ABSTRACT
For 297 penicillin-susceptible, -intermediate, and -resistant pneumococcal strains, the sulopenem MIC50s were 0.008, 0.06, and 0.25, respectively, and the sulopenem MIC90s were 0.016, 0.25, and 0.5 μg/ml, respectively. The MIC50s of amoxicillin for the corresponding strains were 0.03, 0.25, and 2.0 μg/ml, respectively, and the MIC90s were 0.03, 1.0, and 8.0 μg/ml, respectively. The combination of amoxicillin and clavulanate gave MICs similar to those obtained with amoxicillin alone. The sulopenem MICs were similar to those of imipenem and meropenem. The MICs of ß-lactams increased with those of penicillin G, and among the quinolones tested, moxifloxacin had the lowest MICs. Additionally, 45 strains of drug-resistant type 19A pneumococci were tested by agar dilution and gave sulopenem MIC50s and MIC90s of 1.0 and 2.0 μg/ml, respectively. Both sulopenem and amoxicillin (with and without clavulanate) were bactericidal against all 12 strains tested at 2× MIC after 24 h. Thirty-one strains from 10 countries with various penicillin, amoxicillin, and carbapenems MICs, including those with the highest sulopenem MICs, were selected for sequencing analysis of the pbp1a, pbp2x, and pbp2b regions encoding the transpeptidase active site and MurM. We did not find any correlations between specific penicillin-binding protein-MurM patterns and changes in the MICs.
The incidence of pneumococci resistant to penicillin G and other ß-lactam and non-ß-lactam compounds has increased at an alarming rate worldwide, including the United States. The major foci of infections currently include South Africa, Spain, and Central and Eastern Europe (26, 27). A 1997 survey showed that 50.4% of 1,476 clinically significant pneumococcal isolates in the United States were not susceptible to penicillin (26). Of all strains tested, approximately 33% were macrolide resistant, with the highest rate of macrolide resistance being seen among penicillin-resistant strains, for which the MICs were ≥2.0 μg/ml (26). A later survey, conducted as part of the Alexander Project (27), reported that the worldwide prevalence of pneumococci isolated between 1998 and 2000 for which penicillin G MICs were ≥2 μg/ml was 18.2%, with the overall macrolide resistance rate being 24.6%. It is also important to note the higher rates of isolation of penicillin-intermediate and -resistant pneumococci (approximately 30%) in middle ear fluids from patients with refractory otitis media than from other isolation sites (4). Quinolone-resistant pneumococci have also been described (6, 22, 31). The problem of drug-resistant pneumococci is compounded by the ability of resistant clones to spread from country to country and from continent to continent (32, 34). In a recent paper, Pichichero and Casey (40) reported on the emergence in the United States of an otopathogenic strain of pneumococcus type 19A which is not covered by the current pediatric conjugate vaccine and which is resistant to all currently FDA-approved antibiotics for the treatment of acute otitis media in children. Another very recent study has shown that of 393 isolates of pneumococci collected from children in the United States between 2005 and 2006, 30.5% of all the isolates were serotype 19A; >50% of the serotype 19A strains showed multiresistant susceptibility patterns (9).
ß-Lactam resistance in Streptococcus pneumoniae is mediated by stepwise alterations of penicillin-binding proteins (PBPs) that result in decreased antibiotic affinities (21). Pneumococci contain a set of six PBPs. The decreased affinities of PBP 1A, 2X, and 2B for β-lactams have been reported to play an important part in resistance and to cause high-level penicillin G resistance (1, 2, 30, 41, 42, 51). Alterations in the conserved motifs in PBP 2B are associated with penicillin G resistance, and alterations in PBP 2X mediate low-level resistance to cephalosporins (1, 8, 11, 18). The region of PBP 2B from amino acids 538 to 642 also appears to be relevant for resistance, especially among strains with high amoxicillin MICs (5, 29). Additional alterations in PBP 1A raise the penicillin G and cefotaxime MICs (1, 35, 42, 47). These specific PBPs are thought to be the key to resistance to each antibacterial agent. Non-PBP-related resistance determinants are also essential for the development of high-level penicillin G and cephalosporin resistance in pneumococcal isolates. The mechanism used by non-PBP-related resistance determinants involves alterations in MurM, an enzyme involved in the biosynthesis of branched-stem cell wall muropeptides (14).
The stated need for other ß-lactam and non-ß-lactam compounds active against resistant pneumococcal strains isolated from patients with community-acquired respiratory tract infections is as current as ever (17). Amoxicillin is currently the ß-lactam with the most potent activity against penicillin-susceptible and -resistant pneumococci. In combination with clavulanate, amoxicillin also provides excellent coverage against ß-lactamase-positive Haemophilus influenzae and Moraxella catarrhalis strains (26, 27).
