Quinolones as gonadotropin releasing hormone (GnRH) antagonists: Simultaneous optimization of the C(3)-aryl and C(6)-substituents

Article abstract of DOI:10.1016/S0960-894X(00)00318-8

A series of 3-arylquinolones was prepared and evaluated for their ability to act as gonadotropin releasing hormone (GnRH) antagonists. A variety of substitution patterns of the 3-aryl substituent are described. The 3,4,5-trimethylphenyl substituent (23h) was found to be optimal. (C) 2000 Elsevier Science Ltd. All rights reserved.

Full text of DOI:10.1016/S0960-894X(00)00318-8

Bioorganic & Medicinal Chemistry Letters 10 (2000) 1723±1727
Quinolones as Gonadotropin Releasing Hormone (GnRH)
Antagonists: Simultaneous Optimization of the C(3)-Aryl
and C(6)-Substituents
Jonathan R. Young,a,* Song X. Huang,a Irene Chen,a Thomas F. Walsh,a
Robert J. DeVita,a Matthew J. Wyvratt, Jr.,a Mark T. Goulet,a Ning Ren,b
Jane Lo,b Yi Tien Yang,b Joel B. Yudkovitz,b Kang Chengb and Roy G. Smithb
aDepartments of Medicinal Chemistry, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
bDepartment of Biochemistry and Physiology, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
Received 24 April 2000; accepted 30 May 2000
AbstractÐA series of 3-arylquinolones was prepared and evaluated for their ability to act as gonadotropin releasing hormone
(GnRH) antagonists. A variety of substitution patterns of the 3-aryl substituent are described. The 3,4,5-trimethylphenyl substituent
(23h) was found to be optimal. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
chlorides2 were generally prepared by Arndt±Eistert3
homologation of benzoic acids 4 followed by exposure to
oxalyl chloride. Treatment of 6 with base a€orded the
Claisen condensation product 7, which upon alkylation
with alcohol 84 under Mitsunobu conditions gave ether
9. Acid catalyzed N-BOC deprotection provided the tar-
get compounds 10. Those compounds with encouraging
in vitro activities were converted to a variety of urea and
amide derivatives. This involved treatment of 9 with a
combination of hydrazine and catalytic Fe(III), followed
by either isocyanate generation and in situ trapping with
amines, or EDC mediated amide formation with car-
boxylic acids. Deprotection of the BOC moiety with TFA
furnished ureas 12 and amides 14, respectively.
Recent reports from these laboratories described the
GnRH activity of a series of 3-arylquinolones.1 These
earlier e€orts detailed modi®cations of the C(4)-sub-
stituent,1a,b culminating in 1 which was 1000 times more
potent than the screening lead. A subsequent communica-
tion detailed studies on the optimization of the C(6)-posi-
tion.1c In this letter, we wish to report on the continued
advancement of the quinolone structure±activity rela-
tionships (SAR) with emphasis on the C(3)- and C(6)-
positions.
To further expand our investigation, we next examined the
incorporation of polar substituents into 1. Thus, radical
bromination (NBS, benzoyl peroxide, CCl4) of 9 under
high-dilution conditions gave a 4:1 mixture of mono-
and di-bromo adducts (15 and 16, respectively) accom-
panied by recovery of some unreacted starting material
(Scheme 2). Treatment of 15 with acetate or cyanide5
produced 10w and 10aa, after removal of the N-BOC
protecting group. Hydrolysis of nitrile 17 to amide 10y
was achieved with basic peroxide.6 Several sequences were
investigated for conversion of 15 to alcohol 18 of which
the most ecient route involved Kornblum7 oxidation
followed by NaBH4 reduction.8 Re-oxidation of 18 with
tetrapropylammonium perruthenate (TPAP) provided the
aldehyde, which after subjection to either alkyltitanium9
Synthesis
The syntheses of 3-arylquinolones 10, 12, and 14 began
with nitration of 5-chloroanthranillic ester 2 followed
by coupling with acid chloride 5 (Scheme 1). The acid
*Corresponding author. Tel.: +1-732-594-3981; e-mail: jonathan_young
@merck.com
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00318-8

1724
J. R. Young et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1723±1727
Scheme 1. Reagents and conditions: (aꢀ) HNO3, H2SO4; (b) (COCl)2,ꢀDMF(cat), CH2Cl2; CH2N2, Et2O, 0 ꢀC; Ag2O, dioxane, H2O; (c) (COCl)2,
.
