|Home | About | Journals | Submit | Contact Us | Français|
BAL30072 is a new monocyclic β-lactam antibiotic belonging to the sulfactams. Its spectrum of activity against significant Gram-negative pathogens with β-lactam-resistant phenotypes was evaluated and was compared with the activities of reference drugs, including aztreonam, ceftazidime, cefepime, meropenem, imipenem, and piperacillin-tazobactam. BAL30072 showed potent activity against multidrug-resistant (MDR) Pseudomonas aeruginosa and Acinetobacter sp. isolates, including many carbapenem-resistant strains. The MIC90s were 4 μg/ml for MDR Acinetobacter spp. and 8 μg/ml for MDR P. aeruginosa, whereas the MIC90 of meropenem for the same sets of isolates was >32 μg/ml. BAL30072 was bactericidal against both Acinetobacter spp. and P. aeruginosa, even against strains that produced metallo-β-lactamases that conferred resistance to all other β-lactams tested, including aztreonam. It was also active against many species of MDR isolates of the Enterobacteriaceae family, including isolates that had a class A carbapenemase or a metallo-β-lactamase. Unlike other monocyclic β-lactams, BAL30072 was found to trigger the spheroplasting and lysis of Escherichia coli rather than the formation of extensive filaments. The basis for this unusual property is its inhibition of the bifunctional penicillin-binding proteins PBP 1a and PBP 1b, in addition to its high affinity for PBP 3, which is the target of monobactams, such as aztreonam.
There is a paucity of new agents acting against Gram-negative organisms, even though the rate of drug resistance among these bacilli continues to increase and resistance occasionally gives rise to organisms causing infections for which no adequate therapeutic options exist (20, 44). Two recent reports from the Infectious Diseases Society of America have addressed the lack of agents for the treatment of infections caused by three categories of Gram-negative bacilli: extended-spectrum cephalosporin-resistant Escherichia coli and Klebsiella spp., multidrug-resistant Pseudomonas aeruginosa, and carbapenem-resistant Acinetobacter spp. (5, 50).
Strains of Escherichia coli and Klebsiella pneumoniae that have acquired resistance to the extended-spectrum cephalosporins are emerging as threats in hospitals and in the community in many parts of the world (10, 45, 48). The principal mechanism of β-lactam resistance in these two organisms is the expression of extended-spectrum β-lactamases (ESBL) belonging to Ambler class A (43); but other enzymes, such as class B metallo-β-lactamases, class C cephalosporinases, and ESBL variants of class D oxacillinases, may also play a role (11, 34, 54). In addition, both E. coli and K. pneumoniae are known to acquire alterations in the outer membrane proteins that result in limited influx (e.g., by the mutation or deletion of porins) or accelerated efflux (14, 32). These mechanisms of resistance are also affecting the utility of carbapenems. The ability of these organisms to produce class A carbapenemases, in addition to the mechanisms described above, has now become apparent in some regions of the United States, in Israel, and in Greece (31, 38). Acinetobacter baumannii and P. aeruginosa are both intrinsically resistant to a broad variety of antibiotics and have the ability to acquire further mechanisms to the extent that some strains of these pathogens have expressed resistance to all clinically available compounds (4, 15, 22, 29). Resistance to β-lactams in P. aeruginosa is achieved through various combinations of expression of the chromosomally encoded class C β-lactamase; additional enzymes from the other β-lactamase classes; mutation, deletion, or altered expression of the outer membrane proteins involved in the influx of the drug; overexpression of efflux pumps; and mutation of the target penicillin-binding proteins (4, 21, 23, 36, 40, 41, 49). Similar factors play a role in the resistance of Acinetobacter (16, 18), although the variety of β-lactamases implicated in resistance is different. In particular, the class D oxacillinases with carbapenemase activity have acquired particular significance, causing a number of serious outbreaks in Europe and the United States (11).
Monobactams, such as aztreonam (ATM) and carumonam (3, 17, 26, 27), and sulfactams, such as tigemonam (7, 52), make interesting starting points for the development of a new antibiotic targeting Gram-negative bacteria. They have been shown to have potent activity against isolates of the Enterobacteriaceae family and moderate activity against P. aeruginosa (17, 26). In addition, the monocyclic β-lactams are highly resistant to hydrolysis by class B metallo-β-lactamases and are potent inhibitors of class C β-lactamases (3, 27).
