Oral anti-pneumococcal activity and pharmacokinetic profiling of a novel peptide deformylase inhibitor

Transcription

1 Journal of Antimicrobial Chemotherapy (2004) 53, DOI: /jac/dkh108 Advance Access publication 12 February 2004 Oral anti-pneumococcal activity and pharmacokinetic profiling of a novel peptide deformylase inhibitor M. Gross 1 *, J. Clements 2, R. P. Beckett 2, W. Thomas 2, S. Taylor 2, D. Lofland 1, S. Ramanathan- Girish 1, M. Garcia 1, S. Difuntorum 1, U. Hoch 1, H. Chen 1 and K. W. Johnson 1 1 Genesoft Pharmaceuticals, 7300 Shoreline Ct., South San Francisco, CA 94080, USA; 2 British Biotech plc, Watlington Road, Oxford OX4 6LY, UK Received 12 September 2003; returned 3 November 2003; revised 3 December 2003; accepted 8 December 2003 Objective: BB-81384, a novel peptide deformylase (PDF) inhibitor, was characterized in terms of enzyme inhibition profile, antibacterial activity, rodent pharmacokinetics and oral efficacy in murine infection models. Methods: MICs were determined by standard NCCLS broth microdilution. Selectivity of metalloenzyme inhibition was determined with a limited panel of enzymes via standard biochemical assays. Profiling of the pharmacokinetics and select tissue disposition in mice was determined and compared with that of the macrolide, azithromycin. In vivo murine efficacy studies using Streptococcus pneumoniae were conducted using a peritonitis model, as well as lung and thigh burden models of infection. Results: BB selectively inhibited PDF with an IC nm and with MICs < 0.5 mg/l against most S. pneumoniae pathogens. Pharmacokinetic analysis revealed good oral bioavailability and moderate clearance and volume of distribution. BB partitioning to lung tissue was similar in terms of magnitude and kinetics to that of the plasma compartment. Single-administration oral efficacy in a mouse peritonitis model was evident with an ED 50 of 30 mg/kg. BB reduced the bacterial load by 5 and 3 log units in organburden models of lung and thigh infection, respectively. Conclusion: BB-81384, a novel PDF inhibitor with good activity against S. pneumoniae in vitro, was the first compound of this class to be profiled for oral pharmacokinetics and tissue disposition and to demonstrate oral anti-pneumococcal efficacy in mice. Keywords: peptide deformylase, efficacy, pharmacokinetics Introduction The need for antibacterial agents with novel mechanisms of action has been well documented with the percentage of bacteria resistant to currently prescribed antibiotics steadily increasing. 1 3 In the past 30 years, only one new class of antibacterials (oxazolidinones) has emerged in the marketplace, and resistance to this class has already been identified. 4,5 Peptide deformylase (PDF) is an essential bacterial metalloenzyme that deformylates the N-formylmethionine of newly synthesized bacterial polypeptides. Deformylation is an essential post-translational process in prokaryotic, but not eukaryotic, cells. 6,7 The bacterial PDF gene (def) exists in all sequenced pathogenic bacterial genomes and a homologue has been identified in mammalian cell mitochondria, but displayed weak deformylase activity. 8 Interestingly, it has been suggested recently that the natural PDF inhibitor, actinonin, acts as a stimulator of neutrophil-activating peptide secretion via interaction with formyl peptide receptors, thus implicating a second potential antibacterial mechanism. 9 By targeting a unique mechanism for antibacterial activity, such an agent would be expected to lack cross-resistance to existing classes of antibiotics. Indeed, PDF inhibitors have shown potent in vitro and in vivo activity against Gram-positive organisms, including drugresistant isolates For the reasons highlighted above, PDF inhibition is an attractive approach for antibacterial chemotherapy. In fact, a first-generation inhibitor, BB-83698, has been advanced into... *Present address. XOMA Corp., Berkeley, CA 94710, USA. Present address. Emisphere Technologies, Tarrytown, NY 10591, USA; Present address. Sunesis Pharmaceuticals, 341 Oyster Pt, South San Francisco, CA 94080, USA. Corresponding author. Fax: ; kjohnson@genesoft.com JAC vol.53 no.3 The British Society for Antimicrobial Chemotherapy 2004; all rights reserved.