DMF(cat), CH2Cl2; (d) 3+5, DCE, 80 C; (e) NaHMDS, THF, 0±25 C; (f) 8, Ph3P, DEAD, THF; (g) TFA, CH2Cl2; (h) Cl3Fe 6H2O, hydrazine,
6
7
.
MeOH; Cl2CO, pyridine, CH2Cl2; R NH2; (i) Cl3Fe 6H2O, hydrazine, MeOH; EDC, HOBT, R CO2H, CH2Cl2.
Scheme 2. Reagents and conditions: (a) NBS, benzoyl peroxide, CCl4, 80 ꢀC (9 (32%); 15 (41%); 16 (10%)); (b) NaOAc, 18-C-6, MeCN, 60 ꢀC (38%);
(c) TFA, CH2Cl2; (d) KCN, TFA, DMSO, 10 ꢀC (29%); (e) H2O2, K2CO3, DMSO, 50 ꢀC (76%); (f) 2,4,6-collidine, DMSO, 80 ꢀC; NaBH4, MeOH
(51%); (g) TPAP, NMO, CH2Cl2; TiCl4, MeMgBr, Et2O, 78 ꢀC (41%); (h) TPAP, NMO, CH2Cl2; Et2NH, HOAc, NaCNBH3, MeOH (21%).
reagents or reductive-amination conditions gave 10x
and 10z, respectively.
oxidative cleavage of the derived diol followed by amide
formation as described above.
Concurrent with our e€orts to improve the C(6)-urea sub-
stituent, we made the observation that simpli®ed tethers
such as benzamide derivatives1c or the reversed amides
resulted in improved in vitro activity. During this investi-
gation, 4-aminopyrimidine was found to be superior to
cyclopropylamine (i.e., 12A vs 12hh; Table 1). In addition,
the more potent enantiomer of the piperidine was found to
possess the S-stereochemistry.1c These results prompted a
re-examination of the C(3)-aryl position in this more
potent series. Our synthetic e€orts began with electro-
philic iodination of aniline 2 (Scheme 3) which was then
transformed to the fully elaborated 6-iodoquinolone
(21) in close analogy to 9, as described above. Carboxyl-
ation of iodide 21 under the mild palladium catalyzed
conditions of Cacchi et al,10 subsequent EDC mediated
amide formation and N-BOC deprotection a€orded the
target compounds 23a±h. Acetamides such as 26, were
prepared as potential isosteric replacements for the urea
substituent. The acetamide side chain was masked in the
form of an allyl group which was installed via a Stille
cross-coupling reaction between allyltin and iodide 20.
After Claisen condensation and Mitsunobu alkylation,
the allyl moiety was converted to acetamide 26 by
Biological Results and Discussion11
The lead compound 1 bound with high anity to the rat
GnRH receptor (IC50=10 nM).1c Absence of the
methyl groups (10a) established the importance of meta-
substitution, as the binding anity of this analogue was
reduced 40-fold, although most of the potency could be
restored from a single meta-methyl (10b) or halogen
(10c±e) substituent. The meta-electron withdrawing nitro
group (10f) was also bene®cial albeit to a lesser degree.
Comparison of the binding activities of the regioisomeric
chlorophenyls 10d, 10g, and 10h revealed that the opti-
mum order of activity was meta>para>>ortho. The
bulky t-butyl analogue 10i was nearly as potent as the
sterically smaller chloro analogue 10h suggesting a fairly
large steric latitude in the receptor pocket where the aryl
substituent resides.
