The use of esters as prodrugs for oral delivery of beta-lactam antibiotics.

It is apparent that the sequence of events that has been followed in the approach to the discovery and development of a new beta-lactam prodrug has been similar in many of the case histories we have studied and indeed similar to the approach we have followed. Initially, we select a suitable series of prodrug moieties, which either comprises totally novel structures or is deduced from the data bases available (bearing in mind reports of potential toxicity) or both. The successful preparation of these prodrugs and the studies undertaken to ensure they are of known purity and stability is not easy and, as would be expected, is the initial go/no-go decision. Usually, the next stage has involved the assessment of whether or not bioavailability of the parent molecule is increased after administration of the prodrug ester by gavage to laboratory animal species. The selection of which species to use has very often been made according to which has the most information available in those particular laboratories and in the literature. It is this process that can be dishearteningly misleading as was demonstrated in Table IV and Fig. 1. Increasing the range of animal species does not lead to a better ability to predict bioavailability in humans. Hydrolysis studies are important to ensure that any novel prodrug will hydrolyze in human tissues, and also in the clarification of why a particular prodrug is not performing as expected in animals. After selection, it is essential to determine where and how rapidly hydrolysis takes place in the animal species to be used for safety evaluation prior to the first bioavailability studies in humans. The assessment of absolute oral bioavailability has not always been undertaken. This would seem critical for studies in not only the selected animal species but also in humans. In the absence of these data it is difficult to judge whether oral uptake can be increased further by modifying the ester moieties and at the development stage to determine whether or not modifications in formulation could increase bioavailability. When the prodrug is being developed for an injectable beta-lactam already available for humans, there would be no problem, but it would be an important consideration during the development of an entirely novel beta-lactam antibiotic for which no parenteral data are available in humans. Animal data are not totally predictive. The development of prodrugs is not easy, as a consequence of species differences in the properties of the prodrug superimposed on those of the parent compound during the evaluation. However, technical advances have enabled us to assay concentrations more precisely, determine basic physicochemical properties more efficiently, understand absorption processes by the use of in vitro systems, and analyze data far more comprehensively by the use of ever-evolving computer software. The prodrug approach to increasing the oral bioavailability of beta-lactam antibiotics has provided clinically valuable agents and continues. Despite the inherent difficulties, knowledge gained over the years, of the relationships between physicochemical and biological properties of the parent compound and the intact prodrug ester, has contributed to the design of novel prodrugs and a number of novel auxiliaries have been developed.

Activity of amoxicillin-clavulanate against penicillin-resistant Streptococcus pneumoniae in an experimental respiratory infection model in rats.

High doses of amoxicillin, equivalent to those produced by 500- and 750-mg oral doses in humans (area under the plasma concentration-time curve), were effective against a penicillin-resistant strain of Streptococcus pneumoniae in an experimental respiratory tract infection in immunocompromised rats; this superior activity confirms the results of previous studies. An unexpected enhancement of amoxicillin's antibacterial activity in vivo against penicillin-resistant and -susceptible S. pneumoniae strains was observed when subtherapeutic doses of amoxicillin were coadministered with the beta-lactamase inhibitor potassium clavulanate. The reason for this enhancement was unclear since these organisms do not produce beta-lactamase. The differential binding of clavulanic acid and amoxicillin to penicillin-binding proteins may have contributed to the observed effects.

Effect of protein binding on penetration of beta-lactams into rabbit peripheral lymph.

The relevance of protein binding to penetration of beta-lactams into body fluids was investigated by examining the distribution of amoxicillin, ceftriaxone, clavulanic acid, temocillin, and ticarcillin into rabbit peripheral lymph after intravenous administration. The elimination half-lives in rabbit plasma varied between 0.34 h (temocillin) and 1.80 h (ceftriaxone), and the half-lives measured in lymph were similar to those in plasma (0.37 to 1.76 h). The percent penetration (area under the concentration-time curve in lymph/area under the concentration-time curve in plasma x 100) was high for amoxicillin (97.6%), temocillin (89.4%), and clavulanic acid (90.8%) but was lower for ticarcillin (76.0%) and for ceftriaxone (67.3%). There was a direct correlation between plasma protein binding and percent penetration. Correction for plasma and tissue binding increased the percent penetration for all compounds, and figures approached 100%. The results presented demonstrate the use of this model to examine the relationships between plasma pharmacokinetics, protein binding, and distribution of antibiotics.

