In Vitro Activity of Tebipenem, a New Oral Carbapenem Antibiotic, against β-Lactamase-Nonproducing, Ampicillin-Resistant Haemophilus influenzae

Abstract

In vitro activity of tebipenem, a new oral carbapenem antibiotic, against clinical Haemophilus influenzae isolates was compared with those of 8 reference agents. Isolates were classified into 6 resistance classes after PCR identification of β-lactamase genes and ftsI gene mutations. For all isolates, the minimal concentration at which 90% of isolates were inhibited was lower for tebipenem than for the reference oral antibiotics, except for cefditoren. Tebipenem also showed excellent bactericidal activity against β-lactamase-nonproducing, ampicillin-resistant isolates.

Tebipenem-pivoxil (TBM-PI) is a new oral carbapenem agent whose active metabolite, tebipenem (TBM), shows broad-spectrum activity against Gram-positive and -negative bacteria, except for Pseudomonas aeruginosa (4, 7). Unlike other clinically available oral β-lactam antibiotics, TBM displays excellent activity against Streptococcus pneumoniae strains, including penicillin-resistant strains (5). In consideration of these characteristics, the clinical efficacy of TBM-PI for treatment of community-acquired pediatric infections such as pneumonia, acute otitis media (AOM), and sinusitis was investigated. In a phase III clinical study of AOM, the efficacy of TBM-PI (3.5 to 5 mg/kg twice a day [b.i.d.]) was equal to that of high-dose cefditoren-pivoxil (CDN-PI; 4.2 to 6 mg/kg three times a day [t.i.d.]) (11). Among the important pathogens causing pneumonia and AOM, the antimicrobial characteristics of TBM against Haemophilus influenzae are not yet verified. Furthermore, β-lactamase-nonproducing, ampicillin-resistant (BLNAR) strains are increasing in Japan and various other countries (9). In the present study, we evaluated in vitro antibacterial and bactericidal activities of TBM and reference agents against H. influenzae strains, including BLNAR strains.

A total of 232 H. influenzae strains were collected between October 2005 and December 2008 from pediatric patients at random. The details of the strains were as follows: 112 were obtained from tympanic effusions in patients with AOM, 30 were obtained from patients with pneumonia, and 90 were obtained from cerebrospinal fluid samples collected from patients with meningitis. To clarify the genetic background as related to β-lactam resistance, we used PCR identification to place H. influenzae strains in 1 of 6 genetic resistance classes (classes indicated with a “g” prefix) (Table 1) (2). Susceptibility testing was performed using an agar dilution method (3). Muller-Hinton (MH) agar (Becton Dickinson, Sparks, MD) with 0.5% yeast extract, 2% defibrinated, heat-treated horse blood, and 15 μg of β-NAD⁺ per ml (subsequently referred to as “the 3 supplements”) was used. Plates were examined after incubation at 37°C in a 5% CO₂ atmosphere for 20 h. Table 2 shows the MIC ranges, MIC₅₀s, and MIC₉₀s of TBM and 8 reference antibiotics against H. influenzae strains in each of the 6 genetic resistance classes (total n = 232). Antimicrobial activities of TBM for genetically β-lactamase-nonproducing, ampicillin-susceptible (gBLNAS), genetically low β-lactamase-nonproducing, ampicillin-resistant (gLow-BLNAR), and genetically β-lactamase-nonproducing, ampicillin-resistant (gBLNAR) strains were essentially equal to those of meropenem. The increased ratios of MIC₉₀s between gBLNAS and gBLNAR strains for TBM, cefditoren, and meropenem were lower (4 to 8 times) than those for cefotaxime, amoxicillin, and cefdinir (32 to 64 times). The susceptibility distribution of the 232 H. influenzae strains to TBM according to resistance class is shown in Fig. 1. The MIC peaks of TBM against gBLNAS and gBLNAR strains were 0.063 and 0.5 μg/ml, respectively. Antimicrobial activities of cephalosporins, except for cefditoren, were greatly decreased by amino acid substitutions in penicillin-binding protein 3 (PBP3) of BLNAR strains (9). The reason why TBM was affected relatively little by these amino acid substitutions appears to involve the binding affinities of TBM for PBPs; TBM was found to bind not only to PBP3 but also to PBP1B, PBP2, and PBP4 (10).