We turned our attention towards increasing the substitu-
tion around the aryl ring in order to optimize this hydro-
phobic interaction. The 3,4-dimethyl analogue 10j and
dichloro analogues 10l and 10n were nearly equipotent to

J. R. Young et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1723±1727
1725
Scheme 3. Reagents and conditions: (a) I2, AgSO4, MeOH; (b) 4, (COCl)2, DMF, CH2Cl2; 19, DCE, 80 ꢀC; (c) NaHMDS, THF, 0±25 ꢀC; (d) 8,
Ph3P, DEAD, THF; (e) Cl2Pd(dppf), CO, DMSO, 60 ꢀC; (f) 4-aminopyrimidine, EDC, DMAP, NEt3, CH2Cl2; (g) TFA, CH2Cl2; (h) allyltributyltin,
(Ph3P)2PdCl2, DMF, 95 ꢀC; (i) NaHMDS, THF, 0±25 ꢀC; (j) 8, Ph3P, DEAD, THF; (k) OsO4, NMO, t-BuOH-THF-H2O; (l) Pb(OAc)4, pyridine-
MeOH; (m) 4-aminopyrimidine, EDC, DMAP, NEt3, CH2Cl2; (n) TFA, CH2Cl2.
Table 1.11
Entry
Aryl
rGnRH IC50 (nM)
Entry
Aryl
rGnRH IC50 (nM)
200
Entry
Aryl
rGnRH IC50 (nM)
1
12a
10
1
10i
10r1
12r1
53
68
10a
410
10j
12j1
18
6
10s
220
10b
12ba
27
23
10k
400
10t
900
23
10c
10d
53
40
10l
12l1
29
18
10u
10m
10n
10o
900
24
10v
10w
10x
45
63
71
10e
12e1
40
6
10f
100
71
10g
10h
360
160
10p
10q
40
10y
10z
270
18
600
aR6 for compounds 12 (Scheme 1) is cyclopropyl.
1, which were all 3±4 times more potent than the electron-
withdrawing bis-tri¯uoromethyl compound 10o. Electron-
releasing groups also proved to be detrimental. For
example, sequential replacement of the methyl groups of
10j with a methoxy substituent (10p±q) resulted in a 2-
fold decrease in GnRH binding anity for one methoxy
group and a 10-fold loss for the second. Incorporation of
larger aromatic substituents such as the naphthyl

1726
J. R. Young et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1723±1727
derivative 10r or a biaryl ring system 10m served to
further de®ne the emerging SAR. Replacement of the
phenyl ring with the isosteric thiophene residue showed
a marked preference for attachment at the C(2)-position
(10s vs 10t) in which 10s was modestly more potent
than the simple phenyl (10a) case. Further improve-
ments in this design included the pseudo meta-chloro
analogue 10u whose activity was consistent with the
phenyl SAR.
At this stage of the program, several new developments
occurred that impacted our lead development. First, the
individual enantiomers of the piperidine side chain were
independently prepared and incorporated into our lead
design, wherein it was determined that the S-con®guration
was desired. Secondly, the cloned human GnRH receptor
was now available in both a binding (hGnRH) and
functional (hPI) assay, both of which would now serve
as our primary tools to screen new analogues. Table 2
will illustrate this transition.
At this juncture, a preference for 3,4- or 3,5-dialkyl or
dihalogenated phenyls was established. Although electron-
releasing/withdrawing groups were not bene®cial when
directly attached to the aryl ring, we pondered the e€ect
of these groups if deployed o€ the methyl groups of 1.
The results were quite clear. With the exception of the
cyanomethyl analogue (10z), all other variations reduced
receptor-binding anity (cf. acetate 10v, alcohol 10w,
amide 10x, or amine 10y). In conclusion, the results
from Table 1 suggest that the C(3)-aryl group is occupying
a hydrophobic binding region within the GnRH receptor
in which alkyl or halogen substituents are preferred.
Of the simple alkyl ureas prepared, cyclopropyl was
clearly the most potent and displayed good functional
antagonism in both the rat primary pituitary cell assay
(rLH: IC50=375 nM) and in CHO cells expressing the
human receptor (hPI; IC50=12 nM). The single enantio-
mer of 12a with the S-con®guration showed improved
functional antagonism against the human clone (Table
2). The superiority of the cyclopropyl urea motif in the
functional assays over isosteric analogues such as i-propyl
(rGnRH: IC50=25 nM; rLH: IC50=4500 nM), sug-
gested that the sp2 character of the ring may contribute
to the functional potency. Indeed, phenylurea 12bb was
also active in the rLH assay albeit with 10-fold less potency.
