Analogues of estradiol-17 beta (E2) have been evaluated for estrogen receptor (ER) binding affinity and mitogenic potential in the human breast cancer cell line MCF-7. These 42 compounds represent subtle modifications of the natural estrogen structure through the placement of hydroxyl, amino, nitro, or iodo groups around the ring system in addition to, or as replacement of, the 3- and 17 beta-hydroxyls of E2. The mitogenic activity of the analogues was found to be related to ER binding only to a limited extent. In order to elucidate structural features that are uniquely responsible for receptor binding affinity or mitogen potential of estrogens, the three-dimensional quantitative structure-activity (QSAR) method Comparative Molecular Field Analysis (CoMFA) was employed. Separate CoMFA models for receptor binding and cell growth stimulation were optimized through the use of various alignment rules and region step size. Whereas the CoMFA contour plots did outline the shared structural requirements for the two measured biological properties, specific topological features in this set of estrogens were delineated that distinguish mitogenic potential from ER binding ability. In particular, steric interference zones which affected growth extend in a band from above the A-ring to position 4 and below, whereas the ER binding steric interference zones are limited to isolated polyhedra in the 1, 2 and 4 positions and the alpha face of the B-ring. In addition, electronegative features located around the A-, B-, or C-rings contribute to receptor affinity. However, growth is dependent only on electronegative and electropositive properties near the 3-position. In a final QSAR model for the mitogenic response, the value of ER binding was included along with structural features as a descriptor in CoMFA. The resulting 3D-QSAR has the most predictive potential of the models in this study and can be considered a prototype model for the general evaluation of a steroidal estrogen's growth stimulating ability in MCF-7 cells. For example, the location of D-ring contours illustrate the model's preference for 17 beta-hydroxy steroids over the less mitogenic 17 alpha- and 16 alpha-hydroxy compounds. In addition, the enhanced mitogenic effect of steric bulk in the 11 alpha-position is also evident. The QSAR studies in this report illustrate the fact that while ER binding may be a required factor of the estrogen dependent growth response in MCF-7 cells, particular structural characteristics, in addition to those responsible for tight receptor binding, must be present to induce an optimal mitogenic response. Therefore, this report demonstrates that the CoMFA QSAR method can be utilized to characterize structural features of test compounds that account for different types of estrogenic responses.

2-(4-[(7-Chloro-2-quinoxalinyl)oxy]phenoxy)propionic acid (XK469) is among the most highly and broadly active antitumor agents to have been evaluated in our laboratories and is currently scheduled to enter clinical trials in 2001. The mechanism or mechanisms of action of XK469 remain to be elaborated. Accordingly, an effort was initiated to establish a pharmacophore hypothesis to delineate the requirements of the active site, via a comprehensive program of synthesis of analogues of XK469 and evaluation of the effects of structural modification(s) on solid tumor activity. The strategy formulated chose to dissect the two-dimensional parent structure into three regions-I, ring A of quinoxaline; II, the hydroquinone connector linkage; and III, the lactic acid moiety-to determine the resultant in vitro and in vivo effects of chemical alterations in each region. Neither the A-ring unsubstituted nor the B-ring 3-chloro-regioisomer of XK469 showed antitumor activity. The modulating antitumor effect(s) of substituents of differing electronegativities, located at the several sites comprising the A-ring of region I, were next ascertained. Thus, a halogen substituent, located at the 7-position of a 2-(4-[(2-quinoxalinyl)oxy]phenoxy)propionic acid, generated the most highly and broadly active antitumor agents. A methyl, methoxy, or an azido substituent at this site generated a much less active structure, whereas 5-, 6-, 8-chloro-, 6-, 7-nitro, and 7-amino derivatives all proved to be essentially inactive. When the connector linkage (region II) of 1 was changed from that of a hydroquinone to either a resorcinol or a catechol derivative, all antitumor activity was lost. Of the carboxylic acid derivatives of XK469 (region III), i.e., CONH2, CONHCH3, CON(CH3)2, CONHOH, CONHNH2, CN, or CN4H (tetrazole), only the monomethyl- and N,N-dimethylamides proved to be active.

XK469 (1) is among the most highly and broadly active antitumor agents to have been evaluated in our laboratories. Subsequent developmental studies led to the entry of (R)-(+) 1 (NSC 698215) into phase 1 clinical trials (NIH UO1-CA62487). The antitumor mechanism of action of 1 remains to be elucidated, which has prompted a sustained effort to elaborate a pharmacophoric pattern of 1. The present study focused on a strategy of synthesis and biological evaluation of topologically based, bioisosteric replacements of the quinoxaline moiety in the lead compound (1) by quinazoline (4a-d), 1,2,4-benzotriazine (12a-18b), and quinoline (21a-g) ring systems. The synthetic approach to each of the bioisosteres of 1 utilized the methodology developed in previous work (see Hazeldine, S. T.; Polin, L.; Kushner, J.; Paluch, J.; White, K.; Edelstein, M.; Palomino, E.; Corbett, T. H.; Horwitz, J. P. Design, Synthesis, and Biological Evaluation of Analogues of the Antitumor Agent 2-(4-[(7-Chloro-2-quinoxalinyl)oxy]phenoxy)propionic acid (XK469). J. Med. Chem. 2001, 44, 1758-1776.), which is extended to the procurement of the benzoxazole (23a,b), benzthiazole (23c,d), pyridine (25a,b), and pyrazine (27) congeners of 1. Only quinoline analogues, bearing a 7-halo (21a,b,d,e) or a 7-methoxy substituent (21g), showed antitumor activities (Br > Cl > CH(3)O > F approximately I), at levels comparable to or greater than the range of activities manifested by 1 and corresponding analogues. At high individual dosages, the (S)-(-) enantiomers of 1 and 21b,d all produce a reversible slowing of nerve-conduction velocity in the mice, the onset of which is characterized by a distinctive dysfunction of the hind legs, causing uncoordinated movements. The condition resolves within 5-10 min. However, at higher dosages, which approach a lethal level, the behavior extended to the front legs, lasting from 20 min to 1 h. By contrast, the (R)-(+) forms of these same agents did not induce the phenomenon of slowing of nerve-conduction velocity.