Effect of Porins and Plasmid-Mediated AmpC β-Lactamases on the Efficacy of β-Lactams in Rat Pneumonia Caused by Klebsiella pneumoniae

Abstract

The in vivo activities of imipenem, meropenem, and cefepime were studied in a model of rat pneumonia caused by a plasmid-mediated AmpC β-lactamase ACT-1-producing Klebsiella pneumoniae strain (K. pneumoniae strain 12) and a derivative porin-deficient mutant (K. pneumoniae strain 12dp). No differences between these activities were seen with K. pneumoniae 12. Only meropenem showed an activity slightly better than that of imipenem with K. pneumoniae 12dp.

The number of infections caused by gram-negative bacteria producing plasmid-mediated AmpC-type β-lactamases (PACBL) has increased in the last 10 years (3, 7, 15, 18), including three reports of outbreaks (9, 15, 20). The only active β-lactams in those cases are carbapenems (imipenem, meropenem, and ertapenem) and zwitterionic cephalosporins (cefepime and cefpirome). However, cooperation with other mechanisms of resistance, such as decreased permeability, results in the subsequent elevation of the MICs of antibiotics with good activity, in some cases up to resistance level (2, 3, 10).

The aim of this study was to evaluate the in vivo efficacies of human regimens of imipenem, meropenem, and cefepime against a PACBL ACT-1-producing Klebsiella pneumoniae strain (K. pneumoniae 12) and its derived in vitro-obtained mutant deficient in porins (K. pneumoniae 12dp) by using a model of experimental pneumonia in immunocompetent rats.

K. pneumoniae strain 52145 is a clinical isolate that has been described previously (14). K. pneumoniae strain 12 is a derivative of K. pneumoniae 52145 containing the PACBL ACT-1 by conjugation with K. pneumoniae C2 (pMG246) (1). K. pneumoniae strain dp is a porin-deficient mutant derived from K. pneumoniae strain 52145, and it was obtained by overnight culture containing cefoxitin at 64 mg/ml and verified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blot analysis as previously described (8). K. pneumoniae strain 12dp is a porin-deficient derivative of K. pneumoniae 52145 expressing ACT-1. Expression of the AmpC β-lactamase in both K. pneumoniae 12 and K. pneumoniae 12dp was verified by nitrocefin hydrolysis.

Imipenem-cilastatin and cilastatin (Merck Sharp & Dohme-Chibret, Madrid, Spain), meropenem (Astra Zeneca Pharma, Madrid, Spain), and cefepime (Bristol-Myers Squibb, Madrid, Spain) were used according to the manufacturer's instructions. MICs of imipenem, meropenem, cefepime, and ertapenem were determined by the Etest method (AB Biodisk, Solna, Sweden) and are summarized in Table 1.

TABLE 1.

In vitro susceptibilities of K. pneumoniae 52145 and its derivative strains

K. pneumoniae straina	Presence (+) or absence (−) of:		MIC (mg/liter) ofb:
	ACT-1	Porin deficiency	Imipenem (<4/8/≥16)	Meropenem (<4/8/≥16)	Ertapenem (<2/4/≥8)	Cefepime (≤8/16/≥32)
52145	−	−	0.25	0.03	<0.03	0.03
12	+	−	0.25	0.03	<0.03	0.25
dp	−	+	1	0.03	<0.03	0.3
12dp	+	+	16	8	>16	2

^aK. pneumoniae strain 12 is K. pneumoniae 52145 with plasmid-mediated AmpC; K. pneumoniae strain dp is porin (OmpK35 and OmpK36)-deficient K. pneumoniae 52145; K. pneumoniae strain 12dp is K. pneumoniae 52145 with plasmid-mediated AmpC⁺ porin (OmpK35 and OmpK36) deficiency.

^bValues in parentheses are CLSI MIC breakpoints for Enterobacteriaceae as interpretative standard categories of susceptible/intermediate/resistant organisms.

Pharmacokinetic studies were done by a single intraperitoneal dose in male Wistar rats (Harlan Ibérica S. L., Barcelona, Spain) of 40 mg/kg of body weight for imipenem, 40 mg/kg for meropenem, or 60 mg/kg for cefepime. Antibiotic concentrations in blood were determined by the disk plate bioassay method (5). Since rats produce dehydropeptidase, which is able to hydrolyze meropenem (6), cilastatin (1:1) was given together with meropenem. Estimations of the maximum antibiotic plasma concentration (in mg/liter), half-life (in h), and time during which the plasma concentration remained above the MIC (in h) were calculated by a linear regression analysis in an open one-compartment model (Table 2).

TABLE 2.

Pharmacokinetic parameters of the antimicrobials used in animal model and t>MIC of the antibiotics for K. pneumoniae strains 12 and 12dp^a

Antimicrobial (dose in mg/kg)	Cmax (mg/liter)	t1/2 (h)	t>MIC (h) for K. pneumoniae strain:
			12	12dp
Imipenem (40)	49.65	0.52	1.4	1
Meropenem (40)	63.77	0.52	1.4	1.2
Cefepime (60)	209.2	1.1	2.5	2.4

^aC_max, maximum serum concentration; t_1/2, half-life; t>MIC, time during which the plasma concentration remained above the MIC. See footnote a of Table 1 for strain descriptions.

