2, and IBC-2, have occasionally been found (
48). These observations suggest that ESBL-producing enterobacteria, which are often responsible for intra- and interhospital outbreaks due to the spread of strains, plasmids, or genes, might be the sources of these ESBLs for
P. aeruginosa (
33), which could become a reservoir for these types of enzymes. Aminoglycoside resistance is mainly mediated by antibiotic-modifying enzyme production, principally, the type II 6′
-N-aminoglycoside acetyltransferase [AAC(6′)-II] (kanamycin, gentamicin, tobramycin, and netilmicin resistance phenotype) in
P. aeruginosa, while the type II 3′
-N-aminoglycoside acetyltransferase [AAC(3)-II] enzyme (gentamicin, tobramycin, and netilmicin resistance phenotype), the type I 2"
-O-nucleotidyltransferase [ANT(2")-I] enzyme (kanamycin, gentamicin, and tobramycin resistance phenotype), and the type I AAC(6′) enzyme (kanamycin, tobramycin, netilmicin, and amikacin resistance phenotype) are prevalent among members of the family
Enterobacteriaceae (
32).
(This work was presented in part at the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, Ontario, Canada, 18 September 2000.)
DISCUSSION
Between November 1996 and December 2002, infections associated with ESBL-producing isolates of
P. aeruginosa were observed in a nursing home. Most of them were collected from urine specimens. Although
E. coli is the most common organism responsible for UTIs, other enterobacteria, such as
P. mirabilis,
P. stuartii, or
K. pneumoniae, and
P. aeruginosa, occur frequently in older patients, especially those in long-term-care facilities, as in this study (
16,
34,
43,
44). UTIs are the most common bacterial infections in the older population (
34). Moreover, among individuals who use urinary devices, as was the case among all patients in this study, colonization of these devices and subsequent infection by transmission between patients on the hands of caregivers or through contaminated equipment tend to occur (
7,
34). The urinary tract is thus the most prevalent site of nosocomial infections in long-term-care facilities (
16,
43). Although the term “nosocomial infections” has traditionally meant infections acquired in a general hospital, it should be broadened to include infections acquired in nursing homes (
43). UTIs are also one of the most frequent indications for the prescription of systemic antimicrobials in the older population. This contributes to the excessive use of antibiotics and promotes the emergence of antimicrobial resistance (
7). Consistently, older residents of long-term-care facilities have a higher prevalence of resistant organisms than older patients who are not institutionalized (
44,
51).
The
P. aeruginosa isolates involved exhibited an unusual red pigment whose development was delayed and a similar antibiotype. Moreover, the presence of a TEM-21 enzyme was demonstrated by IEF, PCR experiments, and sequencing of representative isolate Pa141 (
14). These data were suggestive of an outbreak of resistant strains due to a clone that persisted over at least a 6-year period in the environment or in the patients of this nursing home. Surprisingly, serotyping revealed that these isolates belonged to four chronologically successive serotypes, O:2, O:16, O:1, and PME. The
blaTEM-21 gene was found to be chromosomally located, as demonstrated by the absence of transfer by conjugation, transformation of a putative plasmid extract, and Southern blot experiments with unrestricted DNA (data not shown). Thus, the plasmid profile could not be used as an epidemiological marker. Consequently, whole-DNA typing was undertaken by RAPD and ERIC-PCR experiments. RAPD analysis is considered a robust, simple, and highly reproducible method and has successfully been used to study the epidemiology of
P. aeruginosa in patients with cystic fibrosis (
8) and during a nosocomial outbreak (
24). Primers ERIC2, 208, and 272 gave the most informative results and indicated a high similarity among the isolates, despite their different serotypes. Thus, the “gold standard” typing method, PFGE, was performed and confirmed that all isolates were closely related, suggesting a progressive genetic drift from the parental isolate. Indeed, members of the species
P. aeruginosa are known to have variable bacterial characteristics and to undergo changes in serotype and O-antigen structure during bacteriophage infection (
27) or antipseudomonal drug use (
25).
