ABSTRACT
Nocardia farcinica, one of the most frequent pathogenic species responsible for nocardiosis, is characterized by frequent brain involvement and resistance to β-lactams mediated by a class A β-lactamase. Kinetic parameters for hydrolysis of various β-lactams by FARIFM10152 from strain IFM 10152 were determined by spectrophotometry revealing a high catalytic activity (kcat/Km) for amoxicillin, aztreonam, and nitrocefin. For cephems, kcat/Km was lower but remained greater than 104 M−1 s−1. A low catalytic activity was observed for meropenem, imipenem, and ceftazidime hydrolysis. FARIFM10152 inhibition by avibactam and clavulanate was compared using nitrocefin as a reporter substrate. FARIFM10152 was efficaciously inhibited by avibactam with a carbamoylation rate constant (k2/Ki) of (1.7 ± 0.3) × 104 M−1 s−1. The 50% effective concentrations (EC50s) of avibactam and clavulanate were 0.060 ± 0.007 μM and 0.28 ± 0.06 μM, respectively. Amoxicillin, cefotaxime, imipenem, and meropenem MICs were measured for ten clinical strains in the presence of avibactam and clavulanate. At 4 μg/ml, avibactam and clavulanate restored amoxicillin susceptibility in all but one of the tested strains but had no effect on the MICs of cefotaxime, imipenem, and meropenem. At 0.4 μg/ml, amoxicillin susceptibility (MIC ≤ 8 μg/ml) was restored for 9 out of 10 strains by avibactam but only for 4 out of 10 strains by clavulanate. Together, these results indicate that avibactam was at least as potent as clavulanate, suggesting that the amoxicillin-avibactam combination could be considered as an option for the rescue treatment of N. farcinica infections if clavulanate cannot be used.
INTRODUCTION
Nocardia spp. are environmental filamentous Gram-positive bacteria causing severe, though rare, opportunistic infections called nocardiosis (1). These infections mostly occur among solid-organ or allogeneic hematopoietic stem transplant recipients or patients receiving corticosteroids with mortality rates of 20 to 30% (1–3). In cases of severe disease, the initial antibiotic regimen for nocardiosis usually includes co-trimoxazole associated with one or two other drugs among the aminoglycoside (amikacin) and the β-lactam (imipenem or cefotaxime) families (4). Such combination therapies are justified by the high diversity of antibiotic resistance profiles among different Nocardia spp. (5–7). The growing use of molecular biology (DNA sequencing) has enabled the establishment of correlations between Nocardia species and antibiotic susceptibility patterns, especially for β-lactams (5–7). For instance, Nocardia farcinica, one of the leading human pathogens among solid-organ transplant patients and frequently involved in brain infection, is almost always resistant to amoxicillin (90 to 100%) and third-generation cephalosporins (55 to 100%). Additional resistance to imipenem (overall frequency of 23 to 67%) (5–9) precludes the use of any β-lactam (coresistance frequency of 23%).
The N. farcinica class A β-lactamase FAR-1 has previously been produced and purified (11). This enzyme preferentially hydrolyses penams (benzylpenicillin, amoxicillin, ticarcillin, and piperacillin), and clavulanate is the most potent inhibitor of FAR-1 among β-lactamase inhibitors belonging to the β-lactam family (clavulanate, tazobactam, and sulbactam) (11). Indeed, ∼80% of N. farcinica isolates are susceptible to the amoxicillin-clavulanate combination (5–7, 9). Avibactam is a non-β-lactam β-lactamase inhibitor belonging to the diazabicyclooctane family, which has been recently approved (12); it inhibits Ambler class A, C, and some class D β-lactamases (13). Here, we have evaluated whether this second-generation β-lactamase inhibitor has the potential to restore the activity of β-lactams against N. farcinica. We report a comparison of the inhibition activity of clavulanate and avibactam based on purification of a close homologue of FAR-1 (FARIFM10152) and determination of the antibacterial activity of various β-lactam–inhibitor combinations against ten N. farcinica clinical isolates.
RESULTS AND DISCUSSION
Kinetic parameters for β-lactam hydrolysis by FARIFM10152.FARIFM10152 showed a high catalytic activity (kcat/Km > 105 M−1 s−1) for hydrolysis of amoxicillin, aztreonam, and nitrocefin (Table 1). For cephems, such as cefuroxime or cefotaxime, kcat/Km was lower but remained higher than 104 M−1 s−1. Meropenem and imipenem were hydrolyzed by FARIFM10152 with low efficacies (kcat/Km of 102 and 103 M−1 s−1, respectively), mostly due to low kcat values of 0.003 and 0.6 s−1, respectively. These values are similar to those previously reported for FAR-1, as expected from the high amino acid identity (98.7%) between FARIFM10152 and FAR-1 (see Fig. S1 in the supplemental material).