Numerous attempts have been made to introduce iron-binding functional groups into β-lactams since the 1980s, in order to circumvent the limitations imposed by porin mutation or deletion (37). BAL30072 is a sulfactam, analogous to tigemonam, with a dihydropyridone iron-chelating group (19, 33). We show here that this combination of features unexpectedly confers markedly different antibiotic properties to BAL30072.
BAL30072 (Fig. (Fig.1)1) was prepared in the research laboratories of Basilea Pharmaceutica International AG, Basel, Switzerland. All other compounds were purchased from commercial sources.
E. coli ATCC 25922 and P. aeruginosa ATCC 27853, routinely used for quality control purposes, were obtained from the American Type Culture Collection. A number of strains with specific β-lactamases were obtained from D. Livermore at the Health Protection Agency (London, United Kingdom) or P. Nordmann at the Hôpital de Bicêtre (Le Kremlin-Bicêtre, France). Otherwise, the isolates used in the test panels were obtained from various hospitals, mainly in Europe, Japan, and the United States, and were collected in more than 25 institutions in at least seven countries over a span of 24 years. They were identified by testing with an API strip (bioMérieux, France) and also, if necessary, by 16S rRNA sequencing and were kept as stock cultures at −70°C or below. The strains comprising the test panels were selected on the basis of their resistotypes, and no isolates from the same source with identical antibiotic resistance patterns were included. Preference was given to strains proven (by sequencing) or suspected (from their antibiotic resistance pattern) to have one or more of class A extended-spectrum β-lactamases, carbapenemases, or inhibitor-resistant variants; class B metallo-β-lactamases; class C cephalosporinases; and class D carbapenemases. The MIC50s and MIC90s of the comparators for these stringent challenge panels may therefore appear to be unusually high.
MICs were determined in duplicate by a broth dilution method in microtiter plates, according to the general recommendation of the CLSI (9), except that Iso-Sensitest broth (Oxoid, Basingstoke, United Kingdom) supplemented with 16 mg/liter 2,2′-bipyridyl (BPL) (hereafter designated IST_BPL medium) to complex iron and thereby induce iron uptake systems was used for all tests. The antibacterial activity of BAL30072 was compared with those of piperacillin-tazobactam (TZP), amoxicillin-clavulanic acid, ticarcillin-clavulanic acid, ATM, cefepime (FEP), ceftazidime (CAZ), imipenem (IPM), and meropenem (MEM). For Acinetobacter spp., the activities of the β-lactam-β-lactamase inhibitor combination ampicillin-sulbactam and polymyxin B were also compared with that of BAL30072, and polymyxin B was also used as a comparator for P. aeruginosa.
Overnight cultures of the test strains were diluted in duplicate into fresh medium to yield inocula of approximately 106 CFU/ml. Drug was added to the inoculum, and incubation at 37°C was initiated. Drug concentrations of 0 (control), 0.5, 1, 2, 4, 8, and 16 times the MIC were used. Samples were taken immediately and at 3, 6, 24, and 28 h after addition of drug. Aliquots of 0.1 ml were plated onto IST_BPL plates, and the numbers of colonies appearing on the plate after 48 h of incubation at 37°C were counted. Each experiment was repeated at least twice, starting from independent inocula.
Overnight cultures of the test strains were diluted into fresh medium to yield inocula of approximately 106 CFU/ml and approximately 109 CFU/ml. Aliquots of 0.1 ml were plated onto IST_BPL plates containing increasing concentrations of the compound to be tested (between 1 and 64 times the MIC). Each experiment was performed in duplicate, the numbers of colonies appearing on the plate after 48 h of incubation at 37°C were counted, and the numbers were averaged. Preliminary counts were also made at 24 h, but many colonies were scarcely visible at that time.
