Potentiation of Antibiotic Activity by a Novel Cationic Peptide: Potency and Spectrum of Activity of SPR741

The emergence of Gram-negative bacteria with increasingly multidrug-resistant (MDR) profiles has become a serious public health concern worldwide (1), and the rapidity with which resistance determinants disseminate globally (2), along with the increasing frequency of reports of clinical isolates that are resistant to carbapenems (3), has led to the reintroduction of older, potentially more toxic antibiotics such as colistin (4, 5) into clinical practice. Despite the fact that no antibiotic class has been spared by the evolution and dissemination of resistance, no totally new class of antibiotics active against Gram-negative bacteria has been introduced for over 40 years due to the manifold challenges of antibacterial drug discovery (6–8). One of the most significant obstacles to developing agents targeting Gram-negative bacteria is the bacterial outer membrane (9), and one novel therapeutic approach to solve the current problem of limited effective treatment options against Gram-negative bacteria is to combine existing antibiotics with a “potentiator” molecule that permeabilizes the outer membrane, granting antibiotics that would otherwise be ineffective access to their intracellular targets (10, 11). SPR741, formerly NAB741 (12), is a cationic peptide derived from polymyxin B and is one such potentiator molecule that is under development for the treatment of serious Gram-negative bacterial infections. Unlike polymyxin B, SPR741 exhibits minimal intrinsic Gram-negative antibacterial activity (12). It retains the cyclic peptide portion of polymyxin B but differs in two important areas: first, SPR741 lacks two cationic diaminobutyryl residues from the linear portion of the peptide, reducing the net positive charge of the molecule to +3, compared to the net charge of +5 in polymyxin B, and second, SPR741 lacks the 6-methyloctanoyl or 6-methylheptanoyl fatty acid tail found in polymyxin B (12). Despite these changes, SPR741 retains the ability to permeabilize the outer membrane of Gram-negative bacteria, thus sensitizing them to hydrophobic antibiotics (12). Importantly, these structural changes also significantly improve the safety profile of SPR741 compared to that of polymyxin B, which suffers severe, dose-limiting nephrotoxicity in humans. In multiday rodent and nonhuman primate studies, polymyxin B exhibited nephrotoxicity at a far-lower exposure-normalized dose than SPR741 (T. Lister, unpublished data).

In the present study, we determine the ability of SPR741 to potentiate the activity of a broad range of conventional antibiotics from multiple classes against strains of Escherichia coli, Klebsiella pneumoniae, and Acinetobacter baumannii and demonstrate that specific combinations with SPR741 are efficacious against both clinical isolates and MDR reference collection isolates of these bacteria.

SPR741 (Fig. 1) was previously shown to be synergistic with rifampin and clarithromycin (12). In the present work, we sought to understand the extent to which SPR741 was able to potentiate the activity of antibiotics more broadly against E. coli, K. pneumoniae, and A. baumannii. To this end, 35 antibiotics from 27 different classes, with diverse mechanisms of action, were tested in combination with SPR741 by using a checkerboard methodology. For each combination, the results were assessed in several ways: first, fractional inhibitory concentration indices (FICIs) were calculated to determine whether there was a synergistic interaction; second, the fold reduction in the MIC (“potentiation factor”) (Table 1) in the presence of either 2, 4, or 8 μg/ml (E. coli) or 4, 8, or 16 μg/ml (K. pneumoniae and A. baumannii) SPR741 was calculated, relative to the MIC of the antibiotic alone; and third, the possible relevance of the MIC of the antibiotic in the presence of SPR741 was inferred by determining whether the MIC had been reduced to a level at or below the clinical breakpoint for susceptibility for each organism-antibiotic combination (13) or, where this information was not available due to the normal spectrum of activity of the antibiotic (i.e., minimal Gram-negative antibacterial activity), below the equivalent breakpoint for Staphylococcus spp. (13).