The influence of uptake from the gastrointestinal tract and first-pass effect on oral bioavailability of (Z)-alkyloxyimino penicillins.

We have investigated the contribution of uptake from the gastrointestinal tract and first-pass effect to the poor oral bioavailability of a series of (Z)-alkyloxyimino penicillins in mice. Investigative studies in gut sacs and perfused small intestine demonstrated that these penicillins were able to pass across the mucosal epithelium although to a lesser extent than amoxycillin and cyclacillin, both of which exhibit excellent oral bioavailability in man and animals. In the jejunal gut sacs the mucosal to serosal flux for BRL 44154 was approximately half that of amoxycillin and four times less than that of cyclacillin, and for all, uptake was pH dependent. The serosal to mucosal fluxes were however similar for these compounds and significantly lower than mucosal to serosal fluxes, suggesting involvement of carrier mechanisms in uptake from the mucosal surface. The order of results for the alkyloxyimino penicillins paralleled that observed for oral bioavailability in the mouse. For the alkyloxyimino penicillins, between 5.5 and 9.9% was taken up from the perfused intestine, values which were significantly less than those for amoxycillin (13.2%) and cyclacillin (33.3%). However, uptake was concentration-dependent for BRL 44154 as it was for amoxycillin, thus confirming the possible use of carrier mechanisms in absorption. These observations suggest that the poor peripheral blood concentrations of the alkyloxyimino penicillins achieved after oral dosing were not a consequence of the inability of the compounds to cross the mucosal epithelium. The biliary clearance of the alkyloxyimino penicillins was, however, considerably greater than for amoxycillin and cyclacillin, a finding which may well have been a contributory factor to the comparatively low peripheral concentrations of BRL 44154 and its analogues achieved after oral administration.

Structure-activity relationships within a series of C(7)-substitutedoxyiminocephalosporins containing the C(3)-methylaminopyridiniumthiomethyl substituent. Synthesis and biological properties of BRL 57342 and some close analogues.

(6R,7R)-7-[2-(2-Amino-4-thiazolyl)-2-[(Z)-[(S)-carboxy(3,4- dihydroxyphenyl)methyl]oxyimino]acetamido]-3-(1-methylaminopyri dinium-4-thiomethyl)ceph-3-em-4-carboxylate sodium salt (BRL 57342, 1f) combines excellent in vitro antibacterial potency against Gram-positive and Gram-negative bacteria, including P. aeruginosa and Acinetobacter spp., with excellent stability to extended spectrum beta-lactamases. This potency is reflected in in vivo efficacy studies.

Simulation of human serum pharmacokinetics of cefazolin, piperacillin, and BRL 42715 in rats and efficacy against experimental intraperitoneal infections.

Studies were performed to determine the effects of BRL 42715, a potent beta-lactamase inhibitor, on the activity of cefazolin and piperacillin against experimental intraperitoneal infections caused by either Escherichia coli or Serratia marcescens in rats. Compounds were administered to rats as a continuous infusion of an exponentially diluted solution to simulate in rat plasma the concentration-versus-time curves obtained for humans following intravenous bolus administration. A simulated 1-g dose of cefazolin was ineffective in reducing the bacterial counts in blood and peritoneal fluid samples of animals infected with S. marcescens US20, which produced class Ia beta-lactamase, and as a result, mortality was similar to that of infected controls. Similarly, a simulated 2-g dose of piperacillin was ineffective in reducing bacterial numbers and mortality in animals infected with E. coli 41548, producing a TEM-1 beta-lactamase. However, when the antibiotics were coadministered with BRL 42715, bacterial numbers were reduced significantly and all animals survived at least 16 h after infection. These data demonstrate the ability of BRL 42715 to potentiate the activity of cefazolin and piperacillin against beta-lactamase-producing bacteria that would otherwise be resistant to these antibiotics and illustrate the application of a model to simulate human serum concentrations in conscious rats.

Efficacy of ticarcillin-clavulanic acid for treatment of experimental Staphylococcus aureus endocarditis in rats.

