ABSTRACT
BAL5788 is the water-soluble prodrug of BAL9141, a novel broad-spectrum cephalosporin with potent bactericidal activities against methicillin-resistant Staphylococcus aureus (MRSA) and penicillin-resistant Streptococcus pneumoniae. We investigated the safety and pharmacokinetics of BAL5788 in a double-blind, single-ascending-dose study with 40 healthy male subjects. The subjects were randomized to receive placebo (n = 2 subjects per dose) or BAL5788 (n = 6 subjects per dose) as a 200-ml intravenous infusion over 30 min. The BAL5788 doses used were 125, 250, 500, 750, and 1,000 mg (BAL9141 equivalents). All doses were well tolerated, with no severe or serious adverse events (AEs). The most frequent AE was taste disturbance. No electrocardiographic abnormalities and no trends or clinically significant changes in laboratory parameters or vital signs were observed. The maximum concentration of drug in serum and the area under the concentration-time curve for BAL9141 were dose proportional over the dosing range. The elimination half-life of BAL9141 was about 3 h. The volume of distribution at steady state was equal to the volume of the adult extracellular water compartment, and the rate of renal clearance of free drug corresponded to the normal glomerular filtration rate for adults. More than 70% of the administered dose was excreted as BAL9141 in the urine, and almost no prodrug was detected. After the infusion of 750 mg, the mean plasma BAL9141 concentrations exceeded the MIC at which 100% of MRSA isolates are inhibited (4 μg/ml) for approximately 7 h, or 58% of a 12-h dosing interval. These results indicate that infusions of 750 mg twice a day should be adequate for the treatment of infections caused by MRSA.
Cephalosporins are widely used for the treatment and prophylaxis of bacterial infections, largely due to their broad-spectrum activities and low risk of toxicity (5). Since their introduction, however, the use of β-lactam antibiotics has been compromised by the emergence of resistance in commonly targeted pathogens. In particular, the growing prevalence (5, 6, 7) of multiresistant pathogens, including methicillin-resistant Staphylococcus aureus (MRSA) and methicillin-resistant Staphylococcus epidermidis, has led some treatment centers to abandon the routine first-line use of cephalosporins, in some cases at the expense of greater toxicity and suboptimal overall coverage (5).
Until recently, efforts to develop β-lactam antibiotics active against MRSA have largely been unsuccessful (10). One exception is BAL9141, a novel pyrrolidinone-3-ylidenemethyl cephalosporin specifically designed to have strong affinity for penicillin-binding proteins (PBPs; PBP 2a and PBP 2x) known to confer resistance in staphylococci and pneumococci, respectively (12). BAL9141 also binds strongly to the relevant PBPs of most gram-positive and gram-negative pathogens and is resistant to many β-lactamases (12, 13). For 77 (12) and 96 (13) MRSA clinical isolates from the United States, Europe, South America, and Japan, the ranges of BAL9141 MICs were 0.5 to 4 and 0.12 to 2 μg/ml, respectively.
In vivo, BAL9141 (molecular weight, 534.6) is administered as a water-soluble prodrug, BAL5788 (molecular weight,712.7), which is rapidly and quantitatively cleaved in human and animal plasma to form BAL9141 plus diacetyl and CO2 (Fig. 1) (12). The prodrug BAL5788 was effective in mouse models of septicemia and subcutaneous abscesses caused by MRSA (12), in a neutropenic mouse model of thigh infection caused by MRSA and penicillin-resistant S. pneumoniae (D. R. Andes and W. A. Craig, Abstr. 40th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 1079, 2000), in a mouse model of acute pneumonia caused by penicillin-resistant Streptococcus pneumoniae also resistant to cephalosporins (E. Azoulay-Dupuis, J. Mohler, J. Bedos, A. Schmitt-Hoffmann, and S. Shapiro, Abstr. 42nd Intersci. Conf. Antimicrob. Agents Chemother., abstr. F335, 2002), and in a rat model of aortic valve endocarditis caused by MRSA (8).
Conversion of prodrug BAL5788 to the active cephalosporin BAL9141.
In view of the promising activities of BAL9141 in vitro and BAL5788 in animal models, we have investigated the safety and pharmacokinetic characteristics of BAL5788 in a single-ascending-dose clinical study. On the basis of data obtained after the administration of single doses, the time (as a percentage of the dosing interval) that the concentration of active drug in plasma exceeds the MICs for target pathogens (T > MIC) was assessed in order to predict the therapeutic doses of BAL5788 (2, 4, 15).
