Identification and optimization of novel partial agonists of Neuromedin B receptor using parallel synthesis

Article abstract of DOI:10.1016/j.bmcl.2004.04.045

The design and parallel synthesis of potent, small molecule partial agonists of Neuromedin B receptor based on the 3-amino-2,3,4,9-tetrahydro-1H- carbazole-3-carboxylic acid amide core is described.

Full text of DOI:10.1016/j.bmcl.2004.04.045

Bioorganic & Medicinal Chemistry Letters 14 (2004) 3037–3042
Identification and optimization of novel partial agonists of
Neuromedin B receptor using parallel synthesis
Stephen J. Shuttleworth,a,* Mike E. Lizarzaburu,a Anne Chaib and Peter Cowardb
aDepartment of Chemistry, Tularik Inc., 1120 Veterans Boulevard, South San Francisco, CA 94080, USA
bDepartment of Biology, Tularik Inc., 1120 Veterans Boulevard, South San Francisco, CA 94080, USA
Received 27January 2004; revised 24 March 2004; accepted 15 April 2004
Abstract—The design and parallel synthesis of potent, small molecule partial agonists of Neuromedin B receptor based on the
3-amino-2,3,4,9-tetrahydro-1H-carbazole-3-carboxylic acid amide core is described.
ꢀ 2004 Elsevier Ltd. All rights reserved.
The bombesin family of G-protein coupled receptors
consists of four subtypes:1 Neuromedin B receptor
(NMBR, BB1),2 Gastrin-Releasing Peptide Receptor
NH
(GRPR, BB2),3 BRS-3,4 and BB4.5 Of these, the former
two receptors have received the greatest level of atten-
O
H
N
N
tion in the scientific literature over recent years. NMB
and GRP, the endogenous ligands for NMBR and
GRPR, respectively, were discovered more than 20 years
ago.6;7 They display a seven amino acid C-terminal
sequence homology, and mediate a range of biological
mechanisms via action at their receptors. These include
CNS-related responses such as thermoregulation,8 sati-
ety,9 control of circadian rhythm,10 and peripheral
functions including macrophage activation,11 and gas-
trointestinal hormone release.12
N
H
N
H
O
1
IC50 (125I-NMB) = 0.35µM
Figure 1.
NMBR than had been reported previously in the liter-
ature.15
Examples of peptoid-based Neuromedin B receptor
antagonists were disclosed several years ago;13 however,
there have been no other small molecule antagonists or
agonists of NMBR reported in the literature. As part of
our investigations into the discovery of novel heterocy-
clic ligands of NMBR, we chose to examine the pub-
lished peptoid PD165929, 1 (Fig. 1), as a starting point
for a targeted library.13 Compound 1 was prepared
according to literature precedent,14 and was used in our
biological assays as a reference compound. Interestingly,
in our hands, 1 showed a lower binding affinity for
In an effort to identify a novel chemotype based on the
structural features of 1, we decided to prepare a con-
formationally-constrained derivative in the form of a
3-amino-2,3,4,9-tetrahydro-1H-carbazole-3-carboxylic
acid amide, 2. The synthesis of this compound was
accomplished in four steps––incorporating a Fischer
cyclization at the penultimate stage––starting from
8-amino-1,4-dioxa-spiro[4.5]decane-8-carboxylic acid,
and proceeded in good overall yield (Scheme 1).16
We observed that 2 displayed a marginal increase in
binding affinity for NMBR over 1, but interestingly was
also a highly selective, weak partial agonist, displaying
no activity against a panel of GPCR targets encom-
passing members of the lipid and neuropeptide receptor
families. Further, 2 showed no cytotoxicity when incu-
bated with HeLa cells for 72 h (Fig. 2).17
Keywords: Neuromedin B receptor; Partial agonist; Lead optimization;
Parallel synthesis.
* Corresponding author.
0960-894X/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.04.045

3038
S. J. Shuttleworth et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3037–3042
Scheme 1.
Table 1.
R
O
HN
O
HN
HN
HN
O
H
N
N
H
N
N
N
H
O
N
H
Compound #
R
IC50 (
125I-NMB), lM
2
IC50 (125I-NMB) = 0.324µM
2
0.32
0.62
Figure 2.
Encouraged by the preliminary data for 2, we con-
structed a targeted library based upon this tetrahydro-
carbazole core. The synthetic protocol used was
identical to that illustrated in Scheme 1, and all of the
analogues were prepared using our parallel synthesis
modules and were purified by automated HPLC.18 The
library was assembled in four stages. Initially we con-
ducted a thorough examination of the exocyclic nitrogen
substitution pattern, and quickly concluded that a
substituted phenylurea was essential for activity, with
the 2,6-diisopropylphenyl urea being the optimal motif.
In addition, we observed that the 4-nitrophenylurea
derivative, 4, also displayed sub-micromolar potency
against the receptor; this nitrophenylurea motif was
present in several of the most potent peptoid ligands
based on compound 1 previously reported by Ashwood
et al.19 that were subsequently examined as part of a
mutagenesis study to elucidate the role of the nitro
NO2
3
NO2
CF3
4
5
0.80
1.11
CO2Et
6
7
1.85
2.27
CO2Et
substituent in hydrogen bonding to the receptor.20
A
representative set of data summarizing our results from
this exercise is shown in Table 1.
O