The development of resistance to BAL30072, meropenem, cefepime, and ciprofloxacin was studied in IST_BPL medium by serial passage over increasing concentrations in 0.1 ml in microtiter plates. An initial inoculum of 106 CFU/ml was introduced into the wells of a microtiter plate containing between 0.125× and 2× MIC of the test antibiotic. Aliquots of 0.01 ml were taken from the well with the highest concentration of drug at which significant turbidity appeared (optical density at 600 nm [OD600], 0.1 to 0.5) after 2 days of incubation. Duplicate experiments were performed in parallel, starting from the same initial inoculum. Two days of incubation was chosen on the basis of the population analysis, in which slowly growing colonies took 24 to 48 h to appear. The aliquots were distributed between a new set of graded concentrations of the drug between 0.125× and 2× the concentration in the well from which the inoculum was taken. The serial transfer was continued until the bacteria were growing in at least 64 μg/ml of the antibiotic.
E. coli W3110 was grown in 5-ml cultures at 37°C while the culture was rotated at about 100 rpm. Samples were transferred at intervals into 1.5-ml Eppendorf tubes and centrifuged at 5,000 rpm for 1 min. The supernatant was removed, and the pellet was resuspended in Dulbecco's phosphate-buffered saline (PBS; the pH was adjusted to 7.1 to 7.5) without CaCl2 or MgCl2 (Sigma-Aldrich) at 4°C. The step of centrifugation and resuspension in PBS was repeated twice. The bacteria in PBS were mixed 1:10 with fixative solution containing 1% glutaraldehyde and 1% formaldehyde in water. Afterwards the mixture was transferred onto glass slides and the slides were placed in a humidity chamber for 20 to 30 min. The slides were washed in 500 mM Na2SO3 for 1 min, dehydrated in methanol at −20°C for 1 min, and dried under N2. The slides were examined with an Axioskop 2 plus microscope (Carl Zeiss micro Imaging), and images were processed with an Axio vision viewer (Carl Zeiss micro Imaging). Fixation was used to enable the direct comparison of samples collected at different times under various conditions. It had no impact on the gross morphological changes observable by phase-contrast microscopy.
The AmpC β-lactamases from E. coli SNO3(pAD7) and Citrobacter freundii 1203 and the P. aeruginosa class C β-lactamases from strains 18S/H and 143724R were purified to homogeneity, as described previously (42). TEM-3 β-lactamase was purified from E. coli K802N(pAT251), SHV-2 was purified from E. coli JC2926(pBP60-1), and SHV-4 was purified from E. coli J53-2(pUD21). KPC-2, IMP-1, and VIM-1 were cloned into plasmid vector pPC56 with chloramphenicol selection; and the plasmids were subsequently fully sequenced. The enzymes were expressed from E. coli M15(pREP4) and purified by ion-exchange chromatography (46). The hydrolytic activities toward BAL30072 (Δ310, −1,800 mol−1·cm−1) and aztreonam (Δ318, −650 mol−1·cm−1) were determined spectrophotometrically. Inhibition by BAL30072 was determined with nitrocefin as the reporter substance and by the use of standard methods (42). The β-lactamase activity in the cell lysates was determined with nitrocefin, in which 1 unit was defined as the hydrolysis of 1 μmol nitrocefin per minute. The protein concentration was determined with Bradford protein assay reagents (Bio-Rad, Basel, Switzerland).
Membranes of E. coli were prepared and the affinity for BAL30072 was determined by counterlabeling with fluorescent penicillin (Bocillin), as described by Davies et al. (12).
The iron-chelating agent BPL was added to the growth medium for all tests to decrease the available iron concentration (52). Under iron limitation, the iron-uptake systems of bacteria are induced and siderophores are expressed in large amounts to compete with the bipyridyl for iron (37). The iron-limited growth condition is believed to mimic the in vivo situation, in which iron is bound by a variety of proteins and has previously been used to assess the activities of siderophore antibiotics. In our investigations of siderophore monobactams (13), we found that concentrations of BPL up to 16 μg/ml had no effect on the growth rates of bacteria or on the MICs of nonsiderophore antibiotics (data not shown). The production of pyoverdine (the most readily monitored siderophore) by P. aeruginosa is induced 8- to 15-fold, according to the strain. The activity of the siderophore monobactam PTX 2416 (19) against some strains of P. aeruginosa (about 20% of all strains tested) was enhanced 2- to 16-fold by the addition of BPL, consistent with earlier observations from a study that used conalbumin to chelate iron (33). The MIC of BAL30072 was lowered 2- to 8-fold among 24/48 strains of A. baumannii and 2- to 32-fold among 16/60 strains of P. aeruginosa in the presence of 16 μg/ml BPL. Therefore, the addition of 16 μg/ml BPL to all growth media was adopted as a standard procedure.
