Synthetic peptide derived from alpha-fetoprotein inhibits growth of human breast cancer: investigation of the pharmacophore and synthesis optimization.

Asynthetic peptide that inhibits the growth of estrogen receptor positive (ER+) human breast cancers, growing as xenografts in mice, has been reported. The cyclic 9-mer peptide, cyclo[EMTOVNOGQ], is derived from alpha-fetoprotein (AFP), a safe, naturally occurring human protein produced during pregnancy, which itself has anti-estrogenic and anti-breast cancer activity. To determine the pharmacophore of the peptide, a series of analogs was prepared using solid-phase peptide synthesis. Analogs were screened in a 1-day bioassay, which assessed their ability to inhibit the estrogen-stimulated growth of uterus in immature mice. Deletion of glutamic acid, Glu1, abolished activity of the peptide, but glutamine (Gln) or asparagine (Asn) could be substituted for Glu1 without loss of activity. Methionine (Met2) was replaced with lysine (Lys) or tyrosine (Tyr) with retention of activity. Substitution of Lys for Met2 in the cyclic molecule resulted in a compound with activity comparable with the Met2-containing cyclic molecule, but with a greater than twofold increase in purity and corresponding increase in yield. This Lys analog demonstrated anti-breast cancer activity equivalent to that of the original Met-containing peptide. Therefore, Met2 is not essential for biologic activity and substitution of Lys is synthetically advantageous. Threonine (Thr3) is a nonessential site, and can be substituted with serine (Ser), valine (Val), or alanine (Ala) without significant loss of activity. Hydroxyproline (Hyp), substituted in place of the naturally occurring prolines (Pro4, Pro7), allowed retention of activity and increased stability of the peptide during storage. Replacement of the first Pro (Pro4) with Ser maintains the activity of the peptide, but substitution of Ser for the second Pro (Pro7) abolishes the activity of the peptide. This suggests that the imino acid at residue 7 is important for conformation of the peptide, and the backbone atoms are part of the pharmacophore, but Pro4 is not essential. Valine (Val5) can be substituted only with branched-chain amino acids (isoleucine, leucine or Thr); replacement by d-valine or Ala resulted in loss of biologic activity. Thus, for this site, the bulky branched side chain is essential. Asparagine (Asn6) is essential for activity. Substitution with Gln or aspartic acid (Asp), resulted in reduction of biologic activity. Removal of glycine (Gly8) resulted in a loss of activity but nonconservative substitutions can be made at this site without a loss of activity indicating that it is not part of the pharmacophore. Cyclization of the peptide is facilitated by addition of Gln9, but this residue does not occur in AFP nor is it necessary for activity. Gln9 can be replaced with Asn, resulting in a molecule with similar activity. These data indicate that the pharmacophore of the peptide includes side chains of Val5 and Asn6 and backbone atoms contributed by Thr3, Val5, Asn6, Hyp7 and Gly8. Met2 and Gln9 can be modified or replaced. Glu1 can be replaced with charged amino acids, and is not likely to be part of the binding site of the peptide. The results of this study provide information that will be helpful in the rational modification of cyclo[EMTOVNOGQ] to yield peptide analogs and peptidomimetics with advantages in synthesis, pharmacologic properties, and biologic activity.