Appendix B12). Some of the similarities and differences are discussed below.

Antiprogestins bind to progesterone-receptor preparations from a variety of species, including rat, rabbit, calf, marmoset, bonnet monkey, and human, but they do not bind to chicken or hamster progesterone receptors (Baulieu, Appendix B1; Weigel, Appendix B2; Van Look and von Hertzen, Appendix B12). In the latter cases, the lack of mifepristone (RU 486) binding to the progesterone receptor has been attributed to a single amino acid change in the hormone-binding domain—the replacement of a glycine by a cysteine at positions 575 and 722 for the chicken and hamster, respectively. All receptors that bind mifepristone, including glucocorticoid and androgen receptors, have a glycine residue at a corresponding position in the hormone-binding domain (Van Look and von Hertzen, Appendix B12). Both the chicken and the hamster progesterone receptors can be modified by a substitution of glycine (but not methionine or leucine) for this key cysteine, and the mutated receptor will bind mifepristone. Conversely, replacing the key glycine in the human progesterone receptor with cysteine renders the receptor incapable of binding mifepristone (Baulieu, Appendix B1). This exquisite sensitivity of progestin-receptor binding to a single amino acid substitution suggests that studies of the molecular actions of antagonists that have potential clinical applications should be conducted by using human steroid receptors, since receptors from other species may respond somewhat differently (Weigel, Appendix B2).

When using animal models, it is important to recognize that there are marked species differences in plasma proteins that can bind to the steroid hormones and to the antihormones. For example, no animal species appears to have the high-affinity binding protein a1-acid glycoprotein, which is found in the human (Baulieu, Appendix B1; Van Look and von Hertzen, Appendix B12). This plasma glycoprotein binds to some, but not all, antiprogestins, and appears to affect clearance rates of these compounds. For example, a1-acid glycoprotein strongly binds to mifepristone and probably lilopristone, but not to onapristone. Research suggests that the low clearance of mifepristone is exacerbated by this tight protein binding and is reflected in the long half-life of mifepristone (20 to 24 hours) versus that for onapristone (2 to 4 hours) (Van Look and von Hertzen, Appendix B12).

The pregnant guinea pig model has been used extensively for the study of abortifacient potency and the mechanism of action of antiprogestins. Studies using this model led to the development and testing of the sequential treatment regimen of mifepristone followed by prostaglandin. However, studies in the guinea pig model have not always correlated with those in humans. For example, in the pregnant guinea pig model, a marked synergism was demonstrated between antipro-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement