Key points are not available for this paper at this time.
A concise modular laboratory construction of the epothilone class of promising antitumor agents has been accomplished. For the first time in the epothilone area, the new synthesis exploits the power of ring-closing olefin metathesis (RCM) in a stereospecific way. Previous attempts at applying RCM to epothilone syntheses have been repeatedly plagued by complete lack of stereocontrol in the generation of the desired 12,13-olefin geometry in the products. The isolation of epothilone 490 (3) prompted us to reevaluate the utility of the RCM procedure for fashioning the 10,11-olefin, with the Z-12,13-olefin geometry already in place. Olefin metathesis of the triene substrate 12 afforded the product diene macrolide in stereoselective fashion. For purposes of greater synthetic convergency, the C3-(S)-alcohol was fashioned late in the synthesis, using chiral titanium-mediated aldol conditions with the entire O-alkyl fragment as a C15 acetate as the enolate component. Examination of the effects of protecting groups on the RCM process showed that deprotection of the C7 alcohol has a beneficial effect on the reaction yield. Performing the RCM as the last synthetic step in the sequence afforded a 64% yield of only the desired E-olefin. Selective diimide reduction of the new 10,11-olefin yielded 12,13-desoxyepothilone B, our current clinical candidate, demonstrating the utility of this new RCM-reduction protocol in efficiently generating the epothilone framework. Furthermore, the new olefin was selectively funtionalized to demonstrate the advantage conferred by this route for the construction of new analogues for SAR studies, in cytoxicity and microtubule affinity screens. Also described is the surprisingly poor in vivo performance of epothilone 490 in xenografts in the light of very promising in vitro data. This disappointing outcome was traced to unfavorable pharmacokinetic features of the drug in murine plasma. By the pharmacokinetic criteria, the prognosis for the effectiveness of 3 in humans is, in principle, much more promising.
Biswas et al. (Sat,) studied this question.