Skip to main page content

• HOME
• Current Issue
• Archive
• Alerts
• About ASM
• Contact us
• Tech Support
• Journals.ASM.org

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Vaara, M. Articles by Vaara, T. Search for Related Content PubMed PubMed Citation Articles by Vaara, M. Articles by Vaara, T.

 Previous Article  |  Next Article 

Antimicrobial Agents and Chemotherapy, September 2008, p. 3229-3236, Vol. 52, No. 9
0066-4804/08/$08.00+0     doi:10.1128/AAC.00405-08
Copyright © 2008, American Society for Microbiology. All Rights Reserved.

Novel Polymyxin Derivatives Carrying Only Three Positive Charges Are Effective Antibacterial Agents

Martti Vaara,^1,2* John Fox,³ Günther Loidl,⁴ Osmo Siikanen,⁵ Juha Apajalahti,⁵ Frank Hansen,⁶ Niels Frimodt-Møller,⁶ Junya Nagai,⁷ Mikihisa Takano,⁷ and Timo Vaara¹

Northern Antibiotics Ltd., FI-00720 Helsinki,¹ Division of Clinical Microbiology, Helsinki University Hospital, FI-00029 HUSLAB, Helsinki,² Alimetrics Ltd., FI-02920 Espoo, Finland,⁵ Alta Bioscience, University of Birmingham, Birmingham B15 2TT, United Kingdom,³ Bachem AG, CH-4416 Bubendorf, Switzerland,⁴ Statens Serum Institut, Copenhagen S, DK-2300, Denmark,⁶ Hiroshima University, Graduate School of Biomedical Sciences, Hiroshima 734-8553, Japan⁷

Received 25 March 2008/ Returned for modification 25 May 2008/ Accepted 22 June 2008

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
The lack of novel antibiotics against gram-negative bacteria has reinstated polymyxins as the drugs of last resort to treat serious infections caused by extremely multiresistant gram-negative organisms. However, polymyxins are nephrotoxic, and this feature may complicate therapy or even require its discontinuation. Like that of aminoglycosides, the nephrotoxicity of polymyxins might be related to the highly cationic nature of the molecule. Colistin and polymyxin B carry five positive charges. Here we show that novel polymyxin derivatives carrying only three positive charges are effective antibacterial agents. NAB739 has a cyclic peptide portion identical to that of polymyxin B, but in the linear portion of the peptide, it carries the threonyl-D-serinyl residue (no cationic charges) instead of the diaminobutyryl-threonyl-diaminobutyryl residue (two cationic charges). The MICs of NAB739 for 17 strains of Escherichia coli were identical, or very close, to those of polymyxin B. Furthermore, NAB739 was effective against other polymyxin-susceptible strains of Enterobacteriaceae and against Acinetobacter baumannii. At subinhibitory concentrations, it dramatically sensitized A. baumannii to low concentrations of antibiotics such as rifampin, clarithromycin, vancomycin, fusidic acid, and meropenem. NAB739 methanesulfonate was a prodrug analogous to colistin methanesulfonate. NAB740 was the most active derivative against Pseudomonas aeruginosa. NAB7061 (linear portion of the peptide, threonyl-aminobutyryl) lacked direct antibacterial activity but sensitized the targets to hydrophobic antibiotics by factors up to 2,000. The affinities of the NAB compounds for isolated rat kidney brush border membrane were significantly lower than that of polymyxin B.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Whereas pharmaceutical companies have in recent years developed a considerable number of new drugs against multiresistant gram-positive bacteria, virtually no novel drugs, with the exception of tigecycline, are suitable for systemic therapy of infections caused by gram-negative bacteria (9, 14, 22, 27). Simultaneously, these bacteria are progressively becoming more and more multiresistant (extremely multiresistant [XMR]), or even panresistant, to the clinically available antibacterial agents (7, 14, 21, 27).

Today, major therapeutic problems are caused by carbapenemase-producing XMR strains of Klebsiella pneumoniae. The other notorious bacteria include the XMR strains of Pseudomonas aeruginosa and Acinetobacter baumannii. However, in the coming years, a far more serious threat will be the emergence of carbapenemase-producing XMR strains of Escherichia coli, a very common opportunistic pathogen. Accordingly, the need for novel agents against gram-negative bacteria is serious. The lack of such agents would result in a literal return to the pre-antibiotic era (23).

The current crisis has reinstated polymyxins as the drugs of last resort for treatment of serious infections caused by gram-negative XMR strains (4, 11, 12, 25, 35). Colistin (polymyxin E) and polymyxin B are cyclic lipodecapeptides, each carrying five free amino groups and, accordingly, five positive charges under physiological conditions (24, 26). Their first target in susceptible gram-negative bacteria is the outer membrane (OM), where these amino groups bind to the acidic lipopolysaccharide (LPS) molecules (18, 30). Polymyxins were discovered as early as 1947 but were largely abandoned owing to their nephrotoxicity and neurotoxicity and the discovery of better-tolerated drugs. Even though the toxicity of polymyxins may not be as high as that observed in the different settings of the past, it might still complicate the patient's therapy or even require its discontinuation (4, 5, 11, 12, 25, 35).

