Rat

Corticotropin releasing factor (CRF) is a 41-peptide amide which stimulates the release of ACTH. (Vale et al. Science 1981,213,1394). CRF has been pos...
14 downloads 0 Views 2MB Size
Download PDF
J . Med. Chem. 1994,37, 1450-1459

1450

Conformationally Restricted Competitive Antagonists of Human/Rat Corticotropin-Releasing Factors Antonio Miranda, Steven C. Koerber, Jozsef Gulyas, Sabine L. Lahrichi, A. Grey Craig, Anne Corrigan, Arnold Hagler,' Catherine Rivier, Wylie Vale, and Jean Rivier' The Clayton Foundation Laboratories for Peptide Biology, The Salk Institute for Biological Studies, 10010 North Torrey Pines Road, La Jolla, California 92037

Received March 11, 19930

Corticotropin releasing factor (CRF) is a 41-peptide amide which stimulates the release of ACTH (Vale et al. Science 1981,213,1394). CRF has been postulated to assume an a-helical conformation upon binding to its pituitary receptor (Hernandez et al. J. Med. Chem. 1993,36,2860). We have exploited this hypothesis in the design of a limited series of cyclic analogues and have taken into consideration the effects of side-chain deletion (Alanine scan, Kornreich et al. J.Med. Chem. 1992, 35,1870) as well as of changes in chirality (Rivier et al. J. Med. Chem. 1993,36,2851), with the rationale that side chains necessary for binding could also be replaced by side-chain bridges. In particular, we have used computer modeling to predict likely side chain bridging opportunities and evaluated the effects of such replacements by correlating biological results with those derived from CD spectroscopy. We have synthesized 38 monocyclic peptide amides, competitive antagonists of h u m a d r a t CRF, using solid-phase methodology on MBHA resin. After purification by preparative RP-HPLC, the peptides were analyzed by RP-HPLC and capillary zone electrophoresis and characterized by mass spectroscopy and amino acid analysis. CRF antagonists were tested for their ability to interfere with CRF-induced release of ACTH by rat anterior pituitary cells. In most cases, one of the bridge heads was located a t a position where substitution by a D-residue was tolerated (i.e., positions 12 and 20). It has become clear that careful optimization of bridge length and chirality is critical. This is best exemplified by the fact that out of the 38 analogues that were h/r.CRF1M1 and cyclosynthesized and tested, only two, (cyclo(20-23) [~Phe~~,Glu~,Lys~,Nle~~~~l were found to be more potent (3 and 2 times, (20-23) [~Phe~~,Glu~~,0rn~~,Nle~~~~Ih/rCRF12-4~), respectively) than [~Phe~~,Nle~~~~lh/rCRF12-41, the parent compound. Six analogues belonging to two different families were found to be half as potent as the standard, 18had 2-20% of the potency of the standard, and the others were significantly less potent. CD results of all analogues in 50% TFE (a concentration of TFE that induced nearly maximum helicity of [ ~ P h e l ~ , N l e h/rCRFlul) ~~B] suggest that while helicity may be an important factor for CRF analogue recognition, little correlation is found between percent helicity as determined by spectral deconvolution and biological activity