PF-03709270 is a novel oral prodrug of sulopenem, a parenteral penem antibiotic (Fig. 1) (16). The current study tested (i) the activities of penicillin, amoxicillin, amoxicillin-clavulanate, imipenem, meropenem, sulopenem, ceftriaxone, cefuroxime, cefpodoxime, cefdinir, ciprofloxacin, levofloxacin, gatifloxacin, moxifloxacin, azithromycin, and clarithromycin against 297 pneumococci and 45 serotype 19A pneumococci resistant to all available oral agents (9, 40) by the agar dilution MIC determination method; (ii) the activities of amoxicillin, amoxicillin-clavulanate, imipenem, meropenem, sulopenem, ertapenem, ceftriaxone, cefuroxime, cefpodoxime, cefdinir, ciprofloxacin, levofloxacin, gatifloxacin, moxifloxacin, azithromycin, and telithromycin against 12 selected pneumococci by time-kill analysis; (iii) whether alterations in PBPs 2B, 2X, and 1A in combination with mutations in the murM gene and specific patterns of the PBP 2B region (amino acids 557 to 647) correlate with resistance to sulopenem and other ß-lactam antibiotics among selected pneumococcal strains; and (iv) the identities of the pneumococcal clones with the highest sulopenem MICs (1 μg/ml) by multilocus sequence typing (MLST).
Sulopenem structure.
MATERIALS AND METHODS
Two hundred ninety-seven selected pneumococcal strains were tested by the agar dilution MIC method. These comprised 80 penicillin-susceptible strains (MICs, ≤0.06 μg/ml), 88 penicillin-intermediate strains (MICs, 0.125 to 1.0 μg/ml), and 129 penicillin-resistant strains (MICs, 2.0 to 16.0 μg/ml). Of these strains, 152 were macrolide resistant (azithromycin MICs, ≥1.0 μg/ml) and had defined macrolide resistance mechanisms. Additionally, 30 strains with levofloxacin MICs of ≥4.0 μg/ml and defined mutations in type II topoisomerase were tested. Strains were selected to include a majority which had penicillin G MICs of ≥2.0 μg/ml. Some strains were macrolide susceptible but penicillin resistant; this does not reflect the current situation in the United States. We also tested an additional 45 strains of drug-resistant type 19A pneumococci (provided by S. Brown and R. Jones) by the agar dilution MIC method. Among the 12 strains tested by time-kill analysis, 2 were penicillin susceptible, 5 were penicillin intermediate resistant, and 5 were penicillin resistant; 6 strains were macrolide resistant: 3 because of erm(B) mutations, 2 because of mef(A) mutations, and 1 because of an L4 ribosomal protein mutation. Two strains were also quinolone resistant and had defined changes in their quinolone resistance-determining region. Thirty-one isolates were selected from among the 297 strains on the basis of the following criteria: all strains with the highest sulopenem MIC of 1 μg/ml (n = 9), 8 strains with amoxicillin MICs equal to their penicillin G MICs, 8 strains with amoxicillin MICs greater than their penicillin G MICs, and 6 strains with penicillin G MICs greater than their amoxicillin MICs. This collection comprised 28 penicillin-resistant S. pneumoniae strains (MICs, ≥2.0 μg/ml) and 3 penicillin-intermediate resistant S. pneumoniae strains (MICs, 1.0 μg/ml) (see Table 5). Identification of the strains as S. pneumoniae was based on optochin susceptibility, bile solubility, the presence of the lytA gene, and sequencing of the 16S rRNA gene (36). All 31 strains were analyzed by pulsed-field gel electrophoresis (PFGE) to determine clonal relatedness. MLST of all strains (n = 9) with sulopenem MICs of 1 μg/ml and one with an MIC of 0.25 μg/ml was performed as described previously (13).