In combinations with SPR741, the MICs of 13 antibiotics (azithromycin, clarithromycin, erythromycin, fidaxomicin, fosfomycin, fusidic acid, mupirocin, novobiocin, quinupristin-dalfopristin, ramoplanin, retapamulin, rifampin, and telithromycin) against E. coli ATCC 25922 were reduced at least 32-fold (Fig. 2A). In line with previously reported results (12), in combination with SPR741, the MIC of rifampin was substantially reduced; in the present work, the MIC was reduced >8,000-fold to 0.002 μg/ml in the presence of 8 μg/ml SPR741 (Fig. 2A and Table 1), substantially below the Clinical and Laboratory Standards Institute (CLSI) breakpoint for susceptibility of 1 μg/ml for Staphylococcus spp. (13). Large reductions in MICs were also observed for the macrolide antibiotics erythromycin (1,024-fold, to 0.03 μg/ml) and clarithromycin (4,096-fold, to 0.016 μg/ml) in the presence of up to 8 μg/ml SPR741 (Fig. 2A and Table 1). Although neither antibiotic would typically be considered for the treatment of Gram-negative bacterial infections, these MIC values are well below the CLSI breakpoints for susceptibility of Staphylococcus spp. to these agents (13). The ketolide antibiotic telithromycin, which also targets protein synthesis, is typically used to treat respiratory pathogens (14) but is less active against members of the enterobacteriaceae and Acinetobacter spp. (15). In combination with SPR741, however, the MIC of telithromycin against E. coli ATCC 25922 was reduced up to 256-fold (Fig. 2A) to 0.06 μg/ml, indicating that the spectrum of activity of the antibiotic can be widened when used in combination with SPR741.

Other agents whose activities were substantially enhanced in the presence of SPR741 included retapamulin and fusidic acid (Fig. 2A). Retapamulin is a semisynthetic pleuromutilin antibiotic that targets the 50S subunit of the bacterial ribosome and is typically active against Gram-positive bacteria (16). In combination with SPR741, the MIC of retapamulin was reduced by 256-fold (Fig. 2A) to 0.03 μg/ml. Fusidic acid is an inhibitor of protein synthesis and typically regarded as a relatively narrow-spectrum antibiotic used for the treatment of skin and soft tissue infections and osteomyelitis caused by staphylococci (17). Fusidic acid alone lacked activity against E. coli ATCC 25922 (MIC > 128 μg/ml), but in combination with 8 μg/ml SPR741, the MIC was reduced by 512-fold (Fig. 2A) to 0.5 μg/ml, again demonstrating the ability of SPR741 to extend the spectrum of activity of an antibiotic when the two agents are used in combination.

In addition to the impermeability of the outer membrane, another contributor to the intrinsic resistance of Gram-negative bacteria to several of the antibiotics tested in the present work is efflux, which prevents the intracellular accumulation of antibiotics (18). Clarithromycin is subject to efflux by the AcrAB-TolC efflux pump in E. coli (19); a tolC mutant, which is deficient in this efflux system, displayed increased susceptibility to clarithromycin relative to its isogenic parent strain (19) (Table 2). Because the disruption of tolC results in pleiotropic effects (20), we also included an acrA mutant, which produced similar results (Table 2). The combination of clarithromycin and 1 μg/ml SPR741 reduced the MIC of the antibiotic against parent strain BW25113 to levels similar to those observed for the tolC and acrA mutants (Table 2); a similar reduction was observed with the clarithromycin-SPR741 combination against E. coli clinical strain ATCC 25922 (Table 2). Fusidic acid, mupirocin, and retapamulin are all substrates of AcrAB-TolC (21, 22) (Table 2), and substantial reductions in the MICs of all three antibiotics against the wild-type strains were also achieved by combining them with SPR741 (Table 2). The lack of intrinsic activity of the oxazolidinone antibiotic linezolid against E. coli is also due to efflux by AcrAB-TolC, but in this case, the outer membrane does not contribute to intrinsic resistance, since permeabilization by a polymyxin B nonapeptide does not reduce the MIC of this antibiotic against E. coli (21, 23). In agreement with these findings, there was little potentiation of the activity of linezolid when it was combined with SPR741 (Table 2 and Fig. 2A). Taken together, these data suggest that the intrinsic resistance of E. coli to certain antibiotics that is mediated by both the outer membrane and the AcrAB-TolC efflux system can be overcome, or circumvented, by combining the antibiotic with SPR741 but that potentiation of intrinsic resistance due primarily to efflux may be limited.