Numerous efforts have been made to render polymyxin less toxic. In colistin methanesulfonate (CMS), the free amino groups are blocked by sulfomethylation to yield an uncharged prodrug. CMS hydrolyzes in vivo as well as in aqueous solutions in vitro to liberate the compound with antibacterial activity (6, 10). In rats, an intravenous bolus of 3 mg of polymyxin B per kg of body weight (3) or 3 mg of colistin per kg (11) caused acute signs of toxicity (dyspnea, piloerection, and decreased movement), while no such signs were observed after an intravenous bolus of 15 mg of CMS per kg (11). CMS is currently used in systemic therapy. However, the nephrotoxicity of CMS does not appear to be lower than that of the unsulfomethylated form of polymyxin B currently used in systemic therapy (5). Furthermore, the degree of sulfomethylation differs between different CMS preparations (6, 10, 11).

Another approach to reducing toxicity has been to treat polymyxin with ficin or related enzymes that remove the lipid moiety (the 6-methyloctanoyl or 6-methylheptanoyl residue). The N-terminal amino acyl residue is lost as well. The resultant polymyxin nonapeptides (colistin nonapeptide and polymyxin B nonapeptide [PMBN]) have less acute toxicity than their parent polymyxins (3, 30) but are devoid of antibacterial activity. However, PMBN and colistin nonapeptide are very effective in synergistic combination with those antibacterial agents that are effectively excluded by the intact OM (8, 20, 30, 32, 33). Unfortunately, PMBN retains the nephrotoxicity of polymyxin B (30). It should be noted that PMBN and colistin nonapeptide carry five positive charges.

In this study, we evaluated the structure-function relationships and antibacterial properties of novel synthetic polymyxin derivatives that differ from all natural polymyxins by possessing no more than three positive charges under physiological conditions. Such derivatives can be expected to have toxicological and pharmacokinetic properties different from those of natural polymyxins.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Peptide synthesis. Polymyxin derivatives were synthesized by conventional solid-phase chemistry using the standard Fmoc (fluorenylmethoxy carbonyl) protection strategy. The -amino function was protected by Fmoc, and Fmoc was removed by 20% piperidine in dimethylformamide. The -amino group of the diaminobutyryl (Dab) residue involved in cyclization of the peptide was protected by t-butoxycarbonyl, an acid-labile group that was removed at the cleavage step. The functional group of asparagine was protected by tritylation. All the other amino acids with functional side chain groups were protected by benzyloxycarbonyl (Z), a group that is stable in the acid cleavage stage but can easily be removed after the cyclization reaction.

The synthesis employed O-(6-chlorobenzotriazol-1-yl)-N,N,N',N'-tetramethyluronium hexafluorophosphate (HCTU) or O-(benzotriazol-1-yl)-N,N,N',N'-tetramethyluronium tetrafluoroborate (TBTU) as an activator, and the steps were performed in a commercial automated synthesizer. The amino acids were purchased already protected from a standard supplier. Acylation was performed for 30 min by using a fourfold molar excess of each amino acid or the fatty acid, a fourfold molar excess of the activator HCTU or TBTU, and an eightfold molar excess of N-methyl morpholine.

The peptide was removed from the resin by reaction with a solution of 95% trifluoroacetic acid and 5% water for 2 h at room temperature, to yield the partially protected product. The resulting peptide was precipitated with diethyl ether.

The cyclization mixture comprised benzotriazol-1-yl-oxy-tris-pyrrolidino-phosphonium hexafluorophosphate (PyBop), N-hydroxybenzotriazole, and N-methyl morpholine at molar excesses of 2, 2, and 4, respectively. The cyclization mixture was added to the peptide (dissolved in dimethylformamide) and allowed to react for 2 h. The cyclized, protected peptide was precipitated by the addition of cold diethyl ether. Any residual PyBop was removed by washing the peptide with water. The remaining side chain protection groups (Z) were removed by catalytic dehydrogenation (i.e., by subjecting the peptide, dissolved in acetic acid-methanol-water [5:4:1], to an atmosphere of hydrogen in the presence of a palladium charcoal catalyst).

The deprotected and cyclized peptides were purified by reversed-phase chromatography using conventional gradients of acetonitrile-water-trifluoroacetic acid. The product was dried by lyophilization. The purity, as estimated by reversed-phase high-performance liquid chromatography, was more than 95%.

Preparation of NAB739 methanesulfonate sodium salt. NAB739 methanesulfonate was synthesized using essentially the method described previously (34) for the synthesis of polymyxin B methanesulfonate. First a neutral formaldehyde solution (400 µl of 30% aqueous formaldehyde [brought to pH 7.2 with 1 N NaHCO₃]) and then a 1 N NaHCO₃ solution (2 ml) were added to NAB739 acetate (100 mg dissolved in 2 ml of water). The precipitated NAB739 formaldehyde derivative was filtered and washed with water. The moist solid was suspended in water (5 ml), and sodium metabisulfite (100 mg) was added. A clear solution was obtained after a few minutes and was freeze-dried. The flocculent white solid was extracted with warm acetone (7.5 ml) and dried in vacuo. The yield was 78 mg. Analysis of the product by electrospray ionization mass spectrometry revealed a predominant peak with a molecular mass of 1,075.3 Da, indicating that most of the derivative was sulfomethylated at each of the three Dab residues of the NAB739 compound. A minor peak representing NAB739 blocked randomly at two of the three Dab residues was also visible.

Microbial strains. E. coli IH3080 (K1:O18) is an encapsulated strain originally isolated from a neonate suffering from meningitis (33). Other E. coli strains included ATCC 25922; three extended-spectrum β-lactamase-producing strains (NCTC13351, NCTC13352, and NCTC13353) from the National Collection of Type Cultures, Colindale, United Kingdom; seven strains isolated from blood cultures in Sweden (from the Culture Collection, University of Gothenburg, Sweden [CCUG]); and five strains isolated from blood cultures at Helsinki University Hospital, Helsinki, Finland, in 2006 to 2007 (the F strains). Other microbial strains were from the ATCC (11 strains), the CCUG (3 strains), and Helsinki University Hospital (8 strains). E. coli SC 9252 and E. coli SC 9253 are well-characterized polymyxin-resistant pmrA mutants of E. coli K-12 (19, 26).