S. J. Shuttleworth et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3037–3042
3039
Table 1 (continued)
determine the requirement for the N-(1-pyridin-2-yl-
cyclohexylmethyl)-acetamide, a residue present in the
most potent analogues in both the peptoid and tetra-
hydrocarbazole series. In this study, we retained the urea
substituents as both 2,6-diisopropylphenyl and 4-nitro-
phenyl, and a range of amide derivatives of 2 were
prepared. These encompassed examples representing
subtle modifications to the lead structure, together with
more random, structurally diverse analogues, prepared
in an attempt to incorporate additional recognition
elements and to reduce molecular weight. The majority
of the analogues showed significantly diminished affinity
for NMBR however, illustrating the importance of the
pyridin-2-yl-cyclohexylmethyl motif in binding to the
receptor; indeed the only derivatives that displayed
an appreciable level of activity were the direct gem-
dimethyl and cyclopropyl analogues, 14 and 15, which
adopt similar conformations to the parental compound
(Table 2).
Compound #
R
IC50 (
125I-NMB), lM
OCF3
8
3.04
3.92
5.49
9
10
11
8.72
OMe
12
13
Et
H
>10
>10
The third stage centered on our examining a broad
range of substitution patterns within the indole ring.
Table 3 summarizes key data obtained for the 2,6-
diisopropylphenylurea derivatives prepared in this
exercise. A general trend of improved affinity with
In the second phase of library construction, we focused
on modifications to the carboxy terminus in order to
Table 2.
HN
R1
HN
O
H
N
R2
O
N
H
Compound #
R1
R2
IC50 (
125I-NMB), lM
N
N
N
14
2,6-Di-isopropyl
3.35
15
16
17
2,6-Di-isopropyl
4.15
4-NO2
10
N
N
4-NO2
>10
>10
18
19
4-NO2
S
O
O
Cl
2,6-Di-isopropyl
>10

3040
S. J. Shuttleworth et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3037–3042
Table 3.
HN
O
HN
H
N
N
R
O
N
H
Compound #
R
IC50 (
125I-NMB), lM
20
21
22
23
24
25
26
27
28
29
5-OMe
5-Me
0.25
0.66
0.70
0.79
1.01
1.89
2.14
3.66
4,6-Di-Me
4,6-Di-Me
4,7-Di-F
5-OCF3
7-Br
5-CF3
5-NO2
5-CO2H
>10
>10
Figure 3.
electron-rich indole systems was observed, with com-
pound 20 being the most potent compound identified.
The final step of our study involved resolving the two
enantiomers of 20, which was accomplished by using
preparative chiral HPLC.21 We observed that the (+)-
enantiomer, 30, was the more potent of the two.22
Further, this compound was very selective for NMBR
over structurally-related GPCRs, and did not display
any cytotoxic effects in HeLa cells (Scheme 2).
As mentioned above, compound 2 showed weak partial
agonist activity in an inositol phosphate production
SPA assay.23 We subsequently observed that the level of
partial agonism in this tetrahydrocarbazole series
increased with improved binding affinity. Figure 3
illustrates the effects of the dose response curves of
Neuromedin B upon treatment with increasing concen-
trations of 30.
Figure 4.
Figure 4 depicts the partial agonist effects of the key
compounds prepared in this study, with levels of IP
accumulation increasing with improved potency, as
determined in the radioligand competition assay.
In summary, a series of novel and potent NMBR partial
agonists has been identified. The compounds were de-
HN
30
HN
O
(+)-enantiomer
Chiral HPLC
IC50 (125I-NMB) = 29nM
IC50 HeLa Cytotoxicity >50µM
O
H
N
N
N
H
O
31
(-)-enantiomer
Scheme 2.