2 M. Gross et al. clinical trials for hospitalized community-acquired pneumonia (CAP) ( product portfolio). Discovery of PDF inhibitors for the treatment of respiratory tract infections (RTIs) in which Streptococcus pneumoniae represents the most frequently isolated organism 15 has been undertaken. Despite numerous advances in PDF inhibitor discovery oral anti-pneumococcal efficacy in vivo has not yet been published. Moreover, the only other N-formylhydroxylamine-type inhibitor previously described (BB-3497) 11 exhibited poor activity against S. pneumoniae. Identification of P2 and P3 substituents that substantially improve antipneumococcal activity relative to BB-3497, has been a required step in therapeutic optimization for respiratory infections. 11,12 A potent, orally active PDF inhibitor for RTI would meet a great clinical need and represent an attractive development opportunity. We set out to identify and characterize orally active, novel PDF inhibitors with potential use against CAP. Demonstrated herein is the antipneumococcal oral efficacy of a novel PDF inhibitor (BB-81384) in multiple animal models of infection and, for the first time for any PDF inhibitor, details of its pharmacokinetics and tissue disposition. Materials and methods Enzyme inhibition The in vitro activity of BB against PDF and the related metalloenzymes collagenase (MMP-1), 72 kda gelatinase (MMP-2) and angiotensin I-converting enzyme was determined as previously described. 11 Briefly, fluorescent-labelled peptide substrates with purified recombinant human enzymes were used with fluorescence as the primary endpoint. Bacterial strains and susceptibility testing The in vitro bacterial susceptibility was determined by standard NCCLS broth microdilution. 16,17 Activity was evaluated against a broad range of organisms including the following subset of strains: methicillin-susceptible Staphylococcus aureus (MSSA) ATCC 29213; penicillin-resistant S. pneumoniae (PRSP) ATCC 9061; penicillin-susceptible S. pneumoniae (PSSP) ATCC 6301; penicillin-susceptible S. pneumoniae (PSSP) ATCC 6303; Haemophilus influenzae (ATCC 31517, ); Moraxella catarrhalis (ATCC 37054); Bacillus cereus (ATCC 11778); and Streptococcus pyogenes (ATCC 49399). Killing kinetics S. pneumoniae ATCC 6301 and 6303 were grown on trypticase soy agar (TSA) with 5% sheep blood (BD Biosciences, Sparks, MD, USA) at 37 C in 5% CO 2. From the overnight growth plate, several colonies were cultured into 3 ml of cation-adjusted Mueller Hinton broth (CAMHB) with 5% lysed horse blood. The broth subcultures were incubated for 2 h at 37 C in 5% CO 2. The 2 h culture of S. pneumoniae 6301 was adjusted to an OD 600 of 0.3 and diluted 1:50 in 5 ml of CAMHB to yield a final inoculum of cfu/ml into test tubes containing 1, 2 or 4 the MIC of antibiotic. A growth control tube without antibiotic was included. Viable counts were determined by removing 100 µl at intervals of 0, 2, 4, 8 and 24 h. Serial dilutions (1:10) were prepared in 96-well microtitre plates. Ten µl was plated on TSA plates containing 5% sheep blood and incubated overnight at 37 C in 5% CO 2. In addition, if tubes were macroscopically clear after overnight incubation, 100 µl was plated on agar and incubated as above. The lower limit of detection for this assay was 10 cfu/ml. Pharmacokinetics and tissue distribution The pharmacokinetics of BB and a reference drug, azithromycin, were characterized in ICR mice. Groups of three female mice per time point were treated via intravenous (iv) bolus (lateral tail vein) or oral gavage. At serial time points (e.g. 5, 20, 60, 120, 300 min), blood was collected in lithium heparin tubes by cardiac puncture and separated by centrifugation. Plasma samples were stored at 80 C until analysis. Tissue distribution. The disposition of BB and the reference drug azithromycin to select tissues following a 10 mg/kg oral dose was determined after single administration in female ICR mice over a period of up to 5 h (5, 20, 60, 120 and 300 min). Mice were exsanguinated and organs (lung, thigh