Fragments of the pbp1A gene (the region from nucleotides 870 to 1950) encoding 350 amino acids, the pbp2B gene (the region from nucleotides 655 to 2028) encoding 458 amino acids, and the pbp2X gene (the region from nucleotides 301 to 2034) encoding 578 amino acids were amplified from chromosomal DNA by PCR with the primers and under the conditions described previously and were directly sequenced (CEQ8000 genetic analysis system; Beckman Coulter, Fullerton, CA) (29). The sequences obtained were analyzed with the BLAST and ClustalW programs (46). The PBP alleles were identified by comparison with the sequences of the analyzed gene from strain R6 and named with a letter of the alphabet on the basis of the degree of homology to the strain R6 sequence; e.g., allele A has the highest degree of homology to the allele in the R6 sequence, and allele L has the lowest degree of homology to the allele in the R6 sequence. The murM gene was amplified from chromosomal DNA by PCR with the primers and under the conditions described previously and was directly sequenced (14). Subsequently, additional primers were designed for amplification of the murein protein variant 149193/50012, as follows: primers mur75up (5′-ATTACAAAGTAGTGCTTGGG) and mur608dn (5′-CGCTTCTCAGTTTTTTTCATCAA) or primers 1491mur135up (5′-CTATGAAGAGGGGAAGTTACTGGCTGTGGCT) and 1491mur549dn (5′-ACCAAATTGAATCTCTACACCCTTA) (43).
Sulopenem susceptibility powder was obtained from Pfizer Central Research, Groton, CT. The other antimicrobials were obtained from their respective manufacturers. The agar dilution method was performed with Mueller-Hinton agar (Becton Dickinson, Sparks, MD) supplemented with 5% sheep blood and inocula of 104 CFU/spot, as previously described by our group (23, 37-39, 44, 45, 48, 49). Clavulanate was combined with amoxicillin in a 1:2 ratio. Standard quality control strains, including S. pneumoniae ATCC 49619, were included in each run of the agar dilution MIC determinations (7).
For the time-kill studies, methods previously described by our group were used (23, 48, 49). The medium was cation-adjusted Mueller-Hinton broth containing 5% lysed horse blood, and the inocula were 5 × 105 to 5 × 106 CFU/ml. The suspensions were incubated at 35°C in a shaking water bath; and viability counts were done after 0, 3, 6, 12, and 24 h. Drugs were considered bactericidal if the original inoculum was reduced by ≥3 log10 CFU/ml (99.9%) at the lowest concentration tested during each of the time periods and bacteriostatic if the inoculum was reduced by 0 to <3 log10 CFU/ml. Drug carryover was addressed by dilution, as reported previously (23, 48, 49).
Nucleotide sequence accession numbers.All sequences of pbp1a, pbp2X, and pbp2B obtained were deposited in the GenBank database and had the following accession numbers: EU863659 to EU863689 for pbp2B, EU863690 to EU863720 for pbp2X, and EU863721 to EU863751 for pbp1A.
RESULTS
The results of the MIC testing of 297 strains can be seen in Table 1. For the 297 penicillin-susceptible, -intermediate, and -resistant pneumococcal strains, sulopenem gave MIC50s of 0.008, 0.06, and 0.25, respectively, and MIC90s of 0.016, 0.25, and 0.5 μg/ml, respectively. The MIC50s of amoxicillin for the corresponding strains were 0.03, 0.25, and 2.0 μg/ml, respectively, and the MIC90s were 0.03, 1.0, and 8.0 μg/ml, respectively. The combination of amoxicillin and clavulanate gave MICs similar to those obtained with amoxicillin alone. The sulopenem MICs were similar to those of imipenem and meropenem. The sulopenem MICs were similar to those of imipenem and meropenem and lower than those of ceftriaxone. All oral cephalosporins tested had MICs higher than those of amoxicillin. The results for the 45 strains of drug-resistant type 19A pneumococci are presented in Table 2. The sulopenem MIC50s and MIC90s for these 45 strains were 1.0 and 2.0 μg/ml, respectively, and the amoxicillin MIC50s and MIC90s for these 45 strains were 8.0 and 8.0 μg/ml, respectively. As can be seen, these strains were highly resistant to all available oral antipneumococcal compounds approved for pediatric use.
Agar dilution MICs for 297 strains
Agar dilution MICs for 45 drug-resistant type 19A pneumococci
The MICs of the 12 strains tested by time-kill analysis are presented in Table 3, and the results of the time-kill analyses are presented in Table 4. As can be seen, all ß-lactams were bactericidal (99.9% killing) against all 12 strains tested at 2× MIC after 24 h and against 8 to 11 strains (the range for the different ß-lactams tested) at 2× MIC after 12 h. All four quinolones were bactericidal at 2× MIC against all 10 strains tested after 24 h; azithromycin was bactericidal against all 8 strains tested, and telithromycin at 2× MIC was bactericidal against 8 of 12 strains after 24 h.
MICs for 12 strains tested by time-kill analysis
Results of killing kinetics
Nine PFGE types (five types represented by 2 closely related strains, three types represented by 3 closely related strains, and one type represented by 6 closely related strains) were found among the 31 isolates characterized by PFGE, and 6 strains had unique patterns (Table 5). Eight sequence types (STs) were found among 10 selected isolates (9 with the highest sulopenem MIC [1 μg/ml] and 1 with the lowest sulopenem MIC [0.125 μg/ml] in the collection of 10 strains) (Table 5).