Evaluation of the pyridine regioisomers in the rat assays
revealed a potency preference wherein the order was
2!3!4-aminopyridine (12cc>12dd>12ee); moreover,
12cc was comparable to 12a in functional potency. These
results led us to consider that other basic heteroaromatics
may lead to further improvements. Indeed, incorporation
of pyrimidine and pyrazine heterocycles into the urea
design provided a major breakthrough in functional
potency. For example, pyrazine 12gg and pyrimidine
A concomitant investigation directed at the C(6) position,
revealed that the cyclopropyl-urea analogue (12a) had a
10-fold potency advantage over the nitro variant (1)
(Tables 1 and 2). Naturally, we decided to survey some
of the more potent C(3)-aryl substituents with this urea
modi®cation. As expected, 12b, e, j, and l demonstrated
improved binding activity over the corresponding nitro
analogues. However, the naphthyl analogue 12f did not
bene®t from this modi®cation. A more detailed account
of the C(6) SAR is provided in Table 2.
Table 2.
Entry
C(6)
C(3)
GnRHa
IC50 (nM)
rLH/hPI
IC50 (nM)
Entry
C(6)
C(3)
GnRHa
IC50 (nM)
rLH/hPI
IC50 (nM)
12a
1.1
0.8/2.3b
375/12
682/5
12hh
12ii
0.6
0.4/1.8b
235/na
52/na
12bb
12cc
12dd
12ee
12€
10
3300/nac
463/na
1300/na
na/na
57b
na/897
2500/227
na/225
1920/na
na/23
15
26
7.9
20
14a12
14b
25b
40
1.8/10
2.0b
0.9b
6
na/na
14c13
23a
12gg
0.3/1.1b
56/9
97/5
aGnRH binding in rat pituitary cell are listed ®rst and GnRH binding using human cloned receptors are in bold.
bDenotes compounds bearing the (S)-piperidine con®guration.
cData not available.

J. R. Young et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1723±1727
1727
Table 3.
substituent was hydrophobic in nature and this ulti-
mately led to the identi®cation of the 3,4,5-trimethyl-
phenyl analogue (23h), a potent GnRH antagonist with
subnanomolar binding anity and excellent functional
antagonism.
23
Aryl
hGnRH hPI 23
IC50
(nM)
Aryl
hGnRH hPI
IC50
(nM)
IC50
(nM)
IC50
(nM)
a
b
c
0.9
1.1
1.0
3.9
5.0
8.5
7.7
53
e
f
0.7
0.5
0.7
0.3
6.5
7.1
Acknowledgements
We would like to thank J. Leone, J. Pisano, S. Fabian,
and G. Reynolds for the preparation of several inter-
mediates. We would also like to thank A. Bernick for
mass spectrometry services.
g
h
16.0
2.2
References and Notes
d
1. (a) DeVita, R. J.; Hollings, D. D.; Goulet, M. T.; Wyvratt,
M. J., Jr.; Fisher, M. H.; Lo, J.; Yang, Y. T.; Cheng, K.;
Smith, R. G. Bioorg. Med. Chem. Lett. 1999, 9, 2615. (b)
DeVita, R. J.; Goulet, M. T.; Wyvratt, M. J., Jr.; Fisher, M.
H.; Lo, J.; Yang, Y. T.; Cheng, K.; Smith, R. G. Bioorg. Med.
Chem. Lett. 1999, 9, 2621. (c) Walsh, T. F.; Toupence, R. B.;
Young, J. R.; Huang, S. X.; DeVita, R. J.; Goulet, M. T.;
Wyvratt, Jr., M. J.; Fisher, M. H.; Lo, J.-L.; Cui, J.; Ren, N.;
Yudkovitz, J. B.; Yang, Y. T.; Cheng, K.; Smith, R. G.
