Decomposition Pathways of the Mono- and Bis(Pivaloyloxymethyl) Esters of Azidothymidine 5′-Monophosphate in Cell Extract and in Tissue Culture Medium: An Application of the ‘on-line ISRP-Cleaning’ HPLC Technique

Abstract

Bis(pivaloyloxymethyl) azidothymidine 5′-monophosphate (piv2-AZTMP) was designed as a cell membrane-permeable precursor of AZTMP. We have reported previously that when incubated with CEM cells deficient in thymidine kinase, piv2-AZTMP gives rise to intracellular AZTMP and the corresponding diphosphate (AZTDP) and triphosphate (AZTTP). Under similar conditions, no intracellular nucleotides were formed with AZT. However, the mechanism by which piv2-AZTMP is converted to AZTMP has not been established. To address this question, we have used the recently developed ‘on-line ISRP-cleaning’ HPLC technique to investigate the stability and metabolic fate of piv2-AZTMP (1) in RPMI 1640 medium, (2) in RPMI containing 10% heat-inactivated fetal calf serum, and (3) in CEM cell extracts. Similar studies were conducted starting with mono(pivaloyloxymethyl) azidothymidine 5′-monophosphate (piv2 AZTMP). From the kinetics of these reactions, it appears that piv2-AZTMP is slowly hydrolyzed to piv1-AZTMP in RPMI and that the metabolic sequence in cell extract and in tissue culture medium is clearly: piv2-AZTMP→ piv1AZTMP→ AZTMP→ AZT. The rate constants are quite different in these three media. Although it is evident that the first step in the metabolism of piv2-AZTMP is catalysed by carboxylate esterase, the enzyme(s) responsible for the second step, piv1-AZTMP→ AZTMP, is less apparent, as carboxylate esterases and/or phosphodiesterases can be taken in account. However, analysis of the kinetic data strongly suggests that carboxylate esterase does not play a significant role and that this second step is mediated by phosphodiesterases. Collectively, these studies demonstrate that piv2-AZTMP is an effective prodrug of AZTMP. They also establish that prv1-AZTMP is an intermediate in this process, and define the sequence of the overall metabolic reaction. With this increased understanding of the metabolism of piv2-AZTMP, it should be possible rationally to design analogues with optimal structural and pharmacological properties for use in vivo.