The goals of new antifolate development are: 1) improved selectivity, 2) improved penetration into pharmacologic sanctuaries, and 3) effectiveness vs. tumors either with intrinsic or acquired resistance to methotrexate (MTX). The major target for antifolate development has been dihydrofolate reductase (DHFR), but other critical folate-dependent enzymes, i.e., thymidylate synthase, methionine synthetase, and folylpolyglutamate synthetase are also important targets for new antifolate development. The possibility that DHFR from tumor tissue differs significantly from normal tissue DHFR now seems improbable, and the ideas of the late Bill Baker to design specific inhibitors of the tumor enzyme vs. the normal tissue DHFR are unlikely to succeed. However, the experience with triazinate (Baker's antifol; TZT) indicates that transport of antifols could be exploited to provide selective toxicity, as well as to provide agents effective vs. MTX-resistant cells. This work led to a second generation of "nonclassical" folate antagonists, of which trimetrexate (JB-11; TMQ) is now in clinical trial. Uptake of TMQ is via an MTX-independent membrane system, and extremely high intracellular levels of this drug are achieved in human leukemia cells.