Antibacterial assays. To screen for both the direct antibacterial activities of the compounds and their abilities to sensitize a set of target bacteria to the hydrophobic model antibiotic (rifampin), the simple agar well diffusion method was used. LB agar plates (LB Agar Lennox; Difco, BD, Sparks, MD) and LB plates containing increasing concentrations (0.1 µg/ml, 0.3 µg/ml, and 1.0 µg/ml) of rifampin (Sigma-Aldrich, St. Louis, MO) were used. The plates were inoculated with bacterial suspensions in 0.9% NaCl (a 0.5 McFarland standard), according to CLSI standard M2-A9 (1) to produce even, confluent growth. Thereafter, small wells were cut (diameter, 2 mm; five wells per plate). Four quantities (0.4 µg, 1 µg, 4 µg, and 10 µg) of each compound dissolved in 0.9% NaCl were tested. Controls included 0.9% NaCl alone, as well as solutions of peptides with previously determined activities, such as polymyxin B sulfate (P0972; Sigma-Aldrich) and PMBN (Sigma-Aldrich). After an 18-h incubation at 37°C, the diameters of the growth inhibition zones were measured, and the corresponding surface areas (in square millimeters) were calculated.

MICs were determined using the agar dilution method according to CLSI protocol M7-A7 (2). Mueller-Hinton agar (Difco 225220; BD, Sparks, MD) plates containing increasing concentrations (0.25 to 32 µg/ml in twofold increments) of the test substances were used. The plates were inoculated (1 µl per spot, corresponding to approximately 10⁴ CFU/spot) using a multipoint inoculator and were incubated for 16 to 20 h at 35°C.

The synergism of the compounds with other antibiotics was studied using E-strips (Biodisk Ltd., Solna, Sweden) on plates containing Mueller-Hinton agar (product no. LabO39; LabM Ltd., Bury, Lancashire, United Kingdom) with increasing concentrations of the compound under study (for NAB739, 0.125 µg/ml to 2 µg/ml in twofold increments; for NAB7061, 1 µg/ml to 16 µg/ml in twofold increments) as well as on Mueller-Hinton agar without the compounds. The inoculum was prepared according to the manufacturer's instructions and corresponded to that in CLSI standard M2-A9 (1). When the synergistic effects of NAB7061 on a large number of bacterial strains were studied, only one concentration of NAB7061 was used (4 µg/ml). Synergism with rifampin (Sigma-Aldrich) was also studied in microtiter plates using the checkerboard method, cation-adjusted Mueller-Hinton II broth (Difco 212322), and an inoculum size of 5 x 10⁵ CFU/ml according to CLSI standard M7-A7 (2).

Affinities of the compounds to the BBM of the renal cortex. The binding of the compounds to the brush border membrane (BBM) of the renal cortex was measured indirectly by their abilities to inhibit the binding of radiolabeled gentamicin to the BBM. BBMs were isolated from the renal cortices of male albino rats by using the Mg²⁺-EGTA precipitation technique as described previously (22). The binding of gentamicin was measured, according to the method described by Nagai et al. (16), by incubating BBM vesicles (20 µl) in 10 mM HEPES (pH 7.5) buffer with 300 mM mannitol in the presence of 20 µM [³H]gentamicin (American Radiolabeled Chemicals, Inc., St. Louis, MO) with or without the compound to be tested or a positive control. After incubation for 60 min at 4°C, 1 ml of this buffer, ice-cold, was added, and the mixture was filtered through a Millipore filter (pore size, 0.45 µm; HAWP). The filter was washed with the buffer, and the radioactivity remaining in the filter was measured using a liquid scintillation counter. Fifty percent inhibitory concentrations (IC₅₀s) were determined as described previously (16) using the Hill equation.

Cytotoxicity assay on V79 cells. Cytotoxicity assays were performed using the Chinese hamster lung fibroblast cell line V79 (European Collection of Cell Cultures, Salisbury, United Kingdom). The cells were grown in Dulbecco's modified Eagle medium supplemented with 1% (vol/vol) antibiotic-antimycotic solution (10,000 U/ml penicillin, 10 µg/ml streptomycin, and 25 µg/ml amphotericin B), glutamine (2 mM), and heat-inactivated fetal bovine serum (10%), all from Sigma-Aldrich. The culture (2 to 3 days old; more than 50% confluent) was trypsinized, and a cell suspension in fresh medium was made and used to seed 60-mm-diameter petri dishes containing 5 ml of culture medium with approximately 200 cells each. Three replicate dishes were prepared for each treatment. After the dishes were incubated for 24 h at 37°C under a humidified atmosphere containing approximately 5% CO₂, the culture medium was replaced with fresh culture medium containing the appropriate concentrations (0.25 µg/ml to 128 µg/ml in twofold increments) of the test item or the positive control (phenol). After the 24-h treatment period under the growth conditions specified above, cultures were washed three times, covered with fresh medium, left to form colonies for 3 days, fixed with methanol, and stained with 5% Giemsa solution; then the colonies were counted.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Structure-function relationships of the polymyxin derivatives. To reduce the number of cationic charges, we removed both cationic Dab residues from the linear portion of the polymyxin molecule. NAB7061 lacks the N-terminal Dab of all polymyxins and carries the neutral aminobutyryl (Abu) residue instead of Dab at the C terminus of the linear portion of the peptide (Table 1). Furthermore, NAB7061 carries octanoyl residues as the fatty acyl residues, instead of a mixture of methyloctanoyl and methylheptanoyl residues (present in polymyxin B). Its cyclic portion is identical to that of polymyxin B.