in vitro. Introduction Corticotropin releasing factor (CRF) was originally characterized from ovine hypothalamus by Vale et al. in 1981.1 Since then, homologous CRF sequences from severalother mammalian species have been either isolated t Current address: Biosym Technologies,Inc.,9685 Scranton Rd.,San Diego, CA 92121-2777. t Abbreviations: The abbreviations for the amino acids are in accord with the recommendations of the IUPAC-IUB Joint Commission on Biochemicalnomenclature(Eur.J. Biochem.1984,138,9-37).Thesymbols represent the h o m e r except when indicated otherwise. In addition: AcOH, acetic acid; ACTH, adrenocorticotropin hormone; Aib, aminoisobutyric acid; Boc, tert-butoxycarbonyl; BOP, (benzotriazol-l-yloxy)triddmethylamino)phosphoniumherafluorophosphate;Bzl, benzyl eater; Cia, octadecyl;CD, circulardichroism;CHaCN, acetonitrile;CNS, central nervous system; CRF, corticotropin releasing factor; CZE, capillary zone electrophoresis;Dbu, 2,4-diaminobutyricacid;DCHA, dicyclohexylamine; DCM, dichloromethane; DIC, 1,3-diisopropylcarbodiimide; DIEA, diisopropylethylamine; DMF, dmethylformamide;Dpr, 2,3-diaminopropionic acid; EDT,ethanedithiol;Fmoc, BH-fluorenylmethoxycarbonyl; HF, hydrogen fluoride; HOBt, 1-hydroxybenzotriazole; HPLC, highperformance liquid chromatography;h/rCRF, human/ratCRF; i-PrOH, 2-propanol; LSIMS, liquid secondary ion maw spectrometry; MBHA, 4-methylbenzhydrylamine; MeOH, methanol; Nle, norleucine; NMP, N-methylpyrrolidone;OcHx,cyclohexyleater;O h ,fluorenylmethyl ester; oCRF, ovine CRF; RPHPLC, reversed-phase high-performance liquid chromatography;RT, retention time; SAR, structure-activity relationship;SEM, standard error of the mean; TEA, triethylamine; TEAP 2.25, triethylammonium phosphate with pH adjusted to 2.25; TFA, trifluoroacetic acid; TFE, 2,2,24rifluoroethanol;Toe, p-toluenesulfonyl;2C1Z, 2-chlorobenzyloxycarbonyl; CNS, central nervous system. * Abstract published in Aduance ACS Abstracts, April 15, 1994.

and characterized- or ~loned.~-lO The primary sequence of humanhat CRF is Ser-Glu-Glu-Pro-Pro-Ile-Ser-Leu-

Asp-Leu1O-Thr-Phe-His-Leu-Leu-Arg-Glu-Val-Leu-GlumMet-Ala-Arg-Ala-Glu-Gln-Leu-Ala-Gln-G l n S Asn-Arg-Lys-Leu-Met-Glu-Ile~-Ile-NH~. Also, nonmammalian peptides with homologous sequences have been isolated and characterized (sauvagine,11J2the urot e n ~ i n s , ' ~and J ~ two diuretic peptides isolated from tobacco hornworm, Manduca sextalsJs)or c10ned.l~These peptides were synthesized and tested for their ability to release ACTH and were all found to be equipotent in the in vitro assay reported here.18 Members of the CRF family have been found to have many additional activities resulting in the alteration of the cardiovascular system, behavior, reproduction,gastrointestinalsecretion,intestinalmotility, and These peptides are also present in a large number of tissues.22.m Most importantly, CRF is involved in a wide spectrum of CNS-mediatedeffects which suggest that this peptide plays an important role within the brain, especially during Because of this diversity of action, potent competitive antagonists of CRF have been extremely useful in the study of the physiologic and pathophysiologic role of CRF.2P28 Ultimately, it is hoped that such analogues will play a major role in the management of some stress-related ~tates.27-3~In view of studies suggestingthat each member of the CRF family possessed

