Synthesis of alkoxy-substituted diaryl compounds and correlation of ring separation with inhibition of tubulin polymerization: differential enhancement of inhibitory effects under suboptimal polymerization reaction conditions.

Abstract

A number of cytostatic compounds (2-4, 7, and 8), which can be described as "diaryl", inhibit tubulin polymerization, cause cells to accumulate in mitotic arrest, and competitively inhibit the binding of colchicine to tubulin. They differ, however, in the separation of the two aryl moieties. To attempt to understand this variability we prepared a series of analogues modeled on 3 and 4 ("benzodioxole series") and on 7 and 8 ("combretastatin series") which differed only in the number of methylene units (ranging from none to four) separating the aryl moieties. These compounds were evaluated for their effects on tubulin polymerization, colchicine binding, and the growth of L1210 murine leukemia cells. In terms of inhibitory effects on tubulin polymerization, for the combretastatin series there was an optimal separation of the two phenyl rings by a two-carbon bridge (compound 24), with progressively decreasing inhibitory activity when the separation was by one carbon (20), three carbons (25), or four carbons (28) (the biphenyl analogue 16 was inactive). The benzodioxole series, however, did not permit us to generalize this finding, because the least active agents prepared (39 and 40) had a two-carbon bridge, while those with one- (5 and 6) and three-carbon (46 and 47) bridges were nearly equivalent in potency. Submicromolar IC50 values for inhibition of L1210 cell growth were only obtained for compounds 20 (IC50, 0.2 microM), 24 (0.07 microM), and 25 (0.4 microM). While evaluating the effects of these agents on tubulin polymerization, we noted with the combretastatin series and with several standard agents that apparent potency (in terms of IC50 values) was always lower if the reaction was performed at 30 degrees C, with 0.25 mM MgCl2, than at 37 degrees C, with 1.0 mM MgCl2. This enhancement of IC50 values in the former system as compared with the latter was particularly dramatic for the less active agents (e.g., 28) as compared with the more active (e.g. 24).