View this table: [in this window] [in a new window]

TABLE 1. Antibacterial activities of the NAB compounds against E coli IH3080

As the test organism, we used the virulent encapsulated E. coli strain IH3080 (O18:K1), the same strain systematically used in our earlier studies. The direct antibacterial activity of NAB7061 was very weak, but the compound remarkably enhanced the activity of rifampin, indicating that it damages the bacterial OM permeability barrier in the same way as PMBN, the best-characterized OM permeability-increasing agent (30).

Essentially similar results were obtained with NAB718 and NAB719. The cyclic portions of these peptides are identical to those of colistin (polymyxin E) and polymyxin T, respectively.

Substituting a neutral amino acyl residue (Abu) for one of the cationic Dab residues in the cyclic portion of the peptide had little effect on the permeability-increasing activity when this change was compensated for by the introduction of one cationic residue in the linear portion of the peptide, as in compounds NAB715, NAB716, and NAB717. Without this compensation, the compound was virtually inert, as shown for NAB708. Substituting two Abu residues for the corresponding Dab residues in the cyclic peptide resulted in inactive compounds, even when the loss of positive charges in the cyclic portion of the peptide was compensated for by the introduction of two Dab residues in the linear portion of the peptide, as in NAB725 and NAB726.

In very sharp contrast to NAB7061 and all the other permeability-increasing polymyxin peptides that possess Abu at the C terminus of the linear portion of the peptide, NAB734, NAB737, and NAB739 all turned out to be very effective as direct antibacterial agents (Table 1). Their activities appeared to be superior even to that of polymyxin B. The linear peptide portions of NAB734 and NAB739 consist of Thr-D-Ser, while that of NAB737 consists of Thr-D-Thr. All these residues carry hydroxyl groups. The cyclic portion of NAB739 is identical to that of polymyxin B, while the cyclic portions of NAB734 and NAB737 are identical to that of polymyxin S.

Compared with NAB739, compounds NAB748 and NAB749, with linear portions Thr-Thr-D-Ser and Ala-Thr-D-Ser, respectively, were as active; NAB743 (with Thr-Ser) was somewhat less active; and compounds with Thr-Asn (NAB746), D-Ala-D-Ala (NAB733), or Thr-D-Ala (NAB738) were much less active as direct antibacterial agents.

Also a direct antibacterial agent, NAB740 differed from NAB739 by only two additional C atoms in its fatty acyl chain.

Similar findings were obtained in screening assays that employed A. baumannii ATCC 19606. Against this strain, NAB7061, NAB733, NAB738, and NAB746 had direct antibacterial activities (expressed as surface areas of growth inhibition) of 0 to 20 mm² only, whereas NAB739, polymyxin B, and NAB734 produced inhibition areas of 169 ± 26, 142 ± 26, and 87 ± 11 mm², respectively (averages ± standard deviations [SD] from two to five independent experiments). Furthermore, very similar differences in activities against E. coli IH3080 were found in screening assays that employed smaller amounts of the same compounds (0.4 µg, 1 µg, and 4 µg per well).

To show that the direct antibacterial activity of NAB739 is due to the presence of free amino groups in the cyclic portion of the peptide, we synthesized NAB739 methanesulfonate, a compound in which the free amino groups are sulfomethylated to yield a labile compound that slowly decomposes to free NAB739 (the nascent compound). In analogy to colistin methanesulfonate (see the introduction), a freshly made aqueous solution of NAB739 methanesulfonate was found to be only weakly bactericidal. When tested at a concentration of 2 µg/ml, it reduced the CFU of E. coli IH3080 only by approximately one-third, while a 2-day-old solution of NAB739 methanesulfonate or the control (free NAB739) reduced the CFU by 90% or 98%, respectively.

Comparison of the properties of NAB739 and NAB7061. NAB739 differs from NAB7061 only in possessing D-Ser instead of Abu in the C terminus of the linear portion of the peptide. The remarkable effect of this difference on antibacterial activity was verified by determining the MICs of these two compounds against E. coli ATCC 25922. Two methods were used: dilution in Mueller-Hinton agar and microdilution in Mueller-Hinton II broth (Table 2). In addition, to assess the observed synergism with rifampin in more detail, the fractional inhibitory concentration indices (FICI) of both compounds with rifampin were determined using two methods: the Etest on Mueller-Hinton agar containing increasing concentrations of the compound and the checkerboard method in Mueller-Hinton II broth. The MIC of NAB739 alone ranged from 1 µg/ml to 2 µg/ml. No measurable synergism with rifampin was found by either of the two tests. On the other hand, the MIC of NAB7061 ranged from 16 µg/ml to >16 µg/ml, and very strong synergism with rifampin was observed, with FICI as low as 0.071 to 0.078 by Etest and 0.128 by the checkerboard method. Accordingly, the alteration of a single amino acyl residue results in a very significant difference in biological activity.

View this table: [in this window] [in a new window]

TABLE 2. MICs of rifampin for E. coli ATCC 25922 in the presence of increasing concentrations of NAB739 or NAB7061a

MICs of NAB739, NAB740, and NAB7061. The MICs of NAB739 against E. coli were very close to those of polymyxin B (Table 3). For clinical isolates of E. coli (15 strains), both the MIC at which 50% of isolates were inhibited (MIC₅₀) and the MIC₉₀ of NAB739 and polymyxin B were 1 µg/ml. For Klebsiella spp., Enterobacter cloacae, and Citrobacter freundii (11 strains altogether), the MICs of NAB739 ranged from 1 µg/ml to 2 µg/ml, whereas for K. pneumoniae CCUG45421, the MIC was 4 µg/ml.