muscle) were rapidly removed and flash frozen. Plasma and tissue samples were analysed internally or at Bay Bioanalytics Laboratory (Hercules, CA, USA) according to the methods described below. Plasma and tissue sample analysis was performed using a high-performance liquid chromatography/mass spectrometry/mass spectrometry (HPLC/MS/MS) method with appropriate quality controls. Plasma analysis. The sample preparation for BB consisted of a protein precipitation with acetonitrile (containing diphenhydramine as internal standard), removal of the precipitate by centrifugation, concentration of the extract under a stream of nitrogen and reconstitution in mobile phase. A standard curve in the range ng/ml was prepared in blank plasma, and treated as described above. Quality controls (1000, 2000, 5000 ng/ml) were prepared and interspersed throughout the sequence run. Back-calculated standards and quality control samples differed by no more than 11% from nominal values. For azithromycin, a similar protein precipitation method was used to process plasma samples, and a plasma standard curve with standards in the range ng/ml was used for quantification. Tissue analysis. To a 50 mg tissue aliquot, 50 µl of water and 150 µl of ice-cold internal standard solution (0.1 µg/ml diphenhydramine in 1:1 methanol/acetonitrile) were added, followed by homogenization at high speed for s with a serrated microtip on a Polytron homogenizer (Brinkman, Westbury, NY, USA). The samples were then spun in a microcentrifuge at rpm for 5 min. An aliquot (40 µl) was transferred from each well to an autosampler microplate to which 200 µl of 0.2% formic acid was added and the plate was sealed with a cap mat. One hundred microlitres from each well was then analysed by HPLC/MS/ MS. Standard curves with standards in the range ng/ml with appropriate quality controls prepared using blank lung and thigh tissues were used for quantification of BB and azithromycin in tissues. Standards and quality controls did not deviate from nominal concentrations by >29%. Murine pneumococcal peritonitis model S. pneumoniae ATCC 6301 (PSSP) was used for these experiments. Fresh colonies were grown in brain heart infusion (BHI) broth (International BioProducts, Bothell, WA, USA) at 37 C with gentle agitation, and adjusted to a concentration of 100 cfu/ml by dilution in sterile 1 PBS (Hyclone, Logan, UT, USA). This represented greater than 10 the LD 50 inoculum, which corresponded to 100% mortality rate by 48 h in vehicle-treated animals (data not shown). Groups of five ICR female mice ( 25 g from Taconic Farms) were inoculated intraperitoneally (ip) with 0.5 ml of the inoculum described above. Groups were treated orally with vehicle, positive control (i.e. azithromycin, co-amoxiclav) or test agents 1 h post-infection. Clinical observations were made over 5 days with a primary endpoint of survival. For all in vivo studies described herein, BB was formulated in a 50 mm citrate buffer containing 140 mm NaCl at ph 4.0. All animal studies strictly adhered to the guidelines set by the US Federal Government and the internal Institutional Animal Care and Use Committee (IACUC) at Genesoft Pharmaceuticals Inc. 488

3 Orally effective novel peptide deformylase inhibitor S. pneumoniae lung infection S. pneumoniae ATCC 6303 (PSSP) was grown overnight in Mueller Hinton broth (BD Biosciences) containing 5% lysed horse blood to a concentration of cfu/ml. Mice were anaesthetized with isofluorane (Baxter, Deerfield, IL, USA) and inoculated with 50 µl of overnight broth culture via the intranasal route. One hour post-inoculation, lungs were collected aseptically into 1 ml of sterile PBS to assess bacterial load. Drug treatments were oral once or twice daily for 3 days beginning 1 h post-inoculation. Survival was monitored over 7 days. At pre-determined time-points post-treatment, lungs were collected aseptically into 1 ml of 1 sterile PBS. Quantification of bacterial load was accomplished by homogenizing lungs with a Polytron homogenizer PT 10/35 (Brinkman) in 1 ml of sterile PBS. Serial