Kinetic parameters for hydrolysis of β-lactams by FARIFM10152
Inhibition of FARIFM10152 by avibactam.FARIFM10152 was efficaciously inhibited by avibactam with a high carbamoylation rate constant (k2/Ki) of (1.7 ± 0.3) × 104 M−1 s−1 combined with a low decarbamoylation rate constant (k−2) of (1.1 ± 0.3) × 10−4 s−1 (see Fig. S2 in the supplemental material). These values are typical for inhibition of class A enzymes from the Enterobacteriaceae, such as KPC-2, TEM-1, and CTX-M-15 (k2/Ki range of 1 × 104 to 10 × 104 M−1 s−1) (13).
Comparison of the activity of FARIFM10152 inhibition by avibactam and clavulanate.Since β-lactamase inhibition by avibactam and clavulanate involves distinct reaction schemes (see Fig. S3 in the supplemental material), we determined the concentration of inhibitor required to reduce the hydrolysis rate of nitrocefin by 50% (EC50s). After a 5-min incubation, EC50s of avibactam and clavulanate were 0.060 ± 0.007 μM and 0.28 ± 0.06 μM, respectively, demonstrating that avibactam was ca. 5-fold more efficacious than clavulanate (see Fig. S4 and Table S1 in the supplemental material). As a comparison, Laurent et al. identified clavulanate as the most potent inhibitor of FAR-1 when compared to tazobactam or sulbactam, with benzylpenicillin EC50s of 0.3 versus 29 and 600 μM, respectively (11).
Avibactam restores N. farcinica susceptibility to amoxicillin.All N. farcinica strains were resistant to amoxicillin with MICs of ≥64 μg/ml (Fig. 1). Only two strains were susceptible to cefotaxime with MICs of 6 and 8 μg/ml. Avibactam reduced amoxicillin MICs (16- to 32-fold) and restored susceptibility to amoxicillin in all but one of the tested strains. Avibactam did not significantly improve the MICs of cefotaxime, imipenem, and meropenem, as expected from the low activity of hydrolysis of these β-lactams by FARIFM10152 (fold reduction of ≤2). Avibactam and clavulanate had a similar impact on the MICs of the four β-lactams. Of note, strain 3712 was more resistant to carbapenems than the other strains, suggesting acquisition of an unknown resistance mechanism. In order to refine the comparison of avibactam and clavulanate, the two inhibitors were tested at various concentrations (Fig. 2). Decreases in the amoxicillin MICs displayed a dose response for both inhibitors, and the fold reductions of the MICs were 2- to 4-fold higher for avibactam than for clavulanate for the intermediary concentration of 0.4 μg/ml. At this inhibitor concentration, susceptibility to amoxicillin (MIC ≤ 8 μg/ml) was restored for 9 out of 10 strains by avibactam but only for 4 out of 10 strains by clavulanate. Thus, avibactam proves to be slightly more efficacious than clavulanate for improving the antibacterial activity of substrates of FARIFM10152.
MICs of amoxicillin (A), cefotaxime (B), imipenem (C), and meropenem (D) for ten clinical isolates of N. farcinica with or without a fixed concentration of avibactam or clavulanate (4 μg/ml). Data are medians of three independent experiments. Dotted lines indicate the MIC susceptibility breakpoints adapted from the CLSI (≤8 μg/ml for amoxicillin and cefotaxime and ≤4 μg/ml for imipenem and meropenem).
MICs of amoxicillin for ten N. farcinica isolates in the presence of various concentrations of avibactam or clavulanate. Data are medians of three independent experiments. Dotted and solid lines represent MIC breakpoints proposed by the CLSI to define susceptible (≤8 μg/ml) and resistant (≥32 μg/ml) strains, respectively.