View this table: [in this window] [in a new window]

TABLE 3. MICs (µg/ml) of selected NAB peptides and polymyxin B

NAB740 was almost as active as NAB739 against Enterobacteriaceae, but for P. aeruginosa and A. baumannii, its MICs (4 µg/ml and 2 to 4 µg/ml, respectively) were slightly lower than those of NAB739. The MICs of NAB7061 were very high for all the strains studied. Bacteria that are known to be resistant to polymyxin B were resistant to NAB739 and NAB740, as well. These included Serratia marcescens, the polymyxin-resistant pmrA mutants of E. coli (19, 26), and microbes that do not belong to the gram-negative bacteria (Staphylococcus aureus, Candida albicans).

Effects of subinhibitory concentrations of NAB739 on the susceptibilities of A. baumannii and P. aeruginosa to antibacterial agents. The MICs of five antibacterial agents were determined for two strains of A. baumannii (ATCC 19606 and F264) and for P. aeruginosa ATCC 27853 by the Etest method on Mueller-Hinton agar containing increasing concentrations of NAB739 (Table 4).

View this table: [in this window] [in a new window]

TABLE 4. MICs of five antibiotics for A. baumannii and P. aeruginosa in the presence of NAB739a

Interestingly, NAB739, at a concentration as low as 0.5 µg/ml, could sensitize both A. baumannii strains to rifampin (sensitization factors, 85 to 128), clarithomycin (sensitization factors, 12 to 16), fusidic acid (sensitization factors, 64 to 85), and vancomycin (sensitization factor, 85). The MICs of vancomycin for both strains were as low as 3 µg/ml in the presence of NAB739 at 0.5 µg/ml. The FICI were very low. Strain F264 was resistant to meropenem, but in the presence of NAB739 (0.5 µg/ml) the MIC of meropenem was as low as 1 µg/ml. These findings were quite unexpected in light of the results obtained with E. coli strains under identical test conditions (Table 2; also data not shown). NAB739 had also some synergism (FICI, 0.46) with clarithromycin against P. aeruginosa.

Abilities of NAB7061 to sensitize E. coli, Klebsiella pneumoniae, and Enterobacter cloacae to antibacterial agents. MICs of a representative set of clinically used antimicrobial agents were determined for two strains of E. coli (ATCC 25922 and IH3080), as well as for K. pneumoniae ATCC 13883 and E. cloacae ATCC 23355, by using Etests on Mueller-Hinton agar containing increasing concentrations of NAB7061. Under the conditions used, no direct growth-inhibitory effect of NAB7061 on E. coli IH3080 or K. pneumoniae ATCC 13883 was detectable, even at 16 µg of NAB7061/ml, whereas 16 µg/ml and 8 µg/ml inhibited the growth of E. coli ATCC 25922 and E. cloacae ATCC 23355, respectively.

NAB7061 exerted strong synergism with rifampin, clarithromycin, azithromycin, erythromycin, and mupirocin against all four strains tested (Table 5). The FICI of the combination of NAB7061 and rifampin ranged from 0.073 to 0.260, and that of NAB7061 and clarithromycin ranged from 0.105 to 0.273. At a concentration of 4 µg/ml, NAB7061 sensitized the strains to rifampin by factors ranging from 170 to 1,500. Extremely high sensitization factors were also observed for clarithromycin (63 to 380), mupirocin (24 to 512), azithromycin (31 to 64), and erythromycin (21 to 48).

View this table: [in this window] [in a new window]

TABLE 5. MICs of rifampin, three macrolides, and mupirocin for four bacterial strains in the presence of increasing concentrations of NAB7061a

Low FICI, determined with E. coli ATCC 25922, were also observed for quinupristin-dalfopristin (0.250), fusidic acid (0.270), linezolid (0.344), clindamycin (0.417), and vancomycin (0.438), although the resulting MICs were reasonably low only in the case of quinupristin-dalfopristin (1.5 to 3 µg/ml). All these antibacterial agents are notably hydrophobic or large (vancomycin) molecules and are effectively excluded by the intact OMs of gram-negative bacteria.

No sensitization (sensitization factor, 2; determined using E. coli ATCC 25922 and NAB7061 at a concentration of 4 µg/ml) to piperacillin, ceftazidime, cefotaxime, levofloxacin, ciprofloxacin, meropenem, or tobramycin was found. All these agents are hydrophilic or relatively hydrophilic, and the intact OM is not an effective permeability barrier against them. Furthermore, no sensitization was found to the large lipoglycopeptide daptomycin or to tigecycline, a molecule that probably penetrates the OM rather easily, as do the tetracyclines.

Susceptibilities of other strains of gram-negative bacteria to NAB7061. The MICs of rifampin and clarithromycin for a representative set of gram-negative bacterial strains, other than those listed in Table 5, were determined in the presence of 4 µg/ml NAB7061. For all strains of E. coli, Klebsiella oxytoca, E. cloacae, and C. freundii (15 strains altogether), the MICs of rifampin were as low as 0.125 µg/ml in the presence of NAB7061, and the sensitization factors ranged from 85 to 2,000 (Table 6). Similar results were obtained with clarithromycin; for 13 out of 15 strains, the MICs of clarithromycin were as low as 0.25 µg/ml, and for all 15 strains, the sensitization factors ranged from 90 to 1,000. Strains of K. pneumoniae remained somewhat more resistant to both antibiotics, and the sensitization factors ranged from 10 to 500.

