Design of potent dicyclic (4-10/5-8) gonadotropin releasing hormone (GnRH)antagonists

With the ultimate goal of identifying a consensus bioactive conformation of GnRH antagonists, the compatibility of a number of side chain to side chai n bridges in bioactive analogues was systematically explored. In an earlier publication, cyclo[Asp(4)-Dpr(10)]GnRH antagonists with high potencies in vitro and in vivo had been identified.(1) Independently from Dutta et al.(2 ) and based on structural considerations, the cyclic [Glu(5)-Lys(8)] constr aint was also found to be tolerated in GnRH antagonists. We describe here a large number of cyclic (4-10) and (5-8) and dicyclic (4-10/5-8) GnRH antag onists optimized for affinity to the rat GnRH receptor and in vivo antiovul atory potency. The most potent monocyclic analogues were cyclo(4-10)-[Ac-DN al(1),DFpa(2),DTrp(3),Asp(4),DArg(6),Xaa(10)]GnRH with Xaa = D/LAgl (1, K-i = 1.3 nM) or Dpr (2, K-i = 0.36 nM), which completely blocked ovulation in cycling rats after sc administration of 2.5 mu g at noon of proestrus. Muc h less potent were the closely related analogues with Xaa = Dbu (3, K-i = 1 0 nM) or cyclo(4-10)[Ac-DNa](1),DFpa(2),DTrp(3),Glu(4),DArg(6),D/LAgl(10)]G nRH (4, K-i = 1.3 nM). Cyclo(5-8)[Ac-DNal(1),DCpa(2),DTrp(3),Glu(5),DArg(6) ,Lys(8),DAla(10)] (13), although at least 20 times less potent in the AOA t han 1 or 2 with similar GnRHR affinity (K-i = 0.84 nM), was found to be one of the most potent in a series of closely related cyclo(5-8) analogues wit h different bridge lengths and bridgehead chirality. The very high affinity of cyclo(5,5'-8)[AcDNal(1),DCpa(2),DPal(3),Glu(5)(beta Ala),DAr6,(D or L)A gl,(8)DAla(10)]GnRH 14 (K-i = 0.15 nM) correlates well with its high potenc y in vivo (full inhibition of ovulation at 25 mu g/rat). Dicyclo(4-10/5-8)[ AcDNal(1),DCpa(2),DTrp(3),Asp(4),Glu(5),DArg(6),Lys(8),Dpr(10)]GnRH (24, K- i = 0.32 nM) is one-fourth as potent as 1 or 2, in the AOA; this suggests t hat the introduction of the (4-10) bridge in 13, while having little effect on affinity, restores functional/conformational features favorable for sta bility and distribution. To further increase potency of dicyclic antagonist s, the size and composition of the (5-8) bridge was varied. For example, th e substitution of Xbb(5)' by Gly (30, K-i = 0.16 nM), Sar (31, K-i = 0.20 n M), Phe (32, K-i = 0.23 nM), DPhe (33, K-i = 120 nM), Arg (36, K-i = 0.20 n M), Nal (37, K-i = 4.2 nM), His (38, K-i = 0.10 nM), and Cpa (39, K-i = 0.2 3 nM) in cyclo(4 - 10/5,5'- 8)[Ac-DNal(1),DCpa(2),DPal(3),Asp(4),Glu(5)(Xbb (5')),DArg(6),Dbu,(8)Dpr(10)] GnRH yielded several very high affinity analo gues that were 10, ca. 10, 4, > 200, 1, ca. 4, >2, and 2 times less potent than 1 or 2, respectively. Other scaffolds constrained by disulfide (7, K-i = 2.4 nM; and 8, K-i = 450 nM), cyclo[Glu(5)-Aph(8)] (16, K-i = 20 nM; and 17, K-i = 0.28 nM), or cyclo[Asp(5)-/Glu(5)-/Asp(5)(Gly(5'))-Amp(8)] (19, K-i = 1.3 nM; 22, K-i = 3.3 nM; and 23, K-i = 3.6 nM) bridges yielded analo gues that were less potent in vivo and had a wide range of affinities. The effects on biological activity of substituting DCpa or DFpa at position 2, DPal or DTrp at position 3, and DArg, DNal, or DCit at position 6 are a lso discussed. Interestingly, monocyclo(5 - 8)[Glu(5),DNal(6), Lys(8)]GnRH (18, K-i = 1.0 nM) and dicyclo(4-10/5-8)[Asp(4),Glu(5),DNal(6),Lys(8),Dpr(1 0)]GnRH (28, K-i = 1.2 nM) contain the native N-terminal pGlu-His-Trp- and are antagonists with relatively high affinity but very low antagonist poten cy in vivo, illustrating an earlier observation that structural constraints alone may lead to partial agonism or competitive antagonism. All of these observations suggest very rigorous requirements for ligand/receptor recogni tion and binding as well as a distinct effect of some substitutions on phar macokinetics.