Design of potent dicyclic (1-5/4-10) gonadotropin releasing hormone (GnRH) antagonists

In three earlier papers, the structures and biological potencies of numerous mono- and dicyclic antagonists of GnRH were reported. Among these, two families, each containing two to four members were identified that had very high antagonist potencies in an antiovulatory assay (within a factor of 2 of those of the most potent linear analogues) and high affinities (K(i) < 0.5 nM) for the rat GnRH receptor (rGnRHR). The most favored cycles bridged the side chains of residues (4-10),(1,2) (5-8),(2) (4-10/5-8),(2) (1-3),(3) and (1-3/4-10).(3) Our goal was to identify a consensus model of bioactive conformations of GnRH antagonists, yet these biocompatible constraints did not sufficiently restrain the spatial location of the N-terminal tripeptide with respect to the C-terminal heptapeptide, due largely to the rotational freedom about the bonds connecting these regions. Examination of models derived from NMR studies of cyclo(4-10) analogues suggested a large number of possible cyclic constraints such as cyclo (0-8), (1-8), or (2-8). All analogues tested with these substitutions were inactive as antiovulatory agents at 1 mg/rat (5-9) and had low affinity for rGnRHR. On the other hand, bridging positions 3 and 8 with a [DAsp(3)] to [Dbu(8)] (12, K(i) = 13 nM) or [Orn(8)] (13, K(i) = 14 nM) in the parent compound cyclo(3-8)[Ac-DNal(1),DCpa(2),DXaa(3), Arg(5),DNal(6),Xbb(8),DAla(10)]GnRH yielded analogues that blocked ovulation at 250 microgram/rat. Analogue 14 (K(i) = 2.3 nM), with a [DAsp(3), Lys(8)] bridge, was fully active at 50 microgram/rat. Loss of potency (>20-fold) was observed with the substitution of [DAsp(3)] in 14 by [DGlu(3)] in 15 (K(i) = 23 nM). Dicyclic analogues possessing the (4-10) cycle and selected (1-6), (2-6), and (2-8) cycles led to analogues that were inactive at doses of 500 microgram/rat or larger. Two analogues with (1-8/4-10) cycles (16, K(i) = 1.1 nM) or (3-8/4-10) cycles (22, K(i) = 17 nM) showed full antiovulatory potency at 250 microgram/rat. None of these substitutions yielded analogues potent enough (>80% inhibition of ovulation at 5 microgram/rat or less and K(i) < 0.5 nM) to be candidates for structural analysis by NMR. On the other hand, four dicyclic (1, 1'-5/4-10) analogues met this criterion: dicyclo(1, 1'-5/4-10)[Ac-Asp(1)(Gly),DCpa(2),DTrp(3),Asp(4),Dbu(5 ), DNal(6), Dpr(10)]GnRH (32, K(i) = 0.22 nM), dicyclo(1, 1'-5/4-10)[Ac-Asp(1)(Gly),DCpa(2),DNal(3),Asp(4),Dbu(5 ), DNal(6), Dpr(10)]GnRH (34, K(i) = 0.38 nM), dicyclo(1, 1'-5/4-10)[Ac-Asp(1)(betaAla),DCpa(2), DTrp(3),Asp(4),Dbu(5),DNal(6), Dpr(10)]GnRH (40, K(i) = 0.15 nM), and dicyclo(1, 1'-5/4-10)[Ac-Glu(1)(Gly), DCpa(2),DTrp(3),Asp(4),Dbu(5),DNal(6), Dpr(10)]GnRH (41, K(i) = 0.24 nM). Since they differed slightly in terms of the (1,1'-5) bridge length (21 and 22 atoms) and bridgehead configuration, we may hypothesize that they assume similar bioactive conformations that satisfy a very discriminating receptor, since many other closely related analogues were significantly less potent.