The MICs determined for BAL30072 and comparators against these specially selected test panels of resistant organisms are summarized in Table Table11 and Fig. Fig.2.2. BAL30072 had activity similar to that of cefepime against the Enterobacteriaceae, with the MIC50s being 2 and 4 μg/ml, respectively, and 67% of the strains were susceptible to each drug at 8 μg/ml (Fig. (Fig.2).2). Elevated MICs for BAL30072 were observed for Citrobacter freundii and Enterobacter cloacae (MIC90s, 16 and >32 μg/ml, respectively) because a number of strains had ESBLs and some strains hyperexpressed the AmpC β-lactamase. Elevated MICs for BAL30072 and cefepime were observed for Enterobacter aerogenes, E. coli, and K. pneumoniae (MIC90s, 16, >32, and >32 μg/ml, respectively, for both BAL30072 and FEP). BAL30072 retained good activity (MIC90 ≤ 4 μg/ml) against Klebsiella oxytoca, Proteus spp., Providencia spp., and Serratia marcescens, despite β-lactamase expression in some of the strains. The MIC90 of BAL30072 was less than or equal to that of imipenem for the panels of E. aerogenes (16 μg/ml for both drugs), Proteus spp. (0.25 and 8 μg/ml for BAL30072 and IPM, respectively), Providencia spp. (≤0.06 and 16 μg/ml for BAL30072 and IPM, respectively), and Serratia marcescens (4 and 8 μg/ml for BAL30072 and IPM, respectively). Imipenem had better activity than BAL30072 against the C. freundii, E. coli, E. cloacae, and K. pneumoniae panels. BAL30072 had MIC90s equal to or lower than those of meropenem against Proteus spp. (0.25 and 0.5 μg/ml for BAL30072 and MEM, respectively), Providencia spp. (≤0.06 and >32 μg/ml for BAL30072 and MEM, respectively), and S. marcescens (4 μg/ml for both drugs), while meropenem had a lower MIC90 than BAL30072 against C. freundii, E. aerogenes, E. cloacae, E. coli, and K. pneumoniae.
BAL30072 had more potent activity against Acinetobacter spp. (MIC90, 4 μg/ml), Burkholderia spp. (MIC90, 0.125 μg/ml), P. aeruginosa (MIC90, 8 μg/ml), and Stenotrophomonas maltophilia (MIC90, 2 μg/ml) than any of the β-lactam comparators (MIC90 ≥ 32 μg/ml for all organisms except Burkholderia spp.; Table Table1).1). BAL30072 had activity comparable to that of polymyxin B against P. aeruginosa (MIC90s, 8 μg/ml for both drugs) and was more active than this agent against Acinetobacter spp. (MIC90s, 4 μg/ml and 32 μg/ml for BAL30072 and polymyxin B, respectively).
BAL30072 had good activity against most strains for which a metallo-β-lactamase largely contributed to the β-lactam resistance (Table (Table2).2). It was also active against Acinetobacter strains with OXA-27 class D carbapenemase but exhibited elevated MICs (4 to 8 μg/ml) against strains with OXA-25 and OXA-26. BAL30072 was also active against strains with the SME-1 and NMCA enzymes, as well as some strains with KPC-2 class A carbapenemases. The activity against strains with KPC-2 depended on the background in which it was expressed: the differences related to the amount of enzyme expressed and the presence of other β-lactamases. The activity against strains with ESBLs or the AmpC β-lactamase also depended on the amount of activity that was expressed. For example, E. coli K802N(pAT251), which has a multicopy plasmid expressing large amounts of TEM-3 β-lactamase (specific activity of the cell lysate, 103 nitrocefin units/mg cell protein, which is about 10% of the total soluble protein), had a MIC of >32 μg/ml for BAL30072 (as well as for most comparators), whereas E. coli CF102, which is a clinical isolate expressing more usual amounts of TEM-3 β-lactamase (specific activity of cell lysate, 0.5 nitrocefin units/mg cell protein), had an MIC for BAL30072 of 4 μg/ml (Table (Table2).2). Similarly, P. aeruginosa strain 18S/H expresses significantly more class C β-lactamase (specific activity of cell lysate, 550 nitrocefin units/mg cell protein) than strain BA (specific activity of cell lysate, 30 nitrocefin units/mg cell protein) and has correspondingly higher MICs not only for BAL30072 but also for the comparators (Table (Table22).