dilutions were performed using PBS as diluent. Three dilutions were plated per homogenate (100 µl each) onto TSA plates containing 5% sheep blood (BD Biosciences). Efficacy was determined by comparing bacterial counts of negative control (vehicle) groups to the test compounds. Blood was collected immediately post-mortem and bacterial counts measured by serial dilution in PBS as above. S. pneumoniae neutropenic mouse thigh infection S. pneumoniae ATCC 6301 was grown overnight in BHI broth at 37 C with gentle agitation and adjusted to a concentration of 10 6 cfu/ml by dilution in sterile 1 PBS. The protocol was essentially as described by Craig & Andes 13 and Andes et al. 18 ICR mice were rendered neutropenic by ip injections of cyclophosphamide 200 mg/kg (Sigma) on days 4 and 1 prior to infection. Groups of three neutropenic mice under isofluorane anaesthesia were inoculated intramuscularly (im) in the anterior thigh (50 µl) with cfu/thigh in both thighs, at time 2 h. At time 0, mice were treated iv or by mouth with test agents, or negative or positive control compounds. At pre-determined time points post-treatment (5 and 24 h), mice were sacrificed and thighs collected aseptically into 10 ml of 1 sterile PBS. Quantification of bacterial load was accomplished as described above for the lung infection model. Data analysis The pharmacokinetic analyses were based on the average plasma concentration for three animals per treatment group at each sample time. Non-compartmental modelling was performed using WinNonlin software (Pharsight, Corp., Mountainview, CA, USA). Data analysis for ED 50 determination and coefficient of variance was performed with the WinNonlin pharmacodynamic (efficacy) analysis package. Results PDF inhibitor Directed optimization efforts involving medicinal chemistry and preliminary, limited-spectrum MIC activity described previously and extended herein, led to the identification of BB depicted in Figure 1. 11,12,19 Relative to the early prototype BB-3497 (S. pneumoniae MIC 8 mg/l), 11 the substituted piperazine moiety at P3 resulted in significantly improved anti-pneumococcal activity as described below. Enzyme inhibition BB exhibited potent inhibition of peptide deformylase from several different bacterial sources, as indicated in Table 1. Also represented is a high degree of selectivity for PDF against a small set of Figure 1. Chemical structure of BB Molecular weight = other relevant mammalian metalloenzymes including MMP-1 and -2 and angiotensin I-converting enzyme. In vitro susceptibility The MICs of BB and reference drugs azithromycin and co-amoxiclav are depicted in Table 2. While an insufficient number of different S. pneumoniae strains were screened for initial MIC 50 and MIC 90 determinations, of eight different strains tested, the MICs of BB were in the range mg/l, with only two strains at the less susceptible end. MICs of BB versus the ATCC strains of H. influenzae were 2 4 mg/l. In addition, BB showed impressive potency against S. pyogenes as well as M. catarrhalis, with MICs of 0.25 and mg/l, respectively. The compound was less potent against B. cereus (MIC = 4 mg/l). Killing kinetics Table 1. BB inhibition of peptide deformylases and select metalloenzymes Enzyme The killing kinetics of BB against S. pneumoniae ATCC 6301 are shown in Figure 2. The viable counts decreased as the antibiotic concentration increased. The difference between the viable count at time 0 and after 24 h at 1, 2 and 4 MIC was 0.3, 2.1 and 2.7 log 10 cfu/ml, respectively. Pharmacokinetics IC 50 (nm) PDF-Ni S. pneumoniae 9 PDF-Ni H. influenzae 11 PDF-Ni Escherichia coli 60 PDF-Ni S. aureus 30 Collagenase (MMP-1) Gelatinase (MMP-2) Angiotensin I-converting enzyme The observed pharmacokinetics of BB in mice are depicted in Table 3 and Figure 3. BB administered iv in mice was characterized by a biphasic disposition with a mean terminal elimination half-life of 2.2 h. The clearance of BB was 1.5 L/h/kg and the volume of distribution at steady state (1.6 L/kg) exceeded total body water in mice. Oral pharmacokinetics were assessed at two dose levels (10 and 50 mg/kg). After oral administration of BB-81384, plasma concentrations increased rapidly and mean maximum plasma concentrations were achieved within 20 min. Maximum plasma concentrations fell steadily in a biphasic manner with a mean terminal 489