Against K. pneumoniae, the extent of the fold reduction in the MICs of antibiotics achieved by combining them with SPR741 was generally lower than that observed against E. coli ATCC 25922 (Fig. 2A and B), with a maximum reduction of 128-fold observed for rifampin, clarithromycin, and retapamulin (Fig. 2B). Indeed, SPR741 was itself intrinsically less active against K. pneumoniae ATCC 43816 (modal MIC of >128 μg/ml) than against E. coli ATCC 25922 (modal MIC of 16 μg/ml). The MICs of 10 antibiotics (azithromycin, clarithromycin, erythromycin, fusidic acid, mupirocin, novobiocin, retapamulin, rifampin, telithromycin, and vancomycin) were reduced at least 32-fold in combination with SPR741. The MICs of clarithromycin and rifampin were reduced to below their respective CLSI breakpoints for susceptibility against Staphylococcus spp. (13) (Table 1). The intrinsic MIC of azithromycin against K. pneumoniae ATCC 43816 was relatively low, at 2 μg/ml, although the MIC was reduced even further in the presence of SPR741 (Fig. 2B and Table 1). The MICs of both erythromycin and telithromycin were reduced to 2 μg/ml in the presence of 16 μg/ml SPR741. Similar to the results with E. coli ATCC 25922 (Fig. 2A), the MIC of retapamulin against K. pneumoniae ATCC 43816 was substantially reduced when combined with SPR741 (Fig. 2B).

A. baumannii NCTC 12156 was also less susceptible to SPR741 (modal MIC of 64 μg/ml) than was E. coli ATCC 25922. The MICs of eight antibiotics (clarithromycin, erythromycin, fusidic acid, quinupristin-dalfopristin, ramoplanin, retapamulin, rifampin, and teicoplanin) were reduced ≥32-fold against this strain (Fig. 2C). Of these antibiotics, the most well-potentiated antibiotics (MIC reductions of 128-fold relative to the MIC of the antibiotic alone) were clarithromycin, erythromycin, fusidic acid, and rifampin. A. baumannii NCTC 12156 was considerably more susceptible to rifampin (MIC of 0.5 μg/ml) than was either E. coli ATCC 25922 or K. pneumoniae ATCC 43816 (MIC of 16 μg/ml), although the MIC was reduced even further when rifampin was combined with SPR741 (Fig. 2C). Similarly, this strain of A. baumannii was substantially more susceptible to fusidic acid (MIC of 2 μg/ml) than was either of the other test strains (MIC > 128 μg/ml), but again, it was possible to reduce this value further in combination with SPR741 (Fig. 2C). The MICs of both clarithromycin and erythromycin were reduced to 0.125 μg/ml, below the CLSI breakpoints for susceptibility of Staphylococcus spp. to these agents (13).

The activities of several antimicrobial agents in combination with SPR741 were not consistently enhanced against all three species. For example, the activities of the aminoglycoside antibiotics gentamicin and tobramycin were not increased against E. coli or K. pneumoniae; the efficacy of the quinolone antibiotics ciprofloxacin and levofloxacin were enhanced against E. coli only (Fig. 2A to C). Simply sharing a mechanism of action or cellular target was therefore not predictive of the ability of SPR741 to potentiate the activity of a given antibiotic. Agents active against pathways involved in cell wall synthesis were potentiated to different extents: for example, dalbavancin lacked activity against E. coli ATCC 25922 and K. pneumoniae ATCC 43816 irrespective of the addition of SPR741; the activity of ramoplanin, however, was substantially enhanced against E. coli ATCC 25922, with a 128-fold reduction in the MIC from 512 to 4 μg/ml. Interestingly, while there was little evidence of a potentiation of dalbavancin against these strains, when combined with SPR741, the activities of the glycopeptide antibiotics vancomycin and teicoplanin and the lipoglycopeptide antibiotic ramoplanin against E. coli and A. baumannii were all enhanced by between 8- and 128-fold (Fig. 2A and C). However, these reductions were not sufficient to reduce the MICs of vancomycin and teicoplanin below the breakpoint for susceptibility against Staphylococcus spp.

Taken together, these data show that SPR741 can potentiate the activities of antibiotics with diverse physicochemical properties and mechanisms of action that are normally ineffective against Gram-negative bacteria. In combination with SPR741, the MICs of clarithromycin, erythromycin, fusidic acid, retapamulin, and rifampin were reduced by at least 32-fold, relative to the MICs of the antibiotics alone, against E. coli ATCC 25922, K. pneumoniae ATCC 43816, and A. baumannii NCTC 12156. Combination with SPR741 also resulted in substantial (≥32-fold) reductions in the MICs of azithromycin, mupirocin, and telithromycin against E. coli ATCC 25922 and K. pneumoniae ATCC 43816. SPR741 is hypothesized to increase the efficacy of antibiotics by permeabilizing the bacterial outer membrane, thereby facilitating the access of the antibiotic to its target; it is not expected that SPR741 would in any way modify the activity of the antibiotic or its interaction with the target. Therefore, point mutations that modify the target would be expected to render bacteria resistant to a given antibiotic irrespective of whether access to that target was enhanced by the presence of a potentiating molecule such as SPR741. Indeed, strains of E. coli bearing defined mutations in rpoB that confer resistance to rifampin were found not to be susceptible to the SPR741-rifampin combination (data not shown).