MATERIALS AND METHODS
Subjects and study procedures.The study described here was a single-center, double-blind, randomized, placebo-controlled, single-ascending-dose study involving five sequential doses of BAL5788. The BAL5788 doses tested were 125, 250, 500, 750, and 1,000 mg (equivalents of the active drug BAL9141 corresponding to 167, 333, 667, 1,000, and 1,333 mg of the prodrug BAL5788, respectively). The study was performed at the Clinical Pharmacology Unit, 1st Department of Medicine, University of Debrecen, Debrecen, Hungary, in accordance with the principles of the Declaration of Helsinki and all of its amendments. All volunteers gave written, informed consent.
Cohorts of healthy male subjects were randomly assigned to receive BAL5788 (six subjects per dose group) or placebo (two subjects per dose group). A prestudy screening was done within 2 weeks before the first study day and included a medical history, physical examination, and a battery of clinical laboratory tests. On the study day, after an overnight fast, BAL5788 or placebo (5% dextrose solution for intravenous infusion) in 200 ml was infused over a period of 30 min into a convenient forearm vein with a constant-rate infusion pump. The subjects remained in the unit for 24 h. Before proceeding to the next higher dose, a full assessment was made of all relevant tolerability and safety data at the previous dose. A follow-up examination was performed on day 7 ± 3 after the drug infusion.
Vital signs (blood pressure and pulse rate) were recorded before dosing and at 30 min and 1, 4, 12, and 24 h after the start of the infusion. A 12-lead electrocardiogram was recorded at approximately 25 min and at 24 h after the start of the infusion. Laboratory tests were conducted 24 h after the start of the infusion. Adverse events (AEs) were recorded throughout the period of dosing and at follow-up.
Blood samples for pharmacokinetic analysis were collected immediately before the start of infusion and at 5, 15, 30, and 45 min and 1, 1.25, 1.5, 2, 3, 4, 6, 8, 10, 12, 14, 16, and 24 h after the start of infusion. Blood samples were treated with EDTA and citric acid to stabilize the samples and to prevent the ex vivo hydrolysis of BAL5788. The subjects were asked to void urine a few minutes before dosing. Subsequently, urine was collected in containers pretreated with citric acid for collection periods of 2 h initially and, later, 4 to 12 h and up to 24 h after dosing.
Sample analysis.BAL9141 and BAL5788 were quantified by gradient reversed-phase liquid chromatography (LC) in the back-flush mode coupled online with a tandem mass spectrometer (LC/LC-mass spectrometry [MS]/MS). Monitoring of selected reactions was performed on a SCIEX API365 mass spectrometer operated in the positive ionization detection mode. The applied transitions were m/z 535 to >308 for BAL9141 and m/z 691 to >535 for BAL5788 at a dwell time of 250 ms.
Sample workup was carried out with 100 μl of plasma or 200 μl of urine. Plasma samples were prepared by precipitation of plasma proteins with 20% perchloric acid, addition of internal standards, and centrifugation of the mixtures. Aliquots of supernatant (75 μl into a 250-μl loop) were injected for analysis. Urine samples were prepared by dilution with 0.1 M ammonium acetate buffer and an internal standard solution and centrifugation. Aliquots of supernatant (30 μl into a 250-μl loop) were injected for analysis.
The limit of quantification (LOQ) was the lowest concentration of analyte in plasma or urine which could be measured with an interassay precision and an accuracy of 100% ± 20%. For plasma, this limit was set to 20.0 ng/ml for both BAL9141 and BAL5788, and all concentrations above 16.0 ng/ml detected are reported. For urine, the LOQ was set to 100 ng/ml for both BAL9141 and BAL5788, and all concentrations above 80.0 ng/ml detected are reported. The lower concentrations detected were reported as below the LOQ.
When mean drug levels were calculated for a subject cohort, if <50% of the samples had concentrations below the LOQ, the drug levels were set equal to zero. If ≥50% of the samples had concentrations below the LOQ, no mean value was calculated.
Pharmacokinetic analysis.The values of the pharmacokinetic parameters maximum concentration in plasma (Cmax), half-life (t1/2) in the distribution phase (t1/2α), t1/2 in the postdistribution phase (t1/2β), the area under the concentration-time curve (AUC) from time zero to the last sampling time (AUC0-t), the AUC from time zero extrapolated to infinity (AUC0-∞), total systemic clearance (CLS), total renal clearance (CLR), and the volume of distribution at steady state (VSS) were derived by noncompartmental and two-compartment methods with WinNonlin (version 2.1) software (Scientific Consulting Inc.).