4 M. Gross et al. Table 2. In vitro antibacterial activity MIC (mg/l) Compound S. aureus (MSSA) S. pneumoniae 9061 (PRSP) S. pneumoniae 6301 (PSSP) S. pneumoniae 6303 (PSSP) S. pyogenes H. influenzae H. influenzae M. catarrhalis B. cereus BB Co-amoxiclav >64 Azithromycin Ofloxacin Table 3. Pharmacokinetic properties of BB in mice Dose (mg/kg) Route C max (mg/l) t 1/2 (h) AUC 0 5 (mg. h/l) CL (ml/min/kg) V ss (L/kg) %F a 10 iv by mouth by mouth a Based on AUC 0 Tissue distribution profile A screen of the tissue partitioning of BB and the reference drug, azithromycin, was undertaken to better understand the in vivo pharmacology of BB-81384, including potential comparisons with azithromycin in vivo potency. The data are represented in Figure 4(a and b). Concentrations of BB in the lung and thigh tissues reached maximum levels within 20 min. Lung and plasma concentrations were comparable, whereas thigh concentrations were approximately one-third of plasma levels. The partitioning of azithromycin into thigh and lung tissue was substantially greater than that of BB Figure 2. Killing kinetics of BB against S. pneumoniae ATCC Each point represents viable bacterial counts at selected time-points. half-life of h. The increase in C max and area under the curve (AUC) values appeared to be more than dose proportional, but may be attributed to limited sampling up to 5 h. A preliminary estimate of bioavailability of BB at 10 and 50 mg/kg dose levels yielded values of 55% and 88%, respectively. Mouse peritonitis efficacy BB exhibited good oral efficacy in a stringent lethal peritonitis model (Table 4). Vehicle-treated control animals all died within 48 h of inoculation. A single oral administration of 30 mg/kg protected 50% of mice from an S. pneumoniae lethal infection. Complete protection from the infection was achieved by a single 90 mg/kg oral dose of BB In subsequent peritoneal sepsis studies, we found that increasing the oral administration frequency to twice daily on the day of infection (i.e. 1 and 5 h post-infection) reduced the ED 50 to 20 mg/kg (data not shown). S. pneumoniae neutropenic mouse thigh infection Two dose levels of BB-83184, 30 and 60 mg/kg, were administered orally in mice following thigh infection with S. pneumoniae. The thigh burden results are shown in Table 5. At 5 h post-treatment, BB reduced the bacterial load relative to the vehicle control by 3.8 log 10 cfu/thigh. At 24 h post-treatment, BB reduced the viable bacteria by 2.6 and 3.3 log 10 cfu/thigh for 30 and 60 mg/kg doses, respectively. In the same study, co-amoxiclav at a 10 mg/kg dose lowered the bacterial load in comparison to the vehicle control by 5 log units at 5 and 24 h post-treatment. Murine lung infection efficacy In the S. pneumoniae lung infection model, all vehicle-treated control animals died within 72 h of infection. At the time of death, animals had cfu/lung and cfu/ml in the blood. BB was tested in a model wherein only robust therapies have proven effi- 490

5 Orally effective novel peptide deformylase inhibitor Table 4. Efficacy of BB and reference drugs in the mouse peritonitis model Compound Dose (mg/kg) 5 day survival (%) BB BB BB Co-amoxiclav Azithromycin Animals treated with test compound orally 1 h post-infection. Results represent two separate studies with n 5/group; ED 50 coefficient of variance 5%. Figure 3. Mouse plasma pharmacokinetics of BB after iv and oral administration. Plasma concentrations of parent BB following a single administration. Each point represents mean plasma concentrations (mg/l) ± S.D. from three mice. Analysis of plasma concentrations was performed by HPLC/ MS/MS. Figure 4. Tissue disposition of BB (a) and azithromycin (b) in mice following a 10 mg/kg oral dose. Each point represents