S. J. Shuttleworth et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3037–3042
3041
removed using a Genevac centrifugal evaporator, and the
residue partitioned between brine and dichloromethane.
The organic phase containing B was isolated, and treated
with piperidine (20% of overall volume); the reaction
mixture was agitated for 1 h at room temperature, and was
then concentrated to dryness. The residue, C, was then
treated with 1 M HCl (aq), and the solution washed with
dichloromethane. The aqueous phase was isolated and
phenylhydrazine (1.0 equiv) was added, together with a
few drops of concentrated HCl, and the reaction mixture
was heated to 85 ꢁC for 2 h. The solution was cooled,
neutralized with saturated sodium bicarbonate solution,
and the product, D, was extracted with dichloromethane.
The overall yields for the synthesis of D typically exceeded
60%. For the final step, stock solutions of the primary
amine D were treated with equal mole quantities of
isocyanate in N,N-dimethylformamide, in the presence of
signed on a conformationally constrained derivative of a
known peptoid-based NMBR antagonist, and were
produced in a matter of weeks using a rapid solution-
based parallel synthesis approach.
Acknowledgements
We would like to thank Jinqian Liu, Richard Connors,
Malgorzata Waska, Shichang Miao, Wayne Inman, Ji
Ma, Adam Park, and Stephen Young for their contri-
butions to this work, and to Juan Jaen and Tim Hoey
for their comments.
€
Hunig’s base. After 16 h stirring at room temperature,
References and notes
conversions in excess of 90% to the desired ureas E were
determined by HPLC at 220 and 254 nm using an Agilent
1100 LC/MSD VL ESI system. The products were then
purified directly using automated preparative HPLC (see
Ref. 19). Note that for several examples, step 3 involved
the use of 1,4,-dioxane as the solvent to circumvent
problems with poor aqueous solubility of a number of
phenylhydrazines used in the library synthesis.
1. Moody, T.; Jensen, R. T. Drug. Fut. 1998, 23, 1305–1315.
2. Wada, E.; Way, J.; Shapira, H.; Kusano, K.; Lebacq-
Verheyden, A. M.; Coy, D.; Jensen, R.; Battey, J. Neuron
1991, 6, 421–423.
3. Spindel, E. R.; Giladi, E.; Brehm, P.; Goodman, R. H.;
Segerson, T. P. Mol. Endocrinol. 1990, 4, 1956–1963.
4. Fathi, J.; Corjay, M. H.; Shapira, H.; Wada, E.; Benya,
R.; Jensen, R.; Viallet, J.; Sausville, E. A.; Battey, J. F.
J. Biol. Chem. 1993, 268, 5979–5984.
5. Nagalla, S. R.; Barry, B. J.; Creswick, K. C.; Eden, P.;
Taylor, J. T.; Spindel, E. R. Proc. Natl. Acad. Sci. U.S.A.
1995, 92, 6205–6209.
6. Minamino, N.; Kangawa, M.; Matsuo, H. Biochem.
Biophys. Res. Commun. 1983, 114, 541–548.
7. McDonald, T. J.; Jornvall, J.; Nilsson, G. Biochem.
Biophys. Res. Commun. 1979, 90, 227–233.
8. Brown, M. R.; Carver, K.; Fisher, L. A. Ann. N.Y. Acad.
Sci. 1988, 547, 174–182.
9. McCoy, J. G.; Avery, D. D. Peptides 1990, 11, 595–607.
10. Albers, H. E.; Liou, S. Y.; Stopa, E. G.; Zoeller, R. T.
J. Neurosci. 1991, 11, 846–851.
11. DelaFuente, M.; DelRio, M.; Hernanz, A. J. Neuroimmu-
nol. 1993, 48, 143–150.
17. Typical protocol for cytotoxicity assay: HeLa cells were
seeded at 5k/well in a 96 well plate (6 plates-for triplicate 0
and 72 h readings). A three-fold dilution series of each
compound was generated, starting at 10 mM. The solution
was diluted 8 times to prepare 9 concentrations ranging
from 10 mM to 150 nM. Compound was added to cells (1–
100 lL total volume), and the final concentration of
compound was 100 lM to 15 nM. Zero and 72 h time
points were recorded as follows: (a) 10 lL Alamar Blue
reagent was added to the wells, and the samples were
incubated at 37 ꢁC for 3 h; (b) fluorescence intensity was
then recorded on an LJL Analyst. The relative growth was
compared to a DMSO control well for each compound.
18. Synthesis was accomplished using 24-position ‘Green-
house’ and 96-position METZ Heater-Shaker modules
purchased from Radleys Ltd. (see www.radleys.com).
Purification was conducted using a Parallex Flexe HPLC
System purchased from Biotage Inc. (see www.Bio-
tage.com).
12. Tache, Y.; Melchiorri, P.; Negri, L. Ann. N.Y. Acad. Sci.
1988, 517, 1–541.
13. Eden, J. M.; Hall, M. D.; Higginbottom, M.; Horwell, D. C.;
Howson, W.; Hughes, J.; Jordan, R. E.; Lewthwaite, R. A.;
Martin, K.; McKnight, A. T.; O’Toole, J. C.; Pinnock, R. D.;
Pritchard, M. C.; Suman-Chauban, N.; Williams, S. C.
Bioorg. Med. Chem. Lett. 1996, 6, 2617–2622.
19. Ashwood, V.; Brownhill, V.; Higginbottom, M.; Horwell,
D. C.; Hughes, J.; Lewthwaite, R. A.; McKnight, A. T.;
Pinnock, R. D.; Pritchard, M. C.; Suma-Chauhan, N.;
Webb, C.; Williams, S. C. Bioorg. Med. Chem. Lett. 1998,
8, 2589–2594.
14. Horwell, D. C.; Pritchard, M. C. PCT/US97/13871;
WO9807718.
20. Tokita, K.; Hocart, S. J.; Katsuno, T.; Mantey, S.; Coy,
D. H.; Jensen, R. T. J. Biol. Chem. 2001, 276, 495–504.
21. Resolution of 20 was accomplished using a Waters 600
Tower system equipped with a Chiralpak AD column
15. Protocol for whole cell radioligand binding assay: Human
embryonic kidney 293 (HEK293) cells were stably trans-
fected with expression plasmids encoding hNMBR. Com-
petition binding assays were performed for 1 h at room
temperature in the presence of 150 pM 125I-[D-Tyr0]NMB
(2200 Ci/mmol, Perkin–Elmer Life Sciences) and 1 · 106
transfected cells in Dulbeco’s Modified Eagle Media
(Mediatech, Inc.) in 100 lL in 96 well plates. Bound
ligand was separated from unbound ligand by filtration
using a Filtermate (Packard) and total counts bound
determined on a TopCount NTX reader (Packard).
16. Typical synthetic procedure for 2 and analogues thereof: A
solution of N-Fmoc-amino-4-(ethylene ketal)cyclohexyl-
carboxylic acid, A (1.0 wt) in N,N-dimethylformamide was
treated with amine (1.0 equiv), HBTU (1.2 equiv) and
(Daicel Chemical Industries
#
AD00CJ-DG008,
25 cm · 2 cm). Protocol: a solution of 20 (8.0 mg in
500 mL isopropyl alcohol) was injected and an isocratic
gradient of 8% isopropyl alcohol in hexanes at 20 mL/min
was maintained. The fractions corresponding to the two
enantiomers (30 ¼ 35 min; 31 ¼ 27min) were isolated,
and were concentrated in vacuo to afford 30 (1.96 mg) and
31 (2.23 mg). LCMS analysis17 confirmed the desired
molecular weights (M + 1 ¼ 636.5) of both products, and
purities were determined by analytical HPLC using a
Hewlett Packard series 1050 Tower system with a Daicel
Chemical Industries Chiralpak AD-H column (# AD-
H0CE-DC041, 25 cm · 0.4 cm), employing an isocratic
gradient of 8% IPA in hexanes at 1 mL/min. Both 30
and 31 were determined to have enantiomeric excesses
€
Hunig’s base (1.2 equiv), and the reaction mixture was
agitated at room temperature for 16 h. The solvent was