The activity of BAL30072 against P. aeruginosa was not affected by deletion of the OprD protein, the loss of which selectively raised the MIC of imipenem (Table (Table3)3) but was affected by the level of expression of the MexAB-OprM and MexEF-OprN efflux pumps (Table (Table3).3). Strains PAO238, PAO253, and PAO267, from which MexAB-OprM was deleted (2, 8), had significantly lower MICs for BAL30072 that parent strain PAO1, while strain PAO253, which overexpresses MexEF-OprN in a ΔMexAB-OprM background, had a higher MIC than the other two strains with the MexAB-OprM deletion. The effect of efflux upregulation did not appear to be as severe for BAL30072 as it was for meropenem, as BAL30072 had a median MIC of 4 μg/ml against a subset of 24 P. aeruginosa strains in the test panel that were identified to have an efflux phenotype (taken as a ciprofloxacin MIC of >2 μg/ml that was decreased at least 4-fold by the addition of the efflux inhibitor phenyl-arginine-β-naphthylamide ), but a median MIC of 2 μg/ml for all 206 P. aeruginosa strains. In contrast, the median MIC for meropenem for the same subset of strains was 16 μg/ml, whereas the median MIC of meropenem was 1 μg/ml for all P. aeruginosa strains.
The activity of BAL30072 against P. aeruginosa was not affected by the absence of one or more of the major siderophore receptors for pyoverdine (FpvA and FpvB proteins), pyochelin (FptA protein), or enterobactin (PirA protein) singly or in several combined deletions (Table (Table33).
BAL30072 was hydrolyzed only very slowly, if at all, by class B metallo-β-lactamases (Table (Table4).4). It was an inhibitor of the class C β-lactamase, although it was not as potent as aztreonam, and was only slowly hydrolyzed by these enzymes. BAL30072 was also a poor substrate for a class A carbapenemase (KPC-2) and had a relatively low turnover number and a high Km. It would appear to be a better substrate for ESBLs, as it had a higher apparent affinity for TEM-3 and SHV-4 than it did for the KPC-2 enzyme.
BAL30072 exhibited a high affinity for E. coli PBP 3 in membrane fragments (Table (Table4),4), comparable to that of aztreonam for this enzyme. Significant inhibition of PBP 1a and PBP 1b was also observed at relatively low concentrations (50% inhibitory concentrations [IC50s], 3.9 and 1.9 μM, respectively), and the inhibition of PBP 2 was detectable (IC50, 30 μM), but not in a range that would be relevant for growth inhibition (the MIC of the E. coli strain used for membrane preparation was 0.25 μg/ml). Aztreonam did not inhibit fluorescent penicillin binding to PBPs 1a, 1b, and 2 up to the highest concentration tested (50 μM).
BAL30072 achieved a 3-log10-unit reduction in the colony count within 24 h, without regrowth, for susceptible strains (MICs < 8 μg/ml) of Acinetobacter, E. cloacae, and S. maltophilia when the inoculum was 105 to 106 CFU/ml (Table (Table5).5). The minimum bactericidal concentrations (MBCs) for these strains were one to eight times the MICs. The MBCs for the P. aeruginosa strains were four to eight times the MICs. The killing curves typically exhibited an initial rapid decrease of 2 to 3 log10 units within 6 h and a slower phase that usually resulted in sterilization of the culture within 24 h (Fig. (Fig.3).3). The killing kinetics against susceptible strains were similar to those observed with meropenem and ceftazidime (Fig. (Fig.3).3). Monitoring of growth turbidimetrically through the apparent change in the optical density at 600 nm showed a rapid onset of lysis and clearance of the suspension after the addition of BAL30072 to a growing culture of E. coli (Fig. (Fig.4a),4a), whereas the turbidity continued to increase after the addition of aztreonam (Fig. (Fig.4b).4b). Examination by phase-contrast microscopy of samples taken from the cultures revealed that a large number of spheroplasts and a number of short filaments 5 to 10 μm long were formed in the culture treated with BAL30072, whereas extremely long filaments up to 200 μm in length were formed after exposure to aztreonam (Fig. (Fig.5).5). The effects of BAL30072 were somewhat concentration dependent: more short filaments were formed at lower concentrations than at high concentrations. Long filaments were formed after exposure to aztreonam at all effective concentrations.