0022-2623/94/1837-1450$04.50/00 1994 American Chemical Society

HumanlRat Corticotropin-Releasing Factor Antagonists

Journal of Medicinal Chemistry, 1994, Vol. 37, No. 10 1451

introduction of the last element of the bridge and prior to completion of the whole sequence. After completion of the synthesis, the protected, cyclized peptide was cleaved from the resin in liquid HF in the presence of a scavenger, extracted, and lyophilized. I t is interesting to note that Leu-Leu-Arg-Glu-Met-Leu-Glu-Met-Ala-Lys-Ala-Glu-GlnGlu-Ala-Glu-Gln-Ala-Ala-Leu-Asn-Arg-Leu-Leu-Leu-Gluthe first strategy seemsto have generated purer compounds Glu-Ala-NH2) was designed.24 This analogue proved resulting in higher yields. The alternate reasoning behind important by showing marked differences in its ability to using these two strategies was that a more flexible peptide antagonize the actions of CRF in three in vivo bioassay (Le., shorter in the case of 9-12, 16, 17, 20, 31, 36-37) systems. This suggested the existence of different CRF would be more likely to cyclize easily or cyclization would receptor subtypes.35 SAR studies in our laboratories led be favored under conditions whereby the peptide would to the observation that the introduction of Nle residues assume a more rigid conformation that would favor at positions 21 and 38 generated analogues that were proximity of those side chains to be cyclized. Because of significantlymore potent than the a-helical-CRFg-41.36One the fact that the location of the bridges was not equivalent such antagonist, [ ~ P h e ' ~ , N l h/rCRF12-41, e ~ ~ ~ ~ ~ ] presently or identical in these two sets of experiments, it is difficult used as our standard, is ca. 18 times more potent than to assess whether one or the other strategy is clearly better a-helical-CRFw1 in inhibiting the release of ACTH by than the other. A source of structure ambiguity was rat pituitary cells in culture.37 pointed out to us by one of the referees who suggested the A complementaryapproach to understanding the mechapossibility of @-aspartyllinkage for compounds 33-38 as nism of CRF/receptor interaction would be to investigate a result of activation of the @-carboxylof the Asp residue the effect of structural constraints on activity and bioat position 20 which may proceed through an aspartimide potency. Three specific issues should be addressed: (i) intermediate. At this point two products may be obtained the positions in a sequence where bridgeheads should be (the desired peptide and the corresponding 8-aspartyl introduced, (ii) the chirality of the bridgeheads, and (iii) derivative) with their ratio depending on steric as well as the length of the bridge. Because physicochemical evielectrophilic considerations. The point was particularly dence (CD), NMR,38and SAR studies (Na-methylation well taken in view of our prior observations (unpublished and C a - m e t h y l a t i ~ n ~suggest ~) that CRF assumes an results in collaboration with L. Gierasch and J. Rizo) that a-helical structure in solution or when interacting with its @-aspartylpeptides readily formed during the synthesis receptor, we have searched for structural constraints that of constrained GnRH analogues as demonstrated by NMR would unequivocally lock such structural motifs into the (manuscript in preparation). In the case of GnRH overall structure. Side chain-side chain bridging (either analogues, however, this side reaction was never prevalent, through disulfide or lactam bridges) offers such an although 5050 mixes of the desired product and its isomer opportunity. Whereas Felix et al.40 took advantage of could be seen in a few instances. In all GnRH analogues, putative salt bridges as the rationale for the location of the two isomers were readily separable by RP-HPLC. bridgeheads in GRF, we independently introduced bridgeAnalogues 33-38 were submitted to 14 Edman cycles in heads in GRF at positions where (at least for one of the order to exclude this possibility. Results clearly indicated bridgeheads) introduction of a D-residue increased POthat we had isolated the desired peptides. Because of the t e n ~ y Here, . ~ ~ we report: (a) the hypotheses that led to complexity of the crude mixes, it is impossible to conclude the synthesis of 38 constrained analogues of human/ whether this side reaction did or did not occur to any ratCRF12-41, (b) their relative biological potencies in significant level in any of the cases presented here. The inhibiting ACTH release from pituitary cells in culture, fact that this reaction was very significant in dicyclic but and ( c )the conclusions that we can draw as to the nature not in monocyclic GnRH analogues, whereby the Asp of CRF's bioactive conformation based on CD studies and containing cycle was formed last, suggests that the molecular modeling. Preliminary results were presented supplemental constraints were responsible for that unat the 12th American Peptide Symp~sium.~' desired reaction. Crude peptides were purified by preparative reverse-phase HPLC (RP-HPLC) and obtained Results as the TFA salts. Compounds were subjected to several Peptides Synthesis, Purification, and CharacterHPLC purification steps using different solvent systems." ization. Stepwise synthesis of the cyclic peptides was Final preparations of the analogues (yields ranged from carried out using manual solid-phase methodology on 3 to 8% of theory based on resin substitution) were methylbenzhydrylamine-resin(MBHA-resin)Qwith tertanalyzed for purity using RP-HPLC and CZE (purity as butoxycarbonyl (Bod-protected amino acids and with 1,3shown in Table 1)and characterized using optical rotation, diisopropylcarbodiimide (DIC) as coupling agent. OFml mass spectrometry (all in Table l ) , and amino acid analysis Fmoc protecting groups were used for the orthogonal (not shown). Noteworthy was the observation that in protection of the' amino acid side chains used as bridge contradistinction with what had been observed with the heads. Recouplings and cyclization on the resin were done introduction of a D-aminO acid,45 the optical rotations of using (benzotriazol-l-yloxy)tris(dimethylamino)phosphothe peptides were significantly influenced by the intronium hexafluorophosphate (BOP)43in the presence of duction of structural constraints (internal bridges). excess diisopropylethylamine (DIEA). Peptides containCircular Dichroism Spectra. The CD spectrum of ing cysteine residues were air-oxidized after HF deprothe standard, [ ~ P h e ' ~ , N l eh/rCRF12-41, ~ ~ l ~ ~ ] was investitection and cleavage from the resin. Lactam formation gated as a function of the concentration of TFE. As shown for the other cyclic analogues used essentially two stratein Figure 1, even in the absence of TFE this peptide gies. In the first strategy (5-8,13-15,18,19,22-30,32-35, assumes a partially helical structure (20 % ) as judged by 38) cyclization was effected upon complete synthesis of spectral deconvolution using the method of Yang et al.4e the peptide. In the second strategy (9-12, 16, 17,20,31, The theoretical helical content at varying concentrations 36-37) cyclization was achieved immediately after the