The efficacy of ticarcillin-clavulanic acid was compared with the efficacies of standard antistaphylococcal agents (flucloxacillin, oxacillin, nafcillin, and vancomycin) and ticarcillin in an experimental model of Staphylococcus aureus endocarditis. Therapy was either initiated soon (8 h) after infection, when numbers of bacteria in aortic valve vegetations were relatively low (approximately 6 to 8 log10 CFU/g), or delayed until 24 h after infection, when the vegetations usually contained greater than 9 log10 CFU/g. Doses of the antibiotic were selected to produce peak concentrations in rat serum similar to those achievable in humans after administration of parenteral therapeutic doses. Ticarcillin-clavulanic acid was more effective overall than ticarcillin alone against endocarditis caused by beta-lactamase-producing strains of S. aureus, illustrating the beta-lactamase-inhibitory activity of clavulanic acid in vivo. Ticarcillin-clavulanic acid was as effective as the standard antistaphylococcal beta-lactam agents flucloxacillin, oxacillin, and nafcillin in these infections, whereas vancomycin was generally less active. These results illustrate the clinical potential of ticarcillin-clavulanic acid in the prophylaxis or therapy of severe staphylococcal infections.

Efficacy of amoxycillin/clavulanic acid in experimental Staphylococcus aureus endocarditis in the rat.

The efficacy of amoxycillin/clavulanic acid was compared with that of flucloxacillin, vancomycin and amoxycillin in an experimental model of Staphylococcus aureus endocarditis. Doses of the antibiotics were selected to produce peak concentrations in rat serum similar to those achievable in man after administration of parenteral therapeutic doses. Amoxycillin clavulanic acid was more effective than amoxycillin alone against endocarditis caused by beta-lactamase producing strains of Staph. aureus, illustrating the beta-lactamase inhibitory activity of clavulanic acid in vivo. Amoxycillin/clavulanic acid was as effective as flucloxacillin in these infections whereas vancomycin was generally less active. These results illustrate the clinical potential of amoxycillin/clavulanic acid in the prophylaxis, or in the therapy of severe staphylococcal infections.

Penetration of amoxycillin, ticarcillin and clavulanic acid into lymph after intravenous infusion in rabbits to simulate human serum pharmacokinetics.

The distribution of amoxycillin, ticarcillin and clavulanic acid into lymph collected from the right lymphatic duct of rabbits was examined after intravenous administration. The compounds were administered to simulate, in the plasma of rabbits, the concentrations of amoxycillin, ticarcillin and clavulanic acid measured in human serum after the administration of either an iv bolus dose of amoxycillin 1.0 g plus clavulanic acid 200 mg, ticarcillin 3.0 g plus clavulanic acid 200 mg, or an iv infusion of amoxycillin 2.0 g plus clavulanic acid 200 mg or ticarcillin 3.0 g plus clavulanic acid 200 mg given over 30 min. Lymph concentrations of the compounds reached a peak rapidly after the simulation of a bolus dose (0-1 h) and the concentration-versus-time profiles in plasma and lymph were generally similar after 45 min. Following simulation of an iv infusion, peak concentrations of amoxycillin and clavulanic acid in lymph were reached at approximately the same time as for the bolus simulation, but that of ticarcillin occurred slightly later. The elimination half-lives of the compounds were similar in plasma and lymph. The percentage penetration values were high (greater than 80%) irrespective of the concentration-versus-time curve simulated. The penetration of clavulanic acid was compatible with that of the coadministered penicillin agent and was similar when given with either amoxycillin or ticarcillin.

Simulation of human serum pharmacokinetics of ticarcillin-clavulanic acid and ceftazidime in rabbits, and efficacy against experimental Klebsiella pneumoniae meningitis.