Statistical analysis was performed with S-Plus 2000 software. The values of AUC and Cmax were logarithmically transformed by assuming a log-normal distribution. Geometric means and 95% confidence intervals were calculated from the means and the 95% confidence intervals of the transformed values by exponentiation. A one-way analysis of variance (ANOVA) with the factor dose was performed on the logarithmically transformed and dose-normalized values of AUC and Cmax. Dose proportionality was tested by comparing the group means of dose-normalized and logarithmically transformed values of AUC and Cmax. For VSS, CL, and t1/2, dose dependency was tested by plotting the values of these variables as a function of dose and by use of ANOVA to determine the effect of slope for the independent variable, dose.
RESULTS
All 40 subjects completed all assessments and were evaluable for safety. One subject in the 500-mg dose group developed leakage at the injection site and received an undefined partial dose of BAL5788. Therefore, data for this subject were excluded from the pharmacokinetic assessment. The remaining 39 subjects completed the dosings as planned, and their data were evaluable for pharmacokinetic analysis.
Conversion of the prodrug BAL5788 to the active cephalosporin BAL9141 was rapid; low levels of prodrug were observed in plasma only during infusion and were insufficient to allow pharmacokinetic evaluations. Figure 2 shows the mean concentration profiles of the active drug BAL9141 in plasma at each dose. Table 1 presents the pharmacokinetic parameter values, derived by use of a noncompartmental model for all parameters except t1/2α, for which the values were derived by use of a two-compartment model. Intersubject variability was low: the coefficients of variation for Cmax and AUC were 7.9 and 13.9%, respectively, for the 125-mg dose; 14.1 and 13.7%, respectively, for the 250-mg dose; 19.1 and 5.1%, respectively, for the 500-mg dose; 18.0 and 20.4%, respectively, for the 750-mg dose; and 12.1 and 6.0%, respectively, for the 1,000-mg dose.
Profiles of mean concentrations of BAL9141 in plasma determined after administration of single ascending intravenous infusions of BAL5788 over 30 min in six subjects for each dose (five subjects were included in the 500-mg dose group).
Values for pharmacokinetic parameters for BAL9141 after intravenous infusion of single ascending dosesa
The peak levels of the active drug BAL9141 in plasma were observed at the end of the 30-min infusion. Afterwards, the concentrations in plasma declined in a biphasic manner consistent with a rapid distribution of BAL9141 from the systemic circulation into other body compartments. The volume of distribution (15 to 18 liter) was about equal to the volume of the extracellular water compartment in adults (16).
Figure 3 shows the results of the assessment of the dose proportionality of AUC. No significant dose effect was found for AUC0-∞ (P = 0.19) or Cmax (P = 0.60), and therefore, the values of both parameters were considered to be dose proportional up to a dose of 1,000 mg. No significant dose dependence was found for VSS (P = 0.404) or CLS (P = 0.550). A borderline effect for t1/2β (P = 0.055) was consistent with the somewhat shorter elimination t1/2 values observed at the lowest dose (Table 1).
Mean AUC0-∞ of BAL9141 plotted against dose. Error bars indicate standard deviations.
BAL9141 was predominantly eliminated in the urine. Table 2 shows the cumulative recovery of drug in urine for each dose, along with the maximal urine drug concentrations. The highest urine drug concentrations were observed in the first collection period, 0 to 2 h after start of the infusion, and the highest concentration (2.2 mg/liter) occurred at the highest dose. The concentrations of the prodrug BAL5788 in urine were very low; the total urinary recovery of prodrug derived from measurable BAL5788 concentrations amounted to 0.7 to 2.2% of the infused dose.
Recovery and concentrations of BAL9141 in urine
The mean CLR of BAL9141 (Table 1) ranged from 4.1 to 5.1 liters/h (68.3 to 85.0 ml/min). Because the free (unbound) fraction of BAL9141 in plasma amounts to 62% (A. Schmitt-Hoffmann, unpublished data), the CLR of the free fraction reached 110 to 136 ml/min, in good agreement with the normal glomerular filtration rate of 125 ml/min in adults.
Table 3 shows the number of hours that the total and unbound concentrations of BAL9141 in plasma exceeded the MIC of 4 μg/ml for each dose. After the infusion of BAL5788 at 500 to 1,000 mg, total concentrations were above this MIC for 5 to 7 h, corresponding to a T > MIC of 42 to 58% of a 12-h dosing interval, assuming twice-daily (b.i.d.) dosing.