The number of colonies recovered after exposure to BAL30072 decreased markedly at concentrations greater than or equal to the MIC for the antibiotic. Higher concentrations of antibiotic were required when 109 CFU rather than 107 CFU was used, possibly reflecting the marked inoculum effect observed with β-lactam antibiotics against many of the strains chosen for evaluation in this study. The numbers of CFU isolated on plates generally decreased to zero at concentrations >4× MIC (Table (Table6).6). No resistant colonies were isolated from A. baumannii HPA74510, Acinetobacter junii 15218, any of the strains of the Enterobacteriaceae, or P. aeruginosa BA. The P. aeruginosa cultures yielded more colonies with elevated MICs, especially when the high inoculum was used. The incidence of such colonies was generally less than that observed with the comparators (e.g., meropenem; Table Table6).6). Typically, the mutants grew very slowly; e.g., the doubling time of P. aeruginosa BA during exponential growth, measured turbidimetrically, was 1.5 ± 0.3 h, whereas the doubling times were 3.8 ± 0.2 h for BA(C1-106) and 4.1 ± 0.3 h for BA(C2-107). Some mutants had elevated levels of expression of β-lactamase, resulting in cross-resistance to ceftazidime and aztreonam [e.g., BA(C1-106) in Table Table3],3], while others did not [e.g., BA(C2-107)]. The underlying mechanism of resistance in the last two strains is under investigation. No hypersensitivity to ciprofloxacin, which might be expected if the strains had been selected for tonB deletions (33, 34), was observed, and the MICs of the mutants for ciprofloxacin were generally unchanged from that of the parent.
BAL30072 alone showed a low propensity to select for resistance in A. junii, E. cloacae S4741B (which has inducible AmpC production), and K. pneumoniae 1338665 (Table (Table7).7). The MIC of BAL30072 rose more quickly in the cultures of E. cloacae S9639, which has constitutive AmpC production and an MIC that was already 4 μg/ml. MICs greater than 4× the original MIC were reached within the first five passages in the cultures of S. maltophilia and two of the P. aeruginosa strains (both of which expressed low levels of chromosomal class C β-lactamase). The rate of resistance development in P. aeruginosa was rather less than that observed with the comparators and did not reach such high levels on extended passaging (Fig. (Fig.6).6). The mutants isolated after serial passage had phenotypes similar to those obtained in the population analysis, exhibiting reduced growth rates and the overexpression of β-lactamase.
There is growing concern over multidrug-resistant strains of E. coli, Klebsiella, P. aeruginosa, and Acinetobacter that have acquired various combinations of β-lactamase, restricted outer membrane permeability, efflux, and target modification mechanisms of resistance (5, 53). In the search for new drugs with activity against Gram-negative pathogens, it is desirable to circumvent as many of these mechanisms of resistance as possible. BAL30072 combines features from earlier attempts to obtain good activity against resistant Gram-negative bacteria (17, 19, 51) that result in a surprisingly different mechanism of action for this compound.
BAL30072 was active against 70% of the carbapenem-resistant Enterobacteriaceae strains tested in this study, including strains with class A, class B, and class D carbapenemases. It was less active against strains with high-level expression of class A and D ESBLs and class C cephalosporinases, and the activity of BAL30072 is therefore complementary to the activities of the carbapenems such as meropenem against Enterobacteriaceae. BAL30072 was more active against the Gram-negative nonfermenters, especially Acinetobacter spp., Burkholderia spp., and Stenotrophomonas maltophilia. It was the most active β-lactam against the multidrug-resistant P. aeruginosa strain in our test panel.