Bioorg. Med. Chem. Lett. 2000, 10, 443.
12hh had binding anities <1 nM in the rat and both
analogues were 7 times more potent than 12a in the rLH
assay. Other pyrimidine isomers, such as 12€ and 12ii,
which have no net dipole, were not active in the rLH
assay (at concentrations of 1 mM). Replacement of the
urea nitrogen directly attached to the quinolone core
with a methylene (26) was not potency enhancing,
although removal of the urea nitrogen attached to the
heterocycle led to amides 14a±c, which were equipotent to
or better than their corresponding urea counterparts.
Finally, reversing the amide connectivity of 14c furnished
23a, which was 2 times more potent in the binding assay
(hGnRH) and 4 times more potent in the PI assay than
14c. This pyrimidine-carboxamide was found to be the
superior C(6)-substituent.
2. For the preparation of the indanephenylacetic acid and the
3,4,5-trimethylphenylacetic acid, the syntheses began with the
phenols which were processed according to the following 4-
step sequence.
3. Nodi€, E. A.; Tanabe, K.; Seyfried, C.; Matsura, S.; Kondo,
Y.; Chen, E. H.; Tyagi, M. P. J. Med. Chem. 1971, 14, 921.
4. Racemic N-BOC-2(20-hydroxyethyl)-piperidine 8 was pre-
pared by protection of the free-amine using standard condi-
tions (BOC2O, NEt3, CH2Cl2). The S-enantiomer of 8 was
prepared by Arndt±Eistert homologation of N-BOC (S)-pipe-
colic acid.
5. Kalir, A.; Mualem, R. Synthesis 1987, 514.
6. Katritzky, A. R.; Pilarski, B.; Urogdi, L. Synthesis 1989,
949.
7. Wilcox, C. F. Jr.; Weber, K. A. J. Org. Chem. 1986, 51,
1088.
8. Due to diculties encountered in puri®cation of the initial
aldehyde product, the crude aldehyde was reduced to the
alcohol which was isolated cleanly.
9. Reetz, M. T.; Kyung, S. H.; Hullman, M. Tetrahedron
1986, 42, 2931.
10. Cacchi, S.; Lupi, A. Tetrahedron Lett. 1992, 33, 3939.
11. For detailed information on rat and human based assays
for receptor binding and function, see ref 1c.
12. Prepared by carbomethoxylation of 5-bromopyrimidine
((Ph3P)2PdCl2, NEt3, DMF-MeOH, CO (1 ATM), 90 ꢀC
(44%)) followed by saponi®cation of the ester (NaOH; HCl
(85%)): Godefroi, E. F. J. Org. Chem. 1962, 27, 2264.
13. 4-Carboxypyrimidine was prepared by reaction of 4-methyl-
pyrimidine with SeO2/pyridine (39%): Archer, G. A.; Kalish, R.
I.; Ning, R. Y.; Sluboski, B. C. J. Med. Chem. 1977, 20, 1312.
With the C(6) position optimized, we returned to the
C(3)-aryl position for further re®nement. Compounds
23b±e are hybrids of the best aryl substitution patterns
found in Table 1 with the preferred C(6)-substituent
from Table 2. The SAR in this series correlated with pre-
vious observations, in which the 3,5- and 3,4-dimethyl
(23a and 23e) and dichloro (23b±c) designs were the most
potent and indeed, nearly equivalent to one another
(Table 3). The activities of iodo-methyl analogue 23f were
indistinguishable from 23a. Attempts to improve the
potency of 23e by incorporating the methyl groups in an
indane framework (23g) resulted in a slight loss of func-
tional antagonism. The similar in vitro pro®les of the 23a
and 23e suggested the preparation of the 3,4,5-trimethyl-
phenyl hybrid (23h).3 Gratifyingly, 23h gave improved
binding and functional antagonist activity compared to
the disubstituted analogues.
In conclusion, a thorough study of the SAR at the C(3)-
aryl position indicated the distinct preference for 3,4- or
3,5-dialkyl or di-halo substituted aromatics. Based upon
the initial SAR, we discovered that the optimal aryl