Doses and times that plasma drug concentrations exceeded the MIC and T > MIC
Overall, eight subjects dosed with BAL5788 experienced 10 AEs. Of these, seven AEs were reported as mild taste disturbances, described by subjects as a caramel-like taste experienced during the infusion period. The remaining three AEs were mild nausea (two subjects) and moderate vomiting (one subject). Except for taste disturbance, the frequency of AEs was not clearly related to the dose of BAL5788. None of the AEs were severe or serious, and none required treatment. No electrocardiographic abnormalities and no clinically relevant changes in laboratory parameters or vital signs were reported.
DISCUSSION
The main finding from this trial was that single doses of BAL5788 are well tolerated when they are given at doses higher than those predicted to be clinically effective for the treatment of infections caused by MRSA. The clinical efficacies of β-lactam antibiotics are correlated to values of T > MIC of at least 40 to 60% (2, 4, 15), although some animal models support a lower T > MIC range (14 to 28%) for BAL9141 for S. aureus and S. pneumoniae species (12; Andes and Craig, 40th ICAAC; Azoulay-Dupuis et al., 42nd ICAAC). According to this criterion, b.i.d. infusions of BAL5788 at 500 to 1,000 mg should be clinically effective because they lead to total concentrations of BAL9141 in plasma above the highest reported MIC for MRSA (4 μg/ml) for 5 to 7 h, which corresponds to 42 to 58% of a 12-h dosing interval.
All BAL5788 doses were well tolerated, with no severe or serious adverse reactions or clinically relevant changes in laboratory parameters. A characteristic event during the infusion period was a caramel-like taste, which was probably attributable to the conversion of the prodrug to produce BAL9141 plus diacetyl as a prodrug moiety (Fig. 1) (12). Diacetyl, a natural product (e.g., from milleri group streptococci) (3, 19), is known as a caramel-like odorant and is further metabolized to acetoin and 2,3-butanedione (1).
The results of this study indicate that infusion of the prodrug BAL5788 rapidly delivers high concentrations of active drug BAL9141 into the systemic circulation and that BAL9141 has pharmacokinetic properties characteristic of β-lactam antibiotics. BAL9141 was rapidly distributed to body compartments. The apparent volume of distribution (18 to 20 liter) was similar to the values reported for other β-lactam antibiotics (14, 22) and was equal to the volume of the extracellular fluid compartment in adults (16). The CLR of the free drug corresponded to the normal glomerular filtration rate in adults. Drug exposure was proportional to the dose from 125 to 1,000 mg. The values of clearance, the volume of distribution, and t1/2β remained constant over this dose range.
The level of systemic drug exposure in a wider population is expected to be predictable, due to the dose linearity of Cmax and AUC, the low intersubject variability in pharmacokinetic parameters, the parenteral route of administration, and renal elimination of the drug. The t1/2β of approximately 3 h appears to be well suited for b.i.d. administration, with little or no drug accumulation under these circumstances. Glomerular filtration of the active metabolite seemed to be predominantly responsible for the removal of free drug from the systemic circulation, and other elimination processes such as metabolic clearance appeared to have a negligible impact.
The prevalence of MRSA isolates among patients with nosocomial infections in the United States has increased markedly, from 5% in 1981 to 29% in 1991 and 35 to 40% in 2001 (11, 17, 18). At present, β-lactam antibiotics are not indicated once a diagnosis of MRSA is established, and antibiotics clinically effective against this pathogen are limited in number. Vancomycin (and teicoplanin in some countries) is widely regarded as the standard therapy for MRSA infections, but clinical failures have been attributed to reduced susceptibility (20) or high-level pathogen resistance (21) to glycopeptides. Recently, the oxazolidinone antibiotic linezolid and the streptogramin combination dalfopristin-quinupristin have become available for the treatment of MRSA infections, but resistance to these agents has also been reported (9, 23). In view of the emerging resistance to the available agents, a broad-spectrum β-lactam with potent bactericidal activity against MRSA would be an attractive treatment option, introducing a new class of antibiotics into this specific therapeutic setting. Accordingly, BAL5788 may represent a substantial advance in antimicrobial therapy. Further clinical studies are warranted to evaluate its efficacy and safety in patients.
ACKNOWLEDGMENTS
We are grateful to I. Heinze-Kraus, M. Page, S. Shapiro, and M. Heep for discussions and critical review of the manuscript and to A. Juhasz, T. Kovacs, A. Ivanov, T. Pataki, and K. Stöckel for participation and assistance in the conduct of the trial. P. H. Joubert provided valuable assistance in preparation of the manuscript.
FOOTNOTES
- Received 30 April 2003.
- Returned for modification 22 December 2003.
- Accepted 4 March 2004.
- American Society for Microbiology
