The Role of the Reduced-Folate Carrier and Metabolism to Intracellular Polyglutamates for the Activity of ICI D1694

Abstract

ICI D1694 is a quinazoline inhibitor of thymidylate synthase (TS) that was developed to act, at least in part, via metabolism to polyglutamate forms ¹. The original rationale was based on the fact that CB3717, which was active in clinical studies, formed intracellular polyglutamates ². These polyglutamates were ~ 100-fold better as inhibitors of TS and were not readily effluxed from the cell. Provided that the life-threatening and dose-limiting nephrotoxicity of CB3717 (see Clarke et al; this volume) was lost it was felt that a useful antitumour agent could be made. As the nephrotoxicity was thought to be due to the poor water-solubility of CB3717, particularly in acid urine, compounds with increased solubility were synthesised (see Clarke et al, this volume). Many of these compounds also had increased cytotoxicity but were generally poorer inhibitors of isolated TS and it was realised that a number of factors were responsible for these observations. Studies suggested that, unlike CB3717, many of the new analogues utilised the reduced-folate/MTX cell membrane transport carrier (RFC) which improved the rate of uptake and, for those that were substrates for folylpolyglutamate synthetase (FPGS), improved the rate of metabolism to polyglutamate forms (e.g. 2-desamino-2-methyl-N ¹⁰-propargyl-5,8-dideazafolate; ICI 198583)³’⁴. The combination of good transport via this carrier and very much improved FPGS substrate activity was a feature of a series of benzoyl ring modified quinazolines (5-membered rings) where the N ¹⁰-substituent was usually methyl ⁵. These compounds, although ~20-fold poorer than CB3717 as TS inhibitors were up to 500-fold more cytotoxic. In mice, improved potency over CB3717 was maintained by either single bolus or repeated once daily bolus administration (depending on the mouse model used, see Clarke et al, this volume) even though clearance of the compounds from the plasma was more rapid than CB3717 (βt¹/₂ of ~30 mins) ⁶ (unpublished observations; D.R. Newell and D.I. Jodrell).