TABLE 1.

Identification of 6 resistance classes based on PCR

Resistance class	Genetic background results
	Amino acid substitutions in PBP3a		β-lactamase gene
	Asn526Lys or Arg517His	Ser385Thr and Asn526Lys or Ser385Thr and Arg517His
gBLNAS	−	−	−
gLow-BLNAR	+	−	−
gBLNAR	−	+	−
gBLPAR	−	−	+
gBLPACR-I	+	−	+
gBLPACR-II	−	+	+

^aAmino acid substitutions in penicillin-binding protein 3 (PBP3) that affect decreases of susceptibilities of H. influenzae strains for β-lactam antibiotics were selected (12).

TABLE 2.

Comparison of in vitro activities of tebipenem and those of reference antibiotics against H. influenzae strains classified genotypically by PCR

Resistance class (no. of H. influenzae strains) and antibiotica	MIC (μg/ml)b
	Range	50%	90%
gBLNAS (65)
Tebipenem	0.008-0.25	0.063	0.12
Ampicillin	0.12-1	0.25	0.5
Amoxicillin	0.12-1	0.5	0.5
Cefdinir	0.12-1	0.25	0.5
Cefditoren	0.002-0.063	0.016	0.031
Cefotaxime	0.004-0.063	0.016	0.031
Meropenem	0.016-0.12	0.063	0.12
Clarithromycin	4-16	8	16
Azithromycin	0.25-8	2	4
gLow-BLNAR (32)
Tebipenem	0.031-0.5	0.25	0.5
Ampicillin	0.5-2	1	1
Amoxicillin	0.5-4	2	4
Cefdinir	0.5-4	0.5	2
Cefditoren	0.016-0.063	0.031	0.063
Cefotaxime	0.016-0.12	0.031	0.12
Meropenem	0.063-0.5	0.12	0.25
Clarithromycin	4-16	8	16
Azithromycin	0.5-4	2	2
gBLNAR (119)
Tebipenem	0.031-1	0.25	1
Ampicillin	0.5-32	2	8
Amoxicillin	0.25-64	8	32
Cefdinir	2-32	8	32
Cefditoren	0.031-1	0.25	0.25
Cefotaxime	0.063-4	0.5	1
Meropenem	0.031-0.5	0.25	0.5
Clarithromycin	4-32	8	16
Azithromycin	0.5-8	2	4
gBLPAR (TEM-1 [6])
Tebipenem	0.063-0.12
Ampicillin	8-32
Amoxicillin	8-16
Cefdinir	0.25-0.5
Cefditoren	0.008-0.031
Cefotaxime	0.008-0.031
Meropenem	0.031-0.063
Clarithromycin	8-16
Azithromycin	1-2
gBLPACR-I (2)
Tebipenem	0.12-0.25
Ampicillin	8-32
Amoxicillin	8
Cefdinir	1
Cefditoren	0.016
Cefotaxime	0.031-0.063
Meropenem	0.063-0.12
Clarithromycin	8
Azithromycin	2
gBLPACR-II (8)
Tebipenem	0.25-0.5
Ampicillin	2->64
Amoxicillin	2->64
Cefdinir	8-16
Cefditoren	0.12
Cefotaxime	0.25-0.5
Meropenem	0.063-0.25
Clarithromycin	8->64
Azithromycin	2-64

^aThe oral antibiotics used in this study were tebipenem, ampicillin, amoxicillin, cefditoren, cefdinir, clarithromycin, and azithromycin. The parenteral antibiotics used in this study were cefotaxime and meropenem. A total of 232 H. influenzae strains were used.