View this table: [in this window] [in a new window]

TABLE 6. Susceptibilities of gram-negative bacteria to rifampin and clarithromycin in the presence of NAB7061a

For the three strains of A. baumannii, the sensitization factors ranged from 21 to 125, and the resulting MICs were quite low (rifampin MIC, 0.125 µg/ml; clarithromycin MIC, 0.5 µg/ml).

As expected, polymyxin-resistant species, i.e., S. marcescens, Proteus mirabilis, and Proteus vulgaris, were resistant to the sensitizing action of NAB7061. Furthermore, NAB7061 did not sensitize P. aeruginosa.

Affinities of NAB739, NAB740, NAB7061, and polymyxin B for the BBM of the renal cortex. The binding of the compounds to the isolated renal BBM was measured indirectly by determining their abilities to inhibit the binding of radiolabeled gentamicin to the BBM. The IC₅₀s for the compounds studied were as follows (averages ± SD for three parallel determinations, if not otherwise stated): 39.3 ± 5.5 µM (two independent experiments, each with three replicates) for polymyxin B, 90.2 ± 9.7 µM for unlabeled gentamicin, 187.3 ± 24.3 µM (two independent experiments, each with three replicates) for NAB7061, 198 ± 49 µM for NAB740, and 264 ± 91 µM for NAB739. The differences between the IC₅₀ of polymyxin B and those of each of the NAB compounds were statistically significant (P < 0.05). Accordingly, it can be concluded that the affinity of NAB739 for the BBM is only approximately one-sixth or one-seventh that of polymyxin B and approximately one-third that of gentamicin. Similarly, the affinity of NAB7061 for the BBM is only one-fifth that of polymyxin B.

Cytotoxicity. NAB7061, NAB739, and NAB740 (all at 0.25 µg/ml to 128 µg/ml) showed no cytotoxicity against V79 Chinese hamster lung fibroblast cells (treatment time, 24 h).

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
The known difficulty in developing antibacterial agents against species of Enterobacteriaceae, P. aeruginosa, A. baumannii, and many other gram-negative bacteria is caused by the OMs of these bacteria. The OM is an effective permeability barrier to hydrophobic antibiotics as well as to large hydrophilic antibiotics that easily enter gram-positive bacteria (18). In addition, effective efflux mechanisms that pump noxious substances out of the cell act synergistically with the OM permeability barrier (13). More than 95% of novel antibacterial agents isolated from nature or synthesized in the laboratory lack activity against E. coli but have activity against gram-positive bacteria, according to an analysis of agents reported from 1981 to 1992 (31).

However, the OM can also be regarded as the Achilles' heel of gram-negative bacteria and an attractive target for antibiotic action. Polymyxins and many of their derivatives, as well as numerous other compounds that carry several positive charges, bind to the acidic LPS molecules on the outer surface of the OM, disorganize the highly ordered structure of the LPS layer, and destroy the permeability barrier function of the OM (30).

The action of polymyxins is rather specific. Typically, the MICs for gram-positive bacteria are high. The enantiomeric mirror analogue of PMBN (D-PMBN) completely lacks the antibacterial properties of PMBN, such as the ability to act synergistically with hydrophobic antibiotics (29). This further substantiates our original view that the antibacterial activity of polymyxins is determined not only by their cationic nature but also by proper conformation (30). Furthermore, all 11 synthetic analogues of PMBN, each having a rather subtle modification, were shown to be much less potent than PMBN (28).

In contrast to polymyxins, many other cationic peptides lack specificity. They are not only antibacterial but also active against eukaryotic microbes and mammalian cells. They are also easily inactivated by serum proteases.

In the present work, we found that several polymyxin derivatives are directly antibacterial even though they carry only three positive charges, all of them in the cyclic portion of the peptide. The MICs of NAB739 for E. coli strains are identical or almost identical to those of polymyxin B. NAB739 shows good activity against other polymyxin-susceptible strains of Enterobacteriaceae, as well. The presence of two hydroxyl groups in the linear portion of the peptide is very advantageous.

NAB739 alone was active against A. baumannii, but at low subinhibitory concentrations it also had remarkable synergism with several other antibiotics, including clarithromycin, rifampin, and vancomycin, against A. baumannii. A combination might be advantageous in preventing the development of resistance.

The methanesulfonate derivative of NAB739 might have some advantages over NAB739. For instance, in analogy to CMS, acute toxicity might be lower for the methanesulfonate derivative than for NAB739 itself.

NAB740 differs from NAB739 in carrying a decanoyl residue instead of an octanoyl residue as the fatty acyl moiety and by being somewhat more active against P. aeruginosa. NAB7061 lacks direct antibacterial activity but has strong synergism with several antibiotics against polymyxin-susceptible strains of Enterobacteriaceae and A. baumannii.

The cyclic portions of the NAB739, NAB740, and NAB7061 molecules are identical to that of polymyxin B. Hence, they could be made semisynthetically by using polymyxin B as the starting material. In addition, the cyclic portions of other naturally occurring polymyxins can be used, as shown in Table 1 for colistin, polymyxin S, and polymyxin T.

Polymyxins as well as other highly cationic agents release histamine from mast cells. This can complicate therapy for cystic fibrosis patients, who receive aerosolized polymyxin. Whether the histamine-releasing activity of the novel compounds is lower than that of polymyxins should be evaluated.