To determine whether these combinations would be more broadly useful against the three organisms for which a single example strain had already been tested, and in order to rank the combinations, the MICs of azithromycin, clarithromycin, fusidic acid, mupirocin, retapamulin, and rifampin against a panel of 25 E. coli, 25 K. pneumoniae, and 17 A. baumannii strains were determined in the presence and absence of SPR741 (Table 3). Meropenem and aztreonam were also included as examples of antibiotics whose activity was increased to a lesser extent by SPR741 (Fig. 2). MIC₉₀ values were calculated for each panel-antibiotic combination in the absence or presence of 2, 4, or 8 μg/ml SPR741 (A. baumannii and K. pneumoniae) or 2 or 4 μg/ml SPR741 (E. coli; 8 μg/ml SPR741 alone was sufficient to inhibit the growth of 90% of this panel of E. coli strains).

By far the most effective combination against the E. coli panel was SPR741-rifampin: the MIC₉₀ of rifampin was reduced by at least 4,096-fold, from >128 μg/ml in the absence of SPR741 to 0.06 μg/ml in the presence of 2 μg/ml SPR741 (Table 3). In agreement with the above-described results, against E. coli ATCC 25922 (Fig. 2A), the activities of azithromycin, fusidic acid, mupirocin, and retapamulin were all increased at least 32-fold (Table 3); the efficacy of meropenem was not increased by combining the antibiotic with SPR741 (Table 3). Interestingly, against this panel of E. coli strains, the efficacy of clarithromycin, as judged by the MIC₉₀ results, was unchanged in combination with 2 μg/ml SPR741 (MIC₉₀ of 32 μg/ml; range, 8 to >128 μg/ml) (Table 3). While this could be the result of the limited concentration of SPR741 used in these experiments, three observations suggest that the efficacy of this combination could be limited by intrinsic resistance to clarithromycin: first, although the MIC₉₀ of clarithromycin in the combination was unchanged, the range of MICs was widened considerably to 0.016 to 32 μg/ml; second, substantial potentiation of other antibiotics had been observed under the same conditions (Table 3); and third, in the presence of 2 μg/ml SPR741, the MIC of clarithromycin against E. coli ATCC 25922 was reduced 512-fold (data not shown). For the A. baumannii panel, 8 μg/ml SPR741 was able to effectively potentiate the activities of fusidic acid and rifampin (Table 3), with reductions in the MIC₉₀s of 512- and 256-fold, respectively, but also of meropenem (256-fold reduction in the MIC₉₀). Against the K. pneumoniae panel, reductions in the MIC₉₀s of each antibiotic were smaller than those against the E. coli and A. baumannii panels (Table 3). Retapamulin was the most effectively potentiated antibiotic in combination with 8 μg/ml SPR741, with a 32-fold reduction in the MIC₉₀ compared to the value for the antibiotic alone (Table 3). SPR741 was also effective in combination with both azithromycin and rifampin (16-fold reductions in MIC₉₀s). Based on the reductions in the MIC₉₀s against each organism, it was possible to rank the effectiveness of each antibiotic-SPR741 combination. Overall, against the three species of bacteria, the magnitude of the potentiation of this series of antibiotics by SPR741 followed the descending order of rifampin, fusidic acid, retapamulin, azithromycin, mupirocin, meropenem, aztreonam, and clarithromycin. Interestingly, retapamulin was equally well potentiated by SPR741 against all three species (Table 3). There were differences in this order among the three species; importantly, these broader strain panels demonstrated that although several antibiotics benefited greatly from combination with SPR741 when assessed against susceptible isolates (Fig. 2), the broader panels tested here revealed that intrinsic resistance determinants that may be present in some strains could limit the effectiveness of certain combinations against clinical and MDR isolates (e.g., clarithromycin).

In summary, the data presented here demonstrate that SPR741 can be used effectively to potentiate and extend the spectrum of activity of several antibiotics with diverse targets and mechanisms of action against E. coli, K. pneumoniae, and A. baumanniiin vitro.

We acknowledge Michael Pucci and Thomas Parr for critical reading of the manuscript.