BAL30072 was expected to be stable against the metallo-β-lactamases of Ambler class B because of its similarity to aztreonam, which is known to be stable toward these enzymes (53). Indeed, we were unable to demonstrate the hydrolysis of BAL30072 by the VIM or IMP β-lactamase, with which a slow reaction of aztreonam was detectable. BAL30072 retained good activity against strains of the Enterobacteriaceae and P. aeruginosa with metallo-β-lactamases that were not susceptible to aztreonam. BAL30072 is an inhibitor of class C β-lactamase, like aztreonam, forming a stable acyl-enzyme complex that is only slowly hydrolyzed (25). This is a common property of monobactams with an (S)-methyl group (as in aztreonam) or a bulky (R)-substituent (as in carumonam) at the C-4 position of the azetidinone ring (22, 25, 51) and is due to interference of the substituent with residues in the active site that prevents a rotation about the C-3—C-4 bond necessary for access of the water molecule that would hydrolyze the acyl-enzyme complex (25). BAL30072 appears to be a less potent inhibitor of class C β-lactamase than aztreonam, which may work in its favor in derepressed strains that produce large amounts of the enzyme, against which BAL30072 is still active but aztreonam is not. Because of the lower affinity, less of the antibiotic is bound to the enzyme and more is free to react with the penicillin-binding proteins. BAL30072 was hydrolyzed by class A and class D enzymes, especially ESBL variants. Nevertheless, it appeared to be a poor substrate for a number of ESBLs, such as TEM-3 and CTX-M 15, as well as carbapenemases from these classes, including KPC-2. The low hydrolysis rates again appeared to allow enough BAL30072 to escape to attack the penicillin-binding proteins, and activity against clinical isolates with such enzymes could be demonstrated for BAL30072 but not for aztreonam (or meropenem, when a carbapenemase was present). On the other hand, BAL30072 appeared to be a better substrate for some SHV derivatives, and its activity was compromised by even small amounts of these enzymes.
BAL30072 has a dihydropyridone siderophore moiety in its side chain which is believed to facilitate uptake by allowing additional uptake routes to be exploited (33, 37). Whether this is actually true for BAL30072 is currently under investigation, but one can envisage that accelerated influx could enhance the activity of BAL30072 against strains producing β-lactamases for which BAL30072 is a poor substrate. Earlier siderophore monobactams, such as pirazmonam and U78,608 (1, 6), have been shown to select for resistant mutants in single-step or passage experiments (30, 39, 47). The mutations that appeared were in siderophore receptors and the TonB protein, which serves to energize the iron-siderophore uptake systems. The latter mutants have a characteristic hypersensitivity to antibiotics such as ciprofloxacin that are efflux substrates, because TonB also energizes some efflux pumps (54). BAL30072 does not readily select for resistant mutants, and so far, all those that we have examined have elevated levels of β-lactamase expression and little or no change in susceptibility to ciprofloxacin. No mutant with an acquired deficiency in tonB or a siderophore receptor has been identified. The activity of BAL30072 against tonB mutants of strains with defined efflux backgrounds is in progress.
BAL30072 inhibited E. coli PBP 1a and PBP 1b, in addition to having a high affinity for PBP 3; this contrasts with the activity of aztreonam, which only inhibits E. coli PBP 3. As can be expected from its different penicillin-binding protein inhibition profile, the effect of BAL30072 on cell morphology was significantly different from that observed with aztreonam, which promotes the formation of very long filaments. Thus, when it is given at concentrations below or close to the MIC, BAL30072 promoted spheroplasting and the formation of short filamentous structures equivalent in length to 4 to 12 cells. When it was given at concentrations above the MIC, only the spheroplasts and incompletely divided structures (equivalent to two to four cells in length) were observed. This behavior is more reminiscent of the effects of bicyclic β-lactams such as imipenem, which causes spheroplasting at all concentrations, and cefotaxime, which causes filamentation and spheroplasting at low concentrations and spheroplasting at high concentrations (24). It is very probable that this novel bactericidal effect of BAL30072 contributes to its potent activity and the low propensity of strains to develop resistance to the drug.
We thank Caroline Müller and Beatrice Hofer for their expert technical assistance. We also thank Jutta Heim for enthusiastic discussion and support. Our thanks go to D. Livermore, P. Nordmann, and H. Schweizer for providing some of the strains used in this study.
Published ahead of print on 22 March 2010.