3042
S. J. Shuttleworth et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3037–3042
greater than 99%. Optical rotations were measured using a
w/o L-Glutamine, I-Inositol (US Biological) supplemented
27
D
Jasco P-1020 polarimeter (30: ½aŠ þ57:8 (c 0.2, CH3OH).
with 2 g/L sodium bicarbonate, 25 mM Hepes, 2% gluta-
mine, 10% dialyzed FBS (Gibco), and 1 lCi/mL myo-
[3H]Inositol (82.0 Ci/mmol, Amersham Pharmacia Biotech)
overnight at 37 ꢁC. Test compounds were then added in
inositol-free DMEM containing 0.3% BSA (Sigma) and
10 mM LiCl (Sigma) for 60 min at 37 ꢁC. The cells were
lysed with 20 mM formic acid for 2 h at 4 ꢁC and added to
RNA-binding Ysi scintillation proximity assay beads
(Amersham Pharmacia Biotech) for 30 min. Plates were
stored overnight at room temperature in the dark and read
the next day on a TopCount NTX.
27
D
31: ½aŠ ꢀ42:8 (c 0.2, CH3OH)).
22. Analytical data for 30: dH (400 MHz, DMSO-d6) 10.90
(1H, s), 8.62 (1H, s), 7.80 (2H, m), 7.55–7.40 (4H, m),
7.20–6.95 (6H, m), 6.20 (1H, s), 3.73 (3H, s), 3.45–3.2 (5H,
m), 2.90–2.70 (5H, m), 2.20–2.10 (2H, m), 2.00 (1H, m),
1.65–1.20 (7H, m), 1.20–1.00 (12H, m). M + 1 found 636.1;
C39H49N5O3 requires 635.38.
23. Protocol for the inositol phosphate (IP) accumulation
assay: hNMBR-transfected HEK293 cells (2.5 · 104 were
incubated in 96 well plates in DMEM Eagle High Glucose,