mean plasma or tissue concentrations ± S.D. from three mice. cacious and, when administered twice daily for 3 days at 100 mg/kg, resulted in long-term survival of 67% (Table 6). Notably, BB reduced the bacterial load in the lungs by 5 log units relative to the vehicle control, which was comparable to or better than that observed with the reference drugs (Figure 5). The reduction in bacterial load at 22 h was greater than that of co-amoxiclav, albeit at a higher dose. The efficacy of BB was comparable with that of ofloxacin in the same dosing regimen, although a dose of 100 mg/kg ofloxacin once daily did not provide sustained protection (Table 6). Discussion PDF inhibitors have been shown previously to possess anti-grampositive activity in vitro and in vivo including parenteral anti-pneumococcal and oral anti-staphylococcal activities. 11,13,14 Herein, we pharmacologically profiled the PDF inhibitor BB to an extent not previously reported for a PDF inhibitor, including demonstration of the first in vivo oral efficacy against S. pneumoniae. The advancement was significant in that PDF inhibitors are being optimized for RTIs, the majority of which are caused by S. pneumoniae, and an orally administered therapeutic would be most desirable for the community setting. The in vitro susceptibility and time kill profiles were encouraging for potential application to community-acquired pneumonia and possibly even acute exacerbation of chronic bronchitis (AECB). It is not uncommon for antibacterials, even PDF inhibitors, to exhibit both bacteriostatic and bactericidal activities against S. pneumoniae. 12,14 In vitro activity against H. influenzae was not as good as desired for AECB, but represented an improvement over the first-generation compound, BB ,21 The pharmacokinetics of BB were attractive as clearance and volume of distribution were moderate and oral bioavailability was good (55% 88%, depending on the dose). Examination of pharmacokinetic parameters and the efficacy profile suggested that the kinetics were generally dose proportional, although more definitive conclusions about pharmacokinetic dose-linearity would require more extensive testing. Disposition to tissues clinically relevant for Gram-positive infection such as lung and muscle has not been published for related PDF inhibitors and was therefore determined for BB Distribution was shown to be rapid, with maximum lung and thigh concentrations being either comparable to plasma levels or three- to five-fold lower. BB cleared from tissues with kinetics similar to elimination from the plasma compartment, which facilitates potential tissue modelling by analysis of the central compartment. Combined with protein binding analysis via an ultracentrifugation method, which suggested that plasma protein binding for BB was 74% (data not shown), the pharmacokinetics of BB were considered favourable. The ED 50 of 30 mg/kg for BB in an S. pneumoniae lethal peritonitis model confirmed significant oral efficacy. The efficacy was dose proportional and potency generally correlated with in vitro 491

6 M. Gross et al. Table 5. Efficacy of BB and reference drugs in a murine neutropenic thigh model Compound Dose Log 10 cfu/thigh reduction relative to vehicle control (t = 5 h) Log 10 cfu/thigh reduction relative to vehicle control (t = 24 h) Vehicle BB BB ND 3.3 Co-amoxiclav ND, not done. Table 6. Efficacy of BB and reference drugs in a murine lung infection model Compound Dose (mg/kg) Regimen Log 10 cfu reduction (22 h) Log 10 cfu reduction (50 h) Day 3 survival (%) Day 6 survival (%) Vehicle 0 twice daily 3 days Co-amoxiclav 10 twice daily 3 days Azithromycin 3 twice daily 3 days Ofloxacin 100 once daily 3 days ND ND Ofloxacin 100 twice daily 3 days ND ND BB once daily 3 days ND ND BB twice daily 3 days BB twice daily 3 days activity (0.5 mg/l for S. pneumoniae 6301). The greater potency of comparator drugs could be attributed to enhanced in vitro potency in the case of co-amoxiclav and similar MICs but high tissue penetration in the case of azithromycin, as further