The penetration into cerebrospinal fluid (CSF) and efficacy of ticarcillin-clavulanic acid, ticarcillin alone, and ceftazidime were compared in rabbits with experimentally induced Klebsiella pneumoniae meningitis. The compounds were administered to simulate in rabbit plasma the concentration-versus-time curves observed in humans after 30-min infusions of Timentin (3 g of ticarcillin plus 100 mg of clavulanic acid), ticarcillin (3 g), and ceftazidime (2 g). Single- and multiple-dosing schedules were used. The penetrations of clavulanic acid into CSF (expressed as [area under the concentration-time curve for CSF/area under the curve for plasma] x 100) after the two dosing schedules were 28 and 24.5%, similar to that for ceftazidime (21%; multiple-dosing only) and greater than those for ticarcillin (8.4 and 9.3%). Ticarcillin was ineffective in reducing viable counts in CSF but, in the presence of clavulanic acid, reduced bacterial numbers by approximately 99% at 4 h after a single dose and by 99.99% at 12 h after three doses given at 4-h intervals. Two doses of ceftazidime given 8 h apart were more effective than the three doses of ticarcillin-clavulanic acid, in keeping with the in vitro activities of these compounds against the infecting organism. These results illustrate the ability of clavulanic acid to penetrate the blood-CSF barrier such that concentrations of the inhibitor in CSF potentiate the activity of ticarcillin against the ticarcillin-resistant, beta-lactamase-producing strain of K. pneumoniae.

Temocillin efficacy in experimental Klebsiella pneumoniae meningitis after infusion into rabbit plasma to simulate antibiotic concentrations in human serum.

An infusion system was developed to simulate in the plasma of rabbits the concentrations of temocillin in human serum measured after administration of a 2-g intravenous bolus dose. The efficacy of therapy with this infusion against experimental Klebsiella pneumoniae meningitis was compared with that of a conventional bolus dose to the animals. The marked difference between the elimination half-life (t1/2) of temocillin in rabbit plasma and human serum (0.3 and 5 h, respectively) was reflected in concentrations in cerebrospinal fluid (CSF). The mean peak concentration after infusion occurred 3.5 h later than after bolus dosing, and levels were more prolonged (t1/2 in CSF was 6.3 h compared with 0.83 h following the bolus dose). After infusion, the mean viable count in CSF decreased by 4 log10 CFU/ml, whereas the bolus dose was ineffective because of the rapid fall to subinhibitory concentrations. These results suggest that the infusion system used is valuable for experimental studies with antibacterial agents whose elimination kinetics differ markedly between animals and humans.

Pharmacokinetics and distribution of ticarcillin-clavulanic acid (Timentin) in experimental animals.

The pharmacokinetics and distribution of ticarcillin and clavulanic acid were studied in rats and rabbits after intravenous coadministration of the compounds. The elimination half-lives for ticarcillin and clavulanic acid were similar in both rats (ticarcillin, 0.22 h; clavulanic acid, 0.24 h) and rabbits (ticarcillin, 0.38 h; clavulanic acid, 0.31 h). Both compounds distributed widely throughout rat tissues, and the patterns of distribution were similar to those observed for other beta-lactams. Values for penetration into rat pleural, peritoneal, and subcutaneous fluids calculated from the equation (AUCfluid/AUCserum) X 100, where AUC is the area under the concentration-time curve, were between 83 and 93% for ticarcillin and 86 and 103% for clavulanic acid. Values for penetration into tissue cages in rabbits were 139% +/- 45% for ticarcillin and 109% +/- 22% for clavulanic acid. The penetration of clavulanic acid into rabbit cerebrospinal fluid was higher (P less than 0.05) (4.0% +/- 0.61%) than that of ticarcillin (1.3% +/- 0.53%). Overall, the results show that ticarcillin and clavulanic acid distribute readily throughout body tissues and fluids and predict that the penicillin and beta-lactamase inhibitor would be present together at sites of infection.

Antibacterial activity of antibiotics in acrylic bone cement.

The release of various penicillins and other antibiotics from two brands of polymerised bone cement has been studied in vitro and in vivo in mice. Bone cement plugs containing antibiotics demonstrated antibacterial activity as a result of diffusion of antibiotic from the plugs into the surrounding medium. With all antibiotics tested, from 2-5 to 10 per cent of the antibiotic in the cement was released in vitro in active form within twenty-four hours. Most of the activity appeared within three hours of the start of the test, but in some cases low levels of activity were detected after four days. Antibiotic cement plugs implanted in mice and rats produced low concentrations of antibiotic in the blood up to two hours after implantation, but activity was seldom detected subsequently. In general, penicillins and non-penicillin antibiotics showed similar diffusion characteristics, and the pattern of release in vitro and in vivo was consistent with the leaching of antibiotic from, or near, the surface of the bone cement.