Conclusions.Here, we show that avibactam is slightly superior to clavulanate for inhibiting the β-lactamase FARIFM10152 and for restoring the activity of β-lactams against β-lactamase-producing strains of N. farcinica. The inhibitory activity of avibactam characterized in this study, associated with a likely satisfying cerebrospinal fluid (CSF) penetration, suggests that the amoxicillin-avibactam combination could be considered as a rescue treatment for N. farcinica infections in cases of clavulanate-related side effects or of brain involvement (14). Indeed, even if amoxicillin-clavulanate appears to be a safe option in the case of infections involving lungs, skin, or soft tissues, its use in the case of brain involvement appears to be more questionable due to penetration issues. Among 18 patients with bacterial meningitis treated with intravenous amoxicillin-clavulanate (200 and 20 mg/kg per day), clavulanate levels in cerebrospinal fluid (CSF) were ≤0.05 μg/ml in 5 out of 18 samples and ranged from 0.1 to 0.8 μg/ml in the others (15). For avibactam, data regarding its use for CNS infections is very limited. Avibactam penetration in rabbit CSF is about 38% (16). Six case reports have described the successful use of intravenous ceftazidime/avibactam for the treatment of CNS infections caused by multidrug-resistant bacteria (17–22). Even if obvious publication bias may preclude the identification of therapeutic failure of avibactam-containing regimens for CNS infections, these data suggest that avibactam penetration in CSF is sufficient for in situ β-lactamase inhibition.
MATERIALS AND METHODS
Bacterial strains.N. farcinica IFM 10152 is a reference strain whose genome has previously been sequenced (GenBank accession number AP006618) (23). Ten N. farcinica clinical isolates were selected from the collection of the Observatoire Français des Nocardioses (see Table S2 in the supplemental material) (5).
Production and purification of FARIFM10152.A portion of the nfa23080 gene of N. farcinica IFM 10152 encoding residues 34 to 313 of FARIFM10152 (see Fig. S1 in the supplemental material) was amplified (primers 5′-AAACATATGGGTTCCGAGCCCGCC-3′ and 5′-AAACTCGAGTCATCGCAGGGCG-3′, which contained NdeI and XhoI restriction sites [underlined]). The PCR product was digested and cloned into the pET-TEV vector (24) in order to generate a translational fusion with a vector-encoded N-terminal 6×His tag followed by a TEV cleavage site (see Fig. S1) (24). FARIFM10152 was produced in Escherichia coli BL21(DE3), as described (24). FARIFM10152 was purified by affinity (Ni-nitrilotriacetic acid [NTA] resin; Sigma-Aldrich) and size exclusion (Superdex 200 HL26/60 column; Amersham Pharmacia Biotech) chromatography in 25 mM Tris-HCl (pH 7.5) containing 300 mM NaCl.
Determination of steady-state kinetic parameters for hydrolysis of β-lactams by FARIFM10152.Hydrolysis of β-lactams was monitored by spectrophotometry (Cary-100 Bio; Agilent) at 20°C in 100 mM 2-(N-morpholino)ethanesulfonic acid (MES) at pH 6.4, as previously described (see Table S3 in the supplemental material) (25). Kinetic parameters kcat and Km were determined by measuring the initial hydrolysis rates for various antibiotic concentrations. Experimental data were fitted to the Michaelis-Menten equation on SigmaPlot software (24) (V, initial velocity; [E] and [S], initial enzyme and substrate concentrations, respectively).
Inhibition of FARIFM10152 by avibactam and clavulanate.The rate constants for carbamoylation (k2/Ki and k−2) of FARIFM10152 by avibactam were determined by using nitrocefin (100 μM) as the reporter substrate (see Fig. S2 in the supplemental material) (26, 27). The efficacies of FARIFM10152 inhibition by clavulanate and avibactam were compared by determining EC50s (inhibitor concentrations that reduce the initial hydrolysis rate of nitrocefin by 50%) (28). FARIFM10152 (0.1 nM) was incubated for 5 or 10 min at 30°C with various concentrations of inhibitors; residual β-lactamase activity was determined by adding nitrocefin (100 μM) (see Fig. S4 in the supplemental material).
In vitro susceptibility testing.MICs were determined by broth microdilution in 96-well plates in cation-adjusted Mueller-Hinton (CA-MH) with a final density of 5 × 105 CFU/ml in each well (29). Microplates were incubated for 72 h at 37°C (for breakpoints, see Table S4 in the supplemental material) (29).
ACKNOWLEDGMENTS
We thank Veronica Rodriguez-Nava for kindly providing the N. farcinica IFM 10152 reference strain and clinical isolates.
Z.E. was supported by the Fondation pour la Recherche Médicale (grant ECO20160736080). C.O. was supported by the “Année Recherche” from the Assistance Publique-Hôpitaux de Paris.
The authors declare no conflict of interest.
FOOTNOTES
- Received 1 August 2019.
- Returned for modification 3 September 2019.
- Accepted 4 November 2019.
- Accepted manuscript posted online 11 November 2019.
Supplemental material is available online only.
- Copyright © 2020 American Society for Microbiology.
