CTX-M β-Lactamase Production and Virulence of Escherichia coli K1

We report a patient with neonatal meningitis caused by a CTX-M-1–producing Escherichia coli K1 strain. The influence of CTX-M production on virulence was investigated in cell culture and a newborn mouse model of meningitis. CTX-M production had no influence on virulence but was a major factor in clinical outcome.


The Study
In April 2007, a 39-year-old pregnant woman with amniotic sac rupture was admitted to a hospital in Orleans, France, at 28 weeks and 4 days of gestation. Treatment was started with betamethasone (12 mg 1×/d) for fetal lung maturation and amoxicillin (1g 3×/d) for 4 days. Because of a high serum level of C-reactive protein, antimicrobial drug therapy was switched to amoxicillin with clavulanic acid (1g 3×/d) for 1 day. A cesarean delivery was performed at 29 weeks and 2 days of gestation. A lumbar puncture sample of the low-weight (1,560 g) newborn female was tinged with blood. Cerebrospinal fl uid (CSF) protein and glucose values were 4.00 g/L and 3.5 mmol/L, respectively. Results of CSF Gram staining were negative.
The infant was admitted to the neonatal critical care unit and received amoxicillin (150 mg), cefotaxime (120 mg), and netilmicin (8 mg) 2×/d for 2 days. Culture of placenta, maternal and infant blood, and infant gastric fl uid yielded E. coli. The isolate was resistant to antimicrobial drugs, including third-generation cephalosporins. Imipenem/cilastatin (25 mg 4×/d) and amikacin (15 mg 2×/d) were given for 2 days. Treatment with imipenem/cilastatin was given for 15 days and then stopped because of the infant's clinical improvement and return of C-reactive protein to the reference level. Similar drug treatment was administrated to the mother.
One week after drug treatment was discontinued, the infant showed signs of septicemia. A second lumbar puncture sample had protein and glucose levels of 4.56 g/L and 0.1 mmol/L, respectively, and a leukocyte count of 4,700 cells/μL (54% polymorphonuclear cells). E. coli were isolated from blood and CSF cultures and showed a resistance pattern identical to that of the previous isolate. Meningitis was a complication of the initial sepsis or a relapse of initial unapparent meningitis (8).
Treatment was started with imipenem/cilastatin (30 mg 4×/d) for 25 days and amikacin (15 mg 2×/d) for 5 days. Because the infant had a seizure, phenobarbital (22.5 mg) and ciprofl oxacin (15 mg 2×/d) were prescribed for 5 additional days. Her condition gradually improved and blood and CSF values returned to reference levels. The infant was discharged from the hospital 1 month later and treatment with the anticonvulsant was discontinued. She showed normal psychomotor development at a regular follow-up pediatric visit.
E. coli strains isolated from the mother and infant were indistinguishable by enterobacterial repetitive intergenic consensus sequence 2 PCR, random amplifi ed polymorphic DNA analysis, and typing with a MALDI BioTyper (Bruker Daltonique, Wissembourg, France) (9). Thus, the isolates corresponded to the same strain, designated Orl-1. PCR-based phylogenetic analysis and serotyping showed that the strain belonged to group B2 and serotype O7:K1:H7, a major O antigen encountered worldwide in NMEC (1).
A derivative strain that did not harbor the 3 plasmids (Orl-c) was obtained by plasmid elimination with ethidium bromide. Orl-1, Orl-c, and E. coli DH5α harboring pOrl-1-CTX-M-1 were tested for invasiveness in human brain microvascular endothelial cells (10) and in a newborn mouse (C57BL/6 wild-type) model of meningitis (R. Mittal et al., unpub. data) to investigate the infl uence of CTX-M-1 production on virulence. E. coli strain E44, a rifampicinresistant mutant of archetypical NMEC K1 strain RS218, was used as a positive control (10).
The difference between Orl-1 and Orl-c in the mouse model may be explained by loss of plasmid pOrl-1-Te from Orl-c. Plasmid pOrl-1-Te is likely similar to pS88-related plasmids of NMEC because they share 3 virulence factor genes (iss, aer, and iroN) and are large and conjugative. These plasmids contribute to virulence of NMEC (11). Orl-1 and Orl-c showed similar behaviors, which suggested that the CTX-M-1-encoding plasmid pOrl-1-CTX-M-1 does not alter virulence of the strain.
Mice with E44-and Orl-1-induced neonatal meningitis were treated with the third-generation cephalosporin cefotaxime, as recommended for humans. Despite antimicrobial drug treatment, Orl-1, but not strain E44, caused meningitis, suggesting that drug resistance is a major factor in clinical outcomes.

Conclusions
Studies have reported emergence of E. coli as the predominant organisms responsible for sepsis at any gestational age and for increased rates of drug-resistant E. coli caused by intrapartum drug prophylaxis (12). Spread of ESBLs in E. coli and intrapartum exposure to antimicrobial drugs may favor emergence of NMEC strains resistant to third-generation cephalosporins.
Two other well-characterized E. coli K1 strains producing ESBLs have been isolated from patients with neonatal meningitis in Algeria and France. The ESBL was identifi ed as CTX-M-15 in both patients, and 1 infection was lethal (13,14). Other putative ESBL-producing E. coli K1 have been recently isolated, especially in developing countries (15).
Emergence of ESBL-producing E. coli strains, which are frequently resistant to fl uoroquinolone (2), highlights the need for possible alternatives to third-generation ce-  phalosporins for treatment of patients with infected with NMEC. Carbapenems are usually recommended for treatment of infections with ESBL-producing E. coli (2). However, this case report shows the role of treatment duration and the need for additional pharmacokinetic and safety studies in neonates and for adjunctive therapies (8). This characterization of a CTX-M-1-producing NMEC strain highlights the emergence of CTX-M-type ESBL in highly virulent E. coli. Because of worldwide spread of CTX-Ms, caution should be exercised in the management of patients with NMEC, and fi rst-line treatment for neonatal meningitis may need to be reconsidered.