The pneumonia model used was a modification of a model previously developed by Mimoz et al. (12, 13). Rats were intraperitoneally anesthetized with ketamine (Pfizer, Madrid, Spain) (100 mg/kg of body weight) plus xylazine (Sigma-Aldrich, Madrid, Spain) (10 mg/kg of body weight) and transtracheally inoculated with 10⁶ CFU of K. pneumoniae 12 or 10⁷ CFU of K. pneumoniae 12dp plus 0.5% agar (Oxoid SA, Madrid, Spain). Bilateral pneumonia was observed by standard techniques (19).

Treatment groups, according to pharmacokinetic parameters, received intraperitoneal injections of 40 mg/kg/2 h for imipenem or meropenem-cilastatin or 60 mg/kg/4 h for cefepime. Therapy was started 4 h after bacterial inoculation and continued for 3 days. Animal sacrifice was performed 8 h after the last antibiotic dose. The lungs were quantitatively cultured, and viable bacteria were expressed as log₁₀ CFU/g of lung. Survival at day 3 was also recorded. A Kruskal-Wallis test and a Mann-Whitney U test were used for statistical analysis by the SPSS 8.0 statistical package (SPSS Inc., Chicago, IL). Emergence of resistance during β-lactam therapy was examined at the end of each experiment by plating 400 μl of the lung homogenates onto MacConkey agar with 4× MIC (mg/liter) of imipenem, cefepime, or meropenem.

In our study, the introduction of the PACBL ACT-1 in the K. pneumoniae 52145 strain (K. pneumoniae strain 12) did not change the MICs of carbapenems but slightly increased the cefepime MIC from 0.03 to 0.25 μg/ml. The same strain with the loss of OmpK35 and OmpK36 porins (K. pneumoniae strain 12dp), however, had an increase in MICs of imipenem from 0.25 to 16 mg/liter, of meropenem from 0.03 to 8 mg/liter, and of cefepime from 0.25 to 2 mg/liter (Table 1). Bradford et al. (4) demonstrated for the first time in vitro imipenem resistance in K. pneumoniae strains with the combination of ACT-1 and the loss of an outer membrane protein. Similar results were later reported with other PACBL (11).

Mortality and bacterial counts in the lungs of the animals in the different groups and with the different strains are summarized in Table 3. The efficacy results in rats with pneumonia infected with K. pneumoniae 12 and K. pneumoniae 12dp strains show that antimicrobial activities of imipenem, meropenem, and cefepime seem to be adequate for the treatment of pneumonia caused by AmpC β-lactamase ACT-1-producing strains with or without porin deficiency, with a reduction of 5 log₁₀ CFU/g and 0% mortality, with no differences between them (P > 0.05). However, for treated animals infected with K. pneumoniae plasmid AmpC without OmpK35 and OmpK36 porins, meropenem seems to be more active than imipenem (P ≤ 0.05). Recently, Pichardo et al. (16, 17) have also demonstrated that carbapenems and cefepime produce similar reductions of the bacterial counts, with imipenem producing the best clearance in an animal model of pulmonary infection with a porin (OmpK35 and OmpK36)-deficient K. pneumoniae strain or with the same strain with a plasmid-mediated AmpC β-lactamase (FOX-5 or CMY-2). In those studies, only true stable mutants were selected by cefepime treatment. We did not observe this phenomenon, even though it may be also possible that the bacterial density of the lungs in treated animals was too low to detect such subpopulations, as these mutants are reported to occur in 1 of 10⁵ to 10⁸ wild-type strains possessing AmpC β-lactamases (11).

TABLE 3.

Bacterial counts in lungs and mortality for bacterial strains studied with different therapeutic groups and placebo^a

Antimicrobial	Result for K. pneumoniae 12			Result for K. pneumoniae 12dp
	No. of animals	% Mortality	Bacterial count in lungs (log10 CFU/g [median ± SD])	No. of animals	% Mortality	Bacterial count in lungs (log10 CFU/g [median ± SD])
Meropenem	24	0	1.71 ± 0.86	20	0	2.42 ± 0.74
Imipenem	24	0	1.87 ± 1.05	20	0	2.98 ± 0.59
Cefepime	23	0	2.00 ± 1.06	20	0	2.77 ± 0.45
None	8	50	8.01 ± 1.49	8	50	9.61 ± 1.05

^aDosages were 40 mg/kg/2 h for imipenem and meropenem and 60 mg/kg/4 h for cefepime. See footnote a of Table 1 for strain descriptions.

We conclude that imipenem, meropenem, and cefepime decrease bacterial lung counts in K. pneumoniae strains producing ACT-1 with or without the presence of porins, probably due to good antibiotic tissue concentrations (over MIC) by the compounds. However, little difference could be seen regardless of the AmpC-type β-lactamase strain used for infection, since some antibiotics could be slightly more effective in decreasing lung bacterial burden.