described below (Tables 2 and 4, and Figure 4b). The BB results are similar to limited efficacy studies described by others, which show altogether good correlation of in vivo efficacy with plasma exposure and in vitro MICs. 13,14 Oral administration of BB also resulted in good efficacy in organ burden models of lung and thigh infection. Long-term survival in the pneumonia model was observed following a 100 mg/kg twicedaily regimen. The reason for the reduced long-term survival for the BB treated animals relative to the co-amoxiclav-treated animals was unclear, although it may have been related to the superior potency and bactericidal nature of co-amoxiclav against the S. pneumoniae strain used in the lung infection model. Importantly, survival correlated well with lung and blood cfu counts. Overall, the results of the S. pneumoniae peritoneal, thigh and lung infection studies demonstrated that BB-81384, an orally efficacious PDF inhibitor, had a dose-related activity profile which was both sufficient for proposed indications and comparable with those of reference drugs. Pharmacokinetic and pharmacodynamic (PK/PD) assessment, performed previously by Craig & Andes 13 on the clinical candidate BB-83698, showed that 24 h AUC/MIC appeared to drive efficacy. The similarities between BB and BB warranted investigation of preliminary PK/PD correlates for BB Profiling of in vivo tissue (lung, muscle) disposition was deemed important as it had not been well studied in general and would facilitate PK/PD projections. Accordingly, given thigh and plasma exposures to BB-81384, estimated AUC/MIC ratios for each compartment were derived. In the context of the thigh infection model, significant oral efficacy was observed at plasma or thigh AUC/MIC ratios in the range and 6 12 for 30 and 60 mg/kg doses, respectively (extrapolated from 10 and 50 mg/kg plasma AUC and 10 mg/kg tissue disposition data). Should the AUC/MIC ratio be validated for additional PDF inhibitors including BB-81384, then it would be pertinent to identify ratios correlating with significant efficacy in other pneumococcal models. C max as a critical determinant would be unlikely as BB iv and oral potencies were identical (ED 50 s 30 mg/kg in the peritonitis model). Pharmacokinetic data project minimal AUC differential whereas C max values would vary five- to 10-fold. Moreover, demonstration that 50 mg/kg twice daily provides better long-term survival in the lung infection model than 100 mg/kg once daily would further support AUC as opposed to C max as a primary PK parameter for BB in vivo pharmacodynamics. Notably, azithromycin had superior potency to BB in the mouse infection models. However, examination of select tissue disposition properties for azithromycin provided an explanation for the improved in vivo potency as compared with BB given similar in vitro MIC and plasma exposure profiles. Partitioning of azithromycin to peripheral tissue (e.g. lung and thigh) was substantial with a tissue/plasma ratio 70-fold greater than BB and with very slow apparent decay. The comparison ultimately becomes most relevant in man and would therefore be dependent upon the pharmacokinetic scaling of a compound like BB in humans. 492

7 Orally effective novel peptide deformylase inhibitor Figure 5. In vivo efficacy of BB-81384, co-amoxiclav and azithromycin in a murine lung infection model. Time course of mean log 10 cfu/lung of S. pneumoniae ATCC 6303 ± S.D. following oral dosing. Each point represents the mean log 10 cfu per lung ± S.D. from >3 mice. In summary, BB has been identified as a potent and selective inhibitor of peptide deformylase with potent anti-pneumococcal activity in vitro against drug-susceptible and -resistant strains. BB demonstrated good oral bioavailability, oral efficacy in multiple murine models of S. pneumoniae infection, and was welldistributed to peripheral tissues. Such results support continued optimization of orally active PDF inhibitors for advancement as novel clinical therapeutics. Acknowledgements We acknowledge the support of our