MICs, PFGE types, and MLST types for strains tested for β-lactam resistancea
The amino acid alterations of the three conserved motifs of PBP 1A, 2X, and 2B are presented in Table 6. During sequence comparisons, we noticed that the sequence numbering for PBP 2B was off by 6 amino acids compared to the original sequence of S. pneumoniae R6 (GenBank accession number AE008520.1) (24).
Amino acid alterations of the three conserved motifs of PBPs 1A, 2X, and 2B in 31 tested strains compared to S. pneumoniae R6 sequence and MurM and PBP allele attributions
Previous reports have described PBP 2B motifs which were off by the 6 amino acids described above, and because of that we have retained the old amino acid numbering to be able to correlate mutations (10, 12, 29, 41).
Analysis of PBP 1A showed that 7 strains (penicillin G MICs, 1 to 4 μg/ml; amoxicillin MICs, 1 to 8 μg/ml) had T371A substitutions in the 370STMK373 motif, 7 strains (penicillin G MICs, 2 to 8 μg/ml; amoxicillin MICs, 0.25 to 4 μg/ml) had T371S substitutions in the 370STMK373 motif, and 14 strains had P432T substitutions (accompanied by T371A in 2 strains and with a T371S mutation in 370STMK373 in all 14 strains) close to the 428SRN430 motif of PBP 1A. PBP 1A was present in 10 different variants (variants A1 to J1) (Table 6).
Ten isolates showed a single PBP 2X mutation, T338A, in the 337STMK340 motif and had penicillin G MICs of 1 to 16 μg/ml and amoxicillin of MICs 1 to 8 μg/ml, and 10 strains showed the T338P substitution and had penicillin G MICs of 1 to 8 μg/ml and amoxicillin MICs of 0.25 to 8 μg/ml. Eight strains had 338TM339 → 338AF339 mutations and had penicillin G MICs of 2 to 8 μg/ml and amoxicillin MICs of 2 to 16 μg/ml. Fourteen of the 31 strains had an L546V substitution close to the 547KSG549 motif of PBP 2X (penicillin G MICs, 1 to 16 μg/ml; amoxicillin MICs, 1 to 8 μg/ml). No change in the 385SVVK388 motif of PBP 2X was noted. PBP 2X showed the presence of 16 different variants (variants A3 to P3) (Table 6).
All strains showed the same substitution T445A in the 442SSNT445 motif in PBP 2B and no change in the 385SVVK388 motif. Sequence analysis of PBP 2B yielded 12 different alleles (alleles A2 to L2) (Table 6). One strain (strain 3374) had an insertion 426YTQ427 (432YTQ433) in PBP 2B. The insertion of three amino acids (425WYT426) has been described previously, and authors have speculated that such alterations may have been responsible for the altered PBP 2B and may have been involved as one of the determinants of penicillin resistance (50).
Six groups were distinguished on the basis of the specific mutation patterns in 90 amino acid fragments from amino acid positions 557 to 647 (Fig. 2).
Comparison of the sequence of PBP 2B (region from amino acid positions 557 to 647 of the deduced amino acid sequence) in all strains tested compared to the PBP 2B sequence from strain R6 of S. pneumoniae.
The murM gene was sequenced from 31 isolates and yielded six different alleles. The deduced amino acid sequences represented six previously published MurM variants: MurMA (14 strains), MurMB6 (5 strains), MurMB1 (5 strains), 149193/5002 (which has 98% homology to allele MurMB1; 3 strains), and MurMB7 and MurMB4 (2 strains each) (Table 6).
DISCUSSION
PF-03709270 is an oral prodrug of sulopenem, an experimental penem (16). In a study of the comparative activities of sulopenem and other drugs against a collection of recently isolated gram-positive and -negative organisms, the sulopenem MIC90s ranged from 0.03 to 1 μg/ml against all clinically significant bacterial species tested. This high in vitro potency was also confirmed by in vitro time-kill studies (25). Our studies confirm the activity of sulopenem against 114 pneumococci of various resistance phenotypes described previously (25) and expand on the information on the activity of sulopenem available.