The nephrotoxicity of aminoglycosides, another group of antibiotics that carry several positive charges under physiological conditions, is manifested in the renal proximal tubules. Aminoglycosides are taken up by the epithelial cells of the proximal tubules and stay there for a long time (17). The binding site for aminoglycosides is believed to be acidic phospholipids in the BBM of the proximal tubular cells. In addition, megalin, a giant endocytic receptor abundantly present in the apical membrane of proximal tubules, plays an important role in the binding and endocytosis of aminoglycosides (17). Polymyxin B binds to megalin (15). Accordingly, the mechanism of nephrotoxicity of polymyxins might be related to that of aminoglycosides. In this paper we showed that reducing the number of positive charges from five to three decreases the affinity of the compound for isolated rat kidney BBMs by a factor of 6 to 7.

Whether the NAB compounds with three positive charges are less nephrotoxic than polymyxin B and colistin remains to be seen. We are currently investigating the pharmacokinetic and toxicological properties of these compounds.

 
J.F. thanks Sat Sandhu, who performed all of the milligram-level peptide synthesis chemistry. M.V., O.S., and J.A. thank Sini Virtanen for excellent technical assistance.

This study was funded by Northern Antibiotics Ltd.

 
* Corresponding author. Mailing address: Northern Antibiotics Ltd., Eskolantie 1, FI-00720 Helsinki, Finland. Phone: 358-50-355 0822. Fax: 358-9-6842 0130. E-mail: martti.vaara{at}northernantibiotics.com

Published ahead of print on 30 June 2008.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