considerable amphiphilic a-helical stretches,12*33?34 Rivier et al.24built a chimeric molecule which was found to be 3 times as potent as CRF. On the basis of this structure, an antagonist, a-helical-CRFwl (Asp-Leu-Thr-Phe-His-

Miranda et al.

1452 Journal of Medicinal Chemistry, 1994,Vol. 37, No.10

Table 1. Percent Purity Determined by HPLC, Isocratic Retention Times, Capillary Zone Electrophoresis, Specific Rotations, Mass

Spectrometry Data, and Relative Potency for the CRF Cyclic Antagonists HPLC

compound

no.

(%)a

is0 RT at CZE [a]=D calcd obsd % CH&N* (%)c (deg)d mass (Da)e (m/z)'

98 99 99 98 95 98 98 95 95 95 90 91 95 9% 9% 100 99 95 98# 95 98 99 96 98 98 98 98# 981 9% 95 98 98 971 95 981 99 88 981

3.8 at 35.4 3.7 at 31.8 4.0 at 31.8 4.5 at 33.0 3.1 at 31.8 3.6 at 36.6 3.7 at 30.0 4.6 at 36.6 4.2at 37.2 6.1at 34.3 4.7at 33.6 3.7at 33.0 4.4at 34.2 3.2 at 36.0 3.9 at 36.6 3.5 at 37.8 4.0 at 34.2 4.0 at 31.8 3.8 at 39.0 3.0at 32.3 3.9 at 30.0 4.3at 38.4 4.6at 36.0 3.6 at 36.0 4.2at 36.6 3.6 at 34.8 3.6 at 40.8 3.9 at 33.6 3.4at 35.4 3.3at 34.8 4.0at 34.8 3.0at 35.4 4.7at 39.0 3.0 at 39.0 3.3at 34.2 3.5 at 34.8 3.6 at 33.6 3.8 at 36.6

99 97 99 98 99 98 99 97 95 96 93 91 95 99 99 99 99 99 99 94 99 99 99 99 98 99 97 99 99 99 98 98 98 97 98 99 85 98

-16 -8 -38 -24 -52 -24 -66 -53 -38 -65 -69 -65 -30: -11:

-21: -41 -42 -55 -381 -53 -278 -46 -51 -41 -49: -65 -64: -50: -4Ot -65 -66 -67 -66 -51 -58 -66 -70 -68

3481.86 3481.86 3481.86 3481.86 3516.97 3473.97 3535.01 3491.98 3477.98 3477.98 3463.97 3463.97 3491.98 3492.00 3492.00 3477.98 3463.97 3519.06 3521.20' 3477.01 3483.93 3519.06 3577.06 3492.00 3492.00 3492.00 3492.00 3492.00 3492.00 3477.98 3449.95 3477.98 3480.10* 3480.10* 3463.97 3445.95 3435.93 3463.97

in vitro

antagonistf

IA8

3481.4 0.015(0.008-0.026) 5 3481.4 0.003 (0.001-0.005) 9 3482.0 0.025(0.009-0.063) ND 3481.6 0.006 (0.002-0.014) ND 0.006 (0.002-0.014) 10 3517.1 0.16 (0.08-0.32) 8 3473.9 4 3535.3