^bThe breakpoint MICs (in μg/ml) of each antibiotic against susceptible strains are as follows: ampicillin, ≤1; amoxicillin-clavulanic acid, ≤4/2; cefdinir, ≤1; clarithromycin, ≤8; azithromycin, ≤4; cefotaxime, ≤2; and meropenem, ≤0.5 (1).

FIG. 1.

Distribution of MICs of TBM for clinical isolates of H. influenzae (n = 232) classified into groups consisting of gBLNAS, genetically β-lactamase-producing ampicillin-resistant (gBLPAR), gLow-BLNAR, gBLNAR, genetically β-lactamase-producing, amoxicillin-clavulanic acid-resistant I (gBLPACR-I), and gBLPACR-II strains (“I” and “II” indicate different substitutions in PBP3) (Table 1).

Time-kill curves for TBM and reference antibiotics in BLNAR strains JPH002 and JPH1306 were determined at concentrations corresponding to the MIC and double the MIC. Colonies precultured on chocolate II agar plates (Nippon Becton Dickinson, Tokyo, Japan) were suspended in tubes of MH broth, with the turbidity adjusted to a 0.5 McFarland standard. This bacterial suspension, diluted 10-fold using MH broth with the 3 supplements, was then grown at 37°C for 120 min. The culture (500 μl) then was inoculated into 9.5 ml of fresh MH broth containing each of the antibiotics with the 3 supplements and 9.5 ml of broth additionally supplemented with 10% fresh human serum to approximate conditions in vivo. Tubes then were incubated without shaking, and cultures were sampled at predetermined intervals.

Figure 2 shows time-kill curves for TBM and β-lactam reference antibiotics at concentrations corresponding to the MIC and double the MIC for gBLNAR strain JPH002. A 3-log₁₀ reduction of bacterial cells by TBM in the supplemented MH broth was achieved at 4 h of exposure at the MIC. The same reduction by cefditoren, ampicillin, or amoxicillin required 6 h or more of exposure at the MIC. Using supplemented MH broth containing 10% human serum, a 3-log₁₀ reduction of bacterial cells by TBM was achieved with 2 h of exposure at the MIC. In contrast, similar reductions by cefditoren, ampicillin, and amoxicillin required 6, 4, and 6 h of exposure at the MIC, respectively. Using this approximation of conditions in vivo, the bactericidal activity of TBM against gBLNAR strains was observed within a short time after initiating exposure to TBM at equal to or greater than the MIC and was higher than those of cefditoren, ampicillin, and amoxicillin. The activity of TBM in the supplemented MH broth including serum was even better than that in supplemented broth without serum. The bactericidal activity of TBM against gBLNAR strain JPH1306 was similar to that against gBLNAR strain JPH002 (data not shown).

FIG. 2.

Time-kill curves for TBM and 3 other β-lactam antibiotics at the MICs (•) and double the MICs (▴) for gBLNAR strain JPH002. ○, control (no drug). HTM, MH broth supplemented with 0.5% yeast extract, 2% defibrinated and heat-treated horse blood, and 15 μg of β-NAD⁺ per ml.

The functions of PBPs in H. influenzae can be deduced from those in Escherichia coli, considering that PBPs in H. influenzae and E. coli show high homology (6). Inhibition of PBP1A and -B in E. coli causes a rapid lysis reaction (8). In the present study, the bactericidal activity of TBM against gBLNAR strains was superior to those of cefditoren, ampicillin, and amoxicillin. This excellent bactericidal activity of TBM against H. influenzae in vitro also may relate to the lysis reaction caused by inhibition of PBP1A and -B.

In summary, the antibacterial and bactericidal activities of TBM against BLNAR H. influenzae were different from those of penicillins and cephalosporins. These phenomena might be related to the characteristic binding affinity of TBM for PBPs.