1
1. Clinical and Laboratory Standards Institute. 2006. M2-A9: performance standards for antimicrobial disk susceptibility tests. Approved standard, 9th ed. CLSI, Wayne, PA.
2
2. Clinical and Laboratory Standards Institute. 2006. M7-A7: methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. Approved standard, 7th ed. CLSI, Wayne, PA.
3
3. Danner, R. L., K. A. Joiner, M. Rubin, W. H. Paterson, N. Johnson, K. M. Ayers, and J. E. Parrillo. 1989. Purification, toxicity, and antiendotoxin activity of polymyxin B nonapeptide. Antimicrob. Agents Chemother. 33:1428-1434.[Abstract/Free Full Text]
4
4. Falagas, M. E., and S. K. Kasiakou. 2005. Colistin: the revival of polymyxins for the management of multidrug-resistant Gram-negative bacterial infections. Clin. Infect. Dis. 40:1333-1341.[CrossRef][Medline]
5
5. Falagas, M. E., and S. K. Kasiakou. 2006. Toxicity of polymyxins: a systematic review of the evidence from old and recent studies. Crit. Care 10:R27.[CrossRef][Medline]
6
6. Garrod, L. P., H. P. Lambert, and F. O'Grady. 1973. Antibiotic and chemotherapy, 4th ed. Churchill Livingstone, London, United Kingdom.
7
7. Giske, C. G., D. L. Monnet, O. Cars, and Y. Carmeli on behalf of ReAct-Action on Antibiotic Resistance. 2008. Clinical and economic impact of common multidrug-resistant gram-negative bacilli. Antimicrob. Agents Chemother. 52:813-821.[Free Full Text]
8
8. Ito-Kagawa, M., and Y. Koyama. 1984. Studies on the selectivity of action of colistin, colistin nonapeptide and colistin heptapeptide on the cell envelope of Escherichia coli. J. Antibiot. 37:926-928.[Medline]
9
9. Levy, S. B., and B. Marshall. 2004. Antibacterial resistance worldwide: causes, challenges and responses. Nat. Med. 10(Suppl):S122-S129.[CrossRef][Medline]
10
10. Li, J., R. W. Milne, R. L. Nation, J. D. Turnidge, and K. Coulthard. 2003. Stability of colistin and colistin methanesulfonate in aqueous media and plasma as determined by high-performance liquid chromatography. Antimicrob. Agents Chemother. 47:1364-1370.[Abstract/Free Full Text]
11
11. Li, J., R. L. Nation, R. W. Milne, J. D. Turnidge, and K. Coulthard. 2005. Evaluation of colistin as an agent against multi-resistant Gram-negative bacteria. Int. J. Antimicrob. Agents 25:11-25.[CrossRef][Medline]
12
12. Li, J., R. L. Nation, J. D. Turnidge, R. W. Milne, K. Coulthard, C. R. Rayner, and D. L. Paterson. 2006. Colistin: the re-emerging antibiotic for multidrug-resistant Gram-negative bacterial infections. Lancet Infect. Dis. 6:589-601.[CrossRef][Medline]
13
13. Li, X. Z., and H. Nikaido. 2004. Efflux-mediated drug resistance in bacteria. Drugs 64:159-204.[CrossRef][Medline]
14
14. Livermore, D. M. 2007. Introduction: the challenge of multiresistance. Int. J. Antimicrob. Agents. 29(Suppl. 3):S1-S7.[Medline]
15
15. Moestrup, S. K., S. Cui, H. Vorum, C. Bregengård, S. E. Bjorn, K. Norris, J. Glieman, and E. I. Christensen. 1995. Evidence that epithelial glycoprotein 330/megalin mediates uptake of polybasic drugs. J. Clin. Investig. 96:1404-1413.[Medline]
16
16. Nagai, J., M. Saito, Y. Adachi, R. Yumoto, and M. Takano. 2006. Inhibition of gentamycin binding to rat renal brush-border membrane by megalin ligands and basic peptides. J. Control Release 112:43-50.[CrossRef][Medline]
17
17. Nagai, J., and M. Takano. 2004. Molecular aspects of renal handling of aminoglycosides and strategies for preventing the nephrotoxicity. Drug Metab. Pharmacokinet. 19:159-170.[CrossRef][Medline]
18
18. Nikaido, H. 2003. Molecular basis of bacterial outer membrane revisited. Microbiol. Mol. Biol. Rev. 67:593-656.[Abstract/Free Full Text]
19
19. Nummila, K., I. Kilpeläinen, U. Zähringer, M. Vaara, and I. Helander. 1995. Lipopolysaccharides of polymyxin B-resistant mutants of Escherichia coli are extensively substituted by 2-aminoethyl pyrophosphate and contain aminoarabinose in lipid A. Mol. Microbiol. 16:271-278.[CrossRef][Medline]
20
20. Ofek, I., S. Cohen, R. Rahmani, K. Kabha, D. Tamarkin, Y. Herzig, and E. Rubinstein. 1994. Antibacterial synergism of polymyxin B nonapeptide and hydrophobic antibiotics in experimental gram-negative infections in mice. Antimicrob. Agents Chemother. 38:374-377.[Abstract/Free Full Text]
21
21. Perez, F., A. M. Hujer, K. M. Hujer, B. K. Decker, P. N. Rather, and R. A. Bonomo. 2007. Global challenge of multidrug-resistant Acinetobacter baumannii. Antimicrob. Agents Chemother. 51:3471-3484.[Free Full Text]
22
22. Projan, S. J., and P. A. Bradford. 2007. Late stage antibacterial drugs in the clinical pipeline. Curr. Opin. Microbiol. 10:441-446.[CrossRef][Medline]
23
23. Spellberg, B., R. Guidos, D. Gilbert, J. Bradley, H. W. Boucher, W. M. Scheld, J. G. Bartlett, and J. Edwards, Jr., for the Infectious Diseases Society of America. 2008. The epidemic of antibiotic-resistant infections: a call to action for the medical community from the Infectious Diseases Society of America. Clin. Infect. Dis. 46:155-164.[CrossRef][Medline]
24
24. Srinivasa, B. R., and L. K. Ramachandran. 1979. The polymyxins. J. Sci. Ind. Res. 38:695-709.
25
25. Stein, A., and D. Raoult. 2002. Colistin: an antibiotic for the 21st century? Clin. Infect. Dis. 35:901-902.[CrossRef][Medline]
26
26. Storm, D. R., K. S. Rosenthal, and P. E. Swanson. 1977. Polymyxin and related peptide antibiotics. Annu. Rev. Biochem. 46:723-763.[CrossRef][Medline]
27
27. Talbot, G. H., J. Bradley, J. E. Edwards, Jr., D. Gilbert, M. Scheld, and J. G. Bartlett. 2006. Bad bugs need drugs: an update on the development pipeline from the Antimicrobial Availability Task Force of the Infectious Diseases Society of America. Clin. Infect. Dis. 42:657-668.[CrossRef][Medline]
28
28. Tsubery, H., I. Ofek, S. Cohen, and M. Fridkin. 2000. Structure-function studies of polymyxin B nonapeptide: implications to sensitization of gram-negative bacteria. J. Med. Chem. 43:3085-3092.[CrossRef][Medline]
29
29. Tsubery, H., I. Ofek, S. Cohen, and M. Fridkin. 2000. The functional association of polymyxin B with bacterial lipopolysaccharide is stereospecific: studies on polymyxin B nonapeptide. Biochemistry 39:11838-11844.
30
30. Vaara, M. 1992. Agents that increase the permeability of the outer membrane. Microbiol. Rev. 56:395-411.[Abstract/Free Full Text]
31
31. Vaara, M. 1993. Antibiotic-supersusceptible mutants of Escherichia coli and Salmonella typhimurium. Antimicrob. Agents Chemother. 37:2255-2260.[Free Full Text]
32
32. Vaara, M., and T. Vaara. 1983. Polycations sensitize enteric bacteria to antibiotics. Antimicrob. Agents Chemother. 24:107-113.[Abstract/Free Full Text]
33
33. Vaara, M., and T. Vaara. 1983. Sensitization of Gram-negative bacteria to antibiotics and complement by a nontoxic oligopeptide. Nature (London) 303:526-528.[CrossRef][Medline]
34
34. Wilkinson, E. S. July 1962. Parenteral administration of methanesulphonate of polymyxin A, B, or E. U.S. patent 3,044,934.
35
35. Zavascki, A. P., L. Z. Goldani, J. Li, and R. L. Nation. 2007. Polymyxin B for the treatment of multidrug-resistant pathogens: a critical review. J. Antimicrob. Chemother. 60:1206-1215.[Abstract/Free Full Text]

---

Antimicrobial Agents and Chemotherapy, September 2008, p. 3229-3236, Vol. 52, No. 9
0066-4804/08/$08.00+0     doi:10.1128/AAC.00405-08
Copyright © 2008, American Society for Microbiology. All Rights Reserved.

This article has been cited by other articles:

• Choi, S.-K., Park, S.-Y., Kim, R., Kim, S.-B., Lee, C.-H., Kim, J. F., Park, S.-H. (2009). Identification of a Polymyxin Synthetase Gene Cluster of Paenibacillus polymyxa and Heterologous Expression of the Gene in Bacillus subtilis. J. Bacteriol. 191: 3350-3358 [Abstract] [Full Text]  

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Vaara, M. Articles by Vaara, T. Search for Related Content PubMed PubMed Citation Articles by Vaara, M. Articles by Vaara, T.