Metabolism of carboxylate drugs (primarily in the liver) often yields acyl glucuronides as the major metabolites. These molecules are chemically unstable, to an extent that depends on the shape and chemistry of the drug. They have the capacity to react with many biomolecules, covalently binding to nucleophiles on proteins, lipids and DNA. This chemical modification is clearly of much interest to pharmacologists as a potential determinant of the observed adverse reactions to these drugs.
A comparison has been made (Chapter Three) of the extent and pattern of covalent adduct formation in plasma and livers of rats dosed with the nonsteroidal anti-inflammatory drugs (NSAIDs) zomepirac (ZP) and diflunisal (DF), the hypolipidaemic agent clofibric acid (CA), and the anti-epileptic agent valproic acid (VPA). These drugs form differing amounts of acyl glucuronides with diverse intrinsic reactivities. A comparison of these drugs revealed a very good correlation between acyl glucuronide reactivity and covalent binding to proteins in plasma. Such a comparison was not feasible for covalent binding to liver proteins. Immunoblotting using rabbit polyclonal antisera raised against synthetic drug-protein adducts showed a very similar pattern of modification in livers of the two NSAID-treated rats. Since these experiments, it has been demonstrated by others that several NSAIDs investigated similarly produced the same adduct pattern as DF and ZP. CA and VPA gave dissimilar patterns of lower intensity. In plasma, the antisera specifically detected drug-modified serum albumin in samples from rats treated with all drugs except VPA. The results with this small series of carboxylate drugs suggested that adduct concentrations in plasma but not liver could be related to acyl glucuronide reactivity. The pattern of modification varied from drug to drug, but was similar for the two NSAIDs.
Many experimental systems have been used to acquire data about drug-protein binding in the liver. Models that do not utilise whole animals can offer advantages, e.g., multiple experiments from a single liver. A comparison of the patterns of covalent binding of ZP to proteins in the livers of intact rats, isolated rat hepatocytes (in culture or suspension), and in rat liver homogenates was undertaken (Chapter Four). Rats were dosed with ZP. Isolated hepatocytes were exposed to ZP in tissue culture medium. Liver homogenates were exposed to ZP and ZP acyl glucuronide. Covalent binding of drug was examined by western blotting. In livers from dosed animals, the familiar band pattern (Chapter Three) was again found. Few similarities existed with the results from isolated hepatocytes and, not surprisingly, liver homogenates. Only a high molecular weight band was common to all treatments. Many proteins seemed to be modified, at least to some extent. The differences in major bands are most likely caused by the loss of liver and hepatocyte architecture. The variability across different model systems in respect to covalent binding to hepatic proteins emphasises the need for care in interpretation of results.
Whether covalent adducts were formed with microtubular protein (MTP, 85% a/p-tubulin) upon incubation with an acyl glucuronide, and if this then influenced the ability of tubulin to assemble into microtubules was investigated (Chapter Five). Bovine brain MTP was purified by assembly-disassembly cycles and incubated with ZP, ZP acyl glucuronide and its rearrangement isomers. Assembly was monitored by change in turbidity. Both the acyl glucuronide and its isomer mixture caused dose-dependent inhibition of assembly, while ZP itself caused only modest inhibition. Electron microscopy directly confirmed the turbidometric measurements. In a slightly different system it was shown that polymerised MTP was much less susceptible to the deleterious effects of modification, possibly because important sites for assembly were not available for reaction with glucuronide metabolites. Immunoblotting showed that the acyl glucuronide and its isomers covalently modified MTP in a dose-dependent manner, while ZP itself caused no modification. Furthermore, ZPmodified tubulin was shown to be present in the cytosol of livers from rats dosed twice daily for 3 days with ZP. These data demonstrated a functional change in an important protein caused by modification by a drug metabolite.
The work presented in this thesis demonstrates the covalent binding of acyl glucuronide metabolites to proteins in vivo, the advantages of using intact animals for examining this covalent binding, and a direct functional change in an important protein after covalent modification by an acyl glucuronide.