colleagues Heinz Moser, Roland Bürli, David Knowles and Gary Patou. Stanley Falkow and Tom Parr are appreciated for their expert advice. Eric Taylor and Jake Pritchett at Bay Bioanalytical Labs (Hercules, CA, USA) provided critical technical assistance. This work was supported by Genesoft Pharmaceuticals and British Biotech Pharmaceuticals. References 1. Dennesen, P., Bonten, M. & Weinstein, R. (1998). Multiresistant bacteria as a hospital epidemic problem. Annals of Medicine 30, Kollef, M. & Fraser, V. (2001). Antibiotic resistance in the intensive care unit. Annals of Internal Medicine 20, Low, D. (2001). Antimicrobial drug use and resistance among respiratory pathogens in the community. Clinical Infectious Diseases 33, Suppl. 3, S Gonzales, R., Schreckenberger, P., Graham, M. et al. (2001). Infections due to vancomycin-resistant Enterococcus faecium resistant to linezolid. Lancet 357, Tsiodras, S., Gold, H., Sakoulas, G. et al. (2001). Linezolid resistance in a clinical isolate of Staphylococcus aureus. Lancet 358, Mazel, D., Pochet, S. & Marliere, P. (1994). Genetic characterization of polypeptide deformylase, a distinctive enzyme of eubacterial translation. EMBO Journal 13, Meinnel, T. & Blanquet, S. (1994). Characterization of the Thermus thermophilus locus encoding peptide deformylase and methionyl-trna f Met formyltransferase. Journal of Bacteriology 176, Nguyen, K. T., Hu, X., Colton, C. et al. (2003). Characterization of human peptide deformylase: implications for antibacterial drug design. Biochemistry 42, Fu, H., Dahlgren, C. & Bylund, J. (2003). Subinhibitory concentrations of the deformylase inhibitor actinonin increase bacterial release of neutrophil-activating peptides: a new approach to antimicrobial chemotherapy. Antimicrobial Agents and Chemotherapy 47, Jones, R. N. & Rhomberg, P. R. (2003). Comparative spectrum and activity of NVP-PDF386 (VRC4887), a new peptide deformylase inhibitor. Journal of Antimicrobial Chemotherapy 51, Clements, J. M., Beckett, R. P., Brown, A. et al. (2001). Antibiotic activity and characterization of BB-3497, a novel peptide deformylase inhibitor. Antimicrobial Agents and Chemotherapy 45, Clements, J. M., Ayscough, A. P., Keavey, K. et al. (2002). Peptide deformylase inhibitors, potential for a new class of broad spectrum antibacterials. Current Medicinal Chemistry Anti-Infective Agents 1, Craig, W. & Andes, D. (2001). In vivo pharmacodynamics of BB , a deformylase inhibitor. In Abstracts of the Forty-first Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, IL, Abstract F-355, p American Society for Microbiology, Washington, DC, USA. 14. Hackbarth, C. J., Chen, D. Z., Lewis, J. G. et al. (2002). N-alkyl urea hydroxamic acids as a new class of peptide deformylase inhibitors with antibacterial activity. Antimicrobial Agents and Chemotherapy 46, Marrie, T. J. (1994). Community-acquired pneumonia. Clinical Infectious Diseases 18, National Committee for Clinical Laboratory Standards. (2000). Methods for Determining Bactericidal Activity of Antimicrobial Agents. Proposed Guideline (M26-P). Approved Standard (M11-A3). NCCLS, Wayne, PA, USA. 17. National Committee for Clinical Laboratory Standards. (1997). Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically Fourth Edition: Approved Standard M7-A4. NCCLS, Wayne, PA, USA. 18. Andes, D., Van Ogtrop, M. L., Peng, J. et al. (2002). In vivo pharmacodynamics of a new oxazolidinone (linezolid). Antimicrobial Agents and Chemotherapy 46, Davies, S. J., Ayscough, A. P., Beckett, R. P. et al. (2003). Structure activity relationships of the peptide deformylase inhibitor BB-3497: modification of the P2 and P3 side chains. Bioorganic and Medicinal Chemistry Letters 13, Wise, R., Andrews, J. M. & Ashby, J. (2002). In vitro activities of peptide deformylase inhibitors against gram-positive pathogens. Antimicrobial Agents and Chemotherapy 46, Lofland, D., Difuntorum, S., Waller, A. et al. (2004). In vitro antibacterial activity of the peptide deformylase inhibitor BB Journal of Antimicrobial Chemotherapy 53, in press. 493