Moreover, because the genetic environment of antibiotic resistance genes may represent an epidemiological marker for closely related strains (
4), the sequences surrounding the
blaTEM-21 gene in
P. aeruginosa isolates were analyzed by PCR amplification and Southern blot experiments. All pseudomonal isolates appeared to possess an identical genetic environment for the
blaTEM-21 gene; i.e., the
blaTEM-21 gene was part of a Tn
801 transposon truncated by an IS
6100 element inserted within the resolvase gene, and the
aac(3)-II gene was adjacent to this structure (
14). In order to elucidate the origin of the
blaTEM-21 gene in the
P. aeruginosa isolates, the ESBL-producing enterobacteria carried by 10 of the 18 patients were examined. These isolates, which belonged to five different species, were found by IEF and PCR experiments to produce the TEM-21 β-lactamase. This enzyme was encoded by the same large conjugative plasmid, according to analysis of the cotransferred drug resistance and restriction profiles. This plasmid carried the
blaTEM-21 and the
aac(3)-II genes and conferred additional resistance to amikacin, conveyed by an AAC(6′)-I enzyme. Indeed, an outbreak of nosocomial infections due to the spread of TEM-21-producing enterobacteria and a TEM-21-encoding plasmid among different species of the family
Enterobacteriaceae has previously been demonstrated for other patients in this nursing home during a 6-month survey in 1999 (
3). Outbreaks of ESBL-producing enterobacteria due to the spread of resistant plasmids have been described in hospitals (
9,
10,
20,
26,
36) as well as in long-term-care facilities (
49), but the persistence of the ESBL-producing enterobacteria in the nursing home described here during at least a 6-year period reflected an endemic situation that has, until now, been reported only in hospitals (
19,
29). Among the ESBLs elaborated by enterobacteria, TEM-21 is less frequent than TEM-24 and TEM-3 in France (
11,
12,
19), but it has been already detected in the Bordeaux area (
3,
46) and was unknown in
P. aeruginosa until our description (
14). Thus, the hypothesis that the
blaTEM-21 gene found in
P. aeruginosa isolates originated from enterobacteria by gene transfer, as previously evidenced for
blaTEM-24 (
30) and
blaSHV-2a (
33) and suspected for
blaSHV-12 and
blaSHV-15 (
9,
38), appeared to be likely. However, cloning and sequencing experiments for a representative isolate, Ec223, followed by PCR amplifications of all transconjugants of the 16 remaining strains, indicated an organization different from that in the
P. aeruginosa isolates. Indeed, the
blaTEM-21 gene was part of Tn
801 and was adjacent to the
aac(3)-II gene, but the transposon was disrupted by IS
4321, inserted between the resolvase and the transposase genes.
Thus, the
blaTEM-21 gene was carried on a conjugative plasmid in the enterobacterial isolates, while it was chromosomally located in the
P. aeruginosa isolates. Differences in origins of replication may lead to the elimination of plasmids from
Enterobacteriaceae in
P. aeruginosa. Therefore, the persistence of the resistance genes requires their chromosomal integration. Since the
blaTEM-21 gene was located within a truncated Tn
801 and the
aac(3)-II gene was situated outside of the transposon in both types of organisms, both genes could not have been transferred by a Tn
801-mediated transposition. Thus, chromosomal integration of the plasmid itself might have happened after intergeneric transfer, as was suggested for
P. aeruginosa isolates containing the
blaSHV-12 gene (
9) and the
blaSHV-5 gene (
38). However, if such chromosomal integration occurred in TEM-21-producing
P. aeruginosa isolates, it was incomplete since the
aac(6′)-I gene present on the plasmid from
Enterobacteriaceae was lacking. Alternatively, a large mobile element encompassing the resistance genes might have ensured this genetic transfer. In this regard, one can notice that IS
4321 is a member of the IS
1111 family and is known to target the terminal inverted repeats of the Tn
21 family of transposons (
37). Moreover, the genetic environment of the
blaTEM-21 gene was different and was characterized by the presence of IS
6100 in
P. aeruginosa isolates, replaced by an IS
4321 at a slightly different position in the enterobacteria. The insertion of IS
4321 in Tn
801 of most
P. aeruginosa isolates indicated that a very large fragment that includes IS
6100 might have been secondarily integrated in the chromosome within Tn
801. Only isolates Pa133 and Pa190, from patient B, lacked this fragment; and both isolates were indistinguishable by PFGE, but their PFGE patterns were slightly different from the outbreak isolate PFGE pattern, suggesting a specific genetic rearrangement.
In conclusion, we report here on an endemic situation in a nursing home due to the dissemination of ESBL-producing strains and an ESBL-encoding plasmid, which shows that such a situation is not restricted to hospitals and clinics. A P. aeruginosa strain once probably acquired the ESBL-encoding plasmid that was epidemic among enterobacteria in this institution, and subsequent chromosomal integration followed by genetic reorganization allowed the persistence of the ESBL-encoding gene. Then, this strain spread among patients of the nursing home, undergoing genetic drift under antibiotic treatment or other pressures which became associated with modifications of the serotype. After a change in the directory staff of the nursing home, the number of prescriptions of microbiological analyses decreased considerably, and no ESBL-producing P. aeruginosa isolates have been isolated from the residents since 2003; but ESBL-producing enterobacteria are still being recovered.