The results of this study indicate that the combination of amoxicillin and clavulanate did not influence the amoxicillin MICs against all pneumococcal groups, and among all oral ß-lactams tested, amoxicillin had the lowest MIC. Sulopenem, imipenem, meropenem, and ceftriaxone all had low MICs against the pneumococci tested, with the oral cephalosporins having higher MICs. Sulopenem in particular was very active, with the MICs ranging from ≤0.004 to 2.0 μg/ml, even against drug-resistant pneumococci with very high penicillin G and amoxicillin MICs. The MICs of the ß-lactams rose with those of penicillin G, and moxifloxacin was the most active quinolone tested (23, 37-39, 44, 45, 48, 49) Against the drug-resistant type 19A pneumococci, sulopenem had MICs lower than those of amoxicillin whether amoxicillin was tested with or without clavulanate. The MICs of sulopenem for these 45 strains were similar to those of imipenem and meropenem. Amoxicillin, with and without clavulanate, and sulopenem showed excellent killing kinetics relative to their MICs, with bactericidal activity against all 12 strains at 2× MIC being achieved after 24 h. Although azithromycin and telithromycin are commonly described to be bacteriostatic against pneumococci, they are, in reality, slowly bactericidal (as it is defined) at multiples of the MIC after 24 h (23, 48, 49). The clinical significance of this phenomenon awaits clarification.
Among the PBPs produced by S. pneumoniae, PBPs 1A, 2B, and 2X are the most important in the development of resistance. Other mechanisms, such as altered protein kinase CiaH and glycosyltransferase CpoA, have been described as potential mechanisms of resistance; but these aspects were not investigated in the current study (19, 20, 28).
Elevated sulopenem MICs have not previously been associated with any MLST pattern. We found that MLST analysis (10 strains) showed a correlation only between the ST and the country of origin. All defined STs, STs 41, 1094, 663 and 610, and 603 (isolated in the United States and South Africa, Russia, and Poland, respectively), have already been described in the MLST database and have previously been isolated in the countries mentioned above. ST81 represents global S. pneumoniae clone 23F-Spain, which has spread in many countries (33).
Among the PBP alleles analyzed, the most variable was PBP 2X, followed by PBP 2B and PBP 1A, with homologies at the protein level with the sequence of strain R6 of 95 to 87%, 96 to 90%, and 93 to 84%, respectively. A similar correlation has been described by Biçmen et al. (3) and del Campo et al. (10).
Altered stem peptide (mur) genes can also affect the activities of β-lactams (15, 43). Six types of MurM were detected in our study, but no correlation to the increased MICs of any antibiotic tested was detected. Each individual clonal group had the same MurM allele (the exceptions being strains 3455 and 3458). Our studies confirm that the role of murM in resistance cannot be explained at this time as a simple correlation between the presence of an altered MurM protein and resistance to ß-lactams: strains with high-level resistance to penicillin G had unaltered MurM proteins, and also, strains with identical PBP patterns but different MurM alleles had no changes in MICs. In general, our data support the hypothesis by du Plessis et al. that the involvement of MurM in penicillin resistance appears to be dependent on specific mutations in PBPs, especially in PBP 2B (12).
In our study, the region of PBP 2B from amino acids 557 to 647 appeared to be relevant for resistance, especially among strains with high amoxicillin MICs, which confirms previous findings (5, 12, 29).
All mutations previously reported to affect β-lactam resistance (e.g., PBP 1A with a T371A or S substitution within the 370STMK373 motif, a T338S or A substitution in PBP 2X, and a T445A substitution in the 442SSNT445 motif of PBP 2B) were found in all isolates tested in the current study, but we were unable to find a clear correlation between those alterations and specific patterns of resistance, including susceptibility to sulopenem (1, 8, 41, 42). This may suggest that at least one change in crucial PBP motifs may be required to develop β-lactam resistance.
The results of the present study indicate that sulopenem (administered either parenterally or as the oral prodrug) has a potential place in the treatment of infections caused by drug-resistant pneumococci, including the drug-resistant nonvaccine type 19A phenotype (40), which is probably spreading throughout all countries in which the vaccine is available. The only current therapeutic options for the latter strains are linezolid or the respiratory quinolones. The clinical use of sulopenem must await the results of pharmacokinetic/pharmacodynamic analyses, as well as data from toxicological, safety, experimental animal, and clinical efficacy studies.
ACKNOWLEDGMENTS
This study was supported by a grant from Pfizer Central Research, Groton, CT.
We thank S. Brown (Clinical Microbiology Institute, Wilsonville, OR) and R. Jones (JMI Laboratories, North Liberty, IA) for the provision of drug-resistant type 19A strains.
FOOTNOTES
- Received 17 November 2008.
- Returned for modification 9 March 2009.
- Accepted 13 March 2009.
- Copyright © 2009 American Society for Microbiology
